Section XII. Hormones and Hormone Antagonists Chapter 56.

Pituitary Hormones and Their Hypothalamic Releasing Factors

Overview This chapter covers the diagnostic and therapeutic uses of some of the pituitary hormones including growth hormone (GH), prolactin, luteinizing hormone (LH), follicle-stimulating hormone (FSH), and oxytocinas well as the therapeutic approaches to conditions of excess secretion of GH and prolactin. Also discussed are the clinical and diagnostic uses of hypothalamic factors that regulate the secretion of pituitary hormones, including growth hormone-releasing hormone (GHRH), somatostatin, and gonadotropin-releasing hormone (GnRH). FSH, LH, and GnRH also are discussed in Chapters 58: Estrogens and Progestins and 59: Androgens. Considered elsewhere are corticotropin and corticotropin-releasing hormone (Chapter 60: Adrenocorticotropic Hormone; Adrenocortical Steroids and Their Synthetic Analogs; Inhibitors of the Synthesis and Actions of Adrenocortical Hormones) and thyrotropin and thyrotropin releasing hormone (Chapter 57: Thyroid and Antithyroid Drugs). Pituitary Hormones and Their Hypothalamic Releasing Factors: Introduction The peptide hormones of the anterior pituitary are essential for the regulation of growth and development, reproduction, responses to stress, and intermediary metabolism. Their synthesis and secretion are controlled by hypothalamic hormones and by hormones from the peripheral endocrine organs. A large number of disease states as well as a diverse group of drugs also affect their secretion. The complex interactions among the hypothalamus, pituitary, and peripheral endocrine glands provide elegant examples of integrated feedback regulation. Clinically, an improved understanding of the mechanisms that underlie these interactions provides the rationale for diagnosing and treating endocrine disorders and for predicting certain side effects of drugs that affect the endocrine system. Moreover, the elucidation of the structures of the anterior pituitary hormones and hypothalamic releasing hormones together with advances in protein chemistry have made it possible to produce synthetic peptide agonists and antagonists that have important diagnostic and therapeutic applications. Ten anterior pituitary hormones have been identified in vertebrates; these can be classified into three different groups based on their structural features (Table 561). Growth hormone (GH) and prolactin belong to the somatotropic family of hormones, which in human beings also includes placental lactogen. The glycoprotein hormonesthyrotropin (TSH), luteinizing hormone (LH), and follicle-stimulating hormone (FSH)share a common -subunit but have different -subunits that determine their distinct biological activities. In human beings, the glycoprotein hormone family also includes placental chorionic gonadotropin (CG). Corticotropin (adrenocorticotrophic hormone; ACTH), the two melanocyte-stimulating hormones ( - and -MSH), and the two lipotropins represent a family of hormones derived from proopiomelanocortin by proteolytic processing. Except for -MSH and the lipotropins, these pituitary hormones all play significant roles in human health and disease. The synthesis and release of anterior pituitary hormones are influenced by the central nervous system. Their secretion is positively regulated by a group of polypeptides referred to as hypothalamic releasing hormones. These hormones are released from hypothalamic neurons in the

region of the median eminence, and they reach the anterior pituitary through the hypothalamicadenohypophyseal portal system. The hypothalamic releasing hormones include growth hormone releasing hormone (GHRH), gonadotropin-releasing hormone (GnRH), thyrotropin-releasing hormone (TRH), and corticotropin-releasing hormone (CRH). Somatostatin, another hypothalamic peptide, negatively regulates the pituitary secretion of growth hormone and thyrotropin. Finally, the catecholamine dopamine inhibits the secretion of prolactin by lactotropes. As discussed further in Chapter 30: Vasopressin and Other Agents Affecting the Renal Conservation of Water, the posterior pituitary gland, also known as the neurohypophysis, contains nerve axons arising from distinct populations of neurons in the supraoptic and paraventicular nuclei that synthesize either arginine vasopressin or oxytocin. Oxytocin plays important roles in labor and parturition and in milk let-down, as discussed below. Growth Hormone The gene encoding human growth hormone (GH) resides on the long arm of chromosome 17, which also contains four related genes: three different variants of placental lactogen and a GH variant expressed in the syncytiotrophoblast (chorionic somatotropin). Secreted GH is a heterogeneous mixture of peptides that can be distinguished on the basis of size or charge; the principal 22,000dalton form is a single polypeptide chain of 191 amino acids that has two disulfide bonds and is not glycosylated. Alternative splicing deletes residues 32 to 46 of the larger form to produce a smaller form ( 20,000 daltons) with equal bioactivity that makes up 5% to 10% of circulating GH. Additional GH species are found in serum, but their physiological significance is unclear. Approximately 45% of the 22,000-dalton and 25% of the 20,000-dalton GH in circulation are bound by a binding protein that contains the extracellular domain of the GH receptor (see below). This GH-binding protein may serve as a reservoir of growth hormone, as the biological half-life of GH complexed to it is approximately 10 times that of unbound GH. Alternatively, the binding protein may decrease GH bioactivity by preventing it from binding to its receptor in target tissues. Regulation of Growth Hormone Secretion Growth hormone, the most abundant anterior pituitary hormone, is synthesized and secreted by somatotropes. These cells account for about 50% of hormone-secreting cells of the anterior pituitary and cluster at its lateral wings. Daily GH secretion varies throughout life; secretion is high in children, reaches maximal levels during adolescence, and then decreases in an age-related manner in adulthood. GH secretion occurs in discrete but irregular pulses. Between these pulses, circulating GH falls to levels that are undetectable with current assays. The amplitude of secretory pulses is maximal at night, and the most consistent period of GH secretion is shortly after the onset of deep sleep. Because of this episodic release, random measurements of GH are of little value in the diagnosis of growth hormone deficiency, and provocative tests are required (see below). The regulation of GH secretion is illustrated in Figure 561. GHRH, produced by hypothalamic neurons found predominantly in the arcuate nucleus, stimulates growth hormone secretion by binding to a specific G proteincoupled receptor on somatotropes, elevating both intracellular cyclic AMP and Ca2+ concentrations. Somatostatin, which is synthesized by more widely distributed neurons as well as by neuroendocrine cells in the gastrointestinal tract and pancreas, inhibits growth hormone secretion. Somatostatin is synthesized from a 92amino acid precursor and processed by proteolytic cleavage to generate two predominant formssomatostatin-14 and somatostatin-28. The somatostatins exert their effects by binding to and activating a family of G proteincoupled receptors. The consequences of receptor activation include inhibition of cyclic AMP accumulation, activation of K+ channels, and activation of tyrosine phosphatase. Five somatostatin receptor

subtypes have been identified, each of which binds somatostatin with nanomolar affinity; whereas receptor types 1 to 4 (abbreviated sst1-4 or SSTR1-4) bind the two somatostatins with approximately equal affinity, type 5 (sst5, SSTR5) has a 10- to 15-fold greater selectivity for somatostatin-28 (Patel, 1999). It appears that the SSTR2 and SSTR5 receptors are most important for regulation of GH secretion. There is evidence supporting both direct effects of somatostatin on somatotropes and indirect effects mediated via GHRH neurons in the arcuate nucleus. As discussed below, somatostatin analogs play an important role in the therapy of syndromes of GH excess such as acromegaly. Figure 561. Growth Hormone Secretion and Actions. Two hypothalamic factors, growth hormonereleasing hormone (GHRH) and somatostatin (SST) stimulate or inhibit the release of growth hormone (GH) from the pituitary, respectively. Insulin-like growth factor 1 (IGF-1), a product of GH action on peripheral tissues, causes negative feedback inhibition of GH release by acting at the hypothalamus and the pituitary. The actions of GH can be direct or indirect and mediated by IGF-1. See text for discussion of the other agents that modulate GH secretion.

Appreciation of a third component of regulation of GH secretion arose from studies of GH secretogogues (Smith et al., 1999). The finding that peptide derivatives of Leu- and Metenkephalins stimulate growth hormone release has led to the development of additional peptide and nonpeptide GH secretogogues that stimulate GH secretion via a G proteincoupled receptor distinct from the GHRH receptor (Howard et al., 1996). This GH-secretogogue receptor is expressed on somatotropes as well as on GHRH neurons in the arcuate nucleus, suggesting that GH secretogogues stimulate GH release both by direct actions on the pituitary and by indirect effects on

GHRH neurons. Intriguingly, both GH and somatostatin inhibit the activation of these neurons. This inhibition by GH indicates a direct feedback action of GH, while the inhibition by somatostatin suggests that an important component of the inhibition of GH secretion by somatostatin is exerted in the hypothalamus rather than in the pituitary. The clinical utility of GH secretogogues in patients with growth hormone deficiency is an area of active investigation, as is the putative endogenous ligand that activates the GH-secretogogue receptor. Although their specific sites of action are not fully understood, several neurotransmitters, drugs, metabolites, and other stimuli also affect GH secretion by modulating the release of GHRH and/or somatostatin. Dopamine, 5-hydroxytryptamine, and 2-adrenergic receptor agonists stimulate GH release, whereas -adrenergic receptor agonists, free fatty acids, and insulin-like growth factor-1 (IGF-1, see below) and GH itself inhibit release. Hypoglycemia stimulates growth hormone release, as do exercise, stress, emotional excitement, and ingestion of protein-rich meals. In contrast, administration of glucose in an oral glucose-tolerance test suppresses GH secretion in normal subjects. These observations form the basis for provocative tests to assess the ability of the pituitary to secrete GH. Provocative stimuli include arginine, glucagon, insulin-induced hypoglycemia, clonidine, and the dopamine precursor levodopa; these agents all increase circulating GH levels in normal subjects within 45 to 90 minutes. At present, insulin-induced hypoglycemia is the test advocated by the Growth Hormone Research Society (Anonymous, 1998), whereas the United States Food and Drug Administration (FDA) requires two independent tests of GH deficiency to establish the diagnosis. When excess GH secretion is suspected (see below), the failure of an oral glucose load to suppress GH is diagnostically useful. Finally, as described below, GH secretion in response to GHRH can be used to distinguish pituitary disease from hypothalamic disease. Molecular and Cellular Bases of Growth Hormone Action All of the effects of GH result from its interactions with the GH receptor, as evidenced by the severe phenotype of rare patients with homozygous mutations of the GH-receptor gene (the Laron syndrome of GH-resistant dwarfism). The GH receptor is a widely distributed cell-surface receptor that belongs to the cytokine receptor superfamily and shares structural similarity with the prolactin receptor, the erythropoietin receptor, and several of the interleukin receptors (Finidori et al., 2000). Like other members of the cytokine receptor family, the GH receptor contains an extracellular domain that binds GH, a single membrane-spanning region, and an intracellular domain that mediates signal transduction. Receptor activation results from the binding of a single GH molecule to two identical receptor molecules (de Vos et al., 1992). The net result is the formation of a ligandoccupied receptor dimer that presumably brings the intracellular domains of the receptor into close proximity, thereby activating cytosolic components critical for cell signaling. As determined from cDNA cloning and sequencing (Leung et al., 1987), the mature human GH receptor contains 620 amino acids, 260 of which are extracellular and 350 of which are cytoplasmic. The formation of the GH-GH receptor ternary complex is initiated by a high-affinity interaction of GH with a receptor monomer, exposing a second site of lower affinity on GH that recruits a second receptor molecule to the complex. Interestingly, GH analogs have been engineered with a disrupted second receptor-binding site; these analogs cannot induce receptor dimerization. One such analog, pegvisomant, behaves as a GH antagonist and has shown promise in the treatment of acromegaly (Trainer et al., 2000; see below). In addition to the full-length GH receptor, truncated forms of the receptor also have been described. A circulating form of the receptor, called GH-binding protein, is formed by proteolytic cleavage of

the extracellular domain of the receptor from its transmembrane segment. GH-binding protein has been reported to delay the clearance of circulating GH and increase its activity in vitro, but its biological role remains unknown. Truncated, membrane-anchored forms of the receptor also have been described. Again, the physiological roles of these proteins, which apparently result from alternative splicing events and constitute a small fraction of the receptor population, are unknown, although they inhibit GH action in cultured cell models. Truncated forms of the GH receptor also have been found in one kindred with growth-hormone insensitivity and short stature (Ayling et al., 1997). These patients are heterozygous for the receptor mutation, suggesting that the truncated receptors behave as dominant negative inhibitors of GH signaling. The ligand-occupied receptor dimer does not have inherent tyrosine kinase activity, but it does provide docking sites for two molecules of Jak2, a cytoplasmic tyrosine kinase of the Janus kinase family. The juxtaposition of two Jak2 molecules leads to trans-phosphorylation and autoactivation of Jak2, with consequent tyrosine phosphorylation of cytoplasmic proteins that mediate downstream signaling events. These include Stat proteins (signal transducers and activators of transcription), Shc (an adapter protein that regulates the Ras/MAP kinase signaling pathway), and IRS-1 and IRS2 (insulin-receptor substrate proteins that activate the phosphatidyl inositol-3 kinase regulatory pathway) (see Figure 562). Figure 562. Mechanism of Growth Hormone Action. The binding of GH to two molecules of the growth hormone receptor (GHR) induces dimerization of JAK2 and its autophosphorylation. JAK2 then phosphorylates cytoplasmic proteins that activate downstream signaling pathways (PI3 kinase, ras, raf, MAPK) that ultimately affect gene expression. The arrows indicate the presumed order of activation in the signaling pathway; the figure does not reflect the localization of the intracellular molecules, which presumably exist in multicomponent signaling complexes. JAK2, janus kinase 2; IRS1, insulin receptor substrate 1; PI3 kinase, phosphatidyl inositol-3 kinase; STAT, signal transducer and activator of transcription; SOS, product of the son of sevenless gene; MAPK, mitogenactivated protein kinase; MEK, MAPK kinase; SHC and Grb2, adapter proteins.

Although GH acts directly on adipocytes to increase lipolysis and on hepatocytes to stimulate gluconeogenesis, its anabolic and growth-promoting effects are mediated indirectly through the induction of insulin-like growth factors (IGFs). There are two members of the IGF family: IGF-1 and IGF-2. IGF-1 is more dependent on GH and is a more potent growth factor postnatally; thus, IGF-1 appears to be the principal mediator of GH action. Most circulating IGF-1 is made in the liver, although IGF-1 produced locally in many tissues also may exert paracrine or autocrine effects on cell growth. Circulating IGF-1 is associated with a family of binding proteins that serve as transport proteins and also may mediate certain aspects of IGF-1 signaling. The essential role of IGF-1 in GH signaling is evidenced by a patient with loss-of-function mutations in both alleles of the IGF1 gene whose severe intrauterine and postnatal growth retardation was unresponsive to GH but responsive to recombinant human IGF-1 (Camacho-Hubner, et al., 1999). Following its synthesis and release, IGF-1 interacts with receptors on the cell surface that mediate its biological activities. The type 1 IGF receptor is closely related to the insulin receptor, consisting of a heterotetramer with intrinsic tyrosine kinase activity. This receptor is present in essentially all tissues and binds IGF-1 and IGF-2 with high affinity; insulin also can activate the type 1 IGF receptor, but with an affinity approximately 100 times less than that of the IGFs. The type 2 IGF receptor encodes a protein that is located predominantly on intracellular membranes and is identical to the mannose-6-phosphate receptor that participates in intracellular targeting of acid hydrolases and other mannose-containing glycoproteins to lysosomes. This receptor apparently is activated specifically by IGF-2. The signal transduction pathway for the insulin receptor is described in detail in Chapter 61: Insulin, Oral Hypoglycemic Agents, and the Pharmacology of the Endocrine Pancreas. Syndromes of Growth Hormone Deficiency GH deficiency in children is a well-accepted cause of short stature, and replacement therapy has been used for more than 30 years to treat children with severe GH deficiency. More recently, GH deficiency in adults has been associated with a defined endocrinopathy that includes increased mortality from cardiovascular causes, probably secondary to deleterious changes in fat distribution and increases in circulating lipids; decreased muscle mass and exercise capacity; and impaired psychosocial function. With the ready availability of recombinant human GH, attention has shifted to the proper role of GH therapy in GH-deficient adults. While this is an area of current debate, the emerging consensus is that at least the most severely affected GH-deficient adults will benefit from GH replacement therapy. GH therapy also is approved by the FDA for AIDS-associated wasting, and its use has resulted in some benefit in patients with this condition. Based on controlled clinical trials showing increased mortality, GH should not be used in patients with acute critical illness due to complications following open heart or abdominal surgery, multiple accidental trauma, or acute respiratory failure. GH also should not be used in patients who have any evidence of neoplasia, and antitumor therapy should be completed prior to initiation of GH therapy. Diagnosis of Growth Hormone Deficiency Clinically, children with GH deficiency present with short stature and a low age-adjusted growth velocity. Most commonly, these children have an isolated deficiency of GH without other documented pathology (i.e., idiopathic, isolated GH deficiency) and are presumed to have a hypothalamic defect. Random sampling of serum GH is insufficient to diagnose GH deficiency; provocative tests are required. After excluding other causes of poor growth, the diagnosis of GH deficiency should be entertained in patients with height 2 to 2.5 standard deviations below normal,

delayed bone age, a growth velocity below the 25th percentile, and a predicted adult height substantially below the mean parental height (Vance and Mauras, 1999). In this setting, a serum GH level of less than 10 g/liter following provocative testing (e.g., insulin-induced hypoglycemia, arginine, levodopa, or glucagon) indicates GH deficiency, with a stimulated value of less than 5 g/liter reflecting severe deficiency. More than 90% of adult patients with GH deficiency have overt pituitary disease due to a functioning or nonfunctioning pituitary adenoma or resulting from surgery or radiotherapy for a pituitary mass. Almost all patients with multiple deficits in other pituitary hormones also will have deficient GH secretion. According to criteria established by the FDA, a normal response to provocative stimuli is an increase in GH to serum levels 5 g/liter by radioimmunoassay or 2.5 g/liter by immunoradiometric or immunochemiluminescent assay. In contrast, the Growth Hormone Research Society has recommended diagnosis based on a stimulated GH serum level of less than 3 g/liter during insulin-induced hypoglycemia (Anonymous, 1998). Treatment of Growth Hormone Deficiency The action of GH is highly species-specific; human beings do not respond to GH from nonprimate species. Therefore, GH for therapeutic use formerly was purified from human cadaver pituitaries in very limited quantities. The production of human GH by recombinant DNA technology not only increased availability of the hormone but also alleviated concerns about Creutzfeldt-Jakob disease associated with use of the hormone purified from cadaver pituitaries. A number of recombinant preparations of human GH are approved for use in many countries. By convention, somatropin refers to GH preparations whose sequence matches that of native GH (SEROSTIM , GENOTROPIN, HUMATROPE , NUTROPIN, SAIZEN), while somatrem refers to a derivative of GH with an additional methionine at the amino terminus (PROTROPIN). Although there are subtle differences in the sources and structures of these preparations, all have similar biological actions and potencies. They typically are administered subcutaneously in the evening; although the circulating half-life of GH is only 20 minutes, its biological half-life is in the range of 9 to 17 hours, and once-daily administration is sufficient. Newer formulations are supplied in prefilled syringes, which may be more convenient for the patient, as the GH does not need refrigeration and the diluent causes less irritation at the injection site. An encapsulated form of somatropin that is injected intramuscularly once or twice per month (NUTROPIN DEPOT) has been approved by the FDA. The relative advantages of any specific formulations over others in clinical use have not been definitively established. In addition to GH, sermorelin acetate (GEREF ), a synthetic form of human GHRH, has received FDA approval for treatment of idiopathic GH deficiency. Sermorelin is a peptide of 29 amino acids that corresponds in sequence to the first 29 amino acids of human GHRH (a 44amino acid peptide) and has full biological activity. Sermorelin generally is well tolerated and is less expensive than somatropin, but at recommended doses it has been less effective than GH in clinical trials. Moreover, this agent will not work in patients whose GH deficiency results from defects in the anterior pituitary (Anonymous, 1999). Therefore, a GH response (>2 g/liter) to a test dose of sermorelin should be documented prior to initiating therapy (30 g/kg per day, given subcutaneously), and the patients must be monitored frequently to ascertain continued growth on therapy. Sermorelin also has been employed diagnostically to distinguish between pituitary and hypothalamic disease; its clinical utility in this setting is not fully established. GH is widely used for replacement therapy in GH-deficient children, whether the deficiency is congenital or acquired. It also is FDA-approved for use in children with chronic renal insufficiency

(although not proven to increase adult height) and for patients with Turner's syndrome (improving adult height significantly). Recommended doses vary with indication and product, but typically a dose of 20 to 40 g/kg is administered subcutaneously either daily or 6 times per week; higher daily doses (e.g., 50 g/kg) are employed for patients with Turner's syndrome, who have partial GH resistance. Initial response and compliance can be monitored with serum IGF-1 levels, while longterm response is monitored by close evaluation of height. Although the most pronounced increase in growth velocity occurs within the first two years of therapy, GH is continued until growth ceases. In view of the increased appreciation of the effects of GH on bone density and the effects of GH deficiency in adults, it seems reasonable to continue therapy into adulthood. However, many patients who clearly were GH deficient in childhoodespecially those with idiopathic, isolated GH deficiencyrespond normally to provocative tests at the cessation of therapy. Thus, it is essential to confirm GH deficiency after optimal growth has been achieved so as to identify patients who will benefit from continuing GH treatment. In adults, previously recommended doses of GH now are viewed as excessive, leading to both an elevated IGF-1 concentration and a greater risk of side effects. The FDA recommends a starting dose of 3 to 4 g/kg, given once daily by subcutaneous injection, with a maximum dose of 25 g/kg in patients 35 years old and 12.5 g/kg in older patients. The Growth Hormone Research Society recommends a starting dose of 150 to 300 g/day regardless of body weight (Anonymous, 1998). Clinical response is monitored by serum IGF-1, which should be restored to the midnormal range adjusted for age and sex. Either an elevated serum IGF-1 or persistent side effects are grounds for decreasing the dose; conversely, the dose can be increased if serum IGF-1 has not reached the normal range after two months of GH therapy. In the setting of AIDS-related wasting, considerably higher doses (e.g., 100 g/kg) have been used in clinical trials. As noted above, a subset of children with growth impairment has elevated GH levels and GH resistance, most frequently secondary to mutations in the GH receptor. These patients can be treated effectively with recombinant human IGF-1 (IGEF), which is administered subcutaneously either once or twice daily in doses ranging from 40 to 120 g/kg (Ranke et al., 1999). Although this therapy clearly is beneficial in promoting growth, the optimal regimen remains to be established. Side Effects of GH Therapy In children, GH therapy is associated with remarkably few side effects. Rarely, generally within the first 8 weeks of therapy, patients develop intracranial hypertension, with papilledema, visual changes, headache, nausea, and/or vomiting. Because of this, funduscopic examination is recommended at the initiation of therapy and at periodic intervals thereafter. Leukemia has been reported in some children receiving GH therapy; a causal relationship has not been established, and conditions associated with GH deficiency (e.g., Down syndrome, cranial irradiation for CNS tumors) probably explain the apparent increased incidence of leukemia. Despite this, the consensus is that GH should not be administered in the first year after treatment of pediatric tumors, including leukemia, or during the first two years after therapy for medulloblastomas or ependymomas (Blethen et al., 1996). An increased incidence of type 2 diabetes mellitus has been reported, presumably secondary to the anti-insulin metabolic effects of GH (Cutfield et al., 2000). In adults, side effects associated with the initiation of GH therapy include peripheral edema, carpal tunnel syndrome, arthralgia, and myalgia. These symptoms, which occur most frequently in patients who are older or more obese, generally respond to a decrease in dose. Although there are potential concerns about impaired glucose tolerance secondary to anti-insulin actions of GH, this has not been a major problem with clinical use at the recommended doses.

Agents Used in Syndromes of Growth Hormone Excess GH excess causes distinct clinical syndromes depending on the age of the patient. If the epiphyses are unfused, GH excess causes increased longitudinal growth, resulting in gigantism. In adults, GH excess causes acromegaly. The symptoms and signs of acromegaly (e.g., arthropathy, carpal tunnel syndrome, generalized visceromegaly, hypertension, glucose intolerance, headache, lethargy, excess perspiration, and sleep apnea) progress slowly, and diagnosis often is delayed. Life expectancy is shortened in these patients; mortality is increased at least two-fold relative to age-matched controls due to increased death from cardiovascular disease, upper airway obstruction, and gastrointestinal malignancies. While the diagnosis of acromegaly should be suspected in patients with the appropriate symptoms and signs, confirmation requires the demonstration of increased circulating GH or IGF-1. Generally, the first screening test is to measure serum IGF-1. Using a good assay with results compared to normal values for age and sex, a normal IGF-1 level argues strongly against the diagnosis of acromegaly. If the IGF-1 is frankly elevated or borderline or if the clinical suspicion is relatively strong, many clinicians also will measure plasma GH following administration of an oral glucose load. Using the standard radioimmunoassay for human GH, the GH level 2 hours after glucose administration normally is less than 2 g/liter in normal subjects; a higher value confirms the diagnosis of acromegaly. Treatment options in acromegaly include transphenoidal surgery, radiation, and drugs that inhibit GH secretion or action. Pituitary surgery traditionally has been viewed as the treatment of choice. In patients with microadenomas (i.e., tumors <1 cm), skilled neurosurgeons can achieve cure rates of up to 80% to 90%; however, the long-term success rate for patients with macroadenomas is considerably lower, often falling below 50%. In addition, there is increasing appreciation that acromegalic patients previously considered cured by pituitary surgery actually have persistent GH excess, with its attendant complications. Thus, more attention has been given to the role of pharmacological management of acromegaly, either as a primary treatment modality or for the treatment of persistent GH excess following transphenoidal surgery (Newman, 1999). Somatostatin Analogs The development of analogs of somatostatin (Table 562) has revolutionized the medical treatment of GH excess. The most widely used analog is octreotide (SANDOSTATIN), an eightamino acid synthetic derivative of somatostatin that has a longer half-life and binds preferentially to SSTR-2 and SSTR-5 receptors on GH-secreting tumors. Typically, octreotide (100 g) is administered subcutaneously three times daily; serum GH and IGF-1 levels are monitored to assess effectiveness of treatment. The goal is to decrease GH levels to less than 2 g/liter following an oral glucosetolerance test and to bring IGF-1 levels to within the normal range for age and sex. Depending on the biochemical response, higher or lower octreotide doses may be used in individual patients. In addition to its effect on GH secretion, octreotide can decrease tumor size in a minority of patients. In these cases, tumor growth generally resumes after octreotide treatment is stopped. Octreotide also has significant inhibitory effects on thryotropin secretion, and it is the treatment of choice for patients who have thryotrope adenomas that oversecrete TSH and who are not good candidates for surgery. The use of octreotide in gastrointestinal disorders is discussed in Chapter 39: Agents Used for Diarrhea, Constipation, and Inflammatory Bowel Disease; Agents Used for Biliary and Pancreatic Disease. Gastrointestinal side effectsincluding diarrhea, nausea, and abdominal painoccur in up to 50%

of patients receiving octreotide. In most patients, these symptoms diminish over time and do not require cessation of therapy. Approximately 25% of patients receiving octreotide develop gallstones, presumably due to decreased gallbladder contraction and gastrointestinal transit time. In the absence of symptoms, gallstones are not a contraindication to continued use of octreotide. Compared to somatostatin, octreotide has much less of an effect on insulin secretion and in clinical studies only infrequently affects glycemic control. The need to inject octreotide three times daily poses a significant obstacle to patient compliance. A long-acting, slow-release form of octreotide (SANDOSTATIN-LAR) is a more convenient alternative that can be administered intramuscularly once every 4 weeks; the recommended dose is 20 or 30 mg. The long-acting preparation is at least as effective as the regular formulation and is used in patients who have responded favorably to a trial of the shorter-acting formulation of octreotide. Like the shorter-acting formulation, the longer-acting formulation of octreotide generally is well tolerated and has a similar incidence of side effects (predominantly gastrointestinal and/or discomfort at injection site) that do not require cessation of therapy. Lanreotide (SOMATULINE LA) is a long-acting octapeptide analog of somatostatin that causes prolonged suppression of GH secretion when administered in a 30-mg dose intramuscularly. Although its efficacy appears comparable to that of the long-acting formulation of octreotide, its duration of action is shorter; thus it must be administered either at 10- or 14-day intervals. One direct comparison with a limited number of patients suggested that the long-acting formulation of octreotide at recommended doses may be somewhat more effective in lowering GH levels than is lanreotide (Turner et al., 1999). The incidence and severity of side effects associated with lanreotide are similar to those of the other somatostatin analogs. Lanreotide has not been approved by the FDA for use in the United States. Somatostatin blocks not only GH secretion, but also the secretion of other hormones, growth factors, and cytokines. Thus, octreotide and the delayed-release somatostatin analogs have been used to treat symptoms associated with metastatic carcinoid tumors (e.g., flushing and diarrhea) and symptoms of adenomas secreting vasoactive intestinal peptide (e.g., watery diarrhea). Octreotide also has been labeled with indium or technetium and used for diagnostic imaging of neuroendocrine tumors such as pituitary adenomas and carcinoids. Based on structure-function studies of somatostatin and its derivatives, the amino acid residues in positions 7 to 10 [FWKT] are the major determinants of biological activity. Residues W8 and K9 appear to be essential, whereas conservative substitutions at F7 and T10 are permissible. Active somatostatin analogs retain this core segment constrained in a cyclic structureformed either by a disulfide bond or amide linkagethat stabilizes the optimal conformation (Patel, 1999). As noted above, the endogenous peptides, somatostatin-14 and somatostatin-28, do not discriminate very well among SSTR subtypes except for SSTR5, which shows some preference for somatostatin-28. Greater selectivity is seen with some of the somatostatin analogs. For example, the octapepetides octreotide, lanreotide, and vapreotide and the hexapeptide seglitide all bind to the SSTR subtypes with the following order of selectivity: SSTR2 > SSTR5 > SSTR3 SSTR1 and SSTR4. The octapeptide analog BIM23268 exhibits modest selectivity for SSTR5, and the undecapeptide CH275 appears to bind preferentially to SSTR1 and 4 (Patel, 1999). More recently, a series of small nonpeptide agonists that exhibit a high degree of SSTR subtype-selectivity has been isolated from combinatorial chemical libraries; these compounds may lead to a new class of highly selective, orally active somatostatin mimetics. Dopamine-Receptor Agonists

The dopamine-receptor agonists are described in more detail below in the section dealing with treatment of prolactinomas. Although dopamine-receptor agonists normally stimulate GH secretion, they cause a paradoxical decrease in GH secretion in some patients with acromegaly. In patients who are unwilling to take injections, the long-acting dopamine-receptor agonist cabergoline (DOSTINEX) may lower GH and IGF-1 levels into the target range. The best responses have been seen in patients whose tumors secreted both GH and prolactin. Doses used in treating acromegaly typically are considerably higher than those employed in prolactinomas. Growth Hormone Antagonists As discussed above, derivatives of GH have been developed that bind the GH receptor but do not induce the formation of receptor dimers or activate Jak/Stat signaling. One such analog, pegvisomant, is now under clinical investigation for the treatment of acromegaly. In a 12-week trial, pegvisomant significantly decreased circulating IGF-1, achieving normal levels in up to 90% of patients at higher doses and causing significant improvement in clinical parameters such as ring size, soft-tissue swelling, and excessive perspiration and fatigue (Trainer et al., 2000). Because pegvisomant differs structurally from native GH, it may induce the formation of specific antibodies that will limit its long-term efficacy. Moreover, it substantially increases GH levels and possibly may have unanticipated side effects. Finally, there are at least theoretical concerns that loss of negative feedback by both growth hormone and IGF-1 may increase the growth of GH-secreting adenomas. Thus, while its ultimate role in the management of acromegaly remains to be determined, pegvisomant represents a novel pharmacologic agent in the management of GH excess. Prolactin As a member of the somatotropin family, prolactin is related structurally to GH and placental lactogen. The human prolactin gene on chromosome 6 encodes a 23,000-dalton polypeptide of 199 amino acids. This polypeptide has three intramolecular disulfide bonds, and a portion of secreted prolactin is glycosylated at a single asparagine residue. In circulation, dimeric and polymeric forms of prolactin also are found, as are degradation products of 16,000 or 18,000 daltons; the biological significance of these different forms is not known. Secretion Prolactin is synthesized in lactotropes. Prolactin synthesis and secretion in the fetal pituitary start in the first few weeks of gestation. Serum prolactin levels decline shortly after birth. Whereas serum prolactin levels remain low throughout life in normal males, they are elevated somewhat in normal cycling females. Prolactin levels rise markedly during pregnancy, reach a maximum at term, and decline thereafter unless the mother breast-feeds the child. In nursing mothers, prolactin secretion is stimulated by the suckling stimulus or breast manipulation, and circulating prolactin levels can rise 10- to 100-fold within 30 minutes of stimulation. This response becomes less pronounced after several months of breast-feeding, and prolactin concentrations eventually decline to prepregnancy levels. Prolactin detected in maternal and fetal blood originates from maternal and fetal pituitaries. Prolactin also is synthesized by decidual cells near the end of the luteal phase of the menstrual cycle and early in pregnancy; the latter source is responsible for the very high levels of prolactin in amniotic fluid during the first trimester. Many of the physiological factors that influence prolactin secretion are similar to those that affect GH secretion. Thus, sleep, stress, hypoglycemia, exercise, and estrogen increase the secretion of

both hormones. Like other anterior pituitary hormones, prolactin is secreted in a pulsatile manner. Prolactin is unique among the anterior pituitary hormones in that hypothalamic regulation inhibits its secretion. The major regulator of prolactin secretion is dopamine, which is released by tuberoinfundibular neurons and interacts with the D2 receptor on lactotropes to inhibit secretion of prolactin (Figure 563). A number of putative prolactin-releasing factors have been described, including TRH, vasoactive intestinal peptide, prolactin-releasing peptide, and pituitary adenylyl cyclase-activating peptide (PACAP), but their physiological roles are unclear. Under certain pathophysiological conditions, such as severe primary hypothyroidism, persistently elevated levels of TRH can induce hyperprolactinemia and galactorrhea. Figure 563. Prolactin Secretion and Actions. Prolactin is the only anterior pituitary hormone for which a unique stimulatory releasing factor (PRH?) has not been identified. Thyrotropin-releasing hormone (TRH) can stimulate prolactin release, however, and dopamine can inhibit it. Prolactin affects lactation and reproductive functions but it also has varied effects on many other tissues. Prolactin is not under feedback control by peripheral hormones.

Molecular and Cellular Bases of Prolactin Action The effects of prolactin result from interactions with specific receptors that are widely distributed among a variety of cell types within many tissues (Bole-Feysot et al., 1998). Whereas prolactin binds specifically to the prolactin receptor and has no GH-like (somatotropic) activity, human GH and placental lactogen bind to the prolactin receptors and are lactogenic. The prolactin receptor is related structurally to receptors for GH and several cytokines and uses similar signaling mechanisms (see above). The prolactin receptor is encoded by a single gene located on chromosome 5. Alternative splicing of this gene gives rise to multiple forms of the receptor, including a short form of 310 amino acids, a long form of 610 amino acids, and an intermediate form of 412 amino acids. In addition, soluble isoforms lacking the transmembrane and cytoplasmic domains bind prolactin in the circulation. Like the GH receptor, the prolactin receptor lacks intrinsic tyrosine kinase activity; hormoneinduced dimerization recruits and activates Jak kinases. Phosphorylation of Jak2 kinase induces phosphorylation, dimerization, and nuclear translocation of the transcription factor Stat5. Physiological Effects of Prolactin A number of hormonesincluding estrogens, progesterone, placental lactogen, and GHstimulate

development of the breast and prepare it for lactation. Prolactin, acting via the prolactin receptor, plays an important role in inducing growth and differentiation of the ductal and lobuloalveolar epithelium, and lactation does not occur in the absence of this hormone. During pregnancy, the high levels of estrogen and progesterone inhibit milk secretion; their declining levels after birth permit prolactin to induce lactation. Prolactin receptors also are present in many other tissues and organs, including the hypothalamus, liver, testes, ovaries, prostate, and immune system. The physiological effects of prolactin at these sites are poorly characterized. Hyperprolactinemia suppresses the hypothalamic-pituitary-gonadal axis, presumably due to inhibitory actions of prolactin on the hypothalamus and/or gonads. The elevated prolactin levels in women who are breast-feeding often suppress the normal menstrual cycle, and pathological hyperprolactinemia is a common cause of infertility in women (see below). Agents Used to Treat Syndromes of Prolactin Excess Prolactin has no therapeutic uses. Hyperprolactinemia is a relatively common endocrine abnormality that can result from hypothalamic or pituitary diseases that interfere with the delivery of inhibitory dopaminergic signals, from renal failure, from primary hypothyroidism associated with increased TRH levels, or from treatment with dopamine-receptor antagonists. Most often, hyperprolactinemia is caused by prolactin-secreting pituitary adenomaseither microadenomas (<1 cm in diameter) or macroadenomas ( 1 cm in diameter). Manifestations of prolactin excess in women include galactorrhea, amenorrhea, and infertility. In men, hyperprolactinemia causes loss of libido, impotence, and infertility. Generally, men seek medical attention considerably later than do women and thus have a higher frequency of macroadenomas. Neurological manifestations such as visual impairment or headache also can be associated with the larger pituitary tumors. Currently, the therapeutic options for patients with prolactinomas include transphenoidal surgery, radiation, and treatment with dopamine-receptor agonists that suppress prolactin production via D2dopamine receptors (Molitch, 1999). Inasmuch as initial surgical cure rates are only 70% with microadenomas and 30% with macroadenomas, most patients with prolactinomas ultimately require drug therapy. Thus, dopamine-receptor agonists have become the initial treatment of choice for many patients. These agents generally decrease both prolactin secretion and the size of the adenoma, thereby improving the endocrine abnormalities as well as the neurologic symptoms caused directly by the adenoma (including visual field deficits). Bromocriptine Bromocriptine (PARLODEL) is the dopamine-receptor agonist most frequently used to treat hyperprolactinemia and has become the standard against which newer agents are compared. Bromocriptine is a semisynthetic ergot alkaloid that interacts with D2dopamine receptors to inhibit both spontaneous and thyrotropin-releasing hormone (TRH)-induced release of prolactin; to a lesser extent, it also activates D1 dopamine receptors. Bromocriptine normalizes the prolactin level in 70% to 80% of patients with prolactinomas and decreases tumor size in more than 50% of patients, including those with macroadenomas. It is worth noting that bromocriptine does not cure the underlying adenoma, and hyperprolactinemia and tumor growth typically recur upon cessation of therapy. Frequent side effects of bromocriptine include nausea and vomiting, headache, and postural hypotensionparticularly on initial use. Less frequent side effects include nasal congestion, digital vasospasm, or CNS effects such as psychosis, hallucinations, nightmares, or insomnia. Patients often develop tolerance to these effects, which can be diminished by starting at a low dose (1.25

mg) administered at bedtime with a snack. After one week, a morning dose of 1.25 mg can be added. If clinical symptoms persist or the prolactin level remains elevated, the dose can be increased gradually, every 3 to 7 days, to 5 mg twice per day or 2.5 mg three times a day as tolerated. Patients who do not respond to bromocriptine or who develop intractable side effects may respond to a different dopamine agonist. Although a high fraction of the oral dose of bromocriptine is absorbed, only 7% of the dose reaches the systemic circulation due to a high extraction rate and extensive first-pass metabolism in the liver. Furthermore, bromocriptine has a relatively short elimination half-life (between 2 and 8 hours). To avoid the need for frequent dosing, a parenteral long-acting form of bromocriptine incorporated into biodegradable microspheres (PARLODEL-LAR) has been developed. Although not available in the United States, this product has produced results in clinical trials comparable to those of oral bromocriptine. Bromocriptine may be administered intravaginally (2.5 mg once daily), reportedly with fewer gastrointestinal side effects. Pergolide Pergolide (PERMAX), an ergot derivative approved by the FDA for treatment of Parkinson's disease, also is used "off label" to treat hyperprolactinemia. If the cost of therapy is an important consideration, pergolide is the cheapest available dopamine-receptor agonist. It induces many of the same side effects as does bromocriptine, but it can be given once a day, starting at 0.025 mg at bedtime and increased gradually to a maximum daily dose of 0.25 mg. Cabergoline Cabergoline (DOSTINEX) is an ergot derivative with a longer half-life (approximately 65 hours) and higher affinity and selectivity for the D2 receptor than bromocriptine (approximately 4-times more potent; Verhelst et al., 1999). Cabergoline has a much lower tendency to induce nausea, although it still may cause hypotension and dizziness. In some clinical trials, cabergoline has been more effective than bromocriptine in decreasing serum prolactin in patients with hyperprolactinemia. Cabergoline has been approved by the FDA for the treatment of hyperprolactinemia, and it likely will play an increasing role in the treatment of this syndrome. As approved by the FDA, therapy is initiated at a dose of 0.25 mg twice a week; a schedule of 0.5 mg once a week also has been used. If the serum prolactin remains elevated, the dose can be increased to a maximum of 1.5 mg two or three times a week as tolerated. The dose should not be increased more often than every 4 weeks. Quinagolide Quinagolide is a nonergot D2dopamine agonist with a half-life (22 hours) intermediate between those of bromocriptine and cabergoline. Quinagolide is administered once daily at doses of 0.1 to 0.5 mg/day. It is not approved by the FDA but has been used extensively in Europe. Patients with prolactinomas who desire to become pregnant make up a special subset of hyperprolactinemic patients. In this setting, drug safety during pregnancy is an important consideration. Bromocriptine, cabergoline, and quinagolide all induce ovulation and permit most patients with prolactinomas to become pregnant without apparent detrimental effects on pregnancy or fetal development. However, experience with cabergoline and quinagolide is much less extensive than that with bromocriptine. Therefore, bromocriptine is recommended as the first-line treatment in this setting, although opinion may change with more experience with cabergoline or quinagolide. Gonadotropin-Releasing Hormone and Gonadotropic Hormones The pituitary hormones, luteinizing hormone (LH) and follicle-stimulating hormone (FSH), as well

as the related placental hormone chorionic gonadotropin (CG), are referred to as the gonadotropic hormones because of their actions on the gonads. These three hormones and TSH consitute the glycoprotein family of pituitary hormones. Each hormone is a glycosylated heterodimer containing a common -subunit and a distinct -subunit that confers specificity of action. While all the subunits of this family are similar structurally, the -subunit of CG is most different, containing a carboxy-terminal extension of 30 amino acids and extra carbohydrate residues. The carbohydrate residues on the gonadotropins influence the rate of their clearance from the circulation, thus extending their serum half-lives; the residues also play a role in signal transduction at gonadotropin receptors. The human FSH gene is located at 11p13, the LH is at 19q12.32, in close proximity to at least seven CG genes, and the gene encoding the -subunit maps to chromosome 6q21-23. Regulation of Gonadotropin Secretion The regulation of gonadotropin secretion is described in detail in Chapters 58: Estrogens and Progestins and 59: Androgens. LH and FSH are synthesized and secreted by gonadotropes, which make up approximately 20% of the hormone-secreting cells in the anterior pituitary. CGproduced only in primates and horsesis made by syncytiotrophoblast cells of the placenta. Pituitary gonadotropin production is stimulated by GnRH and is further regulated by feedback effects of the gonadal hormones (Figure 564; see also Figure 582). Figure 564. The Hypothalamic-Pituitary-Gonadal Axis. A single hypothalamic releasing factor, gonadotropin-releasing hormone (GnRH), controls the synthesis and release of both gonadotropins (LH and FSH) in males and females. Gonadal steroid hormones (androgens, estrogens, and progesterone) cause feedback inhibition at the level of the pituitary and the hypothalamus. The preovulatory surge of estrogen also can exert a stimulatory effect at the level of the pituitary and the hypothalamus. Inhibin, a polypeptide hormone produced by the gonads, specifically inhibits FSH production by the pituitary.

Regulation of Release of Gonadotropin-Releasing Hormone Gonadotropin-releasing hormone (GnRH) regulates the synthesis and secretion of FSH and LH by pituitary gonadotropes. GnRH is encoded by a gene on chromosome 8p21 and is derived by proteolytic processing of a 92amino acid precursor peptide to produce mature GnRH, a decapeptide with blocked amino and carboxyl termini (see Table 563). GnRH release is intermittent and is governed by a neural pulse generator that is located in the mediobasal hypothalamusprimarily in the arcuate nucleusand that controls the frequency and amplitude of GnRH release from neurons in the hypothalamus. Although active late in fetal life and for approximately 1 year after birth, activity of the GnRH pulse generator decreases considerably thereafter, presumably secondarily to inhibition by the CNS. Shortly before puberty, CNS inhibition decreases and there is an increased amplitude and frequency of GnRH pulses, particularly during sleep. As puberty progresses, the GnRH pulses increase further in amplitude and frequency until the normal adult pattern is established. The intermittent release of GnRH is crucial for the proper synthesis and release of the gonadotropins, which also are released in a pulsatile manner. The continuous administration of GnRH leads to desensitization and down-regulation of GnRH receptors on pituitary gonadotropes. The latter actions form the basis for the clinical use of longacting GnRH analogs that suppress gonadotropin secretion (see below for further discussion). These compounds transiently increase LH and FSH secretion, but eventually desensitize gonadotropes to GnRH, thereby inhibiting gonadotropin release. Molecular and Cellular Bases of GnRH Action The GnRH receptor, a member of the family of G proteincoupled receptors, is encoded by a gene

on chromosome 4q21. The binding of GnRH or GnRH agonists to GnRH receptors on the gonadotropes activates Gq11, which in turn stimulates phospholipase activity and increases the intracellular concentration of Ca2+, thereby increasing both the synthesis and secretion of LH and FSH. Although cyclic AMP is not the major mediator of GnRH action, binding of GnRH also modulates adenylyl cyclase activity. GnRH receptors also are present in the ovary and testis, although their physiological significance at these sites remains to be determined. Gonadal steroids also regulate gonadotropin productionat the level of both the pituitary and the hypothalamusbut effects on the hypothalamus predominate. The feedback effects of gonadal steroids are gender-, dosage-, and time-dependent. In women, low levels of estradiol and progesterone inhibit gonadotropin production largely through opioid action on the neural pulse generator that controls GnRH production. Higher and more sustained levels of estradiol have positive feedback effects that ultimately result in the gonadotropin surge that precedes ovulation. In men, testosterone inhibits gonadotropin production, in part through direct actions and in part after its metabolism to estradiol. Another important regulator of gonadotropin production is the gonadal peptide hormone inhibin. Inhibin is made by granulosa cells in the ovary and Sertoli cells in the testis in response to the gonadotropins and local growth factors; it acts directly in the pituitary, selectively inhibiting FSH secretion without affecting that of LH. Inhibin is structurally similar to the family of glycoproteins that includes transforming growth factor and antimllerian hormone. Molecular and Cellular Bases of Gonadotropin Action LH and FSH were named initially based on their actions on the ovary; appreciation of their roles in male reproductive function did not come until later. The actions of LH and CG are mediated by the LH receptor (the gene for which is located on chromosome 2p21) and those of FSH are mediated by the FSH receptor (the gene for which is located on chromosome 2q). Both of these G protein coupled receptors have large, glycosylated extracellular domains that contribute to their affinity and specificity for their ligands. The FSH and LH receptors couple with Gs to activate adenylyl cyclase and raise the intracellular level of cyclic AMP. At higher ligand concentrations, the agonistoccupied gonadotropin receptors also activate protein kinase C and Ca2+ signaling pathways via Gqmediated effects on phospholipase C activity. Since most if not all of the actions of the gonadotropins can be mimicked by cyclic AMP analogs, the precise physiological role of Ca2+ and protein kinase C in gonadotropin action remains to be determined. Physiological Effects of Gonadotropins In men, LH acts on testicular Leydig cells to stimulate the de novo synthesis of androgens, primarily testosterone. Testosterone is required for gametogenesis within the seminiferous tubules and for maintenance of libido and secondary sexual characteristics. FSH acts on the Sertoli cells to stimulate the production of proteins and nutrients required for sperm maturation, thereby indirectly supporting germ cell maturation. The actions of FSH and LH in women are more complicated than those in men. FSH stimulates the growth of developing ovarian follicles and induces the expression of LH receptors on both theca and granulosa cells. FSH also regulates the activity of aromatase in granulosa cells, thereby stimulating the production of 17 -estradiol. LH acts on the theca cells to stimulate the synthesis of androstenedione, the major precursor of ovarian 17 -estradiol in premenopausal women. LH also is required for the rupture of the dominant follicle during ovulation and for the synthesis of progesterone by the corpus luteum. Finally, LH and the LH receptor in women induce the

expression of the FSH receptor by granulosa cells; LH thus plays a permissive role in FSH action. The essential roles of gonadotropins in reproductive physiology are revealed by human subjects with mutations of either the gonadotropin subunits or their cognate receptors (Achermann and Jameson, 1999). Women with mutations in either FSH or its receptor present clinically with primary amenorrhea, infertility, and absent breast development. Histologically, the ovarian follicles fail to mature and corpora lutea are missing. These findings, in conjunction with success in assisted reproductive technologies using FSH alone (see below), establish the critical role of FSH in ovarian function. In men, mutations of FSH or the FSH receptor are associated with decreased testis size and oligospermia, although several subjects have been fertile. The only reported inactivating mutation of LH was in a 46-year-old XY subject with Leydig cell hypoplasia, lack of spontaneous puberty, and infertility. The external genitalia were masculinized, suggesting that CG mediates androgen production in utero. In contrast, apparently complete loss-offunction mutations of the LH receptor cause phenotypes ranging from male hypogonadism to maleto-female sex reversal of the external genitalia and failure to initiate puberty. Presumably, the absence of any virilization of the external genitalia reflects combined loss of both CG and LH signaling in utero. Women with homozygous inactivating mutations of the LH receptor present with primary amenorrhea, oligoamenorrhea, or infertility and have cystic ovaries on histological examination. Mutations leading to a constitutively active LH receptor affect males primarily and are autosomal dominant. These mutations result in precocious puberty due to the uncontrolled production of testosterone in the fetal and prepubertal periods. A subset of these mutations also has been associated with testicular tumors. Diagnostic and Therapeutic Uses of GnRH and Its Analogs As illustrated in Table 563, a number of clinically useful GnRH analogs have been synthesized. These include synthetic GnRH (gonadorelin) and GnRH analogs that contain substitutions at position 6 that protect against proteolysis and substitutions at the C-terminus that improve receptorbinding affinity. The analogs exhibit enhanced potency and a prolonged duration of action compared to GnRH, which has a half-life of approximately 2 to 4 minutes. Pure GnRH antagonists have been developed that do not cause the initial increase in gonadotropin secretion seen with the long-acting GnRH agonists. These newer antagonists apparently do not provoke local and systemic histamine release and the anaphylactoid reactions that hampered the clinical development of earlier analogs. Two different GnRH antagonists, ganirelix (ORGALUTRAN, ANTAGON) and cetrorelix (CETROTIDE), have been used to suppress the LH surge in ovarianstimulation protocols that are part of assisted reproduction techniques. Ganirelix is available in the United States. Cetrorelix is available in Europe, but not in the United States. Although the almost immediate suppression of LH theoretically should result in a decreased duration of the in vitro fertilization cycle and a better-controlled regimen of ovarian stimulation (see below), more clinical trials are needed to define the roles of these compounds in assisted reproduction technologies. Diagnostic Uses Synthetic GnRH (gonadorelin hydrochloride;FACTREL) is marketed for diagnostic purposes to differentiate between pituitary and hypothalamic defects in patients with hypogonadotropic hypogonadism. After a blood sample is obtained for the baseline LH value, a single 100- g dose of GnRH is administered subcutaneously or intravenously and serum LH levels are measured over the

next 2 hours (at 15, 30, 45, 60, and 120 minutes after injection). A normal LH response indicates the presence of functional pituitary gonadotropes. Inasmuch as the long-term absence of GnRH can result in a decreased responsiveness of otherwise normal gonadotropes, the absence of a response does not always indicate intrinsic pituitary disease. GnRH-stimulation testing also can be used to determine whether a subject with precocious puberty has central (i.e., GnRH-dependent) or peripheral precocious puberty. Management of Infertility Gonadorelin acetate (LUTREPULSE ) is a synthetic preparation of GnRH used to treat patients with reproductive disorders secondary to GnRH deficiency or disordered secretion of GnRH. It is administered by an intravenous pump in pulses that promote a physiological cycle, starting at doses of 2.5 g per pulse every 60 to 90 minutes. If necessary, the dose can be increased to 10 g per pulse until ovulation is induced, as described in the manufacturer's manual provided with the kit. Advantages over gonadotropin therapy (see below) include a lower risk of multiple pregnancies and a decreased need to monitor plasma estrogen levels or ovarian ultrasonography. Side effects generally are minimal; the most common is local phlebitis due to the infusion device. In women, normal cycling levels of ovarian steroids can be achieved, leading to ovulation and menstruation. Because of its complexity, however, this regimen is available only in specialized centers in reproductive endocrinology (Hayes et al., 1998). Although growth of testes, appearance of normal levels of gonadal steroids, and induction of spermatogenesis can be achieved in men, GnRH therapy to induce fertility in men is not approved by the FDA, is relatively expensive, and requires that an infusion pump be worn constantly. Therefore, gonadotropins generally are preferred. The long-acting GnRH agonists also have been used in ovulation-induction protocols to suppress the endogenous preovulatory surge of LH and thus prevent premature follicular luteinization. Several treatment regimens have been developed in which the GnRH agonist is given for either short or long periodsin conjunction with gonadotropins to induce follicular maturation (see below)and then ovulation is induced with CG (Lunenfeld, 1999). Suppression of Gonadotropin Secretion As noted above, long-acting GnRH analogs eventually desensitize GnRH receptor-elicited signaling pathways, markedly inhibiting gonadotropin secretion and decreasing the production of gonadal steroids. This "medical castration" has proven to be very useful in disorders that respond to reductions in gonadal steroids. Perhaps the clearest indication for this therapy is in children with gonadotropin-dependent precocious puberty (also called central precocious puberty), whose premature sexual maturation can be arrested with minimal side effects by chronic administration of the GnRH agonists. Long-acting GnRH agonists are used for palliative therapy of hormonally responsive tumors (e.g., prostate or breast cancer), generally in conjunction with agents that block steroid biosynthesis or action to avoid transient increases in hormone levels. The analogs also are used to suppress steroidresponsive conditions such as endometriosis, uterine leiomyomas, and acute intermittent porphyria. Finally, depot preparations of goserelin (ZOLADEX), which can be implanted subcutaneously every 3 months (10.8 mg), may make this drug particularly useful for medical castration in disorders such as pedophilia, where strict patient supervision may be required to ensure compliance. The long-acting agonists generally are well tolerated, and side effects are those that would be

predicted to occur when gonadal steroidogenesis is inhibited (e.g., hot flashes, vaginal dryness and atrophy, decreased bone density). Because of these effects, therapy in settings such as endometrioisis or uterine leiomyomas generally is limited to 6 months unless add-back therapy with estrogens is included to minimize effects on bone density. Diagnostic Uses of Gonadotropins Diagnosis of Pregnancy Significant amounts of CG are present in both the maternal bloodstream and urine during pregnancy and can be detected immunologically with antisera raised against its unique -subunit. This provides the basis for commercial pregnancy kits that qualitatively assay for the presence or absence of CG in the urine. These kits, which offer a rapid, noninvasive means of detecting pregnancy within a few days after a woman's first missed menstrual period, are available in the United States without a prescription. Quantitative measurements of CG concentration in plasma are determined by radioimmunoassay in clinical and research laboratories. These assays typically are used to assess whether or not pregnancy is proceeding normally or to help detect the presence of an ectopic pregnancy, hydatidiform mole, or choriocarcinoma. Prediction of Ovulation Ovulation occurs 36 hours after the onset of the LH surge (10 to 12 hours after the peak of LH). Therefore, urinary concentrations of LH can be used to predict the time of ovulation. Kits are commercially available without a prescription that provide a semiquantitative assessment of LH levels in urine, using LH-specific antibodies that do not recognize other gonadotropins. Urine LH levels are measured every 12 to 24 hours, beginning on day 11 of the menstrual cycle (assuming a 28-day cycle), to detect the rise in LH and thus estimate the time of ovulation. Such estimates facilitate the timing of sexual intercourse to achieve pregnancy. Diagnosis of Diseases of the Male and Female Reproductive System Measurements of plasma LH and FSH levels, as determined by quantitative, subunitspecific radioimmunoassays, are useful in the diagnosis of several reproductive disorders. Low or undetectable levels of LH and FSH are indicative of hypogonadotropic hypogonadism and suggest hypothalamic or pituitary disease, whereas high levels of gonadotropins suggest primary gonadal diseases. Therefore, in cases of amenorrhea in women or delayed puberty in men and women, measurements of plasma gonadotropins can be used to distinguish between gonadal failure and hypothalamic-pituitary failure. The FSH level on day 3 of the menstrual cycle is useful in assessing relative fertility. An FSH level of 15 IU/ml is associated with reduced fertility, even if a woman is menstruating normally, and predicts a lower likelihood of success in assisted reproduction techniques such as in vitro fertilization (see below). CG also is used diagnostically to stimulate testosterone production and thus assess Leydig cell function in men suspected of having Leydig cell failure (for example, in delayed puberty). Serum testosterone levels are assayed after multiple injections of CG. A diminished testosterone response to CG indicates Leydig cell failure; a normal testosterone response suggests a hypothalamic-

pituitary disorder. Therapeutic Uses of Gonadotropins Gonadotropins for clinical use originally came from human pituitaries and women's urine. Pituitary extracts are no longer used because of possible contamination with the Creutzfeldt-Jakob prion. Since their initial introduction, several different preparations of urinary gonadotropins have been developed. Chorionic gonadotropin (PREGNYL, A.P.L., PROFASI, others), which mimics the action of LH, is obtained from the urine of pregnant women. Urine from postmenopausal women is the source of menotropins (PERGONAL, HUMEGON, REPRONEX), which contain roughly equal amounts of FSH and LH as well as a number of other urinary proteins. Because of their relatively low purity, menotropins are administered intramuscularly to decrease the incidence of hypersensitivity reactions. Urofollitropin (uFSH; METRODIN, FERTINEX) is a purified FSH preparation from which most of the LH has been removed by immunodepletion. Finally, a highly purified FSH preparation is prepared by immunoconcentation with monoclonal antibodies (METRODIN HP). This preparation is sufficiently pure that it can be administered subcutaneously. Recombinant FSH (rFSH), prepared by expressing cDNAs encoding the and subunits of FSH in a mammalian cell line, yields products whose glycosylation pattern mimics that of FSH produced by the gonadotropes. The two rFSH preparations that are available [follitropin (GONAL-F) and follitropin (PUREGON, FOLLISTIM)] differ slightly in their carbohydrate structures. Both rFSH preparations can be administered subcutaneously, since they are considerably purer and exhibit less interbatch variability than do preparations purified from urine. The recombinant preparations are considerably more expensive than the naturally derived hormones, and their relative advantages (i.e., efficacy, lower frequency of side effects such as ovarian hyperstimulation) have not been definitively established. Eventually, the recombinant technology is likely to lead to improved forms of gonadotropins with increased half-lives or higher clinical efficacy. Female Infertility Infertility affects approximately 10% of couples of reproductive age. Increasingly, gonadotropins are used in the treatment of infertility (Vollenhoven and Healy, 1998), often in conjunction with assisted reproduction technologies (ART). Although most clearly indicated in anovulatory women with hypogonadotropic hypogonadism secondary to hypothalamic or pituitary dysfunction [patients with World Health Organization (WHO) class I anovulation], gonadotropins also are used to induce ovulation in women with the polycystic ovary syndrome who do not respond to clomiphene citrate (WHO class II; see Chapter 58: Estrogens and Progestins). Finally, gonadotropins also are used in women who are infertile despite normal ovulation, although therapy with clomiphene citrate typically is attempted first. Clinical use of gonadotropins should be limited to physicians who are experienced in the treatment of infertility or endocrine disorders. FSH alone can induce ovulation in most anovulatory women. A typical therapeutic regimen is to administer 75 IU daily. This dosage is given daily until cycle day 6 or 7, when the ovarian response is assessed by determining the number and size of developing follicles by transvaginal ultrasound. Scans typically are performed every 2 to 3 days and focus on identifying intermediate follicles. The finding of a follicle larger than 18 mm in diameter indicates that follicular development has progressed adequately. If three or more follicles >16 mm are present, gonadotropin therapy generally is stopped and pregnancy prevented by barrier contraception to decrease the likelihood of multiple pregnancies or the ovarian hyperstimulation syndrome (OHSS, see below). Measurements of serum estradiol levels also may be helpful. The target estradiol range is from 500 to 1500 pg/ml, with lower levels indicating inadequate gonadotropin stimulation and higher doses portending an

increased risk of OHSS. If laboratory assessment indicates impaired ovarian response, the dose of FSH can be increased to 150 IU daily. To complete follicular maturation and induce ovulation, CG (5000 to 10,000 IU) is given one day after the last dose of gonadotropin. In approximately 10% to 20% of cases, gonadotropin-induced ovulation is associated with multiple births, resulting from nonphysiological development of more than one preovulatory follicle and the release of more than one ovum. Gonadotropin induction also is used in conjunction with ART, including in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI). Again, FSH is administered to induce follicular maturation, CG is given to induce final oocyte development, and then the mature eggs are surgically retrieved from the preovulatory follicles. The retrieved ova are fertilized in vitro with sperm (IVF) or by sperm injection (ICSI), and they are then transferred to the uterus or fallopian tubes. With ART, the increased risk of multiple births is related to the number of fertilized eggs that are transferred to the woman. Aside from the risk of multiple births and its attendant complications, the major side effect of gonadotropin treatment is OHSS. OHSS, which is believed to result from increased ovarian secretion of a substance that increases vascular permeability, is characterized by rapid accumulation of fluid in the peritoneal cavity, thorax, and even the pericardium. Signs and symptoms include abdominal pain and/or distention, nausea and vomiting, diarrhea, marked ovarian enlargement, dyspnea, and oliguria. Consequences of the OHSS include hypovolemia, electrolyte abnormalities, abnormal fluid accumulation (e.g., ascites, pleural effusions, hemoperitoneum), acute respiratory distress syndrome, thromboembolic events, and hepatic dysfunction. If there is clinical suspicion that OHSS is developing before CG administration, then CG must be withheld. Some studies have suggested that gonadotropins are associated with an increased risk of ovarian cancer, but this conclusion is controversial. Importantly, there is no evidence that either menotropins or FSH increases the rate of congenital abnormalities in babies born from oocytes that were stimulated with gonadotropins. Male Infertility In men with impaired fertility secondary to gonadotropin deficiency, gonadotropins can establish or restore fertility. Partly due to expense and partly due to the occasional development of resistance to gonadotropins with prolonged use, standard treatment is to induce sexual development with androgens, reserving gonadotropins until fertility is desired. Treatment typically is initiated with CG (1000 to 5000 IU intramuscularly) three times per week until clinical parameters and the plasma testosterone level indicate full induction of steroidogenesis. Thereafter, the dose of CG is reduced to 2000 IU twice a week, and menotropins are injected three times a week (with typical doses ranging from 75 IU LH/FSH to 150 IU LH/FSH) to fully induce spermatogenesis. The most common side effect of gonadotropin therapy is gynecomastia, which occurs in up to one-third of patients and presumably reflects increased production of estrogens. Maturation of the prepubertal testis typically requires treatment for more than 6 months, and optimal spermatogenesis in some patients may require treatment for up to two years. Once spermatogenesis has been initiated by this combined therapy or in patients who developed hypogonadotropic hypogonadism after sexual maturation, ongoing treatment with CG alone usually is sufficient to support sperm production. As discussed above in the section entitled "Female Infertility," regimens employing recombinant LH, FSH, and CG very likely will play increasing

clinical roles. Cryptorchidism Cryptorchidism, the failure of one or both testes to descend into the scrotum, affects up to 3% of full-term male infants and becomes less prevalent with advancing age. Cryptorchid testes have defective spermatogenesis and are at increased risk to develop germ cell tumors. Hence, the current approach is to reposition the testes as early as possible, typically at 1 year of age but definitely before 2 years of age. As descent of the testes is stimulated by androgens, CG can be used to induce testicular descent if the cryptorchidism is not secondary to anatomical blockage. Therapy usually consists of injections of CG (3000 U/m2 body surface area) intramuscularly every other day for 6 doses. If this does not induce testicular descent, orchiopexy should be performed Oxytocin The structures of the neurohypophyseal hormonesoxytocin and vasopressin (antidiuretic hormone; ADH)and the physiology and pharmacology of vasopressin are presented in Chapter 30: Vasopressin and Other Agents Affecting the Renal Conservation of Water. The following discussion emphasizes the physiology of oxytocin and its use in pregnancy. Biosynthesis of Oxytocin Oxytocin is a cyclic nonapeptide that is structurally similar to vasopressin, differing by only two amino acids. It is synthesized as a larger precursor molecule in cell bodies of the paraventricular nucleus and, to a lesser extent, the supraoptic nucleus in the hypothalamus. The precursor is rapidly broken down to the active hormone and its neurophysin by proteolysis, packaged into secretory granules as an oxytocin-neurophysin complex, and secreted from nerve endings that terminate primarily in the posterior pituitary gland (neurohypophysis). In addition, oxytocinergic neurons that are known to regulate the autonomic nervous system project to regions of the hypothalamus, brainstem, and spinal cord. Other sites of oxytocin synthesis include the luteal cells of the ovary, the uterus, and fetal membranes. Stimuli for oxytocin secretion include sensory stimuli arising from the cervix and vagina and from suckling at the breast. Increases in circulating oxytocin in women in labor are difficult to detect, partly because of the pulsatile nature of oxytocin secretion and partly because of the activity of circulating oxytocinase. Nevertheless, most consistent increases have been observed during the expulsive phase triggered by sustained distension of the uterine cervix and vagina. Estradiol stimulates oxytocin secretion, whereas the ovarian polypeptide relaxin inhibits release. Other factors that primarily affect vasopressin secretion also have some impact on oxytocin release: e.g., ethanol inhibits oxytocin release, and pain, dehydration, hemorrhage, and hypovolemia stimulate oxytocin release. Although peripheral actions of oxytocin appear to play no significant role in the response to dehydration, hemorrhage, or hypovolemia, oxytocin may participate in the central regulation of blood pressure. As described below, pharmacological doses of oxytocin can inhibit free water clearance by the kidney, occasionally causing water intoxication if not used carefully. Physiological Roles of Oxytocin Uterus Oxytocin stimulates both the frequency and force of uterine contractions. These effects are highly dependent on estrogen, and the immature uterus is quite resistant to the effects of oxytocin.

Progesterone antagonizes the stimulant effect of oxytocin in vitro, and the decline in progesterone seen in late pregnancy may play an important role in the normal initiation of human parturition. A very low level of motor activity prevails in the human uterus during the first two trimesters of pregnancy. During the third trimester, spontaneous motor activity progressively increases until the sharp rise that constitutes the initiation of labor and delivery. The responsiveness of the uterus to oxytocin roughly parallels the increase in spontaneous activity. Exogenous oxytocin can initiate or enhance rhythmic contractions at any time, but a considerably higher dose is required in early pregnancy. Thus, an eightfold increase in uterine sensitivity to oxytocin occurs in the last half of pregnancy, mostly in the last 9 weeks, accompanied by a 30-fold increase in oxytocin receptors between early pregnancy and early labor. Because of the difficulties associated with measurements of oxytocin levels (see above) and because loss of pituitary oxytocin seems not to compromise labor and delivery, the physiological role of oxytocin in pregnancy has been highly debated. The finding that the oxytocin antagonist atosiban is effective in suppressing preterm labor (see below) argues in favor of the physiological importance of oxytocin. Breast Oxytocin plays an important physiological role in milk ejection. Stimulation of the breast through suckling or mechanical manipulation induces oxytocin secretion, causing contraction of the myoepithelium that surrounds areolar channels in the mammary gland. This action forces milk from the alveolar channels into large collecting sinuses, where it is available to the suckling infant. Mechanism of Action Oxytocin acts via specific G proteincoupled membrane receptors most closely related to the V1a and V2vasopressin receptors. In the human myometrium, these receptors are coupled to Gq and G11, which upon activation lead the generation of inositol 1,4,5-trisphosphate from phosphoinositide hydrolysis, subsequent mobilization of calcium from intracellular stores, and depolarizationinduced activation of voltage-sensitive calcium channels. The number of oxytocin receptors differs at various stages of pregnancy and increases significantly late in gestation, paralleling the marked increase in myometrial sensitivity to oxytocin. While this increase in receptor number may indicate a role of oxytocin in labor initiation, it also may represent one of the many changes occurring in preparation for uterine involution postpartum. Oxytocin also increases local prostaglandin production, which further stimulates uterine contractions. Clinical Use of Oxytocin Induction of Labor Uterine-stimulating agents are used most frequently to induce or augment labor in selected pregnant women (Dudley, 1997). Indications for induction of labor include situations in which the risk of continued pregnancy to the mother or fetus is considered to be greater than the risks of delivery or of pharmacological induction. Such circumstances include premature rupture of the membranes, isoimmunization, intrauterine growth retardation, and placental insufficiency (as in diabetes, preeclampsia, or eclampsia). Before labor is induced, it is essential to verify that the fetal lungs are sufficiently mature (i.e., the lecithin/cholesterol ratio in amniotic fluid is >2) and to exclude potential contraindications (e.g., abnormal fetal position, evidence of fetal distress, placental abnormalities, or previous uterine surgery that predisposes the uterus to rupture during labor). Oxytocin (PITOCIN, SYNTOCINON) is the drug of choice for induction of labor. It is administered by intravenous infusion of a diluted solution (typically 10 mU/ml), preferably by means of a variable-

speed infusion pump. Although there is continuing debate concerning the optimal dose to induce labor, many physicians use a protocol involving an initial dose of 1 mU/minute, with dose increases of no greater than 1 mU/minute every 30 to 40 minutes. Other authorities advocate a more aggressive approach, with starting doses of 6 mU/minute and increases of up to 2 mU/minute at 20minute intervals. Some published trials have suggested that the higher-dose regimens result in a lower rate of cesarean sections. For the induction of labor, if doses of 30 to 40 mU/minute fail to initiate satisfactory uterine contractions, higher rates of infusion are unlikely to be successful. As labor progresses, the dose of oxytocin required to maintain good uterine contractions may decrease. During labor induction, a physician must be immediately available, and the mother and fetus should be monitored continuously to determine fetal and maternal heart rates, maternal blood pressure, and the strength of uterine contraction. If uterine hyperstimulation occurs, as evidenced by too frequent contractions or the development of uterine tetany, the oxytocin should be discontinued immediately. The half-life of intravenous oxytocin is short ( 3 minutes); thus the hyperstimulatory effects of oxytocin should resolve within several minutes after cessation of the infusion. Because of its structural similarity to vasopressin, oxytocin at higher doses has pronounced antidiuretic effects. Infusions of 20 mU/minute decrease free water clearance by the kidney. Particularly if hypotonic fluids (e.g., dextrose in water) are infused in appreciable amounts, water intoxication may result in convulsions, coma, and even death. Vasodilatory actions of oxytocin also have been noted, particularly at high doses, which may provoke hypotension and reflex tachycardia. Deep anesthesia may exaggerate the hypotensive effect of oxytocin by causing less tachycardia. Augmentation of Labor In most circumstances, oxytocin should not be used to augment labor that is progressing normally because the resulting uterine hyperstimulation often is too forceful and sustained to be compatible with the safety of the mother and fetus. For the augmentation of hypotonic contractions in dysfunctional labor, it rarely is necessary to exceed an infusion rate of 10 mU/minute; doses of >20 mU/minute rarely are effective when lower concentrations fail. Potential complications of overstimulation include trauma of the mother or fetus due to forced passage through an incompletely dilated cervix, uterine rupture, and compromised fetal oxygenation due to loss of placental exchange. In the setting of dysfunctional labor, as seen most frequently in nulliparous women, oxytocin can be used to advantage by experienced obstetricians to facilitate labor progression. Oxytocin usually is effective where there is a very prolonged latent phase of cervical dilation, as well as in cases where there is a significant arrest of dilation or descent. The use of epidural anesthesia can impair the reflex stimulation of endogenous oxytocin during the second stage of labor; in this setting, the cautious administration of oxytocin may facilitate labor progression. Third Stage of Labor and Puerperium After delivery of the fetus or following therapeutic abortion, it is desirable to have the uterus firm and contracted, as this greatly reduces the incidence and extent of hemorrhage. Oxytocin often is given immediately after delivery to help maintain uterine contractions and tone. Typically, 20 mU of oxytocin is diluted in 1 liter of intravenous solution and infused at a rate of 10 ml/minute for a few minutes until the uterus is contracted. Then, the infusion rate is reduced to 1 to 2 ml/minute until the mother is ready for transfer to the postpartum unit. If this is ineffective, ergot alkaloids such as ergonovine maleate (ERGOTRATE ) or methylergonovine maleate (METHERGINE) may be used in nonhypertensive patients. The ergot alkaloids are discussed in more detail in Chapter 11: 5Hydroxytryptamine (Serotonin): Receptor Agonists and Antagonists.

Oxytocin Challenge Test In patients whose pregnancy holds increased risk for maternal or fetal complications (e.g., maternal diabetes mellitus or maternal hypertension), an oxytocin challenge test can be used to assess fetal well-being. Oxytocin is infused intravenously, initially at a rate of 0.5 mU/minute; this rate is increased slowly until 3 uterine contractions occur in 10 minutes. Concurrent monitoring of the fetal heart rate indicates whether or not the uterine contractions are associated with changes in fetal heart rate known to be associated with fetal distress. The outcome of the oxytocin challenge test is helpful in determining the presence of adequate placental reserve for continuation of high-risk pregnancies. Oxytocin-Receptor Antagonists Peptide analogs that competitively inhibit the interaction of oxytocin with its membrane receptor have been of some interest because of their potential use in the treatment of preterm labor (Goodwin and Zograbyan, 1998). The most widely studied oxytocin-receptor antagonist, atosiban, has been evaluated as a potential treatment for preterm labor. Although atosiban significantly decreases the frequency of uterine contractions in women in preterm labor, clinical trials to date have shown no improvement in infant outcomes, and the FDA has not approved atosiban for use in the United States. Prospectus Recent years have witnessed enormous advances in our understanding of the regulation of secretion of anterior pituitary hormones. It seems inevitable that the identification and characterization of additional physiological regulators of pituitary hormone secretion will facilitate the development of new drugs that can manipulate this secretion. For example, the characterization of the receptor for GH secretagoguesas well as the identification of its putative endogenous ligand (Kojima et al., 1999)provide novel approaches to modulate GH secretion. Similarly, the characterization of the different somatostatin receptor subtypes and the identification of agonists with greater selectivity toward these subtypes may provide new drugs that are more efficacious or have fewer side effects. Finally, the observation that dopamine and somatostatin receptors can form heterodimers with enhanced functional activity (Rocheville et al., 2000) may provide novel therapeutic strategies to manipulate growth hormone and/or prolactin secretion. Recent advances in techniques for peptide synthesis and recombinant protein expression have led to the clinical application of a number of drugs used in treating pituitary hormone excess or deficiency. Although the relative advantages of recombinant versus urinary gonadotropin preparations still are being defined, it seems certain that recombinant hormones will be used increasingly in clinical applications. Moreover, an improved knowledge of structure-function relationships of pituitary hormones very likely also will lead to the development of novel therapeutic products. For example, insights from the structure of GH bound to its receptor and sitedirected mutagenesis permitted the development of pegvisomant, a genetically altered variant of human GH that acts as a GH antagonist. Similar approaches very likely will lead to the development of other modified forms of pituitary hormones for clinical application. For further discussion of disorders associated with the pituitary gland and hypothalamus, see Chapters 318 and 319 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005. Acknowledgment

The authors wish to acknowledge Drs. Mario Ascoli and Deborah L. Segaloff, authors of this chapter in the ninth edition of Goodman and Gilman's The Pharmacological Basis of Therapeutics, some of whose text has been retained in this edition.

Chapter 57. Thyroid and Antithyroid Drugs

Overview This chapter discusses the function of the thyroid hormones, thyroxine (T4) and triiodothyronine (T3), in growth and metabolism and the regulation of thyroid function by thyroid-stimulating hormone (TSH) secreted from the pituitary. Calcitonin, also secreted by the thyroid gland, is discussed in Chapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds. Evaluation of free thyroxine and TSH levels as a means to assess thyroid function is provided as a prelude to the discussion of treatment of the hypothyroid patient with hormone replacement and of the hyperthyroid individual with one of a variety of antithyroid drugs, such as propylthiouracil and methimazole, and other thyroid inhibitors, including ionic inhibitors that interfere with the concentration of iodide by the thyroid gland and radioactive iodine, used both for diagnosis as well as treatment of hyperthyroidism. Although disorders of the thyroid are common, effective treatment of most thyroid disorders is available. Thyroid and Antithyroid Drugs: Introduction Thyroid hormones, the only known iodine-containing compounds with biological activity, have two important functions. In developing animals and human beings, they are crucial determinants of normal development, especially in the central nervous system (CNS). In the adult, thyroid hormones act to maintain metabolic homeostasis, affecting the function of virtually all organ systems. To meet these requirements, there are large stores of preformed hormone within the thyroid gland. Metabolism of the thyroid hormones occurs primarily in the liver, although local metabolism within certain target tissues, such as the brain, also occurs. Serum concentrations of thyroid hormones are precisely regulated by the pituitary hormone, thyrotropin, in a classic negative-feedback system. The predominant actions of thyroid hormone are mediated via binding to nuclear thyroid hormone receptors and modulating transcription of specific genes. In this regard, thyroid hormones share a common mechanism of action with steroid hormones, vitamin D, and retinoids, whose receptors make up a superfamily of nuclear receptors (Chin and Yen, 1997; seeChapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). In addition, as with steroid hormones, it has become clear that thyroid hormones have diverse nongenomic actions (Davis and Davis, 1997). Disorders of the thyroid are common. They consist of two general presentations: changes in the size or shape of the gland or changes in secretion of hormones from the gland. Thyroid nodules and goiter in the euthyroid patient are the most common endocrinopathies and can be caused by benign and malignant tumors. The presentation of overt hyper- or hypothyroidism often presents the clinician with dramatic clinical manifestations. While the diagnosis may be clinically obvious, subtle presentations require the use of biochemical tests of thyroid function. Screening of the newborn population for congenital hypothyroidism, followed by the institution of appropriate thyroid hormone replacement therapy, has dramatically decreased the incidence of mental retardation and cretinism in the United States. Worldwide, congenital hypothyroidism due to iodine deficiency remains the major preventable cause of mental retardation, although much progress has

been made to eradicate iodine deficiency. Effective treatment of most thyroid disorders is readily available. Treatment of the hypothyroid patient is straightforward and consists of hormone replacement. There are more options for treatment of the hyperthyroid patient, including the use of antithyroid drugs to decrease hormone synthesis and secretion by the gland and destruction of the gland by the administration of radioactive iodine or by surgical removal. Treatment of thyroid disorders in general is extremely satisfying, as most patients can be either cured or have their diseases controlled (seeBraverman and Utiger, 2000; Braverman and Refetoff, 1997). Thyroid The thyroid gland is the source of two fundamentally different types of hormones. The iodothyronine hormones include thyroxine (T4) and 3,5,3'-triiodothyronine (T3); they are essential for normal growth and development and play an important role in energy metabolism. The other known secretory product of the thyroid, calcitonin, is produced by the parafollicular (C) cells and is discussed in Chapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds. History The thyroid gland was first described by Galen and was named "glandulae thyroidaeae" by Wharton in 1656. Harington (1935) reviewed the many older opinions concerning the function of this gland. Wharton thought, for example, that the viscous fluid within the follicles lubricated the trachea. He also believed that the gland was larger in women, to serve a cosmetic function in giving grace to the contour of the neck. Later observers, influenced by the liberal blood supply of the gland, believed that it provided a vascular shunt for the brain. With this function in mind, Rush in 1820 expressed the belief that the larger size of the gland in women was "necessary to guard the female system from the influence of the more numerous causes of irritation and vexation of mind to which they are exposed than the male sex." However, Hofrichter opposed this theory in the same year by pointing out that "If it were indeed true that the thyroid contains more blood at some times than at others, this effect would be visible to the naked eye; in this case women would certainly have long ceased to go about with bare necks, for husbands would have learned to recognize the swelling of this gland as a danger signal of threatening trouble from their better halves." The thyroid was first recognized as an organ of importance when enlargement was observed to be associated with changes in the eyes and the heart in the condition we now call hyperthyroidism. It is of interest that this condition, the manifestations of which on occasion can be as striking as any in medicine, escaped description until Parry saw his first case in 1786. Parry's account was not published until 1825 and was followed in 1835 and 1840 by those of Graves and Basedow, whose names became applied to the disorder. In 1874, Gull first associated atrophy of the gland with the symptoms now known to be characteristic of thyroid deficiency and hypofunction of the thyroid, hypothyroidism, in adults was known as Gull's disease. The term myxedema was applied to the clinical syndrome in 1878 by Ord in the belief that the characteristic thickening of the subcutaneous tissues was due to excessive formation of mucus. Extirpation experiments to elucidate the function of the thyroid were at first misinterpreted because of the simultaneous removal of the parathyroids. However, the pioneer research in the late 19th century on the latter organs by Gley allowed the functional differentiation of these two endocrine glands. It was not until after calcitonin was discovered in 1961 that it was realized that the thyroid itself also was concerned with the regulation of Ca2+. In 1891, Murray became the first to treat a

case of hypothyroidism by injecting an extract of the thyroid gland; in the following year, Howitz, Mackenzie, and Fox independently discovered that thyroid tissue was fully effective when given by mouth. Magnus-Levy discovered the effect of the thyroid on metabolic rate in 1895; he found that Gull's disease was characterized by a low rate of metabolism and that the administration of thyroid to hypothyroid or normal individuals increased oxygen consumption. Chemistry of Thyroid Hormones The principal hormones of the thyroid gland are the iodine-containing amino acid derivatives of thyronine(T 4 and T 3; Figure 571). Thyroxine was first isolated in crystalline form from a hydrolysate of thyroid by Kendall in 1915; he found that the crystalline product exerted the same physiological effects as the extract from which it was obtained. Eleven years later, the structural formula of thyroxine was elucidated by Harington, and in 1927, Harington and Barger synthesized the hormone. Figure 571. Thyronine, Thyroid Hormones, and Precursors.

Following the isolation and the chemical identification of thyroxine, it was generally believed that all the hormonal activity of thyroid tissue could be accounted for by its content of thyroxine. However, careful studies revealed that crude thyroid preparations possessed greater calorigenic activity than could be accounted for by their thyroxine content. The enigma was resolved with the detection, isolation, and synthesis of triiodothyronine (Gross and Pitt-Rivers, 1952; Roche et al., 1952a,b). Further studies revealed that triiodothyronine is qualitatively similar to thyroxine in its biological action but that it is much more potent on a molar basis (Gross and Pitt-Rivers, 1953a,b). StructureActivity Relationship The stereochemical nature of the thyroid hormones plays an important role in defining hormone activity. A great many structural analogs of thyroxine have been synthesized in order to define the structureactivity relationship, to detect antagonists of thyroid hormones, or to find compounds exhibiting one desirable type of activity while not showing unwanted effects. The only significant success has been the partial separation of the cholesterol-lowering action of thyroxine analogs from their calorigenic or cardiac effects. For example, introduction of specific arylmethyl groups at the 3' position of triiodothyronine resulted in analogs that are liver-selective, cardiac-sparing thyromimetics (Leeson et al., 1989). The D isomer of thyroxine was once used to lower the concentration of cholesterol in plasma, but cardiac side effects resulted in discontinuation of the

clinical uses of this hormone. Thyroid hormone analogs and metabolites offer hope that more useful separation of these activities may yet be achievable. For example, 3,5,3'-triiodothyroacetic acid (triac) has been shown to have less thyromimetic activity in the heart than in other thyroid hormoneresponsive tissues (Liang et al., 1997; Sherman and Ladenson, 1992). The structural requirements for a significant degree of thyroid hormone activity have been defined (seeJorgensen, 1964; Cody, 2000; Wagner et al., 1995). The 3'-monosubstituted compounds are more active than the 3',5'-disubstituted molecules. Thus, triiodothyronine is five times more potent than thyroxine, while 3'-isopropyl-3,5-diiodothyronine has seven times the activity. Although the chemical nature of the 3, 5, 3', and 5' substituents is important, their effects on the conformation of the molecule are even more so. In thyronine, the two rings are angulated at about 120 at the ether oxygen and are free to rotate on their axes. As depicted schematically in Figure 572, when the 3,5 iodines are in place, rotation of the two rings is somewhat restricted, and they tend to take up positions perpendicular to one another. While not potent, even halogen-free derivatives possess some activity if they have the proper conformation. In general, the affinity of iodothyronines for the thyroid hormone receptor parallels their biological potency (Chin and Yen, 1997; Anderson et al., 2000), but additional factors including affinity for plasma proteins, rate of entry into cell nuclei, and rate of metabolism can affect therapeutic potency. Figure 572. Structural Formula of 3,5-Diiodothyronine, Drawn to Show the Conformation in Which the Planes of the Aromatic Rings Are Perpendicular to Each Other. (Adapted fromJorgensen, 1964. See alsoCody, 2000.)

Recent structureactivity correlations indicate that certain plant flavonoids that are long-standing folk remedies can exhibit antihormonal properties, including inhibition of the enzyme that catalyzes 5' (outer, or tyrosyl ring) deiodination of T4 (type I iodothyronine 5'-deiodinase; Cody, 2000). These compounds are also potent competitors of thyroxine binding to transthyretin. Computer graphic modeling suggests that the best structural homology between thyroid hormones and flavonoids involves their respective phenolic rings. Synthesis of Thyroid Hormones The synthesis of the thyroid hormones is unique, complex, and seemingly grossly inefficient. The thyroid hormones are synthesized and stored as amino acid residues of thyroglobulin, a protein constituting the vast majority of the thyroid follicular colloid. The thyroid gland is unique in storing great quantities of potential hormone in this way, and extracellular thyroglobulin can represent a large portion of the mass of the gland. Thyroglobulin is a complex glycoprotein made up of two apparently identical subunits, each with a molecular mass of 330,000 daltons. Interestingly, molecular cloning has revealed that thyroglobulin belongs to a superfamily of serine hydrolases, including acetylcholinesterase (seeChapter 8: Anticholinesterase Agents).

The major steps in the synthesis, storage, release, and interconversion of thyroid hormones are the following: (1) the uptake of iodide ion by the gland, (2) the oxidation of iodide and the iodination of tyrosyl groups of thyroglobulin, (3) coupling of iodotyrosine residues by ether linkage to generate the iodothyronines, (4) the proteolysis of thyroglobulin and the release of thyroxine and triiodothyronine into the blood, and (5) the conversion of thyroxine to triiodothyronine in peripheral tissues as well as in the thyroid. These processes are summarized in Figure 573. Figure 573. Major Pathways of Thyroid Hormone Biosynthesis and Release. Abbreviations are as follows: Tg, thyroglobulin; DIT, diiodotyrosine; MIT, monoiodotyrosine; TPO, thyroid peroxidase; HOI, hypoiodous acid; EOI, enzyme-linked species; PTU, propylthiouracil; MMI, methimazole; ECF, extracellular fluid. (Adapted from Taurog, 2000, with permission.)

Uptake of Iodide Iodine ingested in the diet reaches the circulation in the form of iodide. Under normal circumstances, its concentration in the blood is very low (0.2 to 0.4 g/dl; about 15 to 30 nM), but

the thyroid efficiently and actively transports the ion via a specific, membrane-bound protein, termed the sodium-iodide symporter (NIS) (Eskandari et al., 1997; Dai et al., 1996; Smanik et al., 1996). As a result, the ratio of thyroid to plasma iodide concentration is usually between 20 and 50 and can far exceed 100 when the gland is stimulated. The iodide transport mechanism is inhibited by a number of ions such as thiocyanate and perchlorate (Figure 573). The transport system is stimulated by thyrotropin [thyroid-stimulating hormone (TSH); see below] and also is controlled by an autoregulatory mechanism. Thus, decreased stores of thyroid iodine enhance iodide uptake, and the administration of iodide can reverse this situation. If the further metabolism of iodide is blocked by antithyroid drugs, the iodide-concentrating mechanism (NIS) can be more easily studied. Thus isolated, NIS has been identified in many other tissues, including the salivary glands, gastric mucosa, midportion of the small intestine, choroid plexus, skin, mammary gland, and perhaps the placenta, all of which maintain a concentration of iodide greater than that of the blood (Carrasco, 2000). It has been suggested that the accumulation of iodide by the placenta and the mammary gland may be of importance in providing adequate supplies for the fetus and infant, but no obvious purpose is served by the accumulation of iodide at the other sites. It is evident that NIS in the thyroid is not unique to the gland and does not account for the specific function of synthesizing thyroid hormone. Oxidation and Iodination Consistent with the conditions generally necessary for halogenation of aromatic rings, the iodination of tyrosine residues requires the iodinating species to be in a higher state of oxidation than is the anion. The exact nature of the iodinating species was uncertain for many years. However, Magnusson and coworkers (1984) have provided convincing evidence that it is hypoiodate, either as hypoiodous acid (HOI) or as an enzyme-linked species (EOI). The oxidation of iodide to its active form is accomplished by thyroid peroxidase, a heme-containing enzyme that utilizes hydrogen peroxide (H2O2) as the oxidant (Taurog, 2000; Magnusson et al., 1987). Human thyroid peroxidase has been cloned and identified as an autoantigen in autoimmune thyroid disease (McLachlan and Rapoport, 1992). The peroxidase is membrane-bound and appears to be concentrated at or near the apical surface of the thyroid cell. The reaction results in the formation of monoiodotyrosyl and diiodotyrosyl residues in thyroglobulin just prior to its storage in the lumen of the thyroid follicle. It is thought that the formation of the H2O2 that serves as a substrate for the peroxidase occurs in close proximity to its site of utilization and involves the oxidation of reduced nicotinamide adenine dinucleotide phosphate (NADPH). An increase in the generation of H2O2 may be an important facet of the mechanism by which TSH stimulates the organification of iodide in thyroid cells. This hypothesis has arisen from observations that TSH stimulates the synthesis of inositol trisphosphate and elevates cytosolic concentrations of Ca2+ in thyroid follicular cells (Corda et al., 1985; Field et al., 1987; Laurent et al., 1987). The formation of H2O2 is stimulated by a rise in cytosolic Ca2+ (Takasu et al., 1987). Formation of Thyroxine and Triiodothyronine from Iodotyrosines The remaining synthetic step is the coupling of two diiodotyrosyl residues to form thyroxine or of monoiodotyrosyl and diiodotyrosyl residues to form triiodothyronine. These also are oxidative reactions and appear to be catalyzed by the same peroxidase discussed above. The mechanism involves the enzymatic transfer of groups, perhaps as iodotyrosyl free radicals or positively charged ions, within thyroglobulin. Although many other proteins can serve as substrates for the peroxidase, none is as efficient as thyroglobulin in yielding thyroxine. The configuration of the protein is thus presumed to be important in facilitating this coupling reaction. Thyroxine formation occurs

primarily at a location near the amino terminus of the protein, while most of the triiodotyrosine is synthesized near the carboxy terminus (Dunn et al., 1987). The relative rates of synthetic activity at the various sites depend on the concentration of TSH and the availability of iodide. This may account, at least in part, for the long-known relationship between the proportion of thyroxine and triiodothyronine formed in the thyroid and the availability of iodide or the relative quantities of the two iodotyrosines. For example, when there is a deficiency of iodine in rat thyroid, the ratio of thyroxine to triiodothyronine decreases from 4:1 to 1:3 (Greer et al., 1968). Because triiodothyronine is at least five times as active as thyroxine and contains only three-fourths as much iodine, a decrease in the quantity of available iodine need have little impact on the effective amount of thyroid hormone elaborated by the gland. Although a decrease in the availability of iodide and the associated increase in the proportion of monoiodotyrosine favor the formation of triiodothyronine over thyroxine, a deficiency in diiodotyrosine ultimately can impair the formation of both forms of the hormone. In addition to the coupling reaction, intrathyroidal and secreted triiodothyronine is generated by the 5'-deiodination of thyroxine (Chanoine et al., 1993). Secretion of Thyroid Hormones Since thyroxine and triiodothyronine are synthesized and stored within thyroglobulin, proteolysis is an important part of the secretory process. This process is initiated by endocytosis of colloid from the follicular lumen at the apical surface of the cell. This "ingested" thyroglobulin appears as intracellular colloid droplets, which apparently then fuse with lysosomes containing the requisite proteolytic enzymes. It is generally believed that thyroglobulin must be completely broken down into its constituent amino acids for the hormones to be released. As the molecular mass of thyroglobulin is 660,000 daltons, and the protein is made up of about 300 carbohydrate residues and 5500 amino acid residues, only two to five of which are thyroxine, this is an extravagant process. TSH appears to enhance the degradation of thyroglobulin by increasing the activity of several thiol endopeptidases of the lysosomes (Dunn and Dunn, 1988). The endopeptidases selectively cleave thyroglobulin, yielding hormone-containing intermediates that are subsequently processed by exopeptidases (Dunn and Dunn, 2000). The liberated hormones then exit the cell, presumably at its basal membrane. When thyroglobulin is hydrolyzed, monoiodotyrosine and diiodotyrosine also are liberated, but they usually do not leave the thyroid. Instead, they are selectively metabolized, and the iodine, liberated in the form of iodide, is reincorporated into protein. Normally, all this iodide is reused; however, when proteolysis is activated intensely by TSH, some of the iodide reaches the circulation, at times accompanied by trace amounts of the iodotyrosines. Conversion of Thyroxine to Triiodothyronine in Peripheral Tissues The normal daily production of thyroxine has been estimated to range between 70 and 90 g, while that of triiodothyronine is between 15 and 30 g. Although triiodothyronine is secreted by the thyroid, metabolism of thyroxine by sequential monodeiodination in the peripheral tissues accounts for about 80% of circulating triiodothyronine (Figure 574). Removal of the 5'-, or outer ring, iodine leads to the formation of triiodothyronine and is the "activating" metabolic pathway. The major site of conversion of thyroxine to triiodothyronine outside the thyroid is the liver. Thus, when thyroxine is given to hypothyroid patients in doses that produce normal concentrations of thyroxine in plasma, the plasma concentration of triiodothyronine also reaches the normal range (Braverman et al., 1970). Most peripheral target tissues utilize triiodothyronine that is derived from the circulating hormone. Notable exceptions are the brain and pituitary, for which local generation of triiodothyronine is a major source for the intracellular hormone. Removal of the iodine on position 5 of the inner ring produces the metabolically inactive 3,3',5'-triiodothyronine (reverse T3, rT3; Figure 571). Under normal conditions, about 41% of thyroxine is converted to triiodothyronine, about 38% is converted to reverse T3, and about 21% is metabolized via other pathways, such as

conjugation in the liver and excretion in the bile. Normal circulating concentrations of thyroxine in plasma range from 4.5 to 11.0 g/dl, while those of triiodothyronine are about 100-fold less (60 to 180 ng/dl). Figure 574. Pathways of Iodothyronine Deiodination.

The enzyme responsible for the conversion of thyroxine to triiodothyronine is iodothyronine 5'deiodinase, which exists as two distinct isozymes that are differentially expressed and regulated in peripheral tissues (Figure 575; Leonard and Visser, 1986). Type I 5'-deiodinase (D1) is found in the liver, kidney, and thyroid and generates circulating triiodothyronine that is utilized by most peripheral target tissues. Although 5'-deiodination is the major function of this isozyme, D1 also catalyzes 5-deiodination. D1 is inhibited by a variety of factors (Table 571), including the antithyroid drug propylthiouracil. The decreased plasma triiodothyronine concentrations observed in nonthyroidal illnesses are a result of inhibition of D1 (Farwell, 1999) and decreased entrance of thyroxine into cells. D1 is "up-regulated" in hyperthyroidism and "down-regulated" in hypothyroidism. The cloning of D1 has identified the enzyme as a selenoprotein and demonstrated the presence of a selenocystine at the active site (Berry et al., 1991; Berry and Larsen, 1992). Type II 5'-deiodinase (D2) is distributed in the brain, pituitary, skeletal and cardiac muscle, and, in the rat, brown fat. It functions primarily to supply intracellular triiodothyronine to these tissues (Visser et al., 1982; Bartha et al., 2000). D2 has a much lower K m for thyroxine than does D1 (nM vs. M K m values), and its activity is unaffected by propylthiouracil. D2 is dynamically regulated by its substrate, thyroxine, such that elevated levels of the enzyme are found in hypothyroidism and suppressed levels are found in hyperthyroidism (Leonard et al., 1981; Leonard and Koehrle, 2000). Thus, D2 appears to autoregulate the intracellular supply of triiodothyronine in the brain and pituitary. A D2-like selenoprotein cDNA has been cloned from frog skin (Davey et al., 1995) and

mammalian sources (Croteau et al., 1996; Salvatore et al., 1996), leading to the proposal that the deiodinases belong to a family of selenoproteins (St. Germain and Galton, 1997). However, to date, no native, full-length seleno-D2 translation product has been found in any mammalian tissue despite abundant seleno-D2 gene product (Leonard et al., 1999). Further, the 29,000-dalton substratebinding subunit of rat D2 has been cloned and shown to be a nonselenoprotein (Leonard et al., 2000). Thus, the confirmation of D2 as a selenoprotein is unresolved. Figure 575. Deiodinase Isozymes. Abbreviations are as follows: D1, type I iodothyronine 5'-deiodinase; D2, type II iodothyronine 5'-deiodinase; D3, type III iodothyronine 5-deiodinase; BAT, brown adipose tissue.

Inner ring- or 5-deiodination, a main inactivating pathway for T3, is catalyzed by type III deiodinase (D3), which is found in placenta, skin, and brain. Cloning of D3, like that of D1, has identified the enzyme as a selenoprotein (Croteau et al., 1995). Transport of Thyroid Hormones in the Blood Iodine in the circulation is normally present in several forms, with 95% as organic iodine and approximately 5% as iodide. Most of the organic iodine is thyroxine (90% to 95%), while triiodothyronine represents a relatively minor fraction (about 5%). The thyroid hormones are transported in the blood in strong but noncovalent association with certain plasma proteins. Thyroxine-binding globulin (TBG) is the major carrier of thyroid hormones. It is an acidic glycoprotein with a molecular mass of approximately 63,000 daltons, and it binds one molecule of thyroxine per molecule of protein with a very high affinity (the equilibration association constant, Ka is about 1010 M1). Triiodothyronine is bound less avidly. Thyroxine, but not triiodothyronine, also is bound by transthyretin (formally called thyroxine-binding prealbumin). This protein is present in higher concentration than is TBG and primarily binds thyroxine with an equilibrium association constant near 107 M1. Transthyretin has four apparently identical subunits but only a single high-affinity binding site. Albumin also can serve as a carrier for thyroxine when the more avid carriers are saturated. It is difficult, however, to estimate its quantitative or physiological importance, with the exception of the syndrome known as familial dysalbuminemic hyperthyroxinemia. This is an autosomal dominant hereditary disorder characterized by the increased affinity of albumin for thyroxine (Ruiz et al., 1982) due to a point mutation in the albumin gene (Tang et al., 1999; Sunthornthepvarakul et al., 1994). Thyroxine binds also to the apolipoproteins of the high density lipoproteins, HDL2 and HDL3, the significance of which is unclear at present (Benevenga et al., 1992).

Binding of thyroid hormones to plasma proteins protects the hormones from metabolism and excretion, resulting in their long half-lives in the circulation. The free (unbound) hormone is a small percentage (about 0.03% of thyroxine and about 0.3% of triiodothyronine) of the total hormone in plasma (Larsen et al., 1981). The differential binding affinity for serum proteins also is reflected in the 10- to 100-fold difference in circulating hormone concentrations and half-lives of thyroxine and triiodothyronine. Essential to understanding the regulation of thyroid function is the "free hormone" concept: only the unbound hormone has metabolic activity (Mendel, 1989). Thus, because of the high degree of binding of thyroid hormones to plasma proteins, changes in either the concentrations of these proteins or the binding affinity of the hormones for the proteins would have major effects on the total serum hormone levels. Certain drugs and a variety of pathological and physiological conditions, such as the changes in circulating concentrations of estrogens during pregnancy or during the administration of oral estrogens, can alter both the binding of thyroid hormones to plasma proteins and the amounts of these proteins (Table 572). However, since the pituitary responds to and regulates circulating free hormone levels, minimal changes in free hormone concentrations are seen. Laboratory tests that measure only total hormone levels, therefore, can be subject to misinterpretation. Appropriate tests of thyroid function are discussed later in this chapter. Degradation and Excretion (Figure 576.) Thyroxine is eliminated slowly from the body, with a half-life of 6 to 8 days. In hyperthyroidism, the half-life is shortened to 3 or 4 days, whereas in hypothyroidism it may be 9 to 10 days. These changes presumably are due to altered rates of metabolism of the hormone. In conditions associated with increased binding to TBG, such as pregnancy, clearance is retarded. The increase in TBG is due to the estrogen-induced increase in the sialic acid content of the synthesized TBG resulting in decreased TBG clearance (Ain et al., 1987). The reverse is observed when there is reduced protein binding of thyroid hormones or when binding to protein is inhibited by certain drugs (Table 572). Triiodothyronine, which is less avidly bound to protein, has a half-life of approximately 1 day. Figure 576. Pathways of Metabolism of Thyroxine (T4) and Triiodothyronine (T3). Abbreviations are as follows: DIT, diiodotyrosine; MIT, monoiodotyrosine; T4S, T4 sulfate; T4G, T 4 glucuronide; T3S, T3 sulfate; T3G, T3 glucuronide; T4K, T4 pyruvic acid; T3K, T3 pyruvic acid; Tetrac, tetraiodothyroacetic acid; Triac, triiodothyroacetic acid.

The liver is the major site of nondeiodinative degradation of thyroid hormones; thyroxine and triiodothyronine are conjugated with glucuronic and sulfuric acids through the phenolic hydroxyl group and excreted in the bile. There is an enterohepatic circulation of the thyroid hormones; they are liberated by hydrolysis of the conjugates in the intestine and reabsorbed. A portion of the conjugated material reaches the colon unchanged, is hydrolyzed there, and is eliminated in feces as the free compounds. As discussed above, the major route of metabolism of thyroxine is deiodination to either triiodothyronine or reverse T3. Triiodothyronine and reverse T3 are deiodinated to three different diiodothyronines, which are further deiodinated to two monoiodothyronines (seeFigure 574), inactive metabolites that are normal constituents of human plasma. Additional metabolites (monoiodotyrosine and diiodotyrosine) in which the diphenyl ether linkage is cleaved have been detected both in vitro and in vivo. Regulation of Thyroid Function During the past century, it was appreciated that cellular changes occur in the anterior pituitary in association with endemic goiter or following thyroidectomy. The classical experimental observations of Cushing (1912) and the clinical observations of Simmonds (1914) established that ablation or disease of the pituitary causes thyroid hypoplasia. It eventually was determined that thyrotropes of the anterior pituitary secrete thyrotropin, or TSH. TSH is a glycoprotein hormone with and subunits analogous to those of the gonadotropins. Its structure is discussed with those of other glycoprotein hormones in Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors. Although there was evidence that thyroid hormone or lack of it causes cellular changes in the pituitary, the control of secretion of TSH by the negative-feedback action of thyroid hormone was not appreciated fully until its central role in the pathogenesis of goiter was elucidated in the early 1940s. TSH is secreted in a pulsatile manner and circadian pattern, its levels in the circulation being highest during sleep at night. It is now recognized that the rate of secretion of TSH is delicately controlled by the hypothalamic peptide thyrotropin-releasing hormone (TRH) and the quantity of free thyroid hormones in the circulation. If extra thyroid hormone is given, transcription of both the TRH gene (Wilbur and Xu, 1998) and the thyrotropin gene is decreased (seeSamuels et al., 1988), the secretion of TSH is suppressed, and the thyroid becomes inactive and regresses. Any decrease in the normal rate of secretion of the thyroid evokes an enhanced secretion of TSH in an attempt to stimulate the thyroid to secrete more hormone. Additional mechanisms of the effect of thyroid hormone on TSH secretion appear to be a reduction in TRH secretion by the hypothalamus and a reduction in the number of receptors for TRH on pituitary cells. Thyrotropin-Releasing Hormone (TRH) TRH stimulates the release of preformed TSH from secretory granules and also stimulates the subsequent synthesis of both and subunits of TSH. Somatostatin, dopamine, and pharmacological doses of glucocorticoids inhibit TRH-stimulated TSH secretion. TRH is a tripeptide with both terminal amino and carboxyl groups blocked (L-pyroglutamyl-Lhistidyl-L-proline amide). The mature hormone is derived from a precursor protein that contains six copies of the tripeptide flanked by dibasic residues. TRH is synthesized by the hypothalamus and is released into the hypophysioportal circulation, where it is brought into contact with TRH receptors on thyrotropes. The binding of TRH to its receptor, a G protein-coupled receptor, elicits stimulation of the hydrolysis of polyphosphatidylinositols and activation of protein kinase C (Gershenghorn,

1986). Ultimately, TRH stimulates the synthesis and release of TSH by the thyrotroph. TRH also has been localized in the CNS in regions of the cerebral cortex, circumventricular structures, neurohypophysis, pineal gland, and spinal cord. These findings, as well as its localization in nerve endings, suggest that TRH may act as a neurotransmitter or neuromodulator outside of the hypothalamus. Administration of TRH to animals produces CNS-mediated effects on behavior, thermoregulation, autonomic tone, and cardiovascular function, including increases in blood pressure and heart rate. TRH also has been identified in pancreatic islet cells, heart, testis, and in certain parts of the gastrointestinal tract. Its physiological role there is not known. TRH has been administered both intravenously and intrathecally as a therapeutic agent in refractory depression (Callahan et al., 1997; Marangell et al., 1997). Actions of TSH on the Thyroid When TSH is given to experimental animals, the first effect on thyroid hormone metabolism that can be measured is increased secretion, which can be seen within minutes. All phases of hormone synthesis and release are eventually stimulated: iodide uptake and organification, hormone synthesis, endocytosis, and proteolysis of colloid. There is increased vascularity of the gland and hypertrophy and hyperplasia of thyroid cells. These effects follow the binding of TSH to its receptor on the plasma membrane of thyroid cells. The TSH receptor is a member of the family of G protein-coupled receptors and is structurally similar to the receptors for luteinizing hormone (LH) and follicle-stimulating hormone (FSH) (seeChapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors; Parmentier et al., 1989; Vassart and Dumont, 1992; Nagayama and Rapoport, 1992). These receptors share significant amino acid sequences and have large extracellular domains that are involved in binding of hormone. When TSH binds to its receptor, adenylyl cyclase is stimulated and cyclic AMP levels in the cells increase. At higher concentrations than are required to stimulate cyclic AMP formation, TSH causes activation of phospholipase C, with a resultant hydrolysis of polyphosphatidylinositols, increased cytoplasmic Ca2+, and activation of protein kinase C (Manley et al., 1988; Van Sande et al., 1990). Both the adenylyl cyclase and the phospholipase C signaling pathways appear to mediate effects of TSH on thyroid function in human beings, although the adenylyl cyclase pathway may be the sole mediating pathway in other species (seeVassart and Dumont, 1992). Multiple mutations of the TSH receptor resulting in clinical thyroid dysfunction have been described (Tonacchera et al., 1996b). Germline mutations have presented as congenital, nonautoimmune hypothyroidism (Kopp et al., 1995) and as autosomal, dominant toxic thyroid hyperplasia (Tonacchera et al., 1996a). Somatic mutations that result in constitutive activation of the receptor are recognized as probable causes of hyperfunctioning thyroid adenomas (Paschke et al., 1994b). Finally, resistance to TSH also has been described both in families with mutant TSH receptors (Sunthornthepvarakui et al., 1995) and in those with no apparent mutations in either the TSH receptor or in TSH itself (Xie et al., 1997). Relation of Iodine to Thyroid Function Normal thyroid function obviously requires an adequate intake of iodine; without it, normal amounts of hormone cannot be made, TSH is secreted in excess, and the thyroid becomes hyperplastic and hypertrophies. The enlarged and stimulated thyroid becomes remarkably efficient at extracting the residual traces of iodide from the blood. The iodide-concentrating mechanism develops a gradient for the ion that may be ten times normal, and in mild to moderate iodine

deficiency, the thyroid usually succeeds in producing sufficient hormone. Adult hypothyroidism and cretinism may occur in more severe iodine deficiency. In some areas of the world, simple or nontoxic goiter is prevalent because dietary iodine is not sufficient (Delange et al., 1993). Significant regions of iodine deficiency are present in Central and South America, Africa, Europe, southeast Asia, and China. The daily requirement for iodine in adults is 1 to 2 g/kg body weight. The United States recommended daily allowance for iodine is in the range of 40 to 120 g for children and 150 g for adults, with the addition of 25 g and 50 g recommended during pregnancy and lactation, respectively. Vegetables, meat, and poultry contain minimal amounts of iodine, whereas dairy products and fish are relatively high in iodine content (Table 573; Braverman, 1997). Potable water usually contains negligible amounts of iodine. Iodine has been used empirically for the treatment of iodine-deficiency goiter for 150 years. However, its modern use was the outgrowth of the extensive studies of Marine, which culminated in the use of iodine to prevent goiter in school children in Akron, Ohio, a region where endemic iodine deficiency goiter was prevalent (Marine and Kimball, 1917). The success of these experiments led to the adoption of iodine prophylaxis and therapy in many regions throughout the world where iodine-deficiency goiter is endemic. The most practicable method for providing small supplements of iodine for large segments of the population is the addition of iodide or iodate to table salt; iodate is now preferred. In some countries, the use of iodized salt is required by law; in others, including the United States, the use is optional. In the United States, iodized salt provides 100 g of iodine per gram. However, while the United States population remains iodine-sufficient, iodine intake has steadily decreased over the last twenty years, a trend that needs to be monitored (Hollowell et al., 1998). Other vehicles for supplying iodine to large populations who are iodine-deficient include oral or intramuscular injection of iodized oil (Elnagar et al., 1995), iodized drinking water supplies, iodized irrigation systems (Cao et al., 1994b), and iodized animal feed. Actions of Thyroid Hormones Whereas the precise biochemical mechanisms through which thyroid hormones exert their developmental and tissue-specific effects are only beginning to be understood, the concept that most of the actions of thyroid hormones are mediated by nuclear receptors has been well accepted since the mid-1980s (for review, seeChin and Yen, 1997; Anderson et al., 2000). In this model, triiodothyronine binds to high-affinity nuclear receptors, which then bind to a specific DNA sequence (thyroid hormone response element) in the promoter/regulatory region of specific genes. In this fashion, triiodothyronine modulates gene transcription and, ultimately, protein synthesis. In general, the receptor without hormone is bound to the thyroid response element in the basal state. Typically, this results in repressed gene transcription, although there are some examples of constitutive gene activation. Binding by triiodothyronine may activate gene transcription by releasing the repression. Hormone-associated receptors also may have direct activation or repressive actions. Thyroxine also binds to these receptors, but it does so with a much lower affinity than triiodothyronine. Despite its ability to bind to nuclear receptors, thyroxine has not been shown to alter gene transcription. Thus, it is likely that thyroxine serves principally as a "prohormone," with essentially all actions of thyroid hormone at the transcriptional level being caused by triiodothyronine. Nuclear thyroid hormone receptors were cloned in 1986 by several laboratories (Weinberger et al., 1986; Sap et al., 1986). They were discovered to be the cellular homologs of an avian retroviral oncoprotein, denoted c-erb A. There is considerable homology between the thyroid hormone

receptors and the steroid nuclear receptors, and together they make up a gene superfamily that also includes the retinoic acid and vitamin D nuclear receptors (seeChapters 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect and 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds; Mangelsdorf et al., 1994). The thyroid hormone receptors are derived from two genes, c-erb A (TR ) and c-erb A (TR ), with multiple isoforms identified (Figure 577; Lazar, 1993). TR 1 and TR 1 are found in virtually all tissues that respond to thyroid hormone, whereas the other isoforms exhibit a more tissue-specific distribution. TR 2, for example, is expressed solely in the anterior pituitary. c-erb A 2, an isoform that binds to the thyroid response element but does not bind triiodothyronine, is the most abundant isoform in brain (Strait et al., 1990). Another level of complexity in the regulation of thyroid hormone action at the transcriptional level has been added with the identification of coactivators (Takeshita et al., 1996) and corepressors (Chen and Evans, 1995; Hrlein et al., 1995) that are associated with the T 3receptor complex and serve as mediators of hormone action (Lee and Yen, 1999). Resistance to thyroid hormone has been described in patients with mutations in the TR gene (Brucker-Davis et al., 1995; Adams et al., 1994) and in patients with defective cofactors (Weiss et al., 1996). Figure 577. Thyroid Hormone Receptor Isoforms. The percent of amino acid identity in the DNA binding region is indicated. Identical patterns in the hypervariable and ligand binding regions indicate 100% homology. Three thyroid hormone receptor (TR) isoforms bind thyroid hormone (TR 1, TR 2, and TR 1); c-erb A 2 does not.

Further insight into the mechanisms of thyroid hormone action has been provided by the development of transgenic mice lacking one or more of the thyroid hormone receptor isoforms. Multiple variations of these knockout mice have demonstrated abnormalities in the auditory system, the thyroid-pituitary axis, the heart, the skeletal system and the small intestine (Forrest et al., 1996; Forrest et al., 1990; Fraichard et al., 1997; Wikstrm et al., 1998). Interestingly, despite the recognition of thyroid hormone as an essential regulatory factor during brain development (Oppenheimer and Schwartz, 1997), no obvious abnormalities in brain development have been reported in either single-receptor knockout mice (Hsu and Brent, 1998) or in the recently reported transgenic mice devoid of all known thyroid hormone receptors (Gthe et al., 1999). In addition to nuclear receptor-mediated actions, there are several well-characterized, nongenomic actions of thyroid hormones (Davis and Davis, 1997), including those occurring at the level of the plasma membrane (Davis et al., 1989) and on the cellular cytoarchitecture (Farwell et al., 1990; Siegrist-Kaiser et al., 1990; Farwell and Leonard, 1997). In addition, there are well-characterized thyroid hormone binding sites on the mitochondria (Sterling, 1989). In several of these processes, thyroxine is the hormone that produces the response. Previously, the overall contribution of

nongenomic actions to the general mechanism of thyroid hormone action was considered to be minor. However, this concept, at least as it applies to some species, may need to be reassessed in light of the paucity of abnormalities described in transgenic knockout mice, especially during brain development. Growth and Development As discussed above, it is generally believed that the thyroid hormones exert most of their effects through control of DNA transcription and, ultimately, protein synthesis. This is certainly true for the effects of the hormones on the normal growth and development of the organism. Perhaps the most dramatic example is found in the tadpole, which is almost magically transformed into a frog by triiodothyronine. Not only does the animal grow limbs, lungs, and other terrestrial accoutrements, but the hormone also stimulates the synthesis of a host of enzymes and so influences the tail that it is digested away and used to build new tissue elsewhere. Thyroid hormone plays a critical role in brain development (Bernal and Nunez, 1995; Oppenheimer and Schwartz, 1997; Hendrich, 1997). The appearance of functional, chromatin-bound receptors for thyroid hormone coincides with neurogenesis in the brain (Strait et al., 1990). The absence of thyroid hormone during the period of active neurogenesis (up to 6 months postpartum) leads to irreversible mental retardation (cretinism) and is accompanied by multiple morphological alterations in the brain (Legrand, 1979). These severe morphological alterations result from disturbed neuronal migration, deranged axonal projections, and decreased synaptogenesis. Thyroid hormone supplementation during the first 2 weeks of life prevents the development of these disturbed morphological changes. Myelin basic protein, a major component of myelin, is the product of a specific gene that is regulated by thyroid hormone during development (Farsetti et al., 1991). Decreased expression of myelin basic protein results in defective myelinization in the hypothyroid brain. The appearance of laminin, an extracellular matrix protein that provides key guidance signals to migrating neurons, is delayed and the content is diminished in the developing cerebellum of the hypothyroid rat (Farwell and Dubord-Tomasetti, 1999). Altered expression of laminin is likely to alter neuronal migration and lead to the morphological abnormalities observed in the cretinous brain. Several other brainspecific genes have been reported to be developmentally regulated by thyroid hormone (Bernal and Nunez, 1995). A common characteristic of many of these proteins is that their expression appears to be merely delayed in the hypothyroid animal; normal levels are eventually achieved in the adult. The actions of thyroid hormones on protein synthesis and enzymatic activity are certainly not limited to the brain, and a large number of tissues are affected by the administration of thyroid hormone or by its deficiency. The extensive defects in growth and development that are found in cretins provide a vivid reminder of the pervasive effects of thyroid hormones in normal individuals. Cretinism is usually classified as endemic or sporadic. Endemic cretinism is encountered in regions of endemic goiter and is usually caused by extreme deficiency of iodine. Goiter may or may not be present. Sporadic cretinism is a consequence of failure of the thyroid to develop normally or the result of a defect in the synthesis of thyroid hormone. Goiter is present if a synthetic defect is at fault. While detectable at birth, cretinism often is not recognized until 3 to 5 months of age. When untreated, the condition eventually leads to such gross changes as to be unmistakable. The child is dwarfed, with short extremities, and is mentally retarded, inactive, uncomplaining, and listless. The face is puffy and expressionless, and the enlarged tongue may protrude through the thickened lips of

the half-opened mouth. The skin may have a yellowish hue and feel doughy, and it is dry and cool to the touch. The heart rate is slow, the body temperature may be low, closure of the fontanels is delayed, and the teeth erupt late. Appetite is poor, feeding is slow and interrupted by choking, constipation is frequent, and there may be an umbilical hernia. For treatment to be fully effective, the diagnosis must be made long before these obvious changes have come about. In regions of endemic cretinism due to iodine deficiency, iodine replacement is best instituted prior to pregnancy. However, iodine replacement given to pregnant women up to the end of the second trimester has been shown to enhance the neurologic and psychological development of the children (Cao et al., 1994a). Screening of newborn infants for deficient function of the thyroid is carried out in the United States and in most industrialized countries. Concentrations of TSH and thyroxine are measured in blood from the umbilical cord or from a heel stick. The incidence of congenital dysfunction of the thyroid is about 1 per 4000 births (Fisher, 1991). Calorigenic Effects A characteristic response of homeothermic animals to thyroid hormone is increased oxygen consumption. Most peripheral tissues contribute to this response; heart, skeletal muscle, liver, and kidney are stimulated markedly by thyroid hormone. Indeed, 30% to 40% of the thyroid hormone dependent increase in oxygen consumption can be attributed to stimulation of cardiac contractility. Several organs, including brain, gonads, and spleen, are unresponsive to the calorigenic effects of thyroid hormone. The mechanism of the calorigenic effect of thyroid hormone has been elusive (Silva, 1995). At one time, it was erroneously believed that thyroid hormone uncoupled mitochondrial oxidative phosphorylation. Thyroid hormonedependent lipogenesis may constitute a quantitatively important energy sink, and studies in rats have demonstrated that about 4% of the increased caloric expenditure induced by thyroid hormone is accounted for by lipogenesis. A link between lipogenesis and thermogenesis is the stimulation of lipolysis by triiodothyronine. Further, thyroid hormone induces expression of several lipogenic enzymes, including malic enzyme and fatty acid synthetase. Although the entire picture is not clear, there appears to be an integrated thyroid hormone response program for regulating the set-point of energy expenditure and maintaining the metabolic machinery necessary to sustain it. Indeed, even small changes in Lthyroxine replacement doses may significantly alter the set-point for resting energy expenditure in the hypothyroid patient (al-Adsani et al., 1997). Cardiovascular Effects Thyroid hormone influences cardiac function by direct and indirect actions; changes in the cardiovascular system are prominent clinical consequences in thyroid dysfunctional states. In hyperthyroidism, there is tachycardia, increased stroke volume, increased cardiac index, cardiac hypertrophy, decreased peripheral vascular resistance, and increased pulse pressure. In hypothyroidism, there is bradycardia, decreased cardiac index, pericardial effusion, increased peripheral vascular resistance, decreased pulse pressure, and elevations of mean arterial pressure. (For a review of the effects of thyroid hormone on the heart, seeBraverman et al., 1994.) Thyroid hormones play a direct role in regulating myocardial gene expression. Triiodothyronine regulates genes encoding the isoforms of the sarcomeric myosin heavy chains by increasing the expression of the gene and decreasing the expression of the gene (Everett et al., 1986). A thyroid hormone response element has been located in the 5' upstream region of the myosin heavy chain gene. Triiodothyronine also upregulates the gene encoding myosin Ca2+ATPase, which plays a critical role in myocardial contraction (Rohrer and Dillman, 1988). Regulation of these two genes results in the changes in contractility observed in hyper- and hypothyroidism. Indeed, stress

echocardiography in hyperthyroid patients revealed abnormalities in cardiac contractility that reverted to normal when euthyroidism was restored (Kahaly et al., 1999). Similarly, left ventricular diastolic dysfunction in hypothyroidism was reversed with L-thyroxine replacement therapy (Biondi et al., 1999). Observations in transgenic mice have provided insight into the action of thyroid hormone on heart rate. Previously, alterations in the sensitivity of the cardiac myocyte to catecholamines (enhanced in hyperthyroidism and depressed in hypothyroidism) were considered an indirect effect of thyroid hormone, possibly due to changes in expression of myocardial -adrenergic receptors. This is the basis for the use of -adrenergic receptor antagonists in relieving some of the cardiac manifestations in hyperthyroidism. However, basal heart rate is decreased in mice lacking the TR 1 gene (Johansson et al., 1998) and is increased in mice lacking TR (Johansson et al., 1999), suggesting a more direct role for thyroid hormone in cardiac pacemaking. Finally, T3 leads to hemodynamic alterations in the periphery that result in alterations in the chronotropic and inotropic state of the myocardium. Interestingly, T3 appears to have a direct, nongenomic vasodilatory effect on vascular smooth muscle (Park et al., 1997, Ojamaa et al., 1996). Metabolic Effects Thyroid hormones stimulate metabolism of cholesterol to bile acids, and hypercholesterolemia is a characteristic feature of hypothyroid states. Thyroid hormones have been shown to increase the specific binding of low-density lipoprotein (LDL) by liver cells (Salter et al., 1988), and the concentration of hepatic receptors for LDL is decreased in hypothyroidism (Scarabottolo et al., 1986; Gross et al., 1987). The number of LDL receptors available on the surface of hepatocytes is a strong determinant of the plasma cholesterol concentration (seeChapter 36: Drug Therapy for Hypercholesterolemia and Dyslipidemia). Thyroid hormones enhance the lipolytic responses of fat cells to other hormones, for example, catecholamines, and elevated plasma free fatty acid concentrations are seen in hyperthyroidism. In contrast to other lipolytic hormones, thyroid hormones do not directly increase the accumulation of cyclic AMP. They may, however, regulate the capacity of other hormones to enhance the accumulation of the cyclic nucleotide by decreasing the activity of a microsomal phosphodiesterase that hydrolyzes cyclic AMP (Nunez and Correze, 1981). There also is evidence that thyroid hormones act to maintain normal coupling of the -adrenergic receptor to the catalytic subunit of adenylyl cyclase in fat cells. Fat cells from hypothyroid rats have increased concentrations of guanine nucleotidebinding regulatory proteins (G proteins) that mediate the inhibitory control of adenylyl cyclase (seeChapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). This can account for both the decreased response to lipolytic hormones and the increased sensitivity to inhibitory regulators, such as adenosine, that are found in hypothyroidism (Ros et al., 1988). Thyrotoxicosis is an insulin-resistant state (Gottlieb and Braverman, 1994). Postreceptor defects in the liver and peripheral tissues, manifested by depleted glycogen stores and enhanced glucogenesis, lead to insulin insensitivity. In addition, there is increased absorption of glucose from the gut. Compensatory increases in insulin secretion result in order to maintain euglycemia. This may result in the "unmasking" of clinical diabetes in previously undiagnosed patients and an increase in the insulin requirements of diabetic patients already on insulin. Hypothyroidism results in decreased absorption of glucose from the gut and decreased insulin secretion. Peripheral glucose uptake also is slowed in hypothyroidism, although glucose utilization by the brain is unaffected. Insulin requirements are decreased in the hypothyroid patient with diabetes.

Thyroid Hyperfunction Thyrotoxicosis is a condition caused by elevated concentrations of circulating free thyroid hormones. Various disorders of different etiologies can result in this syndrome. The term hyperthyroidism is restricted to those conditions in which thyroid hormone production and release are increased due to gland hyperfunction. Iodine uptake by the thyroid gland is increased, as determined by the measurement of the percent uptake of 123I or 131I in a 24-hour radioactive iodine uptake (RAIU) test. In contrast, thyroid inflammation or destruction resulting in excess "leak" of thyroid hormones or excess exogeneous thyroid hormone intake results in a low 24-hour RAIU. The term subclinical hyperthyroidism is defined as few if any symptoms with a low serum TSH and normal concentrations of T4 and T3. Graves' disease, or toxic diffuse goiter, is the most common cause of high RAIU thyrotoxicosis. It accounts for 60% to 90% of the cases, depending upon age and geographic region. Graves' disease is an autoimmune disorder characterized by hyperthyroidism, diffuse goiter, and IgG antibodies that bind to and activate the TSH receptor (Burman and Baker, 1985; Bottazzo and Doniach, 1986). This is a relatively common disorder, with an incidence of 0.02% to 0.4% in the United States. Endemic areas of iodine deficiency have a lower incidence of autoimmune thyroid disease. As with most types of thyroid dysfunction, women are affected more than men, with a ratio ranging from 5:1 to 7:1. Graves' disease is more common between the ages of 20 and 50, but may occur at any age. HLA B8 and DR3 haplotypes are associated with Graves' disease in Caucasians. Graves' disease is commonly associated with other autoimmune diseases. The characteristic exophthalmos associated with Graves' disease is an infiltrative ophthalmopathy and is considered an autoimmune-mediated inflammation of the periorbital connective tissue and extraocular muscles. This disorder is clinically evident with various degrees of severity in about 50% of patients with Graves' disease, but it is present on radiological studies, such as ultrasound or CT scan, in almost all patients. The pathogenesis of Graves' ophthalmopathy, including the role of the TSH receptor present in retroorbital tissues (Paschke et al., 1994) and the management of this disorder, is reviewed in a recently published monograph on Graves' disease (Rapoport and McLachlan, 2000). Toxic uninodular and multinodular goiter accounts for 10% to 40% of cases of hyperthyroidism and is more common in older patients. Infiltrative ophthalmopathy is absent. A low RAIU is seen in the destructive thyroiditides and in thyrotoxicosis resulting from exogenous thyroid hormone ingestion. Low RAIU thyrotoxicosis caused by subacute (painful) and silent (painless or lymphocytic) thyroiditis represents about 5% to 20% of all cases. Silent thyroiditis occurs in 7% to 10% of postpartum women in the United States (Emerson and Farwell, 2000). Other causes of thyrotoxicosis are much less common. Most of the signs and symptoms of thyrotoxicosis stem from the excessive production of heat and from increased motor activity and increased activity of the sympathetic nervous system. The skin is flushed, warm, and moist; the muscles are weak and tremulous; the heart rate is rapid, the heartbeat is forceful, and the arterial pulses are prominent and bounding. The increased expenditure of energy gives rise to increased appetite and, if intake is insufficient, to loss of weight. There also may be insomnia, difficulty in remaining still, anxiety and apprehension, intolerance to heat, and increased frequency of bowel movements. Angina, arrhythmias, and heart failure may be present in older patients. Some individuals may show extensive muscular wasting as a result of thyroid myopathy. Patients with long-standing undiagnosed or undertreated thyrotoxicosis may develop osteoporosis due to increased bone turnover (Baran, 2000).

Thyroid Hypofunction Hypothyroidism, known as myxedema when severe, is the most common disorder of thyroid function. Worldwide, hypothyroidism is most often the result of iodine deficiency. In nonendemic areas, where iodine is sufficient, chronic autoimmune thyroiditis (Hashimoto's thyroiditis) accounts for the majority of cases. This disorder is characterized by high levels of circulating antibodies directed against thyroid peroxidase and, less commonly, thyroglobulin. In addition, blocking antibodies directed at the TSH receptor may be present, exacerbating the hypothyroidism (Botero and Brown, 1998). Finally, a cause of thyroid destruction may be apoptotic cell death due to the interaction of Fas and Fas ligand in the thyrocytes (Giordano et al., 1997). Failure of the thyroid to produce sufficient thyroid hormone is the most common cause of hypothyroidism and is referred to as primary hypothyroidism. Central hypothyroidism occurs much less often and results from diminished stimulation of the thyroid by TSH because of pituitary failure (secondary hypothyroidism) or hypothalamic failure (tertiary hypothyroidism). Hypothyroidism present at birth is known as congenital hypothyroidism and is the most common preventable cause of mental retardation in the world. Diagnosis and early intervention with thyroid hormone replacement prevent the development of cretinism, as discussed above. Nongoitrous hypothyroidism is associated with degeneration and atrophy of the thyroid gland. The same condition follows surgical removal of the thyroid or its destruction by radioactive iodine. Since it also may occur years after antithyroid drug therapy for Graves' disease, some have speculated that hypothyroidism can be the end stage of this disorder ("burnt-out" Graves' disease). Goitrous hypothyroidism occurs in Hashimoto's thyroiditis and when there is a severe defect in synthesis of thyroid hormone. When the disease is mild, it may be subtle in its presentation. By the time it has become severe, however, all of the signs are overt. The appearance of the patient is pathognomonic. The face is quite expressionless, puffy, and pallid. The skin is cold and dry, the scalp is scaly, and the hair is coarse, brittle, and sparse. The fingernails are thickened and brittle, the subcutaneous tissue appears to be thickened, and there may be true edema. The voice is husky and low-pitched, speech is slow, hearing is often faulty, mentation is impaired, and depression may be present. The appetite is poor, gastrointestinal activity is diminished, and constipation is common. Atony of the bladder is rare and suggests that the function of other smooth muscles may be impaired. The voluntary muscles are weak and the relaxation phase of the deep-tendon reflexes is delayed. The heart can be dilated, and there is frequently a pericardial effusion, although this is rarely clinically significant. There also may be pleural effusions and ascites. Anemia, most commonly normochromic, normocytic, is often present, although menstrual irregularity with menorrhagia may result in iron deficiency anemia. Hyperlipidemia often is present in hypothyroid patients. Patients are lethargic and tend to sleep a lot and often complain of cold intolerance. Thyroid Function Tests The development of radioimmunoassays and, more recently, chemiluminescent and enzyme-linked immunoassays for T4, T3, and TSH have greatly improved the laboratory diagnosis of thyroid disorders (Nelson and Wilcox, 1996; Klee, 1996; Spencer et al., 1996). However, measurement of the total hormone concentration in plasma may not give an accurate picture of the activity of the thyroid gland. The total hormone concentration changes with alterations in either the amount of TBG or the binding affinity for hormones to TBG in plasma. Although equilibrium dialysis of undiluted serum and radioimmunoassay for free thyroxine in the dialysate represent the gold standard for determining free thyroxine concentrations, this assay is typically not available in routine clinical laboratories (Nelson and Tomei, 1988). The free thyroxine index is an estimation of the free thyroxine concentration and is calculated by multiplying the total thyroxine concentration by the thyroid hormone binding ratio, which estimates the degree of saturation of TBG (Nelson and

Tomei, 1989). Additional assays commonly in use for estimating the free T4 and free T3 concentrations employ labeled analogs of these iodothyronines in chemiluminescence and enzymelinked immunoassays (Klee, 1996; Nelson and Wilcox, 1996). These assays correlate well with free T4 concentrations measured by the more cumbersome equilibrium dialysis method and are easily adaptable to routine clinical laboratory use. However, the analog assays may be affected by a wide variety of nonthyroidal disease states, including acute illness, and by certain drugs to a greater degree than are the free T4 index and free T4 determined by equilibrium dialysis. Estimates of free thyroxine levels should be complemented with serum measurements of TSH. In fact, in individuals whose pituitary function and TSH secretion are normal, serum measurement of TSH is the thyroid function test of choice (Danese et al., 1996; Helfand and Redfern, 1998), because pituitary secretion of TSH is sensitively regulated in response to circulating concentrations of thyroid hormones. Serum measurements of TSH have been available since 1965. The first assays were single antibody radioimmunoassays and remained the standard for 20 years. These assays were useful only for diagnosing primary hypothyroidism, as a lower limit of the normal range could not be reliably measured. The first "sensitive" TSH assay was developed in 1985, utilizing a dual-antibody approach. Application of this method resulted in the expansion of the assay detection limit below the normal range. Thus, any assay of this type is referred to as a sensitive TSH assay (Nicoloff and Spencer, 1990). A major use of the sensitive TSH assay is to differentiate between normal and thyrotoxic patients, who should exhibit suppressed TSH values. Indeed, the sensitive TSH assay has essentially replaced evaluation of the response of TSH to injection of synthetic TRH (TRH stimulation test) in the thyrotoxic patient. While the serum TSH assay is extremely useful in determining the euthyroid state and titrating the replacement dose of thyroid hormone in patients with primary hypothyroidism, abnormal serum TSH concentrations may not always indicate thyroid dysfunction. In such patients, assessment of the circulating thyroid hormone levels will further determine whether or not thyroid dysfunction is truly present. Synthetic preparations of TRH (protirelin, THYREL) are available for the evaluation of pituitary or hypothalamic failure as a cause of secondary hypothyroidism. Recombinant human TSH (thyrotropin alfa, THYROGEN) is now available as an injectable preparation to test the ability of thyroid tissue, both normal and malignant, to take up radioactive iodine and release thyroglobulin (Haugen et al., 1999). This preparation replaces bovine TSH (THYTROPAR ), which was associated with a high incidence of side effects, including anaphylaxis. Therapeutic Uses of Thyroid Hormone The major indications for the therapeutic use of thyroid hormone are for hormone replacement therapy in patients with hypothyroidism or cretinism and for TSH suppression therapy in patients with nontoxic goiter or after treatment for thyroid cancer (Roti et al., 1993; Toft, 1994). While the general consensus has been that thyroid hormone therapy is not indicated for treatment of the "low T4 syndrome" ("sick euthyroid syndrome") that is a result of nonthyroidal illness (Brent and Hershman, 1986, Farwell, 1999), this concept has been challenged recently with the suggestion that severely ill patients may benefit by treatment with T3 (DeGroot, 1999). However, there is no evidence for this recommendation based on published studies, and this suggestion remains a minority opinion. For example, T3 treatment does not decrease mortality in the sick euthyroid syndrome that occurs in patients undergoing coronary artery bypass surgery (Klemperer et al., 1995). The synthetic preparations of the sodium salts of the natural isomers of the thyroid hormones are

available and widely used for thyroid hormone therapy. Levothyroxine sodium (L-T4, SYNTHROID, LEVOXYL, LEVOTHROID, others) is available in tablets and as a lyophilized powder for injection. Liothyronine sodium (L-T3) is the salt of triiodothyronine and is available in tablets (CYTOMEL) and in an injectable form (TRIOSTAT). A mixture of thyroxine and triiodothyronine is marketed as liotrix (THYROLAR ). Desiccated thyroid preparations, derived from whole animal thyroids, contain both thyroxine and triiodothyronine and have highly variable biologic activity, making these preparations much less desirable. Thyroid Hormone Replacement Therapy Thyroxine (levothyroxine sodium) is the hormone of choice for thyroid hormone replacement therapy because of its consistent potency and prolonged duration of action. The absorption of thyroxine occurs in the small intestine and is variable and incomplete, with 50% to 80% of the dose absorbed (Hays, 1991; Hays and Nielson, 1994). Absorption is slightly increased when the hormone is taken on an empty stomach. In addition, certain drugs may interfere with absorption of levothyroxine in the gut, including sucralfate, cholestyramine resin, iron and calcium supplements, and aluminum hydroxide. Enhanced biliary excretion of levothyroxine occurs during the administration of drugs that induce hepatic cytochrome P450 enzymes, such as phenytoin, carbamazepine, and rifampin. This enhanced excretion may necessitate an increase in the dose of orally administered levothyroxine. Triiodothyronine (liothyronine sodium) may be used occasionally when a quicker onset of action is desired, as, for example, in the rare presentation of myxedema coma or for preparing a patient for 131I therapy for treatment of thyroid cancer. It is less desirable for chronic replacement therapy because of the requirement for more frequent dosing, higher cost, and transient elevations of serum triiodothyronine concentrations above the normal range. Combination therapy with levothyroxine and liothyronine has been suggested for use in hypothyroid patients that remain symptomatic on levothyroxine alone and have serum TSH concentrations in the normal range (Bunevicius et al., 1999). However, a definite benefit for this combination therapy has not yet been shown. Furthermore, this combination may lead to transient elevations of circulating T3 concentrations in contrast to the steady levels of T3 during levothyroxine administration due to conversion of T4 to T 3 in peripheral tissues. The average daily adult replacement dose of levothyroxine sodium in a 68-kg person is 112 g as a single dose, while that of liothyronine sodium is 50 to 75 g in divided doses. Institution of therapy in healthy younger individuals can begin at full replacement doses. Because of the prolonged halflife of thyroxine (7 days), new steady-state concentrations of the hormone will not be achieved until 4 to 6 weeks after a change in dose. Thus, reevaluation with determination of serum TSH concentration need not be performed at intervals less than 4 to 6 weeks. The goal of thyroxine replacement therapy is to achieve a TSH value in the normal range, as overreplacement of thyroxine, suppressing TSH values to the subnormal range, may induce osteoporosis and cause cardiac dysfunction (Ross, 1991). In noncompliant, young patients, the cumulative weekly doses of levothyroxine may be given as a single weekly dose, which is safe, effective, and well tolerated (Grebe et al., 1997). In individuals over the age of 60, institution of therapy at a lower daily dose of levothyroxine sodium (25 g per day) is indicated to avoid exacerbation of underlying and undiagnosed cardiac disease. Death due to arrhythmias has been reported during the initiation of thyroid hormone replacement therapy in hypothyroid patients. The dose can be increased at a rate of 25 g per day every few months until the TSH is normalized. For individuals with preexisting cardiac disease, an initial dose of 12.5 g per day, with increases of 12.5 to 25 g per day every 6 to 8 weeks, is indicated. Daily doses of thyroxine may be interrupted periodically because of intercurrent medical or surgical illnesses that prohibit taking medications by mouth. A lapse of several days of hormone replacement is unlikely to have any significant metabolic consequences. However, if more prolonged interruption in oral therapy is necessary, levothyroxine may be given

parenterally at a dose 25% to 50% less than the patient's daily oral requirements. Subclinical hypothyroidism is an asymptomatic state characterized by elevated serum TSH concentrations and serum T4 and T3 concentrations in the normal range (for review, seeSurks and Ocampo, 1996). Population screening has shown that subclinical hypothyroidism is very common, with a prevalence of up to 15% in some populations (Canaris et al., 2000; Tunbridge et al., 1977) and up to 25% in the elderly (Samuels, 1998). The decision to use levothyroxine therapy in these patients to normalize the serum TSH must be made on an individual basis, as treatment may not be appropriate for all patients. However, a recent report strongly suggests that untreated subclinical hypothyroidism is associated with an increased prevalence of aortic atherosclerosis and myocardial infarction (Hak et al., 2000). Patients with subclinical hypothyroidism who are likely to benefit from levothyroxine therapy include those with goiter, autoimmune thyroid disease, hypercholesterolemia, cognitive dysfunction, or pregnancy (see below), and those patients who have symptoms of hypothyroidism. The dose of levothyroxine in the hypothyroid patient who becomes pregnant often needs to be increased, perhaps due to the increased serum concentrations of TBG induced by estrogen (Kaplan, 1992; Glinoer, 1993; Mandel et al., 1990). In addition, pregnancy may "unmask" hypothyroidism in patients with preexisting autoimmune thyroid disease or in those who reside in a region of iodine deficiency (Glinoer et al., 1994). Overt hypothyroidism during pregnancy is associated with fetal distress (Wasserstrum and Anaia, 1995) and impaired psychoneural development in the progeny (Man et al., 1991). Recent studies have suggested that subclinical hypothyroidism during pregnancy also is associated with mildly impaired psychomotor development in the children (Haddow et al., 1999; Pop et al., 1999). These findings strongly suggest that any degree of hypothyroidism, as judged by an elevated serum TSH, should be treated during pregnancy. Thus, serum TSH values should be determined in the first trimester in all patients with preexisting hypothyroidism, as well as in those at high risk for developing hypothyroidism. Therapy with levothyroxine should be administered to keep the serum TSH in the normal range. Any adjustment of the levothyroxine dose should be reevaluated in 4 to 6 weeks to determine if further adjustments are necessary. Comparative Responses to Thyroid Preparations There is no significant difference in the qualitative response of the patient with myxedema to triiodothyronine, thyroxine, or desiccated thyroid. However, there are obvious quantitative differences. Following the subcutaneous administration of a large experimental dose of triiodothyronine, a metabolic response can be detected within 4 to 6 hours, at which time the skin becomes detectably warmer and the pulse rate and temperature increase. With this dose, a 40% decrease in metabolic rate can be restored to normal in 24 hours. The maximal response occurs in 2 days or less, and the effects subside with a half-time of about 8 days. The same single dose of thyroxine exerts much less effect. However, if thyroxine is given in approximately four times the dose of triiodothyronine, a comparable elevation in metabolic rate can be achieved. The peak effect of a single dose is evident in about 9 days, and this declines to half the maximum in 11 to 15 days. In both cases the effects outlast the presence of detectable amounts of hormone; these disappear from the blood with mean half-lives of approximately 1 day for triiodothyronine and 7 days for thyroxine. Myxedema Coma Myxedema coma is a rare syndrome that represents the extreme expression of severe, long-standing hypothyroidism (Emerson, 1999). It is a medical emergency, and even with early diagnosis and treatment, the mortality rate can be as high as 60%. Myxedema coma occurs most often in elderly

patients during the winter months. Common precipitating factors include pulmonary infections, cerebrovascular accidents, and congestive heart failure. The clinical course of lethargy proceeding to stupor and then coma is often hastened by drugs, especially sedatives, narcotics, antidepressants, and tranquilizers. Indeed, many cases of myxedema coma have occurred in hypothyroid patients who have been hospitalized for other medical problems. Cardinal features of myxedema coma are: (1) hypothermia, which may be profound, (2) respiratory depression, and (3) unconsciousness. Other clinical features include bradycardia, macroglossia, delayed reflexes, and dry, rough skin. Dilutional hyponatremia is common and may be severe. Elevated plasma creatine phosphokinase (CPK) and lactate dehydrogenase (LDH) concentrations, acidosis, and anemia are common findings. Lumbar puncture reveals increased opening pressure and high protein content. Hypothyroidism is confirmed by measuring serum free thyroxine index and TSH values. Ultimately, myxedema coma is a clinical diagnosis. The mainstay of therapy is supportive care, with ventilatory support, rewarming with blankets, correction of hyponatremia, and treatment of the precipitating incident. Because of a 5% to 10% incidence of coexisting decreased adrenal reserve in patients with myxedema coma, intravenous steroids are indicated before initiating thyroxine therapy. Parenteral administration of thyroid hormone is necessary due to uncertain absorption through the gut. With intravenous preparations of both levothyroxine and liothyronine now available, a reasonable approach is an initial intravenous loading dose of 200 to 300 g of levothyroxine with a second dose of 100 g given 24 hours later. Alternatively, a bolus of 500 g levothyroxine given orally (by mouth or via nasogastric tube) may be administered to patients <50 years old without cardiac complications (Yamamoto et al., 1999). Simultaneously with the initial dose of levothyroxine, some clinicians recommend adding liothyronine at a dose of 10 g intravenously every 8 hours until the patient is stable and conscious. The dose of thyroid hormone should be adjusted on the basis of hemodynamic stability, the presence of coexisting cardiac disease, and the degree of electrolyte imbalance. Recent studies suggest that over-aggressive treatment with either levothyroxine (>500 g per day) or liothyronine (>75 g) may be associated with an increased mortality (Yamamoto et al., 1999). Treatment of Cretinism Success in the treatment of cretinism depends upon the age at which therapy is started. Because of this, newborn screening for congenital hypothyroidism is routine in the United States, Canada, and many other countries around the world. In cases that do not come to the attention of physicians until retardation of development is clinically obvious, the detrimental effects of thyroid hormone deficiency on mental development will not be overcome. If, on the other hand, therapy is instituted within the first few weeks of life, normal physical and mental development is almost always achieved. Prognosis also depends on the severity of the hypothyroidism at birth and may be worse for babies with thyroid agenesis. The most critical need for thyroid hormone is during the period of myelinization of the central nervous system that occurs about the time of birth. To rapidly normalize the serum thyroxine concentration in the congenitally hypothyroid infant, an initial daily dose of levothyroxine of 10 to 15 g/kg is recommended (Fisher, 1991). This dose will increase the total serum thyroxine concentration to the upper half of the normal range in most infants within 1 to 2 weeks. Individual levothyroxine doses are adjusted at 4- to 6-week intervals during the first 6 months, at 2-month intervals during the 6- to 18-month period, and at 3- to 6-month intervals thereafter to maintain serum thyroxine concentrations in the 10- to 16- g/dl range and serum TSH values in the normal range. The free thyroxine levels should be kept in the upper normal or elevated range. Assessments that are important guides for appropriate hormone replacement include physical growth, motor development, bone maturation, and developmental progress. Management of premature infants with hypothyroxinemia due to the sick euthyroid syndrome ( 50% of those born

at less than 30 weeks of gestation) remains a therapeutic dilemma. Despite impaired psychomotor development in these patients (Reuss et al., 1996; Den Ouden et al., 1996), levothyroxine therapy has not been shown to be beneficial and may be deleterious if overreplacement is administered (van Wassenaer et al., 1997). Nodular Thyroid Disease Nodular thyroid disease is the most common endocrinopathy. The prevalence of clinically apparent nodules is 4% to 7% in the United States, with the frequency increasing throughout adult life. When ultrasound and autopsy data are included, the prevalence of thyroid nodules approaches 50% by age 60. As with other forms of thyroid disease, nodules are more frequent in women. Nodules have been estimated to develop at a rate of 0.1% per year. In individuals exposed to ionizing radiation, the rate of nodule development is 20-fold higher. While the presence of a nodule raises the question of a malignancy, only 8% to 10% of patients with thyroid nodules have thyroid cancer. About 12,000 new cases of thyroid cancer are diagnosed annually, with about 1000 deaths from the disease per year. However, many more people have clinically silent thyroid cancer, as up to 35% of thyroids removed at autopsy or at surgery harbor a small (<1 cm) occult papillary cancer. The evaluation of the patient with nodular thyroid disease includes a careful physical examination, biochemical analysis of thyroid function, and assessment of the malignant potential of the nodule (Mazzaferri, 1993; Gharib and Goellner, 1993). The latter may include examination of a fine-needle aspiration biopsy of the nodule, ultrasound evaluation, and radioisotope scanning with 123I or 131I to determine if a particular nodule is functioning. Fine-needle aspiration biopsy of the nodule is now the most definitive approach to diagnose the pathology of a nodule. TSH suppressive therapy with levothyroxine is an option for the patient diagnosed with a benign solitary nodule and a normal serum TSH. The rationale behind levothyroxine therapy is that the benign nodule will either stop growing or decrease in size after TSH stimulation of the thyroid gland has been suppressed. The success rate of such therapy varies widely (Papini et al., 1998; Zelmanovitz et al., 1998; Gharib and Mazzaferri, 1998). Identification of those patients who are most likely to benefit from thyroid hormone therapy can be achieved through measurement of the serum TSH concentration and radioisotope scanning. Suppression therapy will be of no value if thyroid nodule autonomy exists, as evidenced by a subnormal TSH value and all isotope uptake in the nodule. Functioning nodules are the most likely to respond to suppression therapy. However, once TSH concentrations are suppressed, a repeat radioisotope scan (suppression scan) should be obtained. If significant uptake persists on a suppression scan, the nodule is nonsuppressible and levothyroxine therapy should be discontinued. Suppression therapy needs to be considered carefully in older patients or in those with coronary artery disease; in general, such therapy should be avoided in these patients. Hypofunctioning nodules are much less likely to respond to suppression therapy. However, a 6- to 12-month trial of levothyroxine suppression is reasonable (Hermus and Huysmans, 1998). If levothyroxine is administered, therapy should be continued for as long as the nodule is decreasing in size. Once the size of a nodule remains stable for a 6- to 12-month period, therapy may be discontinued and the nodule observed for recurrent growth. Any nodule that grows while on suppression therapy should be rebiopsied and/or surgically excised. Antithyroid Drugs and Other Thyroid Inhibitors A large number of compounds are capable of interfering, directly or indirectly, with the synthesis, release, or action of thyroid hormones (Table 574). Several are of great clinical value for the temporary or extended control of hyperthyroid states. These are discussed in detail below, while others are primarily of research or toxicological interest and are only mentioned briefly. The major inhibitors may be classified into four categories: (1) antithyroid drugs, which interfere directly with

the synthesis of thyroid hormones; (2) ionic inhibitors, which block the iodide transport mechanism; (3) high concentrations of iodine itself, which decrease release of thyroid hormones from the gland and also may decrease hormone synthesis; and (4) radioactive iodine, which damages the gland with ionizing radiation. Adjuvant therapy with drugs that have no specific effects on thyroid gland hormonogenesis is useful in controlling the peripheral manifestations of thyrotoxicosis. These drugs include inhibitors of the peripheral deiodination of thyroxine to the active hormone, triiodothyronine; -adrenergic receptor antagonists; and Ca 2+ channel blockers. The antithyroid drugs have been reviewed by Cooper (1998). Adrenergic receptor antagonists are discussed more fully in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists and Ca 2+ channel blockers in Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia and 35: Antiarrhythmic Drugs. Antithyroid Drugs The antithyroid drugs that have clinical utility are the thioureylenes, which belong to the family of thionamides. Propylthiouracil may be considered as the prototype. History Studies on the mechanism of the development of goiter began with the observation that rabbits fed a diet composed largely of cabbage often developed goiters. This result was probably due to the presence of precursors of the thiocyanate ion in cabbage leaves (see below). Later, two pure compounds were shown to produce goiter, sulfaguanidine and phenylthiourea. Investigation of the effects of thiourea derivatives revealed that rats became hypothyroid despite hyperplastic changes in their thyroid glands that were characteristic of intense thyrotropic stimulation. After treatment was begun, no new hormone was made, and the goitrogen had no visible effect upon the thyroid gland following hypophysectomy or the administration of thyroid hormone. This suggested that the goiter was a compensatory change resulting from the induced state of hypothyroidism and that the primary action of the compounds was to inhibit the formation of thyroid hormone (Astwood, 1945). The therapeutic possibilities of such agents in hyperthyroidism were evident, and the substances so used became known as antithyroid drugs. StructureActivity Relationship The two goitrogens found in the early 1940s proved to be prototypes of two different classes of antithyroid drugs. These two, with one later addition, made up three general categories into which the majority of the agents can be assigned: (1) thioureylenes include all the compounds currently used clinically (Figure 578); (2) aniline derivatives, of which the sulfonamides make up the largest number, embrace a few substances that have been found to inhibit thyroid hormone synthesis; and (3) polyhydric phenols, such as resorcinol, which have caused goiter in human beings when applied to the abraded skin. A few other compounds, mentioned briefly below, do not fit into any of these categories. Figure 578. Antithyroid Drugs of the Thiamide Type.

Thiourea and its simpler aliphatic derivatives and heterocyclic compounds containing a thioureylene group make up the majority of the known antithyroid agents that are effective in human beings. Although most of them incorporate the entire thioureylene group, in some a nitrogen atom is replaced by oxygen or sulfur so that only the thioamide group is common to all. Among the heterocyclic compounds, active representatives are the sulfur derivatives of imidazole, oxazole, hydantoin, thiazole, thiadiazole, uracil, and barbituric acid.
L-5-Vinyl-2-thiooxazolidone

(goitrin) is responsible for the goiter that results from consuming turnips or the seeds or green parts of cruciferous plants. These plants are eaten by cows, and the compound is found in cow's milk in areas of endemic goiter in Finland; it is about as active as propylthiouracil in human beings. As the result of industrial exposure, toxicological studies, or clinical trials for various purposes, several other compounds have been noted to possess antithyroid activity (De Rosa et al., 1998). Thiopental and oral hypoglycemic drugs of the sulfonylurea class have weak antithyroid action in experimental animals. This is not significant at usual doses in human beings. However, antithyroid effects in human beings have been observed from dimercaprol, aminoglutethimide, and lithium salts. Polychlorinated biphenyls bear a striking structural resemblance to the thyroid hormones and may function as either agonists or antagonists of thyroid hormone action (De Rosa et al., 1998). Altered circulating concentrations of thyroid hormones and thyrotropin and impaired brain development have been attributed to exposure to polychlorinated biphenyls (Porterfield and Hendry, 1998; Sher et al., 1998). Amiodarone, the iodinerich drug used in the management of cardiac arrhythmias, has complex effects on thyroid function (Harjai and Licata, 1997). In areas of iodine sufficiency, amiodarone-induced hypothyroidism due to the excess iodine is not uncommon, whereas in iodine-deficient regions, amiodarone-induced thyrotoxicosis predominates, whether because of the excess iodine or the thyroiditis induced by the drug. Amiodarone and its major metabolite, desethylamiodarone, are potent inhibitors of iodothyronine deiodination, resulting in decreased conversion of thyroxine to triiodothyronine. In addition, desethylamiodarone decreases binding of triiodothyronine to its nuclear receptors. Recommendations recently have been made as to screening methods to identify chemicals that may alter thyroid hormone action or homeostasis (DeVito et al., 1999). Mechanism of Action The mechanism of action of the thiourylene drugs has been thoroughly discussed by Taurog (2000). Antithyroid drugs inhibit the formation of thyroid hormones by interfering with the incorporation of iodine into tyrosyl residues of thyroglobulin; they also inhibit the coupling of these iodotyrosyl residues to form iodothyronines. This implies that they interfere with the oxidation of iodide ion and iodotyrosyl groups. Taurog (2000) proposed that the drugs inhibit the peroxidase enzyme, thereby preventing oxidation of iodide or iodotyrosyl groups to the required active state. The antithyroid drugs bind to and inactivate the peroxidase only when the heme of the enzyme is in the oxidized state. Over a period of time, the inhibition of hormone synthesis results in the depletion of

stores of iodinated thyroglobulin as the protein is hydrolyzed and the hormones are released into the circulation. Only when the preformed hormone is depleted and the concentrations of circulating thyroid hormones begin to decline do clinical effects become noticeable. There is some evidence that the coupling reaction may be more sensitive to an antithyroid drug, such as propylthiouracil, than is the iodination reaction (Taurog, 2000). This may explain why patients with hyperthyroidism respond well to doses of the drug that only partially suppress organification. When Graves' disease is treated with antithyroid drugs, the concentration of thyroid-stimulating immunoglobulins in the circulation often decreases. This has prompted some to propose that these agents act as immunosuppressants. Burman and Baker (1985) point out that perchlorate, which acts by an entirely different mechanism, also decreases thyroid-stimulating immunoglobulins, suggesting that improvement in hyperthyroidism may, itself, favorably affect the abnormal humoral immune state. In addition to blocking hormone synthesis, propylthiouracil inhibits the peripheral deiodination of thyroxine to triiodothyronine. Methimazole does not have this effect and can antagonize the inhibitory effect of propylthiouracil. Although the quantitative significance of this inhibition has not been established, it does provide a theoretical rationale for the choice of propylthiouracil over other antithyroid drugs in the treatment of severe hyperthyroid states or of thyroid storm. In this acute situation, a decreased rate of conversion of circulating thyroxine to triiodothyronine would be beneficial. Absorption, Metabolism, and Excretion The antithyroid compounds currently used in the United States are propylthiouracil (6-npropylthiouracil) and methimazole (1-methyl-2-mercaptoimidazole; TAPAZOLE). In Great Britain and Europe, carbimazole (NEO -MERCAZOLE), a carbethoxy derivative of methimazole, is available, and its antithyroid action is due to its conversion to methimazole after absorption. Some pharmacological properties of propylthiouracil and methimazole are shown in Table 575. Measurements of the course of organification of radioactive iodine by the thyroid show that absorption of effective amounts of propylthiouracil follows within 20 to 30 minutes of an oral dose. They also show that the duration of action of the compounds used clinically is brief. The effect of a dose of 100 mg of propylthiouracil begins to wane in 2 to 3 hours, and even a 500-mg dose is completely inhibitory for only 6 to 8 hours. As little as 0.5 mg of methimazole similarly decreases the organification of radioactive iodine in the thyroid gland, but a single dose of 10 to 25 mg is needed to extend the inhibition to 24 hours. The half-life of propylthiouracil in plasma is about 75 minutes, whereas that of methimazole is 4 to 6 hours. The drugs are concentrated in the thyroid, and methimazole, derived from the metabolism of carbimazole, accumulates after carbimazole is administered. Drugs and metabolites appear largely in the urine. Propylthiouracil and methimazole cross the placenta equally and also can be found in milk; methimazole does so to a greater degree than propylthiouracil. The use of these drugs during pregnancy is discussed below. Untoward Reactions The incidence of side effects from propylthiouracil and methimazole as currently used is relatively

low. The overall incidence as compiled from published cases by early investigators was 3% for propylthiouracil and 7% for methimazole, with 0.44% and 0.12% of cases, respectively, developing the most serious reaction, agranulocytosis (Meyer-Gessner et al., 1994). The development of agranulocytosis with methimazole may be dose-related, but no such relationship exists with propylthiouracil. Further observations have found little, if any, difference in side effects between these two agents and suggest that an incidence of agranulocytosis of approximately 1 in 500 is a maximal figure. Agranulocytosis usually occurs during the first few weeks or months of therapy but may occur later. Because agranulocytosis can develop rapidly, periodic white-cell counts usually are of little help. Patients should immediately report the development of sore throat or fever, which usually heralds the onset of this reaction. Agranulocytosis is reversible upon discontinuation of the offending drug, and the administration of recombinant human granulocyte colony-stimulating factor may hasten recovery (Magner et al., 1994). Mild granulocytopenia, if noted, may be due to thyrotoxicosis or may be the first sign of this dangerous drug reaction. Caution and frequent leukocyte counts are then required. The most common reaction is a mild, occasionally purpuric, urticarial papular rash. It often subsides spontaneously without interrupting treatment, but it sometimes calls for the administration of an antihistamine, corticosteroids, or changing to another drug, because cross-sensitivity is uncommon. Other less frequent complications are pain and stiffness in the joints, paresthesias, headache, nausea, skin pigmentation, and loss of hair. Drug fever, hepatitis, and nephritis are rare, although abnormal liver function tests are not infrequent with higher doses of propylthiouracil. Therapeutic Uses The antithyroid drugs are used in the treatment of hyperthyroidism in the following three ways: (1) as definitive treatment, to control the disorder in anticipation of a spontaneous remission in Graves' disease; (2) in conjunction with radioactive iodine, to hasten recovery while awaiting the effects of radiation; and (3) to control the disorder in preparation for surgical treatment. There is no uniformity of opinion as to which form of treatment is the most desirable (Trring et al., 1996), and this is often influenced by a variety of considerations, as discussed below. The usual starting dose for propylthiouracil is 100 mg every 8 hours or 150 mg every 12 hours. When doses larger than 300 mg daily are needed, further subdivision of the time of administration to every 4 to 6 hours is occasionally helpful. Methimazole is effective when given as a single daily dose because of its relatively long plasma and intrathyroidal half-life, as well as its long duration of action. Failures of response to daily treatment with 300 to 400 mg of propylthiouracil or 30 to 40 mg of methimazole are most commonly due to noncompliance. Delayed responses also are noted in patients with very large goiters or those in whom iodine in any form has been given beforehand. Once euthyroidism is achieved, usually within 12 weeks, the dose of antithyroid drug can be reduced. Response to Treatment Hyperthyroidism may be of two kindsGraves' disease and hyperthyroidism from one or more hyperfunctioning thyroid nodules; whichever the cause, the hyperthyroidism seems to respond to antithyroid drugs in the same way. Improvement in the thyrotoxic state usually is noted within three to six weeks after the initiation of antithyroid drugs. The clinical response is related to the dose of antithyroid drug, the size of the goiter, and pretreatment serum T3 concentrations (Benker et al., 1995). The rate of response is determined by the quantity of stored hormone, the rate of turnover of hormone in the thyroid, the half-life of the hormone in the periphery, and the completeness of the block in synthesis imposed by the dosage given. When large doses are continued, and sometimes

with the usual dose, hypothyroidism may develop as a result of overtreatment. The earliest signs of hypothyroidism call for a reduction in dose; if by chance they have advanced to the point of discomfort, thyroid hormone can be given to hasten recovery. A full dose of levothyroxine can be given. The lower maintenance dose of antithyroid drug discussed above is instituted for continued therapy. Initial reports suggested that concomitant use of levothyroxine therapy along with antithyroid drugs increased rates of remission of Graves' disease in Japan (Hashizume et al., 1991). However, subsequent studies have shown no benefit of combination levothyroxine and methimazole therapy on either remission rates (McIver, 1996; Rittmaster et al., 1998) or on changes in serum concentrations of thyroid-stimulating immunoglobulins (Rittmaster et al., 1996). After treatment is initiated, patients should be examined and thyroid function tests (serum free thyroxine index and total triiodothyronine concentrations) measured every 2 to 4 months. Once euthyroidism is established, follow-up every 4 to 6 months is reasonable. Control of the hyperthyroidism usually is associated with a decrease in goiter size, but if the thyroid enlarges, hypothyroidism probably has been induced. When this occurs, the dose of the antithyroid drug should be significantly decreased and/or levothyroxine can be added once hypothyroidism is confirmed. Remissions The antithyroid drugs have been used in many patients to control the hyperthyroidism of Graves' disease until a remission occurs. Early investigators reported that 50% of patients so treated for one year remained well without further therapy for long periods, perhaps indefinitely. More recent reports have indicated that a much smaller percentage of patients sustain remissions after such treatment (Maugendre et al., 1999; Benker et al., 1998). Increased dietary iodine has been implicated in the latter, less favorable rates. Unfortunately, there is no way of predicting before treatment is begun which patients will eventually achieve a lasting remission and who will relapse. It is clear that a favorable outcome is unlikely when the disorder is of long standing, the thyroid is quite large, and various forms of treatment have failed. To complicate the issue further, remission and eventual hypothyroidism may represent the natural history of Graves' disease. During treatment, a positive sign that a remission may have taken place is a reduction in the size of the goiter. The persistence of goiter often indicates failure, unless the patient becomes hypothyroid. Another favorable indication is continued freedom from all signs of hyperthyroidism when the maintenance dose is small. Finally, a decrease in thyroid-stimulating immunoglobulins, suppression of 123I thyroid uptake when thyroxine or triiodothyronine is given, and a normal serum TSH response to TRH may be helpful in predicting a remission in some patients, although these tests are not routinely carried out. The Therapeutic Choice Because antithyroid drug therapy, radioactive iodine, and subtotal thyroidectomy all are effective treatments for Graves' disease, there is no worldwide consensus among endocrinologists as to the best approach to therapy (Franklyn, 1994; Klein et al., 1994; Trring et al., 1996). Prolonged drug therapy of Graves' disease in anticipation of a remission is most successful in patients with small goiters or mild hyperthyroidism. Those with large goiters or severe disease usually require definitive therapy with either surgery or radioactive iodine (131 I). Radioactive iodine remains the treatment of choice of many endocrinologists in the United States (Soloman et al., 1990). Many

investigators consider coexisting ophthalmopathy to be a relative contraindication for radioactive iodine therapy, since worsening of ophthalmopathy has been reported after radioactive iodine (Bartalena et al., 1998). Others suggest that development of hypothyroidism, regardless of the treatment, is the strongest risk factor for progression of ophthalmopathy (Manso et al., 1998). Depleting the thyroid gland of preformed hormone by treatment with antithyroid drugs is advisable in older patients prior to therapy with radioactive iodine so as to prevent a severe exacerbation of the hyperthyroid state during the subsequent development of radiation thyroiditis. Subtotal thyroidectomy is advocated for Graves' disease in young patients with large goiters, children who are allergic to antithyroid drugs, pregnant women (usually in the second trimester) who are allergic to antithyroid drugs, and patients who prefer surgery over antithyroid drugs or radioactive iodine (Zimmerman, 1999; Mestman, 1997). Radioactive iodine or surgery is indicated for definitive therapy in toxic nodular goiter, since remissions following antithyroid drug therapy do not occur. Thyrotoxicosis in Pregnancy Thyrotoxicosis occurs in about 0.2% of pregnancies and is caused most frequently by Graves' disease. Antithyroid drugs are the treatment of choice; radioactive iodine is clearly contraindicated. Historically, propylthiouracil has been preferred over methimazole because of lower transplacental passage. However, more recent data suggest that either may be used safely in the pregnant patient (Momotani et al., 1997; Mortimer et al., 1997; Mestman, 1997). The antithyroid drug dosage should be minimized in order to keep the serum free thyroxine index in the upper half of the normal range or slightly elevated. As pregnancy progresses, Graves' disease often improves. Indeed, it is not uncommon for patients either to be on very low doses or off all antithyroid drugs completely by the end of pregnancy. Therefore the antithyroid drug dose should be reduced, and maternal thyroid function should be frequently monitored in order to decrease chances of fetal hypothyroidism. Relapse or worsening of Graves' disease is common after delivery, and patients should be monitored closely. Propylthiouracil is the drug of choice in nursing women, since very small amounts of the drug appear in breast milk and do not appear to affect thyroid function in the suckling baby. However, doses of methimazole up to 20 mg daily in nursing mothers have been shown to have no effect on fetal thyroid function (Azizi, 1996). Adjuvant Therapy Several drugs that have no intrinsic antithyroid activity are useful in the symptomatic treatment of thyrotoxicosis. -Adrenergic receptor antagonists (Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists) are effective in antagonizing the catecholaminergic effects of thyrotoxicosis by reducing the tachycardia, tremor, and stare and relieving palpitations, anxiety, and tension. Either propranolol, 20 to 40 mg four times daily, or atenolol, 50 to 100 mg daily, is usually given initially. Propranolol and esmolol can be given intravenously if needed. Propranolol, in addition to its -adrenergic receptor antagonist action, has weak inhibitory effects on peripheral conversion of thyroxine to triiodothyronine. Ca 2+ channel blockers (diltiazem, 60 to 120 mg four times daily) can be used to control tachycardia and decrease the incidence of supraventricular tachyarrhythmias (seeChapter 35: Antiarrhythmic Drugs). These drugs should be discontinued once the patient is euthyroid. Other drugs that are useful in the rapid treatment of the severely thyrotoxic patient are agents that inhibit the peripheral conversion of thyroxine to triiodothyronine. Dexamethasone (0.5 to 1 mg two to four times daily) and the iodinated radiological contrast agents, iopanoic acid (TELEPAQUE, 500 to 1000 mg once daily) and sodium ipodate (ORAGRAF IN, 500 to 1000 mg once daily) are effective in preoperative preparation and should not be used chronically. Sodium ipodate recently has been removed from the United States market. Cholestyramine has been used in severely toxic patients to

bind thyroid hormones in the gut and thus block the enterohepatic circulation of the iodothyronines (Mercardo et al., 1996). Preoperative Preparation To reduce operative morbidity and mortality, patients must be rendered euthyroid prior to subtotal thyroidectomy as definitive treatment for hyperthyroidism. It is possible to bring virtually 100% of patients to a euthyroid state; the operative mortality in these patients in the hands of an experienced thyroid surgeon is extremely low. Prior treatment with antithyroid drugs usually is successful in rendering the patient euthyroid for surgery. Iodide is added to the regimen for 7 to 10 days prior to surgery to decrease the vascularity of the gland, making it less friable and decreasing the difficulties for the surgeon. In the patient who is either allergic to antithyroid drugs or is noncompliant, a euthyroid state usually can be achieved by treatment with iopanoic acid, dexamethasone, and propranolol for 5 to 7 days prior to surgery. All of these drugs should be discontinued after surgery. Thyroid Storm Thyroid storm is an uncommon but life-threatening complication of thyrotoxicosis in which a severe form of the disease is usually precipitated by an intercurrent medical problem (Abend and Braverman, 1999). It occurs in untreated or partially treated thyrotoxic patients. Precipitating factors associated with thyrotoxic crisis include infections, stress, trauma, thyroidal or nonthyroidal surgery, diabetic ketoacidosis, labor, heart disease, and, rarely, radioactive iodine treatment. Clinical features are similar to those of thyrotoxicosis, but more exaggerated. Cardinal features include fever (temperature usually over 38.5C) and tachycardia out of proportion to the fever. Nausea, vomiting, diarrhea, agitation, and confusion are frequent presentations. Coma and death may ensue in up to 20% of patients. Thyroid function abnormalities are similar to those found in uncomplicated hyperthyroidism. Therefore, thyroid storm is primarily a clinical diagnosis. Treatment includes supportive measures such as intravenous fluids, antipyretics, cooling blankets, and sedation. Antithyroid drugs are given in large doses. Propylthiouracil is preferred over methimazole because of its additional action of impairing peripheral conversion of thyroxine to triiodothyronine. The recommended initial dose of propylthiouracil is 200 to 400 mg every 4 hours. Propylthiouracil and methimazole can be administered by nasogastric tube or rectally if necessary. Neither of these preparations is available for parenteral administration in the United States. Iodides, orally or intravenously, are used after the first dose of an antithyroid drug has been administered (see below). The radiographic contrast dyes may be used to block thyroid hormone release (as a result of the iodide released from these agents) and to inhibit thyroxine to triiodothyronine conversion. -Adrenergic receptor antagonists, such as propranolol and esmolol, and Ca2+ channel blockers also may be used to control tachyarrhythmias. Dexamethasone (0.5 to 1 mg intravenously every 6 hours) is recommended both as supportive therapy of possible relative adrenal insufficiency and as an inhibitor of conversion of thyroxine to triiodothyronine. Finally, treatment of the underlying precipitating illness is essential. Ionic Inhibitors The term ionic inhibitors designates the substances that interfere with the concentration of iodide by the thyroid gland. The effective agents are themselves anions that in some ways resemble iodide; they are all monovalent, hydrated anions of a size similar to that of iodide. The most studied example, thiocyanate, differs from the others qualitatively; it is not concentrated by the thyroid

gland, but in large amounts may inhibit the organification of iodine. Thiocyanate is produced following the enzymatic hydrolysis of certain plant glycosides. Thus, certain foods (e.g., cabbage) and cigarette smoking result in an increased concentration of thiocyanate in the blood and urine, as does the administration of sodium nitroprusside. Indeed, cigarette smoking has been reported to worsen both subclinical hypothyroidism (Mller et al., 1995) and Graves' ophthalmopathy (Bartelena et al., 1998b). Dietary precursors of thiocyanate may be a contributing factor in endemic goiter in certain parts of the world, especially in Central Africa, where the intake of iodine is very low (Delange et al., 1993). Among other anions, perchlorate (ClO 4) is ten times as active as thiocyanate (Wolff, 1998). Perchlorate blocks the entrance of iodide into the thyroid by competitively inhibiting the NIS (Carrasco, 2000). Although perchlorate can be used to control hyperthyroidism, it has caused fatal aplastic anemia when given in excessive amounts (2 to 3 g daily). Over the past few years, however, percholorate in doses of 750 mg daily has been used in the treatment of Graves' disease and amiodarone-induced thyrotoxicosis. Perchlorate can be used to "discharge" inorganic iodide from the thyroid gland in a diagnostic test of iodide organification. Other ions, selected on the basis of their size, also have been found to be active; fluoborate (BF4) is as effective as perchlorate. Lithium has a multitude of effects on thyroid function; its principal effect is decreased secretion of thyroxine and triiodothyronine (Takami, 1994). Iodide Iodide is the oldest remedy for disorders of the thyroid gland. Before the antithyroid drugs were used, it was the only substance available for control of the signs and symptoms of hyperthyroidism. Its use in this way is indeed paradoxical, and the explanation for this paradox is still incomplete. Mechanism of Action High concentrations of iodide appear to influence almost all important aspects of iodine metabolism by the thyroid gland (seeRoti and Vagenakis, 2000). The capacity of iodide to limit its own transport has been mentioned above. Acute inhibition of the synthesis of iodotyrosines and iodothyronines by iodide also is well known (the Wolff-Chaikoff effect). This transient, 2-day inhibition is observed only above critical concentrations of intracellular rather than extracellular concentrations of iodide. With time there is "escape" from this inhibition that is associated with an adaptive decrease in iodide transport and a lowered intracellular iodide concentration, most likely due to a decrease in NIS mRNA and protein (Eng et al., 1999). The mechanism of the acute WolffChaikoff effect remains elusive and has been postulated to be due to the generation of organic iodocompounds within the thyroid (Pisarev and Grtner, 2000). A very important clinical effect of high plasma iodide concentration is an inhibition of the release of thyroid hormone. This action is rapid and efficacious in severe thyrotoxicosis. The effect is exerted directly on the thyroid gland, and it can be demonstrated in the euthyroid subject and experimental animals as well as in the hyperthyroid patient. Studies in a cultured thyroid cell line suggest that some of the inhibitory effects of iodide on thyrocyte proliferation may be mediated by actions of iodide on crucial regulatory points in the cell cycle (Smerdely et al., 1993). In euthyroid individuals, the administration of doses of iodide from 1.5 to 150 mg daily results in small decreases in plasma thyroxine and triiodothyronine concentrations and small compensatory increases in serum TSH values, with all values remaining in the normal range. However, euthyroid patients with a history of a wide variety of underlying thyroid disorders may develop iodineinduced hypothyroidism when exposed to large amounts of iodine present in many commonly

prescribed drugs (Table 576), and these patients do not escape from the acute Wolff-Chaikoff effect (Roti et al., 1997). Among the disorders that predispose patients to iodine-induced hypothyroidism are treated Graves' disease, Hashimoto's thyroiditis, postpartum lymphocytic thyroiditis, subacute painful thyroiditis, and lobectomy for benign nodules. The most commonly prescribed iodine-containing drugs are certain expectorants, topical antiseptics, and radiologic contrast agents. Response to Iodide in Hyperthyroidism The response to iodides in patients with hyperthyroidism is often striking and rapid. The effect usually is discernible within 24 hours, and the basal metabolic rate may fall at a rate comparable to that following thyroidectomy. This provides evidence that the release of hormone into the circulation is rapidly blocked. Furthermore, thyroid hormone synthesis also is mildly decreased. The maximal effect is attained after 10 to 15 days of continuous therapy, when the signs and symptoms of hyperthyroidism may have greatly improved. The changes in the thyroid gland have been studied in detail; vascularity is reduced, the gland becomes much firmer, the cells become smaller, colloid reaccumulates in the follicles, and the quantity of bound iodine increases. The changes are those that would be expected if the excessive stimulus to the gland had somehow been removed or antagonized. Unfortunately, iodide therapy usually does not completely control the manifestations of hyperthyroidism, and after a variable period of time, the beneficial effect disappears (Emerson et al., 1975). With continued treatment, the hyperthyroidism may return in its initial intensity or may become even more severe than it was at first. It is for this reason that, when iodide was the only agent available for the treatment of hyperthyroidism, its use was usually restricted to preparation of the patient for thyroidectomy. Therapeutic Uses The uses of iodide in the treatment of hyperthyroidism are in the preoperative period in preparation for thyroidectomy and, in conjunction with antithyroid drugs and propranolol, in the treatment of thyrotoxic crisis. Prior to surgery, iodide is sometimes employed alone, but more frequently it is used after the hyperthyroidism has been controlled by an antithyroid drug. It is then given for 7 to 10 days immediately preceding the operation. Optimal control of hyperthyroidism is achieved if antithyroid drugs are first given alone. If iodine also is given from the beginning, variable responses are observed; sometimes the effect of iodide predominates, storage of hormone is promoted, and prolonged antithyroid treatment is required before the hyperthyroidism is controlled. These clinical observations may be explained by the ability of iodide to prevent the inactivation of thyroid peroxidase by antithyroid drugs (Taurog, 2000). Another use of iodine is to protect the thyroid from radioactive iodine fallout following a nuclear accident. Because the uptake of radioactive iodine is inversely proportional to the serum concentration of stable iodine, the administration of 30 to 100 mg of iodide daily will markedly decrease the thyroid uptake of radioisotopes of iodine. Following the Chernobyl nuclear reactor accident in 1986, approximately 10 million children and adults in Poland were given stable iodide to block the thyroid exposure to radioactive iodine from the atmosphere and from dairy products from cows that ate contaminated grass (Naumann and Wolff, 1993). This prevented the occurrence of radiation-induced thyroid cancer, as observed in children residing near Chernobyl. The dosage or form in which iodide is administered bears little relationship to the response achieved

in hyperthyroidism, provided that not less than the minimal effective amount is given; this dosage is 6 mg per day in most, but not all, patients. Strong iodine solution (Lugol's solution) is widely used and consists of 5% iodine and 10% potassium iodide, which yields a dose of 6.3 mg of iodine per drop. The iodine is reduced to iodide in the intestine before absorption. Saturated solution of potassium iodide also is available, containing 38 mg per drop. Typical doses include 3 to 5 drops of Lugol's solution or 1 to 3 drops of saturated solution of potassium iodide 3 times a day. These doses have been determined empirically and are far in excess of that needed. Untoward Reactions Occasional individuals show marked sensitivity to iodide or to organic preparations that contain iodine when they are administered intravenously. The onset of an acute reaction may occur immediately or several hours after administration. Angioedema is the outstanding symptom, and swelling of the larynx may lead to suffocation. Multiple cutaneous hemorrhages may be present. Also, manifestations of the serum-sickness type of hypersensitivitysuch as fever, arthralgia, lymph node enlargement, and eosinophiliamay appear. Thrombotic thrombocytopenic purpura and fatal periarteritis nodosa attributed to hypersensitivity to iodide also have been described. The severity of symptoms of chronic intoxication with iodide (iodism) is related to the dose. The symptoms start with an unpleasant brassy taste and burning in the mouth and throat as well as soreness of the teeth and gums. Increased salivation is noted. Coryza, sneezing, and irritation of the eyes with swelling of the eyelids are commonly observed. Mild iodism simulates a "head cold." The patient often complains of a severe headache that originates in the frontal sinuses. Irritation of the mucous glands of the respiratory tract causes a productive cough. Excess transudation into the bronchial tree may lead to pulmonary edema. In addition, the parotid and submaxillary glands may become enlarged and tender, and the syndrome may be mistaken for mumps parotitis. There also may be inflammation of the pharynx, larynx, and tonsils. Skin lesions are common and vary in type and intensity. They usually are mildly acneform and distributed in the seborrheic areas. Rarely, severe and sometimes fatal eruptions (ioderma) may occur after the prolonged use of iodides. The lesions are bizarre, resemble those caused by bromism, a rare problem, and, as a rule, involute quickly when iodide is withdrawn. Symptoms of gastric irritation are common, and diarrhea, which is sometimes bloody, may occur. Fever is occasionally observed, and anorexia and depression may be present. The mechanisms involved in the production of these derangements remain unknown. Fortunately, the symptoms of iodism disappear spontaneously within a few days after stopping the administration of iodide. The renal excretion of I can be increased by procedures that promote Cl excretion (e.g., osmotic diuresis, chloruretic diuretics, and salt loading). These procedures may be useful when the symptoms of iodism are severe. Radioactive Iodine Chemical and Physical Properties Although iodine has several radioactive isotopes, greatest use has been made of 131I. It has a halflife of 8 days; therefore, more than 99% of its radiation is expended within 56 days. Its radioactive emissions include both rays and particles. The short-lived radionuclide of iodine, 123 I, is primarily a -emitter with a half-life of only 13 hours. This permits a relatively brief exposure to radiation during thyroid scans. Effects on the Thyroid Gland

The chemical behavior of the radioactive isotopes of iodine is identical to that of the stable isotope, 127 131 I. I is rapidly and efficiently trapped by the thyroid, incorporated into the iodoamino acids, and deposited in the colloid of the follicles, from which it is slowly liberated. Thus, the destructive particles originate within the follicle and act almost exclusively upon the parenchymal cells of the thyroid, with little or no damage to surrounding tissue. The radiation passes through the tissue and can be quantified by external detection. The effects of the radiation depend upon the dosage. When small tracer doses of 131I are administered, thyroid function is not disturbed. However, when large amounts of radioactive iodine gain access to the gland, the characteristic cytotoxic actions of ionizing radiation are observed. Pyknosis and necrosis of the follicular cells are followed by disappearance of colloid and fibrosis of the gland. With properly selected doses of 131I, it is possible to destroy the thyroid gland completely without detectable injury to adjacent tissues. After smaller doses, some of the follicles, usually in the periphery of the gland, retain their function. Therapeutic Uses Sodium iodide I 131 (IODOTOPE THERAPEUTIC) is available as a solution or in capsules containing essentially carrier-free 131I suitable for oral administration. Sodium iodide I 123 is available for scanning procedures. Radioactive iodine finds its widest use in the treatment of hyperthyroidism and in the diagnosis of disorders of thyroid function. Discussion here is limited to the uses of 131I. Hyperthyroidism Radioactive iodine is highly useful in the treatment of hyperthyroidism; in many circumstances it is regarded as the therapeutic procedure of choice for this condition (Soloman et al., 1990; for review, seeLevy, 1997). The use of stable iodide as treatment for hyperthyroidism, however, may preclude treatment and certain imaging studies with radioactive iodine for weeks after the iodide has been discontinued. Dosage and Technique
131

I is administered orally, and the effective dose differs for individual patients. It depends primarily upon the size of the thyroid, the iodine uptake of the gland, and the rate of release of radioactive iodine from the gland subsequent to its deposition in the colloid. To determine these variables insofar as possible, many investigators administer a tracer dose of 131 I and calculate the 131I accumulated by the gland and the rate of loss therefrom. The weight of the gland is estimated by palpation or by ultrasound. From these data, the dose of isotope necessary to provide from 7000 to 10,000 rad per gram of thyroid tissue is determined. Even when dosage is controlled in this manner, it is difficult to predict the response of an individual to a given amount of the isotope. Indeed, comparison studies have shown little advantage of a standardized dose, based on gland weight and radioactive iodine uptake, over a fixed dose (Jarlv et al., 1995; de Bruin et al., 1994). For these reasons, the optimal dose of 131I, expressed in terms of microcuries taken up per gram of thyroid tissue, varies in different laboratories from 80 to 150 Ci. The usual total dose is 4 to 15 mCi. Lower-dosage 131I therapy (80 Ci/g thyroid) has been advocated to reduce the incidence of subsequent hypothyroidism. While the incidence of hypothyroidism in the early years after such therapy is lower, many patients with late hypothyroidism may go undetected, and the ultimate incidence of hypothyroidism is probably no less than with the larger doses. In addition, relapse of the hyperthyroid state, or initial failure to alleviate the hyperthyroid state, is increased in patients receiving lower doses of 131I. There also is evidence that pretreatment with propylthiouracil reduces the therapeutic efficacy of 131I, necessitating a higher dose for a desired effect (Imseis et al., 1998; Tuttle et al., 1995). Methimazole appears not to share this effect of propylthiouracil (Imseis et al.,

1998). Course of Disease The course of hyperthyroidism in a patient who has received an optimal dose of 131 I is characterized by progressive recovery. It is very unusual for any tenderness to be noted in the thyroid region, and most observers have failed to detect any exacerbation of hyperthyroidism from loss of hormone from the damaged gland in patients whose preformed hormone stores have been depleted by antithyroid drug therapy. Beginning a few weeks after treatment, the symptoms of hyperthyroidism gradually abate over a period of 2 to 3 months. If therapy has been inadequate, the necessity for further treatment is apparent within 6 to 12 months. It is not uncommon, however, for the serum TSH to remain low for several months after 131I therapy, especially if the patient was not pretreated to euthyroidism prior to receiving the radioactive iodine (Uy et al., 1995). Occasionally, this delayed recovery of the hypothalamic-pituitarythyroid axis results in a picture of central hypothyroidism, with low circulating thyroid hormones. Thus, assessing radioactive iodine failure based on TSH concentrations alone may be misleading and should always be accompanied by determination of a free T4 index and serum T3 concentrations. Furthermore, transient hypothyroidism, lasting up to 6 months, may occur in up to 50% of patients receiving a dose of 131I calculated to result in euthyroidism (Aizawa et al., 1997). This is less of a problem if the patient receives a higher, ablative dose of 131I, since hypothyroidism occurs far more frequently and persists. Depending to some extent upon the dosage schedule adopted, one-half to two-thirds of patients are cured by a single dose, one-third to one-fifth require two doses, and the remainder require three or more doses before the disorder is controlled. Patients treated with larger doses of 131I almost always develop hypothyroidism within a few months. Propranolol, antithyroid drugs, or both, or stable iodide, can be used to hasten the control of hyperthyroidism while awaiting the full effects of the radioactive iodine. However, the antithyroid drugs should be withheld for a few days before and after the therapeutic dose of 131I. Advantages The advantages of radioactive iodine in the treatment of Graves' disease are many. No death as a direct result of the use of the isotope has been reported, and only by a gross miscalculation of dose could such an event conceivably occur. There have been reports of increased mortality from cardiovascular and cerebrovascular disease in the first year after radioactive iodine therapy (Franklyn et al., 1998). However, there is no evidence that the increased mortality was related to the radioactive iodine itself, and long-term follow-up of radioactive iodine therapy for Graves' disease has demonstrated no increase in overall cancer mortality in patients treated with 131I (Ron et al., 1998). In the nonpregnant patient, no tissue other than the thyroid is exposed to sufficient ionizing radiation to be detectably altered. Nevertheless, the continuing concern about potential effects of radiation on germ cells prompts some endocrinologists to advocate antithyroid drugs or surgery in younger patients who are acceptable operative risks (Zimmerman, 1999). Hypoparathyroidism is a small risk of surgery. With radioactive iodine treatment, the patient is spared the risks and discomfort of surgery. The cost is low, hospitalization is not required, and patients can indulge in their customary activities during the entire procedure. Disadvantages The chief disadvantage of the use of radioactive iodine is the high incidence of delayed

hypothyroidism that is induced. Even when elaborate procedures are used to estimate iodine uptake and gland size, a certain percentage of patients will be overtreated. A distressing feature of this complication is its rising prevalence with the passage of time; the longer the interval after treatment, the higher the incidence. Several analyses of groups of patients treated 10 or more years previously suggest that the eventual rate may exceed 80%. However, it now appears that the incidence of hypothyroidism also increases progressively after subtotal thyroidectomy or after antithyroid drug therapy, and such failure of glandular function is probably part of the natural progression of Graves' disease, no matter what the therapy. Although it is often said that hypothyroidism is not a serious complication because it can be treated so easily with thyroid hormone, its onset may be insidious and overlooked for some time. Also, once diagnosed, it is difficult to ensure that patients who need the hormone actually take it. Since the health risks of untreated subclinical hypothyroidism are becoming increasingly evident (Hak et al., 2000; Surks and Ocampo, 1996), hypothyroidism, either subclinical or overt, is a serious complication and requires long-term follow-up to ensure that optimal replacement therapy be provided. Another disadvantage of radioactive iodine therapy is the long period of time that is sometimes required before the hyperthyroidism is controlled. When a single dose is effective, the response is most satisfactory; however, when multiple doses are needed, it may be many months or a year or more before the patient is well. This disadvantage can be largely overcome if the initial dose is sufficiently large. Other disadvantages include possible worsening of ophthalmopathy after treatment, although this is controversial (DeGroot et al., 1995). Although extremely rare, there have been reported cases of thyroid storm after therapy with 131I. Importantly, the cases of thyroid storm after radioactive iodine therapy occur in most cases in patients who have not received pretreatment with antithyroid drugs. Indications The clearest indication for this form of treatment is hyperthyroidism in older patients and in those with heart disease. Radioactive iodine also is the best form of treatment when Graves' disease has persisted or recurred after subtotal thyroidectomy and when prolonged treatment with antithyroid drugs has not led to remission. Finally, radioactive iodine is indicated in patients with toxic nodular goiter, since the disease does not go into spontaneous remission. The risk of inducing hypothyroidism is less in nodular goiter than in Graves' disease, perhaps because of the normal progression of the latter and the preservation of nonautonomous thyroid tissue in the former. Usually, larger doses of radioactive iodine are required in the treatment of toxic nodular goiter than in the treatment of Graves' disease. Recently, radioactive iodine has been used to decrease the size of large, nontoxic, multinodular goiters that are causing compressive symptoms in patients who are otherwise poor operative risks (Huysmans et al., 1997). While surgery remains the treatment of choice for the young patient with compressive multinodular goiters, radioactive iodine therapy may be of benefit in elderly patients, especially those with cardiopulmonary disease. Contraindications The main contraindication for the use of 131 I therapy is pregnancy. After the first trimester, the fetal thyroid will concentrate the isotope and thus suffer damage, but even during the first trimester, radioactive iodine is best avoided because there may be adverse effects of radiation on fetal tissues. The risk of causing neoplastic changes in the thyroid gland has been constantly under consideration since radioactive iodine was first introduced, and only small numbers of children have been treated in this way. Indeed, many clinics have declined to treat younger patients for fear of causing cancer

and have reserved radioactive iodine for patients over some arbitrary age, such as 25 or 30 years. Since experience with 131I is now vast, these age limits are lower than they were in the past. The most recent report by the Cooperative Thyrotoxicosis Therapy Follow-up Study Group, which began tracking patients in 1961, shows no increase in total cancer mortality following 131 I treatment for Graves' disease (Ron et al., 1998). Furthermore, there was no increase in the occurrence of leukemia following large dose 131 I therapy for thyroid cancer, although there was an increase in colorectal cancers in this population (de Vathaire et al., 1997). These data strongly suggest that laxatives be given to all patients receiving 131 I therapy for treatment of thyroid cancer to decrease the risk of future digestive tract malignancies. Transient abnormalities in testicular function have been reported following 131I therapy for treatment of thyroid cancer, but no long-term effects on fertility in either men or women have been demonstrated (Pacini et al., 1994a; Dottorini et al., 1995). Metastatic Thyroid Carcinoma While most well-differentiated thyroid carcinomas accumulate very little iodine, stimulation of iodine uptake with TSH often is used effectively to treat metastases. Follicular carcinomas, which account for 10% to 15% of thyroid malignancies, are especially amenable to this treatment. Currently, endogenous TSH stimulation is evoked by withdrawal of thyroid hormone replacement therapy in patients previously treated with near-total thyroidectomy with or without radioactive ablation of residual thyroid tissue. Total body 131 I scanning and measurement of serum thyroglobulin when the patient is hypothyroid (TSH > 35 mU/liter) should be performed to identify metastatic disease or residual thyroid bed tissue. Depending upon the residual uptake, or the presence of metastatic disease, an ablative dose of 131I ranging from 30 to 150 mCi is administered, and a repeat total body scan is obtained 1 week later. The precise amount of 131I needed to treat residual tissue and metastases is controversial. Recombinant human TSH (THYROGEN ) is now available to test the ability of thyroid tissue, both normal and malignant, to take up radioactive iodine and to secrete thyroglobulin (Haugen et al., 1999). The major advantage to the use of this medication is that patients do not have to stop their suppressive levothyroxine therapy and become clinically hypothyroid for the presence of persistent or metastatic disease to be assessed. Recombinant human TSH is not yet approved for treatment prior to therapeutic administration of 131 I. TSH-suppressive therapy with levothyroxine is indicated in all patients after treatment for thyroid cancer. The goal of therapy usually is to keep serum TSH levels in the subnormal range (Burmeister et al., 1992). Follow-up evaluation every 6 months is reasonable, along with determination of serum thyroglobulin concentrations. A rise in serum thyroglobulin concentration is often the first indication of recurrent disease. Prognosis in patients with thyroid cancer depends upon the pathology and size of the tumor and is generally worse in the elderly (seeMazzaferri, 2000). Overall, the vast majority of patients with thyroid cancer will not die of their disease. Papillary cancer is not an aggressive tumor. It metastasizes locally and has a 10-year survival rate of greater than 90%. Lymph node metastases at the time of diagnosis do little to alter the prognosis. Follicular cancer is more aggressive and can metastasize via the bloodstream. Still, prognosis is fair, and longterm survival is common. It is important to realize that, even in patients with metastatic, differentiated thyroid cancer, 131I therapy is very effective and may be even curative (Pacini et al., 1994b). Anaplastic cancer is the exception, as it is highly malignant with survival usually less than 1 year. Diagnostic Uses Tracer studies with radioactive iodine have found wide application in studies of disorders of the

thyroid gland. Measurement of the thyroidal accumulation of a tracer dose is helpful in the differential diagnosis of thyrotoxicosis and nodular goiter. The response of the thyroid to TSHsuppressive doses of thyroid hormone can be evaluated in this way. Following the administration of a tracer dose, the pattern of localization in the thyroid gland can be depicted by a special scanning apparatus, and this technique is sometimes useful in defining thyroid nodules as functional ("hot") or nonfunctional ("cold") and in finding ectopic thyroid tissue and occasionally metastatic thyroid tumors. For further discussion of diseases of the thyroid, see alsoChapter 320 in Harrison's Principles of Internal Medicine, 16th ed., McGraw Hill, New York, 2005.

Chapter 58. Estrogens and Progestins

Overview Estrogens and progestins are among the most widely prescribed drugs. This chapter covers the major uses of estrogens and progestins, alone or in combination, for contraception and for hormonereplacement therapy in postmenopausal women. The less-frequent use of estrogen, sometimes in conjunction with growth hormone or gonadotropins, for treatment of developmental delay or hypogonadism also is discussed (Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors provides additional discussion of this topic). The use of estrogen-receptor antagonists or progestins and aromatase inhibitors as functional estrogen antagonists is described for treatment of estrogen-dependent neoplasms. The use and potential use of progesterone antagonists, such as mifepristone (RU 486), also is discussed. Cancer chemotherapeutic strategies based on blockade of estrogen and/or progesterone receptor functions are considered in further detail in Chapter 52: Antineoplastic Agents. Complementary therapeutic strategies based on suppression of gonadotropin secretion by long-acting gonadotropin-releasing hormone agonists are discussed in Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors. Estrogens and Progestins: Introduction Estrogens and progestins are endogenous hormones that produce numerous physiological actions. In women, these include developmental effects, neuroendocrine actions involved in the control of ovulation, the cyclical preparation of the reproductive tract for fertilization and implantation, and major actions on mineral, carbohydrate, protein, and lipid metabolism. Many features of the female habitus also are influenced by these hormones. It has become clear more recently that estrogens have important actions in males, including effects on bone, spermatogenesis, and behavior. The basic features of the biosynthesis, biotransformation, and disposition of these agents are well established, and the nuclear receptor system for these hormones is well characterized. This knowledge provides a firm conceptual basis for understanding the physiological and pharmacological activities of both hormones. The therapeutic use of estrogens and progestins is widespread, and their pharmacological actions largely reflect extensions of their physiological activities. The most common uses of these agents are hormone-replacement therapy in postmenopausal women and contraception, but the specific compounds and dosages used in these two settings are substantially different. Although oral contraceptives are used primarily to prevent pregnancy, they also have significant health benefits beyond contraception. Naturally occurring and synthetic compounds are available for oral and parenteral uses.

Estrogen- and progesterone-receptor antagonists also are available. The main use of antiestrogens is in the treatment of hormone-responsive breast cancer. Treatment of female infertility is the most common use of these antagonists in gynecology. The main use of antiprogestins to date has been for medical abortion, but other uses are being developed. A number of naturally occurring and synthetic environmental chemicals mimic, antagonize, or otherwise affect the actions of estrogens in experimental test systems. While the precise effect of these environmental agents on human beings is not known, this is an area of active investigation. History It has long been known that removal of the ovaries results in uterine atrophy and a loss of sexual functions. The hormonal nature of the ovarian control of the female reproductive system was established in 1900 by Knauer when he found that ovarian transplants prevented the symptoms of gonadectomy. This observation was extended by Halban (1900), who showed that, if the glands were transplanted even in immature animals, normal sexual development and function were assured. In 1923, Allen and Doisy devised a simple bioassay for ovarian extracts based upon changes produced in the vaginal smear of the rat. Loewe (1925) first reported a female sex hormone in the blood of various species, and shortly thereafter Frank and associates (1925) detected an active sex principle in the blood of sows in estrus. Of even greater significance was the discovery by Loewe and Lange (1926) of a female sex hormone in the urine of menstruating women and the observation that the concentration of the hormone in the urine varied with the phase of the menstrual cycle. The excretion of large amounts of estrogen in the urine during pregnancy also was reported (Zondek, 1928). This finding was a boon to chemists, who soon isolated an active substance in crystalline form (Butenandt, 1929; Doisy et al., 1929, 1930). A few years later, its chemical structure was elucidated. The results of early investigations indicated that the ovary secretes two substances. Beard (1897) had postulated that the corpus luteum serves a necessary function during pregnancy, and Fraenkel (1903) showed that destruction of the corpora lutea in pregnant rabbits causes abortion. The contributions of Corner and Allen (1929) firmly established the hormonal function of the corpus luteum. These investigators showed that the abortion following extirpation of the corpora lutea in pregnant rabbits can be prevented by the injection of luteal extracts. In the early 1960s, pioneering studies by Jensen and colleagues suggested the presence of intracellular receptors for estrogens in the target tissues (Jensen and Jacobsen, 1962). This was historically important, because it was the first demonstration of receptors of the steroid/thyroid superfamily and because it provided the experimental approaches used to identify similar receptors for the other steroid hormones (Jensen and DeSombre, 1972). A second form of the estrogen receptor recently has been identified and termed estrogen receptor (ER ) to distinguish it from the first receptor, which is now referred to as estrogen receptor (ER ). Estrogens Chemistry Estrogenic activity is shared by many steroidal and nonsteroidal compounds, some of which are shown in Table 581 and Figure 581. The most potent naturally occurring estrogen in human beings is 17 -estradiol, followed by estrone and estriol. Each of these molecules is an 18-carbon steroid, containing a phenolic A ring (an aromatic ring with a hydroxyl group at carbon 3) and a hydroxyl group or ketone in position 17 of ring D. The phenolic A ring is the principal structural

feature responsible for selective, high-affinity binding to estrogen receptors. Most alkyl substitutions on the phenolic A ring impair such binding, but substitutions on ring C or D may be tolerated. Ethinyl substitutions at the C 17 position greatly increase oral potency by inhibiting firstpass hepatic metabolism. A model for the estrogen receptor ligand-binding site has been developed from structure-activity studies (Anstead et al., 1997), and the crystal structures of estrogen-receptor complexes have been reported (Brzozowski et al., 1997). Figure 581. The Biosynthetic Pathway for the Estrogens.

One of the first nonsteroidal estrogens to be synthesized was diethylstilbestrol or DES (see Table 581), which is structurally similar to estradiol when viewed in the trans conformation. DES is as potent as estradiol in most assays, but is orally active and has a longer half-life in the body. DES no longer has widespread use, but it is important historically because its introduction as a cheap, plentiful, orally active estrogen at a time when the natural products were scarce was a milestone in the development of effective endocrine therapy (Dodds et al., 1938). Nonsteroidal compounds with estrogenic or antiestrogenic activityincluding flavones, isoflavones (e.g., genistein), and coumestan derivativesoccur naturally in a variety of plants and fungi. A number of synthetic agentsincluding pesticides (e.g., p,p'-DDT), plasticizers (e.g., bisphenol A), and a variety of other industrial chemicals (e.g., polychlorinated biphenyls)also have hormonal or antihormonal activity. Many of these polycyclic compounds contain a phenolic ring that mimics the A ring of steroids. While the affinity of these "environmental estrogens" for the estrogen receptor is relatively weak, their large number, bioaccumulation in adipose tissue, and persistence in the environment have raised concerns about their potential toxicity in human beings and wildlife (Mkelet al., 1999). Both over-the-counter and prescription preparations containing naturally

occurring, estrogen-like compounds from plants (i.e., phytoestrogens) now are available. There also have been reports that phytoestrogens such as genistein may exhibit relative selectivity for ER (Kuiper et al., 1998); this possibility is being actively investigated. Biosynthesis Steroidal estrogens are formed from either androstenedione or testosterone as immediate precursors (see Figure 581). The reaction involves aromatization of the A ring, and it is catalyzed in three steps by a cytochrome P450 monooxygenase enzyme complex (aromatase or CYP19) that uses NADPH and molecular oxygen as cosubstrates (Simpson et al., 1994). In the first step of this reaction, C 19 (the angular methyl group on C 10 of the androgen precursors) is hydroxylated. A second hydroxylation results in the elimination of the newly formed C 19 hydroxymethyl group, and a final hydroxylation of C 2 results in the formation of an unstable intermediate that rearranges to form the phenolic A ring. The entire reaction consumes three molecules of oxygen and three molecules of NADPH. Aromatase activity resides within a transmembrane glycoprotein (cytochrome P450 family of monooxygenases); a ubiquitous flavoprotein, NADPHcytochrome P450 reductase, also is essential. Both proteins are localized in the endoplasmic reticulum of ovarian granulosa cells, testicular Sertoli and Leydig cells, stromal cells of adipose tissue, placental syncytiotropho-blasts, the preimplantation blastocyst, bone, and various brain regions (Simpson et al., 1999). The ovaries are the principal source of circulating estrogen in premenopausal women. The major secretory product is estradiol, synthesized by granulosa cells from androgenic precursors provided by theca cells. Aromatase activity is induced by gonadotropins, which act via plasma membrane receptors to elevate intracellular concentrations of adenosine 3',5'-monophosphate (cyclic AMP). Gonadotropins and cyclic AMP also increase the activity of the cholesterol side-chain cleavage enzyme and facilitate the transport of cholesterol (the precursor of all steroids) into the mitochondria of cells that synthesize steroids. The ovary contains a form of 17 -hydroxysteroid dehydrogenase (type 1) that favors the production of testosterone and estradiol from androstenedione and estrone, respectively. However, in the liver, another form of this enzyme (type 2) favors oxidation of circulating estradiol to estrone (Peltoketo et al., 1999), and both of these steroids are then converted to estriol (see Figure 581). All three of these estrogens are then excreted in the urine along with their glucuronide and sulfate conjugates. In postmenopausal women, the principal source of estrogen is adipose tissue stroma and other nonovarian sites, where estrone is synthesized from dehydroepiandrosterone, secreted by the adrenal cortex. In men, estrogens are produced by the testes, but extragonadal production by aromatization of circulating C 19 steroids such as androstenedione and dehydroepiandrosterone appears to account for the majority of circulating estrogenic hormones. Thus, the level of estrogens is regulated in part by the availability of androgenic precursors (Mendelson and Simpson, 1987). Estrogenic effects most often have been attributed to circulating hormones, but locally produced estrogens also may have important actions (Simpson et al., 1999). For example, estrogens may be produced from androgens by the actions of aromatase or from estrogen conjugates by hydrolysis. Such local production of estrogens could play a causal role in the development of certain diseases such as breast cancer, since mammary tumors contain both aromatase and hydrolytic enzymes. Estrogens also may be produced from androgens via aromatase present in the central nervous system (CNS) and other tissues and exert local effects near the site of their production; in the testes, they affect spermatogenesis, and in bone, they have major effects on bone mineral density. Large quantities of estrogens are synthesized by the placenta, which uses fetal

dehydroepiandrosterone and its 16 -hydroxyl derivative to produce estrone and estriol, respectively; human urine of pregnancy is thus an abundant source of natural estrogens. The pregnant mare excretes more than 100 mg daily, a record exceeded only by the stallion, who, despite clear manifestations of virility, excretes into his environment more estrogen than any other living creature. Physiological and Pharmacological Actions Developmental Actions The estrogens are largely responsible for the changes that take place at puberty in girls and account for the secondary sexual characteristics of females. By a direct action, they cause growth and development of the vagina, uterus, and fallopian tubes. Estrogens act in concert with other hormones to cause enlargement of the breasts through promotion of ductal growth, stromal development, and the accretion of fat. They also contribute in a poorly understood manner to molding the body contours, shaping the skeleton, and bringing about the pubertal growth spurt of the long bones and its culmination by fusion of the epiphyses. Growth of axillary and pubic hair and pigmentation of the genital region also are effects of estrogen, as are the regional pigmentation of the nipples and areolae that occur after the first trimester of pregnancy. While sexual development in females appears to be due primarily to estrogens, androgens may play a secondary role. Testosterone and androstenedione are normally found in venous ovarian blood (see Chapter 59: Androgens); these may contribute to pubertal changes in girls, such as growth spurts, the full development of pubic and axillary hair, and the appearance of acne due to growth and secretions from the sebaceous glands. It has been recognized only recently that estrogens appear to play important developmental roles in males. In boys, estrogen deficiency does not affect the age of pubertal onset, but the pubertal growth spurt is diminished, skeletal maturation and epiphyseal closure are delayed, and linear growth continues into adulthood. Estrogen deficiency in men also leads to hypergonadotropism, macroorchidism, and increased testosterone levels and also may affect carbohydrate and lipid metabolism and fertility in some individuals (Grumbach and Auchus, 1999). Neuroendocrine Control of the Menstrual Cycle The menstrual cycle in women is controlled by a neuroendocrine cascade involving the hypothalamus, pituitary, and ovaries, as illustrated in Figure 582. A neuronal oscillator or "clock" in the hypothalamus fires at regular intervals, resulting in the periodic release of gonadotropinreleasing hormone (GnRH) into the hypothalamic-pituitary portal vasculature (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). The GnRH then interacts with its cognate receptor on gonadotropes and causes the release of luteinizing hormone (LH) and folliclestimulating hormone (FSH) from the anterior pituitary. The gonadotropins (LH and FSH) are responsible for the growth and maturation of the graafian follicle in the ovary and for the ovarian production of estrogen and progesterone, which exert feedback regulation on the pituitary and hypothalamus.

Figure 582. Neuroendocrine Control of Gonadotropin Secretion in Females. The hypothalamic pulse generator located in the arcuate nucleus of the hypothalamus functions as a neuronal "clock" that fires at regular hourly intervals (A). This results in the periodic release of gonadotropin-releasing hormone (GnRH) from GnRH-containing neurons into the hypothalamic-pituitary portal vasculature (B). GnRH neurons (B) receive inhibitory input from opioid, dopamine, and gammaaminobutyric acid (GABA) neurons and stimulatory input from noradrenergic neurons (NE, norepinephrine). The pulses of GnRH trigger the intermittent release of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) from pituitary gonadotropes (C), resulting in a pulsatile plasma profile (D). FSH and LH regulate ovarian production of estrogen and progesterone, which exert feedback controls (E). (See text and Figure 583 for additional details.)

Figure 583. Hormonal Relationships of the Human Menstrual Cycle. A. Average daily values of LH, FSH, estradiol (E2), and progesterone in plasma samples from women exhibiting normal 28-day menstrual cycles. Changes in the ovarian follicle (top) and endometrium (bottom) also are illustrated schematically. Frequent plasma sampling reveals pulsatile patterns of gonadotropin release. Characteristic profiles are illustrated schematically for the follicular phase (day 9, inset on left) and luteal phase (day 17, inset on right). Both the frequency (number of pulses per hour) and amplitude (extent of change of hormone release) of pulses vary throughout the cycle. (Redrawn with permission from Thorneycroft et al., 1971.) B. Major regulatory effects of ovarian steroids on hypothalamic-pituitary function. Estrogen decreases the amount of follicle stimulating hormone (FSH) and luteinizing hormone (LH) released (i.e., gonadotropin pulse amplitude) during most of the cycle and triggers a surge of LH release only at midcycle. Progesterone decreases the frequency of GnRH release from the hypothalamus and thus decreases the frequency of plasma gonadotropin pulses. Progesterone also increases the amount of LH released (i.e., the pulse amplitude) during the luteal phase of the cycle.

Because the release of GnRH is intermittent, LH and FSH release is pulsatile as determined by the neural "clock" (Figure 582), which is referred to as the hypothalamic GnRH pulse generator (Knobil, 1981; Wilson et al., 1984). The intermittent, pulsatile nature of hormone release is essential for the maintenance of normal ovulatory menstrual cycles, since constant infusion of GnRH results in a cessation of LH and FSH release, a decrease of estradiol and progesterone production, and amenorrhea (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). It is not clear if the regular, intermittent discharges of the pulse generator are an intrinsic property of GnRH neurons or if other neurons that synapse on GnRH cells exert the pacemaker function. Neuroanatomically, the pulse generator resides in the arcuate nucleus of the hypothalamus, and this region of the brain contains the highest concentration of GnRH neurons. The pulse generator is not

dependent on afferent input from other regions of the brain to maintain its pulsatile activity. The hypothalamus has relatively few GnRH-containing cells, and there is no obvious GnRH network. It is thus unclear how this small number of cells scattered bilaterally throughout the arcuate nucleus fires simultaneously. Most GnRH cells appear to be devoid of estrogen or progesterone receptors, but they may receive synaptic input from opioid, catecholamine, and gamma-aminobutyric acid (GABA) neurons that express receptors for the ovarian steroids (see Figure 582). Prior to puberty, the hypothalamic GnRH pulse generator does not function, gonadotropin secretion is absent, and menstrual cycles do not occur. Unknown physiological mechanisms that take place at the onset of puberty activate the pulse generator. Following this activation, the LH, FSH, estradiol, and progesterone profiles seen in the menstrual cycle occur. Figure 583 provides a schematic diagram of the profiles of gonadotropin and gonadal steroid secretion during the menstrual cycle. The "average" plasma levels of LH throughout the cycle are shown in panel A of Figure 583; panel B illustrates the pulsatile patterns of LH in more detail. Note that the average LH levels are similar throughout the early (follicular) and late (luteal) phases of the cycle, but the frequency and amplitude of the LH pulses are quite different in the two phases. This characteristic pattern of hormone secretions results from complex positive and negative feedback mechanisms (for a more comprehensive review, see Hotchkiss and Knobil, 1994). In the early follicular phase of the cycle: (1) the pulse generator produces a burst of neuronal activity with a frequency of about 1 per hour resulting in the liberation of GnRH; (2) this causes a corresponding pulsatile release of LH and FSH from pituitary gonadotropes, and FSH in particular; (3) which causes the graafian follicle to mature and secrete estrogen. The effects of estrogens on the pituitary are inhibitory at this time. Therefore, as estrogen levels increase, the steroid reduces the amount of LH and FSH released from the pituitary (i.e., the amplitude of the LH pulse), and gonadotropin levels gradually decline, as seen in Figure 583. Inhibin, produced by the ovary, also exerts a negative feedback and decreases serum FSH levels at this time (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). At midcycle, a different set of regulatory interactions comes into play. At this time, serum estradiol rises above a threshold level of 150 to 200 pg/ml for approximately 36 hours. This sustained elevation of estrogen no longer inhibits gonadotropin release but exerts a brief positive feedback effect on the pituitary to trigger the preovulatory surge of LH and FSH. This effect involves a change in pituitary responsiveness to GnRH, but whether or not estrogens also exert a positive effect on hypothalamic neurons that contributes to the "surge" of GnRH release at midcycle in primates is not yet resolved. The actions of estrogen and progesterone on the pituitary are the major factors that regulate the amount of LH released in each pulse (i.e., the amplitude of LH pulses). However, only progesterone has a physiological effect on the frequency of LH release; it decreases the frequency of firing of the hypothalamic pulse generator. These feedback effects of steroids, coupled with the intrinsic activity of the hypothalamic GnRH pulse generator, produce relatively frequent LH pulses of small amplitude in the follicular phase of the cycle and less frequent pulses of larger amplitude in the luteal phase. In males, the hypothalamic pulse generator also fires and releases GnRH in an episodic fashion, which causes the pulsatile release of LH necessary for normal testosterone production by the Leydig cells of the testis. Testosterone regulates the hypothalamic-pituitary-gonadal axis at both the hypothalamic and pituitary levels, and its negative feedback effect is mediated to a substantial degree by estrogen formed via aromatization. Thus, exogenous estrogen administration decreases

LH and testosterone levels in men, and antiestrogens such as clomiphene cause an elevation of serum LH, which can be used as a provocative test to evaluate the reproductive axis in men. In cycling women, the midcycle surge in gonadotropins stimulates follicular rupture and ovulation within 1 to 2 days. The ruptured follicle then develops into the corpus luteum, which produces large amounts of progesterone and estrogen under the influence of LH during the second half of the cycle. In the absence of pregnancy, the corpus luteum ceases to function after several days, steroid levels drop, and menstruation occurs. The luteal phase of the cycle is thus regulated by the limited 14-day functional lifetime of the corpus luteum. When steroid levels drop, the pulse generator reverts to a firing pattern characteristic of the follicular phase, the entire system then resets, and a new ovarian cycle occurs. Increased progesterone levels during the luteal phase of the cycle affect both the frequency and amplitude of LH pulses. Progesterone directly decreases the frequency of the hypothalamic pulse generator, which in turn decreases the frequency of LH pulses released from the pituitary. Progesterone also exerts a direct effect on the pituitary to oppose the inhibitory actions of estrogens and thus increase the amount of LH released (i.e., the amplitude of the LH pulses). When the ovaries are removed or cease to function, there is overproduction of FSH and LH, which are excreted in the urine. Measurement of urinary or plasma LH is a valuable clinical test and can be used to assess pituitary function and to show the effectiveness of replacement doses of estrogen, which will elicit a decline in LH levels. Although FSH levels will decline once hormonereplacement therapy is initiated, they do not return to normal, secondary to production of inhibin by the ovary (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). Consequently, the measurement of FSH levels as a means to monitor the effectiveness of hormonereplacement therapy is not clinically useful. Additional features of the regulation of gonadotropin secretion and actions are discussed in Chapters 56: Pituitary Hormones and Their Hypothalamic Releasing Factors and 59: Androgens. Effects of Cyclical Gonadal Steroids on the Reproductive Tract The cyclical changes in estrogen and progesterone production by the ovaries regulate corresponding events in the fallopian tubes, uterus, cervix, and vagina. Physiologically, these changes prepare the uterus for implantation if the ovum is fertilized, and the proper timing of events in these tissues is essential for a successful pregnancy. If pregnancy does not occur, the endometrium is shed and is visible externally as the menstrual discharge. The uterus is composed of an endometrium and a myometrium. The endometrium contains an epithelium lining the uterine cavity and an underlying stroma; the myometrium is the smooth muscle component responsible for uterine contractions. These cell layers, the fallopian tubes, cervix, and vagina display a characteristic set of responses to both estrogens and progestins. The changes typically associated with menstruation occur largely in the endometrium, which is shed during the menstrual discharge or menses (see Figure 583). The endometrium is the mucosa that lines the uterine cavity. The luminal surface of the endometrium is a layer of simple columnar epithelial secretory and ciliated cells. This epithelium is continuous with the openings of numerous glands that extend through the underlying stroma to the myometrial border. The endometrial epithelium connects distally to the mucus-secreting epithelium of the endocervix and proximally to the epithelium of the fallopian tube. Fertilization normally occurs in the fallopian tubes, so ovulation, transport of the fertilized ovum through the fallopian

tube, and preparation of the endometrial surface must be temporally coordinated if successful implantation is to occur. The endometrial stroma is a highly cellular connective tissue layer containing a variety of blood vessels that undergo cyclic changes associated with menstruation. The predominant cells in the stroma are fibroblasts, but substantial numbers of macrophages, lymphocytes, and other resident and migratory cell types also are present. By convention, menstruation is considered to mark the start of the menstrual cycle. During the follicular (or proliferative) phase of the cycle, estrogen begins the rebuilding of the endometrium by stimulating proliferation and differentiation: numerous mitoses become visible, the thickness of the layer increases, and characteristic changes occur in the glands and blood vessels of the tissue. These and subsequent effects of estrogens and progesterone are thought to be mediated in large part by the steroidal regulation of peptide growth factors and their cognate receptors that exert autocrine and paracrine actions in the endometrium. An important response to estrogen in the endometrium and other tissues is induction of the progesterone receptor, which enables cells to respond to this hormone during the second half of the cycle. In the luteal (or secretory) phase of the cycle, progesterone levels increase sharply due to secretion from the corpus luteum, and estrogens remain elevated. Progesterone limits the proliferative effect of estrogens on the endometrium by stimulating differentiation. Major effects include stimulation of secretions of the epithelium important for implantation of the blastocyst (the fertilized ovum at this stage of development) and its growth and promotion of the characteristic growth of the endometrial blood vessels seen at this time. Progesterone is thus important in preparation for implantation and for the changes that take place in the uterus at the implantation site (i.e., the decidual response). There is a rather narrrow "window of implantation," generally considered to span days 19 or 20 through day 24 of the endometrial cycle, when the epithelial cells of the endometrium are receptive to implantation of the blastocyst. Since endometrial status is regulated by estrogens and progestins, the efficacy of some contraceptives may be due in part to production of an endometrial surface that is not receptive to implantation. If pregnancy does not occur, the corpus luteum regresses due to lack of continued LH secretion, estrogen and progesterone levels fall, the endometrium cannot be maintained, and it is shed, resulting in the menstrual discharge as illustrated in Figure 583. If implantation occurs, human chorionic gonadotropin (hCG; see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors), produced initially by the trophoblast cells of the blastocyst and later by the placenta, interacts with the LH receptor of the corpus luteum to maintain steroid hormone synthesis during the early stages of pregnancy. In the later stages of pregnancy, the placenta becomes the major site of estrogen and progesterone synthesis. Estrogens and progesterone have important effects on the fallopian tube, myometrium, and cervix. In the fallopian tube, estrogens stimulate proliferation and differentiation, whereas progesterone inhibits these processes. Also, estrogens increase and progesterone decreases tubal muscular contractility, which affects transit time of the ovum to the uterus. Estrogens increase the amount and water content of cervical mucus and facilitate sperm penetration of the cervix, while progesterone generally has opposite effects. Estrogens favor rhythmic contractions of the uterine myometrium, while progesterone diminishes contractions. These effects are physiologically important and may contribute to some of the contraceptive actions of estrogens and progestins. Metabolic Effects Estrogens affect many tissues and have many metabolic actions in human beings and animals. It is

not clear in all cases if effects result directly from hormone actions on the tissue in question or secondarily from actions at other sites. However, it is now clear that many nonreproductive tissues (e.g., bone, vascular endothelium, liver, CNS, and heart) express low levels of estrogen receptors. Many metabolic effects of estrogens thus may result directly from receptor-mediated events in affected organs. While estrogens produce many metabolic responses, their effects on selected aspects of mineral, lipid, carbohydrate, and protein metabolism are particularly important for understanding their pharmacological actions. It has long been recognized that estrogens have positive effects on bone mass (reviewed by Riggs and Melton, 1992). Bone is continuously remodeled at sites called bone-remodeling units by the resorptive action of osteoclasts and the bone-forming action of osteoblasts (see Chapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds). Maintenance in total bone mass requires equal rates of formation and resorption as occurs in early adulthood (18 to 40 years); thereafter resorption predominates. Osteoclasts and osteoblasts contain functional ERs, androgen receptors (ARs), and progesterone receptors (PRs). Estrogens directly regulate osteoblasts and increase the synthesis of collagen type I, osteocalcin, osteopontin, osteonectin, alkaline phosphatase, and other markers of differentiated osteoblasts. However, the major effect of estrogens is to decrease the numbers and activity of osteoclasts. Much of the action of estrogens on osteoclasts appears to be mediated by altering cytokine (both paracrine and autocrine) signals from osteoblasts. Estrogens decrease osteoblast and stromal cell production of the osteoclast-stimulating cytokines interleukin (IL)-1, IL6, and tumor necrosis factor (TNF)- and increase the production of insulin-like growth factor (IGF)-1, bone morphogenic protein (BMP)-6, and transforming growth factor (TGF)- which are antiresorptive (reviewed by Spelsberg et al., 1999). Estrogens also increase osteoblast production of the cytokine osteoprotegrin (OPG), a soluble non-membrane-bound member of the TNF superfamily. OPG acts as a "decoy" receptor and antagonizes the binding of osteoprotegrin-ligand (OPG-L) to its receptor (termed RANK, or receptor activator of NF-Kappa B) and prevents the differentiation of osteoclast precursors to mature osteoclasts (Hofbauer et al., 2000). In addition, there are direct actions of estrogens on osteoclasts to increase the rate of apoptosis, which leads to a reduced number of osteoclasts. Estrogens affect bone growth and epiphyseal closure in both sexes. The importance of estrogen in the male skeleton is illustrated in a man with a completely defective ER who had osteoporosis, unfused epiphyses, increased bone turnover, and delayed bone age (Smith et al., 1994) and in the observation that male idiopathic osteoporosis is associated with reduced ER- expression (Braidman et al., 2000) in both osteocytes and osteoblasts. Estrogens have many effects on lipid metabolism; of major interest are their effects on serum lipoprotein and triglyceride levels (Lobo, 1991; Walsh et al., 1994). In general, estrogens slightly elevate serum triglycerides and slightly reduce total serum cholesterol levels. However, the more important actions are thought to be an increase in high-density lipoprotein (HDL) levels and a decrease in low-density lipoprotein (LDL) values and lipoprotein (a) [Lp(a)] (see Chapter 36: Drug Therapy for Hypercholesterolemia and Dyslipidemia). This beneficial ratio of HDL to LDL is an attractive side effect of estrogen therapy in postmenopausal women. The presence of estrogen receptors in the liver suggests that the beneficial effects of estrogen on lipoprotein metabolism are due in part to direct hepatic actions. Other sites of action, however, cannot be excluded. In addition to these effects on serum lipids, estrogens alter bile composition by increasing cholesterol secretion and decreasing bile acid secretion. This leads to increased saturation of bile with cholesterol and appears to be the basis for increased gallstone formation in some women receiving estrogens. Estrogen alone appears to slightly decrease fasting levels of glucose and insulin (Barrett-Connor and Laakso, 1990) but does not appear to have major effects on carbohydrate metabolism. Some older studies of combined oral contraceptives (which contained higher levels of both estrogens and

progestins than available now) suggested that estrogens might impair glucose tolerance, but it is uncertain whether these effects were due to the progestational or the estrogenic component of the older combined oral contraceptives. Estrogens have effects on many serum proteins, particularly those involved in hormone binding and clotting cascades. In general, estrogens tend to increase plasma levels of cortisol-binding globulin (CBG or transcortin), thyroxine-binding globulin (TBG), and sex steroidbinding globulin (SSBG), which binds both androgens and estrogens. Estrogens alter a number of metabolic pathways that affect the cardiovascular system (see Mendelsohn and Karas, 1999). Systemic effects include changes in lipoprotein metabolism and in hepatic production of plasma proteins. Estrogens cause a small increase in coagulation factors VII and XII and decrease the anticoagulation factors protein C, protein S, and antithrombin III. Fibrinolytic pathways also are affected, and several studies in women treated with estrogen alone or estrogen with a progestin have demonstrated decreased levels of plasminogen-activator inhibitor protein I (PAI-1) with a concomitant increase in fibrinolysis (Koh et al., 1997). Thus, both coagulation and fibrinolytic pathways are increased by estrogens, and imbalance in these two opposing activities may cause adverse effects. At relatively high concentrations, estrogens have antioxidant activity and may inhibit the oxidation of LDL (Sack et al., 1994) by affecting superoxide dismutase. Long-term administration of estrogen is associated with decreased plasma renin, angiotensin converting enzyme, and endothelin-1; angiotensin-1 receptor expression is decreased. Estrogen actions on the vascular wall include increased production of nitric oxide, which occurs within minutes, and induction of inducible nitric oxide synthase and increased production of prostacyclin, which occur more slowly. All of these changes promote vasodilation. Estrogens also promote endothelial cell growth while inhibiting the proliferation of vascular smooth muscle cells. Estrogen Receptors Estrogens exert their effects by interaction with receptors that are members of the superfamily of nuclear receptors (Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). There are two distinct estrogen receptors, ER and ER , which are products of separate genes. ER , the first discovered, is located in highest abundance in the female reproductive tractespecially the uterus, vagina, and ovaryas well as in the mammary gland, the hypothalamus, endothelial cells, and vascular smooth muscle. ER is expressed with a somewhat different tissue distribution, with highest expression in the prostate and ovaries and less expression in lung, brain, and vasculature. The two human ERs are 44% identical in overall aminoacid sequence (Figure 584) and share the domain structure common to members of this family. The estrogen receptor is divided into six functional domains: the NH2-terminal A/B domain contains the activation function-1 (AF-1) segment, which can activate transcription independently of ligand; the highly conserved C domain comprises the DNA binding domain, which contains four cysteines arranged in two zinc fingers; the D domain, frequently called the "hinge region," contains the nuclear localization signal; and the E/F domain has multiple functions, including ligand binding, dimerization, and ligand-dependent transactivation, mediated by the AF-2 domain. As illustrated in Figure 584, there are significant differences between the two receptor isoforms in the ligand binding domains and in both transactivation domains. The receptors appear to have different biological functions and may respond differently to various estrogenic compounds. For example, while both receptors bind 17 estradiol with the same K D (about 0.3 nM), the phytoestrogen genistein (Table 581) binds to ER with about fivefold higher affinity than to ER (Kuiper et al., 1997). However, the high degree of identity in the DNA-binding domains indicates that both receptors probably recognize similar DNA sequences and hence regulate many of the same target genes.

Figure 584. Domain Structure and Sequence Identity between Estrogen Receptor (ER) and ER . Nuclear hormone receptors can be divided into common functional domains. Activation function (AF)-1 and AF-2 represent transactivation domains. AF-2 is part of the carboxy terminal portion of the receptor, which is the region that also binds ligand. The highly conserved DNAbinding domain interacts with specific DNA sequences. Numbers within the boxes represent the last amino acid in a domain. The percentage of identical amino acids between functional domains in ER and ER is indicated.

Feamle transgenic mice homozygous for a disruption of ER are infertile; they have atrophic uteri and hyperemic ovaries that appear to lack corpora lutea (Lubahn et al., 1993) and lack uterotrophic responses to estradiol. Males lacking ER also are infertile, with abnormal testes and seminiferous tubules and inactive sperm, reduced bone density, and cardiovascular abnormalities. Female ER null animals are infertile, with an arrest in follicular development; males are fertile (Korach, 2000). Several variants of both ER and ER have been described, most frequently in breast cancer cells. These variants arise from use of alternate promoters or alternate splicing (Murphy et al., 1997). Some of these variants may act in an estrogen-independent manner (Fuqua et al., 1991), but their physiological significance is unclear. A truncated form of ER that suppresses transactivation of the wild-type ER and ER has been identified in the rat pituitary (Resnick et al., 2000), although its biological role and species distribution also are uncertain. Other variants of ER are caused by polymorphisms in the genes encoding the receptors, but attempts to correlate specific polymorphisms with the frequency of breast cancer (Roodi et al., 1995), bone mass (Kobayashi et al., 1996; Vandevyver et al., 1999), or endometrial cancer (Weiderpass et al., 2000) have led to contradictory results. Additional studies thus are required to determine how polymorphisms affect receptor structure and function and responses to estrogens. Mechanism of Action Both estrogen receptors are ligand-activated transcription factors that increase or decrease the synthesis of mRNA from target genes. After entering the cell by passive diffusion through the plasma membrane, the hormone binds to an ER in the nucleus. In the nucleus, the ER is present as a monomer, and upon binding estrogen, a change in conformation occurs that results in dimerization, which increases the affinity and the rate of receptor binding to DNA (Cheskis et al., 1997). The ER binds to estrogen response elements (EREs) in target genes with the consensus sequence GGTCANNNTGACC. The ER/DNA complex recruits one or more coactivator proteins to the promoter region (Figure 58 5B). The best-characterized coactivators are SRC-1 (steroid-receptor coactivator-1) and CBP (cyclic AMP response-element binding protein) (Collingwood et al., 1999). The coactivators have histone acetylase activity and/or recruit other proteins with this activity to the complex. Acetylation

of histones alters chromatin structure in the promoter region of target genes and allows the proteins that make up the general transcription apparatus to assemble and initiate transcription. Interaction of ER with an antagonist also promotes dimerization and DNA binding. However, an antagonist produces a conformation of ER that is different from an agonist-occupied receptor (Wijayaratne et al., 1999). The antagonist-induced conformation facilitates binding of corepressors such as NcoR/SMRT (nuclear hormone receptor corepressor/silencing mediator of retinoid and thyroid receptors) (Figure 585C). The corepressor/ER complex then further recruits other proteins with histone deacetylase activity such as HDAC1. Deacetylation of histones alters chromatin conformation and reduces the ability of the general transcription apparatus to form initiation complexes. The differences in the amino acid sequences of AF-1 and AF-2 in ER and ER suggest they interact with different specificity and affinity to coactivators and corepressors. Combined with the observation that ER and ER can form homo- and heterodimers (Cowley et al., 1997), a diverse array of different ER/coactivator or ER/corepressor proteins can be assembled on a target promoter. In cells that express both receptor isoforms, the action of ER appears to oppose the activity of ER (Hall and McDonnell, 1999). Figure 585. A. Unliganded estrogen receptor (ER) exists as a monomer within the nucleus. B. Agonist (blue oval) binds to ER and causes a ligand-directed change in conformation that facilitates dimerization and interaction with specific estrogen response element (ERE) sequences in DNA. The ER-DNA complex recruits coactivators such as steroid-receptor coactivator-1 (SRC-1) and other proteins such as cyclic AMPelement binding protein (CBP) which have histone acetylase activity (HAT). Acetylation of histones can cause a rearrangement of the nucleosomes and facilitates the interaction of the proteins that make up the general transcription apparatus (GTA) with subsequent synthesis of mRNA. C. Antagonist (blue triangles) binding to ER produces a different receptor conformation. The antagonist-induced conformation also facilitates dimerization and interaction with DNA, but a different set of proteins called corepressors, such as nuclear-hormone receptor corepressor (NcoR), are recruited to the complex. NcoR further recruits proteins such as histone deacetylase I (HDAC1) that act on histone proteins to stabilize nucleosome structure and prevent interaction with the GTA.

Besides coactivators and corepressors, both ER and ER can interact physically with other transcription factors such as Sp1 (Saville et al., 2000) or AP-1 (Paech et al., 1997), and these protein-protein interactions provide an alternate mechanism of action. Binding of ER-ligand complexes to a target gene in these circumstances is directed by Sp1 or AP-1 binding to their specific regulatory sequences, rather than ER binding to ERE sequences. This may explain how estrogens are able to regulate genes that lack a consensus ERE. Responses to agonists and antagonists mediated by these protein-protein interactions also are ER isoform- and promoterspecific. For example, 17 -estradiol induces transcription of a target gene controlled by an AP-1 site in the presence of an ER /AP-1 complex, but inhibits transcription in the presence of an ER /AP-1 complex. Conversely, antiestrogens are potent activators of ER /AP-1 but not of ER /AP-1 complexes. Other signaling systems may activate nuclear ER by ligand-independent mechanisms. Kato et al. (1995) demonstrated that phosphorylation of ER at serine 118 by microtubule-associated protein kinase (MAPK) activates the receptor. This provides a means of cross-talk between membranebound receptor pathways (i.e., EGF/IGF-1) that activate MAPK and the nuclear ER. Several studies have suggested that a form of estrogen receptor other than the nuclear receptors might be located on the plasma membrane, but it is unclear whether or not these receptors are encoded by the same gene that encodes the nuclear ERs (Razandi et al., 1999). These membranelocalized ERs may mediate the rapid activation of some proteins such as MAPK, which has been shown to be phosphorylated in several cell types within 5 to 10 minutes of 17 -estradiol addition (Endoh et al., 1997), or the rapid increase in cyclic AMP caused by the hormone (Aronica and Katzenellenbogen, 1993). The finding that MAPK is activated by estradiol provides an additional level of cross-talk with growth factors such as IGF-1 and EGF that activate this kinase pathway.

Membrane-localized ER also may be responsible for the rapid release of nitric oxide produced by estradiol treatment of endothelial cells. Absorption, Fate, Elimination Various estrogens are available for oral, parenteral, transdermal, or topical administration (see Table 581 for the structures of these agents). Given the lipophilic nature of estrogens, absorption generally is good with the appropriate preparation. Several preparations are available for oral use. Aqueous or oil-based esters of estradiol and estrone are available for intramuscular injection ranging in frequency from every several days to once per month. Transdermal patches that are changed once or twice weekly deliver estradiol (or a combination of estradiol and norethindrone acetate) continuously through the skin. Several preparations are available for topical use in the vagina. Oral administration is common and may utilize estradiol, conjugated estrogens, esters of estrone and other estrogens, and ethinyl estradiol. Estradiol itself was not used orally frequently because of extensive first-pass hepatic metabolism. A micronized preparation of estradiol (ESTRACE ) that yields a large surface for rapid absorption is available now for oral administration, although high doses must be used because absolute bioavailability remains low (Fotherby, 1996). Ethinyl estradiol frequently is used orally, either alone (ESTINYL) or with a progestin in oral contraceptives; the ethinyl substitution in the C 17 position inhibits first-pass hepatic metabolism. The absorption of these estrogens from the gastrointestinal tract is rapid and generally complete. Other common oral preparations contain conjugated equine estrogens (PREMARIN), which are primarily the sulfate esters of estrone, equilin, and other naturally occurring compounds, esterified esters (ESTRATAB ), or mixtures of conjugated estrogens prepared from plant-derived sources (CENESTIN ). These are hydrolyzed by enzymes present in the lower gut that remove the charged sulfate groups and allow absorption of estrogen across the intestinal epithelium. In another oral preparation (estropipate, ORTHO-EST, OGEN ), estrone is solubilized as the sulfate and stabilized with piperazine. Due largely to differences in metabolism, the potencies of various oral preparations differ widely; ethinyl estradiol, for example, is much more potent than conjugated estrogens. A number of foodstuffs and plant-derived products, largely from soy, are available as nonprescription items and often are touted as providing benefits similar to those from compounds with established estrogenic activity. These products may contain flavonoids such as genistein (Figure 581), coumestans, and lignans, which have been reported to possess estrogenic activity in laboratory tests, albeit generally much less than that of estradiol. In theory, these preparations could produce appreciable estrogenic effects. However, their efficacy at relevant doses has not been established in human trials, and many of these products contain many other compounds besides phytoestrogens, which could contribute to any effects (Mkelet al., 1999; Tham et al., 1998). Administration of estradiolvia transdermal patches (ESTRADERM, VIVELLE, ALORA, others) provides slow, sustained release of the hormone, systemic distribution, and more constant blood levels than are obtained with oral doses. In addition, the transdermal route does not lead to the high level of the drug that enters the liver via the portal circulation after oral administration, which may explain why the two routes are associated with different effects on lipoprotein profiles (Walsh et al., 1994, and the following section). Other preparations are available for intramuscular injection. When dissolved in oil and injected, esters of estradiol are well absorbed. The aryl and alkyl esters of estradiol become less polar as the size of the substituents increases; correspondingly, the rate of absorption of oily preparations is progressively slowed, and the duration of action can be prolonged. A single therapeutic dose of

compounds such as estradiol valerate (DELESTROGEN, VALERGEN, others) or estradiol cypionate (DEPO-ESTRADIOL CYPIONATE , others) may be absorbed over several weeks following a single intramuscular injection. Suspensions containing estrone or a combination of esters (primarily estrone and equilin sulfates) also may be given via intramuscular injection. Preparations of estradiol and conjugated estrogens are available for topical administration to the vagina. These are effective locally, but systemic effects also are possible, since significant absorption can occur from this site (Rigg et al., 1978). A 3-month vaginal ring (ESTRING ) also is available for slow release of estradiol. Estradiol, ethinyl estradiol, and other compounds exist in blood plasma extensively bound to plasma proteins. Estradiol and other naturally occurring estrogens are bound mainly to sex steroidbinding globulin (SSBG) and to a lesser degree to serum albumin. In contrast, ethinyl estradiol is bound extensively to serum albumin but not SSBG. Due to their size and lipophilic nature, unbound estrogens readily exit the plasma space and distribute extensively to tissue compartments. Variations in estradiol metabolism occur and depend upon the stage of the menstrual cycle and whether the individual is pre- or postmenopausal. In general, the hormone undergoes rapid hepatic biotransformation, with a plasma half-life measured in minutes. Estradiol is converted primarily by 17 -hydroxysteroid dehydrogenase to estrone, which undergoes conversion by 16 -hydroxylation and 17-keto reduction to estriol, which is the major urinary metabolite. A variety of sulfate and glucuronide conjugates also are excreted in the urine. Lesser amounts of estrone or estradiol are oxidized to the 2-hydroxycatechols by CYP3A in the liver and by CYP1A in extrahepatic tissues or to 4-hydroxycatechols by CYP1B1 in extrahepatic sites, with the 2-hydroxycatechol being formed to a greater extent. The 2- and 4-hydroxycatechols are largely inactivated by catechol-O-methyl transfereases (COMT). However, smaller amounts may be converted by cytochrome P450- or peroxidase-catalyzed reactions to yield semiquinones or quinones that are capable of forming DNA adducts or of generating (via redox cycling) reactive oxygen species that could oxidize DNA bases (Liehr, 2000). Estrogens also undergo enterohepatic recirculation via (1) sulfate and glucuronide conjugation in the liver, (2) biliary secretion of the conjugates into the intestine, and (3) hydrolysis in the gut followed by reabsorption. Ethinyl estradiol is cleared much more slowly than is estradiol due to decreased hepatic metabolism, and the elimination phase half-life has been reported in various studies to be 13 to 27 hours. Unlike estradiol, the primary route of biotransformation of ethinyl estradiol is via 2hydroxylation and subsequent formation of the corresponding 2- and 3-methyl ethers. Mestranol, another semisynthetic estrogen and a component of some combination oral contraceptives, is the 3methyl ether of ethinyl estradiol. In the body it undergoes rapid hepatic demethylation to ethinyl estradiol, which is its active form. Measurements of plasma and urinary estrogens and their metabolites are used for a number of purposes. In the normal menstrual cycle, the mean daily excretion of estrogens at the midcycle ovulatory maximum is 25 to 100 g; the second rise during the luteal phase is more prolonged, but the maximal rates of excretion are somewhat smaller (10 to 80 g per day). After menopause, the average excretion of estrogens in normal women is about 5 to 10 g daily, and estrogen synthesis occurs primarily from androgenic precursors in nonovarian tissues. The normal values for men are 2 to 25 g per day, quantities about equal to the urinary estrogens of women during the first week of the menstrual cycle. No estrogen is detectable in the urine of young children. During the first trimester of pregnancy, the placenta becomes the primary source of the urinary estrogens, which continue to increase and reach levels of about 30 mg per day near term. In the past, serial estrogen determinations were used to assess placental and fetal function. However, with the advent of fetal

monitors, this is no longer done. Untoward Responses Estrogens are highly efficacious for most of their therapeutic purposes. Hence the decision about their use is largely a matter of analyzing the risk-to-benefit ratio for each patient. Historically, the major concerns about the use of estrogens have been cancer, thromboembolic disease, changes in carbohydrate and lipid metabolism, hypertension, gallbladder disease, nausea, migraine, changes in mood, and several lesser side effects. The literature in this area is voluminous, complex, and often difficult to interpret, and a historical perspective is helpful to analyze risk-benefit issues for current preparations. First, many concerns arose initially from studies of early oral contraceptives, which contained higher doses of estrogens than used for other purposesfor example, hormone-replacement therapy in postmenopausal women. Since the untoward effects of estrogens are dose-dependent, the extrapolation of oral contraceptive side effects to other settings may not be appropriate. Second, it is recognized now that some of the deleterious effects of oral contraceptives originally attributed to estrogens are due to the progestational component. As a result of the recognition of the above two points, the amount of estrogens (and progestins) in oral contraceptives has been markedly decreased, and this has dramatically diminished the risks associated with current oral contraceptive preparations. Reports based on usage in the 1960s and 1970s established that unopposed estrogens caused endometrial carcinoma. Since these reports, however, estrogens have been used with progestins, which greatly diminish this risk. Finally, it is now recognized that the two major uses of estrogens, postmenopausal hormone replacement and contraception, have many substantial health benefits that previously were not appreciated. Concern About Carcinogenic Actions The possibility of developing cancer is probably the major concern for the use of estrogens and oral contraceptives. In several mammalian species, the administration of estrogens is followed by the development of certain tumors. Since the early studies of Lacassagne (1936), it has been known that estrogens can induce tumors of the breast, uterus, testis, bone, kidney, and several other tissues in various animal species. These early studies disseminated a fear of cancer resulting from estrogen use. In later reports (Greenwald et al., 1971; Herbst et al., 1971), an increased incidence of vaginal and cervical adenocarcinoma was noted in female offspring of mothers who had taken diethylstilbestrol (DES) during the first trimester of pregnancy. This may have resulted from the inability of the fetus to metabolize DES, leading to its accumulation in the fetus. The incidence of clear-cell vaginal and cervical adenocarcinoma in women who were exposed to estrogens in utero is only 0.01% to 0.1% (FDA Drug Bulletin, 1985), but these findings established for the first time that developmental exposure to estrogens was associated with an increase in a human cancer. Estrogen use during pregnancy also can increase the incidence of nonmalignant genital abnormalities in both male and female offspring. Thus, pregnant patients should not be given estrogens because of the possibility of such reproductive tract toxicities. While DES and other estrogens are no longer given intentionally to pregnant women, there is a concern that exposure during pregnancy to environmental substances with estrogenic activity may cause developmental abnormalities in the fetus (Mkelet al., 1999). Other studies established that the use of unopposed estrogen for hormone replacement in postmenopausal women is associated with the development of endometrial carcinoma (Shapiro et al., 1985). The risk is estimated to be increased as much as 5- to 15-fold by estrogen, depending

upon the dose and duration of use, but it declines to normal several years after discontinuation of estrogen. Other studies indicate a lower incidence of endometrial carcinoma when low doses of estrogen are combined with a progestin (Pike et al., 1997). The association between estrogen use and breast cancer remains uncertain. Part of the difficulty arises because of the frequency of the disease (about 1 in 8 in women who live 85 years) and the finding that 50% of women who develop the disease have no identifiable risk factors other than being female and aging. This makes determination of the association between estrogen use and breast cancer more difficult. An analysis of data from 51 epidemiologic studies of more than 150,000 women (Collaborative Group on Hormonal Factors in Breast Cancer, 1997) found that the risk of breast cancer increased slightly more than 2% for each year of estrogen-replacement therapy if treatment lasted between 1 and 4 years; treatment for more than 5 years had an accumulated risk of 35%. Lean women were at greater risk than heavier women. Recent studies have established that the progestin component in hormone-replacement therapy may play a major role in this increased risk of breast cancer (Schairer et al., 2000; Ross et al., 2000). Inclusion of the progestin component, administered either continuously or during only part of the estrogen treatment cycle, increased risk by about fourfold over estrogen-only use. Importantly, excess risk of breast cancer appears to be eliminated 5 years after cessation of hormone-replacement therapy. Historically, the carcinogenic actions of estrogens were thought to be related to their trophic effects by one of two mechanisms. First, an increase in cell proliferation would be expected to cause an increase in spontaneous errors associated with DNA replication. Second, after mutations were introduced into target cells by this or other mechanisms (e.g., chemical carcinogens), estrogens would enhance the replication of clones carrying such genetic errors. More recently, a third potential mechanism related to estrogen metabolism has been proposed. If catechol estrogens, especially the 4-hydroxycatechols, are converted to semiquinones or quinones prior to "inactivation" by COMT, these products, or reactive oxygen species generated during subsequent biotransformations, may cause direct chemical damage to DNA bases. In this regard, CYP1B1, which has specific estrogen-4-hydroxylase activity, is present in tissues such as uterus, breast, ovary, and prostate, which often give rise to hormone-responsive cancers (Liehr, 2000). Metabolic and Cardiovascular Effects Estrogens themselves generally have favorable overall effects on plasma lipoprotein profiles, although they may slightly elevate plasma triglycerides. (However, as noted in a later section dealing with hormone-replacement regimens, progestins may reduce the favorable actions of estrogens.) In contrast, estrogens do increase cholesterol levels in bile and cause a relative increase of two- to threefold in gallbladder disease. Oral estrogens increase the risk of venous thromboembolism about threefold in healthy women (Jick et al., 1996), appear to reduce the risk of cardiovascular disease (Grodstein et al., 1996; Henderson et al., 1991), and increase the relative risk of stroke by a factor of 1.35. In women with a history of cardiovascular disease, a threefold increase in the risk of venous thromboembolism but no reduction in the incidence of secondary cardiovascular events has been reported (Grady et al., 2000; Hulley et al., 1998). Currently prescribed doses of estrogens do not increase the risk of hypertension. Other Potential Untoward Effects Nausea and vomiting are an initial reaction to estrogen therapy in some women, but these effects may disappear with time and may be minimized by taking estrogens with food or just prior to

sleeping. Fullness and tenderness of the breasts and edema may occur but sometimes can be diminished by lowering the dose. A more serious concern is that estrogens may cause severe migraine in some women. Estrogens also may reactivate or exacerbate endometriosis and its attendant pain. Therapeutic Uses Estrogens are among the most commonly prescribed drugs in the United States. By far the two major uses are as a component of combination oral contraceptives (covered in a following section) and for hormone-replacement therapy in postmenopausal women. In addition they are used less frequently for a variety of other purposes. As already noted, the risk-to-benefit ratio is generally the major issue when considering their therapeutic use, as they are generally highly efficacious. The pharmacological considerations for estrogen use in oral contraceptives and postmenopausal hormone replacement are substantially different, primarily because of the doses used in the two settings. Historically, conjugated estrogens have been the most common agents for postmenopausal use, and 0.625 mg/day is effective in most women (although 1.25 mg is used if needed in some patients). In contrast, most combination oral contraceptives in current use employ 20 to 35 g/day of ethinyl estradiol. Conjugated estrogens and ethinyl estradiol differ widely in their oral potencies; for example, a dose of 0.625 mg of conjugated estrogens generally is considered equivalent to 5 to 10 g of ethinyl estradiol. It is important to recognize that the dose of estrogen used for postmenopausal hormone-replacement therapy is substantially less than that used in oral contraception, taking into account the different potencies of the drugs normally employed in the two settings. Since the untoward effects of estrogens are dose-dependent, the incidence and severity of side effects reported for oral contraceptives may be far greater than those for hormone replacement. In addition, much of the epidemiological literature examining the side effects of oral contraceptives is derived from studies of older preparations that generally contained 50 to 150 g of mestranol or ethinyl estradiol rather than 20- to 35- g doses most commonly used at present. Postmenopausal Hormone-Replacement Therapy The decline in the secretion of estrogen by the ovary is a slow and gradual process that continues for some years after menstruation has ceased (see Eskin, 1978). It is a frequent observation that menopausal symptoms are more severe following abrupt removal of estrogens, such as with oophorectomy, than with natural menopause. Of primary importance in the treatment of postmenopausal women with estrogens are the prevention of bone loss and the amelioration of vasomotor systems, which are established benefits of replacement therapy. Osteoporosis Osteoporosis is a disorder of the skeleton associated with the loss of both hydroxyapatite (calcium phosphate complexes) and protein matrix or colloid (see Chapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds for additional information). The result is thinning and weakening of the bones and an increased incidence of fractures, particularly compression fractures of the vertebrae and minimal trauma fractures of the hip and wrist. The frequency and severity of these fractures and their associated complications are a major public health problem, especially as the population continues to age. Twenty percent of patients who sustain hip fractures die within the first 12 months following the fracture, and many others are permanently disabled. Osteoporosis is a major indication for

estrogen replacement therapy, and it is clearly efficacious for this purpose. The primary mechanism by which estrogens act is to decrease bone resorption, and consequently estrogens are more effective at preventing rather than restoring bone loss (Prince et al., 1991; Belchetz, 1994). Estrogens are most effective if treatment is initiated before significant bone loss occurs, and their beneficial effects require continuous use; bone loss resumes when treatment is discontinued. An appropriate diet with adequate calcium and vitamin D intake and weight-bearing exercise support the effects of estrogen treatment. Higher doses of estrogens may lead to some increase in bone mass, and combinations of estrogens with calcium, fluoride, and/or other agents that increase bone mass are under study. Public health efforts to improve diet and exercise patterns in girls and young women also are rational approaches to increase bone mass prior to the onset of postmenopausal osteoporosis. Vasomotor Symptoms The decline in ovarian function at menopause is associated with vasomotor symptoms in most women due to deficiency of estrogen. The characteristic hot flashes may alternate with chilly sensations, inappropriate sweating, and paresthesias. Treatment with estrogen is specific and very effective (Belchetz, 1994), but if the drug is contraindicated or otherwise undesirable, other options may be considered (Young et al., 1990). Medroxyprogesterone acetate (discussed in the later section on progestins) may provide some relief of vasomotor symptoms for certain patients, and the 2-adrenergic agonist clonidine diminishes vasomotor symptoms in some women, presumably by blocking the CNS outflow that regulates blood flow to cutaneous vessels. In many women, hot flashes diminish within several years. Prevention of Cardiovascular Disease In past years, prevention of osteoporosis, hot flashes, and other symptoms associated with menopause were considered the major benefits of estrogen replacement. However, it is now clear that the leading cause of death in women in the United States over 65 years old is cardiovascular disease, particularly myocardial infarction. The incidence of cardiovascular disease due to atherosclerosis is low in premenopausal women, rises after menopause, and is reduced to premenopausal levels after estrogen-replacement therapy (see Mendelsohn and Karas, 1999). The protective effects of estrogen are mediated in part by systemic changes in lipoprotein metabolism as well as by direct effects of estrogens on blood vessels. Estrogens promote vasodilation, inhibit the response to vascular injury, and reduce atherosclerosis (see Mendelsohn and Karas, 1999). Estrogens accelerate endothelial cell growth both in vivo and in vitro; this may be due to estrogeninduced production of vascular endothelial growth factor within vessels. Estrogens participate in vascular protection by stimulating the proliferation and activity of endothelial cells and inhibiting the growth and migration of vascular smooth muscle cells. Numerous retrospective and prospective studies have concluded that in normal healthy women, estrogens reduce cardiovascular disease by 35% to 50% (Grodstein et al., 1996; Henderson et al., 1991). However, The Heart and Estrogen/Progestin Replacement Study (Hulley et al., 1998), which studied older women with documented cardiovascular disease, reported no change in the incidence of myocardial infarction despite an 11% reduction in LDL cholesterol. Thromboembolic disease and disease of the gallbladder were increased. These studies suggest that the beneficial effects of hormonereplacement therapy on risk of heart disease may be dependent on the age and the cardiovascular status of the patient. Neuroprotective Effects

Several retrospective studies have suggested that estrogens had beneficial effects on cognition and delayed the onset of Alzheimer's disease (see Green and Simpkins, 2000). Estrogens have been shown to exert neuroprotective effects both in vivo and in vitro, and various mechanisms, including activation of the MAPK pathway, inhibition of apoptosis, and decreased neurotoxicity of amyloid peptide have been suggested. Recent prospective studies have not confirmed these beneficial effects and have found no improvement in the symptoms of Alzheimer's disease (Henderson et al., 2000) or in the slow progression of the disease (Mulnard et al., 2000). Further large-scale studies are required to assess the neurological benefits of estrogens. Urogenital Atrophy Loss of tissue lining the vagina or bladder leads to a variety of symptoms in a very high percentage of postmenopausal women. These include dryness and itching of the vagina, pain during sexual intercourse, swelling of the tissues in the genital region such as the entrance to the vagina, pain during urination, a need to urinate urgently or often, and sudden or unexpected urinary incontinence. These symptoms are effectively treated by estrogens, administered either orally or locally as a vaginal cream (PREMARIN , others) or ring device (ESTRING ). Other Therapeutic Effects Many other changes occur in postmenopausal women, including a general thinning of the skin; changes in the urethra, vulva, and external genitalia; a decrease in height, a "camel hump" on the back, and a protuberant abdomen secondary to osteoporosis precipitated by estradiol deficiency; and a variety of changes including headache, tiredness, and difficulty concentrating, many of which may derive from the chronic lack of sleep created by hot flashes and other vasomotor symptoms. Estrogen replacement may help alleviate or lessen some of these via direct actions (e.g., improvement of vasomotor symptoms, direct skeletal effects) or secondary effects resulting in an improved feeling of well-being (Belchetz, 1994). Hormone-Replacement Regimens In the 1960s and 1970s there was an increase in estrogen-replacement therapy (i.e., estrogens alone) in postmenopausal women, primarily to reduce vasomotor symptoms, vaginitis, and osteoporosis. About 1980, epidemiological studies indicated that this treatment was associated with a large increase in the incidence of endometrial carcinoma, presumably due in part to the continuous stimulation of endometrial hyperplasia by unopposed estrogens. This realization has led to the use of hormone-replacement therapy that includes both an estrogen, for its beneficial effects, and a progestin to limit endometrial hyperplasia. While the actions of progesterone on the endometrium are complex, its effects on estrogen-induced hyperplasia may involve a decrease in estrogen receptor content, increased local conversion of estradiol to the less potent estrone via the induction of 17 -hydroxysteroid dehydrogenase in the tissue, and/or the conversion of the endometrium from a proliferative to a secretory state. Hormone-replacement therapy with both an estrogen and progestin now is recommended for most postmenopausal women with a uterus (Belchetz, 1994). For those women with a uterus who are unable to tolerate progestins or have a high risk of cardiovascular disease due to unfavorable lipoprotein profiles (see below), estrogens alone may be preferable. For women who have undergone a hysterectomy, endometrial carcinoma is not a concern, and estrogen alone is more commonly used because of possible deleterious effects of progestins (see below). Medroxyprogesterone acetate (MPA) is the progestin that, historically, has been most commonly

used in hormone-replacement regimens. This is a C 21 derivative of progesterone, which has less androgenic activity than other progestins such as the 19-nor compounds commonly used in combination oral contraceptives (see following section). This choice of a highly selective progestin such as MPA is an important pharmacological consideration, since 19-nor compounds may have undesirable effects on lipid and carbohydrate metabolism due in part to their androgenic activity. There has been considerable concern that the inclusion of a progestin would oppose the beneficial effects of estrogens on lipoprotein profiles and/or have other undesirable metabolic effects, and this is an area of current investigation. Several hormone-replacement regimens have been utilized. An example of a "cyclic" regimen is as follows: (1) the administration of an estrogen for 25 days; (2) the inclusion of MPA for the last 10 to 13 days of estrogen treatment; and (3) 5 to 6 days with no hormone treatment, during which withdrawal bleeding normally occurs due to breakdown and shedding of the endometrium. Many physicians do not use regimens that include days without hormone. The inclusion of the progestin during only a portion of each treatment cycle effectively decreases endometrial hyperplasia and the incidence of endometrial carcinoma yet minimizes the total amount of progestin administered. PREMPRO (PREMARIN plus MPA, given as a fixed dose daily) or PREMPHASE (PREMARIN given for 28 days plus MPA given for 14 out of 28 days) are widely used combination formulations. Two other combination products that recently became available in the United States are FEM HRT (estradiol plus norethindrone acetate) and ORTH PREFEST (estradiol and norgestimate). Efforts continue to develop regimens (e.g., FEM HRT ) that provide the beneficial effects of estrogen without the withdrawal bleeding that may limit compliance in some women. Another pharmacological consideration is the route of estrogen administrationi.e., oral versus transdermal. Oral administration exposes the liver to high concentrations of estrogens via the portal circulation and causes a more rapid conversion of estradiol or conjugated estrone to estrone. Both of these effects are lessened with transdermal estradiol. Either route effectively relieves vasomotor symptoms and protects against bone loss. Oral but not transdermal estrogen may increase SSBG, other binding globulins, and renin substrate; the oral route might be expected to cause greater increases in the cholesterol content of the bile. Transdermal estrogen appears to cause smaller beneficial changes in lipoprotein profiles (approximately 50% of those seen with the oral route), presumably because the liver is not exposed to the high estrogen levels seen after oral dosing (Walsh et al., 1994). Some women may have skin reactions to the transdermal patch. Estrogen Treatment in the Failure of Ovarian Development In several conditions, the ovaries do not develop and puberty does not occur. In ovarian dysgenesis with dwarfism (Turner's syndrome), therapy with estrogen at the appropriate time replicates the events of puberty except for the growth spurt and the changes in the ovary. The genital structures grow to normal size, the breasts develop, axillary and pubic hair grows, and the body assumes the normal feminine contour. Androgens (Chapter 59: Androgens) and/or growth hormone (Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors) may be used to promote normal growth. Failure of ovarian development also is associated with hypopituitarism in childhood. Deficiency of the thyroid and the adrenal cortex is corrected easily by replacement therapy, and the failure of sexual development is treated with estrogen. Administration of growth hormone permits achievement of near normal adult stature (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). Treatment with estrogen at the normal age of puberty can be expected to cause a small acceleration of growth, but the addition of small doses of androgen has a greater growth-promoting effect. While estrogens and androgens promote bone growth, they also

accelerate epiphyseal fusion, and their premature use can thus result in a shorter ultimate height. Earlier Uses of Estrogens Estrogens, particularly diethylstilbestrol, have been historically important in the treatment of testosterone-dependent prostate carcinoma due to the ability of estrogens to suppress pituitary LH secretion via negative feedback effects and thus inhibit testosterone production in the testes. More recently, the use of GnRH analogs has gained widespread acceptance for this purpose, since they may have fewer untoward side effects than estrogens in the treatment of male patients (see Chapter 59: Androgens). Selective Estrogen Receptor Modulators (SERMs) and Antiestrogens Recent advances in understanding how ligands alter the conformation of ERs has brought about a fundamental conceptual change in how estrogen agonists and antagonists act and made possible the revolutionary development of compounds with uniquely selective estrogenic properties. The simple model of an agonist binding to a single ER that subsequently affects transcription by the same molecular mechanism in all target tissues, and of antagonists that act by simple competition with agonist for binding, is no longer valid. By altering the conformation of the two different ERs and thereby changing interactions with coactivators and corepressors in a cell- and promoter-specific context, ligands may have a broad spectrum of activities from purely antiestrogenic in all tissues, to partially estrogenic in some tissues with antiestrogenic or no activities in others, to purely estrogenic activities in all tissues. The elucidation of these concepts has been a major breakthrough in estrogen pharmacology in the past decade and should permit the rational design of drugs with very selective patterns of estrogenic activity. SERMs: Tamoxifen, Raloxifene, and Toremifene SERMs are compounds whose estrogenic activities are tissue-selective. The pharmacological goal of these drugs is to produce estrogenic actions in those tissues where these actions are beneficial (e.g., bone, brain, liver during postmenopausal hormone replacement) and to have either no activity or antagonist activity in tissues such as breast and endometrium, where estrogenic actions (e.g., cellular proliferation) might be deleterious. Currently approved drugs in the United States in this class are tamoxifen citrate (NOLVADEX ) and raloxifene hydrocloride (EVISTA). Toremifene (FARESTON) is chemically related to tamoxifen (see below) and has similar actions. Toremifene is approved by the United States Food and Drug Administration (FDA) for treatment of metastatic breast cancer in postmenopausal women with ER-positive or ER-unknown tumors. Tamoxifen was originally classified as an antiestrogen, but subsequent experience has shown that it has agonist activity in bone, liver, and the endometrium. It currently is approved for four uses: (1) as an adjuvant for the treatment of axillary nodenegative breast cancer in women after total mastectomy or segmental mastectomy, axillary dissection, and breast irradiation; (2) as therapy following total mastectomy or segmental mastectomy, axillary dissection, and breast irradiation in postmenopausal women with node-positive disease; (3) in the treatment of women and men with advanced or metastatic breast cancer; and (4) as a preventative agent for women at high risk for breast cancer. Raloxifene is approved for the prevention and treatment of osteoporosis in postmenopausal women. Antiestrogens: Clomiphene and ICI 182,780 These compounds are distinguished from the SERMs in that they are pure antagonists in all tissues

studied. They may have inverse agonist activity in some settings. Clomiphene (CLOMID, SEROPHENE, others) is approved for the treatment of anovulatory women desiring pregnancy. ICI 182,780 (FASLODEX) is in clinical trials for the treatment of breast cancer and may be efficacious in women who are resistant to tamoxifen. Chemistry The structures of the trans-isomer of tamoxifen, and of raloxifene, trans-clomiphene, (enclomiphene), and ICI 182,780 are as follows:

Tamoxifen belongs to the triphenylethylene class of compounds derived from the same stilbene nucleus as diethylstilbestrol; compounds of this class display a variety of estrogenic and antiestrogenic activities. In general, the trans conformations have antiestrogenic activity, whereas the cis conformations display estrogenic activity. However, the pharmacological activity of the trans compound depends upon the species, target tissue, and gene. Hepatic metabolism produces Ndesmethyltamoxifen, which has affinity for ER comparable to that of tamoxifen, and the highly active 4-hydroxy metabolite, which has a much (25- to 50-fold) higher affinity for both ER and ER than does tamoxifen (Kuiper et al., 1997). In addition, the 4-hydroxy metabolites formed in vivo isomerize readily, and this complicates the comparison of in vivo effects of the drugs with their in vitro actions. Tamoxifen is marketed as the pure trans isomer. Toremifene also is a

triphenylethelene with a chlorine substitution at the R2 position. Raloxifene is a polyhydroxylated nonsteroidal compound with a benzothiophene core. Raloxifene binds with high affinity for both ER and ER (Kuiper et al., 1997). Clomiphene citrate is a triphenylethylene, and its two isomers, zuclomiphene (cis-chlomiphene) and enclomiphene (trans-clomiphene) are a weak estrogen agonist and a potent antagonist, respectively. Clomiphene binds to both ER and ER , but the individual isomers were not examined (Kuiper et al., 1997). ICI 182,780 (FASLODEX) is a 7 -alkylamide derivative of estradiol that interacts with both ER and ER (Van Den Bemd et al., 1999). Pharmacological Effects Tamoxifen exhibits antiestrogenic, estrogenic, or mixed activity depending upon the species and target gene measured. In clinical tests or laboratory studies with human cells, the drug's activity depends upon the tissue and endpoint measured (Jordan and Murphy, 1990). For example, tamoxifen inhibits the proliferation of cultured human breast cancer cells and reduces tumor size and number in women (reviewed in Jiayesimi et al., 1995), yet it stimulates proliferation of endometrial cells and causes endometrial thickening (Lahti et al., 1993). The drug has an antiresorptive effect on bone, and it decreases total cholesterol, LDL, and lipoprotein (a) but does not increase HDL and triglycerides (Love et al., 1994) in human beings. The lipid-lowering effect appears to be greater in postmenopausal than premenopausal women (Ilanchezhian et al., 1995). Tamoxifen treatment causes a two- to threefold increase in the relative risk of deep vein thrombosis and pulmonary embolism (Fisher et al., 1998). Some investigators have reported that tamoxifen decreases overall cardiovascular risk (Rutqvist and Mattsson, 1993), while others have seen no change (Fisher et al., 1998). Tamoxifen produces hot flashes in some women, which is the expected vasomotor effect of antiestrogens; other adverse effects include vaginal discharge or dryness, cataracts, and nausea. Raloxifene is an estrogen agonist in bone, where it exerts an antiresorptive effect. Results of several large clinical trials have shown that raloxifene reduces the number of vertebral fractures by up to 50% in a dose-dependent manner (Delmas et al., 1997; Ettinger et al., 1999). The drug also acts as an estrogen agonist in reducing total cholesterol and LDL, but it does not increase HDL or normalize plasminogen-activator inhibitor 1 in postmenopausal women (Walsh et al., 1998). Raloxifene does not cause proliferation or thickening of the endometrium. Preclinical studies indicate that raloxifene has an antiproliferative effect on ER-positive breast tumors and on proliferation of ER-positive breast cancer cell lines (see Hol et al., 1997) and significantly reduces the risk of ER-positive but not ER-negative breast cancer (Cummings et al., 1999). Raloxifene does not alleviate the vasomoter symptoms associated with menopause. Adverse effects include hot flashes and leg cramps; more serious adverse effects include about a threefold increase in deep vein thrombosis and pulmonary embolism (Cummings et al., 1999). Initial animal studies with clomiphene showed slight estrogenic activity, now thought to be due to the cis isomer, and moderate antiestrogenic activity, but the most striking effect was the inhibition of the pituitary's gonadotropic function. In both male and female animals, the compound thus acted as a contraceptive. In contrast, the most prominent effect in women was enlargement of the ovaries, and Greenblatt and coworkers (1962) found that the drug induced ovulation in many patients with amenorrhea, Stein-Leventhal syndrome, and dysfunctional bleeding with anovulatory cycles. This is the basis for clomiphene's major pharmacological use, which is the induction of ovulation in women

with a functional hypothalamic-hypophyseal-ovarian system and adequate endogenous estrogen production. In some cases, clomiphene is used in conjunction with human menotropins and hCG (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors) to induce ovulation. ICI 182,780 and its less potent forerunner ICI 164,384 have been purely antiestrogenic in studies to date. Effects of ICI 182,780 include inhibition of the activity of P-glycoprotein and of gene expression of aromatase, IGF-1, and insulin receptor substrate 1 and antiprogestin activity (see Ibrahim and Hortobagyi, 1999). In vitro, ICI 782,780 was more potent than 4-hydroxytamoxifen (DeFriend et al., 1994) in inhibiting proliferation of breast cancer cells, and in a small clinical trial, ICI 782,780 was partially successful in treating tamoxifen-resistant breast cancers (Howell et al., 1995). All of these agents bind to the ligand-binding pocket of both ER and ER and competitively block estradiol binding. However, the conformation of the ligand-bound ER is different with different ligands (see McDonnell, 2000). Thus, the conformation of ER when occupied by 17 -estradiol is different from its conformation when occupied by tamoxifen, raloxifene, or ICI 182,780 (Wijayaratne et al., 1999), and this has two important mechanistic consequences. These distinct ERligand conformations appear to recruit different coactivators and corepressors onto the promoter of a target gene by differential protein-protein interactions at the receptor surface. The tissue-specific actions of SERMs thus can be explained in part by the distinct conformation of the ER when occupied by different ligands, in combination with the repertoire of coactivators and corepressors present in a specific cell type that also will affect the nature of ER complexes formed. The ER contains multiple transcriptional activation functions (i.e., AF-1 and AF-2), which can be differentially regulated by individual SERMs and contribute to the drug's pharmacological activities. For example, tamoxifen does not produce an ER conformation that allows activation of the AF-2 function of the ER, but binding of the drug allows the AF-1 domain to become functional. Tamoxifen may thus be an agonist for activation of transcription of some genes in tissues, such as the endometrium, which either do not require the AF-2 function for transcriptional control or contain other transcription factors that can substitute for the AF-2 function of the ER. However, the drug also may be an antagonist for transcription of certain genes in tissues, such as the breast, which require the AF-2 function for expression and do not contain other factors that can provide this activity to the transcription machinery. Additionally, the presence of ER and the opposing ER suggest that the ratio of the two receptors also is important in specifying the response of a cell. This model might explain the eventual resistance to tamoxifen that develops with time in breast cancer; i.e., changes in the amount or the nature of coactivators and corepressors expressed or changes in ratio of ER to ER may occur during tumor progression. In fact, overexpression of the coactivator SRC-1 in HeLa cells changes tamoxifen from an antagonist to an agonist (Smith et al., 1997). Similarly, raloxifene acts as a partial agonist in bone but does not stimulate endometrial proliferation in postmenopausal women. Presumably this is due to some combination of differential expression of transcription factors in the two tissues and the effects of this SERM on ER conformation. Clomiphene acts to oppose the negative feedback effect of endogenous estrogens to increase gonadotropin secretion and stimulate ovulation. Most studies have found that clomiphene increases the amplitude of LH and FSH pulses, without a change in pulse frequency (Kettel et al., 1993). This suggests that the drug is acting largely at the pituitary level to block inhibitory actions of estrogen on gonadotropin release from the gland, and/or is somehow affecting the hypothalamus to release larger amounts of GnRH per pulse. Clomiphene also has been used in men to stimulate

gonadotropin release and enhance spermatogenesis. ICI 182,780 binds to ER and ER and represses transactivation, but it also appears to increase intracellular proteolytic degradation of ER , while apparently protecting ER from degradation (Van Den Bemd et al., 1999). Absorption, Fate, and Excretion Tamoxifen is given orally, and peak plasma levels are reached within 4 to 7 hours after treatment. This drug displays two elimination phases with half-lives of 7 to 14 hours for the first phase and 4 to 11 days for the second. Due to the prolonged half-life, 3 to 4 weeks of treatment may be required to reach steady-state plasma levels. The parent drug is converted largely to metabolites within 4 to 6 hours after oral administration. In human beings and other species, 4-hydroxytamoxifen is produced via hepatic metabolism, and this compound is considerably more potent than the parent drug as an antiestrogen. The major route of elimination from the body involves N-demethylation and deamination. The drug undergoes enterohepatic circulation, and excretion is primarily in the feces as conjugates of the deaminated metabolite. Raloxifene is adsorbed rapidly after oral administration and has an absolute bioavailability of about 2%. The drug has a half-life of about 28 hours and is eliminated primarily in the feces after hepatic glucuronidation. Clomiphene is well absorbed following oral administration, and the drug and its metabolites are eliminated primarily in the feces and to a lesser extent in the urine. The long plasma half-life (approximately 5 to 7 days) is due largely to plasma-protein binding, enterohepatic circulation, and accumulation in fatty tissues. Other active metabolites with long half-lives also may be produced. ICI 182,780 is administered intramuscularly with monthly (depot) injections being favored. Longacting preparations are available for investigational use. Therapeutic Uses Breast Cancer Tamoxifen is widely used in the treatment of breast cancer, and numerous studies have established its beneficial effects in this setting. It is used alone for palliation of advanced breast cancer in women with ER-positive tumors, and it is now indicated as the hormonal treatment of choice for both early and advanced breast cancer in women of all ages (Jiayesimi et al., 1995). Response rates are approximately 50% in women with ER-positive tumors and 60% to 70% in tumors that are both ER- and progestin-receptor (PR)-positive; response in ER-negative tumors falls to 10%. Tamoxifen has been shown consistently to increase disease-free survival and overall survival; treatment for 5 years has reduced cancer recurrance by 47% to 50% and death by 26% to 28%. The incidence of contralateral breast cancer has been reduced by 47% (Early Breast Cancer Trialists' Group, 1998). Tamoxifen has been approved by the FDA for primary prevention of breast cancer in women at high risk, based on the results of the National Surgical Adjuvant Breast and Bowel Project (Fisher et al., 1998), where the drug caused a 49% decrease in the incidence of invasive breast cancer and a 50% reduction of noninvasive breast cancer. Treatment efficacy decreases after 5 years due to acquisition of drug resistance by the tumors. Tamoxifen has estrogenic activity in the uterus, increases the risk of endometrial cancer by two- to threefold, and also causes a comparable increase in the risk of thromboembolemic disease.

Osteoporosis Raloxifene reduces the rate of bone loss at both distal sites and in the spinal column and may increase bone mass at certain sites. The rate of vertebral fractures can be reduced by as much as 50% (Delmas et al., 1997, Ettinger et al., 1999). While peripheral bone mass was increased by more than 2%, there was no decrease in nonvertebral fractures during the 3-year study period. Raloxifene does not appear to increase the risk of developing endometrial cancer. Raloxifene has beneficial actions on lipoprotein metabolism, reducing both total cholesterol and LDL; however, HDL is not increased, as with estrogen-replacement therapy. Adverse effects include deep vein thrombosis and leg cramps. Ovulatory disturbances are present in 15% to 25% of couples with infertility. For more than 25 years, clomiphene has been used in these cases because it has a low cost, is orally active, and requires less extensive monitoring than do other treatment protocols. However, the drug may exhibit untoward effects including ovarian hyperstimulation; increased incidence of multiple births; ovarian cysts; antiestrogenic effects on the developing follicle, endometrium, and cervical mucus that may counteract its beneficial effects on gonadotropin release; hot flashes; and blurred vision. In addition, clomiphene-induced cycles have a relatively high incidence of luteal phase dysfunction due to inadequate progesterone production. For these reasons, other treatment strategies such as human menopausal gonadotropins in combination with long-acting GnRH analogs and human chorionic gonadotropin may be more favorable (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). There also are reports of teratogenic effects of clomiphene in animals, and the drug should not be administered to pregnant women, even though there is no evidence of human fetal abnormalities in cases where the drug has been used to induce ovulation. Clomiphene also may be used to evaluate the male reproductive system, since testosterone feedback on the hypothalamus and pituitary is mediated to a large degree by estrogens formed from aromatization of the androgen. In normal individuals, once-daily administration of clomiphene for 7 days produces a doubling in plasma LH and a 50% increase in FSH. Estrogen-Synthesis Inhibitors Several agents can be used to decrease the effects of endogenous estrogens by blocking their biosynthesis. One such option is the continual administration of GnRH or the use of long-acting GnRH agonists, either of which prevents ovarian synthesis of estrogens but not the peripheral synthesis of estrogens from adrenal androgens (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). Aminoglutethimide inhibits aromatase activity, and thus blocks estrogen biosynthesis from all precursors. This agent is not selective, however, as it inhibits other cytochrome P450s involved in steroidogenesis, including those in the adrenal. There is increasing interest in the use of aromatase inhibitors to block specifically the local production of estrogens that may contribute substantially to hormone-responsive diseases such as breast cancer (see Chapter 52: Antineoplastic Agents). Third-generation aromatase inhibitors now are available and are more potent and selective than earlier agents such as aminoglutethimide (Brodie and Njar, 2000). These include both steroidal [e.g., formestane and exemestane (AROMASIN)] and nonsteroidal agents [e.g., anastrozole (ARIMIDEX), letrozole (FEMARA), and vorozole], which have been used for the second-line treatment of breast cancer in patients for whom tamoxifen therapy is unsuccessful. The structures of exemestane and anastrozole are as follows:

Progestins The progestins (see Figure 586) include the naturally occurring hormone progesterone, 17 acetoxyprogesterone derivatives in the pregnane series, 19-nortestosterone derivatives (estranes), and norgestrel and related compounds in the gonane series. (Note: Compounds with biological activities similar to those of progesterone have been variously referred to in the literature as progestins, progestational agents, progestagens, progestogens, gestagens, or gestogens.) Medroxyprogesterone acetate (MPA) and megestrol acetate are C 21 steroids with selective activity very similar to that of progesterone itself. MPA and oral micronized progesterone are widely used with estrogens for postmenopausal hormone replacement and other situations where a selective progestational effect is desired, and a depot form of MPA is used as a long-acting injectable contraceptive. The 19-nortestosterone derivatives were developed for use as progestins in oral contraceptives, and while their predominant activity is progestational, they exhibit androgenic and other activities. The gonanes are a more recently developed series of "19-nor" compounds, containing an ethyl rather than a methyl substituent in the 13-position, and they have diminished androgenic activity. These two classes of 19-nortestosterone derivatives are the progestational components of all oral and some long-acting injectable contraceptives. Figure 586. Structural Features of Various Progestins.

History Corner and Allen originally isolated a hormone in 1933 from the corpora lutea of sows and named it progestin. The next year, several European groups also isolated the crystalline compound and gave it the name luteo-sterone, unaware of the previous name given by Corner and Allen. This difference in nomenclature was resolved in 1935 at a garden party in London given by the famous English pharmacologist Sir Henry Dale, who helped persuade all parties that the name progesterone was a suitable compromise incorporating elements of the two previous designations. Studies with progesterone initially were hampered because production from animal sources was extremely difficult and time-consuming. Prices of progesterone were as high as $1000 per gram, which were astronomical in the worldwide economy of the 1930s. In addition, the fact that progesterone itself is not orally active, due to extensive first-pass hepatic metabolism, limited its pharmacological utility. These difficulties were overcome by two major advances in steroid chemistry made by several brilliant chemists (see Perone, 1994). The first advance was the synthesis of progesterone by Russel Marker from the plant product diosgenin in the 1940s. The synthesis was a real breakthrough, as it provided large amounts of relatively inexpensive progesterone and eliminated cross-contamination with compounds that might copurify with progesterone from animal sources. The second major

chemical advance was the synthesis of 19-nor compounds, the first orally active progestins, in the early 1950s by Carl Djerassi, who synthesized norethindrone at Syntex, and Frank Colton, who synthesized the isomer norethynodrel at Searle. These are undoubtedly some of the most important advances in synthetic organic chemistry in the twentieth century, since they eventually led to the development of effective oral contraceptives, agents that have had an enormous impact on society. Chemistry Unlike the estrogen receptor, which requires a phenolic A ring for high-affinity binding, the progesterone receptor favors a 4-3-one A-ring structure in an inverted 1 ,2 -conformation (Duax et al., 1988). Other steroid hormone receptors also bind this nonphenolic A-ring structure, although the optimal conformation differs from that for the progesterone receptor. Thus, some synthetic progestins (especially the 19-nor compounds) display limited binding to glucocorticoid, androgen, and mineralocorticoid receptors, a property that probably accounts for some of their nonprogestational activities. The spectrum of activities of these compounds is highly dependent upon specific substituent groups, especially the nature of the C 17 substitutent in the D ring, the presence of a C 19 methyl group, and the presence of an ethyl group at position C 13. One major class of agents is similar to progesterone and its metabolite 17 -hydroxyprogesterone (see Figure 586). 17 -Hydroxyprogesterone itself is inactive, but some of its ester derivatives have progestational activity. Compounds such as hydroxyprogesterone caproate have progestational activity but must be used parenterally due to first-pass hepatic metabolism. However, substitutions of such 17-esters at the 6-position of the B ring yield orally active compounds such as medroxyprogesterone acetate and megestrol acetate. Compounds in this chemical class display relatively selective progestational activity. The second major class of compounds is the 19-nor group. 19-norprogesterone (similar to progesterone except that the C 19 methyl group is replaced by a hydrogen atom) has potent progestational activity, and 19-nortestosterone derivatives display primarily progestational rather than androgenic activity. An ethinyl substituent at C 17 decreases hepatic metabolism and yields 19-nortestosterone analogs such as norethindrone, norethindrone acetate, norethynodrel, and ethynodiol diacetate, which are orally active. The activity of the latter three compounds is due primarily to their rapid in vivo conversion to norethindrone. These compounds are less selective than the 17 -hydroxyprogesterone derivatives mentioned above and have varying degrees of androgenic activity and, to a lesser extent, estrogenic and antiestrogenic activities. Replacement of the 13-methyl group of norethindrone with a 13-ethyl substituent yields the gonane norgestrel, which is a more potent progestin than the parent compound and has less androgenic activity. Norgestrel is a racemic mixture of an inactive dextrorotary isomer and the active levorotary isomer, levonorgestrel. Preparations containing half as much levonorgestrel as norgestrel thus have equivalent pharmacological activity. More recently, other gonanesincluding gestodene, norgestimate, and desogestrelhave become available and are reported to have very little if any androgenic activity at therapeutic doses (Rebar and Zeserson, 1991). The latter are used as the progestin component in the so-called third-generation combination oral contraceptives. Synthesis and Secretion Progesterone is secreted by the ovary mainly from the corpus luteum during the second half of the menstrual cycle, as illustrated in Figure 583. Secretion actually begins just before ovulation from the follicle that is destined to release an ovum. The formation of progesterone from steroid precursors is presented in detail in Chapter 60: Adrenocorticotropic Hormone; Adrenocortical

Steroids and Their Synthetic Analogs; Inhibitors of the Synthesis and Actions of Adrenocortical Hormones and occurs in the ovary, testis, adrenal cortex, and placenta. The stimulatory effect of LH on progesterone synthesis and secretion by the corpus luteum is mediated by a membrane-bound receptor linked to a G proteincoupled signal transduction pathway that increases the synthesis of cyclic AMP by stimulation of adenylyl cyclase (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). If the ovum is fertilized, implantation takes place about 7 days later, and almost at once the developing trophoblast begins secreting hCG into the maternal circulation, thereby sustaining the functional life of the corpus luteum. hCG, detectable in urine several days before the expected time of the next menstrual period, is excreted in progressively increasing amounts for the next 5 weeks or so and in reduced quantities thereafter throughout pregnancy. During the second or third month of pregnancy, the developing placenta begins to secrete estrogen and progesterone in collaboration with the fetal adrenal glands, and thereafter the corpus luteum is not essential to continued gestation. Estrogen and progesterone continue to be secreted in large amounts by the placenta up to the time of delivery. Measurements of the rate of secretion of progesterone suggest that, from a few milligrams per day secreted during the follicular phase of the cycle, the rate increases to 10 to 20 mg during the luteal phase and to several hundred mg during the latter part of pregnancy. Rates of 1 to 5 mg per day have been measured in men and are comparable to the values in women during the follicular phase of the cycle. Physiological and Pharmacological Actions Neuroendocrine Actions As discussed in a previous section, progesterone produced in the luteal phase of the cycle has several physiological effects. It decreases the frequency of the hypothalamic pulse generator and increases the amplitude of LH pulses released from the pituitary. Reproductive Tract Progesterone released during the luteal phase of the cycle decreases estrogen-driven endometrial proliferation and leads to the development of a secretory endometrium (see Figure 583). The abrupt decline in the release of progesterone from the corpus luteum at the end of the cycle is the main determinant of the onset of menstruation. If the duration of the luteal phase is artificially lengthened, either by sustaining luteal function or by treatment with progesterone, decidual changes in the endometrial stroma similar to those seen in early pregnancy can be induced. Under normal circumstances, estrogen antecedes and accompanies progesterone in its action upon the endometrium and is essential to the development of the normal menstrual pattern. Progesterone also influences the endocervical glands, and the abundant watery secretion of the estrogen-stimulated structures is changed to a scant, viscid material. As noted previously, these and other effects of progestins decrease penetration of the cervix by sperm. The estrogen-induced maturation of the human vaginal epithelium is modified toward the condition of pregnancy by the action of progesterone, a change that can be detected in cytological alterations in the vaginal smear. If the quantity of estrogen concurrently acting is known to be adequate, or if it is assured by giving estrogen, the cytological response to a progestin can be used to evaluate its

progestational potency. Progesterone is very important for the maintenance of pregnancy. Major actions of the hormone are to suppress menstruation and uterine contractility, but other effects also may be important. These effects to maintain pregnancy have led to the historical use of progestins to prevent threatened abortion. However, such treatment is of questionable benefit, probably because diminished progesterone is infrequently the cause of spontaneous abortion (see below). Mammary Gland During pregnancy and to a minor degree during the luteal phase of the cycle, progesterone, acting with estrogen, brings about a proliferation of the acini of the mammary gland. Toward the end of pregnancy, the acini fill with secretions, and the vasculature of the gland is notably increased; however, only after the levels of estrogen and progesterone decrease at parturition does lactation begin. During the normal menstrual cycle, mitotic activity in the breast epithelium is very low in the follicular phase and then peaks in the luteal phase. This pattern is due to progesterone, which triggers a single round of mitotic activity in the mammary epithelium. This effect is transient, however, and continued exposure to the hormone is rapidly followed by arrest of growth of the epithelial cells (Clarke and Sutherland, 1993). This is in contrast to the endometrium, where proliferation is greatest in the follicular phase due to increasing estrogen levels and is opposed by progesterone in the second half of the cycle. The hormonal control of proliferation is thus different in these two tissues, and these cell-specific effects should be kept in mind when therapeutic and untoward effects of the two agents are being interpreted. CNS Effects If the body temperature is carefully measured each day throughout the normal menstrual cycle, an increase of about 1F (0.56C) may be noted at midcycle; this correlates with ovulation. The temperature rise persists for the remainder of the cycle until the onset of the menstrual flow. This increase in temperature clearly is due to progesterone, as can be shown by administration of the hormone. The exact central mechanism of this effect is unknown at present, but an alteration of the temperature regulatory center in the hypothalamus may be involved. Progesterone also increases the ventilatory response of the respiratory centers to carbon dioxide and leads to reduced arterial and alveolar PCO2 in the luteal phase of the menstrual cycle and during pregnancy. Progesterone also may have depressant and hypnotic actions in the CNS, which may account for reports of drowsiness after hormone administration. This potential untoward effect may be abrogated by giving progesterone preparations at bedtime, which may even help some patients sleep. Metabolic Effects Progestins have numerous metabolic actions. Progesterone itself increases basal insulin levels and the rise in insulin after carbohydrate ingestion, but it does not normally cause a change in glucose tolerance. However, long-term administration of more potent progestins, such as norgestrel, may decrease glucose tolerance (Godsland, 1996). Progesterone stimulates lipoprotein lipase activity and seems to enhance fat deposition. Progesterone and analogs such as MPA have been reported to increase LDL and cause either no effects or modest reductions in serum HDL levels. The 19norprogestins may have more pronounced effects on plasma lipids because of their androgenic

activity. In this regard, a large prospective study has shown that MPA decreases the favorable HDL increase caused by conjugated estrogens during postmenopausal hormone replacement but does not significantly affect the beneficial effect of estrogens to lower LDL. In contrast, micronized progesterone does not significantly affect beneficial estrogen effects on either HDL or LDL profiles (The Writing Group for the PEPI Trial, 1995). Progesterone also may diminish the effects of aldosterone in the renal tubule and cause a decrease in sodium reabsorption that may increase mineralocorticoid secretion from the adrenal cortex. Mechanism of Action There is a single progesterone receptor (PR) gene that produces two isoforms of the progesterone receptor, PR-A and PR-B. The first 164 amino acids of PR-B are missing from PR-A; this occurs by use of two distinct estrogen-dependent promoters in the PR gene. Both PRs have a modular, domain structure common to all members of the nuclear receptor subfamily, as illustrated in Figure 584. Since the ligand-binding domain is identical in both isoforms of PR, there is no difference in ligand binding, as is seen with the two isoforms of ER. In the absence of ligand, PR is present in the nucleus in an inactive monomeric state bound to heat-shock proteins (HSP-90, HSP-70 and p59). Upon binding progesterone, the heat shock proteins dissociate, and the receptors are phosphorylated and subsequently form dimers (homo- and heterodimers) that bind with high selectivity to PREs (progesterone response elements) located on target genes (see Giangrande and McDonnell, 1999). Transcriptional activation by PR occurs primarily via recruitment of coactivators such as SRC-1, NcoA-1 or NcoA-2 (see Collingwood et al., 1999) or by direct interaction with general transcription factors such as TFIIB (Ing, 1992). The receptor coactivator complex then favors further interactions with yet additional proteins such as CBP and p300, which have histone acetylase activity. Histone acetylation causes a remodeling of chromatin that increases the accessibility of general transcriptional proteins, including RNA polymerase II, to the target promoter. Progesterone antagonists also facilitate receptor dimerization and DNA binding, but, as with ER, the conformation of antagonist-bound PR is different from that of agonist-bound PR. This different conformation favors PR interaction with corepressors NcoR/SMRT, which recruit histone deacetylases. Histone deacetylation increases DNA interaction with nucleosomes and renders a target promoter inaccessible to the general transcription apparatus. The biological activities of PR-A and PR-B are distinct and depend on the target gene in question. The shorter PR-A acts as a transcriptional inhibitor of other steroid receptors. Specifically, stimulation of target genes by estrogen, glucocorticoid, mineralocorticoid, and androgen receptors is repressed by liganded PR-A (McDonnell and Goldman, 1994). In most cells, PR-B mediates the stimulatory activities of progesterone; PR-A also strongly inhibits this action of PR-B (Vegeto et al., 1993). An inhibitory domain that is responsible for the transrepression caused by PR-A has been localized in the amino-terminus of PR-A. This inhibitory domain is present in both PR-B and PR-A, but for unknown reasons has repressor activity only in the context of PR-A. Current data suggest that different coactivators and corepressors interact with PR-A and PR-B, e.g., the corepressor SMRT (silencing mediator of retinoid and thyroid receptors) binds much more tightly to PR-A than to PR-B (Giangrande et al., 2000), and this may account, at least in part, for the differential activity of the two isoforms. Certain effects of progesterone, such as increased Ca2+ mobilization in sperm, can be seen in as little as 3 minutes (Blackmore, 1999), and these effects are thought to be caused by nongenomic mechanisms involving membrane-bound progesterone receptors. Several candidates for this receptor have been identified, and expression of a cDNA for one of these putative membrane PRs leads to production of a membrane-bound protein that binds progesterone relatively specifically (Falkenstein et al., 1999). The pharmacological importance of these membrane-bound receptors has not been determined.

Absorption, Fate, and Excretion Progesterone itself undergoes rapid first-pass metabolism; historically, this low oral bioavailability limited the administration of the natural hormone to intramuscular injections in oil or to vaginal suppositories and was an impetus to develop 17 -hydroxyprogesterone analogs such as medroxyprogesterone acetate (MPA) and 19-nor steroids for oral use. More recently, high-dose (e.g., 100 to 200 mg) preparations of micronized progesterone (PROMETRIUM) containing small particles (<10 m) suspended in oil and packaged in gelatin capsules have been developed. Although the absolute bioavailability of these preparations is low (Fotherby, 1996), efficacious plasma levels nevertheless may be obtained. Progesterone also is available in oil solution for injection, as a vaginal gel (CRINONE ), and as a slow-release, intrauterine device (PROGESTASERT ) for contraception. In addition to progesterone itself, a number of progestins are available. Esters such as hydroxyprogesterone caproate (HYALUTIN ) and MPA (DEPO-PROVERA) are available for intramuscular administration; MPA (PROVERA, CYCRIN , others) and megestrol acetate (MEGACE ) may be used orally, because their hepatic metabolism is substantially reduced relative to the parent hormone. The 19-nor steroids have good oral activity, because the ethinyl substituent at C 17 significantly slows hepatic metabolism. Implants (NORPLANT SYSTEM ) and depot preparations of synthetic progestins are available for release over very long periods of time (e.g., see later section on contraceptives). In the plasma, progesterone is bound by albumin and CBG, but is not appreciably bound to SSBG. 19-Nor compounds, such as norethindrone, norgestrel, and desogestrel, bind to SSBG and albumin, and esters such as MPA bind primarily to albumin. Total binding of all these synthetic compounds to plasma proteins is extensive, 90% or more, although the distribution of binding to the different proteins is compound-specific. The elimination half-life of progesterone is approximately 5 minutes, and the hormone is metabolized primarily in the liver to hydroxylated metabolites and their sulfate and glucuronide conjugates, which are eliminated in the urine. A major urinary metabolite specific for progesterone metabolism is pregnane-3 ,20 -diol; its measurement in urine and plasma is used as an index of endogenous progesterone secretion. The synthetic progestins have much longer half-lives, e.g., approximately 7 hours for norethindrone, 16 hours for norgestrel, 12 hours for gestodene, and 24 hours for MPA. The metabolism of synthetic progestins is thought to be primarily hepatic, and elimination is generally via the urine as conjugates and various polar metabolites. Therapeutic Uses The two most frequent uses of progestins are for contraception, either alone or with an estrogen in oral contraceptives, and in combination with estrogen for hormone replacement therapy of postmenopausal women. These two uses are discussed in detail elsewhere in this chapter. Progestins also are used in several settings for secondary amenorrhea, uterine bleeding disorders, luteal-phase support to treat infertility, and premature labor. There is interest in their potential use in mood disorders and the premenstrual syndrome. Among the oral progestins used besides MPA in these settings is norethindrone acetate (AYGESTIN). In general, these uses of oral progestins are extensions of the physiological actions of progesterone on the neuroendocrine control of ovarian function and on the endometrium. Progesterone can be used diagnostically to test for estrogen secretion and for responsiveness of the

endometrium. After administration of progesterone for 5 to 7 days to amenorrheic women, withdrawal bleeding will occur if the endometrium has been stimulated by endogenous estrogens. Combinations of estrogens and progestins also can be used to test endometrial responsiveness in patients with amenorrhea. Progestins have been used as a palliative measure for metastatic endometrial carcinoma, and megestrol acetate (MEGACE ) is used as a second-line treatment for breast cancer. Antiprogestins Although antiestrogens have been available since the late 1950s, the first report of an antiprogestin did not appear until 1981, when the glucocorticoid antagonist RU 38486 (now referred to as RU 486), or mifepristone, was reported to show antigestagenic properties (Philibert et al., 1981). In 1982, the first report of interruption of the human menstrual cycle and early pregnancy by this compound was presented before the French Acadmie des Sciences (Hermann et al., 1982). Mifepristone has been available since 1988 in France and several other countries for the termination of pregnancy, and in 2000 the FDA approved the drug for this use in the United States. While termination of pregnancy has been the main focus of mifepristone use to date, antiprogestins also have several other potential applications including uses as contraceptives, as agents to induce labor, and as agents to treat uterine leiomyomas, endometriosis, meingiomas, and breast cancer (Cadepond et al., 1997). Mifepristone Chemistry Mifepristone (RU 486) is a derivative of the 19-norprogestin norethindrone containing a dimethylaminophenyl substituent at the 11 -position. It is a potent, competitive antagonist of both progesterone and glucocorticoid binding to their respective receptors. Many other compounds with similar activity now have been synthesized and most contain a similar 11 -aromatic group as in mifepristone. Another widely studied antiprogestin is onapristone (or ZK 98 299), which is similar in structure to mifepristone but contains a methyl substituent in the 13 orientation rather than the 13 configuration present in mifepristone. Mifepristone and onapristone have the following structures:

Pharmacological Actions In the presence of progestins, mifepristone acts as a competitive receptor antagonist for both the A and B forms of the progesterone receptor. While mifepristone acts as an antagonist in vivo, it exhibits some agonist activity in certain in vitro test systems. In contrast, onapristone appears to be a pure progesterone antagonist both in vivo and in vitro. Binding of the two antiprogestins appears to cause different conformations of the progesterone receptor, which may account for the differences in their activities (Gass et al., 1998). When administered in the early stages of pregnancy, mifepristone causes decidual breakdown by blockade of uterine progesterone receptors. This leads to detachment of the blastocyst, which decreases hCG production. This in turn causes a decrease in progesterone secretion from the corpus luteum, which further accentuates decidual breakdown. Decreased endogenous progesterone coupled with blockade of progesterone receptors in the uterus increases uterine prostaglandin levels and sensitizes the myometrium to the contractile actions of prostaglandins. A separate effect of the compound is to cause cervical softening, which facilitates expulsion of the detached blastocyst. Mifepristone also has effects on ovulation. If given acutely in the mid to late follicular phase, it delays follicle maturation and the LH surge, and ovulation occurs later than normal. If the drug is given intermittently (e.g., once a week) or continuously, ovulation is prevented in most but not all cases. These effects are due largely to actions on the hypothalamus and pituitary rather than the ovary, although the mechanisms are unclear. If administered for one or several days in the mid- to late-luteal phase, mifepristone impairs the development of a secretory endometrium and produces menses. Progesterone-receptor blockade at this time is the pharmacological equivalent of progesterone withdrawal, and bleeding normally ensues within several days and lasts for 1 to 2 weeks after antiprogestin treatment. Mifepristone also binds to glucocorticoid and androgen receptors and exerts antiglucocorticoid and antiandrogenic actions. A predominant effect in human beings is blockade of the feedback inhibition by cortisol of ACTH secretion from the pituitary, thus increasing both corticotropin and adrenal steroid levels in the plasma. Onapristone also binds to both glucocorticoid and androgen

receptors, but has less antiglucocorticoid activity than does mifepristone. Absorption, Fate, and Excretion Mifepristone is orally active and has good bioavailability by this route. Peak plasma levels occur within several hours after administration, and the drug is slowly cleared, with a plasma half-life of 20 to 40 hours being reported. In plasma, mifepristone is bound with high affinity by 1-acid glycoprotein, and this may contribute to its long half-life. Metabolites are primarily the mono- and didemethylated products (which are thought to have pharmacological activity) formed via CYP3A4-catalyzed reactions and to a lesser extent hydroxylated compounds. The drug undergoes hepatic metabolism and enterohepatic circulation, and metabolic products are found predominantly in the feces (Jang and Benet, 1997). Therapeutic Uses and Prospects Mifepristone has been available for several years for clinical use in many European countries, China, and Israel. In September 2000, the FDA approved mifepristone (MIFEPREX), in combination with misoprostol (see below), for the termination of early pregnancy (defined as 49 days or less, counting from the beginning of the last menstrual period). Terms of the approval limit the distribution of mifepristone to physicians who can determine the duration of pregnancy, detect an ectopic pregnancy, and provide surgical intervention in cases of incomplete abortion or severe bleeding. The major use of mifepristone is to produce medical abortion in the first trimester of pregnancy. When mifepristone is used to produce a medical abortion, a prostaglandin is given 48 hours after the antiprogestin to further increase myometrial contractions and ensure expulsion of the detached blastocyst. Intramuscular sulprostone, intravaginal gemeprost, and oral misoprostol have been used. The success rate with such regimens is >90% among women with pregnancies of 49 days' duration or less. The most severe untoward effect is vaginal bleeding, which most often lasts from 8 to 17 days but is only rarely (0.1% of patients) severe enough to require blood transfusions. High percentages of women also have experienced abdominal pain and uterine cramps, nausea, vomiting, and diarrhea, which are due to the use of the prostaglandin. Many patients receive one or more medications for pain relief from these untoward effects. Women with adrenal failure or severe asthma or who are receiving long-term glucocorticoid therapy should not be given mifepristone because of its antiglucocorticoid activity, and the drug should be used very cautiously in women with anemia or those being treated with anticoagulants. Women over 35 years old with cardiovascular risk factors should not be given sulprostone, because the drug has been associated with heart failure in such individuals (Christin-Maitre et al., 2000). Mifepristone also has been used as a postcoital contraceptive, and it may be slightly more effective than high-dose estrogen-progestin combinations. The mechanism of action in this case is thought to be prevention of implantation. It also has been proposed that regular use of an antiprogestin in the late luteal phase would be an effective contraceptive because it would ensure that the endometrium was shed and menstruation occurred in each cycle. Other investigational or potential uses for mifepristone include the induction of labor after fetal death, the induction of labor at the end of the third trimester, treatment of endometriosis and leiomyomas, breast cancer, and meningiomas (Cadepond et al., 1997). Given the multiple potential uses of these agents for clinical and experimental purposes, this is expected to remain a major area of therapeutic development in the future. Hormonal Contraceptives

Oral contraceptives are among the most widely used agents in the United States and throughout the world. Since they became available in 1960, they have influenced the lives of untold millions of individuals and have had a revolutionary impact on global society. For the first time in history, a convenient, affordable, and completely reliable means of contraception was available for family planning and the avoidance of unplanned pregnancies. It is important to consider several key points as a prelude to the pharmacology of specific hormonal contraceptives: (1) Hormonal contraceptives are among the most effective drugs available. (2) A variety of agents with substantially different components, doses, and side effects are available and provide real therapeutic options. (3) In addition to contraceptive actions, these agents have substantial health benefits. (4) Because of differences in doses and specific compounds used, it is not appropriate to extrapolate directly untoward effects of hormonal contraceptives to hormone replacement therapy, or vice versa. Oral contraceptives are completely effective and have a low incidence of untoward effects for most women. History Around the beginning of the twentieth century, a number of European scientists including Beard, Prenant, and Loeb developed the concept that secretions of the corpus luteum suppressed ovulation during pregnancy. These and other workers focused on understanding basic relationships in reproductive physiology, and the Austrian physiologist Haberlandt extended this concept to advance the idea that hormones could be used for purposes of sterilization (see Perone, 1994). In 1927 he published a report entitled "On the Hormonal Sterilization of Female Animals," in which he reported producing temporary sterility in rodents by feeding ovarian and placental extracts (Haberlandt, 1927)i.e., a clear example of an oral contraceptive! After progesterone was isolated from the corpus luteum, Makepeace and colleagues reported in 1937 that the pure hormone blocked ovulation in rabbits (Makepeace et al., 1937). Two years later, Astwood and Fevold (1939) found a similar effect in rats. In the 1950s, Pincus, Garcia, and Rock found that progesterone and 19-nor progestins prevented ovulation in women (Rock et al., 1957). Ironically, this finding grew out of their attempts to treat infertility with progestins or estrogen-progestin combinations. The initial findings were that either treatment effectively blocked ovulation in the majority of women. However, concern about cancer and other possible side effects of the estrogen (diethylstilbestrol) led to the use of a progestin alone in their studies. One of the compounds used was norethynodrel, and early batches of this compound were contaminated with a small amount of mestranol. When mestranol was removed, it was noted that treatment with pure norethynodrel led to increased breakthrough bleeding and less consistent inhibition of ovulation. Mestranol was thus reincorporated into the preparation, and this combination was employed in the first large-scale clinical trial of combination oral contraceptives. Clinical studies performed by Pincus, Rock, Garcia, and their colleagues in the mid- to late 1950s in Puerto Rico and Haiti established the virtually complete contraceptive success of the norethynodrelmestranol combination (Pincus et al., 1959). In late 1959, ENOVID (norethynodrel plus mestranol) was the first "Pill" approved by the FDA for use as a contraceptive agent in the United States, and this was followed in 1962 by approval for ORTHO-NOVUM (norethindrone plus mestranol). By 1966 there were approximately a dozen "first-generation" preparations on the market in the United States utilizing either mestranol or ethinyl estradiol in combination with one of several different 19-nor progestins. In the 1960s, the progestin-only minipill and long-acting injectable preparations were developed and subsequently used throughout much of the world, but they were not approved for use

in the United States until the 1990s. Millions of women in the United States and elsewhere began using oral contraceptives, and frequent reports of untoward effects began appearing in the 1970s (see Kols et al., 1982). The recognition that these side effects were dose-dependent and the realization that estrogens and progestins synergistically inhibited ovulation led to the reduction of doses and the development of so-called low-dose or second-generation contraceptives. The increasing use of biphasic and triphasic preparations throughout the 1980s further reduced steroid dosages; it may be that current doses available commercially are the lowest that will provide reliable contraception. In the 1990s, the "third-generation" oral contraceptives, containing progestins with reduced androgenic activity [e.g., norgestimate (ORTHO TRI-CYCLEN ) desogestrel (DESOGEN)], became available in the United States after being used in Europe for some time. Products containing gestodene as a progestin with reduced androgenic activity also are available in Europe. Another major development in the 1980s was the widespread realization that oral contraceptives have a number of substantial health benefits (Kols et al., 1982). Types of Hormonal Contraceptives Combination Oral Contraceptives The most frequently used agents in the United States are combination oral contraceptives containing both an estrogen and a progestin. They are highly efficacious, with a theoretical effectiveness generally considered to be 99.9% and a use effectiveness of 97% to 98%. Ethinyl estradiol and mestranol are the two estrogens used (with ethinyl estradiol being much more frequently used), and several progestins currently are used. The progestins are 19-nor compounds in the estrane or gonane series, and each has varying degrees of androgenic, estrogenic, and antiestrogenic activities that may be responsible for some of their side effects. Compounds such as desogestrel and norgestimate are the most recently developed and have less androgenic activity than other 19-nor compounds (Shoupe, 1994; Archer, 1994; Rebar and Zeserson, 1991). The absorption, fate, and excretion of the individual components have been discussed in previous sections. Combination oral contraceptives are available as monophasic, biphasic, or triphasic preparations, generally provided in 21-day packs. For the monophasic agents, fixed amounts of the estrogen and progestin are present in each pill, which is taken daily for 21 days, followed by a 7-day "pill-free" period. (Virtually all preparations come as 28-day packs, with the pills for the last 7 days containing only inert ingredients.) The biphasic and triphasic preparations provide two or three different pills containing varying amounts of active ingredients, to be taken at different times during the 21-day cycle. This reduces the total amount of steroids administered and more closely approximates the estrogen to progestin ratios that occur during the menstrual cycle (such as a generally higher ratio in the luteal phase; see Figure 583). Phasic preparations were developed in the 1980s, largely to reduce the dose of progestins in oral contraceptives when it was recognized these might have untoward cardiovascular effects. In 2000, the FDA approved a once-monthly injectable preparation containing estradiol cypionate and medroxyprogesterone acetate. The estrogen content of current preparations ranges from 20 to 50 g; the majority contain 30 to 35 g. Preparations containing 35 g or less of an estrogen are generally referred to as "low-dose" or "modern" pills. The dose of progestin is more variable because of differences in potency of the compounds used. For example, monophasic pills currently available in the United States contain 0.4 to 1 mg of norethindrone, 0.1 to 0.15 mg of levonorgestrel, 0.3 to 0.5 mg of norgestrel, 1 mg of ethynodiol diacetate, 0.25 mg of norgestimate, and 0.15 mg of desogestrel, with slightly different dose ranges in biphasic and triphasic preparations. The first agent available (ENOVID) contained 10

mg of norethynodrel and 150 g of mestranol. In 1966, most first-generation preparations on the market contained 50 to 100 g of an estrogenic component and 2 to 10 mg of a progestin. These large differences in doses complicate extrapolation of data from early epidemiological studies on the side effects of "high-dose" oral contraceptives to the "low-dose" preparations now used. Progestin-Only Contraceptives Several agents are available for progestin-only contraception. They are slightly less efficacious than combination oral contraceptives, with reports of theoretical effectiveness of 99% and a useeffectiveness of 96% to 97.5%. Specific preparations include the "minipill" or oral preparations of low doses of progestins, e.g., 350 g of norethindrone (NOR-QD , MICRONOR ) or 75 g of norgestrel (OVRETTE ) taken daily without interruption; subdermal implants of 216 mg of norgestrel (NORPLANT SYSTEM) for slow release and resultant long-term contraceptive action (e.g., up to 5 years); and crystalline suspensions of medroxyprogesterone acetate (DEPO-PROVERA) for intramuscular injection of 150 mg of drug, which provides effective contraception for 3 months. An intrauterine device (PROGESTASERT ) that releases low amounts of progesterone locally is available for insertion on a yearly basis. Its effectiveness is considered to be 97% to 98%, and contraceptive action probably is due to local effects on the endometrium. Postcoital or Emergency Contraceptives High doses of diethylstilbesterol and other estrogens once were used for postcoital contraception (the "morning-after pill") but never received FDA approval for this indication. Clinical trials (Task Force on Postovulatory Methods of Fertility Regulations, 1998) have led to FDA approval and marketing of two preparations for postcoital contraception. For PLAN-B, the total treatment is two one-pill doses (0.75 mg levonorgestrel per pill) separated by 12 hours. For PREVEN, it is two twopill doses (0.25 mg of levonorgestrel and 0.05 mg of ethinyl estradiol per pill) separated by 12 hours. These preparations are basically high-dose oral contraceptives, and other products with the same or very similar composition have been declared safe and effective for use as emergency contraceptive pills by the FDA. The first dose of such preparations should be taken within 72 hours of intercourse, and this should be followed 12 hours later by a second dose. This treatment reduces the risk of pregnancy by approximately 75%. Roughly 8 of 100 women having unprotected intercourse during the second or third week of their cycles will become pregnant; emergency contraceptive pills reduce this to 2 women per 100. Mechanism of Action Combination Oral Contraceptives Combination oral contraceptives act by preventing ovulation (Lobo and Stanczyk, 1994). Direct measurements of plasma hormone levels indicate that LH and FSH levels are suppressed, a midcycle surge of LH is absent, endogenous steroid levels are diminished, and ovulation does not occur. While either component alone can be shown to exert these effects in certain situations, the combination synergistically decreases plasma gonadotropin levels and suppresses ovulation more consistently than either alone. Given the multiple actions of estrogens and progestins on the hypothalamic-pituitary-ovarian axis during the menstrual cycle, several effects probably contribute to the blockade of ovulation. In

addition, the prolonged administration of these drugs may bring into play other mechanisms that do not operate physiologically in the menstrual cycle. It seems likely that the reason these drugs are so extraordinarily effective is that they produce their contraceptive action via multiple mechanisms. Hypothalamic actions of steroids play a major role in the mechanism of oral contraceptive action. Progesterone clearly diminishes the frequency of GnRH pulses. Since the proper frequency of LH pulses is essential for ovulation, this effect of progesterone likely plays a major role in the contraceptive action of these agents. In monkeys and women with normal menstrual cycles, estrogens do not affect the frequency of the pulse generator. However, in the prolonged absence of a menstrual cycle (e.g., in ovariectomized monkeys and postmenopausal women; see Hotchkiss and Knobil, 1994), estrogens markedly diminish pulse generator frequency, and progesterone enhances this effect. In theory, this hypothalamic effect of estrogens could come into play when oral contraceptives are used for extended times. Pituitary effects also are likely to contribute to the actions of oral contraceptives. Administration of exogenous GnRH to women receiving oral contraceptives increases plasma LH levels. However, the increase is much smaller than that seen in control subjects, indicating that oral contraceptives decrease pituitary responsiveness to GnRH (Mishell et al., 1977). Estrogens normally suppress FSH release from the pituitary during the follicular phase of the menstrual cycle, and this effect could contribute to the lack of follicular development observed in oral-contraceptive users. Sustained elevation of estrogen levels above a threshold also triggers the midcycle surge of LH required for ovulation. Physiologically, progesterone does not affect this process, but its pharmacological administration inhibits the estrogen-induced LH surge. Multiple pituitary effects of both estrogen and progestin components thus contribute to oral contraceptive action. In addition to prevention of ovulation, other effects are suspected to contribute to the extraordinary efficacy of oral contraceptives. Transit of sperm, the egg, and fertilized ovum is important to establish pregnancy, and steroids are likely to affect transport in the fallopian tube. Progestin effects also are likely to be dominant in the cervix to produce a thick, viscous mucus to reduce sperm penetration and in the endometrium to produce a state that is not receptive to implantation. However, it is difficult to assess quantitatively the contributions of these effects, because the drugs block ovulation so effectively. Progestin-Only Contraceptives The doses of progestins in minipills and in subcutaneous implants of levonorgesterel are sufficient to block ovulation in only 60% to 80% of cycles. The effectiveness of these preparations is thus thought to be due largely to a thickening of cervical mucus, which decreases sperm penetration, and to endometrial alterations that impair implantation. Depot injections of MPA are thought to exert these latter effects, but they also yield plasma levels of drug high enough to prevent ovulation in virtually all patients. Observed decreases in ovulation are thought to be due to a slowing of the frequency of the GnRH pulse generator, which prevents the LH surge required for ovulation. Emergency Contraceptive Pills Multiple mechanisms are likely to contribute to the efficacy of these agents. Some studies have shown that ovulation is inhibited or delayed if the agents are taken in the first half of the cycle, but other mechanisms are likely to be involved as well because of the high degree of efficacy. Additional mechanisms, some of which are speculative, may include alterations in endometrial receptivity for implantation; interference with functions of the corpus luteum that maintain pregnancy; production of a cervical mucus that decreases sperm penetration; alterations in tubular

transport of sperm, egg, or embryo; or effects on fertilization. However, emergency contraceptives do not interrupt an established pregnancy defined as beginning with implantation. Untoward Effects Combination Oral Contraceptives Shortly after the introduction of oral contraceptives, reports of adverse side effects associated with their use began to appear (see Kols et al., 1982). Many of the side effects were found to be dosedependent, and this led to the development of current low-dose preparations. Untoward effects of early hormonal contraceptives fell into several major categories: adverse cardiovascular effects, including hypertension, myocardial infarction, hemorrhagic or ischemic stroke, and venous thrombosis and embolism; breast, hepatocellular, and cervical cancers; and a number of endocrine and metabolic effects. The current consensus is that low-dose preparations pose minimal health risks in women who have no predisposing risk factors, and these drugs also provide many beneficial health effects (Baird and Glasier, 1993). Cardiovascular Effects The question of cardiovascular side effects has been reexamined for the newer low-dose oral contraceptives (Baird and Glasier, 1993; Mischell, 1999; Castelli, 1999; Sherif, 1999). For nonsmokers without other risk factors, there is no significant increase in risk of myocardial infarction or stroke. There still is an increase in relative risk for venous thromboembolism, but the estimated absolute increase is very small, because the incidence of these events in women without other predisposing factors is low, e.g., roughly half that associated with the risk of venous thromboembolism in pregnancy. Nevertheless, the risk may be increased in women who smoke or have other factors that predispose to thrombosis or thromboembolism (Castelli, 1999). Early highdose, combination oral contraceptives caused hypertension in 4% to 5% of normotensive women and increased blood pressure in 10% to 15% of those with preexisting hypertension. This incidence is much lower with newer, low-dose preparations, and most reported changes in blood pressure are not significant. The cardiovascular risk associated with oral contraceptive use does not appear to persist after use is discontinued. There were several reports in 1995 that use of oral contraceptives containing the third-generation progestins gestodene and desogestrel caused a substantial increase in venous thromboembolism in European users relative to preparations with levonorgestrel and norethindrone; this has been a contentious issue, however, and other analyses have attributed the reported differences to confounding variables (Barbieri et al., 1999). In general, however, there have not been major reported differences in cardiovascular parameters associated with different progestins used in modern low-dose preparations. As noted previously, estrogens increase serum HDL and decrease LDL levels, and progestins tend to have the opposite effect. Recent studies of several low-dose preparations have not found significant change in total serum cholesterol or lipoprotein profiles, although slight increases in triglycerides have been reported. Cancer Given the growth-promoting effects of estrogens, there has been a long-standing concern that oral contraceptives might increase the incidence of endometrial, ovarian, breast, and other cancers. These concerns were further heightened in the late 1960s by reports of endometrial changes caused

by sequential oral contraceptives, which have since been removed from the market in the United States. However, it is now clear that there is not a widespread association between oral contraceptive use and cancer (see Baird and Glasier, 1993; Sherif, 1999; Westhoff, 1999). Combination oral contraceptives do not increase the incidence of endometrial cancer but actually cause a 50%decrease in the incidence of this disease, which lasts 15 years after the pills are stopped. This is thought to be due to the inclusion of a progestin, which opposes estrogen-induced proliferation, throughout the entire 21-day cycle of administration. Similarly, these agents also decrease the incidence of ovarian cancer, and decreased ovarian stimulation by gonadotropins provides a logical basis for this effect. There have been reports of increases in the incidence of hepatic adenoma and hepatocellular carcinoma in oral contraceptive users, but these are relatively rare diseases, and analysis of their incidence in oral contraceptive users is complicated by numerous factors. There also have been reports of increased cervical cancer in oral contraceptive users, but confounding factors have precluded a definitive association with this disease. The major present concern about the carcinogenic effects of oral contraceptives is focused on breast cancer. Numerous studies have dealt with this issue, and the following general picture has emerged. The risk of breast cancer in women of childbearing age is very low, and current oral contraceptive users in this group have only a very small increase in relative risk of 1.1 to 1.2, depending on other variables. This small increase is not substantially affected by duration of use, dose or type of component, age at first use, or parity. Importantly, 10 years after discontinuation of oral contraceptive use, there is no difference in breast cancer incidence between past users and never users. In addition, breast cancers diagnosed in women who have ever used oral contraceptive are more likely to be localized to the breast and thus easier to treat, because they are less likely to have spread to other sites (Westhoff, 1999). Overall, there is thus no significant difference in the cumulative risk of breast cancer between those who have ever used oral contraceptives and those who have never used them. Metabolic and Endocrine Effects The effects of sex steroids on glucose metabolism and insulin sensitivity are complex (Godsland, 1996) and may differ among agents in the same class e.g., the 19-nor progestins. Early studies with high-dose oral contraceptives generally reported impaired glucose tolerance as demonstrated by increases in fasting glucose and insulin levels and responses to glucose challenge. These effects have decreased as steroid dosages have been lowered, and current low-dose combination contraceptives may even improve insulin sensitivity. Similarly, the high-dose progestins in early oral contraceptives did raise LDL and reduce HDL levels, but modern low-dose preparations do not produce unfavorable lipid profiles (see Sherif, 1999). There also have been periodic reports that oral contraceptives increase the incidence of gallbladder disease, but any such effect appears to be weak and limited to current or very long term users (Grodstein et al., 1994). The estrogenic component of oral contraceptives may increase hepatic synthesis of a number of serum proteins, including those that bind thyroid hormones, glucocorticoids, and sex steroids. While physiological feedback mechanisms generally adjust hormone synthesis to maintain normal "free" hormone levels, these changes can affect the interpretation of endocrine function tests that measure total plasma hormone levels. The ethinyl estradiol present in oral contraceptives appears to cause a dose-dependent increase in several serum factors known to increase coagulation. However, in healthy women who do not

smoke, there also is an increase in fibrinolytic activity, which exerts a counter effect so that overall there is a minimal effect on hemostatic balance. In women who smoke, however, this compensatory effect is diminished, which may shift the hemostatic profile toward a hypercoagulable condition (Fruzzetti, 1999). Miscellaneous Effects Nausea, edema, and mild headache occur in some individuals, and more severe migraine headaches may be precipitated by oral contraceptive use in a smaller fraction of women. Some patients may experience breakthrough bleeding during the 21-day cycle when the active pills are being taken. Withdrawal bleeding may fail to occur in a small fraction of women during the 7-day "off" period, thus causing confusion about a possible pregnancy. Weight gain, acne, and hirsutism are thought to be mediated by the androgenic activity of the 19-nor progestins. Progestin-Only Contraceptives Episodes of irregular, unpredictable spotting and breakthrough bleeding are the most frequently encountered untoward effect and the major reason women discontinue use of all three types of progestin-only contraceptives. With time, the incidence of these bleeding episodes decreases, especially with the long-acting preparations, and amenorrhea becomes common after a year or more of use. There is no evidence that the progestin-only minipill preparations increase thromboembolic events, which are thought to be related to the estrogenic component of combination preparations; blood pressure does not appear to be elevated; and nausea and breast tenderness do not occur. Acne may be a problem, however, because of the androgenic activity of norethindrone-containing preparations. These preparations may be attractive for nursing mothers, because they do not decrease lactation as do products containing estrogens. Aside from bleeding irregularities, headache is the most commonly reported untoward effect of depot MPA (medroxyprogesterone acetate). Mood changes and weight gain also have been reported, but controlled clinical studies of these effects are not available. It is of more concern that a number of studies have found decreases in HDL levels and increases in LDL levels and that there have been several reports of decreased bone density. These effects may be due to reduced endogenous estrogens, because depot MPA is particularly effective in lowering gonadotropin levels. An early study found that MPA caused breast cancer in beagle dogs, but this was subsequently found to be due to a unique species-specific metabolism of the drug to estrogens, and numerous human studies have not found any increases in breast, endometrial, cervical, or ovarian cancer in women receiving MPA. Because of the time required to completely eliminate the drug, the contraceptive effect of this agent may remain for 6 to 12 months after the last injection. Implants of norethindrone may be associated with infection, local irritation, pain at the insertion site, and, rarely, expulsion of the inserts. Headache, weight gain, and mood changes have been reported, and acne is a concern in some patients. A number of metabolic studies have been performed in NORPLANT users, and in most cases only minimal changes have been observed in lipid, carbohydrate, and protein metabolism and in serum chemistry. In women desiring pregnancy, ovulation occurs soon after implant removal, reaching 50% in 3 months and almost 90% within 1 year. Emergency Contraceptive Pills

Nausea and vomiting are the main untoward effects, with an incidence of roughly 50% and 20%, respectively, for combined estrogen-levonorgestrel combinations and 23% and 6% for levonorgestrel alone (Task Force on Postovulatory Methods of Fertility Regulation, 1998). No changes in clotting factors have been reported for the combined regimen, but based on concerns with combination oral contraceptives, levonorgestrel alone might be considered for women who smoke or have a history of blood clots. Emergency contraceptive pills are contraindicated in cases of confirmed pregnancy. Contraindications While the use of modern oral contraceptives is considered generally safe in most healthy women, these agents can contribute to the incidence and severity of certain diseases if other risk factors are present. The following conditions are thus considered absolute contraindications for combination oral contraceptive use: the presence or history of thromboembolic disease, cerebral vascular disease, myocardial infarction, coronary artery disease, or congenital hyperlipidemia; known or suspected carcinoma of the breast, carcinoma of the female reproductive tract, or other hormonedependent/responsive neoplasias; abnormal undiagnosed vaginal bleeding; known or suspected pregnancy; and past or present liver tumors or impaired liver function. The risk of serious cardiovascular side effects is particularly marked in women over 35 years of age who smoke heavily (e.g., over 15 cigarettes per day); even low-dose oral contraceptives are contraindicated in such patients. Several other conditions are relative contraindications and should be considered on an individual basis. These include migraine headaches, hypertension, diabetes mellitus, obstructive jaundice of pregnancy or prior oral contraceptive use, and gallbladder disease. If elective surgery is planned, many physicians recommend discontinuation of oral contraceptives for several weeks to a month to minimize the possibility of thromboembolism after surgery. These agents should be used with care in women with prior gestational diabetes or uterine fibroids, and low-dose pills should generally be used in such cases. Progestin-only contraceptives are contraindicated in the presence of undiagnosed vaginal bleeding, benign or malignant liver disease, and known or suspected breast cancer. Depot medroxyprogesterone acetate and levonorgestrel inserts are contraindicated in women with a history or predisposition to thrombophlebitis or thromboembolic disorders. Choice of Contraceptive Preparations Many preparations that differ substantially in dose and specific components are available, providing the option to select the preparation best suited to each individual. The general feeling is that treatment should be started with preparations containing the minimum dose of steroids that provides effective contraceptive coverage. This often is a pill with 30 to 35 g of estrogen, but preparations with 20 g may be adequate for women who weigh much less than average or who are over 40 years of age. A newer use of 20- g pills is in the treatment of perimenopausal menstrual disorders. A preparation containing 50 g of estrogen may be required for heavier women. Breakthrough bleeding may occur in some women if the estrogen:progestin ratio is too low to produce a stable endometrium, and this may be prevented by switching to a pill with a higher ratio. In women for whom estrogens are contraindicated or undesirable, progestin-only contraceptives may be an option. The progestin-only minipill may have an enhanced effectiveness in several such types of women, e.g., nursing mothers and women over 40, in whom fertility may be decreased. In contrast to estrogen-containing contraceptives, progestins can be used in nursing mothers without

affecting lactation. Another consideration is the administration of medications that may increase metabolism of estrogens (e.g., rifampicin, barbiturates, and phenytoin) or reduce their enterohepatic recycling (e.g., tetracyclines and ampicillin). Antimicrobials may decrease intestinal flora that produce enzymes required for hydrolysis and reuptake of conjugated metabolites initially secreted into the intestine via the bile. In these situations, a low-dose pill may not be 99.9% effective due to decreased plasma levels of the estrogenic component. The choice of a preparation also may be influenced by the specific 19-nor progestin component, since this component may have varying degrees of androgenic and other activities. The androgenic activity of this component may contribute to untoward effects such as weight gain, acne due to increased sebaceous gland secretions, and unfavorable lipoprotein profiles. These side effects are greatly reduced in newer, low-dose contraceptives, but any patients exhibiting such side effects may benefit by switching to pills that contain a progestin with less androgenic activity. Of the progestins found in oral contraceptives, norgestrel is generally considered to have the most androgenic activity; norethindrone and ethynodiol diacetate to have more moderate androgenic activity; and desogestrel and norgestimate to have the least androgenic activity. A triphasic, low-dose combination oral contraceptive (ORTHO TRI-CYCLEN ) containing ethinyl estradiol and norgestimate has been approved by the FDA for the treatment of moderate acne vulgaris. Similar preparations (DEMULEN 1/35, DESOGEN, ORTHOCEPT) also are effective. The mechanism appears to be a decrease in free plasma testosterone due to an increase in plasma SSBG, since total testosterone levels are unchanged (Redmond et al., 1997). In summary, for a given individual, both the efficacy and side effects of hormonal contraceptives may vary considerably among preparations. A number of choices are available, and changing preparations may decrease the incidence of side effects in a given patient without decreasing contraceptive efficacy. Noncontraceptive Health Benefits It has been accepted for well over a decade that combination oral contraceptives have substantial health benefits unrelated to their contraceptive use (see Kols et al., 1982; Goldzieher, 1994; Baird and Glasier, 1993). These include effects on endometrial and ovarian cancer, a variety of common menstrual disorders, and several other diseases. Oral contraceptives significantly reduce the incidence of ovarian and endometrial cancer within 6 months of use, and the incidence is decreased 50% after 2 years of use. Furthermore, this protective effect persists for up to 15 years after oral contraceptive use is discontinued. These agents also decrease the incidence of ovarian cysts and benign fibrocystic breast disease. Oral contraceptives have major benefits related to menstruation in many women. These include more regular menstruation, reduced menstrual blood loss and less iron-deficiency anemia, less premenstrual tension, and decreased frequency of dysmenorrhea. There also is a decreased incidence of pelvic inflammatory disease and ectopic pregnancies, and endometriosis may be ameliorated. Some women also may obtain these benefits with progestin-only contraceptives, and there are suggestions that MPA may improve hematological parameters in women with sickle-cell disease (Cullins, 1996). There is now a consensus that combination oral contraceptives prevent thousands of deaths,

episodes of various diseases, and cases of hospitalization each year in the United States alone. Approximately 20% of all pregnant women are hospitalized before delivery because of complications, and the incidence of death related to childbirth (approximately 20 per 100,000 births to women under 35 in developed countries) is not insignificant. Thus, from a purely statistical perspective, fertility regulation by oral contraceptives is substantially safer than pregnancy or childbirth for most women (Grimes, 1994), even without considering the additional health benefits of these agents. Prospectus Estrogens and progestins currently are among the most widely used prescription medications. This heavy use is likely to continue and possibly increase, because, as the population ages, increasing numbers of women will be in the age bracket that traditionally has received hormone-replacement therapy. Such therapy will remain a major use of these agents, and there will be an intense effort to develop an ideal SERM for this purpose. Such an agent would have agonist activity necessary to provide relief of vasomotor symptoms, maintain bone mass, prevent urogenital atrophy, and yield favorable profiles of lipoprotein and hemostatic factors, but be an antagonist in the breast and devoid of tropic actions in the endometrium. Whether or not a single agent can produce all these desirable actions remains to be determined; if not, combinations of agents may be investigated to elicit the desired spectrum of activities. These efforts will be aided by advances in structural biology and molecular pharmacology that will provide the molecular topography of hormone-binding sites and other domains involved in steroid-receptor functions. Advances in molecular genetics also will determine how polymorphisms in genes encoding receptors, enzymes involved in biotransformation reactions, or coactivators/corepressors affect responses to estrogens and progestins. In terms of postmenopausal hormone-replacement therapy, increasing attention also will be focused on various routes of delivery (e.g., transdermal and intravaginal) to minimize systemic concentrations of hormones and/or to increase or decrease exposure of individual tissues to the hormone. Increased effort also will be devoted to reexamining dosing regimens for progestins used in combination with estrogens, especially in light of historic questions about adverse effects of these agents on cardiovascular health, and in light of more recent reports that women receiving estrogenplus-progestin-replacement therapy are at greater risk for developing breast cancer than are those receiving estrogen alone. Prospective clinical studies now in progress also should provide more definitive information about the risk of breast cancer and other diseases associated with hormonereplacement therapy; the effectiveness of such therapy in the primary prevention of cardiovascular disease; and whether or not estrogen treatment slows the onset or progression of Alzheimer's disease and other neurodegenerative disorders. Prospective trials comparing the beneficial effects of tamoxifen, raloxifene, and possibly other SERMs also will be conducted within the next 5 years. The effects of estrogens in men (e.g., on growth of the long bones, in diseases such as prostate cancer, on behavior, and on the reproductive system) will receive increased attention, since estrogen-receptor knockout animals have led to the realization that estrogens may have major effects on these tissues/systems in males as well as females. Considerable efforts also will be focused on the pharmacology of antiestrogens and antiprogestins. Tamoxifen has proven efficacy in the prophylaxis of breast cancer but increases the risk of uterine cancer and venous thromboembolism; attempts thus will be focused on identifying other SERMs with the former but not the latter actions. Antiprogestins will receive increased attention in the possible treatment of breast cancer, the induction of labor, and as contraceptives; both antiestrogens and antiprogestins may find new applications in the treatment of endometriosis and uterine fibroids. The actions of progesterone itself as opposed to synthetic progestins, including the "third

generation" gonane compounds, also will be evaluated, as will various non-oral routes of progestin administration. The role of selective aromatase inhibitors in the treatment of estrogen-responsive diseases such as breast cancer will receive increasing attention with the recognition that substantial amounts of estrogens may be locally produced. New developments in contraceptives may include a second generation of long-acting progestin (norethindrone) implants; new biodegradable implants; long-acting, timed-release injectable preparations such as microspheres containing estrogens and progestins; and transdermal delivery devices. For further discussion of disorders of the ovary and female reproductive tract see Chapter 326 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 59. Androgens

Overview Testosterone is the principal circulating androgen in men. It is secreted by the Leydig cells of the testes in response to luteinizing hormone (LH) from the pituitary gland. The varied effects of testosterone are due to its ability to act by at least three different mechanisms: by binding to the androgen receptor; by conversion in certain tissues to dihydrotestosterone, which also binds to the androgen receptor; and by conversion to estradiol, which binds to the estrogen receptor. Testosterone is responsible for male sexual differentiation in utero and for male pubertal changes. Consequently, failure of a male fetus to secrete testosterone or to have functional androgen receptors during the first trimester results in incomplete male sexual differentiation; failure of testosterone secretion before puberty results in incomplete pubertal changes; and failure during adulthood results in a diminution, at different rates, of some aspects of virilization. In women the physiological role of testosterone and the consequences of its deficiency are not yet understood, but it is possible that it contributes to libido, energy, muscle mass and strength, and bone strength. Oral administration of testosterone leads to absorption into the hepatic circulation but rapid catabolism by the liver, so oral ingestion is ineffective in delivering testosterone systemically. Most attempts to devise pharmacological testosterone preparations, therefore, have involved finding ways of bypassing hepatic catabolism. The 17 -alkylated androgens can be administered orally and are not catabolized as rapidly as testosterone itself, but they tend to cause cholestasis. Esters of testosterone and a fatty acid, when injected, produce serum testosterone concentrations that remain within the normal range for one to several weeks. Transdermal preparations of testosterone deliver testosterone itself into the systemic circulation and, when applied daily, produce relatively even serum testosterone concentrations. The major indication for testosterone treatment is male hypogonadism, for which a testosterone ester or transdermal preparation should be used. Treatment should be monitored for efficacy by measurement of the serum testosterone concentration and for deleterious effects by evaluating for obstruction to urine flow due to benign prostatic hyperplasia, for prostate cancer, and for erythrocytosis. Athletes have used androgens to attempt to improve their performance. Androgens have been used to attempt to develop a male contraceptive. For this purpose they have been used alone or in combination with a gonadotropin-releasing hormone (GnRH) antagonist or a progestin to suppress endogenous testosterone production and thereby spermatogenesis. The 17 -alkylated

androgens are used to treat angioneurotic edema, because they stimulate C1 esterase inhibitor. Some drugs are antiandrogens that are used intentionally to inhibit undesirable effects of androgens; other drugs, used for nonhormonal purposes, have side effects as a consequence of their antiandrogenic properties. Analogs of GnRH inhibit LH secretion and thereby reduce testosterone synthesis. They are used to treat metastatic prostate cancer. A side effect of the antifungal agents of the imidazole class (see Chapter 49: Antimicrobial Agents: Antifungal Agents) is direct inhibition of cortisol synthesis in the adrenal glands and testosterone synthesis in the testes. Flutamide and bicalutamide are androgen receptor antagonists that are used in combination with GnRH analogs in the treatment of metastatic prostate cancer because they block the effects of adrenal androgens. Spironolactone (see Chapter 29: Diuretics) is an aldosterone receptor antagonist and also a weak androgen receptor antagonist that causes gynecomastia when used as a diuretic in men. Finasteride is an inhibitor of the 5 -reductase enzyme, which is used to treat benign prostatic hyperplasia. Testosterone and Other Androgens Synthesis of Testosterone In men, testosterone is the principal secreted androgen. The Leydig cells synthesize the majority of testosterone by the pathways shown in Figure 591. In women, testosterone also is probably the principal androgen and is synthesized both in the corpus luteum and the adrenal cortex by similar pathways. The testosterone precursors androstenedione and dehydroepiandrosterone are weak androgens. Figure 591. Pathway of Synthesis of Testosterone in the Leydig Cells of the Testes. Bold arrows indicate favored pathways. DHEA, dehydroepiandrosterone. (Adapted from Santen, 1995, with permission.)

Secretion and Transport of Testosterone

The magnitude of testosterone secretion is greater in men than in women at almost all stages of life, a difference that explains almost all other differences between men and women. In the first trimester in utero, the fetal testes begin to secrete testosterone, which is the principal factor in male sexual differentiation, probably stimulated by human chorionic gonadotropin from the placenta. By the beginning of the second trimester, the value is close to that of midpuberty, about 250 ng/dl (Figure 592) (Dawood and Saxena, 1977; Forest, 1975). Testosterone production then falls by the end of the second trimester, but by birth the value is again about 250 ng/dl (Forest and Cathiard, 1975; Forest, 1975; Dawood and Saxena, 1977), possibly due to stimulation of the fetal Leydig cells by luteinizing hormone (LH) from the fetal pituitary gland. The testosterone value falls again in the first few days after birth, but it rises and peaks again at about 250 ng/dl at two to three months after birth and falls to <50 ng/dl by six months, where it remains until puberty (Forest, 1975). During puberty, from about age 12 to 17 years, the serum testosterone concentration in males increases to a much greater degree than in females, so that by early adulthood the serum testosterone concentration is 500 to 700 ng/dl in men, compared to 30 to 50 ng/dl in women. The magnitude of the testosterone concentration in the male is responsible for the pubertal changes that further differentiate men from women. As men age, their serum testosterone concentrations gradually decrease, which may contribute to other effects of aging in men. Figure 592. Schematic Representation of the Serum Testosterone Concentration from Early Gestation to Old Age.

LH, secreted by the gonadotroph cells of the pituitary (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors), is the principal stimulus of testosterone secretion in men, perhaps potentiated by follicle stimulating hormone (FSH), also secreted by the gonadotroph cells. GnRH from the hypothalamus (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors), in turn, stimulates LH secretion, and testosterone inhibits it, acting directly on the gonadotroph cell. LH is secreted in pulses, which occur approximately every two hours and are greater in magnitude in the morning. The pulsatility appears to result from pulsatile secretion of GnRH from the hypothalamus. Pulsatile administration of GnRH to men who are hypogonadal due to hypothalamic disease results in normal LH pulses and testosterone secretion, but continuous administration does not (Crowley et al., 1985). Testosterone secretion is likewise pulsatile and diurnal, with the highest plasma concentrations occurring at about 8 A.M. and the lowest at about 8 P.M. The morning peaks diminish as men age (Bremner et al., 1983).

In women, LH stimulates the corpus luteum (formed from the follicle after release of the ovum) to secrete testosterone. Under normal circumstances, however, estradiol and progesterone, not testosterone, are the principal inhibitors of LH secretion in women. Sex hormone binding globulin (SHBG) binds about 40% of circulating testosterone with high affinity. Albumin binds almost 60% of circulating testosterone with low affinity. Approximately 2% of testosterone is unbound or free. Metabolism of Testosterone to Active and Inactive Compounds Testosterone has many different effects in many different tissues. One of the mechanisms by which the varied effects are mediated is the metabolism of testosterone to two other active steroids, dihydrotestosterone and estradiol (Figure 593). Some effects of testosterone appear to be mediated by testosterone itself, some by dihydrotestosterone, and some by estradiol. Figure 593. Metabolism of Testosterone to Its Major Active and Inactive Metabolites.

The enzyme 5 -reductase irreversibly catalyzes the conversion of testosterone to dihydrotestosterone. Although both testosterone and dihydrotestosterone act via the same receptor, the androgen receptor, dihydrotestosterone binds with higher affinity (Wilbert et al., 1983) and activates gene expression more efficiently (Deslypere et al., 1992). As a result, testosterone, acting via dihydrotestosterone, is able to have effects in tissues that express 5 -reductase which it could not have if it were present only as testosterone. Two forms of 5 -reductase have been identified: type I, which is found predominantly in nongenital skin and the liver, and type II, which is found predominantly in urogenital tissue in men and genital skin in both men and women. The effects of dihydrotestosterone in these tissues are described below.

The enzyme complex aromatase, which is present in many tissues, especially the liver and adipose tissue, catalyzes the irreversible conversion of testosterone to estradiol. This conversion results in approximately 85% of circulating estradiol in men; the remainder is secreted directly by the testes, probably the Leydig cells (MacDonald et al., 1979). The effects of testosterone thought to be mediated via estradiol are described below. Testosterone is metabolized in the liver to androsterone and etiocholanolone (Figure 593), which are biologically inactive. Dihydrotestosterone is metabolized to androsterone, androstanedione, and androstanediol. Physiological and Pharmacological Effects of Androgens The biological effects of testosterone can be considered by the mechanisms by which they occur and by the tissues in which they occur at various stages of life. Testosterone can act as an androgen either directly by binding to the androgen receptor or indirectly by conversion to dihydrotestosterone, which also binds to the androgen receptor as described above. Testosterone also can act as an estrogen by conversion to estradiol, which binds to the estrogen receptor (Figure 594). Figure 594. Direct Effects of Testosterone and Effects Mediated Indirectly via Dihydrotestosterone or Estradiol.

Effects That Occur via the Androgen Receptor Testosterone and dihydrotestosterone both act as androgens via a single androgen receptor (Figure 595). The androgen receptor is a member of the superfamily of nuclear receptors, which includes steroid hormone receptors, thyroid hormone receptors, and orphan receptors (see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). Both testosterone and dihydrotestosterone bind to the hormone-binding domain of the androgen receptor, allowing the ligand-receptor complex to bind, via the DNAbinding domain of the receptor, to certain responsive genes. The ligand-receptor complex acts as a transcription factor complex and stimulates expression of those genes (Brinkmann and Trapman, 2000). Figure 595. Structure of the Androgen Receptor.

For many years, the mechanisms by which androgens had so many different actions in so many different tissues were not understood, but recently these mechanisms have become clearer. One mechanism is the higher affinity with which dihydrotestosterone binds to and activates the androgen receptor compared to testosterone (Deslypere et al., 1992; Wilbert et al., 1983). Another mechanism, postulated more recently, involves transcription cofactors, both coactivators and corepressors, that are tissue specific. The importance of the androgen receptor is illustrated by the consequences of its mutations. Predictably, mutations that either alter the primary sequence of the protein or cause a single amino acid substitution in the hormone- or DNA-binding domains result in resistance to the action of testosterone, beginning in utero (McPhaul and Griffin, 1999). Male sexual differentiation is, therefore, incomplete, as is pubertal development. Another kind of mutation occurs in patients who have spinal and bulbar muscular atrophy, known as Kennedy's disease. These patients have an expansion of the CAG repeat, which codes for glutamine, at the amino terminus of the molecule (Laspada et al., 1991). The result is very mild androgen resistance but progressively severe motor neuron atrophy. The mechanism by which the neuron atrophy occurs is unknown. Yet other kinds of androgen receptor mutations may explain why prostate cancer that is treated by androgen deprivation eventually becomes androgen-independent. Prostate cancer is initially at least partially androgen-sensitive, which is the basis for the initial treatment of metastatic prostate cancer by androgen deprivation. Metastatic prostate cancer often regresses initially in response to this treatment, but then becomes unresponsive to continued deprivation. Several mutations of the androgen receptor have been described in these patients, and it has been postulated that these mutations might allow the receptor to respond to ligands other than androgens or to act without ligand activation (Visakorpi et al., 1995). Effects That Occur via the Estrogen Receptor The effects of testosterone on at least one tissue are mediated by its conversion to estradiol, catalyzed by the aromatase enzyme complex. In the rare cases in which a male does not express aromatase (Carani et al., 1997; Morishma et al., 1995) or the estrogen receptor (Smith et al., 1994), the epiphyses do not fuse and long bone growth continues indefinitely. In addition, the patients are osteoporotic. Administration of estradiol corrects the bone abnormalities in patients with an aromatase defect (Bilezikian et al., 1998) but not an estrogen-receptor defect. There is evidence suggesting that conversion of testosterone to estradiol mediates male sexual behavior in rats, but similar evidence has not yet been found in human beings. Effects of Androgens at Different Stages of Life In Utero When the fetal testes, stimulated by human chorionic gonadotropin, begin to secrete testosterone at

about the eighth week of gestation, the high local concentration of testosterone around the testes stimulates the nearby Wolffian ducts to differentiate into the male internal genitalia: the epididymis, vas deferens, and seminal vesicles (George and Wilson, 1992). Further away, in the anlage of the external genitalia, testosterone is converted to dihydrotestosterone, which causes the development of the external genitalia: the penis, scrotum, and prostate (George and Wilson, 1992). The increase in testosterone at the end of gestation might result in further phallic growth. Infancy The consequences of the increase in testosterone secretion by the testes during the first few months of life are not yet known. Puberty Puberty in the male begins at a mean age of 12 years with an increase in the secretion of FSH and LH from the gonadotroph cells, stimulated by increased secretion of GnRH from the hypothalamus. The increased secretion of FSH and LH stimulate the testes, so, not surprisingly, the first sign of puberty is an increase in testicular size. The increase in testosterone production within the testes, along with the effect of FSH on the Sertoli cells, stimulates the development of the seminiferous tubules, which eventually produce mature sperm. Increased secretion of testosterone into the systemic circulation affects many tissues virtually simultaneously, and the changes in most of them occur gradually during the course of several years. The phallus enlarges in length and width, the scrotum becomes rugated, and the prostate begins secreting the fluid it contributes to the semen. The skin becomes coarser and oilier due to increased sebum production, which contributes to the development of acne. Sexual hair begins to grow, initially pubic and axillary hair, then hair on the lower legs, and finally other body hair and facial hair. Full development of the latter two may not occur until ten years after the start of puberty and marks the completion of puberty. Muscle mass and strength, especially of the shoulder girdle, increase, and subcutaneous fat decreases. Epiphyseal bone growth accelerates, resulting in the pubertal growth spurt, but epiphyseal maturation leads eventually to a slowing and then cessation of growth. Bone also becomes thicker. The increase in muscle mass and bone result in a pronounced increase in weight. Erythropoiesis increases, resulting in higher hematocrit and hemoglobin concentrations in men than boys or women. The larynx thickens, resulting in a lower voice. Libido develops. Other changes also may be the result of the increase in testosterone during puberty. Men tend to have a better sense of spatial relations than do women and to exhibit behavior that is different in some ways from that of women, including being more aggressive. Adulthood The serum testosterone concentration and the characteristics of the adult male are maintained largely during early adulthood and midlife. One change during this time is the gradual development of male pattern baldness, beginning with recession of hair at the temples and/or at the vertex. Two changes that can occur in the prostate gland during adulthood are of much greater medical significance. One is the gradual development of benign prostatic hyperplasia, which occurs to a variable degree in almost all men, sometimes to the degree of obstructing urine outflow by compressing the urethra as it passes through the prostate. This development is mediated by the conversion of testosterone to dihydrotestosterone within prostatic cells (Wilson, 1980). One current treatment of benign prostatic hyperplasia is based on inhibiting 5 -reductase II, which mediates this

conversion (McConnell et al., 1998), as discussed below. The other change that can occur in the prostate during adulthood is the development of cancer. Although no direct evidence suggests that testosterone causes the disease, prostate cancer is dependent on testosterone, at least to some degree and at some time in its course. This dependency is the basis of treating metastatic prostate cancer by lowering the serum testosterone concentration (Huggins and Hodges, 1941; Iversen et al., 1990). Senescence As men age, the serum testosterone concentration gradually declines (Figure 592) and the sex hormone-binding globulin concentration gradually increases, so that by age 80, the total testosterone concentration is approximately 85% and the free testosterone is approximately 40% of those at age 20 (Purifoy et al., 1981; Deslypere and Vermeulen, 1984). This fall in serum testosterone could contribute to several other changes that occur with increasing age in men, including decreases in energy, libido, muscle mass (Forbes, 1976) and strength (Murray et al., 1980), and bone mineral density (Riggs et al., 1982). The possibility of such a relationship is suggested by the occurrence of similar changes when men develop hypogonadism at a younger age due to known diseases, as discussed below. Consequences of Androgen Deficiency The consequences of androgen deficiency depend on the stage of life during which the deficiency first occurs and the degree of the deficiency. During Fetal Development Testosterone deficiency in a male fetus during the first trimester in utero causes incomplete sexual differentiation. Testosterone deficiency in the first trimester results only from testicular disease, such as deficiency of 17 -oxidoketoreductase; deficiency of LH secretion due to pituitary or hypothalamic deficiency does not result in testosterone deficiency during the first trimester, because Leydig-cell secretion of testosterone at that time is under the control of hCG from the placenta. Complete deficiency of testosterone secretion results in entirely female external genitalia; less severe testosterone deficiency results in incomplete virilization of the external genitalia proportionate to the degree of deficiency. Testosterone deficiency at this stage of development also leads to failure of the Wolffian ducts to differentiate into the male external genitalia, such as the vas deferens and seminal vesicles, but the mllerian ducts do not differentiate into the female external genitalia as long as testes are present and secrete mllerian inhibitory substance. Similar changes occur if testosterone is secreted normally, but its action is diminished because of an abnormality of the androgen receptor or of the 5 -reductase enzyme. Abnormalities of the androgen receptor can be quite variable. The most severe form results in complete absence of androgen action and a female phenotype; moderately severe forms result in partial virilization of the external genitalia; and the mildest forms permit normal virilization in utero and result only in impaired spermatogenesis in adulthood (McPhaul and Griffin, 1999). Abnormal 5 -reductase results in incomplete virilization of the external genitalia in utero but normal development of the male internal genitalia, which depends on testosterone per se (Wilson et al., 1993). Testosterone deficiency during the third trimester, due either to a testicular disease or a deficiency of fetal LH secretion, appears to have two known consequences. One is failure of the phallus to grow as much as it would normally. The result, called microphallus, is a common occurrence in boys later discovered to be unable to secrete LH due to abnormalities of GnRH synthesis. The other

consequence is failure of the testes to descend into the scrotum, called cryptorchidism, also a common occurrence in boys whose LH secretion is subnormal. Before Completion of Puberty When a boy can secrete testosterone normally in utero but loses the ability to do so before the anticipated age of puberty, the result is failure to complete puberty. All of the pubertal changes described above, including those of the external genitalia, sexual hair, muscle mass, voice, and behavior, fail to occur to a degree proportionate to the abnormality of testosterone secretion. In addition, if growth hormone secretion is normal when testosterone secretion is subnormal during the years of expected puberty, the long bones continue to lengthen because the epiphyses do not close. The result is longer arms and legs relative to the trunk; these proportions are referred to as eunuchoid. Another consequence of subnormal testosterone secretion during the age of expected puberty is enlargement of glandular breast tissue, called gynecomastia. After Completion of Puberty When the ability to secrete testosterone becomes impaired after the completion of puberty, regression of the pubertal effects of testosterone depends on both the degree and the duration of testosterone deficiency. When the degree of testosterone deficiency is substantial, libido and energy decrease within a week or two, but other testosterone-dependent characteristics decline more slowly. Decreases in muscle mass and strength probably can be detected by testing groups of men within a few months, but a clinically detectable decrease in muscle mass in an individual does not occur for several years. A pronounced decrease in hematocrit and hemoglobin will occur within several months. A decrease in bone mineral density probably can be detected by dual energy absorptiometry within two years, but an increase in fracture incidence likely would not occur for many years. A loss of sexual hair takes many years. In Women Loss of androgen secretion in women results in a decrease in sexual hair, but not for many years. Androgens may have other important effects in women, and the loss of androgens (especially severe loss of both ovarian and adrenal androgens, as occurs in panhypopituitarism) may result in the loss of these effects. Testosterone preparations that can yield serum testosterone concentrations in the physiological range in women currently are being developed. The availability of such preparations will allow determining if replacement of testosterone in androgen-deficient women will improve their libido, energy, muscle mass and strength, and bone mineral density. Therapeutic Androgen Preparations The need for a creative approach to pharmacotherapy with androgens arises from the fact that ingestion of testosterone is not an effective means of replacing testosterone deficiency. The reason is that, even though ingested testosterone is readily absorbed into the hepatic circulation, the hormone is catabolized so rapidly by the liver that it is not practical for a hypogonadal man to ingest it in sufficient amounts and with sufficient frequency to maintain a normal serum testosterone concentration. Most pharmaceutical preparations of androgens, therefore, are designed to bypass hepatic catabolism of testosterone. Another goal of androgen pharmacotherapy is to separate certain effects from others. Testosterone Esters

Esterifying a fatty acid to the 17 hydroxyl group of testosterone creates a compound that is even more lipophilic than testosterone itself. When an ester, such as testosterone enanthate (heptanoate) or cypionate (cyclopentylpropionate) (Figure 596) is dissolved in oil and administered intramuscularly every two to four weeks to hypogonadal men, the ester hydrolyzes in vivo and results in serum testosterone concentrations that range from higher than normal in the first few days after the injection to low-normal just before the next injection (Snyder and Lawrence, 1980; Figure 597). Attempts to decrease the frequency of injections by increasing the amount of each injection result in wider fluctuations and poorer therapeutic effects. The undecanoate ester of testosterone (Figure 596), when dissolved in oil and ingested orally, is absorbed into the lymphatic circulation, thus bypassing initial hepatic catabolism. Testosterone undecanoate in oil also can be injected and produces stable serum testosterone concentrations for a month (Zhang et al., 1998). The undecanoate ester of testosterone is not marketed in the United States. Figure 596. Structures of Androgens Available for Therapeutic Use.

Figure 597. Pharmacokinetic Profiles of Three Testosterone Preparations during Their Chronic Administration to Hypogonadal Men. Doses of each were given at time 0. [Adapted from Snyder and Lawrence (1980) (A): Yu et al., (1997) (B): and Wang et al., (2000) (C).] Dashed lines indicate range of normal levels.

Alkylated Androgens Several decades ago, chemists found that adding an alkyl group to the 17 position of testosterone (Figure 596) retarded hepatic catabolism of the molecule. Consequently, 17 -alkylated androgens do have an androgenic effect when administered orally. However, they do not appear to be as fully androgenic as testosterone itself, and they cause hepatotoxicity (Petera et al., 1962; Cabasso, 1994), whereas native testosterone does not. Transdermal Delivery Systems Recent attempts to avoid the destructive "first pass" of testosterone through the liver have employed novel delivery systems, instead of chemically modified testosterone, that release native testosterone across the skin in a controlled fashion. When these transdermal preparations are applied once a day,

they result in serum testosterone concentrations that fluctuate less than when testosterone esters are administered systemically. The first such preparation was a skin patch ( TESTODERM) designed to be applied to the scrotal skin (Findlay et al., 1989). The rationale for that location is that the scrotal skin is so thin that sufficient testosterone can be absorbed without the need for chemicals to facilitate its absorption. Subsequent patches were designed to be applied to nonscrotal skin (ANDRODERM, TESTODERM TTS ) and therefore employ chemicals to facilitate absorption (Yu et al., 1997; Dobs et al., 1999). A newer transdermal preparation (ANDROGEL ) employs a hydroalcoholic gel which is applied to nonscrotal skin (Wang et al., 2000). All of these preparations are applied once a day, and all produce serum testosterone concentrations within the normal range in the majority of hypogonadal men (Figure 597). Attempts to Design Selective Androgens Alkylated Androgens Decades ago, investigators attempted to synthesize analogs of testosterone that possessed greater anabolic effects than androgenic effects compared to native testosterone. Several compounds appeared to have such differential effects, based on a greater effect on the levator ani muscle compared to the ventral prostate of the rat (Hershberger and Meyer, 1953). These compounds were called anabolic steroids, and most are 17 -alkylated androgens, described above. None of these compounds, however, has been demonstrated to have such a differential effect in human beings. Nonetheless, they have enjoyed popularity among athletes who are attempting to improve their performance, as described below. Another alkylated androgen, 7 -methyl-19-nortestosterone, is poorly converted to dihydrotestosterone (Kumar et al., 1992). Selective Androgen-Receptor Modulator Stimulated by the development of selective estrogen-receptor modulators, which have estrogenic effects in some tissues but not others, investigators are now attempting to develop selective androgen-receptor modulators (Negro-Vilar, 1999). However, the selective effect of raloxifene (EVISTA ), the first estrogen-receptor modulator to be developed for clinical use, derives from its much greater affinity for the form of estrogen receptor expressed in certain tissues, such as bone and cardiac muscle, than for the form expressed in other tissues, such as breast and uterus. Because only one form of the androgen receptor is expressed, development of compounds that have certain androgen effects but not others is based, instead, on tissue specificity of coactivators and corepressors of androgen-receptor transcriptional activity. Endogenous protein coactivators and corepressors of androgen receptor-dependent transcription have been demonstrated (Moilanen et al., 1999), and a family of quinolinones that has selective androgen properties has been synthesized (Zhi et al.,1999). Therapeutic Uses of Androgens The clearest indication for administration of androgens is testosterone deficiency in men, i.e., treatment of male hypogonadism. Androgens also have been used in other situations in the past and likely will be used in yet others in the future. Male Hypogonadism Any of the transdermal testosterone preparations or testosterone esters described above can be used to treat testosterone deficiency. Monitoring treatment for beneficial and deleterious effects differs somewhat in adolescents and the elderly from that in other men.

Monitoring for Efficacy The goal of administering testosterone to a hypogonadal man is to mimic the normal serum concentration as closely as possible. Therefore, measuring the serum testosterone concentration during treatment is the most important aspect of monitoring testosterone treatment for efficacy. When the serum testosterone concentration is measured depends on the testosterone preparation used. When a transdermal preparation is used, the serum testosterone concentration can be measured on any day at any time, recognizing that, when a patch is used, the peak value will be found 2 to 4 hours after application of the patch for scrotal skin (Findlay et al., 1987), 2 to 4 hours after application of one patch for nonscrotal skin (TESTODERM TTS; Yu et al., 1997), and 6 to 9 hours after application of another patch for nonscrotal skin (ANDRODERM ; Dobs et al., 1999). The nadir, before the next application, will be about 60% to 70% of the peak value (Findlay et al., 1987). When the testosterone gel is used, there is no appreciable fluctuation during the course of the day, but steady-state values may not be reached for a month after the initiation of treatment. When the enanthate or cypionate esters of testosterone are administered once every two weeks, the serum testosterone concentration should be measured midway between doses. At each of these times, the serum testosterone concentration should be normal, and if not, the dosage schedule should be adjusted accordingly. If the cause of the testosterone deficiency is testicular disease, as indicated by an elevated serum LH concentration, adequacy of testosterone treatment also can be judged by its reduction to normal within two months of initiation of treatment (Snyder and Lawrence, 1980; Findlay et al., 1989). Normalization of the serum testosterone concentration results in normal virilization in men who are not normally virilized and maintenance of virilization in those who already are. Libido and energy in hypogonadal men should increase within a few weeks (Davidson et al., 1979). Muscle mass should increase, fat mass should decrease, and muscle strength should increase within a few months (Katznelson et al., 1996). Bone mineral density should increase to a maximum within two years (Snyder et al., 2000). Monitoring for Deleterious Effects When testosterone itself is administered, as in one of the transdermal preparations or as an ester that is hydrolyzed to testosterone (Caminos-Torres et al., 1977), it has no "side effects"i.e., no effects that endogenously secreted testosterone does not have, as long as the dose is not excessive. Modified testosterone compounds, such as the 17 -alkylated androgens, do have side effects. Even replacement of endogenously secreted testosterone levels, however, can have effects that are undesirable. Some effects occur shortly after testosterone administration is initiated, whereas others usually do not occur until administration has been continued for many years. Raising the serum testosterone concentration from prepubertal or midpubertal levels to that of an adult male at any age can result in undesirable effects similar to those that occur during puberty, including acne, gynecomastia, and more aggressive sexual behavior. Physiological amounts of testosterone do not appear to affect serum lipids or apolipoproteins. Replacement of physiological levels of testosterone occasionally may have undesirable effects in the presence of concomitant illnesses. For example, stimulation of erythropoiesis would increase the hematocrit from subnormal to normal in a healthy man, but would raise the hematocrit above normal in a man with a predisposition to erythrocytosis, such as in chronic pulmonary disease. Similarly, the mild degree of sodium and water retention with testosterone replacement would have no clinical effect in a healthy man but would exacerbate preexisting congestive heart failure. If the testosterone dose is excessive, erythrocytosis and, uncommonly, salt and water retention and peripheral edema occur even in men who have no predisposition to these conditions. When a man's serum testosterone concentration has been in the normal adult male range for many years, whether from endogenous secretion or exogenous

administration, and he is over age 40, he is subject to certain testosterone-dependent diseases, including benign prostatic hyperplasia and prostate cancer, as discussed above. The principal side effects of the 17 -alkylated androgens are hepatic, including cholestasis and, uncommonly, peliosis hepatis, blood-filled hepatic cysts. Hepatocellular cancer has been reported rarely, so that an etiologic link is uncertain. The 17 -alkylated androgens, especially in large amounts, may lower serum high-density-lipoprotein cholesterol. Monitoring at the Anticipated Time of Puberty Administration of testosterone to testosterone-deficient boys at the anticipated time of puberty should be guided by the considerations above, but also by the fact that testosterone accelerates epiphyseal maturation, leading initially to a growth spurt but then to epiphyseal closure and permanent cessation of linear growth. Consequently, the height and growth-hormone status of the boy must be considered. Boys who are short because of growth-hormone deficiency should be treated with growth hormone before their hypogonadism is treated with testosterone. Male Senescence Preliminary evidence suggests that increasing the serum testosterone concentration of men whose serum levels are subnormal for no reason other than their age will increase their bone mineral density and lean mass and decrease their fat mass (Snyder et al., 1999a; Snyder et al., 1999b). It is entirely uncertain at this time, however, if such treatment will worsen benign prostatic hyperplasia or increase the incidence of clinically detectable prostate cancer. Female Hypogonadism It remains to be determined if increasing the serum testosterone concentrations of women whose serum testosterone concentrations are below normal will improve their libido, energy, muscle mass and strength, and bone mineral density. Enhancement of Athletic Performance Some athletes take drugs, including androgens, to attempt to improve their performance. Because androgens taken for this purpose usually are taken surreptitiously, information about their possible effects is not as good as that for androgens taken for treatment of male hypogonadism. Kinds of Androgens Used Virtually all androgens produced for human or veterinary purposes have been taken by athletes. When use by athletes began more than two decades ago, 17 -alkylated androgens and other compounds that were thought to have greater anabolic effects than androgen effects relative to testosterone (so-called "anabolic steroids") were used most commonly. Because these compounds can be detected readily by organizations that govern athletic competitions, preparations that increase the serum concentration of testosterone itself, such as the testosterone esters or human chorionic gonadotropin, have increased in popularity. Testosterone precursors, such as androstenedione and dehydroepiandrosterone (DHEA), also have increased in popularity recently because they are not regulated by national governments or athletic organizations. Efficacy

Most studies of the effects of pharmacological doses of androgens on muscle strength have been uncontrolled, but in one study, testosterone or placebo was administered in a double-blind fashion. In that study, 43 men were randomized to one of four groups: strength training exercise with either 600 mg of testosterone enanthate once a week (more than six times a replacement dose) or placebo for testosterone; or no exercise with either testosterone or placebo. The men who received testosterone experienced an increase in fat-free mass and muscle strength compared to those who received placebo treatment, and the men who exercised simultaneously experienced even greater increases (Bhasin et al., 1997). In another double-blind study, men who took 100 mg of androstenedione three times a day for eight weeks did not experience an increase in muscle strength compared to men who took placebo. Failure of this treatment to increase muscle strength is not surprising, because it also did not increase the mean serum testosterone concentration (King et al., 1999). Side Effects Some side effects of taking pharmacological doses of androgens occur with all androgens and all circumstances, but others occur only with certain androgens or in certain circumstances. All androgens suppress gonadotropin secretion when taken in high doses and thereby suppress endogenous testicular function. The result is a decrease in both endogenous testosterone and sperm production, resulting in diminished fertility. If administration continues for many years, testicular size may diminish. Testosterone and sperm production usually return to normal within a few months of discontinuation but may take longer. High doses of androgens also causes erythrocytosis (Drinka et al., 1995). Androgens that can be converted to estrogens, such as testosterone itself, cause gynecomastia when administered in high doses. Androgens whose A ring has been modified so that it cannot be aromatized, such as dihydrotestosterone, do not cause gynecomastia even in high doses. The 17 -alkylated androgens are the only androgens that cause hepatotoxicity, as discussed above. These androgens also appear to be much more likely than others, when administered in high doses, to affect serum lipid concentrations, specifically to decrease high-density-lipoprotein (HDL) cholesterol and increase low-density-lipoprotein (LDL) cholesterol. Other side effects have been suggested by many anecdotes but have not been confirmed, including psychological disorders and sudden death due to cardiac disease, possibly related to changes in lipids or to coagulation activation. Certain side effects occur specifically in women and children. Both experience virilization, including facial and body hirsutism, temporal hair recession in a male pattern, and acne. Boys experience phallic enlargement and women clitoral enlargement. Boys and girls whose epiphyses have not yet closed experience premature closure and stunting of linear growth. Male Contraception Attempts currently are being made to develop androgens alone or in combination with other drugs as male contraceptives based on their ability to inhibit secretion of LH by the pituitary, which in turn decreases endogenous testosterone production. Because the concentration of testosterone within the testes is normally approximately one hundred times that in the peripheral circulation, and that concentration is necessary for spermatogenesis, suppression of endogenous testosterone production greatly diminishes spermatogenesis. Initial use of testosterone alone to suppress spermatogenesis, however, required administration of approximately twice as much testosterone

enanthate as would be used for physiological replacement, and even then spermatogenesis was not entirely suppressed in all men (WHO Task Force for the Regulation of Male Fertility, 1996). Other early attempts to suppress spermatogenesis employed a GnRH antagonist to suppress LH secretion combined with a physiological dose of testosterone to maintain a normal serum testosterone concentration (Pavlou et al., 1991). That combination is not practical for widespread use because existing GnRH antagonists require daily injection and have strong histamine-releasing properties. A more promising approach is the combination of a progestin with a physiological dose of testosterone to suppress LH secretion and spermatogenesis but provide a normal serum testosterone concentration (Bebb et al., 1996). Androgens currently being tested as part of male contraceptive regimens include an injectable form of testosterone undecanoate, which appears to produce a relatively stable serum testosterone concentration for a month (Zhang et al., 1999), and 7 -methyl19-nortestosterone, a synthetic androgen that cannot be metabolized to dihydrotestosterone (Cummings et al., 1998). Catabolic and Wasting States Testosterone, because of its anabolic effects, has been used in attempts to ameliorate catabolic and muscle-wasting states, but it has not been effective in most of these states. One exception is in the treatment of muscle wasting associated with acquired immunodeficiency syndrome (AIDS), which is accompanied by hypogonadism. Treatment of men with AIDS-related muscle wasting and subnormal serum testosterone concentrations increases their muscle mass and strength (Bhasin et al., 2000). Angioneurotic Edema Chronic androgen treatment of patients with angioneurotic edema effectively prevents attacks. The disease is caused by hereditary impairment of C1-esterase inhibitor or acquired development of antibodies against it (Cicardi et al., 1998). The 17 -alkylated androgens, such as stanozolol and danazol, act by stimulating the hepatic synthesis of the esterase inhibitor. In women, virilization is a potential side effect. In children virilization and premature epiphyseal closure prevent chronic use of androgens for prophylaxis, although they are used occasionally for treatment of acute episodes. Blood Dyscrasias Androgens once were employed to attempt to stimulate erythropoiesis in patients with anemias of various etiologies, but the availability of erythropoietin has supplanted that use. Androgens, such as danazol, still are used occasionally as adjunctive treatment for hemolytic anemia and idiopathic thrombocytopenic purpura that are refractory to first-line agents. Antiandrogens Because certain effects of androgens are undesirable, at least under certain circumstances, agents have been developed specifically to inhibit androgen synthesis or effects. Other drugs, originally developed for other purposes, have been found to be antiandrogens. When these drugs are used for their originally intended purposes, their antiandrogenic effects can be undesirable side effects, but some are used intentionally as antiandrogens. Inhibitors of Testosterone Synthesis Analogs of GnRH effectively inhibit testosterone synthesis by inhibiting LH secretion. GnRH antagonists block the action of endogenous GnRH at the gonadotroph cell's GnRH receptor.

Antagonists that are currently available require daily injection and have significant histaminereleasing properties, so their therapeutic use is not practical. GnRH "superactive" analogs, given repeatedly, down-regulate the GnRH receptor, and currently are available for treatment of metastatic prostate cancer (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). Some antifungal drugs of the imidazole family, such as ketoconazole (see Chapter 49: Antimicrobial Agents: Antifungal Agents), block the synthesis of steroids, including testosterone and cortisol (Feldman, 1986). Because of the inhibition of cortisol and hepatotoxicity, these drugs are not generally useful to inhibit androgen synthesis intentionally. Inhibitors of Androgen Action These drugs act by inhibiting the binding of androgens to the androgen receptor or by inhibiting 5 reductase. Androgen Receptor Antagonists Flutamide and Bicalutamide These are relatively potent androgen receptor antagonists which are limited in their effectiveness when used alone, because increased secretion of LH stimulates higher serum testosterone concentrations. They are used primarily in conjunction with a GnRH analog in the treatment of metastatic prostate cancer (see Chapter 52: Antineoplastic Agents). In this situation, they block the action of adrenal androgens, which are not inhibited by GnRH analogs. Survival rates in groups of patients with metastatic prostate cancer treated with a combination of a GnRH agonist and either flutamide (EVLEXIN) or bicalutamide (CASODEX) are similar to each other (Schellhammer, Sharifi, et al., 1995) and to survival rates in those treated by castration (Iversen et al., 1990). Bicalutamide is replacing flutamide for this purpose, because it appears to have less hepatotoxicity and needs to be taken only once a day instead of three times a day. Flutamide also has been used to treat hirsutism in women, and it appears to be as effective as any other treatment (Venturoli et al., 1999), but its hepatotoxicity cautions against its use for this cosmetic purpose. Spironolactone Spironolactone (ALDACTONE ; see Chapter 29: Diuretics) is an inhibitor of aldosterone which also is a weak inhibitor of the androgen receptor and a weak inhibitor of testosterone synthesis. When it is used for treatment of fluid retention or hypertension in men, gynecomastia is a common side effect (Caminos-Torres et al., 1977). Conversely, it can be used intentionally in women to treat hirsutism, for which it is approved by the U.S. Food and Drug Administration and is moderately effective (Cumming et al., 1982), but it may cause irregular menses. Cyproterone Acetate Cyproterone acetate is a progestin and a weak antiandrogen by virtue of binding to the androgen receptor. It is moderately effective in reducing hirsutism alone or in combination with an oral contraceptive (Venturoli et al., 1999), but it is not approved for use in the United States. Selective Androgen-Receptor Antagonists A group of quinoline derivatives has been developed that act as antagonists at the androgen receptor

in rat prostate glands but not in the pituitary. Analogous effects have not yet been demonstrated in human beings, but these compounds suggest the possible development of selective androgenreceptor antagonists. 5 -Reductase Inhibitors Finasteride (PROSCAR ) is an antagonist of 5 -reductase, especially the type II, so it blocks the conversion of testosterone to dihydrotestosterone, especially in the male external genitalia. It was developed as a treatment for benign prostatic hyperplasia, and it is approved in the United States and many other countries for this purpose. When it is administered to men with moderately severe symptoms due to obstruction of urinary tract outflow, serum and prostatic concentrations of dihydrotestosterone decrease, prostatic volume decreases, and urine flow rate increases (McConnell et al., 1998). Impotence is a well-documented although infrequent side effect of this use, although the mechanism is not understood. Finasteride also is approved for use in the treatment of male pattern baldness under the trade name PROPECIA , even though that effect is presumably mediated via the type I enzyme. It appears to be as effective as flutamide and the combination of estrogen and cyproterone in the treatment of hirsutism (Venturoli et al., 1999), but is not approved in the United States for this purpose. For further discussion of disorders of the testes and of sexual differentiation, see Chapters 325 and 328 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 60. Adrenocorticotropic Hormone; Adrenocortical Steroids and Their Synthetic Analogs; Inhibitors of the Synthesis and Actions of Adrenocortical Hormones
Overview Adrenocorticotropic hormone (ACTH, also called corticotropin) and the steroid hormone products of the adrenal cortex are considered together in this chapter, because the major physiological and pharmacological effects of ACTH result from its action to increase the circulating levels of adrenocortical steroids. Synthetic derivatives of ACTH are used principally in the diagnostic assessment of adrenocortical function. As all of the known therapeutic effects of ACTH can be achieved with corticosteroids, synthetic steroid hormones generally are used instead of ACTH for therapeutic applications. Corticosteroids and their biologically active synthetic derivatives differ in their metabolic (glucocorticoid) and electrolyte-regulating (mineralocorticoid) activities. These agents are employed at physiological doses for replacement therapy when endogenous production is impaired. In addition, glucocorticoids are potent suppressors of inflammation, and their use in a wide variety of inflammatory and autoimmune diseases makes them among the most frequently prescribed classes of drugs. Because they exert effects on almost every organ system, the clinical use of and withdrawal from corticosteroids are complicated by a number of serious side effects, some of which are life-threatening. Therefore, the decision to institute therapy with corticosteroids always requires careful consideration of the relative risks and benefits in each patient. Agents that inhibit various reactions in the steroidogenic pathway and thus alter the patterns of secretion of adrenocortical steroids are discussed, as are synthetic steroids, such as mifepristone

(see also Chapter 58: Estrogens and Progestins), that inhibit glucocorticoid action. Agents that inhibit the action of aldosterone are presented in Chapter 29: Diuretics; agents used to inhibit growth of steroid-dependent tumors are discussed in Chapter 52: Antineoplastic Agents. Adrenocorticotropic Hormone; Adrenocortical Steroids and Their Synthetic Analogs; Inhibitors of the Synthesis and Actions of Adrenocortical Hormones: Introduction History The clinical importance of the adrenal glands was first appreciated by Addison, who described fatal outcomes in patients with adrenal destruction in a presentation to the South London Medical Society in 1849. These studies, published subsequently (Addison, 1855), were soon extended by Brown-Squard, who demonstrated that bilateral adrenalectomy was fatal in laboratory animals. It later was shown that the adrenal cortex, rather than the medulla, was essential for survival in these experiments. Further studies demonstrated that the adrenal cortex regulated both carbohydrate metabolism and fluid and electrolyte balance. Efforts by a number of investigators ultimately led to the isolation and characterization of the various adrenocorticosteroids. Studies of the factors that regulated carbohydrate metabolism (termed glucocorticoids) culminated with the synthesis of cortisone, the first pharmacologically effective glucocorticoid to be available in large amounts. Subsequently, Tate and colleagues isolated and characterized a distinct corticosteroid, aldosterone, that had potent effects on fluid and electrolyte balance (and therefore was termed a mineralocorticoid). The isolation of distinct corticosteroids that regulated carbohydrate metabolism or fluid and electrolyte balance ultimately led to the concept that the adrenal cortex comprises two largely independent units: an outer zone that produces mineralocorticoids and an inner region that synthesizes glucocorticoids and weak androgens. Studies of the adrenocortical steroids also played a key part in delineating the role of the anterior pituitary in endocrine function. As early as 1912, Cushing described patients with hypercorticism, and later recognized that pituitary basophilism represented the cause of the adrenal overactivity (Cushing, 1932), thus establishing the link between the anterior pituitary and adrenal function. These studies ultimately led to the purification of ACTH (Astwood et al., 1952) and the determination of its chemical structure. ACTH was further shown to be essential in maintaining the structural integrity and steroidogenic capacity of the inner cortical zones. The role of the hypothalamus in pituitary control was established by Harris (1948), who further postulated that a soluble factor produced by the hypothalamus activated ACTH release. These investigations culminated with the determination of the structure of corticotropin-releasing hormone (CRH), a hypothalamic peptide that regulates secretion of ACTH from the pituitary (Vale et al., 1981). Shortly after synthetic cortisone became available, Hench and colleagues demonstrated the dramatic effect of glucocorticoids and ACTH in the treatment of rheumatoid arthritis (Hench et al., 1949). These studies set the stage for the clinical use of corticosteroids in a wide variety of diseases, as discussed below. Adrenocorticotropic Hormone (ACTH; Corticotropin) The sequence of human ACTH, a peptide of 39 amino acids, is shown in Figure 601. Whereas removal of a single amino acid at the amino terminus considerably impairs biological activity, a number of amino acids can be removed from the carboxy-terminal end without a marked effect. The structureactivity relationships of ACTH have been studied extensively, and it is believed that a stretch of four basic amino acids at positions 15 to 18 is an important determinant of high-affinity binding to the ACTH receptor, whereas amino acids 6 to 10 are important for receptor activation

(Imura, 1994). As discussed in Chapter 23: Opioid Analgesics and as schematized in Figure 601, ACTH is synthesized as part of a larger precursor protein, pro-opiomelanocortin (POMC), and is liberated from the precursor through proteolytic cleavage at dibasic residues by the enzyme prohormone convertase 1. Impaired processing of POMC due to a mutation in prohormone convertase 1 has been implicated in the pathogenesis of a human disorder presenting with adrenal insufficiency. Intriguingly, these patients also exhibit childhood obesity, hypogonadotropic hypogonadism, and diabetes (Jackson et al., 1997), suggesting other proteolytic targets for prohormone convertase 1. A number of other biologically important peptides, including endorphins, lipotropins, and the melanocyte-stimulating hormones (MSH), also are produced from the same prcursor. Figure 601. Processing of POMC to ACTH and the Sequence of ACTH. A schematic overview of the pathway by which pro-opiomelanocortin (POMC) is converted to ACTH and other peptides in the anterior pituitary is shown. The light blue boxes behind the ACTH structure indicate regions identified as important for steroidogenic activity (residues 610) and binding to the ACTH receptor (1518). The amino acid sequence of human ACTH is shown. LPH, lipotropin; MSH, melanocyte-stimulating hormone; PC1, prohormone convertase 1.

Actions on the Adrenal Cortex ACTH stimulates the adrenal cortex to secrete glucocorticoids, mineralocorticoids, and weak androgens such as androstenedione and dehydroepiandrosterone, which can be converted peripherally into more potent androgens. Based on histological analyses, the adrenal cortex originally was separated into three zones: the zona glomerulosa, zona fasciculata, and zona reticularis. Functionally, it is more useful to view the adrenal cortex as two discrete compartments: the outer zona glomerulosa, which secretes the mineralocorticoid aldosterone, and the inner zonae fasciculata/reticularis, which secrete the glucocorticoid cortisol as well as the adrenal androgens (Figure 602). The biochemical basis for these differences in steroidogenic output has been defined in considerable detail. Cells of the outer zone have receptors for angiotensin II and express aldosterone synthase, an enzyme that catalyzes the terminal reactions in mineralocorticoid biosynthesis. In contrast, cells of the inner zones lack receptors for angiotensin II and express two enzymes, steroid 17 -hydroxylase (P45017 ) and 11 -hydroxylase (P45011 ), that catalyze the production of glucocorticoids.

Figure 602. The Adrenal Cortex Contains Two Anatomically and Functionally Distinct Compartments. The major functional compartments of the adrenal cortex are shown, along with the steroidogenic enzymes that determine the unique profiles of corticosteroid products. Also shown are the predominant physiologic + regulators of steroid production: angiotensin II (A II) and K for the zona glomerulosa and ACTH for the zonae fasciculata/reticularis.

In the absence of the adenohypophysis, the inner zones of the cortex atrophy, and the production of glucocorticoids and adrenal androgens is markedly impaired. Although ACTH acutely can stimulate mineralocorticoid production by the zona glomerulosa, this zone is regulated predominantly by angiotensin II and extracellular K+ (see Chapter 31: Renin and Angiotensin) and does not undergo atrophy in the absence of ongoing stimulation by the pituitary gland. In the setting of persistently elevated ACTH, mineralocorticoid levels initially increase and then return to normal (a phenomenon termed ACTH escape). Persistently elevated levels of ACTH, due either to repeated administration of large doses of ACTH or to excessive endogenous ACTH production, induce hyperplasia and hypertrophy of the inner zones of the adrenal cortex, with overproduction of cortisol and adrenal androgens. Adrenal hyperplasia is most marked in congenital disorders of steroidogenesis, where ACTH levels are continuously elevated as a secondary response to impaired cortisol biosynthesis. Mechanism of Action ACTH stimulates the synthesis and release of adrenocortical hormones. As specific mechanisms for steroid hormone secretion have not been defined and since steroids do not accumulate appreciably in the gland, it is believed that the actions of ACTH to increase steroid hormone production are predominantly mediated at the level of de novo biosynthesis. ACTH, like most peptide hormones, interacts with a specific membrane receptor. As determined by gene cloning and sequencing, the human ACTH receptor is a member of the G proteincoupled receptor family, closely resembling in its structure the receptors for melanocyte-stimulating hormones (Cone and Mountjoy, 1993). ACTH acts through the G protein Gs to activate adenylyl cyclase and increase intracellular cyclic AMP content. Cyclic AMP is an obligatory second messenger for most, if not all, effects of ACTH on steroidogenesis. Mutations in the ACTH receptor have been associated with rare syndromes leading

to familial resistance to ACTH (Clark and Weber, 1998). Temporally, the response of adrenocortical cells to ACTH has two phases: an acute phase, which occurs within seconds to minutes, largely reflects an increased supply of cholesterol substrate to the steroidogenic enzymes; a chronic phase, which occurs over hours to days, results largely from increased transcription of the steroidogenic enzymes. A summary of the pathways of adrenal steroid biosynthesis and the structures of the major steroid intermediates and products of the human adrenal cortex are shown in Figure 603. The rate-limiting step in steroid hormone production is the conversion of cholesterol to pregnenolone, a reaction catalyzed by the cholesterol side-chain cleavage enzyme, designated P450scc. Most of the enzymes required for steroid hormone biosynthesis, including P450scc, are members of the cytochrome P450 superfamily, a related group of mixed-function oxidases that play important roles in the metabolism of xenobiotics such as drugs and environmental pollutants as well as in the biosynthesis of such endogenous compounds as steroid hormones, vitamin D, bile acids, fatty acids, prostaglandins, and biogenic amines (see Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination). The rate-limiting components in this reaction regulate the mobilization of substrate cholesterol and its delivery to the P450scc, located in the inner mitochondrial matrix. Figure 603. Pathways of Corticosteroid Biosynthesis. The steroidogenic pathways used in the biosynthesis of the corticosteroids are shown, along with the structures of the intermediates and products. The pathways that are unique to the zona glomerulosa are shown in blue, whereas those unique to the zonae fasciculata/reticularis are shown in gray. P450scc , cholesterol side-chain cleavage enzyme; 3 -HSD, 3 -hydroxysteroid dehydrogenase; P45017 , steroid 17 hydroxylase; P45021, steroid 21-hydroxylase; P450aldo, aldosterone synthase; P45011 , steroid 11 -hydroxylase.

The adrenal cortex uses multiple sources of cholesterol to ensure an adequate supply of substrate for steroidogenesis. These sources include (1) circulating cholesterol and cholesterol esters taken up via the low-density lipoprotein (LDL)- and high-density lipoprotein (HDL)-receptor pathways, (2) liberation of cholesterol from endogenous cholesterol ester stores via activation of cholesterol esterase, and (3) increased de novo biosynthesis. The mechanism(s) by which ACTH stimulates the translocation of cholesterol to the inner mitochondrial matrix are not well defined. Several candidate mediators of the acute delivery of

cholesterol to the mitochondria have been proposed, including a 30,000 dalton phosphoprotein induced by ACTH in all primary steroidogenic tissues, the peripheral benzodiazepine receptor, and sterol carrier protein-2. The cDNA encoding the 30,000 dalton phosphoprotein (designated the Steroidogenic Acute Regulatory Protein, or StAR) has been cloned and shown to activate steroidogenesis (Stocco and Clark, 1996). Significantly, mutations in the gene encoding StAR are found in patients with congenital lipoid adrenal hyperplasia, a rare congenital disorder in which adrenal cells become engorged with cholesterol deposits secondary to an inability to synthesize any steroid hormones (Lin et al., 1995). This finding points to a key role of StAR in the regulated delivery of cholesterol to the steroid biosynthetic pathway. An important component of the trophic effect of ACTH is the enhancement of transcription of the genes that encode the individual steroidogenic enzymes, with associated increases in the steroidogenic capacity of the gland. Although the molecular mechanisms are still under investigation, it appears that a variety of transcriptional regulators mediate the induction of the steroid hydroxylases by ACTH (Parker and Schimmer, 1995). Extraadrenal Effects of ACTH In large doses, ACTH causes a number of metabolic changes in adrenalectomized animals, including ketosis, lipolysis, hypoglycemia (immediately after treatment), and resistance to insulin (later after treatment). Because of the large doses of ACTH required, the physiological significance of these extraadrenal effects is doubtful. ACTH also improves learning in experimental animals; this latter effect appears to be nonendocrine and mediated via distinct receptors in the central nervous system. Patients with primary adrenal insufficiency and persistently elevated ACTH levels classically are hyperpigmented. This hyperpigmentation probably results from ACTH activating the MSH receptor on the melanocytes, perhaps a consequence of the identity of ACTH and MSH in the first 13 amino acids of each of their sequences. Regulation of ACTH Secretion Hypothalamic-Pituitary-Adrenal Axis The rates of secretion of glucocorticoids are determined by fluctuations in the release of ACTH by the pituitary corticotropes. These corticotropes, in turn, are regulated by corticotropin-releasing hormone (CRH), a peptide hormone released by CRH neurons of the endocrine hypothalamus. These three organs collectively are referred to as the hypothalamic-pituitary-adrenal (HPA) axis, an integrated system that maintains appropriate levels of glucocorticoids (see Figure 604 for an overview of this axis). There are three characteristic modes of regulation of the HPA axis: diurnal rhythm in basal steroidogenesis, negative feedback regulation by adrenal corticosteroids, and marked increases in steroidogenesis in response to stress. The diurnal rhythm is entrained by higher neuronal centers in response to sleep-wake cycles, such that levels of ACTH peak in the early morning hours, causing the circulating glucocorticoid levels to peak at approximately 8 A.M. As discussed below, negative feedback regulation occurs at multiple levels of the HPA axis and is the major mechanism that operates to maintain circulating glucocorticoid levels in the appropriate range. Stressful stimuli can override these normal negative feedback control mechanisms, leading to marked increases in plasma concentrations of adrenocortical steroids. Figure 604. Overview of the Hypothalamic-Pituitary-Adrenal (HPA) Axis and Its Bidirectional Communication with the Immune System. The complex regulatory interactions between the HPA axis and the immune/inflammatory network are shown. indicates a positive regulator, indicates a negative

regulator. IL-1, interleukin-1; IL-2, interleukin-2; IL-6, interleukin-6; TNF- , tumor necrosis factor ; CRH, corticotropin-releasing hormone.

Central Nervous System The central nervous system (CNS) integrates a number of different positive and negative influences on ACTH release (Figure 604). These signals converge on the CRH neurons, which are clustered largely in the parvocellular region of the paraventricular hypothalamic nucleus and make axonal connections to the median eminence of the hypothalamus (see Chrousos, 1995, see also Chapter 12: Neurotransmission and the Central Nervous System). Following release into the hypophyseal plexus, CRH is transported via this portal system to the pituitary, where it binds to specific membrane receptors on corticotropes. Upon CRH binding, the CRH receptor activates adenylyl cyclase and increases cyclic AMP levels within corticotropes, ultimately increasing both ACTH biosynthesis and secretion. The human CRH receptor has been cloned and shown to resemble most closely in sequence the calcitonin/vasoactive intestinal peptide/growth hormonereleasing hormone family of G proteincoupled receptors (Chen et al., 1993). Arginine Vasopressin Argininevasopressin (AVP) also acts as a secretagogue for corticotropes, significantly potentiating the effects of CRH. Animal studies have revealed that the potentiation of CRH action by AVP likely plays a physiologically significant role in the full magnitude of the stress response. AVP is produced in the parvocellular neurons of the paraventricular nucleus, like CRH, as well as by magnocellular neurons of the supraoptic nucleus; it is secreted into the pituitary plexus from the median eminence. After binding to specific G proteincoupled receptors of the V1b subtype, AVP activates phospholipase C, producing diacylglycerol and 1,4,5-inositol trisphosphate as messengers to release ACTH (see Chapters 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect and 12: Neurotransmission and the Central Nervous System); in contrast to CRH, AVP apparently does not increase ACTH synthesis. Negative Feedback of Glucocorticoids Glucocorticoids inhibit ACTH secretion via direct and indirect actions on CRH neurons to decrease

CRH mRNA levels and CRH release and via direct effects on corticotropes. The effect on CRH release may be mediated by specific corticosteroid receptors in the hippocampus, which are proposed to play important roles in negative feedback inhibition exerted by glucocorticoids. At lower cortisol levels, the mineralocorticoid (type I) receptor, which has a higher affinity for glucocorticoids and is the predominant form found in the hippocampus, is the major receptor species occupied. As glucocorticoid concentrations rise, the glucocorticoid (type II) receptor also becomes occupied as the capacity of the mineralocorticoid receptor is exceeded. Basal activity of the HPA axis apparently is controlled by both classes of receptor, whereas feedback inhibition by glucocorticoids predominantly involves the glucocorticoid receptor. In the pituitary, glucocorticoids act through the glucocorticoid receptor to inhibit the expression of POMC in corticotropes as well as the release of ACTH. These effects are both rapid (occurring within seconds to minutes and possibly mediated by glucocorticoid receptorindependent mechanisms) and delayed (requiring hours and involving changes in gene transcription mediated through the glucocorticoid receptor). The Stress Response Circumstances of stress overcome negative feedback regulation of the HPA axis, leading to a marked rise in the production of corticosteroids. Examples of stress signals include injury, hemorrhage, severe infection, major surgery, hypoglycemia, cold, pain, and fear. Although the precise mechanisms that underlie this stress response and the essential actions played by the glucocorticoids are not defined fully, it is clear that glucocorticoid secretion is vital for maintaining homeostasis in these stressful settings. As discussed below, complex interactions between the HPA axis and the immune system may be a fundamental physiological component of this stress response (see Sapolsky et al., 2000; Turnbull and Rivier, 1999). Assays for ACTH Initially, ACTH levels were measured by bioassays that measured induced steroid production or the depletion of adrenal ascorbic acid; such assays have been used to standardize ACTH amounts in different preparations used for both diagnostic and therapeutic purposes. Radioimmunoassays were developed to quantitate ACTH levels in individual patients, but they were not always reproducible, and their sensitivity did not always clearly differentiate between low and normal levels of the hormone. An immunoradiometric assay, which reliably measures ACTH levels, is now widely available. This assay, which uses two separate antibodies directed at distinct epitopes on the ACTH molecule, considerably increases the ability to differentiate between primary hypoadrenalism due to intrinsic adrenal disease and secondary forms of hypoadrenalism due to hypothalamic or pituitary disorders. Patients with primary adrenal insufficiency have high ACTH levels because they lack normal glucocorticoid feedback inhibition, whereas patients with secondary adrenal insufficiency have pituitary or hypothalamic disease resulting in low levels of ACTH. The immunoradiometric ACTH assay also is useful in differentiating between ACTH-dependent and ACTH-independent forms of hypercorticism: high ACTH levels are seen when pituitary adenomas (i.e., Cushing's disease) or nonpituitary tumors that secrete ACTH (i.e., ectopic ACTH) underlie the hypercorticism, whereas very low ACTH levels are seen in patients with excessive glucocorticoid production due to primary adrenal disorders. Despite its considerable strengths, one problem with the immunoradiometric ACTH assay is that its specificity for intact ACTH can lead to false, low values in patients with ectopic ACTH secretion; these tumors often secrete aberrantly processed forms of ACTH that have biological activity but do not react in the antibody assay. Therapeutic Uses and Diagnostic Applications of ACTH

There are anecdotal reports that selected conditions respond better to ACTH than to corticosteroids (e.g., multiple sclerosis), and some clinicians continue to advocate therapy with ACTH. Despite this, ACTH currently has only limited utility as a therapeutic agent. Therapy with ACTH is both less predictable and less convenient than is therapy with appropriate steroids. In addition, ACTH stimulates mineralocorticoid and adrenal androgen secretion and may therefore cause acute retention of salt and water as well as virilization. While ACTH and the corticosteroids are not pharmacologically equivalent, all of the known therapeutic effects of ACTH also can be achieved with appropriate doses of corticosteroids at a lesser risk of side effects. Testing the Integrity of the HPA Axis At present, the major clinical use of ACTH is in testing the integrity of the HPA axis to identify those patients needing supplemental steroid coverage in stressful situations. Other tests used to assess the HPA axis include the insulin tolerance test (see Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors) and the metyrapone test (discussed later in this chapter). ACTH purified from animal pituitary glands is available in long-lasting injectable gel preparations as a gelatin solution (H.P. ACTHAR GEL; 40 or 80 IU/vial). Cosyntropin (CORTROSYN) is a synthetic peptide that corresponds to residues 1 to 24 of human ACTH. At the considerably supraphysiological dose of 250 g, cosyntropin maximally stimulates adrenocortical steroidogenesis. In the rapid cosyntropin stimulation test, 250 g of cosyntropin is administered either intramuscularly or intravenously, with cortisol measured just before administration (baseline) and 30 to 60 minutes after cosyntropin administration. An increase in circulating cortisol to levels greater than 18 to 20 g/dl indicates a normal response (others also have included an increase of 7 g/dl over the baseline value as a positive response, although this is less widely accepted). In patients with pituitary or hypothalamic disease of recent onset or shortly after surgery for pituitary tumors, this standard cosyntropin stimulation test may be misleading, as the duration of ACTH deficiency may have been insufficient to cause significant adrenal atrophy with frank loss of steroidogenic capacity. For this latter group of patients, some experts advocate a "low-dose" cosyntropin stimulation test, in which 1 g of cosyntropin is administered intravenously, with cortisol measured just before and 30 minutes after cosyntropin administration (Abdu et al., 1999). Because cosyntropin is not generally available in a 1- g dose, the standard ampoule of cosyntropin (250 g) is diluted to permit accurate delivery of the 1- g challenge dose, with the cutoff for a normal response being the same as that for the standard test. Care must be taken to avoid adsorption of the ACTH to plastic tubing and to measure the plasma cortisol precisely at 30 minutes after the cosyntropin injection. Although some studies indicate that the low-dose test is more sensitive than the standard 250- g test, others report that this test also may fail to detect secondary adrenal insufficiency. As noted above, primary and secondary adrenocortical diseases are reliably distinguished using currently available sensitive assays for ACTH. Thus, longer-course ACTH stimulation tests rarely are used to differentiate between these disorders. CRH Stimulation Test Ovine CRH, also termed corticorelin (ACTHREL), is available for diagnostic testing of the HPA axis. In patients with documented ACTH-dependent hypercorticism, CRH testing may help differentiate between a pituitary source (i.e., Cushing's disease) and an ectopic source of ACTH. After two baseline blood samples are obtained fifteen minutes apart, corticorelin (1 g/kg) is administered intravenously over a 30- to 60-second interval, and blood samples are obtained at 15, 30, and 60 minutes for ACTH measurement. It is important that the blood samples be handled as recommended for the ACTH assay. At the recommended dose, CRH generally is well tolerated,

although flushing may occur, particularly if the dose is administered as a bolus. Patients with Cushing's disease respond to CRH with either a normal or an exaggerated increase in ACTH, whereas ACTH levels do not increase in patients with ectopic sources of ACTH. It should be noted that this test is not perfect: ACTH levels are induced by CRH in occasional patients with ectopic ACTH, whereas approximately 5% to 10% of patients with Cushing's disease fail to respond. To improve the diagnostic accuracy of the CRH stimulation test, some authorities advocate sampling of blood from the inferior petrosal sinus following peripheral administration of CRH. When performed by a skilled neuroradiologist, this procedure may increase diagnostic accuracy with a tolerable risk of complications from the catheterization procedure. Absorption and Fate ACTH is readily absorbed from parenteral sites. The hormone rapidly disappears from the circulation following intravenous administration; in human beings, the half-life in plasma is about 15 minutes, primarily due to rapid enzymatic hydrolysis. Toxicity of ACTH Aside from rare hypersensitivity reactions, the toxicity of ACTH is primarily attributable to the increased secretion of corticosteroids. Cosyntropin is generally less antigenic than native ACTH. Moreover, ACTH isolated from animal pituitaries contains significant amounts of vasopressin, which can lead to life-threatening hyponatremia. These factors make cosyntropin the preferred agent for clinical use. Adrenocortical Steroids The adrenal cortex synthesizes two classes of steroids: the corticosteroids (glucocorticoids and mineralocorticoids), which have 21 carbon atoms, and the androgens, which have 19 (Figure 603). The actions of corticosteroids historically were described as glucocorticoid (carbohydrate metabolismregulating) and mineralocorticoid (electrolyte balanceregulating), reflecting their preferential activities. In human beings, hydrocortisone (cortisol) is the main glucocorticoid, and aldosterone is the main mineralocorticoid. The mechanisms by which glucocorticoid biosynthesis is regulated by ACTH have been discussed above, and the regulation of aldosterone production is described in Chapter 31: Renin and Angiotensin. Table 601 shows typical rates of secretion of the physiologically most significant corticosteroids in human beingscortisol and aldosteroneas well as their normal concentrations in peripheral plasma. Although earlier studies had suggested that cortisol was produced at a daily rate of 20 mg, more recent studies indicate that the actual rate is closer to 10 mg/day (Esteban et al., 1991). Although the adrenal cortex is an important source of circulating androgens in women, patients with adrenal insufficiency can be restored to normal life expectancy by replacement therapy with glucocorticoid and mineralocorticoid. Thus, adrenal androgens are not essential for survival. There are, however, age-related changes in dehydroepiandrosterone (DHEA) levels, which peak in the third decade of life and decline progressively thereafter. Moreover, patients with a number of chronic diseases have very low DHEA levels, leading some to propose that DHEA treatment might at least partly alleviate the adverse consequences of aging. Based on these issues, there has been considerable discussion about the need for DHEA therapy in patients with primary or secondary adrenal insufficiency. In one study, the addition of DHEA (50 mg orally each morning) to the standard replacement regimen in women with adrenal insufficiency led to improved subjective wellbeing and sexuality (Arlt et al., 1999).

Physiological Functions and Pharmacological Effects Physiological Actions The effects of corticosteroids are numerous and widespread. Their diverse effects include alterations in carbohydrate, protein, and lipid metabolism; maintenance of fluid and electrolyte balance; and preservation of normal function of the cardiovascular system, the immune system, the kidney, skeletal muscle, the endocrine system, and the nervous system. In addition, by mechanisms that are still not fully understood, corticosteroids endow the organism with the capacity to resist stressful circumstances such as noxious stimuli and environmental changes. In the absence of the adrenal cortex, survival is made possible only by maintaining an optimal environment, including adequate and regular feedings, ingestion of relatively large amounts of sodium chloride, and maintenance of an appropriate environmental temperature. Until recently, corticosteroid effects were viewed as physiological (reflecting actions of corticosteroids at doses corresponding to normal daily production levels) or pharmacological (representing effects seen only at doses exceeding the normal daily production of corticosteroids). More recent concepts suggest that the antiinflammatory and immunosuppressive actions of corticosteroids, one of the major "pharmacological" uses of this class of drugs, also provide a protective mechanism in the physiological setting, since many of the immune mediators associated with the inflammatory response decrease vascular tone and could lead to cardiovascular collapse if unopposed by the adrenal corticosteroids. This hypothesis is supported by the fact that the daily production rate of cortisol can rise markedly (at least 10-fold) in the setting of severe stress. In addition, as discussed below, the pharmacological actions of corticosteroids in different tissues and many of their physiological effects seem to be mediated by the same receptor. Thus, the various glucocorticoid derivatives used currently as pharmacological agents have side effects on physiological processes that parallel their therapeutic effectiveness. The actions of corticosteroids are related in complex ways to those of other hormones. For example, in the absence of lipolytic hormones, cortisol has virtually no effect on the rate of lipolysis by adipocytes. Likewise, in the absence of glucocorticoids, epinephrine and norepinephrine have only minor effects on lipolysis. Administration of a small dose of a glucocorticoid, however, markedly potentiates the lipolytic action of these amines. These effects of corticosteroids that involve concerted actions with other hormonal regulators are termed permissive and most likely reflect steroid-induced changes in protein synthesis that, in turn, modify tissue responsiveness. Corticosteroids are grouped according to their relative potencies in Na+ retention, effects on carbohydrate metabolism (i.e., hepatic deposition of glycogen and gluconeogenesis), and antiinflammatory effects. In general, potencies of steroids as judged by their ability to sustain life in adrenalectomized animals closely parallel those determined for Na+ retention. Potencies based on effects on glucose metabolism closely parallel those for antiinflammatory effects. The effects on Na+ retention and the carbohydrate/antiinflammatory actions are not closely related and reflect selective actions at distinct receptors, as noted above. Based on these differential potencies, the corticosteroids traditionally are divided into mineralocorticoids and glucocorticoids. Estimates of potencies of representative steroids in these actions are listed in Table 602. Several steroids that are classified predominantly as glucocorticoids (e.g., cortisol and prednisone) also possess modest but significant mineralocorticoid activity and thus may affect fluid and electrolyte handling in the clinical setting. At doses used for replacement therapy in patients with primary adrenal insufficiency (see below), the mineralocorticoid effects of these "glucocorticoids" are insufficient to replace that of aldosterone, and concurrent therapy with a more potent mineralocorticoid generally + is needed. In contrast, aldosterone is exceedingly potent with respect to Na retention but has only

modest potency for effects on carbohydrate metabolism. At normal rates of secretion by the adrenal cortex or in doses that maximally affect electrolyte balance, aldosterone has no significant glucocorticoid activity and thus acts as a pure mineralocorticoid. General Mechanisms for Corticosteroid Effects Corticosteroids interact with specific receptor proteins in target tissues to regulate the expression of corticosteroid-responsive genes, thereby changing the levels and array of proteins synthesized by the various target tissues (see Figure 605). As a consequence of the time required for changes in gene expression and protein synthesis, most effects of corticosteroids are not immediate but become apparent after several hours. This fact is of clinical significance, because a delay generally is seen before beneficial effects of corticosteroid therapy become manifest. Although corticosteroids predominantly act to increase expression of target genes, there are well-documented examples where glucocorticoids decrease transcription of target genes, as discussed below. In contrast to these genomic effects, some actions of corticosteroids may be immediate and are mediated by membrane-bound receptors (Christ et al., 1999). Figure 605. Intracellular Mechanism of Action of the Glucocorticoid Receptor. The molecular pathway by which glucocorticoid steroids (labeled S) enter cells and interact with the glucocorticoid receptor to change the GR conformation (indicated by the change in shape of the GR), induce GR nuclear translocation, and activate transcription of target genes is shown. The example shown is one in which glucocorticoids activate expression of target genes; the expression of certain genes, including proopiomelanocortin (POMC) expression by corticotropes, is inhibited by glucocorticoid treatment. CBG, corticosteroid binding globulin; GR, glucocorticoid receptor; S, steroid hormone; HSP90, the 90-kDa heat shock protein; HSP70, the 70-kDa heat shock protein; IP, the 56kDa immunophilin; GRE, glucocorticoid-response elements in the DNA that are bound by GR, thus providing specificity to induction of gene transcription by glucocorticoids. Within the gene are introns (unshaded) and exons (shaded); transcription and mRNA processing leads to splicing and removal of introns and assembly of exons into mRNA.

Through the use of molecular biological approaches, the receptors for the corticosteroid hormones have been cloned and their structures determined. These receptors are members of a superfamily of structurally related proteins, the nuclear receptors, that transduce the effects of a diverse array of small, hydrophobic ligands, including the steroid hormones, thyroid hormone, vitamin D, and retinoids (Mangelsdorf et al., 1995). These receptors share two highly conserved domains: a region of approximately 70 amino acids forming two zinc-binding domains, termed zinc fingers, that are essential for the interaction of the receptor with specific DNA sequences, and a region at the carboxy terminus that interacts with ligand (the ligand-binding domain). Glucocorticoid Receptor As shown in Figure 605, the glucocorticoid receptor (GR) resides predominantly in the cytoplasm in an inactive form until it binds the glucocorticoid steroid ligand, denoted as S in the figure. Steroid binding results in receptor activation and translocation to the nucleus. The inactive GR is found as a complex with other proteins, including heat shock protein (HSP) 90, a member of the heat-shock family of stress-induced proteins; HSP70; and a 56,000 dalton immunophilin, one of the group of intracellular proteins that bind the immunosuppressive agents cyclosporine and tacrolimus (see Chapter 53: Immunomodulators: Immunosuppressive Agents, Tolerogens, and Immunostimulants for a discussion of these agents). HSP90, through interactions with the steroidbinding domain, may facilitate folding of GR into an appropriate conformation that is believed to be essential for ligand binding. Regulation of Gene Expression by Glucocorticoids Following ligand binding, the GR dissociates from its associated proteins and translocates to the nucleus. There, it interacts with specific DNA sequences within the regulatory regions of affected genes. The short DNA sequences that are recognized by the activated GR are termed glucocorticoid responsive elements (GREs) and provide specificity to the induction of gene transcription by

glucocorticoids. The consensus GRE sequence is an imperfect palindrome (GGTACAnnnTGTTCT, where n is any nucleotide) to which the GR binds as a receptor dimer. The mechanisms by which GR activates transcription are complex and not completely understood, but they appear to involve the interaction of the GR with transcriptional cofactors and with proteins that make up the basal transcription apparatus. Genes that are negatively regulated by glucocorticoids also have been identified (Webster and Cidlowski, 1999). One well-characterized example is the proopiomelanocortin gene (POMC), whose negative regulation in corticotropes by glucocorticoids is an important part of the negative feedback regulation of the HPA axis. In this case, the GR appears to inhibit transcription by a direct interaction with a GRE in the POMC promoter. Although glucocorticoids, and presumably the GR, are essential for survival, interactions of the GR with specific GREs apparently are not. These conclusions are supported by the findings that genetically engineered mice completely lacking GR function die immediately after birth, whereas mice harboring a mutated GR incapable of binding to DNA are viable (Reichardt et al., 1998). These observations imply that the critical function of GR involves proteinprotein interactions with other transcription factors (Xu et al., 1999). Indeed, proteinprotein interactions have been observed between the GR and the transcription factors NF- B and AP-1, which regulate the expression of a number of components of the immune system (McKay and Cidlowski, 1999). Such interactions repress the expression of genes encoding a number of cytokinesregulatory molecules that play key roles in the immune and inflammatory networksand enzymes, such as collagenase and stromelysin, that are proposed to play key roles in the joint destruction seen in inflammatory arthritis. Thus, these negative effects on gene expression appear to contribute significantly to the antiinflammatory and immunosuppressive effects of the glucocorticoids. Regulation of Gene Expression by Mineralocorticoids Like the glucocorticoid receptor, the mineralocorticoid receptor also is a ligand-activated transcription factor and binds to a very similar, if not identical, hormone-responsive element. Although its actions have been studied in less detail than the glucocorticoid receptor, the basic principles of action appear to be similar; in particular, the mineralocorticoid receptor also associates with HSP90 and also activates the transcription of discrete sets of genes within target tissues. Studies to date have not identified differences in the DNA recognition motifs for the glucocorticoid and mineralocorticoid receptors that would explain their differential abilities to activate discrete sets of target genes. Glucocorticoid and mineralocorticoid receptors differ in their ability to inhibit AP1mediated gene activation (Pearce and Yamamoto, 1993), suggesting that differential interactions with other transcription factors may underlie their distinct effects on cell function. In addition, unlike the glucocorticoid receptor, the mineralocorticoid receptor has a restricted expression; it is expressed principally in the kidney (distal cortical tubule and cortical collecting duct), colon, salivary glands, sweat glands, and hippocampus. Aldosterone exerts its effects on Na+ and K+ homeostasis primarily via its actions on the principal cells of the distal renal tubules and collecting ducts, while the effects on H+ secretion largely are exerted in the intercalated cells. Recent studies have identified some of the mechanisms by which aldosterone alters fluid and electrolyte transport. After binding to mineralocorticoid receptors in responsive cells, aldosterone initiates a sequence of events that includes the rapid induction of serum- and glucocorticoid-regulated kinase, which in turn phosphorylates and activates amiloride+ sensitive epithelial Na channels in the apical membrane (Chen et al., 1999). Thereafter, increased + + + Na influx stimulates the Na ,K ATPase in the basolateral membrane. In addition to these rapid actions, aldosterone also increases the synthesis of the individual components of these membrane proteins.

Further insights into the roles of the mineralocorticoid receptor and its target genes in fluid and electrolyte balance have emerged from analyses of patients with rare genetic disorders of mineralocorticoid action, such as pseudohypoaldosteronism and pseudoaldosteronism. Despite elevated levels of mineralocorticoids, patients with pseudohypoaldosteronism present with clinical manifestations suggestive of deficient mineralocorticoid action (i.e., volume depletion, hypotension, hyperkalemia, and metabolic acidosis). Molecular analyses have defined discrete subpopulations of patients with this disorder. One form is an autosomal recessive disease resulting from loss-offunction mutations in genes encoding subunits of the amiloride-sensitive epithelial sodium channel. A second, autosomal dominant form of pseudohypoaldosteronism is caused by mutations in the mineralocorticoid receptor that impair its activity (Geller et al., 1998). Pseudoaldosteronism, also termed Liddle's syndrome, is an autosomal dominant disease that results from activating mutations in the amiloride-sensitive Na+ channel (Shimkets et al., 1994). The constitutive activity of this channel leads to hypertension, hypokalemia, and metabolic alkalosis despite low levels of plasma renin and aldosterone. Receptor-Independent Mechanism for Corticosteroid Specificity The availability of cloned genes encoding the glucocorticoid receptor and mineralocorticoid receptor led to the surprising finding that aldosterone (a classic mineralocorticoid) and cortisol (generally viewed as predominantly glucocorticoid) bound the mineralocorticoid receptor with equal affinity. This raised the question of how the apparent specificity of the mineralocorticoid receptor for aldosterone was maintained in the face of much higher levels of circulating glucocorticoids. At least part of the answer came with the discovery of the type 2 isozyme of 11 hydroxysteroid dehydrogenase, a steroid-metabolizing enzyme that plays a key role in corticosteroid specificity, particularly in the kidney, colon, and salivary glands. This enzyme forms a barrier in certain mineralocorticoid-responsive tissues by metabolizing glucocorticoids such as cortisol to receptor-inactive 11-keto derivatives such as cortisone (Figure 606). Aldosterone escapes metabolism by 11 -hydroxysteroid dehydrogenase, because its predominant form in physiological settings is the hemiacetal derivative, which is resistant to 11 -hydroxysteroid dehydrogenase action. In the absence of 11 -hydroxysteroid dehydrogenase, as occurs in an inherited disease called the syndrome of apparent mineralocorticoid excess, the mineralocorticoid receptor is swamped by cortisol, leading to severe hypokalemia and mineralocorticoid-related hypertension. A state of hypermineralocorticism also can be induced by the inhibition of 11 hydroxysteroid dehydrogenase with glycyrrhizic acid, a component of licorice implicated in licorice-induced hypertension. Figure 606. Receptor-Independent Mechanism for Conferring Specificity of Glucocorticoid Action. By converting cortisol (which binds the mineralocorticoid receptor) to cortisone (which does not bind to the mineralocorticoid receptor), 11 -hydroxysteroid dehydrogenase protects the mineralocorticoid receptor from the high circulating concentrations of glucocorticoids, thereby allowing specific responses to aldosterone in classic mineralocorticoid-responsive cells.

Carbohydrate and Protein Metabolism Corticosteroids have profound effects on carbohydrate and protein metabolism. Teleologically, these effects of glucocorticoids on intermediary metabolism can be viewed as protecting glucosedependent tissues (e.g., the brain and heart) from starvation. This is achieved by stimulating the liver to form glucose from amino acids and glycerol and by stimulating the deposition of glucose as liver glycogen. In the periphery, glucocorticoids diminish glucose utilization, increase protein breakdown, and activate lipolysis, thereby providing amino acids and glycerol for gluconeogenesis. The net result is to increase blood glucose levels. Because of these effects on glucose metabolism, treatment with glucocorticoids can worsen control in patients with overt diabetes and can precipitate the onset of hyperglycemia in patients who are otherwise predisposed. The mechanisms by which glucocorticoids inhibit glucose utilization in peripheral tissues are not fully understood. Glucocorticoids decrease glucose uptake in adipose tissue, skin, fibroblasts, thymocytes, and polymorphonuclear leukocytes; these effects are postulated to result from translocation of the glucose transporters from the plasma membrane to an intracellular location. These peripheral effects are associated with a number of catabolic actions, including atrophy of lymphoid tissue, decreased muscle mass, negative nitrogen balance, and thinning of the skin. Similarly, the mechanisms by which the glucocorticoids promote gluconeogenesis are not fully defined. Amino acids mobilized from a number of tissues in response to glucocorticoids reach the liver and provide substrate for the production of glucose and glycogen. In the liver, glucocorticoids induce the transcription of a number of enzymes involved in gluconeogenesis and amino acid metabolism, including phosphoenolpyruvate carboxykinase (PEPCK), glucose-6-phosphatase, and fructose-2, 6-bisphosphatase. Analyses of the molecular basis for regulation of PEPCK gene expression have identified complex regulatory influences involving an interplay among glucocorticoids, insulin, glucagon, and catecholamine. The effects of these hormones and amines on PEPCK gene expression mirror the complex regulation of gluconeogenesis in the intact organism. Lipid Metabolism Two effects of corticosteroids on lipid metabolism are firmly established. The first is the dramatic redistribution of body fat that occurs in settings of hypercorticism, such as Cushing's syndrome. The other is the permissive facilitation of the effect of other agents, such as growth hormone and adrenergic receptor agonists, in inducing lipolysis in adipocytes, with a resultant increase in free fatty acids following glucocorticoid administration. With respect to fat distribution, there is increased fat in the back of the neck ("buffalo hump"), face ("moon facies"), and supraclavicular

area, coupled with a loss of fat in the extremities. One hypothesis for this phenomenon is that peripheral and truncal adipocytes differ in their relative sensitivities to insulin and to glucocorticoid-facilitated lipolytic effects. According to this hypothesis, truncal adipocytes respond predominantly to elevated levels of insulin resulting from glucocorticoid-induced hyperglycemia, whereas peripheral adipocytes are less sensitive to insulin and respond mostly to the glucocorticoid-facilitated effects of other lipolytic hormones. Electrolyte and Water Balance Aldosterone is by far the most potent naturally occurring corticosteroid with respect to fluid and electrolyte balance. Evidence for this comes from the relatively normal electrolyte balance found in hypophysectomized animals, despite the loss of glucocorticoid production by the inner cortical zones. Mineralocorticoids act on the distal tubules and collecting ducts of the kidney to enhance reabsorption of Na+ from the tubular fluid; they also increase the urinary excretion of both K + and H+. Conceptually, it is useful to think of aldosterone as stimulating a renal exchange between Na+ and K+ or H+ , although the molecular mechanism of monovalent cation handling is not a simple 1:1 exchange of cations in the renal tubule. These renal actions on electrolyte transport, in conjunction with similar effects in other tissues (e.g., colon, salivary glands, sweat glands), appear to account for the physiological and pharmacological activities that are characteristic of mineralocorticoids. Thus, the primary features of hyperaldosteronism are positive Na+ balance with consequent expansion of the extracellular fluid volume, normal or slight increases in plasma Na+ concentration, hypokalemia, and alkalosis. Mineralocorticoid deficiency, in contrast, leads to Na+ wasting and contraction of the extracellular fluid volume, hyponatremia, hyperkalemia, and acidosis. Chronically, hyperaldosteronism can cause hypertension, whereas aldosterone deficiency can lead to hypotension and vascular collapse. Because of the effects of mineralocorticoids on electrolyte handling by sweat glands, patients who are adrenal-insufficient are especially predisposed to Na+ loss and volume depletion through excessive sweating in hot environments. Glucocorticoids also exert effects on fluid and electrolyte balance, largely due to permissive effects on tubular function and actions that maintain glomerular filtration rate. Glucocorticoids play a permissive role in the renal excretion of free water; the ability to excrete a water challenge was used at one time to diagnose adrenal insufficiency. In part, the inability of Addisonian patients to excrete free water results from the increased secretion of AVP, which stimulates water reabsorption in the kidney. In addition to their effects on monovalent cations and water, glucocorticoids also exert multiple effects on Ca2+ metabolism. In the gut, steroids interfere with Ca2+ uptake by undefined mechanisms, while there is increased Ca2+ excretion at the level of the kidney. These effects collectively lead to decreased total body Ca2+ stores. Cardiovascular System As noted above, the most striking effects of corticosteroids on the cardiovascular system result from mineralocorticoid-induced changes in renal Na+ excretion, as is evident in primary aldosteronism. The resultant hypertension can lead to a diverse group of adverse effects on the cardiovascular system, including increased atherosclerosis, cerebral hemorrhage, stroke, and hypertensive cardiomyopathy. The mechanism underlying the hypertension remains incompletely understood,

but restriction of dietary Na+ can lower the blood pressure considerably. The second major action of corticosteroids on the cardiovascular system is to enhance vascular reactivity to other vasoactive substances. Hypoadrenalism generally is associated with hypotension and reduced response to vasoconstrictors such as norepinephrine and angiotensin II. This diminished pressor response is explained partly by studies in experimental systems showing that glucocorticoids increase expression of adrenergic receptors in the vascular wall. Conversely, hypertension is seen in patients with excessive glucocorticoid secretion, occurring in most patients with Cushing's syndrome and in a subset of patients treated with synthetic glucocorticoids (even those lacking any significant mineralocorticoid action). The underlying mechanisms in glucocorticoid-induced hypertension also are unknown; in hypertension related to the endogenous secretion of cortisol, as seen in patients with Cushing's syndrome, it is not known if the effects are mediated by the glucocorticoid or mineralocorticoid receptor. Unlike hypertension caused by high aldosterone levels, the hypertension secondary to excess glucocorticoids is generally resistant to Na+ restriction. Studies also have shown direct effects of aldosterone on both the heart and vascular lining; treating rats with aldosterone induced hypertension and interstitial cardiac fibrosis (Funder et al., 1997). The increased cardiac fibrosis was proposed to result from direct mineralocorticoid actions in the heart rather than from the effect of hypertension, because treatment with spironolactone, a mineralocorticoid antagonist, blocked the fibrosis without altering blood pressure. Similar effects of mineralocorticoids on cardiac fibrosis in human beings may explain, at least in part, the beneficial effects of the mineralocorticoid receptor antagonist spironolactone in patients with congestive heart failure (Pitt et al., 1999). Skeletal Muscle Permissive concentrations of corticosteroids are required for the normal function of skeletal muscle; diminished work capacity is a prominent sign of adrenocortical insufficiency. In patients with Addison's disease, weakness and fatigue are frequent symptoms and are believed to reflect mostly an inadequacy of the circulatory system. Excessive amounts of either glucocorticoids or mineralocorticoids also impair muscle function. In primary aldosteronism, muscle weakness results primarily from hypokalemia rather than from direct effects of mineralocorticoids on skeletal muscle. In contrast, glucocorticoid excess over prolonged periods, either secondary to glucocorticoid therapy or endogenous hypercorticism, tends to cause skeletal muscle wasting via unknown mechanisms. This effect, termed steroid myopathy, accounts in part for the weakness and fatigue noted in Cushingoid patients and is discussed in more detail below. Central Nervous System Corticosteroids exert a number of indirect effects on the CNS, through maintenance of blood pressure, plasma glucose concentrations, and electrolyte concentrations. Improved awareness of the distribution and function of steroid receptors in the brain has led to increasing recognition of direct effects of corticosteroids on the CNS, including effects on mood, behavior, and brain excitability. Patients with Addison's disease can exhibit a diverse array of psychiatric manifestations, including apathy, depression, and irritability; some patients are frankly psychotic. Appropriate replacement therapy corrects these abnormalities. Of greater clinical consequence, glucocorticoid administration can induce multiple CNS reactions. Most patients respond with mood elevation, which may impart a sense of well-being despite the persistence of underlying disease. Some patients exhibit more

pronounced behavioral changes, such as euphoria, insomnia, restlessness, and increased motor activity. A smaller but significant percentage of patients treated with glucocorticoids become anxious, depressed, or overtly psychotic. A high incidence of neuroses and psychoses has been noted among patients with Cushing's syndrome. These abnormalities usually disappear after cessation of glucocorticoid therapy or treatment of the Cushing's syndrome. The mechanisms by which corticosteroids affect neuronal activity are unknown, but it should be noted that steroids produced locally in the brain (termed neurosteroids) may regulate neuronal excitability (Baulieu, 1998). Studies in rodent models have long suggested that glucocorticoids deleteriously affect survival and function of hippocampal neurons, and that these changes are associated with diminished memory (Lupien and McEwan, 1997). A study in human beings used basal cortisol levels over time to establish a correlation between increased cortisol levels and hippocampal atrophy and memory deficits (Lupien et al., 1998). To the extent that these results can be confirmed, they have important prognostic implications for age-related memory decline, and they suggest therapeutic approaches directed at diminishing the negative effects of glucocorticoids on hippocampal neurons with aging. Formed Elements of Blood Glucocorticoids exert minor effects on hemoglobin and erythrocyte content of blood, as evidenced by the frequent occurrence of polycythemia in Cushing's syndrome and of normochromic, normocytic anemia in Addison's disease. More profound effects are seen in the setting of autoimmune hemolytic anemia, where the immunosuppressive effects of glucocorticoids can diminish the self-destruction of erythrocytes. Corticosteroids also affect circulating white blood cells. Addison's disease, as noted by Addison in his initial report, is associated with an increased mass of lymphoid tissue and lymphocytosis. In contrast, Cushing's syndrome is characterized by lymphocytopenia and decreased mass of lymphoid tissue. The administration of glucocorticoids leads to a decreased number of circulating lymphocytes, eosinophils, monocytes, and basophils. A single dose of hydrocortisone leads to a decline of these circulating cells within 4 to 6 hours; this effect persists for 24 hours and results from the redistribution of cells away from the periphery rather than from increased destruction. In contrast, glucocorticoids increase circulating polymorphonuclear leukocytes as a result of increased release from the marrow, diminished rate of removal from the circulation, and increased demargination from vascular walls. Certain lymphoid malignancies, however, are destroyed by glucocorticoid treatment. This latter effect may be related to the ability of glucocorticoids to activate programmed cell death in certain lymphoid tissues. Antiinflammatory and Immunosuppressive Actions In addition to their effects on lymphocyte number, corticosteroids profoundly alter the immune responses of lymphocytes. These effects are an important facet of the antiinflammatory and immunosuppressive actions of the glucocorticoids. Glucocorticoids can prevent or suppress inflammation in response to multiple inciting events, including radiant, mechanical, chemical, infectious, and immunological stimuli. Although the use of glucocorticoids as antiinflammatory agents does not address the underlying cause of the disease, the suppression of inflammation is of enormous clinical utility and has made these drugs among the most frequently prescribed agents. Similarly, glucocorticoids are of immense value in treating diseases that result from undesirable immune reactions. These diseases range from conditions that predominantly result from humoral immunity, such as urticaria (see Chapter 65: Dermatological Pharmacology), to those that are mediated by cellular immune mechanisms, such as transplantation rejection (see Chapter 53:

Immunomodulators: Immunosuppressive Agents, Tolerogens, and Immunostimulants). The immunosuppressive and antiinflammatory actions of glucocorticoids are inextricably linked, perhaps because they both involve inhibition of leukocyte functions (Chrousos, 1995). Multiple mechanisms are involved in the suppression of inflammation by glucocorticoids. It is now clear that glucocorticoids inhibit the production by multiple cells of factors that are critical in generating the inflammatory response. As a result, there is decreased release of vasoactive and chemoattractive factors, diminished secretion of lipolytic and proteolytic enzymes, decreased extravasation of leukocytes to areas of injury, andultimatelydecreased fibrosis. Some of the cell types and mediators that are inhibited by glucocorticoids are summarized in Table 603. The net effect of these actions on various cell types is to diminish markedly the inflammatory response. The influence of stressful conditions on immune defense mechanisms is well documented, as is the contribution of the HPA axis to the stress response (Sapolsky et al., 2000). This has led to a growing appreciation of the importance of glucocorticoids as physiological modulators of the immune system, where glucocorticoids appear to protect the organism against life-threatening consequences of a full-blown inflammatory response. Stresses such as injury, infection, and disease result in the increased production of cytokines, a network of signaling molecules that integrate actions of macrophages/monocytes, T lymphocytes, and B lymphocytes in mounting immune responses. Among these cytokines, interleukin (IL)-1, IL6, and tumor necrosis factor- (TNF- ) stimulate the HPA axis, with IL-1 having the broadest range of actions. IL-1 stimulates the release of CRH by hypothalamic neurons, interacts directly with the pituitary to increase the release of ACTH, and may directly stimulate the adrenal gland to produce glucocorticoids (Turnbull and Rivier, 1999). As detailed above, the increased production of glucocorticoids, in turn, leads to a profound inhibition of the immune system at multiple sites. Factors that are inhibited include components of the cytokine network, including interferon gamma (INF- ), granulocyte/monocyte colony-stimulating factor (GM-CSF) interleukins (IL-1, IL-2, IL-3, IL-6, IL-8, IL-12), and TNF- . Thus, the HPA axis and the immune system are capable of bidirectional interactions in response to stress, and these interactions appear to be important for homeostasis. Although glucocorticoids traditionally have been considered as immunosuppressive agents, there are intriguing observations suggesting that glucocorticoids produced as part of the physiological response to stress may upregulate the humoral arm of the immune response (e.g., antibody production) while suppressing cellular immunity (Elenkov and Chrousos, 1999). The mechanisms underlying this glucocorticoid-induced switch are unclear but seem to involve inhibition of T-helper (Th-1) cells and activation of Th-2 cells (see Chapter 53: Immunomodulators: Immunosuppressive Agents, Tolerogens, and Immunostimulants). Absorption, Transport, Metabolism, and Excretion Absorption Hydrocortisone and numerous congeners, including the synthetic analogs, are effective when given by mouth. Certain water-soluble esters of hydrocortisone and its synthetic congeners are administered intravenously to achieve high concentrations of drug rapidly in body fluids. More prolonged effects are obtained by intramuscular injection of suspensions of hydrocortisone, its congeners, and its esters. Minor changes in chemical structure may markedly alter the rate of absorption, time of onset of effect, and duration of action.

Glucocorticoids also are absorbed systemically from sites of local administration, such as synovial spaces, the conjunctival sac, skin, and respiratory tract. When administration is prolonged, when the site of application is covered with an occlusive dressing, or when large areas of skin are involved, the absorption may be sufficient to cause systemic effects, including suppression of the HPA axis. Transport, Metabolism, and Excretion Following absorption, 90% or more of cortisol in plasma is reversibly bound to protein under normal circumstances. Only the fraction of corticosteroid that is unbound can enter cells to mediate corticosteroid effects. Two plasma proteins account for almost all of the steroid-binding capacity: corticosteroid-binding globulin (CBG; also called transcortin), and albumin. CBG is an -globulin secreted by the liver that has high affinity for steroids but relatively low total binding capacity, whereas albumin, also produced by the liver, has low affinity but relatively large binding capacity. At normal or low concentrations of corticosteroids, most of the hormone is protein-bound. At higher steroid concentrations, the capacity of protein binding is exceeded, and a significantly greater fraction of the steroid exists in the free state. Corticosteroids compete with each other for binding sites on CBG. CBG has relatively high affinity for cortisol and most of its synthetic congeners and low affinity for aldosterone and glucuronide-conjugated steroid metabolites; thus, greater percentages of these latter steroids are found in the free form. During pregnancy or estrogen treatment, CBG, total plasma cortisol, and free cortisol increase severalfold. The physiological significance of these changes remains to be established. All of the biologically active adrenocortical steroids and their synthetic congeners have a double bond in the 4,5 position and a ketone group at C 3. As a general rule, the metabolism of steroid hormones involves sequential additions of oxygen or hydrogen atoms, followed by conjugation to form water-soluble derivatives. Reduction of the 4,5 double bond occurs at both hepatic and extrahepatic sites, yielding inactive compounds. Subsequent reduction of the 3-ketone substituent to the 3-hydroxyl derivative, forming tetrahydrocortisol, occurs only in the liver. Most of these A ringreduced steroids are conjugated through the 3-hydroxyl group with sulfate or glucuronide by enzymatic reactions that take place in the liver and, to a lesser extent, in the kidney. The resultant sulfate esters and glucuronides are water-soluble and are the predominant forms excreted in the urine. Neither biliary nor fecal excretion is of quantitative importance in human beings. Synthetic steroids with an 11-keto substituent, such as cortisone and prednisone, must be enzymatically reduced to the corresponding 11 -hydroxy derivative before they are biologically active. This reaction is catalyzed in the liver by the type 1 isozyme of 11 -hydroxysteroid dehydrogenase, which operates in a reductive mode. In settings where this enzymatic activity is impaired, such as severe hepatic failure, or in rare patients who lack this enzyme, it is prudent to use steroids that do not require enzymatic activation (e.g., hydrocortisone and prednisolone rather than cortisone or prednisone). StructureActivity Relationships Chemical modifications to the cortisol molecule have generated derivatives with greater separations of glucocorticoid and mineralocorticoid activity; for a number of synthetic glucocorticoids, the effects on electrolytes are minimal even at the highest doses used. In addition, these modifications have led to derivatives with greater potencies and with longer durations of action. A vast array of different steroid preparations is therefore available for oral, parenteral, and topical use. Some of these agents are summarized in Table 604. However, because the antiinflammatory and metabolic effects of glucocorticoids are mediated by the same glucocorticoid receptor, the various derivatives

do not effectively separate antiinflammatory effects from effects on carbohydrate, protein, and fat metabolism or from suppressive effects on the HPA axis. The structures of hydrocortisone (cortisol) and some of its major derivatives are shown in Figure 607. Changes in chemical structure may bring about changes in specificity and/or potency as a result of changes in affinity and intrinsic activity at corticosteroid receptors, alterations in absorption, protein binding, rate of metabolic transformation, rate of excretion, or membrane permeability. The effects of various substitutions on glucocorticoid and mineralocorticoid activity and on duration of action are summarized in Table 602. The 4,5 double bond and the 3-keto group on ring A are essential for both glucocorticoid and mineralocorticoid activity; an 11 -hydroxyl group on ring C is required for glucocorticoid activity but not mineralocorticoid activity; a hydroxyl group at C 21 on ring D is present on all natural corticosteroids and on most of the active synthetic analogs and seems to be an absolute requirement for mineralocorticoid activity, but not glucocorticoid activity. The 17 -hydroxyl group on ring D is a substituent on cortisol and on all of the currently used synthetic glucocorticoids. While steroids without the 17 -hydroxyl group (e.g., corticosterone) have appreciable glucocorticoid activity, the 17 -hydroxyl group gives optimal potency. Figure 607. Structure and Nomenclature of Corticosteroid Products and Selected Synthetic Derivatives. The structure of hydrocortisone is represented in two dimensions. It should be noted that the steroid ring system is not completely planar and that the orientation of the groups attached to the steroid rings is an important determinant of the biological activity. The methyl groups at C 18 and C 19 and the hydroxyl group at C 11 project upward (forward in the twodimensional representation and shown by a solid line connecting the atoms) and are designated . The hydroxyl at C 17 projects below the plane (behind in the two-dimensional representation, and represented by the dashed line connecting the atoms) and is designated .

Introduction of an additional double bond in the 1,2 position of ring A, as in prednisolone or prednisone, selectively increases glucocorticoid activity (approximately fourfold compared to hydrocortisone), resulting in an enhanced glucocorticoid to mineralocorticoid potency ratio. This modification also results in compounds that are metabolized more slowly than hydrocortisone.

Fluorination at the 9 position on ring B enhances both glucocorticoid and mineralocorticoid activity and possibly is related to an electron-withdrawing effect on the nearby 11 -hydroxyl group. Fludrocortisone (9 -fluorocortisol) has enhanced activity at the glucocorticoid receptor (10-fold relative to cortisol) but even greater activity at the mineralocorticoid receptor (125-fold relative to cortisol). It is used in mineralocorticoid replacement therapy (see below) and has no appreciable glucocorticoid effect at usual daily doses of 0.05 to 0.2 mg. When combined with the 1,2 double bond in ring A and other substitutions at C 16 on ring D (Figure 607), the 9 -fluoro derivatives formed (e.g., triamcinolone, dexamethasone, betamethasone) have marked glucocorticoid activity. The substitutions at C 16 virtually eliminate mineralocorticoid activity. Other Substitutions 6 Substitution on ring B has somewhat unpredictable effects. 6 -Methylcortisol has increased glucocorticoid and mineralocorticoid activity, whereas 6 -methylprednisolone has somewhat greater glucocorticoid activity and somewhat less mineralocorticoid activity than prednisolone. A number of modifications convert the glucocorticoids to more lipophilic molecules with enhanced topical to systemic potency ratios. Examples include the introduction of an acetonide between hydroxyl groups at C 16, C 17, esterification of the hydroxyl group with valerate at C 17, esterification of hydroxyl groups with propionate at C 17 and C 21, and substitution of the hydroxyl group at C 21 with chlorine. Other approaches to achieve local glucocorticoid activity while minimizing systemic effects involve the formation of analogs that are rapidly inactivated following absorption. Examples of this latter group include C 21 carboxylate or carbothioate glucocorticoid esters, which are rapidly metabolized to inactive 21-carboxylic acids. Toxicity of Adrenocortical Steroids Two categories of toxic effects result from the therapeutic use of corticosteroids: those resulting from withdrawal of steroid therapy and those resulting from continued use of supraphysiological doses. The side effects from both of these categories are potentially life-threatening and mandate a careful assessment of the risks and benefits in each patient. Withdrawal of Therapy Withdrawal of corticosteroid therapy poses a number of difficult decisions. It is important to remember that the most frequent problem in steroid withdrawal is flare-up of the underlying disease for which steroids were prescribed. There are several complications associated with steroid withdrawal, as discussed by Sullivan (1982). The most severe complication of steroid cessation, acute adrenal insufficiency, results from too rapid withdrawal of corticosteroids after prolonged therapy, where the HPA axis has been suppressed. The therapeutic approach to acute adrenal insufficiency is detailed below. There is significant variation among patients with respect to the degree and duration of adrenal suppression following corticosteroid therapy, making it difficult to establish the relative risk in any given patient. Many patients recover from corticosteroid-induced HPA suppression within several weeks to months; however, in some individuals, the time to recovery can be one year or longer. In an effort to diminish the risk of iatrogenic acute adrenal insufficiency, protocols for discontinuing corticosteroid therapy in patients receiving long-term treatment with corticosteroids have been proposed (for example, see Byyny, 1976). In general, patients who have received supraphysiological doses of glucocorticoids for a period of two weeks within the preceding year should be considered to have some degree of HPA impairment in settings of acute stress and should

be treated accordingly. In addition to this most severe form of withdrawal, a characteristic glucocorticoid withdrawal syndrome consists of fever, myalgias, arthralgias, and malaise, which may be difficult to differentiate from some of the underlying diseases for which steroid therapy was instituted. Finally, pseudotumor cerebri, a clinical syndrome that includes increased intracranial pressure with papilledema, is a rare condition that sometimes is associated with reduction or withdrawal of corticosteroid therapy. Continued Use of Supraphysiological Corticosteroid Doses Besides the consequences that result from the suppression of the HPA axis, there are a number of other complications that result from prolonged therapy with corticosteroids. These include fluid and electrolyte abnormalities, hypertension, hyperglycemia, increased susceptibility to infection, osteoporosis, myopathy, behavioral disturbances, cataracts, growth arrest, and the characteristic habitus of steroid overdose including fat redistribution, striae, ecchymoses, acne, and hirsutism. Fluid and Electrolyte Handling Alterations in fluid and electrolyte handling can cause hypokalemic alkalosis, edema, and hypertension, particularly in patients with primary hyperaldosteronism secondary to an adrenal adenoma or in patients treated with potent mineralocorticoids. Similarly, hypertension is a relatively common manifestation in patients with endogenous glucocorticoid excess and also can be seen in patients treated with glucocorticoids lacking appreciable mineralocorticoid activity. Hyperglycemia with glycosuria usually can be managed with diet and/or insulin, and its occurrence should not be a major factor in the decision to continue corticosteroid therapy or to initiate therapy in diabetic patients. Immune Responses Because of their multiple effects to inhibit the immune system and the inflammatory response, glucocorticoid use also is associated with an increased susceptibility to infection and a risk for reactivation of latent tuberculosis. In the presence of known infections of some consequence, glucocorticoids should be administered only if absolutely necessary and concomitantly with appropriate and effective antimicrobial or antifungal therapy. Possible Risk of Peptic Ulcers There is considerable debate about the association between peptic ulcers and glucocorticoid therapy. The possible onset of hemorrhage and perforation in these ulcers and their insidious onset make peptic ulcers serious therapeutic problems (see Chapter 37: Agents Used for Control of Gastric Acidity and Treatment of Peptic Ulcers and Gastroesophageal Reflux Disease); estimating the degree of risk from corticosteroids has received much study. One report indicates that most patients who develop gastrointestinal bleeding while receiving corticosteroids also received nonsteroidal antiinflammatory drugs, which are known to promote ulceration; the pathogenic role of corticosteroids thus remains open to debate (Piper et al., 1991). Nonetheless, it is prudent to be especially vigilant for peptic ulcer formation in patients receiving therapy with corticosteroids, especially when administered concomitantly with nonsteroidal antiinflammatory drugs. Myopathy

Myopathy, characterized by weakness of proximal limb muscles, occasionally is seen in patients taking large doses of corticosteroids and also is part of the clinical picture in patients with endogenous Cushing's syndrome. It can be of sufficient severity to impair ambulation and is an indication for withdrawal of therapy. Attention also has focused on steroid myopathy of the respiratory muscles in patients with asthma or chronic obstructive pulmonary disease (see Chapter 28: Drugs Used in the Treatment of Asthma); this complication can diminish respiratory function. Recovery from the steroid myopathies may be slow and incomplete. Behavioral Changes Behavioral disturbances are seen commonly after administration of corticosteroids and in patients who have Cushing's syndrome secondary to endogenous hypercorticism; these disturbances may take many forms, including nervousness, insomnia, changes in mood or psyche, and overt psychosis (Haskett, 1985). Suicidal tendencies are not uncommon. A history of previous psychiatric illness does not preclude the use of steroids in patients for whom they are otherwise indicated. Conversely, the absence of a history of previous psychiatric illness does not guarantee that a given patient will not develop psychiatric disorders while on steroids. Cataracts Cataracts are a well-established complication of glucocorticoid therapy and are related both to dosage and duration of therapy. Children appear to be particularly at risk. Cessation of therapy may not lead to complete resolution of opacities, and the cataracts may progress despite reduction or cessation of therapy. Patients on long-term glucocorticoid therapy at doses of prednisone of 10 to 15 mg/day or greater should receive periodic slit-lamp examinations to detect glucocorticoid-induced posterior subcapsular cataracts. Osteoporosis Osteoporosisa frequent serious complication of glucocorticoid therapyoccurs in patients of all ages and is related to both dosage and duration of therapy (Lane and Lukert, 1998). A reasonable estimate is that 30% to 50% of all patients who receive chronic glucocorticoid therapy ultimately will develop osteoporotic fractures. Glucocorticoids preferentially affect trabecular bone and the cortical rim of the vertebral bodies; the ribs and vertebrae are the most frequent sites of fracture. Glucocorticoids decrease bone density by multiple mechanisms, including inhibition of gonadal steroid hormones, diminished gastrointestinal absorption of calcium, and inhibition of bone formation due to suppressive effects on osteoblasts. In addition, glucocorticoid inhibition of intestinal calcium uptake may lead to secondary increases in parathyroid hormone, thereby increasing bone resorption. The considerable morbidity of glucocorticoid-related osteoporosis has led to efforts to identify patients at risk for fractures and to prevent or reverse bone loss in patients requiring chronic glucocorticoid therapy. The initiation of glucocorticoid therapy is considered an indication for bone densitometry, preferably with techniques such as dual-energy X-ray absorptiometry of the lumbar spine or hip that most sensitively detect abnormalities in trabecular bone. Because bone loss associated with glucocorticoids predominantly occurs within the first six months of therapy, densitometric evaluation and prophylactic measures should be initiated coincident with therapy or shortly thereafter. Most authorities advocate maintaining a calcium intake of 1500 mg/day by diet plus calcium supplementation and vitamin D intake of 400 IU/day, assuming that these measures do not increase urinary calcium excretion above the normal range. Unless contraindicated, gonadal hormone replacement therapy is indicated in specific groups of patients receiving chronic

glucocorticoid therapy, including postmenopausal females, premenopausal females with decreased estradiol levels, and males with decreased testosterone levels. The most important advance in the prevention of glucocorticoid-related osteoporosis is the successful use of bisphosphonates. Several different agents have been shown to decrease the decline in bone density in patients receiving glucocorticoid therapy. In particular, both alendronate (Saag et al., 1998) and cyclical etidronate have been shown to be effective both in primary prevention and in restoration of bone density in patients receiving chronic therapy with glucocorticoids. Additional discussion of these issues is found in Chapters 58: Estrogens and Progestins and 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds. Osteonecrosis Osteonecrosis (also known as avascular or aseptic necrosis) is a relatively common complication of glucocorticoid therapy (Lane and Lukert, 1998). The femoral head is affected most frequently, but this process also may affect the humeral head and distal femur. Joint pain and stiffness are usually the earliest symptoms, and this diagnosis should be considered in patients receiving glucocorticoids who abruptly develop hip, shoulder, or knee pain. Although the risk increases with both the duration and dose of glucocorticoid therapy, osteonecrosis also can occur when high doses of glucocorticoids are given for short periods of time. Osteonecrosis generally progresses, and most affected patients ultimately require joint replacement. Regulation of Growth and Development Growth retardation can result from administration of relatively small doses of glucocorticoids to children. Although the precise mechanism is unknown, there are reports that collagen synthesis and linear growth in these children can be restored by treatment with growth hormone; further studies are needed to define the role of concurrent treatment with growth hormone in this setting. Further studies also are needed to explore the possible effects of exposure to corticosteroids in utero. Studies in experimental animals have shown that antenatal exposure to glucocorticoids is clearly linked to cleft palate and altered neuronal development, ultimately resulting in complex behavioral abnormalities. Thus, although the actions of glucocorticoids to promote cellular differentiation play important physiological roles in human development in the neonatal period (e.g., induction of the hepatic gluconeogenic enzymes and surfactant production in the lung), the possibility remains that antenatal steroids can lead to subtle abnormalities in fetal development. Therapeutic Uses With the exception of replacement therapy in deficiency states, the use of glucocorticoids largely is empirical. Based on extensive clinical experience, a number of therapeutic principles can be proposed. First, given the number and severity of potential side effects, the decision to institute therapy with glucocorticoids always requires a careful consideration of the relative risks and benefits in each patient. For any disease and in any patient, the appropriate dose to achieve a given therapeutic effect must be determined by trial and error and must be reevaluated periodically as the activity of the underlying disease changes or as complications of therapy arise. A single dose of glucocorticoid, even a large one, is virtually without harmful effects, and a short course of therapy (up to 1 week), in the absence of specific contraindications, is unlikely to be harmful. As the duration of glucocorticoid therapy is increased beyond 1 week, there are time- and dose-related increases in the incidence of disabling and potentially lethal effects. Except in patients receiving replacement or substitution therapy, glucocorticoids are neither specific nor curative; instead, they

provide palliation by virtue of their antiinflammatory and immunosuppressive actions. Finally, abrupt cessation of glucocorticoids after prolonged therapy is associated with a significant risk of adrenal insufficiency, which may be fatal. These principles have several implications for clinical practice. When glucocorticoids are to be given over long periods, the dose must be determined by trial and error and must be the smallest one that will achieve the desired effect. When the therapeutic goal is relief of painful or distressing symptoms not associated with an immediately life-threatening disease, complete relief is not sought, and the steroid dose is reduced gradually until worsening symptoms indicate that the minimal acceptable dose has been found. Where possible, the substitution of other medications, such as nonsteroidal antiinflammatory drugs, may facilitate the tapering process once the initial benefit of glucocorticoid therapy has been achieved. When therapy is directed at a life-threatening disease (e.g., pemphigus), the initial dose should be a large one aimed at achieving rapid control of the crisis. If some benefit is not observed quickly, then the dose should be doubled or tripled. After initial control in a potentially lethal disease, reduction of dose should be carried out under conditions that permit frequent, accurate observations of the patient. It is always essential to weigh carefully the relative dangers of therapy and of the disease being treated. The lack of demonstrated deleterious effects of a single dose of glucocorticoids within the conventional therapeutic range justifies their administration to critically ill patients who may have adrenal insufficiency. If the underlying condition does result from deficiency of glucocorticoids, then a single intravenous injection of a soluble glucocorticoid may prevent immediate death and allow time for a definitive diagnosis to be made. If the underlying disease is not adrenal insufficiency, the single dose will not harm the patient. In the absence of specific contraindications, short courses of high-dose, systemic glucocorticoids also may be given for diseases that are not life-threatening, but the general rule is that long courses of therapy at high doses should be reserved for life-threatening disease. In selected settings, as when a patient is threatened with permanent disability, this rule is justifiably violated. In an attempt to dissociate therapeutic effects from undesirable side effects, various regimens of steroid administration have been utilized. In an attempt to diminish HPA axis suppression, alternate-day therapy with relatively short-lived glucocorticoids (e.g., prednisone) has been employed. Certain patients obtain adequate therapeutic responses on this regimen. Alternatively, pulse therapy with higher glucocorticoid doses (e.g., doses as high as 1.0 to 1.5 g/day of methylprednisolone for three days) frequently is used to initiate therapy in patients with fulminant, immunologically related disorders such as acute transplantation rejection, necrotizing glomerulonephritis, and lupus nephritis (Boumpas et al., 1993). The benefit of such pulse therapy in long-term maintenance regimens remains to be defined. Replacement Therapy Adrenal insufficiency can result from structural or functional lesions of the adrenal cortex (primary adrenal insufficiency) or from structural or functional lesions of the anterior pituitary or hypothalamus (secondary adrenal insufficiency). In developed countries, primary adrenal insufficiency most frequently is secondary to autoimmune adrenal disease, whereas tuberculous adrenalitis is the most frequent etiology in underdeveloped countries. Other causes include adrenalectomy, bilateral adrenal hemorrhage, the acquired immunodeficiency syndrome, and Xlinked adrenoleukodystrophy (Carey 1997). Secondary adrenal insufficiency resulting from pituitary or hypothalamic dysfunction generally presents in a more insidious manner than does the

primary disorder. Acute Adrenal Insufficiency This life-threatening disease is characterized by gastrointestinal symptoms (nausea, vomiting, and abdominal pain), dehydration, hyponatremia, hyperkalemia, weakness, lethargy, and hypotension. It usually is associated with disorders of the adrenal rather than the pituitary or hypothalamus, and it frequently follows abrupt withdrawal of glucocorticoids used at high doses or for prolonged periods. The presence of pigmentation is diagnostically useful in identifying patients with primary adrenal disease. The immediate management of patients with acute adrenal insufficiency includes intravenous therapy with isotonic sodium chloride solution supplemented with 5% glucose and corticosteroids and appropriate therapy for precipitating causes such as infection, trauma, or hemorrhage. Because cardiovascular function often is reduced in the setting of adrenocortical insufficiency, the patient should be monitored for evidence of volume overload such as rising central venous pressure or pulmonary edema. After an initial intravenous bolus of 100 mg, hydrocortisone (cortisol) should be given by continuous infusion at a rate of 100 mg every 8 hours. In this dose, which approximates the maximum daily rate of cortisol secretion in response to stress, hydrocortisone alone has sufficient mineralocorticoid activity to meet all requirements. As the patient stabilizes, intramuscular hydrocortisone may be used in a dose of 25 mg every 6 to 8 hours. Thereafter, patients are treated in the same fashion as those with chronic adrenal insufficiency (see below). For the treatment of suspected but unconfirmed acute adrenal insufficiency, 4 mg of dexamethasonesodium phosphate can be substituted for hydrocortisone, since dexamethasone does not cross-react in the cortisol assay and will not interfere with the measurement of cortisol (either basally or in response to the cosyntropin stimulation test). A failure to respond to cosyntropin in this setting is diagnostic of adrenal insufficiency. A sample for the measurement of plasma ACTH often also is obtained, as it provides information about the underlying etiology if the diagnosis of adrenocortical insufficiency is established. Chronic Adrenal Insufficiency Patients with chronic adrenal insufficiency present with many of the same manifestations seen in adrenal crisis, but with lesser severity. These patients require daily treatment with corticosteroids. Traditional replacement regimens have used hydrocortisone in doses of 20 to 30 mg/day. Cortisone acetate, which is inactive until converted to cortisol by 11 -hydroxysteroid dehydrogenase, also has been used in doses ranging from 25 to 37.5 mg/day. In an effort to mimic the normal diurnal rhythm of cortisol secretion, these glucocorticoids generally have been given in divided doses, with twothirds of the dose given in the morning and one-third given in the afternoon. Based on revised estimates of daily cortisol production (Esteban et al., 1991) and clinical studies showing that subtle degrees of glucocorticoid excess can decrease bone density in patients on conventional replacement regimens (Zelissen et al., 1994), many authorities advocate a daily hydrocortisone dose of 20 mg/day divided into either two doses (e.g., 15 mg on awakening and 5 mg in late afternoon) or three doses (e.g., 10 mg on awakening, 5 mg at lunch, and 5 mg in late afternoon). Others prefer to use long-acting glucocorticoids, such as prednisone or dexamethasone, since no regimen employing shorter-acting steroids can reproduce the peak serum cortisol levels that normally occur prior to awakening in the morning. The superiority of any one of these regimens has not been rigorously demonstrated. Although some patients with primary adrenal insufficiency can be maintained on hydrocortisone and liberal salt intake, most of these patients also require mineralocorticoid replacement; fludrocortisone acetate generally is used in doses of 0.05 to 0.2 mg/day. For patients

with secondary adrenal insufficiency, the administration of a glucocorticoid alone is generally adequate, as the zona glomerulosawhich makes mineralocorticoidsis intact. When initiating treatment in patients with panhypopituitarism, it is important to administer glucocorticoids first before initiating treatment with thyroid hormone, because the administration of thyroid hormone may precipitate acute adrenal insufficiency. The adequacy of corticosteroid replacement therapy is judged by clinical criteria and biochemical measurements. The subjective well-being of the patient is an important clinical parameter in both primary and secondary disease. In primary adrenal insufficiency, the disappearance of hyperpigmentation and the resolution of electrolyte abnormalities are valuable indicators of adequate replacement. Overtreatment may cause manifestations of Cushing's syndrome in adults and decreased linear growth in children. Plasma ACTH levels may be used to monitor therapy in patients with primary adrenal insufficiency; the early morning ACTH level should not be suppressed, but should be less than 100 pg/ml (20 pmol/liter). Although advocated by some endocrinologists, assessments of daily profiles of cortisol based on multiple blood sampling or measurements of urinary free cortisol have been used more frequently as research tools than as a routine part of clinical practice. Standard doses of glucocorticoids often must be adjusted upward in patients who also are taking drugs that increase their metabolic clearance (e.g., phenytoin, barbiturates, rifampin). Dosage adjustments also are needed to compensate for the stress of intercurrent illness, and proper patient education is essential for the execution of these adjustments. All patients with adrenal insufficiency should wear a medical alert bracelet or tag that lists their diagnosis and carries information about their steroid regimen. During minor illness, the glucocorticoid dose should be doubled. Patients should be instructed to contact their physician if nausea and vomiting preclude the retention of oral medications. It also is highly recommended that the patient and family members be instructed so that they can administer parenteral dexamethasone (4 mg subcutaneously or intramuscularly) in the event that severe nausea or vomiting precludes the oral administration of medications. They then should seek medical attention immediately. Based largely on empirical data, glucocorticoid doses also are adjusted when patients with adrenal insufficiency undergo either elective or emergency surgery. In this setting, the doses are designed to approximate or exceed the maximal cortisol secretory rate of 200 mg/day; a standard regimen is hydrocortisone, 100 mg parenterally every 6 to 8 hours. Following surgery, the dose is halved each day until it is reduced to routine maintenance levels. Although some data suggest that increases in dose to this degree are not essential for survival even in major surgery (Glowniak and Loriaux, 1997), this approach remains the standard clinical practice at present. Congenital Adrenal Hyperplasia This term denotes a group of genetic disorders in which the activity of one of the several enzymes required for the biosynthesis of corticosteroids is deficient. The impaired production of cortisol, aldosterone, or both and the consequent lack of negative feedback inhibition lead to increased release of ACTH and/or angiotensin II. As a result, other hormonally active steroids that are proximal to the enzymatic block in the steroidogenic pathway are overproduced. Congenital adrenal hyperplasia (CAH) includes a spectrum of disorders whose precise clinical presentation, laboratory findings, and treatment depend on which of the steroidogenic enzymes is deficient (see Donohoue et al., 2000, for a general discussion of the various forms of CAH). In approximately 90% of patients, congenital adrenal hyperplasia (CAH) results from mutations in CYP21, the enzyme that carries out the 21-hydroxylation reaction (New, 1998). Clinically, patients are divided into those with classical CAH, who have severe defects in enzymatic activity and first

present during childhood, and those with nonclassical CAH, who present after puberty with signs and symptoms of mild androgen excess such as hirsutism, amenorrhea, infertility, and acne. Female patients with classical CAH, if not treated in utero with glucocorticoids, frequently are born with virilized external genitalia (female pseudohermaphroditism), which results from elevated production of adrenal androgens at critical stages of sexual differentiation in utero. Males appear normal at birth and later may have precocious development of secondary sexual characteristics (isosexual precocious puberty). In both sexes, linear growth is accelerated in childhood, but the height at maturity is reduced by premature closure of the epiphyses. In a subset of patients with classical CAH, the enzymatic deficiency is sufficiently severe to compromise aldosterone production. Such patients are unable to conserve Na+ normally and thus are termed "salt wasters." These patients can present with cardiovascular collapse secondary to volume depletion; in an effort to prevent such life-threatening events, especially in males who appear normal at birth, some locations mandate routine screening of all babies for elevated levels of 17-hydroxyprogesterone, the immediate steroid precursor to the enzymatic block. All patients with classical CAH require substitution therapy with hydrocortisone or a suitable congener, and those with salt wasting also require mineralocorticoid replacement. The goals of therapy are to restore levels of physiological steroid hormones to the normal range, as well as to suppress ACTH and thereby abrogate the effects of overproduction of adrenal androgens. The typical oral dose of hydrocortisone is approximately 0.6 mg/kg daily in two or three divided doses. The mineralocorticoid used is fludrocortisone acetate (0.05 to 0.2 mg/day). Many experts also administer table salt to infants (one-fifth of a teaspoon dissolved in formula daily) until the child is eating solid food. Therapy is guided by gain in weight and height, by plasma levels of 17hydroxyprogesterone, and by blood pressure. Elevated plasma renin activity suggests that the patient is receiving an inadequate dose of mineralocorticoid. Sudden spurts in linear growth often indicate inadequate pituitary suppression and excessive androgen secretion, whereas growth failure often suggests overtreatment with glucocorticoid. The development of methods to detect classical CAH (21-hydroxylase deficiency) prenatally has made possible the treatment of affected females with glucocorticoids in utero, thereby eliminating the need for genital surgery to correct the virilization of the external genitalia (New, 1998). Glucocorticoid therapy (e.g., dexamethasone, 20 g/kg taken daily orally by mothers at risk) must be initiated before 10 weeks' gestation, before a definitive diagnosis of CAH can be made, to suppress effectively fetal adrenal androgen production. The genotype and sex of the fetus then are determined: If the sex is male or there is at least one wild-type allele for 21-hydroxylase, steroid therapy is stopped. If genotyping reveals an affected female, steroid therapy is continued until delivery. Potential maternal side effects include hypertension, weight gain, edema, and mood changes. Although it theoretically is possible that exposure to glucocorticoids in utero may have developmental consequences, adverse effects have not yet been described. Therapeutic Uses in Nonendocrine Diseases Given below are brief outlines of important uses of glucocorticoids in diseases that do not directly involve the HPA axis. The disorders discussed are not inclusive; rather, they illustrate the principles governing glucocorticoid use in selected diseases for which they are more frequently employed. The dosage of glucocorticoids varies considerably depending on the nature and severity of the underlying disorder. For convenience, approximate doses of a representative glucocorticoid (generally prednisone) are provided in the following discussion. This choice is not an endorsement of one particular glucocorticoid preparation over other congeners but is made for illustrative

purposes only. Rheumatic Disorders Glucocorticoids are used widely in the treatment of a variety of rheumatic disorders and are a mainstay in the treatment of the more serious inflammatory rheumatic diseases, such as systemic lupus erythematosus, and a variety of vasculitic disorders, such as polyarteritis nodosa, Wegener's granulomatosis, and giant cell arteritis. For these more serious disorders, the starting dose of glucocorticoids should be sufficient to suppress the disease rapidly and minimize resultant tissue damage. Initially prednisone (1 mg/kg per day in divided doses) often is used, generally followed by consolidation to a single daily dose, with subsequent tapering to a minimal effective dose as determined by clinical variables. There is controversy regarding the role of glucocorticoids in rheumatoid arthritis, particularly because of the serious and debilitating side effects associated with chronic use. Some authorities recommend glucocorticoids only as temporizing agents for progressive disease that fails to respond to first-line treatments such as physiotherapy and nonsteroidal antiinflammatory agents. In this case, glucocorticoids provide relief until other, slower-acting antirheumatic drugs, such as methotrexate or gold, take effect. A typical starting dose is 5 to 10 mg of prednisone per day. In the setting of an acute exacerbation, higher doses of glucocorticoids may be employed (typically 20 to 40 mg/day of prednisone or equivalent), with rapid taper thereafter. Complete relief of symptoms is not sought, and the symptomatic effect of small reductions in dose (decreases of perhaps 1 mg/day of prednisone every 2 to 3 weeks) should be tested frequently, while concurrent therapy with other measures is continued, to maintain the lowest possible prednisone dose. Alternatively, patients with major symptomatology confined to one or a few joints may be treated with intraarticular steroid injections. Depending on joint size, typical doses are 5 to 20 mg of triamcinolone acetonide or its equivalent. In noninflammatory degenerative joint diseases (e.g., osteoarthritis) or in a variety of regional pain syndromes (e.g., tendonitis or bursitis), glucocorticoids may be administered by local injection for the treatment of episodic disease flare-up. It is important to minimize the frequency of local steroid administration whenever possible. In the case of repeated intraarticular injection of steroids, there is a significant incidence of painless joint destruction, resembling Charcot's arthropathy. It is recommended that intraarticular injections be performed with intervals of at least three months to minimize complications. Glucocorticoids are an important component of treatment for most of the vasculitic syndromes, often in conjunction with other immunosuppressive agents such as cyclophosphamide. Caution should be exercised in the use of glucocorticoids in some forms of vasculitis (e.g., polyarteritis nodosa), where underlying infections with hepatitis viruses may play a pathogenetic role. Although glucocorticoids are indicated in these cases, there is at least a theoretical consideration that glucocorticoids may complicate the course of the viral infection by suppressing the immune system. The shorter-acting glucocorticoids, such as prednisone and methylprednisolone, are preferred over longer-acting steroids, such as dexamethasone, to facilitate drug tapering and/or conversion to alternate-day treatment regimens. Guidelines for treatment of the major vasculitic syndromes have been proposed by Weisman and Weinblatt (1995). Renal Diseases The utility of glucocorticoids in renal disease also has been the subject of considerable debate. Patients with nephrotic syndrome secondary to minimal change disease generally respond well to

steroid therapy, and glucocorticoids are now accepted uniformly as first-line treatment in both adults and children. Initial daily doses of prednisone are 1 to 2 mg/kg for 6 weeks, followed by a gradual tapering of the dose over 6 to 8 weeks, although some nephrologists advocate alternate-day therapy. Objective evidence of response, such as diminished proteinuria, is seen within 2 to 3 weeks in 85% of patients, and more than 95% of patients will have remission within three months. Cessation of steroid therapy frequently is complicated by disease relapse, as manifested by recurrent proteinuria. Patients who relapse repeatedly are termed steroid-resistant and often are treated with other immunosuppressive drugs such as azathioprine or cyclophosphamide. Patients with renal disease secondary to systemic lupus erythematosus also are generally given a therapeutic trial of glucocorticoids. Studies with other forms of renal disease, such as membranous and membranoproliferative glomerulonephritis and focal sclerosis, have provided conflicting data on the role of glucocorticoids. In clinical practice, patients with these disorders often are given a therapeutic trial of glucocorticoids with careful monitoring of laboratory indices of response. In the case of membranous glomerulonephritis, many nephrologists recommend a trial of alternate-day glucocorticoids for 8 to 10 weeks (e.g., prednisone, 120 mg every other day), followed by a 1- to 2month period of tapering. Allergic Disease It must be emphasized that the onset of action of glucocorticoids in allergic diseases is delayed, and patients with severe allergic reactions such as anaphylaxis require immediate therapy with epinephrine: for adults, 0.5 ml of a 1:1000 solution intramuscularly or subcutaneously (repeated as often as every 15 minutes for up to three additional doses if necessary). The manifestations of allergic diseases of limited durationsuch as hay fever, serum sickness, urticaria, contact dermatitis, drug reactions, bee stings, and angioneurotic edemacan be suppressed by adequate doses of glucocorticoids given as supplements to the primary therapy. In severe disease, intravenous glucocorticoids (methylprednisolone 125 mg intravenously every 6 hours, or equivalent) are appropriate. In less severe disease, antihistamines are the drugs of first choice. In allergic rhinitis, intranasal steroids also may provide symptomatic relief. Bronchial Asthma Corticosteroids frequently are used in bronchial asthma (see Chapter 28: Drugs Used in the Treatment of Asthma). They sometimes are employed in chronic obstructive pulmonary disease (COPD), particularly when there is some evidence of reversible obstructive disease. Data supporting the efficacy of corticosteroids are much more convincing for bronchial asthma than for COPD. The increased use of corticosteroids in asthma reflects an increased appreciation of the role of inflammation in the immunopathogenesis of this disorder (Goldstein et al., 1994). In severe asthmatic attacks requiring hospitalization, aggressive treatment with parenteral glucocorticoids is considered essential even though their onset of action is delayed for 6 to 12 hours. Intravenous administration of 60 to 120 mg of methylprednisolone (or equivalent) every 6 hours is used initially, followed by daily oral doses of prednisone (40 to 60 mg) as the acute attack resolves. The dose then is tapered gradually, with withdrawal planned for 10 days to 2 weeks after initiation of steroid therapy. In general, patients subsequently can be managed on their prior medical regimen. Less severe, acute exacerbations of asthma (as well as acute flares of COPD) often are treated with brief courses of oral glucocorticoids. In adult patients, 40 to 60 mg of prednisone is administered daily for five days; an additional week of therapy at lower doses also may be required. Upon resolution of the acute exacerbation, the glucocorticoids generally can be rapidly tapered without

significant deleterious effects. Any suppression of adrenal function usually dissipates within 1 to 2 weeks. In the treatment of severe chronic bronchial asthma (or, less frequently, COPD) that is not controlled by other measures, the long-term administration of glucocorticoids may be necessary. As with other long-term uses of these agents, the lowest effective dose is used, and care must be exercised when withdrawal is attempted. Given the risks of long-term treatment with glucocorticoids, it is especially important to document objective evidence of a response (e.g., an improvement in pulmonary function tests). In addition, these risks dictate that long-term glucocorticoid therapy be reserved for those patients who have failed to respond to adequate regimens of other medications (see Chapter 28: Drugs Used in the Treatment of Asthma). In many patients, the use of inhaled steroids (most frequently beclomethasone dipropionate, triamcinolone acetonide, flunisolide, or budesonide) either can reduce the need for oral corticosteroids or replace them entirely (see Barnes, 1995). In addition, many physicians recommend inhaled glucocorticoids over previously recommended oral theophylline in the treatment of children with moderately severe asthma, in part because of the behavioral toxicity associated with chronic theophylline administration (see Chapter 28: Drugs Used in the Treatment of Asthma). When used as recommended, inhaled glucocorticoids are effective in reducing bronchial hyperreactivity with less suppression of adrenal function than with oral glucocorticoids. Dysphonia or oropharyngeal candidiasis may develop, but the incidence of such side effects can be reduced substantially by maneuvers that reduce drug deposition in the oral cavity, such as spacers and mouth rinsing. The evolving status of glucocorticoids in asthma therapy is discussed in detail in Chapter 28: Drugs Used in the Treatment of Asthma. Infectious Diseases Although it would seem paradoxical to use immunosuppressive glucocorticoids in infectious diseases, there are a limited number of settings where they are indicated in the therapy of specific infectious pathogens (McGowan et al., 1992). One dramatic example of such beneficial effects is seen in AIDS patients with Pneumocystis carinii pneumonia and moderate to severe hypoxia; addition of glucocorticoids to the antibiotic regimen increases oxygenation and lowers the incidence of respiratory failure and mortality. Similarly, glucocorticoids clearly decrease the incidence of long-term neurological impairment associated with Haemophilus influenzae type b meningitis in infants and children two months of age or older. Ocular Diseases Ocular pharmacology, including some consideration of the use of glucocorticoids, is discussed in Chapter 66: Ocular Pharmacology. Glucocorticoids frequently are used to suppress inflammation in the eye and can lead to the preservation of sight when used properly. They are administered topically for diseases of the outer eye and anterior segment and attain therapeutic concentrations in the aqueous humor following instillation into the conjunctival cul-de-sac. For diseases of the posterior segment, systemic administration is required. It is generally recommended that ocular use of glucocorticoids be under the supervision of an ophthalmologist. A typical prescription is 0.1%dexamethasonesodium phosphate solution (ophthalmic), 2 drops in the conjunctival sac every 4 hours while awake, and 0.05% dexamethasone sodium phosphate ointment (ophthalmic) at bedtime. For inflammations of the posterior segment, systemic therapy is required, and typical doses are 30 mg of prednisone or equivalent per day, administered orally in divided doses. Topical glucocorticoid therapy frequently increases intraocular pressure in normal eyes and

exacerbates intraocular hypertension in patients with antecedent glaucoma. The glaucoma is not always reversible on cessation of glucocorticoid therapy. Intraocular pressure should be monitored when glucocorticoids are applied to the eye for more than 2 weeks. Topical administration of glucocorticoids to patients with bacterial, viral, or fungal conjunctivitis can mask evidence of progression of the infection until sight is irreversibly lost. Glucocorticoids are contraindicated in herpes simplex keratitis, because progression of the disease may lead to irreversible clouding of the cornea. Topical steroids should not be used in treating mechanical lacerations and abrasions of the eye because they delay healing and promote the development and spread of infection. Skin Diseases Glucocorticoids are remarkably efficacious in the treatment of a wide variety of inflammatory dermatoses. As a result, a large number of different preparations and concentrations of topical glucocorticoids of varying potencies are available. A typical regimen for an eczematous eruption is 1%hydrocortisone ointment applied locally twice daily. Effectiveness is enhanced by application of the topical steroid under an occlusive film, such as plastic wrap; unfortunately, the risk of systemic absorption also is increased by occlusive dressings, and this can be a significant problem when the more potent glucocorticoids are applied to inflamed skin. Glucocorticoids are administered systemically for severe episodes of acute dermatologic disorders and for exacerbations of chronic disorders. The dose in these settings is usually 40 mg/day of prednisone. Systemic steroid administration can be lifesaving in pemphigus, which may require daily doses of up to 120 mg of prednisone. Further discussion of the treatment of skin diseases is given in Chapter 65: Dermatological Pharmacology. Gastrointestinal Diseases Glucocorticoid therapy is indicated in selected patients with inflammatory bowel disease (chronic ulcerative colitis and Crohn's disease; see Chapter 39: Agents Used for Diarrhea, Constipation, and Inflammatory Bowel Disease; Agents Used for Biliary and Pancreatic Disease). Patients who fail to respond to more conservative management (i.e., rest, diet, and sulfasalazine) may benefit from glucocorticoids; steroids are most useful for acute exacerbations (Stein and Hanauer, 1999). In mild ulcerative colitis, hydrocortisone (100 mg) can be administered as a retention enema with beneficial effects. In more severe acute exacerbations, oral prednisone (10 to 30 mg/day) frequently is employed. For severely ill patientswith fever, anorexia, anemia, and impaired nutritional status larger doses should be used (60 to 120 mg prednisone per day). Major complications of ulcerative colitis or Crohn's disease may occur despite glucocorticoid therapy, and glucocorticoids may mask signs and symptoms of complications such as intestinal perforation and peritonitis. Budesonide, a highly potent synthetic glucocorticoid that is inactivated by first-pass hepatic metabolism, has diminished systemic side effects commonly associated with glucocorticoids. Oral administration of budesonide in delayed-release capsules (9 mg/day) facilitates drug delivery to the ileum and ascending colon (Greenberg et al., 1994); the drug also has been used as a retention enema in the treatment of ulcerative colitis. These dosage forms, however, are not yet available in the United States. Currently, budesonide is not approved for treatment of inflammatory bowel disease in the United States and is available only as an antiinflammatory agent for inhalation therapy. Hepatic Diseases

The use of corticosteroids in hepatic disease has been highly controversial. Glucocorticoids are clearly of benefit in autoimmune, chronic active hepatitis, where as many as 80% of patients show histological remission when treated with prednisone (40 to 60 mg daily initially, with tapering to a maintenance dose of 7.5 to 10 mg daily after serum transaminase levels fall). The role of corticosteroids in alcoholic liver disease is not fully defined; the most recent studies, including meta-analysis of previously published reports, suggest a beneficial role of prednisolone (40 mg/day for 4 weeks) in patients with severe disease indicators (e.g., hepatic encephalopathy) without active gastrointestinal bleeding (McCullough and O'Connor, 1998). Further studies are needed to confirm or refute the role of steroids in this setting. In the setting of severe hepatic disease, prednisolone should be used instead of prednisone, which requires hepatic conversion to be active. Malignancies Glucocorticoids are used in the chemotherapy of acute lymphocytic leukemia and lymphomas because of their antilymphocytic effects. Most commonly, glucocorticoids are one component of combination chemotherapy administered under scheduled protocols. Further discussion of the chemotherapy of malignant disease is given in Chapter 52: Antineoplastic Agents. Glucocorticoids once were frequently employed in the setting of hypercalcemia of malignancy, but they have been largely supplanted by more effective agents such as the bisphosphonates. Cerebral Edema Corticosteroids are of value in the reduction or prevention of cerebral edema associated with parasites and with neoplasms, especially those that are metastatic. Although frequently used for the treatment of cerebral edema caused by trauma or cerebrovascular accidents, controlled clinical trials do not support their use in these settings. Miscellaneous Diseases and Conditions Sarcoidosis Sarcoidosis is treated with corticosteroids (approximately 1 mg/kg per day of prednisone, or equivalent dose of alternative steroids) to induce remission. Maintenance doses, which often are required for long periods of time, may be as low as 10 mg/day of prednisone. These patients, like all patients who require chronic glucocorticoid therapy at doses exceeding the normal daily production rate, are at increased risk for secondary tuberculosis; therefore, patients with a positive tuberculin reaction or other evidence of tuberculosis should receive prophylactic antituberculosis therapy. Thrombocytopenia In thrombocytopenia, prednisone (0.5 mg/kg) is used to decrease the bleeding tendency. In more severe cases, and for initiation of treatment of idiopathic thrombocytopenia, daily doses of prednisone (1 to 1.5 mg/kg) are employed. Patients with refractory idiopathic thrombocytopenia may respond to pulsed, high-dose glucocorticoid therapy. Autoimmune Destruction of Erythrocytes Patients with autoimmune destruction of erythrocytes (i.e., hemolytic anemia with a positive Coombs test) are treated with prednisone (1 mg/kg per day). In the setting of severe hemolysis, higher doses may be used, with tapering as the anemia improves. Small maintenance doses may be

required for several months in patients who respond. Organ Transplantation In organ transplantation, high doses of prednisone (50 to 100 mg) are given at the time of transplant surgery, in conjunction with other immunosuppressive agents, and most patients are kept on a maintenance regimen that includes lower doses of glucocorticoids (see Chapter 53: Immunomodulators: Immunosuppressive Agents, Tolerogens, and Immunostimulants). Spinal Cord Injury Multicenter trials have shown significant decreases in neurological defects in patients with acute spinal cord injury treated within 8 hours of injury with large doses of methylprednisolone [30 mg/kg initially followed by an infusion of 5.4 mg/kg per hour for 23 hours (Bracken et al., 1997)]. The ability of corticosteroids at these high doses to decrease neurological injury may reflect inhibition of free radicalmediated cellular injury, as occurs following ischemia and reperfusion. Diagnostic Applications of Adrenocortical Steriods In addition to their therapeutic uses, glucocorticoids also are used for diagnostic purposes. To determine if patients with clinical manifestations suggestive of hypercortisolism have biochemical evidence of increased cortisol biosynthesis, an overnight dexamethasone test has been devised. Patients are given 1 mg of dexamethasone orally at 11 P.M., and cortisol is measured at 8 A.M. the following morning. Suppression of plasma cortisol to less than 5 g/dl suggests strongly that the patient does not have Cushing's syndrome. The formal dexamethasone suppression test is used in the differential diagnosis of biochemically documented Cushing's syndrome. Following determination of baseline cortisol levels for 48 hours, dexamethasone (0.5 mg every 6 hours) is administered orally for 48 hours. This dose markedly suppresses cortisol levels in normal subjects, including those who have nonspecific elevations of cortisol due to obesity or stress, but does not suppress levels in patients with Cushing's syndrome. In the high-dose phase of the test, dexamethasone is administered orally at 2 mg every 6 hours for 48 hours. Patients with pituitary-dependent Cushing's syndrome (i.e., Cushing's disease) generally respond with decreased cortisol levels. In contrast, patients with ectopic production of ACTH or with adrenocortical tumors generally do not exhibit decreased cortisol levels. Despite these generalities, dexamethasone may suppress cortisol levels in some patients with ectopic ACTH production, particularly with tumors such as bronchial carcinoids. Inhibitors of the Biosynthesis and Action of Adrenocortical Steroids Five pharmacological agents have been useful as inhibitors of adrenocortical secretion. Mitotane (o,p'-DDD), an adrenocorticolytic agent, is discussed in Chapter 52: Antineoplastic Agents. The other inhibitors of steroid hormone biosynthesismetyrapone, aminoglutethimide, ketoconazole, and trilostaneare discussed here. Metyrapone, aminoglutethimide, and ketoconazole act by inhibiting cytochrome P450 enzymes in-volved in adrenocorticosteroid biosynthesis. Differential selectivity of these agents for the different steroid hydroxylases provides some degree of specificity to their actions. Trilostane is a competitive inhibitor of the conversion of pregnenolone to progesterone, a reaction catalyzed by 3 -hydroxysteroid dehydrogenase. In addition, agents that act as glucocorticoid receptor antagonists (antiglucocorticoids) are discussed here (mineralocorticoid antagonists are discussed in Chapter 29: Diuretics). All of these agents pose the common risk of precipitating acute adrenal insufficiency; thus, they must be used in appropriate doses, and the

status of the patient's HPA axis must be carefully monitored. Aminoglutethimide Aminoglutethimide ( -ethyl-p-aminophenyl-glutarimide; CYTADREN) primarily inhibits P450 scc , which catalyzes the initial and rate-limiting step in the biosynthesis of all physiological steroids. As a result, the production of all classes of steroid hormones is impaired. Aminoglutethimide also inhibits P45011 and the enzyme aromatase, which converts androgens to estrogens. Aminoglutethimide has been used to decrease hypersecretion of cortisol in patients with Cushing's syndrome secondary to autonomous adrenal tumors and hypersecretion associated with ectopic production of ACTH. Because of its actions to inhibit aromatase, aminoglutethimide also has been evaluated as a therapeutic agent for the treatment of hormonally responsive tumors such as prostate and breast cancer (see Chapter 52: Antineoplastic Agents). Dose-dependent gastrointestinal and neurological side effects are relatively common, as is a transient, maculopapular rash. The usual dose is 250 mg every 6 hours, with gradual increases of 250 mg per day at 1- to 2-week intervals until the desired biochemical effect is achieved, side effects prohibit further increases, or a daily dose of 2 g is reached. Since aminoglutethimide can cause frank adrenal insufficiency, sometimes associated with signs of mineralocorticoid deficiency, glucocorticoid replacement therapy is necessary, and mineralocorticoid supplements may be indicated. Because aminoglutethimide accelerates the metabolism of dexamethasone, this steroid should not be used for glucocorticoid replacement in patients receiving aminoglutethimide. Ketoconazole Ketoconazole (NIZORAL) is an antifungal agent, and this remains its most important clinical role (see Chapter 49: Antimicrobial Agents: Antifungal Agents). In doses higher than those employed in antifungal therapy, it is an effective inhibitor of adrenal and gonadal steroidogenesis, primarily because of its inhibition of the C1720 lyase activity of P45017 . At even higher doses, ketoconazole also inhibits P450scc, effectively blocking steroidogenesis in all primary steroidogenic tissues. Ketoconazole has been reported to be the most effective inhibitor of steroid hormone biosynthesis in patients with Cushing's disease (see Sonino and Boscaro, 1999, for a review of the medical management of Cushing's disease). In most cases, a dosage regimen of 600 to 800 mg/day (in two divided doses) is required, and some patients may require up to 1200 mg/day given in two to three doses. Side effects include hepatic dysfunction, which ranges from asymptomatic elevations of transaminase levels to severe hepatic injury. The potential of ketoconazole to interact with cytochrome P450 enzymes can lead to drug interactions of serious consequence. For example, rare but potentially life-threatening interactions were reported between ketoconazole and nonsedating antihistamines (e.g., terfenadine, astemizole) due to inhibition of hepatic CYP3A4 and decreased metabolism of these antihistamines. Terfenadine and astemizole have since been withdrawn from the market. Further studies are needed to define the precise role of ketoconazole in the medical management of patients with excessive steroid hormonal production, and ketoconazole currently is not approved by the FDA for this indication. Metyrapone Metyrapone (METOPIRONE) is a relatively selective inhibitor of 11 -hydroxylase, which converts 11-deoxycortisol to cortisol in the terminal reaction of the glucocorticoid biosynthetic pathway. Because of this inhibition, the biosynthesis of cortisol is markedly impaired, and the levels of steroid precursors (e.g., 11-deoxycortisol) are markedly increased. Although the biosynthesis of aldosterone also is impaired, the elevated levels of 11-deoxycortisol sustain mineralocorticoiddependent functions. In a diagnostic test of the entire HPA axis, metyrapone (30 mg/kg, maximum

dose of 3 g) is administered orally with a snack at midnight, and plasma cortisol and 11deoxycortisol are measured at 8 A.M. the next morning. A plasma cortisol of less than 8 g/dl validates adequate inhibition of 11 -hydroxylase; in this setting, an 11-deoxycortisol level of less than 7 g/dl is highly suggestive of impaired hypothalamic-pituitary-adrenal function. An abnormal response does not identify the site of the defecteither hypothalamic CRH release, ACTH production, or adrenal biosynthetic capacity could be impaired. Some authorities avoid overnight metyrapone testing in outpatients thought to have a reasonable probability of impaired HPA function, as there is some risk of precipitating acute adrenal insufficiency. Others believe that the ability to assess the entire HPA axis with a relatively easy test justifies the use of metyrapone testing in outpatients. Metryapone also is used to diagnose patients with Cushing's syndrome who respond equivocally to the high-dose dexamethasone suppression test. Those with pituitary-dependent Cushing's syndrome exhibit a normal response, whereas those patients with ectopic secretion of ACTH exhibit no changes in ACTH or 11-deoxycortisol levels. Therapeutically, metyrapone has been used to treat the hypercorticism resulting from either adrenal neoplasms or tumors producing ACTH ectopically. Maximal suppression of steroidogenesis requires doses of 4 g/day. More frequently, metyrapone is used as adjunctive therapy in patients who have received pituitary irradiation or in combination with other agents that inhibit steroidogenesis. In this setting, a dose of 500 to 750 mg three or four times daily is employed. The use of metyrapone in the treatment of Cushing's syndrome secondary to pituitary hypersecretion of ACTH is more controversial. Chronic administration of metyrapone can cause hirsutism, which results from increased synthesis of adrenal androgens prior to the enzymatic block, and hypertension, which results from elevated levels of 11-deoxycortisol. Other side effects include nausea, headache, sedation, and rash. Antiglucocorticoids The progesterone receptor antagonist mifepristone [RU-486; (11 -4-dimethylaminophenyl)-17 hydroxy-7 -(propyl-1-ynyl)estra-4,9-dien-3-one] has received considerable attention because of its use as an antiprogestagen that can terminate early pregnancy (see Chapter 58: Estrogens and Progestins). At higher doses, however, mifepristone also inhibits the glucocorticoid receptor, blocking feedback regulation of the HPA axis and secondarily increasing endogenous ACTH and cortisol levels. Because of its ability to inhibit glucocorticoid action, mifepristone also has been studied as a potential therapeutic agent in patients with hypercorticism. Prospectus Many of the current clinical uses of corticosteroids are based on empirical approaches, rather than on detailed understanding of the mechanisms by which these drugs act. Our understanding of the pathways of corticosteroid actions within target cells has increased dramatically within recent years; these advances, while not yet directly translatable to clinical medicine, hold promise for the development of new therapeutic approaches with greater selectivity and less toxicity. We now recognize the importance of glucocorticoid receptor coactivators and corepressors in glucocorticoid action and have gained new insights into the interactions of glucocorticoids with other signal transduction pathways that may provide new avenues to manipulate glucocorticoid signaling. The potent immunosuppressive agents cyclosporine and tacrolimus are examples of drugs that at least partly intersect with the AP-1 and NF- B transcriptional activation pathways in lymphocytes, providing a rationale for their common immunosuppressive actions. These agents have markedly improved success rates in organ transplantation, permitting the use of lower doses of

glucocorticoids with diminished long-term complications (see Chapter 53: Immunomodulators: Immunosuppressive Agents, Tolerogens, and Immunostimulants). Novel drugs that impinge upon the lymphocyte activation cascade also may provide alternative antiinflammatory and immunosuppressive drugs with fewer side effects. Finally, new insights from the crystal structures of the ligand-binding regions of the steroid hormone receptors may facilitate a more rational approach to developing novel compounds with glucocorticoid activity. Inasmuch as the antiinflammatory and metabolic effects of glucocorticoids are mediated by the same receptor, efforts to separate desired therapeutic effects from undesirable side effects largely have been unsuccessful to date. Nevertheless, the development of selective estrogen receptor modulators that act as receptor agonists in certain target tissues and as antagonists in others provides hope that glucocorticoid analogs may be developed with more favorable therapeutic profiles. Indeed, several dissociated glucocorticoids have been identified that act selectively as antiinflammatory agents (Vayssiere et al., 1997; Vanden Berghe et al., 1999). These agents apparently exert their antiinflammatory effects by preferentially promoting the direct interaction of the glucocorticoid receptor with transcription factors such as AP-1 and NF- B, while having little effect on the metabolic and other activities mediated by glucocorticoid receptor dimers at specific GREs (e.g., Figure 605). It remains to be seen if these experimental agents develop into clinically useful drugs. For further discussion of diseases of the adrenal cortex see Chapter 321 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 61. Insulin, Oral Hypoglycemic Agents, and the Pharmacology of the Endocrine Pancreas
Overview This chapter provides background on the pharmacological actions of insulin, glucagon, somatostatin, and hypoglycemic agents. The discovery of insulin in 1921 allowed the previously fatal disorder of insulin-dependent diabetes mellitus (type 1 diabetes mellitus) to be treated and represents a landmark in medical history. In the first part of this chapter, the diverse physiological functions of insulin are described at the cellular and whole-body levels. This section establishes the role of insulin in the treatment of diabetes mellitus. The next section describes the pharmacodynamics and pharmacokinetics of exogenously administered insulin and highlights the benefits of intensive insulin therapy in limiting long-term tissue complications of diabetes. The chapter continues with descriptions of the pharmacology of orally effective agents. These drugs have an increasingly important role in the treatment of non-insulin-dependent or type 2 diabetes mellitus, the most common form of diabetes. The final part of the chapter describes the physiology and pharmacology of glucagon and somatostatin, with emphasis on the expanding use of somatostatin analogs in clinical medicine. Insulin History Few events in the history of medicine are more dramatic than the discovery of insulin. Although the

discovery is appropriately attributed to Banting and Best, several other investigators and collaborators provided important observations and techniques that made it possible. In 1869, a German medical student, Paul Langerhans, noted that the pancreas contains two distinct groups of cellsthe acinar cells, which secrete digestive enzymes, and cells that are clustered in islands, or islets, which he suggested served a second function. Direct evidence for this function came in 1889, when Oskar Minkowski and Joseph von Mering showed that pancreatectomized dogs exhibit a syndrome similar to diabetes mellitus in human beings (see Minkowski, 1989). There were numerous attempts to extract the pancreatic substance responsible for regulating blood glucose. In the early 1900s, Gurg Ludwig Zuelzer, an internist in Berlin, attempted to treat a dying diabetic patient with extracts of pancreas. Although the patient improved temporarily, he sank back into coma and died when the supply of extract was exhausted. E. L. Scott, a student at the University of Chicago, made another early attempt to isolate an active principle in 1911. Using alcoholic extracts of the pancreas (not so different from those eventually used by Banting and Best), Scott treated several diabetic dogs with encouraging results; however, he lacked clear measures of control of blood glucose concentrations, and his professor considered the experiments inconclusive at best. Between 1916 and 1920, the Romanian physiologist Nicolas Paulesco conducted a series of experiments in which he found that injections of pancreatic extracts reduced urinary sugar and ketones in diabetic dogs. Although he published the results of his experiments, their significance was fully appreciated only many years later. Unaware of much of this previous work, in 1921 Frederick G. Banting, a young Canadian surgeon, convinced a professor of physiology in Toronto, J. J. R. Macleod, to allow him access to a laboratory to search for the antidiabetic principle of the pancreas. Banting assumed that the islet tissues secreted insulin but that the hormone was destroyed by proteolytic digestion prior to or during extraction. Together with Charles H. Best, a fourth-year medical student, he attempted to overcome the problem by tying the pancreatic ducts. The acinar tissue degenerated, leaving the islets undisturbed; the remaining tissue was then extracted with ethanol and acid. Banting and Best thus obtained a pancreatic extract that was effective in decreasing the concentration of blood glucose in diabetic dogs. The first patient to receive the active extracts prepared by Banting and Best was Leonard Thompson, aged 14 (Banting et al., 1922). He appeared at the Toronto General Hospital with a blood glucose level of 500 mg/dl (28 mM), and he was excreting 3 to 5 liters of urine per day. Despite rigid control of diet (450 kcal per day), he continued to excrete large quantities of glucose, and, without insulin, the most likely outcome was death after a few months. The administration of Banting and Best's extracts induced a reduction in the plasma concentration and urinary excretion of glucose. Daily injections were then begun, and there was immediate improvement. The excretion of glucose was reduced from over 100 to as little as 7.5 g per day. Furthermore, "the boy became brighter, looked better and said he felt stronger." Thus, replacement therapy with the newly discovered hormone, insulin, had interrupted what was clearly an otherwise fatal metabolic disorder (Banting et al., 1922). Banting and Best faced many trials and tribulations during the subsequent year. It was difficult to obtain active extracts reproducibly. This led to a greater involvement of Macleod; Banting also sought help from J. B. Collip, a chemist with expertise in extraction and purification of epinephrine. Stable extracts eventually were obtained, and patients in many parts of North America soon were being treated with insulin from porcine and bovine sources. Now, as a result of recombinant DNA technology, human insulin is used for therapy. The Nobel Prize in Medicine or Physiology was awarded to Banting and Macleod with remarkable rapidity in 1923, and a furor over credit followed immediately. Banting announced that he would share half of his prize with Best; Macleod did the same with Collip. The early history of the

discovery of insulin has been reviewed by Bliss (1982). Chemistry Insulin was purified and crystallized by Abel within a few years of its discovery. The amino acid sequence of insulin was established by Sanger in 1960, and this led to the complete synthesis of the protein in 1963 and the elucidation of its three-dimensional structure by Hodgkin and coworkers in 1972. Insulin was the first hormone for which a radioimmunoassay was developed (Yalow, 1978). The cells of pancreatic islets synthesize insulin from a single-chain precursor of 110 amino acids termed preproinsulin. After translocation through the membrane of the rough endoplasmic reticulum, the 24amino acid N-terminal signal peptide of preproinsulin is rapidly cleaved off to form proinsulin (see Figure 611). Here the molecule folds and the disulfide bonds are formed. On conversion of human proinsulin to insulin in the Golgi complex, four basic amino acids and the remaining connector or C peptide are removed by proteolysis. This gives rise to the two peptide chains (A and B) of the insulin molecule, which contains one intrasubunit and two intersubunit disulfide bonds. The A chain usually is composed of 21 amino acid residues, and the B chain has 30; the molecular mass is thus about 5734 daltons. Although the amino acid sequence of insulin has been highly conserved in evolution, there are significant variations that account for differences in both biological potency and immunogenicity (De Meyts, 1994). There is a single insulin gene and a single protein product in most species. However, rats and mice have two genes that encode insulin, and they synthesize two molecules that differ from each other by two amino acid residues in the B chain. Figure 611. Human Proinsulin and Its Conversion to Insulin. The amino acid sequence of human proinsulin is shown. By proteolytic cleavage, four basic amino acids (residues 31, 32, 64, 65) and the connecting peptide are removed, converting proinsulin to insulin. The sites of action of the endopeptidases PC2 and PC3 are shown.

The crystal structure, now resolved to 1.5 , reveals that the two chains of insulin form a highly ordered structure with several a-helical regions in both the A and B chains. The isolated chains of insulin are inactive. In solution, insulin can exist as a monomer, dimer, or hexamer. Two molecules of Zn2+ are coordinated in the hexamer, and this form of insulin presumably is stored in the granules of the pancreatic cell. It is believed that Zn 2+ has a functional role in the formation of crystals and that crystallization facilitates the conversion of proinsulin to insulin, as well as storage of the hormone. Traditional insulin is hexameric in most of the highly concentrated preparations used for therapy. When the hormone is absorbed and the concentration falls to physiological levels (nanomolar), the hormone dissociates into monomers, and the monomer is most likely the biologically active form of insulin. Monomeric insulin is now available for therapy. A great deal of information about the structureactivity relationship of insulin has been obtained by study of insulins purified from a wide variety of species and by modification of the molecule. A dozen invariant residues in the A and B chains form a surface that interacts with the insulin receptor (Figure 612). These residuesGlyA1, GluA4, GlnA5, TyrA19, AsnA21, ValB12, TyrB16, GlyB23 , Phe B24, PheB25, and TyrB26 overlap with domains that also are involved in insulin dimerization (de Meyts, 1994). The Leu A13 and LeuB17 residues may form part of a second binding surface (de Meyts, 1994). Insulin binds to surfaces located at the N-terminal and C-terminal regions of the subunit of the receptor. A cysteine-rich region in the receptor chain appears to be involved in the binding of insulin. In most cases, there is a very close correlation between the affinity of insulin for the insulin receptor and its potency for eliciting effects on glucose metabolism. Compared with human insulin, bovine and porcine insulins are equipotent; South American guinea pig insulin is much less potent, while certain avian insulins are significantly more so. Figure 612. Model of the Three-Dimensional Structure of Insulin. The shaded area indicates the receptor-binding face of the insulin molecule. (See Pullen et al., 1976).

Insulin is a member of a family of related peptides termed insulin-like growth factors (IGFs). The two IGFs (IGF-1 and IGF-2) have molecular masses of about 7500 daltons and structures that are homologous to that of proinsulin (Cohick and Clemmons, 1993). However, the short equivalents of the C peptide in proinsulin are not removed from the IGFs. In contrast to insulin, the IGFs are produced in many tissues, and they may serve a more important function in regulation of growth than in regulation of metabolism. These peptides, particularly IGF-1, are the presumed mediators of the action of growth hormone, and they originally were called somatomedins. The uterine hormone

relaxin also may be a distant relative of this family of polypeptides. The receptors for insulin and IGF-1 are also closely related (Duronio and Jacobs, 1988). Thus, insulin can bind to the receptor for IGF-1 with low affinity and vice versa. The growth-promoting actions of insulin appear to be mediated in part through the IGF-1 receptor, and there may be discordance between the metabolic potency of an insulin analog and its ability to promote growth. For example, proinsulin has only 2% of the metabolic potency of insulin in vitro, but it is half as potent as insulin as a stimulator of mitogenesis (King and Kahn, 1981). This fact could be important in selecting insulins for therapy, since the mitogenic activity of insulin may contribute to an increased risk of atherosclerosis. Synthesis, Secretion, Distribution, and Degradation of Insulin Insulin Production The molecular and cellular events involved in the synthesis, storage, and secretion of insulin by the cell and the ultimate degradation of the hormone by its target tissues have been studied in great detail and have served as a model for study of other cell types in the pancreatic islet (Orci, 1986). The islet of Langerhans is composed of four types of cells, each of which synthesizes and secretes a distinct polypeptide hormone: insulin in the (B) cell, glucagon in the (A) cell, somatostatin in the (D) cell, and pancreatic polypeptide in the PP or F cell. The cells make up 60% to 80% of the islet and form its central core. The , , and F cells form a discontinuous mantle, one to three cells thick, around this core. The cells in the islet are connected by tight junctions that allow small molecules to pass and make possible coordinated control of groups of cells (Orci, 1986). Arterioles enter the islets and branch into a glomerular-like capillary mass in the -cell core. Capillaries then pass to the rim of the islet and coalesce into collecting venules (Bonner-Weir and Orci, 1982). Blood flows in the islet from the cells to and cells (Samols et al., 1986). Thus, the cell is the primary glucose sensor for the islet, and the other cell types are presumably exposed to particularly high concentrations of insulin. As noted above, insulin is synthesized as a single-chain precursor in which the A and B chains are connected by the C peptide. The initial translation product, preproinsulin, contains a sequence of 24 primarily hydrophobic amino acid residues attached to the amino terminus of the B chain. This signal sequence is required for the association and penetration of nascent preproinsulin into the lumen of the rough endoplasmic reticulum. This sequence is rapidly cleaved, and proinsulin is then transported in small vesicles to the Golgi complex. Here, proinsulin is packaged into secretory granules along with the enzyme(s) responsible for its conversion to insulin (Orci, 1986). The conversion of proinsulin to insulin begins in the Golgi complex, continues in the secretory granules, and is nearly complete at the time of secretion. Thus, equimolar amounts of C peptide and insulin are released into the circulation. The C peptide has no known biological function, but it can serve as a useful index of insulin secretion (Polonsky and Rubenstein, 1986). Small quantities of proinsulin and des-31,32 proinsulin also are released from cells. This presumably reflects either exocytosis of granules in which the conversion of proinsulin to insulin is not complete or secretion by another pathway. Since the half-life of proinsulin in the circulation is much longer than that of insulin, up to 20% of immunoreactive insulin in plasma is, in reality, proinsulin and intermediates. Two distinct Ca2+-dependent endopeptidases, which are found in the islet cell granules and in other neuroendocrine cells, are responsible for the conversion of proinsulin to insulin. These endoproteases, PC2 and PC3, have catalytic domains related to that of subtilisin and cleave at

lysine-arginine or arginine-arginine sequences (Steiner et al., 1992). PC2 selectively cleaves at the C peptideA chain junction (see Figure 611). PC3 preferentially cleaves at the C peptideB chain junction but has some action at the A chain junction as well. Although there are at least two other members of the family of endoproteases (PC1 and furin), PC2 and PC3 appear to be the enzymes responsible for processing proinsulin to insulin. Regulation of Insulin Secretion Insulin secretion is a tightly regulated process, designed to provide stable concentrations of glucose in blood during both fasting and feeding. This regulation is achieved by the coordinated interplay of various nutrients, gastrointestinal hormones, pancreatic hormones, and autonomic neurotransmitters. Glucose, amino acids, fatty acids, and ketone bodies promote the secretion of insulin. The islets of Langerhans are richly innervated by both adrenergic and cholinergic nerves. Stimulation of 2-adrenergic receptors inhibits insulin secretion, whereas 2-adrenergic receptor agonists and vagal nerve stimulation enhance release. In general, any condition that activates the autonomic nervous system (such as hypoxia, hypothermia, surgery, or severe burns) suppresses the secretion of insulin by stimulation of 2-adrenergic receptors. Predictably, 2-adrenergic receptor antagonists increase basal concentrations of insulin in plasma, and 2-adrenergic receptor antagonists decrease them (Porte and Halter, 1981). Glucose is the principal stimulus to insulin secretion in human beings and is an essential permissive factor for the actions of many other secretagogues (Matschinsky, 1996). The sugar is more effective in provoking insulin secretion when taken orally than when administered intravenously. This is true because the ingestion of glucose (or food) induces the release of gastrointestinal hormones and stimulates vagal activity (Malaisse, 1986; Brelje and Sorenson, 1988). Several gastrointestinal hormones promote the secretion of insulin (Ebert and Creutzfeldt, 1987). The most potent of these are gastrointestinal inhibitory peptide and glucagon-like peptide-1. Insulin release also is stimulated by gastrin, secretin, cholecystokinin, vasoactive intestinal peptide, gastrin-releasing peptide, and enteroglucagon. When evoked by glucose, insulin secretion is biphasic: The first phase reaches a peak after 1 to 2 minutes and is short-lived, whereas the second phase has a delayed onset but a longer duration. The exact mechanism by which glucose stimulates insulin release is not fully understood, but its entry into the cell and metabolism is required (Matschinsky, 1996). Glucose enters the cell by facilitated transport, which is mediated by GLUT2, a specific subtype of glucose transporter (see below). The sugar is then phosphorylated by glucokinase. In contrast to other hexokinases, which have a wide tissue distribution, expression of glucokinase is primarily limited to cells and tissues involved in the regulation of glucose metabolism, such as the liver and pancreatic cells. Its relatively high K m (10 to 20 mM) gives it an important regulatory role at physiological concentrations of glucose. The capacity of sugars to undergo phosphorylation and subsequent glycolysis correlates closely with their ability to stimulate insulin release. This fact has led to the hypothesis that one or more glycolytic intermediates or enzyme cofactors is the actual stimulator of insulin secretion (Matschinsky, 1996). The role of glucokinase as the glucose sensor was solidified by the recent association of mutations of the glucokinase gene with a form of maturity-onset diabetes of the young (MODY2; see below), a relatively uncommon form of diabetes. These mutations, which compromise the ability of glucokinase to phosphorylate glucose, raise the threshold for glucose-stimulated insulin release (Gidh-Jain et al., 1993). Insulin secretion ultimately depends on the intracellular concentration of Ca2+ (Wolf et al., 1988). Glucose metabolism, initiated by glucokinase, results in a change in the ATP/ADP ratio. This

results in the inhibition of an ATP-sensitive K+ channel and depolarization of the cell. A compensatory activation of a voltage-dependent Ca 2+ channel results in the influx of Ca2+ into the cell. Ca2+ activates phospholipase A2 and phospholipase C, which results in the formation of arachidonic acid, inositol polyphosphates, and diacylglycerol. Inositol-1,4,5-trisphosphate mobilizes Ca2+ from an endoplasmic reticulumlike compartment, further elevating the cytosolic concentration of the cation. Intracellular Ca2+ acts as the insulin secretagogue. Elevation of free Ca2+ concentrations also occurs in response to stimulation of phospholipase C by acetylcholine and cholecystokinin and by hormones that increase intracellular concentrations of cyclic AMP (Ebert and Creutzfeldt, 1987). -Cell adenylyl cyclase, the enzyme that synthesizes cyclic AMP, is activated by glucagon, gastrointestinal inhibitory peptide, and glucagon-like peptide-1, and is inhibited by somatostatin and 2-adrenergic receptor agonists (Fleischer and Erlichman, 1989). Most of the nutrients and hormones that stimulate insulin secretion also enhance the biosynthesis of the hormone (Gold et al., 1982). Although there is a close correlation between the two processes, some factors affect one pathway but not the other. For example, lowering extracellular concentrations of Ca2+ inhibits secretion of insulin without affecting biosynthesis. There usually is a reciprocal relationship between the rates of secretion of insulin and glucagon from the pancreatic islet (Unger, 1985). This reciprocity reflects both the influence of insulin on the cell and the level of glucose and other substrates (see below). In addition, somatostatin, a third islet-cell hormone, can modulate the secretion of both hormones (see below). Glucagon stimulates the release of somatostatin, and the latter may suppress the secretion of insulin but is not a major physiological influence. Since the blood supply in the islet flows from the cell core to the and cells (Samols et al., 1986), insulin can act as a glucagon-release-inhibiting paracrine hormone, but somatostatin must pass through the circulation to reach the and cells. Thus, while insulin affects the secretion of glucagon and pancreatic polypeptide, the role of islet somatostatin is not clear. Distribution and Degradation of Insulin Insulin circulates in blood as the free monomer, and its volume of distribution approximates the volume of extracellular fluid. Under fasting conditions, the pancreas secretes about 40 g [1 unit (U)] of insulin per hour into the portal vein, to achieve a concentration of insulin in portal blood of 2 to 4 ng/ml (50 to 100 U/ml) and in the peripheral circulation of 0.5 ng/ml (12 U/ml) or about 0.1 nM. After ingestion of a meal, there is a rapid rise in the concentration of insulin in portal blood, followed by a parallel but smaller rise in the peripheral circulation. A goal of insulin therapy is to mimic this pattern, but this is difficult to achieve with subcutaneous injections. The half-life of insulin in plasma is about 5 to 6 minutes in normal subjects and patients with uncomplicated diabetes (Sodoyez et al., 1983). This value may be increased in diabetics who develop anti-insulin antibodies. The half-life of proinsulin is longer than that of insulin (about 17 minutes), and this protein usually accounts for about 10% of the immunoreactive "insulin" in plasma (Robbins et al., 1984). In patients with insulinoma, the percentage of proinsulin in the circulation usually is increased and may be as much as 80% of immunoreactive "insulin." Since proinsulin is only about 2% as potent as insulin, the biologically effective concentration of insulin is somewhat lower than estimated by immunoassay. C peptide is secreted in equimolar amounts with insulin; however, its molar concentration in plasma is higher because of its lower hepatic clearance and considerably longer half-life (about 30 minutes) (Robbins et al., 1984). C-peptide serves as a marker for acute insulin secretion.

Degradation of insulin occurs primarily in liver, kidney, and muscle (Duckworth, 1988). About 50% of the insulin that reaches the liver via the portal vein is destroyed and never reaches the general circulation. Insulin is filtered by the renal glomeruli and is reabsorbed by the tubules, which also degrade it. Severe impairment of renal function appears to affect the rate of disappearance of circulating insulin to a greater extent than does hepatic disease (Rabkin et al., 1984). Hepatic degradation of insulin operates near its maximal capacity and cannot compensate for diminished renal breakdown of the hormone. The oral administration of glucose appears to reduce hepatic extraction of insulin (Hanks et al., 1984). Peripheral tissues such as fat also inactivate insulin, but this is of less significance quantitatively. Proteolytic degradation of insulin in the liver occurs primarily after internalization of the hormone and its receptor and, to a lesser extent, at the cell surface (Berman et al., 1980). The primary pathway for internalization is receptor-mediated endocytosis. The complex of insulin and its receptor is internalized into small vesicles termed endosomes, where degradation is initiated (Duckworth, 1988). Some insulin also is delivered to lysosomes for degradation. The extent to which internalized insulin is degraded by the cell varies considerably with the cell type. In hepatocytes, over 50% of the internalized insulin is degraded, whereas most internalized insulin is released intact from endothelial cells. In the latter case, this finding appears to be related to the role of these cells in transcytosis of insulin molecules from the intravascular to the extracellular space (King and Johnson, 1985). Transcytosis has an important role in the delivery of insulin to its target cells in tissues where endothelial cells form tight junctions, including skeletal muscle and adipose tissue. Several enzymes have been implicated in insulin degradation. The primary insulin-degrading enzyme is a thiol metalloproteinase. It is primarily localized in hepatocytes (Shii and Roth, 1986), but immunologically related molecules also have been found in muscle, kidney, and brain (Duckworth, 1988). Most insulin-degrading enzyme activity appears to be cytosolic, raising the question of how the internalized, vesicular insulin becomes associated with the degrading enzyme, although this activity also has been found in endosomes (Hamel et al., 1991). A second insulindegrading enzyme also has been described (Authier et al., 1994). The relative roles of these enzymes have not been established. Insulin-degrading enzyme also may have a role in the degradation of other hormones, including glucagon. Molecular Mechanisms of Insulin Action Cellular Actions of Insulin Insulin elicits a remarkable array of biological responses. The important target tissues for regulation of glucose homeostasis by insulin are liver, muscle, and fat, but insulin exerts potent regulatory effects on other cell types as well. Insulin is the primary hormone responsible for controlling the uptake, utilization, and storage of cellular nutrients. Insulin's anabolic actions include the stimulation of intracellular utilization and storage of glucose, amino acids, and fatty acids, while it inhibits catabolic processes, such as the breakdown of glycogen, fat, and protein. It accomplishes these general purposes by stimulating the transport of substrates and ions into cells, promoting the translocation of proteins between cellular compartments, activating and inactivating specific enzymes, and changing the amounts of proteins by altering the rates of gene transcription and specific mRNA translation (see Figure 613).

Figure 613. Model of Insulin Action at the Cellular and Molecular Level. Insulin signaling pathways. Binding of insulin to its specific cell-membrane receptor results in a cascade of intracellular events. The stimulation of the intrinsic tyrosine kinase activity of the insulin receptor marks the initial event, resulting in increased tyrosine phosphorylation (Y Y-P) of both the receptor and specific signaling molecules. This increase in phosphotyrosine stimulates the activity of many intracellular molecules such as GTPases, protein kinases, and lipid kinases, all of which have a role to play in certain metabolic actions of insulin. The two best-described pathways are shown. First, phosphorylation of IRS-2 results in the activation of the lipid kinase, PI 3-kinase, and generates novel inositol lipids that may act as "second messenger" molecules, which, in turn, activate a variety of poorly described signaling pathways (e.g., p70S6 kinase). Second, phophorylation of IRS-1 results in the activation of the small GTPase, p21RAS, and stimulates a protein kinase cascade that activates the p42/p44 MAP kinase isoforms, protein kinases that are important in the regulation of proliferation and differentiation of several cell types. Each of these cascades may influence different physiological processes, as shown. (IRS-1, insulin receptor substrate-1; IRS-2, insulin receptor substrate-2; GRB2, growth factor receptor binding protein 2; mSOS, mammalian son of sevenless; MEK, MAP kinase kinase and ERK kinase; MAP kinase, mitogen-activated protein kinase; p90rsk, p90 ribosomal protein S6 kinase; MAPKAP K2, MAP kinaseactivated protein kinase-2; PI 3-kinase, phosphatidylinositol 3-kinase; p70 S6, p70 ribosomal protein S6 kinase; akt (PKB), protein kinase B.) (Modified from Granner, 2000, with permission.)

Some effects of insulin occur within seconds or minutes, including the activation of glucose and ion transport systems, the covalent modification (i.e., phosphorylation or dephosphorylation) of enzymes, and some effects on gene transcription (i.e., inhibition of the phosphoenolpyruvate carboxykinase gene) (Granner, 1987; O'Brien and Granner, 1996). Other effects, such as those on protein synthesis and gene transcription, may take a few hours. Effects of insulin on cell proliferation and differentiation may take days. It is not clear whether these kinetic differences result from the use of different mechanistic pathways or from the intrinsic kinetics of the various processes. Regulation of Glucose Transport Stimulation of glucose transport into muscle and adipose tissue is a crucial component of the physiological response to insulin. Glucose enters cells by facilitated diffusion through one of a family of glucose transporters. Five of these (GLUT1 through GLUT5) are thought to be involved in Na+-independent facilitated diffusion of glucose into cells (Shepherd and Kahn, 1999). The glucose transporters are integral membrane glycoproteins with molecular masses of about 50,000 daltons, and each has 12 membrane-spanning -helical domains. Insulin stimulates glucose transport at least in part by promoting the energy-dependent translocation of intracellular vesicles that contain the GLUT4 and GLUT1 glucose transporters to the plasma membrane (Suzuki and Kono, 1980; Simpson and Cushman, 1986; see Figure 613). This effect is reversible; the

transporters return to the intracellular pool upon removal of insulin. Faulty regulation of this process may contribute to the pathophysiology of type 2 diabetes (Shepherd and Kahn, 1999). Regulation of Glucose Metabolism The facilitated diffusion of glucose into cells along a downhill gradient is assured by glucose phosphorylation. This enzymatic reaction, the conversion of glucose to glucose 6-phosphate (G6P), is accomplished by one of a family of hexokinases. The four hexokinases (I through IV), like the glucose transporters, are distributed differently in tissues, and two are regulated by insulin. Hexokinase IV, a 50,000-dalton enzyme more commonly known as glucokinase, is found in association with GLUT2 in liver and pancreatic cells. There is one glucokinase gene, but different first exons and promoters are employed in the two tissues (Printz et al., 1993a). The liver glucokinase gene is regulated by insulin (Magnuson et al., 1989). Hexokinase II, a 100,000-dalton enzyme, is found in association with GLUT4 in skeletal and cardiac muscle and in adipose tissue. Like GLUT4, hexokinase II is regulated at the transcriptional level by insulin (Printz et al., 1993b). G6P is a branch-point substrate. It can enter the glycolytic pathway and lead to the production of ATP through a series of enzymatic reactions, many of which are promoted by insulin. The effects of insulin on this pathway are exerted on gene transcription or through alteration of enzyme activity by phosphorylation or dephosphorylation on serine and/or threonine residues. Alternatively, G6P can be incorporated into glycogen after isomerization to glucose 1-phosphate (G1P). Insulin promotes glycogen deposition by stimulating the activity of glycogen synthase, the rate-limiting enzyme in glycogen synthesis, and by inhibiting phosphorylase, the rate-controlling enzyme in glycogen degradation. As in glycolysis, these effects of insulin are mediated through changes in the phosphorylation state of the enzymes. Covalent modification by phosphorylation/dephosphorylation is a major mechanism of action of insulin. For example, phosphorylation increases the activity of acetyl-CoA carboxylase and citrate lyase, whereas glycogen synthase and pyruvate dehydrogenase are activated by dephosphorylation. The latter occurs as a result of the activation of phosphatases by insulin. Dozens of proteins are so modified, with resulting changes in their activity (Denton, 1986). Regulation of Gene Transcription It is now clear that a major action of insulin is the regulation of transcription of specific genes. The first example of this activity to be identified was the inhibition of phosphoenolpyruvate carboxykinase transcription by insulin (Granner et al., 1983). This finding helped explain how insulin inhibits gluconeogenesis (Sasaki et al., 1984) and may explain why the liver overproduces glucose in the insulin-resistant state that is characteristic of non-insulin-dependent diabetes mellitus (Granner and O'Brien, 1992). There are now more than 100 examples of genes that are regulated by insulin (O'Brien and Granner, 1996), and the list continues to grow. The exact mechanism by which these effects are accomplished is not known. The Insulin Receptor Insulin initiates its actions by binding to a cell-surface receptor. Such receptors are present in virtually all mammalian cells, including not only the classic targets for insulin action (liver, muscle, and fat) but also such nonclassic targets as circulating blood cells, brain cells, and gonadal cells. The number of receptors varies from as few as 40 per cell on erythrocytes to 300,000 per cell on adipocytes and hepatocytes. The insulin receptor is a large transmembrane glycoprotein composed of two 135,000-dalton subunits (719 or 731 amino acids, depending on whether a 12-amino-acid insertion has occurred

through alternate splicing of mRNA) and two 95,000-dalton subunits (620 amino acids); the subunits are linked by disulfide bonds to form a - - - heterotetramer (Figure 613) (Virkamki et al., 1999). Both subunits are derived from a single-chain precursor molecule that contains the entire sequence of the and subunits, separated by a processing site consisting of four basic amino acid residues. These two subunits are specialized to perform the two functions of the receptor. The subunits are entirely extracellular and contain the insulin-binding domain (see above), while the subunits are transmembrane proteins that possess tyrosine protein kinase activity. After insulin is bound, receptors aggregate and are rapidly internalized. Since bivalent (but not monovalent) antiinsulin receptor antibodies cross-link adjacent receptors and mimic the rapid actions of insulin, it has been suggested that aggregation of the receptor is essential for signal transduction. After internalization, the receptor may be degraded or recycled back to the cell surface. Tyrosine Phosphorylation and the Insulin Action Cascade The insulin receptor and the receptors for several other growth factors are ligand-activated tyrosine protein kinases (Virkamki et al., 1999). Other growth factor receptors that exhibit such activity include those for epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and colony-stimulating factor-1 (Yarden and Ullrich, 1988). The large family of tyrosine protein kinases also includes several retrovirus-encoded proteins that cause cellular transformation (e.g., Src). Binding of hormone to the subunits of the heterotetrameric insulin receptor leads to the rapid intramolecular autophosphorylation of several tyrosine residues in the subunits. Phosphorylation of the receptor is autocatalytic and results in substantial enhancement of the receptor's tyrosine kinase activity toward other substrates. In intact cells, the insulin receptor also is phosphorylated on serine and threonine residues, presumably by protein kinase C and cyclic AMP-dependent protein kinase. Such phosphorylation inhibits the tyrosine kinase activity of the insulin receptor (Cheatham and Kahn, 1995). The tyrosine kinase activity of the insulin receptor is required for signal transduction. Mutation of the insulin receptor with modification of the ATP-binding site or replacement of the tyrosine residues at major sites of autophosphorylation results in a decrease both of insulin-stimulated kinase activity and of the cellular response to insulin (Ellis et al., 1986). An insulin receptor incapable of autophosphorylation is biologically inert. The activated receptor kinase initiates a cascade of events by first phosphorylating one of a family of proteins called insulin receptor substrates (IRS-1 to 4) (White, et al., 1985). Phosphorylated IRS2 serves as a docking protein for other proteins that contain so-called Src homology 2 (SH2) domains. One of these SH2-domain proteins is phosphoinositide (PI) 3-kinase. PI 3-kinase is a heterodimer consisting of a 110,000-dalton (p110) catalytic subunit and an 85,000-dalton (p85) regulatory subunit. The p85 subunit contains two SH2 domains, and these bind to IRS-1. PI 3kinase catalyzes the addition of phosphate to phosphoinositides on the 3-position of the D-myoinositol ring, and these compounds apparently are involved in signal transduction. PI 3-kinase is activated by a number of hormones that stimulate mitogenesis, including PDGF, EGF, and interleukin-4 (IL-4) (Virkamki et al., 1999). PI 3-kinase is not thought to be the final mediator of mitogenesis; other steps, including the activation of one or more kinases, including protein kinase B (akt/PKB), appear to be involved. The Ras oncoprotein is one of the most potent mitogens. Ras has been linked to the insulin-action pathway because it is known to activate the cascade of mitogen-activated protein (MAP) kinases, and the MAP kinases are among the many that insulin is known to activate (Avruch et al., 1994). The biochemistry of this association has been clarified. Although many points remain obscure,

activation of receptor tyrosine kinases, such as the insulin receptor, result in the association of another SH2 domain-containing protein, Grb2, with phosphorylated IRS-1. Grb2 binds to the guanine nucleotide exchange factor mSOS, and this complex increases the affinity of Ras for GTP. Activated Ras binds to Raf-1, a serine/threonine protein kinase that activates the MAP kinase cascade. Alternatively, the SH2 domain-containing protein Shc is phosphorylated by the activated insulin receptor. Phospho-Shc also binds to Grb2 and activates the MAP kinase cascade through Ras and Raf-1, presumably by enhancing mSOS association with the surface membrane. Although the exact mechanism of mitogenesis in response to insulin is unclear, it appears that multiple, and possibly redundant, pathways are involved (Avruch et al., 1994). The metabolic actions of insulin appear to be mediated by the IRS-2 pathway. Translocation of the adipocyte and muscle cell glucose transporter, and attendant increases in glucose uptake, are a major action of the hormone. Translocation of the glucose transporter is inhibited by wortmannin, which is an inhibitor of PI 3-kinase. The effects of insulin on metabolic gene transcription also are inhibited by wortmannin, and presumably are mediated by the IRS-2 pathway and downstream targets of PI 3-kinase. Diabetes Mellitus and the Physiological Effects of Insulin Diabetes mellitus is a group of syndromes characterized by hyperglycemia; altered metabolism of lipids, carbohydrates, and proteins; and an increased risk of complications from vascular disease. Most patients can be classified clinically as having either type 1 diabetes mellitus (type 1 DM, formerly known as insulin-dependent diabetes or IDDM) or type 2 diabetes mellitus (type 2 DM, formerly known as non-insulin-dependent diabetes or NIDDM) (Alberti and Zimmet, 1998; Expert Committee, 1997). Diabetes mellitus or carbohydrate intolerance also is associated with certain other conditions or syndromes (see Table 611). The incidence of each type of diabetes varies widely throughout the world. In the United States, about 5% to 10% of all diabetic patients have type 1 DM, with an incidence of 18 per 100,000 inhabitants per year. This is similar to the incidence found in the United Kingdom (17 per 100,000). The incidence of type 1 DM in Europe varies with latitude. The highest rates occur in northern Europe (Finland, 43 per 100,000), and the lowest in the south (France, Italy, and Israel, 8 per 100,000). The one exception to this rule is the small island of Sardinia, close to Italy, which has an incidence of 30 per 100,000. However, the relatively low incidence rates of type 1 DM in southern Europe are far higher than the rates in Japan, which are only about 1 per 100,000 inhabitants. The vast majority of diabetic patients have type 2 DM. In the United States, about 90% of all diabetic patients have type 2 DM. Incidence rates of type 2 DM increase with age, with a mean rate of about 440 per 100,000 per year by the sixth decade in males in the United States. Ethnicity within a country also can influence the incidence of type 2 DM; the mean rate in African-American males is 540 per 100,000, and that in Pima Indians is about 5000 per 100,000. Unlike those for type 1 DM, the incidence rates for type 2 DM are lower in northern Europe (100 to 250 per 100,000) than in the south (Israel, 800 per 100,000). Although prevalence data exist for type 2 DM, it should be noted that there is an equal number of undiagnosed cases. There are more than 125 million persons with diabetes in the world today, and by 2010 this number is expected to approach 220 million (Amos et al., 1997). Some investigators expect the incidence to double by 2025. Types 1 and 2 are both increasing in frequency. The reason for the increase of type 1 DM is not known. The genetic basis for type 2 DM cannot change in such a short time; thus, other contributing factors including increasing age, obesity, sedentary lifestyle, and low birth weight must account for this dramatic increase. In addition, type 2 DM now is being diagnosed with remarkable

frequency in preadolescents and adolescents. In certain tropical countries, the most common cause of diabetes is chronic pancreatitis associated with nutritional or toxic factors (a form of secondary diabetes). Also, on rare occasions, diabetes results from point mutations in the insulin gene (Chan et al., 1987). Amino acid substitutions from such mutations may result in insulins with lower potency or may alter the processing of proinsulin to insulin (see above). Other single-gene mutations cause the several types of MODY (Hattersley, 1998) and maternally inherited diabetes and deafness (MIDD, van den Ouwenland et al., 1992) (see Table 611). There are genetic and environmental components to both type 1 DM and type 2 DM. A number of factors place persons at high risk for developing type 2 DM. A positive family history is predictive for the disease. Studies of identical twins show 70% to 80% concordance for developing type 2 DM (Newman et al., 1987). Furthermore, there is a high prevalence of type 2 DM in offspring of parents with the disease (up to 70%) and also in siblings of affected individuals. Persons more than 20% over their ideal body weight also have a greater risk of developing type 2 DM. In fact, 70% of type 2 DM subjects in the United States are obese. Certain ethnic groups have a higher incidence of type 2 DM (American Indians, African-Americans, Hispanics, Polynesian Islanders). In addition, previously identified impaired glucose tolerance, gestational diabetes, hypertension, or significant hyperlipidemia are associated with an increased risk of type 2 DM. These data suggest that there is a strong genetic basis for type 2 DM, but the genetic mechanism(s) involved are not known. A pancreatic -cell defect and a reduction in tissue sensitivity to insulin both are required before phenotypic type 2 DM is apparent. However, type 2 DM is an extremely heterogenous disease, and it is likely that a variety of different genes are involved. In addition, environmental factors could play a role. Type 2 DM thus is considered to be a multifactorial disease. Any combination of genetic and environmental factors that exceeds a threshold can result in type 2 DM. The genetic basis for type 2 DM in a small subset of patients has been established. One-half of patients with a rare type of type 2 DM called MODY2 (maturity-onset diabetes of the young) have a mutation of the glucokinase gene as the primary cause of diabetes. Because of decreased glucokinase activity, these patients have an increase in the glycemic threshold for insulin release. This, in turn, results in persistent mild hyperglycemia. This form of MODY is familial, with autosomal dominant inheritance, and appears to be quite distinct from the usual type of type 2 DM as are the other forms of MODY (see Table 611). With type 1 DM, the concordance rate for identical twins is only 25% to 50%; this suggests that environmental as well as genetic influences have an important role in the disease. However, the genetic factors in type 1 DM are well characterized and relate to the genes that control the immune response. There is considerable evidence that type 1 DM can be caused by an autoimmune disease of the pancreatic cell. Antibodies to components of islet cells are detected in up to 80% of patients with type 1 DM early during the onset or prior to the onset of clinical disease. The antibodies are directed at both cytoplasmic and membrane-bound antigens and include islet-cell antibodies and antibodies directed against insulin, glutamic acid decarboxylase-65 and -67 (GAD-65 and -67), heat-shock protein-65 (HSP-65), bovine serum albumin, and tyrosine phosphatase-like protein (IA2 or IA-2B). Although it is now accepted that these antibodies are correlated with the clinical expression of type 1 DM, it is controversial whether or not the presence of autoantibodies can predict the development of clinical diabetes. Most prospective studies designed to determine if type 1 DM can be predicted on the basis of antibodies have been performed in healthy first-degree relatives of diabetic patients. These studies have determined that the presence of insulin autoantibodies (IAA) confers only a small risk for the development of type 1 DM. On the other hand, the presence of high-titer islet-cell

antibodies (ICA) and GAD antibodies, or ICA combined with IAA, confers a very high risk for the development of type 1 DM in first-degree relatives (Verge et al., 1996). As most of the studies aimed at predicting the development of type 1 DM have been carried out in first-degree relatives of diabetic patients, it is not known whether or not the occurrence of ICA in individuals from the general population confers a similar risk for development of clinical diabetes. Most available data indicate that the presence of ICA in individuals from the general population is associated with a lower risk of developing type 1 DM. However, as in first-degree relatives of type 1 DM patients, it may be that the presence of more than one form of autoantibody in individuals from the general population can be a more powerful predictor of the development of clinical diabetes (Bingley et al., 1993). Individuals with type 1 DM also tend to have antibodies directed toward other endocrine tissues, including the adrenal, parathyroid, and thyroid glands, which can be clinically significant. They also have a higher than normal incidence of other autoimmune diseases. There is an association of type 1 DM with specific human leukocyte antigen (HLA) types, especially at the B and Dr loci. Approximately 90% of patients with type 1 diabetes are positive for HLA-Dr3 and/or Dr4, as compared with only 40% of the general population (Nerup et al., 1984). In contrast, the haplotype HLA-Dr2 appears to be negatively associated with the occurrence of the disease. A polymorphism of the HLA-DQ chain at position 57 correlates even more closely with susceptibility to diabetes (Todd et al., 1987). Type 1 DM is associated with alleles coding for alanine, valine, or serine at position 57 in the HLA-DQ chain, while aspartic acid in this position is negatively correlated with the disease in Caucasians (see Dotta and Eisenbarth, 1989). These findings implicate both humoral and cell-mediated immune mechanisms in the etiology of type 1 DM. The trigger for the immune response remains unknown. The identification of triggering agents is difficult, since autoimmune destruction of pancreatic cells may occur over a period of many months or several years before the onset of overt disease (Srikanta et al., 1983). In about 10% of new cases of type 1 DM, there is no evidence of autoimmune insulitis (Imagawa et al., 2000). The American Diabetes Association and the World Health Organization subdivide this disease into autoimmune (1A) and idiopathic (1B) subtypes. Whatever the causes, the final result in type 1 DM is an extensive and selective loss of pancreatic cells and a state of absolute insulin deficiency. The situation in type 2 DM is not so clear-cut. Most studies indicate that there is a reduction in cell mass in type 2 DM patients. Obesity, duration of diabetes, and prevailing hyperglycemia potentially can confound interpretation of data, but studies that have controlled for these variables have reported an approximately 50% reduction in -cell volume in type 2 DM patients compared to nondiabetic control subjects (Leahy, 1990). Owing to the heterogeneous nature of type 2 DM, mean 24-hour plasma concentrations of insulin in patients have been reported to vary from low to normal to even increased relative to values in control subjects. It is important to realize, however, that routine radioimmunoassay of insulin detects precursor (proinsulin) and intermediate forms of proinsulin (32/33 and 64/65 split proinsulin). Studies in which specific insulin and proinsulin assays have been used (Temple et al., 1989) have revealed that "true" insulin values in "hyperinsulinemic" type 2 DM patients are, in fact, no greater or distinctly less than values in control subjects. Therefore, increased amounts of proinsulin have confounded the appreciation of subnormal insulin levels in type 2 DM patients. In healthy persons, the contribution of proinsulin to basal immunoreactive insulin levels is low. Proinsulin intermediates make up about 10% of the total immunoreactive insulin in the portal vein. However, owing to its long half-life (about 44 minutes) and tenfold slower metabolic clearance, proinsulin and intermediates make up about 20% of circulating immunoreactive insulin. This

amount is physiologically trivial, as proinsulin has only about 5% the metabolic effect of insulin (Davis et al., 1991b). Nevertheless, recent data indicate that plasma proinsulin-like molecules are increased in type 2 DM to about 20% or more of total immunoreactive insulin. Furthermore, proinsulin levels increase in response to any -cell stimulation. Type 2 DM also is associated with several distinct defects in insulin secretion. The earliest manifestation is a loss of the regular periodicity of insulin secretion. At diagnosis, virtually all persons with type 2 DM have a profound defect in first-phase insulin secretion in response to an intravenous glucose challenge. The responses to other secretagogues (e.g., isoproterenol or arginine) are preserved, although there is less potentiation by glucose (Weir et al., 1986; Leahy et al., 1987). Some of these abnormalities of the cell in type 2 DM are in part secondary to desensitization by chronic hyperglycemia. The relationship between fasting glycemia and insulinemia in type 2 DM subjects is complex. Patients who have fasting blood glucose levels of 6 to 10 mM (108 to 180 mg/dl) have fasting and stimulated insulin values equal to those of euglycemic control subjects. More severely hyperglycemic subjects are frankly hypoinsulinemic. Insulin levels in type 2 DM patients with mild hyperglycemia, although similar to those in euglycemic control subjects, are in fact inappropriately low, as they should be increased commensurate with the hyperglycemic stimulus. Virtually all forms of diabetes mellitus are caused by a decrease in the circulating concentration of insulin (insulin deficiency) and a decrease in the response of peripheral tissues to insulin (insulin resistance). These abnormalities lead to alterations in the metabolism of carbohydrates, lipids, ketones, and amino acids; the central feature of the syndrome is hyperglycemia (see Figure 614). Figure 614. Overview of Insulin Action. Insulin stimulates the storage of glucose in the liver as glycogen and in adipose tissue as triglycerides and the storage of amino acids in muscle as protein; it also promotes utilization of glucose in muscle for energy. These pathways, which also are enhanced by feeding, are indicated by the solid blue arrows. Insulin inhibits the breakdown of triglycerides, glycogen, and protein and the conversion of amino acids to glucose (gluconeogenesis), as indicated by the open arrows. These pathways are increased during fasting and in diabetic states. The conversion of amino acids to glucose and of glucose to fatty acids occurs primarily in the liver.

Insulin lowers the concentration of glucose in blood by inhibiting hepatic glucose production and by

These two important effects occur at different concentrations of insulin. Production of glucose is inhibited half maximally by an insulin concentration of about 20 U/ml, while glucose utilization is stimulated half maximally at about 50 U/ml. In both types of diabetes, glucagon (levels of which are elevated in untreated patients) opposes the effect of insulin on the liver by stimulating glycogenolysis and gluconeogenesis, but it has relatively little effect on peripheral utilization of glucose. Thus, in the diabetic patient with insulin deficiency or insulin resistance and hyperglucagonemia, there is an increase in hepatic glucose production, a decrease in peripheral glucose uptake, and a decrease in the conversion of glucose to glycogen in the liver (DeFronzo et al., 1992). Alterations in secretion of insulin and glucagon also have profound effects on lipid, ketone, and protein metabolism. At concentrations below those required to stimulate glucose uptake, insulin inhibits the hormone-sensitive lipase in adipose tissue and thus inhibits the hydrolysis of triglycerides stored in the adipocyte. This counteracts the lipolytic action of catecholamines, cortisol, and growth hormone and reduces the concentrations of glycerol (a substrate for gluconeogenesis) and free fatty acids (a substrate for production of ketone bodies and a necessary fuel for gluconeogenesis). These actions of insulin are deficient in the diabetic patient, leading to increased gluconeogenesis and ketogenesis. The liver produces ketone bodies by oxidation of free fatty acids to acetyl CoA, which is then converted to acetoacetate and -hydroxybutyrate. The initial step in fatty-acid oxidation is transport of the fatty acid into the mitochondria. This involves the interconversion of the CoA and carnitine esters of fatty acids by the enzyme acylcarnitine transferase. The activity of this enzyme is inhibited by intramitochondrial malonyl CoA, one of the products of fatty-acid synthesis. Under normal conditions, insulin inhibits lipolysis, stimulates fatty-acid synthesis (thereby increasing the concentration of malonyl CoA), and decreases the hepatic concentration of carnitine; all these factors decrease the production of ketone bodies. Conversely, glucagon stimulates ketone-body production by increasing fatty-acid oxidation and decreasing concentrations of malonyl CoA. In the diabetic patient, particularly the patient with type 1 DM, the consequences of insulin deficiency and glucagon excess provide a hormonal milieu that favors ketogenesis and, in the absence of appropriate treatment, may lead to ketonemia and acidosis (see Foster, 1984). Insulin also enhances the transcription of lipoprotein lipase in the capillary endothelium. This enzyme hydrolyzes triglycerides present in very-low-density lipoproteins (VLDL) and chylomicrons, resulting in release of intermediate-density lipoprotein (IDL) particles (see also Chapter 36: Drug Therapy for Hypercholesterolemia and Dyslipidemia). The IDL particles are converted by the liver to the more cholesterol-rich low-density lipoproteins (LDL). Thus, in the untreated or undertreated diabetic patient, hypertriglyceridemia and hypercholesterolemia often occur. In addition, deficiency of insulin may be associated with increased production of VLDL. The important role of insulin in protein metabolism usually is evident clinically only in diabetic patients with persistently poor control of their disease. Insulin stimulates amino acid uptake and protein synthesis and inhibits protein degradation in muscle and other tissues; it thus causes a decrease in the circulating concentrations of most amino acids. Glutamine and alanine are the major amino acid precursors for gluconeogenesis. Insulin lowers alanine concentrations during hyperinsulinemic euglycemic conditions. The rate of appearance of alanine is maintained in part by the enhanced rate of transamination of pyruvate to alanine. However, alanine utilization greatly exceeds production (owing to increased hepatic uptake and fractional extraction of the amino acid), and this results in a fall of peripheral alanine levels. In a poorly controlled hyperglycemic diabetic subject, there is increased conversion of alanine to glucose, contributing to the enhanced rate of

gluconeogenesis. The conversion of larger amounts of amino acids to glucose also results in increased production and excretion of urea and ammonia. In addition, there are increased circulating concentrations of the branched-chain amino acids as a result of increased proteolysis, decreased protein synthesis, and increased release of branched-chain amino acids from the liver. An almost pathognomonic feature of diabetes mellitus is thickening of the capillary basement membrane and other vascular changes that occur during the course of the disease. The cumulative effect is progressive narrowing of the vessel lumina, causing inadequate perfusion of critical regions of certain organs. The matrix is expanded in many vessel walls, in the basement membrane of the retina, and in the mesangial cells of the renal glomerulus (McMillan, 1997). Cellular proliferation in many large vessels further contributes to luminal narrowing. These pathological changes contribute to some of the major complications of diabetes, including premature atherosclerosis, intercapillary glomerulosclerosis, retinopathy, neuropathy, and ulceration and gangrene of the extremities. It has been hypothesized that the factor responsible for the development of most complications of diabetes is the prolonged exposure of tissues to elevated concentrations of glucose. Prolonged hyperglycemia results in the formation of advanced glycation end products (AGE) (Beisswenger, et al., 1995). These macromolecules are thought to induce many of the vascular abnormalities that result in the complications of diabetes (Brownlee, 1995). The results from the Diabetes Control and Complications Trial (DCCT) have definitively answered this question in the affirmative: Most diabetic complications arise from prolonged exposure of tissue to elevated glucose concentrations. The DCCT (DCCT Research Group, 1993) was a multicenter, randomized clinical trial designed to compare intensive with conventional diabetes therapy with regard to their effects on the development and progression of the early vascular and neurologic complications of type 1 DM. The intensive therapy regimen was designed to achieve blood glucose values as close to the normal range as possible with three or more daily insulin injections or with an external insulin pump. Conventional therapy consisted of one or two daily insulin injections. Two groups of patients were studied to answer separate but related questions. The first question was whether or not intensive therapy could prevent the development of diabetic tissue complications such as retinopathy, nephropathy, and neuropathy (primary prevention). The second was whether or not intensive therapy could slow the progression of existing tissue complications of diabetes (secondary intervention). The results of the DCCT were definitive. In the primary prevention group, intensive therapy reduced the mean risk for the development of retinopathy by 76% compared to conventional therapy. In the secondary intervention group, intensive therapy slowed the progression of retinopathy by 54%. Intensive therapy reduced the risk of nephropathy by 34% in the primary prevention group and by 43% in the secondary intervention group. Similarly, neuropathy was reduced by about 60% in both the primary prevention and secondary intervention groups. Intensive therapy reduced the development of hypercholesterolemia by 34% in the combined groups. Because of the relative youth of the patients, it was predicted that the detection of treatment-related differences in rates of macrovascular events would be unlikely. However, intensive therapy reduced the risk of macrovascular disease by 41% in the combined groups. Thus, it is clear that improving day-to-day glycemic control in type 1 DM patients can dramatically reduce and slow the development of tissue complications of diabetes. A follow-up study showed that the reduction in the risk of progressive retinopathy and nephropathy persists for at least 4 years, even if glycemic control has not been well maintained (DCCT Research Group, 2000). A serious complication of intensive therapy was an increased incidence of severe hypoglycemia. Patients receiving intensive therapy had a threefold greater incidence of severe hypoglycemia

(blood glucose below 50 mg/dl or 2.8 mM and needing external resuscitative assistance) and hypoglycemic coma than did conventionally treated subjects. Therefore, the present guidelines for treatment given by the American Diabetes Association include a contraindication for implementing tight metabolic control in infants less than 2 years old and an extreme caution in children between 2 and 7 years, as hypoglycemia may impair brain development. Older patients with significant arteriosclerosis also may be vulnerable to permanent injury from hypoglycemia. The DCCT was performed in relatively young type 1 DM patients. The question was asked whether or not intensive therapy would provide similar benefits to the typical middle-aged or elderly person with type 2 DM. The results of the DCCT indeed have been found to apply to patients with type 2 DM [UK Prospective Diabetes Study (UKPDS) Group, 1998a,b]. The eye, kidney, and nerve abnormalities appear similar in type 1 DM and type 2 DM, and it is likely that the same or similar underlying mechanisms of disease apply. However, because of a higher prevalence of macrovascular disease, older patients with type 2 DM may be more vulnerable to serious consequences of hypoglycemia. Thus, as is the case for everyone with diabetes, treatment of type 2 DM patients has to be tailored to the individual. Nevertheless, the results of the DCCT and UKPDS suggest that many otherwise healthy patients with type 2 DM should attempt to achieve tight metabolic control. The toxic effects of hyperglycemia may be the result of accumulation of nonenzymatically glycosylated products and osmotically active sugar alcohols such as sorbitol in tissues; the effects of glucose on cellular metabolism also may be responsible (Brownlee, 1995). The covalent reaction of glucose with hemoglobin provides a convenient method to determine an integrated index of the glycemic state. Hemoglobin undergoes glycosylation on its amino-terminal valine residue to form the glucosyl valine adduct of hemoglobin, termed hemoglobin A1c (Brownlee, 1995). The half-life of the modified hemoglobin is equal to that of the erythrocyte (about 120 days). Since the amount of glycosylated protein formed is proportional to the glucose concentration and the time of exposure of the protein to glucose, the concentration of hemoglobin A1c in the circulation reflects the severity of the glycemic state over an extended period (4 to 12 weeks) prior to sampling. Thus, a rise in hemoglobin A1c from 5% to 10% suggests a prolonged doubling of the mean blood glucose concentration. Although this assay is applied widely, measurement of the glycosylation of proteins with somewhat shorter survival times (e.g., albumin) also has proven useful in the management of pregnant diabetic patients. Glycosylated products accumulate in tissues and may eventually form cross-linked proteins termed advanced glycosylation end products (Beisswenger et al., 1995). It is possible that nonenzymatic glycosylation is directly responsible for expansion of the vascular matrix and the vascular complications of diabetes. The modified cellular proliferative activity in vascular lesions of diabetic patients also might be explained by this process, since macrophages appear to have receptors for advanced glycosylation end products. Binding of such proteins to macrophages in these lesions may stimulate the production of cytokines such as tumor necrosis factor and interleukin-1, which in turn induce degradative and proliferative cascades in mesenchymal and endothelial cells, respectively. Other explanations for the toxic manifestations of hyperglycemia may exist. Intracellular glucose is reduced to its corresponding sugar alcohol, sorbitol, by the enzyme aldose reductase (Burg and Kador, 1988), and the rate of production of sorbitol is determined by the ambient glucose concentration. This is particularly true in tissues such as the lens, retina, arterial wall, and Schwann cells of peripheral nerves. In diabetic human beings and rodents, these tissues have increased intracellular concentrations of sorbitol, which may contribute to an increased osmotic effect and tissue damage. Inhibitors of aldose reductase currently are being evaluated for treatment of diabetic neuropathy and retinopathy. The results of studies with these agents thus far have been somewhat

conflicting and inconclusive (reviewed by Frank, 1994). In neural tissue and perhaps in other tissues, glucose competes with myoinositol for transport into cells (Greene et al., 1987). Reduction of cellular concentrations of myoinositol may contribute to altered nerve function and neuropathy. Hyperglycemia also may enhance the de novo synthesis of diacylglycerol, which could facilitate persistent activation of protein kinase C (Lee et al., 1989). Insulin Therapy Insulin is the mainstay for treatment of virtually all type 1 DM and many type 2 DM patients. When necessary, insulin may be administered intravenously or intramuscularly; however, long-term treatment relies predominantly on subcutaneous injection of the hormone. Subcutaneous administration of insulin differs from physiological secretion of insulin in at least two major ways: The kinetics do not reproduce the normal rapid rise and decline of insulin secretion in response to ingestion of nutrients, and the insulin diffuses into the peripheral circulation instead of being released into the portal circulation; the direct effect of secreted insulin on hepatic metabolic processes is thus eliminated. Nonetheless, when such treatment is performed carefully, considerable success is achieved. Preparations of insulin can be classified according to their duration of action into short-, intermediate-, and long-acting and by their species of originhuman, porcine, bovine, or a mixture of bovine and porcine. Human insulin (HUMULIN , NOVOLIN) is now widely available as a result of its production by recombinant DNA techniques. Porcine insulin differs from human insulin by one amino acid (alanine instead of threonine at the carboxy terminal of the B chain, i.e., in position B30), and bovine insulin differs by two additional alterations of the A chain (threonine and isoleucine in positions A8 and A10 are replaced by alanine and valine, respectively). Prior to the mid-1970s, commercially available insulin preparations contained proinsulin or glucagon-like substances, pancreatic polypeptide, somatostatin, and vasoactive intestinal peptides. These contaminants were avoided with the advent of monocomponent porcine insulins. During the late 1970s, intense work was carried out on the development of biosynthetic human insulin. During the last decade, the use of human insulin has rapidly become the standard form of therapy. The physicochemical properties of human, porcine, and bovine insulins differ owing to their different amino acid sequences. Human insulin, produced using recombinant DNA technology, is more soluble than porcine insulin in aqueous solutions, owing to the presence of threonine (instead of alanine), with its extra hydroxyl group. The vast majority of preparations now are supplied at neutral pH, which improves stability and permits storage for several days at a time at room temperature. Unitage For therapeutic purposes, doses and concentrations of insulin are expressed in units (U). This tradition dates to the time when preparations of the hormone were impure and it was necessary to standardize them by bioassay. One unit of insulin is equal to the amount required to reduce the concentration of blood glucose in a fasting rabbit to 45 mg/dl (2.5 mM). The current international standard is a mixture of bovine and porcine insulins and contains 24 U/mg. Homogeneous preparations of human insulin contain between 25 and 30 U/mg. Almost all commercial preparations of insulin are supplied in solution or suspension at a concentration of 100 U/ml, which is about 3.6 mg of insulin per milliliter (0.6 mM). Insulin also is available in a more concentrated solution (500 U/ml) for patients who are resistant to the hormone.

Classification of Insulins Shortand rapid-acting insulins are solutions of regular, crystalline zinc insulin (insulin injection) dissolved usually in a buffer at neutral pH. These preparations have the most rapid onset of action but the shortest duration (see Table 613). Short-acting insulin (i.e., regular or soluble) usually should be injected 30 to 45 minutes before meals (Dimitriadis and Gerich, 1983). Regular insulin also may be given intravenously or intramuscularly. After intravenous injection, there is a rapid fall in the blood glucose concentration, which usually reaches a nadir in 20 to 30 minutes. In the absence of a sustained infusion of insulin, the hormone is rapidly cleared, and counterregulatory hormones (glucagon, epinephrine, norepinephrine, cortisol, and growth hormone) restore plasma glucose to baseline in 2 to 3 hours. In the absence of a normal counterregulatory response (e.g., in diabetic patients with autonomic neuropathy), plasma glucose will remain suppressed for many hours following an insulin bolus of 0.15 U/kg, because the cellular actions of insulin are prolonged far beyond its clearance from plasma. Intravenous infusions of insulin are useful in patients with ketoacidosis or when requirements for insulin may change rapidly, as during the perioperative period, during labor and delivery, and in intensive-care situations (see below). When metabolic conditions are stable, regular insulin usually is given subcutaneously in combination with an intermediateor long-acting preparation. Short-acting insulin is the only form of the hormone that can be used in subcutaneous infusion pumps. Special buffered formulations of regular insulin have been made for the latter purpose; these are less likely to crystallize in the tubing during the slow infusion associated with this type of therapy (Lougheed et al., 1980). The native insulin monomers are associated as hexamers in currently available insulin preparations. These hexamers slow the absorption and reduce postprandial peaks of subcutaneously injected insulin. This unsatisfactory situation has stimulated the development of a number of short-acting insulin analogs that retain a monomeric or dimeric configuration. A large number of compounds have been investigated during the last decade (Brange et al., 1990). Of the analogs tested, two, insulin lispro (HUMALOG) and insulin aspart (NOVOLOG), have demonstrated clinical effectiveness (Kang et al., 1991). These analogs are absorbed three times more rapidly from subcutaneous sites than is human insulin. Consequently there is a more rapid increase in plasma insulin concentrations and an earlier hypoglycemic response. Injection of the analogs 15 minutes before a meal affords glycemic control similar to that from an injection of human insulin given 30 minutes before the meal. The first commercially available short-acting analog was human insulin lispro. This analog is identical to human insulin except at positions B28 and B29, where the sequence of the two residues has been reversed to match the sequence in IGF-1, a polypeptide that does not self-associate. Like regular insulin, lispro exists as a hexamer in commercially available formulations. Unlike regular insulin, lispro dissociates into monomers almost instantaneously following injection. This property results in the characteristic rapid absorption and shorter duration of action compared to regular insulin. A review of clinical experience with insulin lispro has been published (Bolli et al., 1999). Two therapeutic advantages have emerged with lispro as compared to regular insulin. First, the prevalence of hypoglycemia is reduced by 20% to 30% with lispro; second, glucose control, as assessed by hemoglobin A1c, is modestly but significantly improved (0.3% to 0.5%) with lispro as compared to regular insulin. Insulin aspart is formed by the replacement of proline at B28 with aspartic acid. This results in a reduction of self-association to that observed with lispro. Like lispro, insulin aspart rapidly dissociates into monomers following injection. Intermediate-acting insulins are formulated to dissolve more gradually when administered subcutaneously; thus, their durations of action are longer. The two preparations most frequently used are neutral protamine Hagedorn (NPH) insulin (isophane insulin suspension) and lente insulin (insulin zinc suspension). NPH insulin is a suspension of insulin in a complex with zinc and

protamine in a phosphate buffer. Lente insulin is a mixture of crystallized (ultralente) and amorphous (semilente) insulins in an acetate buffer, which minimizes the solubility of insulin. The pharmacokinetic properties of human intermediate-acting insulins are slightly different from those of porcine preparations. Human insulins have a more rapid onset and shorter duration of action than do porcine insulins. This difference may be related to the more hydrophobic nature of human insulin, or human and porcine insulins may interact differently with protamine and zinc crystals. This difference may create a problem with optimal timing for evening therapy; human insulin preparations taken before dinner may not have a duration of action sufficient to prevent hyperglycemia by morning. It should be noted that there is no evidence that lente or NPH insulin has different pharmacodynamic effects when used in combination with regular (soluble) insulin in a twice-a-day dosage regimen (Tunbridge et al., 1989). Intermediate-acting insulins usually are given either once a day before breakfast or twice a day. In patients with type 2 DM, intermediate-acting insulin given at bedtime may help normalize fasting blood glucose (Riddle, 1985). When lente insulin is mixed with regular insulin, some of the regular insulin may form a complex with the 2+ protamine or Zn after several hours, and this may slow the absorption of the fast-acting insulin (Colagiuri and Villalobos, 1986). NPH insulin does not retard the action of regular insulin when the two are mixed vigorously by the patient or when they are commercially available as a mixture (see below; Davis et al., 1991a). Ultralente insulin (extended insulin zinc suspension) and protamine zinc insulin suspension are long-acting insulins; they have a very slow onset and a prolonged, relatively "flat" peak of action. These insulins have been advocated to provide a low basal concentration of insulin throughout the day. The long half-life of ultralente insulin makes it difficult to determine the optimal dosage, since several days of treatment are required before a steady-state concentration of circulating insulin is achieved. As with the intermediate-acting insulins, bovine-porcine ultralente insulin has an even more prolonged course of action than does human ultralente insulin. Doses given once or twice daily are adjusted according to the fasting blood glucose concentration. Protamine zinc insulin rarely is used today because of its very unpredictable and prolonged course of action, and it is no longer available in the United States. Preparations of insulin that are available for clinical use in the United States are shown in Table 614. In the vast majority of patients, insulin replacement therapy includes intermediateor long-acting insulin. A search for an ideal intermediate-acting insulin also has been in progress for the last 15 years. A compound that demonstrated considerable early promise in this regard was human proinsulin (HPI). Animal studies using porcine proinsulin indicated that the compound was a soluble, intermediate-acting insulin agonist that had a greater suppressive effect on hepatic glucose production than on stimulation of peripheral glucose disposal. This profile of action appeared favorable for clinical use in diabetic subjects, since unrestrained hepatic glucose production is a hallmark of the disease, and a hepatospecific insulin would tend to reduce peripheral hyperinsulinemia and the attendant risk of hypoglycemia. Early studies with HPI in human beings confirmed its relatively hepatospecific action and demonstrated that it had a duration of action similar to that of NPH insulin. Preliminary results from clinical trials, however, indicated that HPI conferred no additional benefit over currently available human insulins, and all clinical studies soon were suspended because of a high incidence of myocardial infarction in HPI-treated subjects. Because of pharmacokinetic limitations of ultralenteinsulin, there is a great clinical need for an insulin analog that does not have a significant peak in its action. Considerable research has been directed to the development of such a product. Insulin glargine (LANTUS) is the first long-acting analog of human insulin to be approved for clinical use in the United States. Insulin glargine is produced following two alterations of human insulin (Rosskamp and Park, 1999). Two arginine residues are added to the C terminus of the B chain, and an asparagine molecule in position A21 on

the A chain is replaced with glycine. Glargine is a clear solution with a pH of 4.0. This pH stabilizes the insulin hexamer and results in a prolonged and predictable absorption from subcutaneous tissues. Due to insulin glargine's acidic pH, it cannot be mixed with currently available short-acting insulin preparations (regular insulin or lispro) that are formulated at a neutral pH. Thus far, clinical studies have revealed that insulin glargine may cause less hypoglycemia, result in a sustained "peakless" absorption profile, and provide a better once-daily, 24-hour insulin coverage than ultralente insulin. Other approaches to prolong the action of soluble insulin analogs are under investigation. One approach is the addition of a saturated fatty acid to the amino group of LysB29 (Kurtzhals et al., 1997), yielding an acylated insulin. Clinical trials with such compounds are in progress. The wide variability in the kinetics of insulin action between and even within individuals must be emphasized. The time to peak hypoglycemic effect and insulin levels can vary by 50%. This variability is caused, at least in part, by large variations in the rate of subcutaneous absorption and often has been said to be more noticeable with the intermediate- and long-acting insulins. However, more recent data have demonstrated that the administration of regular insulin can result in similar variability (Davis et al., 1991a). When this variability is coupled with normal variations in diet and exercise, it is sometimes surprising how many patients do achieve good control of blood glucose concentrations. Indications and Goals for Therapy Subcutaneous administration of insulin is the primary treatment for all patients with type 1 DM, for patients with type 2 DM that is not adequately controlled by diet and/or oral hypoglycemic agents, and for patients with postpancreatectomy diabetes or gestational diabetes (American Diabetes Association, 1999). In addition, insulin is critical for the management of diabetic ketoacidosis, and it has an important role in the treatment of hyperglycemic, nonketotic coma and in the perioperative management of both type 1 DM and type 2 DM patients. In all cases, the goal is the normalization not only of blood glucose but also of all aspects of metabolism; the latter is difficult to achieve. Optimal treatment requires a coordinated approach to diet, exercise, and the administration of insulin. A brief overview of the principles of therapy is given below. (For a more detailed description, see LeRoith et al., 2000.) Near-normoglycemia can be attained in patients with multiple daily doses of insulin or with socalled pump therapy. The goal is to achieve a fasting blood glucose concentration between 90 and 120 mg/dl (5 to 6.7 mM) and a 2-hour postprandial value below 150 mg/dl (8.3 mM). In less disciplined patients or in those with defective responses of counterregulatory hormones, it may be necessary to accept higher fasting blood glucose concentrations [e.g., 140 mg/dl (7.8 mM)] and 2hour postprandial concentrations [200 to 250 mg/dl (11.1 to 13.9 mM)]. Daily Requirements Insulin production by a normal, thin, healthy person is between 18 and 40 U per day or about 0.2 to 0.5 U per kilogram of body weight per day (Polonsky and Rubenstein, 1986). About half of this amount is secreted in the basal state and about half in response to meals. Thus, basal secretion is about 0.5 to 1 U per hour; after an oral glucose load, insulin secretion may increase to 6 U per hour (Waldhausl et al., 1979). In nondiabetic, obese, insulin-resistant individuals, insulin secretion may be increased fourfold or more. Insulin is secreted into the portal circulation, and about 50% is destroyed by the liver before reaching the systemic circulation.

In a mixed population of type 1 DM patients, the average dose of insulin is usually 0.6 to 0.7 U/kg body weight per day, with a range of 0.2 to 1 U/kg per day. Obese patients generally require more (about 2 U/kg per day) because of resistance of peripheral tissues to insulin. Patients who require less insulin than 0.5 U/kg per day may have some endogenous production of insulin, or they are more sensitive to the hormone because of good physical conditioning. As in nondiabetics, the daily requirement for insulin can be divided into basal and postprandial needs. The basal dose suppresses hepatic output of glucose; it is usually 40% to 60% of the daily dose. The dose necessary for disposition of nutrients after meals usually is given before meals. Insulin often has been administered as a single daily dose of an intermediate-acting insulin, alone or in combination with regular insulin. This is rarely sufficient to achieve true euglycemia, and in view of the DCCT evidence that hyperglycemia is the major determinant of the long-term complications of diabetes, more complex regimens that include combinations of intermediateor long-acting insulins with regular insulin are used to reach this goal. A number of commonly used dosage regimens that include mixtures of insulin given in two or three daily injections are depicted in Figure 615 (LeRoith et al., 2000). The most frequently used is the so-called "split-mixed" regimen, involving the prebreakfast and presupper injection of a mixture of regular and intermediate-acting insulins (Figure 615A). When the presupper NPH or lente insulin is not sufficient to control hyperglycemia throughout the night, the evening dose may be divided into a presupper dose of regular insulin followed by NPH or lente insulin at bedtime (Figure 61 5B). Both normal and diabetic individuals have an increased requirement for insulin in the early morning; this has been termed the "dawn phenomenon" (Blackard et al., 1989). It makes the kinetics and timing of the evening dose of insulin extremely important. Figure 615. Common Multidose Insulin Regimens. A. A typical "split-mixed" regimen consisting of twice-daily injections of a mixture of regular (regular or lispro) and intermediate-acting (NPH or lente) insulin. B. A variation in which the evening dose of intermediate-acting insulin is delayed until bedtime to increase the amount of insulin available the next morning. C. A regimen that incorporates ultralente or glargine insulin. D. A variation that includes premeal short-acting insulin with intermediate-acting insulin at breakfast and bedtime. E. Patterns of insulin administration with a regimen of continuous subcutaneous insulin infusion.

An alternative regimen that is gaining widespread use involves multiple daily injections consisting of basal administration of an intermediateor long-acting insulin (either before breakfast or bedtime or both) and preprandial injections of a short-acting insulin (Figure 615C). This dosage regimen is very similar to the pattern of insulin administration achieved with a subcutaneous infusion pump (Figure 615E), except that with a pump it is possible to control and vary the basal rate of insulin infusion more precisely (Kitabchi et al., 1983). In all patients, the exact dose of insulin is chosen by monitoring therapeutic endpoints carefully. This approach is facilitated by the use of home glucose monitors and measurements of hemoglobin A1c concentrations. Special care must be taken when the patient has other underlying diseases, deficiencies in other endocrine systems (e.g., adrenocortical or pituitary failure), or substantial resistance to insulin.

Factors That Affect Insulin Absorption The degree of control of plasma glucose concentrations may be modified by changes in insulin absorption, factors that alter insulin action, diet, exercise, and other factors, many of which are probably undefined. Factors that determine the rate of absorption of insulin after subcutaneous administration include the site of injection, the type of insulin, subcutaneous blood flow, regional muscular activity at the site of the injection, the volume and concentration of the injected insulin, and depth of injection (insulin will have a more rapid onset of action if delivered intramuscularly rather than subcutaneously). When insulin is injected subcutaneously, there can be an initial "lag phase" followed by a slow but steadily increasing rate of absorption. The initial lag phase almost disappears when a reduced concentration or volume of insulin is injected. Insulin usually is injected into the subcutaneous tissues of the abdomen, buttock, anterior thigh, or dorsal arm. Absorption is usually most rapid from the abdominal wall, followed by the arm, buttock, and thigh (Galloway et al., 1981). Rotation of insulin injection sites traditionally has been advocated to avoid lipohypertrophy or lipoatrophy, although these conditions are less likely to occur with highly purified preparations of insulin. If a patient is willing to inject into the abdominal area, injections can be rotated throughout the entire area, thereby eliminating the injection site as a cause of variability in the rate of absorption. The abdomen currently is the preferred site of injection in the morning, as insulin is absorbed about 20% to 30% faster from that site than from the arm. If the patient refuses to inject into the abdominal area, it is preferable to select a consistent injection site for each component of insulin treatment (e.g., prebreakfast dose into the thigh, evening dose into the arm). Several other factors may affect the absorption of insulin. Increased subcutaneous blood flow (brought about by massage, hot baths, and exercise) increases the rate of absorption. In the upright posture, subcutaneous blood flow diminishes considerably in the legs and to a lesser extent in the abdominal wall. An altered volume or concentration of injected insulin affects the rate of absorption and the duration of action. When regular insulin is mixed with lente insulin, some of the regular insulin becomes modified, causing a partial loss of the rapidly acting component (Galloway et al., 1981). This problem is even more severe if regular insulin is mixed with ultralente insulin. Injections of mixtures of insulin preparations thus should be made without delay. There is less delay in absorption of regular insulin when it is mixed with NPH insulin. Stable, mixed combinations of NPH and regular insulin in proportions of 50:50, 60:40, 70:30, and 80:20, respectively, are commercially available; in the United States only the 70:30 and 50:50 combinations are available. Combinations of lispro and NPH insulin also are available in the United States (Table 614). "Pen devices" containing prefilled regular, lispro, NPH, or premixed regular/NPH or lispro/NPH insulin have proven to be popular with many diabetic patients. In a small group of patients, subcutaneous degradation of insulin has been observed, and this has necessitated the injection of large amounts of insulin for adequate metabolic control (Schade and Duckworth, 1986). Jet injector systems that enable patients to receive subcutaneous insulin"injections" without a needle are available. These devices are rather expensive and cumbersome, but they are preferred by a small number of patients. Dispersal of insulin throughout an area of subcutaneous tissue should increase the rate of absorption of both regular and intermediate insulins (Malone et al., 1986); this result has not always been observed, however (Galloway et al., 1981). Subcutaneous insulin administration results in anti-insulin IgG antibody formation. Older, impure preparations of animal insulins resulted in far greater antibody production than do the more recent

purified porcine or bovine and recombinant human preparations. It is disputed whether or not chronic therapy with human insulin reduces antibody production compared to monocomponent porcine insulin. Regardless, it is clear that human insulin is immunogenic. In the vast majority of patients receiving insulin treatment, circulating anti-insulin antibodies do not alter the pharmacokinetics of the injected hormone. In rare patients who have a high titer of anti-insulin antibodies, the kinetics of action of regular insulin may resemble those of an intermediate-acting insulin, which itself may become longer acting. Such effects could lead to increased postprandial hyperglycemia (due to decreased action of regular insulin) but nighttime hypoglycemia (due to the prolonged action of intermediate insulin). IgG antibodies can cross the placenta, raising the possibility that anti-insulin antibodies could cause fetal hyperglycemia by neutralizing fetal insulin. On the other hand, fetal or neonatal hypoglycemia could result from the undesirable and unpredictable release of insulin from insulinantibody complexes. Switching from bovine/porcine to monocomponent insulin preparations has been shown to reduce anti-insulin antibodies, leading to the recommendation that only human insulin be used during pregnancy (Chertow et al., 1988). Continuous Subcutaneous Insulin Infusion A number of pumps are available for continuous subcutaneous insulin infusion (CSII) therapy (Kitabchi et al., 1983). CSII or "pump" therapy is not suitable for all patients, since it demands considerable attention, especially during the initial phases of treatment. However, for patients interested in intensive insulin therapy, a pump may be an attractive alternative to several daily injections. Most modern pumps provide a constant basal infusion of insulin and have the option of different infusion rates during the day and night to help avoid the dawn phenomenon and bolus injections that are programmed according to the size and nature of a meal. Pump therapy presents some unique problems. Since all of the insulin used is short-acting and there is a minimal amount of insulin in the subcutaneous pool at any given time, insulin deficiency and ketoacidosis with unexpected high levels of potassium may develop rapidly if therapy is accidentally interrupted. Although modern pumps have warning devices that detect changes in line pressure, mechanical problems such as pump failure, dislodgement of the needle, aggregation of insulin in the infusion line, or accidental kinking of the infusion catheter may occur. There also is a possibility of subcutaneous abscesses and cellulitis. Selection of the most appropriate patients is extremely important for success with pump therapy. Offsetting the above potential problems, pump therapy is capable of producing a more physiological profile of insulin replacement during exercise (where insulin production is decreased) and therefore less hypoglycemia than do traditional subcutaneous insulin injections. Adverse Reactions Hypoglycemia The most common adverse reaction to insulin is hypoglycemia. This may result from an inappropriately large dose, from a mismatch between the time of peak delivery of insulin and food intake, or from superimposition of additional factors that increase sensitivity to insulin (adrenal insufficiency, pituitary insufficiency) or that increase insulin-independent glucose uptake (exercise). The more vigorous the attempt to achieve euglycemia, the more frequent the episodes of hypoglycemia. In the Diabetes Control Complications Trial, the incidence of severe hypoglycemic reactions was three times higher in the intensive insulin therapy group than in the conventional

therapy group (DCCT Research Group, 1993). Milder but significant hypoglycemic episodes were much more common than were severe reactions, and their frequency also increased with intensive therapy. Hypoglycemia is the major risk that must be weighed against any benefits of intensive therapy. There is a hierarchy of physiological responses to hypoglycemia. The initial response is a reduction of endogenous insulin secretion, following which, at a plasma glucose level of about 70 mg/dl (3.9 mM), the counterregulatory hormonesepinephrine, glucagon, growth hormone, cortisol, and norepinephrineare released. The symptoms of hypoglycemia are first discerned at a plasma glucose level of 60 to 80 mg/dl (3.3 to 3.9 mM). Sweating, hunger, paresthesias, palpitations, tremor, and anxiety, principally of autonomic origin, usually are seen first. Difficulty in concentrating, confusion, weakness, drowsiness, a feeling of warmth, dizziness, blurred vision, and loss of consciousness are referred to as neuroglycopenic symptoms and usually occur at lower plasma glucose levels than do autonomic symptoms. In a normal individual, plasma glucose levels are tightly regulated, and it is only under rare conditions that hypoglycemia occurs. Glucagon is the predominant counterregulatory hormone in acute hypoglycemia in newly diagnosed type 1 DM patients and normal human beings. When hypoglycemia is prolonged, catecholamines, cortisol, and growth hormone become more important. In subjects with type 1 DM of longer duration, the glucagon secretory response to hypoglycemia becomes deficient, but effective glucose counterregulation still occurs because epinephrine plays a compensatory role. Type 1 DM subjects thus become dependent on epinephrine for counterregulation, and if this mechanism becomes deficient, the incidence of severe hypoglycemia increases. This occurs in patients with diabetes of long duration who have autonomic neuropathy. The absence of both glucagon and epinephrine can lead to prolonged hypoglycemia, particularly during the night, when some individuals can have extremely low plasma glucose for several hours. Severe hypoglycemia can lead to convulsions and coma. In addition to autonomic neuropathy, several related syndromes of defective counterregulation contribute to the increased incidence of severe hypoglycemia in intensively treated type 1 DM patients. These include hypoglycemic unawareness, altered thresholds for release of counterregulatory hormones, and deficient secretion of counterregulatory hormones (reviewed by Cryer 1992, 1993). With the ready availability of home glucose monitoring, hypoglycemia can be documented in most patients who experience suggestive symptoms. Hypoglycemia that occurs during sleep may be difficult to detect but should be suspected from a history of morning headaches, night sweats, or symptoms of hypothermia. Nocturnal hypoglycemia has been proposed as a cause of morning hyperglycemia in type 1 DM patients. This syndrome, known as the Somogyi phenomenon, is reputedly due to an elevation of counterregulatory hormones in response to nocturnal hypoglycemia. The existence of the Somogyi phenomenon recently has been questioned, as several groups of investigators have not been able to reproduce it. Moreover, neuroendocrine counterregulatory responses now are known to be severely diminished with disease duration and intensive control. Therefore, it is unlikely that, in patients with reduced neuroendocrine responses to hypoglycemia, nocturnal counterregulatory responses to hypoglycemia could be responsible for morning hyperglycemia. The practice of reducing nighttime insulin doses in type 1 DM subjects with morning hyperglycemia thus cannot now be recommended. It is more likely that a reduced action of injected intermediate-acting insulin that occurs in concert with the dawn phenomenon is the cause of morning hyperglycemia. The current recommended therapeutic approach to treating morning hyperglycemia is to administer more intermediate acting insulin the night before, perhaps at bedtime, or to increase the basal rate of a CSII pump between the hours of 3 and 7 A.M.

All diabetic patients who receive insulin should be aware of the symptoms of hypoglycemia, carry some form of easily ingested glucose, and carry an identification card or bracelet containing pertinent medical information. When possible, patients who suspect that they are experiencing hypoglycemia should document the glucose concentration with a measurement. Mild to moderate hypoglycemia may be treated simply by ingestion of glucose. When hypoglycemia is severe, it should be treated with intravenous glucose or an injection of glucagon (see below). Insulin Allergy and Resistance Although there has been a dramatic decrease in the incidence of resistance and allergic reactions to insulin with the use of human insulin or highly purified preparations of the hormone, these reactions still occur as a result of reactions to the small amounts of aggregated or denatured insulin in all preparations, to minor contaminants, or because of sensitivity to one of the components added to insulin in its formulation (protamine, Zn2+, phenol, etc.). The most frequent allergic manifestations are IgE-mediated local cutaneous reactions, although on rare occasions patients may develop lifethreatening systemic responses or insulin resistance due to IgG antibodies (Kahn and Rosenthal, 1979). Attempts should be made to identify the underlying cause of the hypersensitivity response by measuring insulin-specific IgG and IgE antibodies. Skin testing also is useful; however, many patients exhibit positive reactions to intradermal insulin without experiencing any adverse effects from subcutaneous insulin. If patients have allergic reactions to mixed bovine/porcine insulin, human insulin should be used. If allergy persists, desensitization may be attempted; it is successful in about 50% of cases. Antihistamines may provide relief in patients with cutaneous reactions, while glucocorticoids have been used in patients with resistance to insulin or more severe systemic reactions. Lipoatrophy and Lipohypertrophy Atrophy of subcutaneous fat at the site of insulin injection (lipoatrophy) is probably a variant of an immune response to insulin, whereas lipohypertrophy (enlargement of subcutaneous fat depots) has been ascribed to the lipogenic action of high local concentrations of insulin (LeRoith et al., 2000). Both of these problems may be related to some contaminant in insulin; they are rare with more purified preparations. However, hypertrophy occurs frequently with human insulins if patients inject themselves repeatedly in the same site. When these problems occur, they may cause irregular absorption of insulin as well as a cosmetic problem. The recommended treatment is to avoid the hypertrophic areas by using other injection sites, and to inject insulin into the periphery of the atrophic sites in an attempt to restore the subcutaneous adipose tissue. Insulin Edema Some degree of edema, abdominal bloating, and blurred vision develops in many diabetic patients with severe hyperglycemia or ketoacidosis that is brought under control with insulin (Wheatley and Edwards, 1985). This is associated with a weight gain of 0.5 to 2.5 kg. The edema usually disappears spontaneously within several days to a week unless there is underlying cardiac or renal disease. Edema is attributed primarily to retention of Na+, although increased capillary permeability associated with inadequate metabolic control also may contribute. Insulin Treatment of Ketoacidosis and Other Special Situations Acutely ill diabetic patients may have metabolic disturbances that are sufficiently severe or labile to justify intravenous administration of insulin. Such treatment is most appropriate in patients with ketoacidosis (Schade and Eaton, 1983; Kitabchi, 1989). Although there has been some controversy

over appropriate dosage, infusion of a relatively low dose of insulin (0.1 U/kg per hour) will produce plasma concentrations of insulin of about 100 U/mla level sufficient to inhibit lipolysis and gluconeogenesis completely and to produce near-maximal stimulation of glucose uptake in normal individuals. In most patients with ketoacidosis, blood glucose concentrations will fall by about 10% per hour; the acidosis is corrected more slowly. As treatment proceeds, it may be necessary to administer glucose along with the insulin to prevent hypoglycemia but to allow clearance of all ketones. Some physicians prefer to initiate therapy with a loading dose of insulin, but this tactic appears unnecessary as steady-state concentrations of the hormone are achieved within 30 minutes with a constant infusion. Patients with nonketotic, hyperglycemic coma frequently are more sensitive to insulin than are those with ketoacidosis. Appropriate replacement of fluid and electrolytes is an integral part of the therapy in both situations, since there is always a major deficit. Regardless of the exact insulin regimen used, the key to effective therapy is careful and frequent monitoring of the patient's clinical status, glucose, and electrolytes. A frequent error in the management of such patients is the failure to administer insulin subcutaneously at least 30 minutes before intravenous therapy is discontinued. This is necessary because of the very short halflife of insulin. Intravenous administration of insulin also is well suited to the treatment of diabetic patients during the perioperative period and during childbirth. There is debate, however, about the optimum route of insulin administration during surgery. Although some clinicians advocate subcutaneous insulin administration, more now recommend intravenous insulin infusion. The two most widely used protocols for intravenous insulin administration are the variable-rate regimen (Watts et al., 1987) and the glucose, insulin, and potassium infusion (GIK) method (Thomas et al., 1984). Both protocols provide stable plasma glucose, fluid, and electrolyte levels during the operative and postoperative period. Despite these recommendations, many physicians give patients half of their normal daily dose of insulin as intermediate-acting insulin subcutaneously on the morning before an operation, and then administer 5%dextrose infusions during surgery to maintain glucose concentrations. Although this may be satisfactory in some patients, use of an insulin with an intermediate duration of action provides less minute-to-minute control than is possible with an intravenous regimen. The limited data available on this subject indicate that intravenous regimens are superior to subcutaneous insulin injection in patients undergoing surgery. Drug Interactions and Glucose Metabolism A large number of drugs can cause hypoglycemia or hyperglycemia or may alter the response of diabetic patients to their existing therapeutic regimens (see Koffler et al., 1989; Seltzer, 1989). Some drugs with hypoglycemic or hyperglycemic effects and their presumed sites of action is given in Table 615. Aside from insulin and oral hypoglycemic drugs, the most common drug-induced hypoglycemic states are those caused by ethanol, -adrenergic receptor antagonists, and salicylates. The primary action of ethanol is to inhibit gluconeogenesis. This effect is not an idiosyncratic reaction but is observed in all individuals. In diabetic patients, -adrenergic receptor antagonists pose a risk of hypoglycemia because of their capacity to inhibit the effects of catecholamines on gluconeogenesis and glycogenolysis. These agents also may mask the sympathetically mediated symptoms associated with the fall in blood glucose (e.g., tremor and palpitations). Salicylates, on the other hand, exert their hypoglycemic effect by enhancing pancreatic -cell sensitivity to glucose and potentiating insulin secretion. These agents also have a weak insulin-like action in the periphery. Pentamidine, an antiprotozoal agent now used frequently for the treatment of infections caused by Pneumocystis carinii, apparently can cause both hypoglycemia and hyperglycemia. The hypoglycemic effect results from destruction of cells and release of insulin; continuation of use

may cause secondary hypoinsulinemia and hyperglycemia. An equally large number of drugs may cause hyperglycemia in normal individuals or impair metabolic control in diabetic patients. Many of these are agents with direct effects on peripheral tissues that counter the actions of insulin; examples include epinephrine and glucocorticoids. Other drugs cause hyperglycemia by inhibiting insulin secretion directly (phenytoin, clonidine, Ca2+ channel blockers) or indirectly via depletion of K+ (diuretics). A number of drugs have no direct hypoglycemic action but may potentiate the actions of sulfonylureas (see below). It is important to be aware of such interactions and to modify treatment regimens for diabetic patients accordingly. New Forms of Insulin Therapy There are a number of experimental approaches to delivery of insulin, including the use of new insulins, new routes of administration, intraperitoneal delivery devices, implantable pellets, the closed-loop artificial pancreas, islet-cell and pancreatic transplantation, and gene therapy. New Routes of Delivery Attempts have been made to administer insulin orally, nasally, rectally, by inhalation, and by subcutaneous implantation of pellets. The most promising of these alternatives is by inhalation, which can be achieved by addition of various adjuvants such as mannitol, glycine, and sodium citrate to insulin to increase its absorption through the pulmonary mucosa (Skyler et al., 2001; Cefalu et al., 2001). The kinetics of absorption are rapid and approach the rate achieved with subcutaneous administration of regular insulin. Further work is under way with the aim of reducing the size and increasing the convenience of the inhaled delivery systems. Implantable pellets have been designed to release insulin slowly over days or weeks. Although oral delivery of insulin would be preferred by patients and would provide higher relative concentrations of insulin in the portal circulation, attempts to increase intestinal absorption of the hormone have met with only limited success. Efforts have focused on protection of insulin by encapsulation or incorporation into liposomes. Intraperitoneal infusion of insulin into the portal circulation has been used experimentally in human subjects for periods of several months. Transplantation and Gene Therapy Transplantation and gene therapy are provocative approaches to replacement of insulin. Segmental pancreatic transplantation has been employed successfully in several hundred patients (Sutherland et al., 1989). However, the surgery is technically complex and usually is considered only in patients with advanced disease and complications. Islet-cell transplants are theoretically less complicated. They have been accomplished in experimental rodent models of diabetes, and recently in a small group of type 1 DM patients along with a novel glucocorticoid-free immunosuppressive regimen (Shapiro et al., 2000). Introduction of an active insulin gene into cells such as fibroblasts, which can then be reintroduced into the host, also has been achieved in rodents. Oral Hypoglycemic Agents History In contrast to the systematic studies that led to the isolation of insulin, the sulfonylureas were discovered accidentally. In 1942, Janbon and colleagues noted that some sulfonamides caused hypoglycemia in experimental animals. These observations were soon extended, and 1-butyl-3sulfonylurea (carbutamide) became the first clinically useful sulfonylurea for the treatment of

diabetes. This compound was later withdrawn because of adverse effects on the bone marrow, but it led to the development of the entire class of sulfonylureas. Clinical trials of tolbutamide, the first widely used member of this group, were instituted in type 2 DM patients in the early 1950s. Since that time, approximately 20 different agents of this class have been in use worldwide. In 1997, the first member of a new class of oral insulin secretagogues called meglitinides (benzoic acid derivatives) was approved for clinical use. This agent, repaglinide, has gained acceptance as a fast-acting, premeal therapy to limit postprandial hyperglycemia. A plant (Galega officinalis, goat's rue) used to treat diabetes in Europe in medieval times was found in the early part of this century to contain guanadine. Guanadine has hypoglycemic properties but is too toxic for clinical use. During the 1920s, biguanides were investigated for use in diabetes, but they were overshadowed by the discovery of insulin. Later, the antimalarial agent chloroguanide was found to have weak hypoglycemic action. Shortly after the introduction of the sulfonylureas, the first biguanides became available for clinical use. However, phenformin, the primary drug in this group, was withdrawn from the market in the United States and Europe because of an increased frequency of lactic acidosis associated with its use. Another biguanide, metformin, has been used extensively in Europe without significant adverse effects and was approved for use in the United States in 1995. Thiazolidinediones were introduced in 1997 as the second major class of "insulin sensitizers." These agents bind to peroxisome proliferatoractivated receptors (principally PPAR ), resulting in increased glucose uptake in muscle and reduced endogenous glucose production. The first of these agents, troglitazone, was withdrawn from use in the United States in 2000 because of an association with hepatic toxicity. Two other agents of this class, rosiglitazone and pioglitazone, have not been associated with widespread liver toxicity and are used worldwide. Sulfonylureas Chemistry The sulfonylureas are divided traditionally into two groups or generations of agents. Their structural relationships are shown in Table 616. All members of this class of drugs are substituted arylsulfonylureas. They differ by substitutions at the para position on the benzene ring and at one nitrogen residue of the urea moiety. The first group of sulfonylureas includes tolbutamide, acetohexamide, tolazamide, and chlorpropamide. A second generation of hypoglycemic sulfonylureas has emerged. These drugs [glyburide (glibenclamide), glipizide, gliclazide, and glimepiride] are considerably more potent than the earlier agents. Mechanism of Action Sulfonylureas cause hypoglycemia by stimulating insulin release from pancreatic cells. Their effects in the treatment of diabetes, however, are more complex. The acute administration of sulfonylureas to type 2 DM patients increases insulin release from the pancreas. Sulfonylureas also may further increase insulin levels by reducing hepatic clearance of the hormone. In the initial months of sulfonylurea treatment, fasting plasma insulin levels and insulin responses to oral glucose challenges are increased. With chronic administration, circulating insulin levels decline to those that existed before treatment, but, despite this reduction in insulin levels, reduced plasma glucose levels are maintained. The explanation for this is not clear, but it may relate to reduced plasma glucose allowing circulating insulin to have more pronounced effects on its target tissues, and to the fact

that chronic hyperglycemia per se impairs insulin secretion (glucose toxicity). It should be noted that there is no measurable acute stimulatory effect of sulfonylureas on insulin secretion during chronic treatment. This is thought to be due to downregulation of cell-surface receptors for sulfonylureas on the pancreatic cell. If chronic sulfonylurea therapy is discontinued, pancreatic -cell responsiveness to acute administration of the drug is restored. Sulfonylureas also stimulate release of somatostatin, and they may suppress the secretion of glucagon slightly (Krall, 1985). The effects of the sulfonylureas are initiated by binding to and blocking an ATP-sensitive K + channel, which has been cloned (Aguilar-Bryan et al., 1995; Philipson and Steiner, 1995). The drugs thus resemble physiological secretagogues (e.g., glucose, leucine), which also lower the conductance of this channel (Ribalet and Ciani, 1987; Boyd, 1988). Reduced K+ conductance causes membrane depolarization and influx of Ca2+ through voltage-sensitive Ca2+ channels. There has been controversy about whether or not sulfonylureas have clinically significant extrapancreatic effects (Beck-Nielsen, 1988). The concentration of insulin receptors increases in the monocytes, adipocytes, and erythrocytes of type 2 DM patients who receive oral hypoglycemic agents (Olefsky and Reaven, 1976). Sulfonylureas enhance insulin action in cells in culture and stimulate the synthesis of glucose transporters (Jacobs et al., 1989). Sulfonylureas also have been shown to suppress hepatic gluconeogenesis (Blumenthal, 1977); however, it is not clear if this is a direct effect of the drug or a reflection of increased sensitivity to insulin. In general, attempts to ascribe the long-term blood glucose-lowering effects of sulfonylureas to specific changes in insulin action on target tissues are confounded by the effects of a lowered prevailing blood glucose level. Although extrapancreatic effects of sulfonylureas can be demonstrated, they are of minor clinical significance in the treatment of type 2 DM patients. Absorption, Fate, and Excretion The sulfonylureas have similar spectra of activities; thus, their pharmacokinetic properties are their most distinctive characteristics (see Appendix II). Although there are differences in the rates of absorption of the different sulfonylureas, all are effectively absorbed from the gastrointestinal tract. However, food and hyperglycemia can reduce the absorption of sulfonylureas. (Hyperglycemia per se inhibits gastric and intestinal motility and thus can retard the absorption of many drugs.) In view of the time required to reach an optimal concentration in plasma, sulfonylureas with short half lives may be more effective when given 30 minutes before eating. Sulfonylureas in plasma are largely (90% to 99%) bound to protein, especially albumin; plasma protein binding is least for chlorpropamide and greatest for glyburide. The volumes of distribution of most of the sulfonylureas are about 0.2 liter/kg. The first-generation sulfonylureas vary considerably in their half-lives and extents of metabolism. The half-life of acetohexamide is short, but the drug is reduced to an active compound with a halflife that is similar to those of tolbutamide and tolazamide (4 to 7 hours). It may be necessary to take these drugs in divided daily doses. Chlorpropamide has a long half-life (24 to 48 hours). The second-generation agents are approximately 100 times more potent than are those in the first group (Lebovitz and Feinglos, 1983). Although their half-lives are short (3 to 5 hours), their hypoglycemic effects are evident for 12 to 24 hours, and it is often possible to administer them once daily. The reason for the discrepancy between the half-life and duration of action of these drugs is not clear. All of the sulfonylureas are metabolized by the liver, and the metabolites are excreted in the urine. Metabolism of chlorpropamide is incomplete, and about 20% of the drug is excreted unchanged.

Thus, sulfonylureas should be administered with caution to patients with either renal or hepatic insufficiency. Adverse Reactions Adverse effects of the sulfonylureas are infrequent, occurring in about 4% of patients taking firstgeneration drugs and perhaps slightly less often in patients receiving second-generation agents (Paice et al., 1985). Not unexpectedly, sulfonylureas may cause hypoglycemic reactions, including coma (Ferner and Neil, 1988; Seltzer, 1989). This is a particular problem in elderly patients with impaired hepatic or renal function who are taking longer-acting sulfonylureas. Sulfonylureas can be ranked in order of decreasing risk of causing hypoglycemia based on their half-lives. The longer the half-life, the more likely an agent will induce hypoglycemia. Severe hypoglycemia in the elderly can present as an acute neurological emergency that may mimic a cerebrovascular accident. Thus, it is important to check the plasma glucose of any elderly patient presenting with acute neurological symptoms. Because of the long half-life of some sulfonylureas, it may be necessary to treat an elderly hypoglycemic patient for 24 to 48 hours with an intravenous glucose infusion. A number of other drugs may potentiate the effects of the sulfonylureas, particularly the firstgeneration agents, by inhibiting their metabolism or excretion. Some drugs also displace the sulfonylureas from binding proteins, thereby increasing the free concentration transiently (Seltzer, 1989). These include other sulfonamides, clofibrate, dicumarol, salicylates, and phenylbutazone. Other drugs, including ethanol, may enhance the action of sulfonylureas by causing hypoglycemia. Other side effects of sulfonylureas include nausea and vomiting, cholestatic jaundice, agranulocytosis, aplastic and hemolytic anemias, generalized hypersensitivity reactions, and dermatological reactions. About 10% to 15% of patients who receive these drugs, particularly chlorpropamide, develop an alcohol-induced flush similar to that caused by disulfiram (see Chapter 18: Ethanol). Sulfonylureas, especially chlorpropamide, also may induce hyponatremia by potentiating the effects of antidiuretic hormone on the renal collecting duct (Paice et al., 1985). This undesirable side effect occurs in up to 5% of all patients; it is less frequent with glyburide and glipizide. This side effect has been used to therapeutic advantage in patients with mild forms of diabetes insipidus (see Chapter 29: Diuretics). A long-running debate centered around whether or not treatment with sulfonylureas is associated with increased cardiovascular mortality; this possibility was suggested by a large multicenter trial (the University Group Diabetes Program or UGDP). The UGDP was designed to compare the effect of diet, oral agents (tolbutamide or phenformin), and fixed-dose insulin therapy on the development of vascular complications in type 2 DM. During an 8-year period of observation, patients who received tolbutamide had a twofold higher rate of cardiovascular death than patients treated with placebo or insulin (Meinert et al., 1970). A 10-year debate followed on the validity of this conclusion, because the observation was unexpected, the study had not been designed to test this question, and all of the excess mortality occurred in only three centers. The recent UK Prospective Diabetes Study Group (UK Prospective Diabetes Study Group, 1998a), however, clearly demonstrated no excess cardiovascular mortality over a 14-year period in patients receiving firstor second-generation sulfonylureas. Therapeutic Uses Sulfonylureas are used to control hyperglycemia in type 2 DM patients who cannot achieve appropriate control with changes in diet alone. In all patients, however, continued dietary restrictions are essential to maximize the efficacy of the sulfonylureas. Some physicians still

consider treatment with insulin to be the preferred approach in such patients. Patients with type 2 DM whose disease is controlled with relatively low doses of insulin (less than 40 U per day) are more likely to respond to sulfonylureas, as are those who are obese and/or older than 40 years of age. Contraindications to the use of these drugs include type 1 DM, pregnancy, lactation, and significant hepatic or renal insufficiency. Between 50% and 80% of properly selected patients will respond initially to an oral hypoglycemic agent (Krall, 1985). All of the drugs appear to be equally efficacious. Concentrations of glucose often are lowered sufficiently to relieve symptoms of hyperglycemia, but they may not reach normal levels. To the extent that complications of diabetes may be related to hyperglycemia, the goal of treatment should be normalization of both fasting and postprandial glucose concentrations. About 5% to 10% of patients per year who respond initially to a sulfonylurea become secondary failures, as defined by unacceptable levels of hyperglycemia. This may occur as a result of a change in drug metabolism, progression of -cell failure, change in dietary compliance, or misdiagnosis of a patient with slow-onset type 1 DM. Additional oral agent(s) can produce a satisfactory response, but most of these patients will eventually require insulin. The usual initial daily dose of tolbutamide is 500 mg, while 3000 mg is the maximally effective total dose; corresponding doses for acetohexamide are 250 and 1500 mg. Tolazamide and chlorpropamide usually are administered in a daily dose of 100 to 250 mg, while 1000 (tolazamide) to 750 mg (chlorpropamide) is maximal. Tolbutamide, acetohexamide, and tolazamide often are taken twice daily, 30 minutes before breakfast and dinner. The initial daily dose of glyburide is 2.5 to 5 mg, while daily doses of more than 20 mg are not recommended. Therapy with glipizide usually is initiated with 5 mg given once daily. The maximal recommended daily dose is 40 mg; daily doses of more than 15 mg should be divided. The starting dose of gliclazide is 40 to 80 mg per day, and the maximal daily dose is 320 mg. Glimepiride therapy can begin with doses as low as 0.5 mg once per day. The maximal effective daily dose of the agent is 8 mg. Treatment with the sulfonylureas must be guided by the individual patient's response, which must be monitored frequently. Combinations of insulin and sulfonylureas have been used in some patients with type 1 DM and type 2 DM. Studies in type 1 DM patients have not provided any evidence that glucose control is improved by combination therapy. The results in type 2 DM patients are more provocative but inconclusive. Some studies have revealed no benefits with combination therapy, while others have shown an improvement in metabolic control. A prerequisite for a beneficial effect of combination therapy is residual -cell activity, and a short duration of diabetes also has been suggested to predict a good response. Repaglinide Repaglinide (PRANDIN) is an oral insulin secretagogue of the meglitinide class. This agent is a derivative of benzoic acid, and its structure (shown below) is unrelated to that of the sulfonylureas.

However, like sulfonylureas, repaglinide stimulates insulin release by closing ATP-dependent potassium channels in pancreatic cells. The drug is absorbed rapidly from the gastrointestinal tract; peak blood levels are obtained within one hour. The half-life of the drug is about one hour. These features of the drug allow for multiple preprandial use, as compared to the classical onceor twicedaily dosing of sulfonylureas. Repaglinide is metabolized primarily by the liver. Metabolites of the drug do not have a hypoglycemic action. Repaglinide should be used cautiously in patients with hepatic insufficiency. Because a small proportion (about 10%) of repaglinide is metabolized by the kidney, increased dosing of the drug in patients with renal insufficiency also should be performed cautiously. As with sulfonylureas, the major side effect of repaglinide is hypoglycemia. Nateglinide Nateglinide (STARLIX) is an orally effective insulin secretagogue derived from D-phenylalanine. Like sulfonylureas and repaglinide, nateglinide stimulates insulin secretion by blocking ATPsensitive potassium channels in pancreatic cells. Nateglinide promotes a more rapid but less sustained secretion of insulin than do other available oral antidiabetic agents (Kalbag et al., 2001). The drug's major therapeutic effect is reducing postprandial glycemic elevations in type 2 diabetic patients. Nateglinide recently has been approved by the United States Food and Drug Administration (FDA) for use in type 2 DM and is most effective if administered 1 to 10 minutes before a meal in a dose of 120 mg. Nateglinide is metabolized primarily by the liver and thus should be used cautiously in patients with hepatic insufficiency. About 16% of an administered dose is excreted by the kidney as unchanged drug. Dosage adjustment is unnecessary in renal failure. Early studies have suggested that nateglinide therapy may produce fewer episodes of hypoglycemia than do other currently available oral insulin secretagogues (Horton et al., 2001). Biguanides Metformin (GLUCOPHAGE ) and phenformin were introduced in 1957 and buformin was introduced in 1958. The latter was of limited use, but metformin and phenformin were widely used. Phenformin was withdrawn in many countries during the 1970s because of an association with lactic acidosis. Metformin has been associated only rarely with that complication and has been widely used in Europe and Canada; it became available in the United States in 1995. Metformin given alone or in combination with a sulfonylurea improves glycemic control and lipid concentrations in patients who respond poorly to diet or to a sulfonylurea alone (DeFronzo et al., 1995). Metformin is absorbed mainly from the small intestine. The drug is stable, does not bind to plasma proteins, and is excreted unchanged in the urine. It has a half-life of about 2 hours. The maximum

recommended daily dose of metformin in the United States is 2.5 g, taken in three doses with meals. Metformin is antihyperglycemic, not hypoglycemic (see Bailey, 1992). It does not cause insulin release from the pancreas and does not cause hypoglycemia, even in large doses. Metformin has no significant effects on the secretion of glucagon, cortisol, growth hormone, or somatostatin. Metformin reduces glucose levels primarily by decreasing hepatic glucose production and by increasing insulin action in muscle and fat. The mechanism by which metformin reduces hepatic glucose production is controversial, but the preponderance of data indicates an effect on reducing gluconeogenesis (Stumvoll et al., 1995). Metformin also may decrease plasma glucose by reducing the absorption of glucose from the intestine, but this action has not been shown to have clinical relevance. Patients with renal impairment should not receive metformin. Hepatic disease, a past history of lactic acidosis (of any cause), cardiac failure requiring pharmacological therapy, or chronic hypoxic lung disease also are contraindications to the use of the drug. The drug also should be withheld for 48 hours after administration of intravenous contrast media. The drug should not be readministered until renal function is normal. These conditions all predispose to increased lactate production and hence to the fatal complications of lactic acidosis. The reported incidence of lactic acidosis during metformin treatment is lower than 0.1 case per 1000 patient years, and the mortality risk is even lower. Acute side effects of metformin, which occur in up to 20% of patients, include diarrhea, abdominal discomfort, nausea, metallic taste, and anorexia. These usually are minimized by increasing the dosage of the drug slowly and taking it with meals. Intestinal absorption of vitamin B12 and folate often is decreased during chronic metformin therapy. Calcium supplements reverse the effect of metformin on vitamin B12 absorption (Bauman et al., 2000). Consideration should be given to stopping treatment with metformin if the plasma lactate level exceeds 3 mM. Similarly, decreased renal or hepatic function also may be a strong indication for withholding treatment. It also would be prudent to stop metformin if a patient is undergoing a prolonged fast or is treated with a very-low-calorie diet. Myocardial infarction or septicemia mandate stopping the drug immediately. Metformin usually is administered in divided doses either two or three times daily. The maximum effective dose is 2.5 g daily. Metformin lowers hemoglobin A1c values to a similar extent as do sulfonylureas (about 2.0%). Metformin does not promote weight gain and can reduce plasma triglycerides by 15% to 20%. There is a strong consensus that reduction in hemoglobin A1c by any therapy (insulin or oral agents) can lead to diminished microvascular complications. Metformin, however, is the only therapeutic agent that has been demonstrated to reduce macrovascular events in type 2 DM (UK Prospective Diabetes Study Group, 1998b). Metformin can be administered in combination with sulfonylureas, thiazolizinediones, and/or insulin. A fixed-combination tablet containing glyburide (glibenclamide) and metformin (GLUCOVANCE ) is available. Thiazolidinediones Three thiazolidinediones have been used in clinical practice (troglitazone, rosiglitazone, and pioglitazone). However, the first of these agents to be introduced (troglitazone) has been withdrawn from use because it was associated with severe hepatic toxicity. The structures of rosiglitazone and pioglitazone are shown below.

Thiazolidinediones are selective agonists for nuclear peroxisome proliferatoractivated receptorgamma (PPAR ). These drugs bind to PPAR , which, in turn, activates insulin-responsive genes that regulate carbohydrate and lipid metabolism. Thiazolidinediones require insulin to be present for their action. Thiazolidinediones exert their principal effects by lowering insulin resistance in peripheral tissue, but an effect to lower glucose production by the liver also has been reported. Thiazolidinediones increase glucose transport into muscle and adipose tissue by enhancing the synthesis and translocation of specific forms of the glucose transporter proteins. The thiazolidinediones also can activate genes that regulate free fatty-acid metabolism in peripheral tissue. Studies are in progress to determine if these agents reduce insulin resistance primarily by their actions on free fatty-acid metabolism. Rosiglitazone (AVANDIA) and pioglitazone (ACTOS ) are taken once a day. Both agents are absorbed within about 2 hours, but the maximum clinical effect is not observed for 6 to 12 weeks. The thiazolidinediones are metabolized by the liver and may be administered to patients with renal insufficiency, but these agents should not be used if there is active hepatic disease or if there are significant elevations of serum liver transaminases. Regular monitoring of liver funtion should be instituted in patients receiving thiazolidinediones. Thiazolidinediones also have been reported to cause anemia, weight gain, edema, and plasma volume expansion. These drugs generally are not indicated for patients with New York Heart Association class 3 and 4 heart failure. Rosiglitazone and pioglitazone can lower hemoglobin A1c levels by 1.0% to 1.5% in patients with type 2 DM. These drugs can be combined with insulin or other classes of oral glucose-lowering agents. The thiazolidinediones tend to lower triglycerides (10% to 20%) but increase both HDL (up to 19%) and LDL (up to 12%) cholesterol. The increased LDL has been reported to reflect a change in particle size from a dense to a more buoyant, less atherogenic compound. Both available thiazolidinediones are metabolized by cytochrome P450 enzymes in the liver. Rosiglitazone is metabolized by CYP2C8 and pioglitazone by CYP3A4 and CYP2C8. Metabolism by these hepatic enzymes provides the potential for interactions with other classes of drugs that are metabolized via these pathways. To date, no clinically significant interactions have been identified between the available thiazolidinediones and other drug classes, but further studies are in progress. -Glucosidase Inhibitors -Glucosidase inhibitors reduce intestinal absorption of starch, dextrin, and disaccharides by

inhibiting the action of intestinal brush border -glucosidase. Inhibition of this enzyme slows the absorption of carbohydrates; the postprandial rise in plasma glucose is blunted in both normal and diabetic subjects. Acarbose (PRECOSE), an oligosaccharide of microbial origin, and miglitol (GLYSET), a desoxynojirimycin derivative, also competitively inhibit glucoamylase and sucrase but have weak effects on pancreatic -amylase. They reduce postprandial plasma glucose levels in type 1 DM and type 2 DM subjects. -Glucosidase inhibitors can have profound effects on hemoglobin A1c levels in severely hyperglycemic type 2 DM patients. However, in patients with mild to moderate hyperglycemia, the glucose-lowering potential of -glucosidase inhibitors (assessed by hemoglobin A1c levels) is about 30% to 50% of that of other oral antidiabetic agents. -Glucosidase inhibitors do not stimulate insulin release and therefore do not result in hypoglycemia. These agents may be considered as monotherapy in elderly patients or in patients with predominantly postprandial hyperglycemia. -Glucosidase inhibitors typically are used in combination with other oral antidiabetic agents and/or insulin. The drugs should be administered at the start of a meal. They are poorly absorbed. -Glucosidase inhibitors cause dose-related malabsorption, flatulence, diarrhea, and abdominal bloating. Titrating the dose of drug slowly (25 mg at the start of a meal for 4 to 8 weeks followed by increases at 4-to 8-week intervals up to 75 mg before each meal) will reduce gastrointestinal side effects. Smaller doses are given with snacks. Acarbose is most effective when given with a starchy, high-fiber diet with restricted amounts of glucose and sucrose (Bressler and Johnson, 1992). If hypoglycemia occurs when -glucosidase inhibitors are used with insulin or an insulin secretagogue, glucose rather than sucrose, starch, or maltose should be administered. Glucagon History Distinct populations of cells were identified in the islets of Langerhans before the discovery of insulin. Glucagon itself was discovered by Murlin and Kimball in 1923, less than 2 years after the discovery of insulin. In contrast to the excitement caused by the discovery of insulin, few were interested in glucagon, and it was not recognized as an important hormone for more than 40 years. Glucagon is now known to have a significant physiological role in the regulation of glucose and ketone body metabolism, but it is only of minor therapeutic interest for the short-term management of hypoglycemia. It also is used in radiology for its inhibitory effects on intestinal smooth muscle. Chemistry Glucagon is a 29-amino-acid, single-chain polypeptide (Figure 616). It shows significant homology with several other polypeptide hormones, including secretin, vasoactive intestinal peptide, and gastrointestinal inhibitory polypeptide. The primary sequence of glucagon is highly conserved in mammals, and it is identical in human beings, cattle, pigs, and rats. Figure 616. The Amino Acid Sequence of Glucagon.

Glucagon is synthesized from preproglucagon, a 180-aminoacid precursor with five separately processed domains (Bell et al., 1983). An amino-terminal signal peptide is followed by glicentinrelated pancreatic peptide, glucagon, glucagon-like peptide-1, and glucagon-like peptide-2. Processing of the protein is sequential and occurs in a tissue-specific fashion; this results in different secretory peptides in pancreatic cells and intestinal -like cells (termed L cells) (Mojsov et al., 1986). Glicentin, a major processing intermediate, consists of glicentin-related pancreatic polypeptide at the amino terminus and glucagon at the carboxyl terminus, with an Arg-Arg pair between. Enteroglucagon (or oxyntomodulin) consists of glucagon and a carboxyl-terminal hexapeptide linked by an Arg-Arg pair. The biological roles of these precursor peptides are uncertain, but the highly controlled nature of the processing suggests that these peptides may have distinct biological functions. In the pancreatic cell, the granule consists of a central core of glucagon surrounded by a halo of glicentin. Intestinal L cells contain only glicentin and presumably lack the enzyme required to process this precursor to glucagon. Enteroglucagon binds to hepatic glucagon receptors and stimulates adenylyl cyclase with 10% to 20% of the potency of glucagon. Glucagon-like peptide-1 is an extremely potent potentiator of insulin secretion, although it apparently lacks significant hepatic actions. Glicentin, enteroglucagon, and the glucagon-like peptides are found predominantly in the intestine, and their secretion continues after total pancreatectomy. Regulation of Secretion The secretion of glucagon is regulated by dietary glucose, insulin, amino acids, and fatty acids; glucose is a potent inhibitor. As in insulin secretion, glucose is a more effective inhibitor of glucagon secretion when taken orally than when administered intravenously, suggesting a possible role for some gastrointestinal hormone in the response. The effect of glucose is lost in the untreated or undertreated type 1 DM patient and in isolated pancreatic cells, indicating that at least part of the effect is secondary to stimulation of insulin secretion. Somatostatin also inhibits glucagon secretion, as do free fatty acids and ketones. Most amino acids stimulate the release of both glucagon and insulin. This coordinated response to amino acids may prevent insulin-induced hypoglycemia in individuals who ingest a meal of pure protein. Like glucose, amino acids are more potent when taken orally and thus may exert some of their effects via gastrointestinal hormones. Secretion of glucagon also is regulated by the autonomic innervation of the pancreatic islet. Stimulation of sympathetic nerves or administration of sympathomimetic amines increases glucagon secretion. Acetylcholine has a similar effect. Glucagon in Diabetes Mellitus Plasma concentrations of glucagon are elevated in poorly controlled diabetic patients. In view of its

capacity to enhance gluconeogenesis and glycogenolysis, glucagon exacerbates the hyperglycemia of diabetes. However, this abnormality of glucagon secretion appears to be secondary to the diabetic state and is corrected with improved control of the disease (Unger, 1985). The importance of the hyperglucagonemia in diabetes has been evaluated by administration of somatostatin (Gerich et al., 1975). Although somatostatin does not restore glucose metabolism to normal, it significantly slows the rate of development of hyperglycemia and ketonemia in insulinopenic type 1 DM subjects. In normal individuals, glucagon secretion increases in response to hypoglycemia, but in type 1 DM patients this important defense mechanism (against insulin-induced hypoglycemia) is lost early in the course of the disease. Degradation Glucagon is extensively degraded in liver, kidney, and plasma, as well as at its sites of action (Peterson et al., 1982). Its half-life in plasma is approximately 3 to 6 minutes. Proteolytic removal of the amino-terminal histidine residue leads to loss of biological activity. Cellular and Physiological Actions Glucagon interacts with a 60,000-dalton glycoprotein receptor on the plasma membrane of target cells (Sheetz and Tager, 1988). Although the exact structure of this receptor is not yet known, it interacts with the stimulatory guanine-nucleotide-binding regulatory protein, Gs, which activates adenylyl cyclase (see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). The primary effects of glucagon on the liver are mediated by cyclic AMP. In general, modifications of the amino-terminal region of glucagon (e.g., [Phe1]glucagon and des-His1-[Glu9]glucagon amide) result in molecules that behave as partial agoniststhey retain some affinity for the glucagon receptor but have a markedly reduced capacity to stimulate adenylyl cyclase (Unson et al., 1989). Phosphorylase, the rate-limiting enzyme in glycogenolysis, is activated by glucagon as a result of cyclic AMP-stimulated phosphorylation, while concurrent phosphorylation of glycogen synthase inactivates the enzyme; glycogenolysis is enhanced and glycogen synthesis is inhibited. Cyclic AMP also stimulates transcription of the gene for phosphoenolpyruvate carboxykinase, a ratelimiting enzyme in gluconeogenesis (Granner et al., 1986). These effects are normally opposed by insulin, and when maximal concentrations of both hormones are present, insulin is dominant. Cyclic AMP also stimulates phosphorylation of the bifunctional enzyme 6-phosphofructo-2kinase/fructose-2,6-bisphosphatase (Pilkis et al., 1981; Foster, 1984). This enzyme determines the cellular concentration of fructose-2,6-bisphosphate, which acts as a potent regulator of gluconeogenesis and glycogenolysis. When the concentration of glucagon is high relative to that of insulin, this enzyme is phosphorylated and acts as a phosphatase, reducing the concentration of fructose-2,6-bisphosphate in the liver. When the concentration of insulin is high relative to that of glucagon, the enzyme is dephosphorylated and acts as a kinase, raising fructose-2,6-bisphosphate concentrations. Fructose-2,6-bisphosphate interacts allosterically with phosphofructokinase-1, the rate-limiting enzyme in glycolysis, increasing its activity. Thus, when glucagon concentrations are high, glycolysis is inhibited and gluconeogenesis is stimulated. This also leads to a decrease in the concentration of malonyl CoA, stimulation of fatty-acid oxidation, and production of ketone bodies. Conversely, when insulin concentrations are high, glycolysis is stimulated and gluconeogenesis and ketogenesis are inhibited (see Foster, 1984). Glucagon exerts effects on tissues other than liver, especially at higher concentrations. In adipose tissue, it stimulates adenylyl cyclase and increases lipolysis. In the heart, glucagon increases the

force of contraction. Glucagon has relaxant effects on the gastrointestinal tract; this has been observed with analogs that apparently do not stimulate adenylyl cyclase. Some tissues (including liver) possess a second type of glucagon receptor that is linked to generation of inositol trisphosphate, diacylglycerol, and Ca2+ (Murphy et al., 1987). The role of this receptor in regulation of metabolism remains uncertain. Therapeutic Use Glucagon is used to treat severe hypoglycemia, particularly in diabetic patients when intravenous glucose is not available; it also is used by radiologists for its inhibitory effects on the gastrointestinal tract. All glucagon used clinically is extracted from bovine and porcine pancreas; its sequence is identical to that of the human hormone. For hypoglycemic reactions, 1 mg is administered intravenously, intramuscularly, or subcutaneously. Either of the first two routes is preferred in an emergency. Clinical improvement is sought within 10 minutes to minimize the risk of neurological damage from hypoglycemia. The hyperglycemic action of glucagon is transient and may be inadequate if hepatic stores of glycogen are depleted. After the initial response to glucagon, patients should be given glucose or urged to eat to prevent recurrent hypoglycemia. Nausea and vomiting are the most frequent adverse effects. Glucagon also is used to relax the intestinal tract to facilitate radiographic examination of the upper and lower gastrointestinal tract with barium and retrograde ileography (Monsein et al., 1986) and in magnetic resonance imaging of the gastrointestinal tract (Goldberg and Thoeni, 1989). Glucagon has been used to treat the spasm associated with acute diverticulitis and disorders of the biliary tract and sphincter of Oddi, as an adjunct in basket retrieval of biliary calculi, and for impaction of the esophagus and intussusception (Friedland, 1983; Mortensson et al., 1984; Kadir and Gadacz, 1987). It has been used for diagnostic purposes to distinguish obstructive from hepatocellular jaundice (Berstock et al., 1982). Glucagon releases catecholamines from a pheochromocytoma and has been used experimentally as a diagnostic test for this disorder. The hormone also has been used as a cardiac inotropic agent for the treatment of shock, particularly when prior administration of a -adrenergic receptor antagonist has rendered -adrenergic receptor agonists ineffective. Somatostatin Somatostatin was first isolated and synthesized in 1973, following a search for hypothalamic factors that might regulate secretion of growth hormone from the pituitary gland (Brazeau et al., 1973; see also Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). A potential physiological role for somatostatin in the islet was suggested by the observation that somatostatin inhibits secretion of insulin and glucagon (Alberti et al., 1973; Gerich et al., 1974). The peptide subsequently was identified in the D cells of the pancreatic islet, in similar cells of the gastrointestinal tract, and in the central nervous system (Dubois, 1975). Somatostatin, the name originally given to a cyclic peptide containing 14 amino acids, is now known to be one of a group of related peptides. These include the original somatostatin (S-14), an extended 28-amino-acid peptide molecule (S-28), and a fragment containing the initial 12 amino acids of somatostatin-28 [S-28(112)]. Somatostatin-14 is the predominant form in the brain, whereas somatostatin-28 is the main form in the gut. Somatostatin inhibits the release of thyroidstimulating hormone and growth hormone from the pituitary gland, of gastrin, motilin, vasoactive

intestinal peptide (VIP), glicentin, and gastrointestinal polypeptide from the gut, and of insulin, glucagon, pancreatic polypeptide, and somatostatin from the pancreas. Somatostatin secreted from the pancreas can regulate pituitary function, thereby acting as a true neurohormone. In the gut, however, somatostatin acts as a paracrine agent by influencing the functions of adjacent cells. It also can act as an autocrine agent by inhibiting its own release at the pancreas. The D cell is the last to receive blood flow in the islet; that is, it is downstream from the and cells (Samols et al., 1986). Thus, somatostatin may regulate the secretion of insulin and glucagon only via the systemic circulation. Somatostatin is released in response to many of the nutrients and hormones that stimulate insulin secretion, including glucose, arginine, leucine, glucagon, vasoactive intestinal polypeptide, cholecystokinin, and even tolbutamide (Ipp et al., 1977; Weir et al., 1979). The physiological role of somatostatin has not been defined precisely. When administered in pharmacological amounts, somatostatin inhibits virtually all endocrine and exocrine secretions of the pancreas, gut, and gallbladder. Somatostatin also can inhibit secretion of the salivary glands and, under some conditions, can block parathyroid, calcitonin, prolactin, and ACTH secretion. The cell is about 50 times more sensitive to somatostatin than is the cell, but inhibition of glucagon secretion is more transient. Somatostatin also inhibits nutrient absorption from the intestine, decreases intestinal motility, and reduces splanchnic blood flow. The therapeutic uses of somatostatin are confined mainly to blocking hormone release in endocrinesecreting tumors, including insulinomas, glucagonomas, VIPomas, carcinoid tumors, and somatotropinomas (causing acromegaly). Because of its short biological half-life (3 to 6 minutes), substantial effort has been directed toward the production of a longer-acting analog. One such agent, octreotide (SANDOSTATIN ), is now available in the United States for treatment of carcinoid tumors, glucagonomas, VIPomas, and acromegaly (see also Chapter 56: Pituitary Hormones and Their Hypothalamic Releasing Factors). Octreotide successfully controls excess secretion of growth hormone in most patients and has been reported to reduce the size of pituitary tumors in about onethird of cases. Octreotide also has been used to reduce the disabling form of diarrhea that occasionally occurs in diabetic autonomic neuropathy. As octreotide also can decrease blood flow to the gastrointestinal tract, the agent has been used to treat bleeding esophageal varices, peptic ulcers, and postprandial orthostatic hypotension. Gallbladder abnormalities (stones and biliary sludge) occur frequently with chronic use of the peptide, and abnormal cardiac rhythms and gastrointestinal symptoms also occur commonly. Hypoglycemia, hyperglycemia, hypothyroidism, and goiter have been significant complications in patients being treated with octreotide for acromegaly. Diazoxide Diazoxide is an antihypertensive, antidiuretic benzothiadiazine derivative with potent hyperglycemic actions when given orally (see also Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension). Hyperglycemia results primarily from inhibition of insulin secretion (Levin et al., 1975). Diazoxide interacts with an ATP-sensitive K+ channel and either prevents its closing or prolongs the open time; this effect is opposite to that of the sulfonylureas (Panten et al., 1989). The drug does not inhibit insulin synthesis, and thus there is an accumulation of insulin within the cell. Diazoxide also has a modest capacity to inhibit peripheral glucose utilization by muscle and to stimulate hepatic gluconeogenesis. Diazoxide (PROGLYCEM) has been used to treat patients with various forms of hypoglycemia (Grant

et al., 1986). The usual oral dose is 3 to 8 mg/kg per day in adults and 8 to 15 mg/kg daily in infants and neonates. The drug has a tendency to cause nausea and vomiting and thus is usually given in divided doses with meals. Diazoxide circulates largely bound to plasma proteins and has a half-life of about 48 hours. Thus, the patient should be maintained at any dosage for several days before evaluating the therapeutic result. Diazoxide has a number of adverse effects that sometimes limit its use in the treatment of hypoglycemia. These include retention of Na+ and fluid, hyperuricemia, hypertrichosis (especially in children), thrombocytopenia, and leukopenia. Despite these side effects, the drug may be quite useful in patients with inoperable insulinomas (Schein, 1973) and in children with hyperinsulinism due to nesidioblastosis (Grant et al., 1986). Prospectus The dramatic increase in the prevalence of type 2 DM has refocused clinical strategies to control plasma glucose levels and prevent complications of the disease. The DPP trial is a large, multicenter clinical study in the United States that is aimed at determining whether or not lifestyle changes or therapeutic intervention (metformin) at the stage of impaired glucose tolerance (IGT) can prevent the onset of diabetes. Other large trials (BARI 2, VA Cooperative Study) are focused on determining whether or not tight metabolic control and the class of therapeutic agents (insulin sensitizers vs. sulfonylureas vs. insulin or combinations thereof) can reduce macrovascular disease in type 2 diabetes. Currently, no drug is approved in the United States to treat IGT. The number of individuals with IGT worldwide is enormous (20 million in the United States alone). Approximately 5% of individuals with IGT develop diabetes each year. Thus, any therapeutic option that can prevent the transformation of IGT into diabetes is eagerly awaited. Atherosclerotic disease is by far the primary cause of death in patients with diabetes. The pathophysiology resulting in atherosclerosis is complex and multifactoral. For nearly three decades, there has been considerable debate regarding whether hyperglycemia, hyperinsulinemia, and/or sulfonylurea therapy promotes atherosclerosis and accelerates cardiovascular disease. Recent data from the United Kingdom Prospective Diabetes Study demonstrate that neither insulin nor sulfonylurea therapy is associated with a higher incidence of macrovascular disease. Interestingly, a subgroup of patients treated primarily with metformin experienced a significant reduction in macrovascular events. The mechanism responsible for this finding is unknown, but the study raises intriguing possibilities that insulin resistance or associated factors (cytokines or plasminogen activator inhibitor I) may be causally implicated in the development of macrovascular complications in type 2 DM. The results from the above trials should clarify these most important questions and provide additional novel therapeutic targets to reduce the devastating consequences of this common disease. For further discussion of diabetes mellitus, see Chapter 323, and for information about hypoglycemia, see Chapter 324, in Harrison's Principles of Internal Medicine, 16th ed., McGrawHill, New York, 2005.

Chapter 62. Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds

Overview In earlier editions, this chapter focused on hormones involved with calcium homeostasis, mechanisms by which they act to maintain blood Ca2+ concentrations within normal limits, and the derangements in calcium physiology associated with insufficiency or excess of these hormones. In recent years, there has been a shift in the relative importance of the prevalence and severity of these disorders. Primary hyperparathyroidism is more commonly diagnosed than in years past, but most commonly appears today as a mild disorder that does not necessarily require treatment. By contrast, osteoporotic fracture, particularly of the hip, has emerged as a major public health problem and an important contributor to disability, mortality, and health care costs in industrialized countries. Considerable information has been obtained regarding the acquisition and subsequent loss of bone, as well as the contributions of genetics, diet, physical activity, and reproductive hormone status to skeletal health. Important knowledge also has been obtained concerning the central role of bone remodeling as the final pathway of adult bone loss. A rapidly increasing body of evidence supports the concept that regular physical activity, adequate calcium intake, either through diet or supplements, and timely use of estrogen replacement therapy will decrease bone remodeling, constrain bone loss, and reduce fracture risk. However, treatment of established osteoporosis remains a formidable challenge. Like estrogen and calcium, other approved therapies, such as bisphosphonates, act by slowing bone resorption rather than by stimulating new bone formation and therefore do not solve the problem of restoring normal bone mass. In fact, since bone remodeling is a coupled process, agents that suppress bone resorption ultimately decrease bone formation. The primary challenge for future research in this field is to develop agents that safely increase bone mass. At present, there is considerable interest in developing analogs of parathyroid hormone, vitamin D, and various bone morphogenetic proteins as potential therapies for osteoporosis. Another recent development is the recognition that vitamin D plays an important role as a cellular differentiation factor in systems not directly related to calcium metabolism. Calcitriol, the hormonal form of vitamin D, shows considerable promise as a treatment for psoriasis and also is under study for several malignancies. Therapeutic utility of calcitriol is limited by its calcemic effects, but noncalcemic calcitriol analogs are under development. Such analogs may offer a new approach to manage patients with diverse conditions, ranging from primary and secondary hyperparathyroidism to cancer and leukemia. Calcium Ca2+ is the major extracellular divalent cation. The normal adult man and woman possess about 1300 and 1000 g of calcium, respectively, of which more than 99% is in bone. Ca2+ is present in small amounts in extracellular fluids and to a minor extent within cells, where its ionized concentration under basal conditions is about 0.1 M. In response to hormonal, electrical, or mechanical stimuli, a temporary increase in Ca2+ flux raises this concentration toward 1 M, permitting interactions with specific Ca2+-binding proteins that activate numerous processes. The major Ca2+-binding protein in all organisms is calmodulin, a highly conserved protein that binds four moles of Ca2+ per mole of protein. Ca2+ is essential for many important processes, including neuronal excitability, neurotransmitter release, muscle contraction, membrane integrity, and blood 2+ coagulation. In addition, Ca serves a second messenger function for the actions of many hormones. To carry out these various roles, Ca2+ must be available in the proper concentration. In human plasma, calcium circulates concentration about to 10.4 mg/dl (2.1 to 2.6 mM). Of this, about 45% is

bound to plasma proteins, primarily albumin, and about 10% is complexed with anionic buffers, such as citrate and phosphate. The remaining fraction, ionized Ca2+, is the component that exerts physiological effects and, when reduced, produces hypocalcemic symptoms. Hence, interpretation of any given value for total plasma calcium is impossible without correction for the concentration of plasma proteins. As an approximation, a change in plasma albumin concentration of 1.0 g/dl from a nominal value of 4.0 g/dl can be expected to change total calcium by 0.8 mg/dl. Regulation of the extracellular Ca2+ concentration is under tight endocrine control that affects its entry at the intestine and its exit at the kidney, and which regulates a large skeletal reservoir for withdrawals at times of need. Calcium Stores The skeleton contains 99% of total body calcium in a crystalline form resembling the mineral hydroxyapatite [Ca10(PO4)6(OH)2]; but other ionsincluding Na +, K+, Mg2+, and Falso are present in the crystal lattice. The steady-state content of calcium in bone reflects the net effect of bone resorption and bone formation, two coupled aspects of bone remodeling (see below). In addition, a labile pool of bone Ca2+ is readily exchangeable with interstitial fluid. The rates of exchange are modulated by drugs, hormones, vitamins, and other factors that directly alter bone turnover or that influence the level of Ca2+ in interstitial fluid. Calcium Absorption and Excretion In the United States, about 75% of dietary calcium is obtained from milk and dairy products. The adequate intake value for calcium in adolescents is 1300 mg/day and in adults to age 24 years is 1000 mg/day. The adequate intake for men and women age 50 and older is 1200 mg/day (see Section XIII: "The Vitamins, Introduction") (Institute of Medicine, 1997). The median intakes of calcium for boys and girls aged 9 years and older are 865 and 625 mg, respectively. After age 50, median daily calcium intake declines for women (to 517 mg) (Institute of Medicine, 1997). Figure 621 demonstrates the elements of whole body daily calcium turnover. Ca 2+ enters the body only through the intestine. Two different mechanisms contribute to this relatively inefficient process. Active vitamin Ddependent transport occurs in the proximal duodenum. In addition, accounting for a large fraction of total Ca2+ uptake, facilitated diffusion takes place throughout the small intestine. There is also an obligatory daily intestinal calcium loss of about 150 mg/day, reflecting the mineral contained in mucosal and biliary secretions and in sloughed intestinal cells. Figure 621. Schematic Representation of the Whole Body Daily Turnover of Calcium. (Adapted with permission from Yanagawa and Lee, 1992.)

Intestinal Ca2+ absorption efficiency is inversely related to calcium intake, so that a diet low in calcium leads to a compensatory increase in fractional absorption, due in part to activation of vitamin D. The strength of this response decreases substantially with age. Drugs such as glucocorticoids and phenytoin depress intestinal Ca 2+ transport. Some dietary constituents, e.g., phytate and oxalate, depress Ca2+ absorption by promoting the formation of nonabsorbable complexes. Disease states associated with steatorrhea, diarrhea, or chronic intestinal malabsorption also promote fecal loss of calcium. Urinary excretion of Ca2+ is the net result of the quantity filtered at the glomerulus and the amount reabsorbed. About 9 g of Ca2+ is filtered each day. Tubular reabsorption is very efficient, more than 98% of filtered Ca2+ returning to the circulation. Reabsorption efficiency is highly regulated by parathyroid hormone (PTH) but also is influenced by filtered Na+, the presence of nonreabsorbed anions, and diuretic agents. Sodium intake, and therefore sodium excretion, is directly related to urinary calcium excretion. Diuretics acting on the ascending limb of the loop of Henle increase calciuresis. By contrast, thiazide diuretics uniquely uncouple the relationship between Na+ and Ca 2+ excretion, leading to reduced calciuria (Lemann et al., 1985). Dietary protein is directly related to urine calcium excretion, presumably as an effect of sulfur-containing amino acids on renal tubular function. Urinary Ca2+ is only slightly influenced by dietary calcium in normal people. Significant amounts of calcium are secreted in milk during lactation; sweat also makes a small contribution to daily losses. Bone Remodeling Growth and development of endochondral bone are driven by a process called modeling. Once new bone is laid down, it is subject to a continuous process of breakdown and renewal called remodeling that continues throughout life. After linear growth has ceased and peak bone mass has been approached, remodeling becomes the final common pathway by which bone mass is adjusted throughout adult life. Remodeling is carried out by myriad individual and independent "bone

remodeling units" throughout the skeleton (Figure 622). It takes place on bone surfaces, about 90% of which are normally inactive, covered by a thin layer of lining cells. In response to physical or biochemical signals, recruitment of marrow precursor cells to a site at the bone surface results in their fusion into the characteristic multinucleated osteoclasts that resorb, or dig a cavity into the bone. Figure 622. The Bone Remodeling Cycle. A. Resting trabecular surface. B.Multinucleated osteoclasts dig a cavity of approximately 20 m. C. Resorption to 60 m is completed by mononuclear phagocytes.D.Osteoblast precursors are recruited to the base of the resorption cavity. E. New matrix is secreted by osteoblasts.F. Matrix continues to be secreted, with the initiation of calcification. G. Mineralization of the new matrix is completed. Bone has returned to a quiescent state, but a small deficit in bone mass persists. (Adapted from Marcus, 1987, with permission.)

Osteoclast production is regulated by cytokines, such as interleukins-1 and -6, produced by osteoblasts. Recent studies have begun to clarify the mechanisms through which osteoclast production is regulated (see Suda et al., 1999). RANK (receptor for activating NF B) is the name given to an osteoclast protein whose expression is required for osteoclastic bone resorption. Its natural ligand, osteoclast differentiation factor (ODF, also called RANK ligand), is a membranespanning osteoblast protein. Upon binding to RANK, ODF induces osteoclast formation. The requirement for this interaction is revealed by findings that antibodies against ODF inhibit bone resorption induced by multiple hormones and other regulators of bone turnover (Yasuda et al., 1998). ODF initiates the activation of mature osteoclasts as well as the differentiation of osteoclast precursors (Jimi et al., 1999). Osteoblasts also produce an inhibitor of ODF action, called osteoprotegerin (OPG), which acts as a decoy ligand for ODF. Under conditions favoring increased bone resorption, as during estrogen deprivation, OPG is suppressed, ODF binds to RANK, and osteoclast production increases. When estrogen sufficiency is reestablished, OPG increases and competes effectively with ODF for binding to RANK. In cortical bone, resorption creates tunnels within Haversian canals, whereas trabecular resorption creates scalloped areas of the bone surface called Howship's lacunae. On termination of the resorption phase, a cavity remains that is about 60 m deep and is bordered at its deepest extent by a cement line, a region of loosely organized collagen fibrils. Completion of the resorption phase is followed by ingress of preosteoblasts derived from marrow stroma into the base of the resorption cavity. These cells develop the characteristic osteoblastic phenotype and begin to replace the resorbed bone by elaborating new bone matrix constituents, such as collagen, osteocalcin, and other proteins. Once the newly formed osteoid reaches a thickness of about 20 m, mineralization begins. Completion of a remodeling cycle normally requires about 6 months. If the replacement of resorbed bone matched the amount that was removed, remodeling would lead to no net change in bone mass. However, small bone deficits persist on completion of each cycle, reflecting an inefficiency in remodeling dynamics. Consequently, lifelong accumulation of remodeling deficits underlies the well-documented phenomenon of age-related bone loss, a process

that begins shortly after growth stops. Alterations in remodeling activity represent the final pathway through which diverse stimuli, such as dietary insufficiency, hormones, and drugs, affect bone balance. A change in whole-body remodeling rate can be brought about through distinct perturbations in remodeling dynamics. Changes in hormonal milieu often lead to an increase in the activation, or birthrate, of remodeling units. Examples include hyperthyroidism, hyperparathyroidism, and hypervitaminosis D. Other factors may impair osteoblastic functional adequacy, such as high doses of corticosteroids or ethanol. Finally, it appears that estrogen deficiency may augment osteoclastic resorptive capacity (see Marcus, 1987; Dempster, 1992). At any given time, a transient deficit in bone exists called the remodeling space, representing sites of bone resorption that have not yet filled in. In response to any stimulus that alters the birthrate of new remodeling units, the remodeling space will either increase or decrease accordingly until a new steady state is established, and this adjustment will be seen as an increase or decrease in bone mass. Physiological and Pharmacological Actions Neuromuscular System Moderate elevations of the concentration of Ca2+ in the extracellular fluid may have no clinically detectable influences on the neuromuscular apparatus. However, when hypercalcemia becomes severe, the threshold for excitation of nerve and muscle is increased. This is manifested clinically by muscle weakness, lethargy, and even coma. In contrast, modest reductions in Ca2+ activity may decrease excitation thresholds, leading to positive Chvostek and Trousseau signs, tetanic seizures, and laryngospasm. Ca2+ influx into cells is thought to be by means of carrier-mediated facilitated diffusion and by exchange of Ca2+ for Na+. Several Ca2+ channels in cell membranes are regulated by hormones and neurotransmitters and membrane potential. In liver and skeletal muscle, intracellular Ca2+ is reversibly sequestered by endoplasmic and sarcoplasmic reticulum, respectively. Ca2+ plays an important role in muscular excitationcontraction coupling. The action potential stimulates Ca2+ release from the sarcoplasmic reticulum. The released Ca2+ activates contraction by binding to troponin, abolishing the inhibitory effect of troponin on the actinmyosin interaction. Muscle relaxation occurs when Ca2+ is pumped back into the sarcoplasmic reticulum, restoring troponin inhibition. Ca2+ is necessary for exocytosis and thus has an important role in stimulussecretion coupling in most exocrine and endocrine glands. Release of catecholamines from the adrenal medulla, neurotransmitters at synapses, and certain autacoids (e.g., histamine from mast cells) requires Ca2+. Cardiovascular System Ca2+ is essential for excitationcontraction coupling in cardiac muscle, as well as for the conduction of electrical impulses in certain regions of the heart, particularly through the AV node. Depolarization of myocardial fibers opens voltage-regulated Ca2+ channels and causes the "slow" inward current that occurs during the action potential plateau. This current allows permeation of Ca2+ sufficient to trigger the release of additional Ca2+ from the sarcoplasmic reticulum, thereby causing contraction. Passage of Ca2+ through similar channels in tissues such as the AV node carries virtually all the inward (depolarizing) current during the action potential. Ca2+ is responsible for the initiation of contraction in vascular and other smooth muscles, and it frequently carries an important fraction of depolarizing currents in these tissues. Hence Ca2+

channel blockers have profound effects on the contractility of cardiac and vascular smooth muscle as well as on the conduction of impulses within the heart. These drugs have important uses in the treatment of angina, cardiac arrhythmias, and hypertension (see Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia, 33: Antihypertensive Agents and the Drug Therapy of Hypertension, and 35: Antiarrhythmic Drugs). Miscellaneous Effects Ca2+ plays a role in maintaining the integrity of mucosal membranes, cell adhesion, and functions of individual cell membranes as well. Ca2+ is involved in blood coagulation, but the ion is not used to treat disorders of coagulation. Calcium chloride is an acidifying salt and will promote diuresis; however, ammonium salts are much more effective acidifying agents. Abnormalities of Calcium Metabolism Hypocalcemic States The prominent signs and symptoms of hypocalcemia include tetany and related phenomena such as paresthesias, increased neuromuscular excitability, laryngo-spasm, muscle cramps, and tonic-clonic convulsions. Some causes of hypocalcemia are discussed below. Combined deprivation of Ca and vitamin D readily promotes hypocalcemia. This combination of events is observed in the various malabsorption states and also occurs from inadequate diets. When due to malabsorption, hypocalcemia is accompanied by low concentrations of phosphate, total 2+ plasma proteins, and magnesium. During Mg deficiency, hypocalcemia may be accentuated by diminished secretion and action of PTH (see below). Hypocalcemia stimulates the release of PTH, which increases bone turnover, resulting in increased delivery of skeletal calcium to the extracellular fluid. In infants with malabsorption or inadequate calcium intake, Ca2+ concentrations are usually depressed, there is hypophosphatemia, and the resultant bone disease is rickets (see"Vitamin D," below). Hypoparathyroidism is most often a consequence of thyroid or neck surgery, but it also may be due to genetic or autoimmune disorders. In hypoparathyroidism, hypocalcemia is accompanied by hyperphosphatemia, reflecting decreased PTH action on renal phosphorus handling. Although other conditions of hypocalcemia may be associated with lens opacity, papilledema, and calcification of the basal ganglia, these conditions occur more often with hypoparathyroidism. Pseudohypoparathyroidism (PHP) is characterized by multiple somatic defects and a failure to respond to exogenous PTH. The somatic features include a round face, short stature, and shortening of metacarpal and metatarsal bones (Albright's hereditary osteodystrophy). In its classic form, PHP is due to a mutant guanine nucleotidebinding protein that normally mediates hormone-induced activation of adenylyl cyclase (see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). An assortment of hormonal abnormalities has been associated with this type of PHP, but none is so severe as the deficient response to PTH. PHP has been reviewed recently by Levine (1999). In the first several days following removal of a parathyroid adenoma, hypocalcemia is not unusual. This may be due to temporary failure of the remaining parathyroid glands to compensate for the missing adenomatous tissue. In this case, hyperphosphatemia also is seen, and the condition is one of functional hypoparathyroidism. In patients with parathyroid bone disease, postoperative hypocalcemia may reflect rapid uptake of calcium into bone, the so-called hungry bone syndrome. Here, the serum inorganic phosphate concentration also is low, reflecting its concurrent uptake into
2+

bone, and persistent, severe hypocalcemia may require administration of vitamin D and supplemental calcium for several months. Neonatal tetany sometimes is observed in infants of mothers with hyperparathyroidism; indeed, it may be the tetany that calls attention to the mother's disorder. This problem usually is transient, disappearing when the infant's own parathyroid glands respond appropriately. Hypocalcemia is associated with advanced renal insufficiency accompanied by hyperphosphatemia. Many patients with this condition do not develop tetany unless the severe accompanying acidosis is treated. High concentrations of phosphate in plasma inhibit the conversion of 25hydroxycholecalciferol to 1,25-dihydroxycholecalciferol (Haussler and McCain, 1977). Hypocalcemia also can occur following massive transfusions with citrated blood. Treatment of Hypocalcemia and Other Therapeutic Uses of Calcium Calcium is used in the treatment of calcium deficiency states and as a dietary supplement. Ca2+ salts are specific in the immediate treatment of hypocalcemic tetany regardless of etiology. In severe manifest tetany, symptoms are best brought under control by intravenous medication. Calcium chloride (CaCl22H2O) contains 27% Ca2+; it is valuable in the treatment of hypocalcemic tetany and laryngospasm. The salt is given intravenously, but it must never be injected into tissues. Injections of calcium chloride are accompanied by peripheral vasodilation and a cutaneous burning sensation. The salt usually is given intravenously in a concentration of 10% (equivalent to 1.36 meq Ca2+/ml). The rate of injection should be slow (not over 1 ml per minute) to prevent a high concentration of Ca2+ from causing a cardiac arrhythmia. A moderate fall in blood pressure due to vasodilation may attend the injection. Since calcium chloride is an acidifying salt, it is usually undesirable in the treatment of the hypocalcemia caused by renal insufficiency. Calcium gluceptate injection (a 22% solution; 18 mg or 0.9 meq of Ca2+/ml) is administered intravenously in a dose of 5 to 20 ml for the treatment of severe hypocalcemic tetany; the injection produces a transient tingling sensation when given too rapidly. When the intravenous route is not possible, injections may be given intramuscularly in a dose up to 5 ml, which may produce a mild local reaction. Calcium gluconate injection (a 10% solution; 9.3 mg of Ca 2+/ml) is a readily available source of calcium, and the intravenous administration of this salt is the treatment of choice for severe hypocalcemic tetany. Patients with moderate to severe hypocalcemia may be treated by infusing calcium gluconate at a dose of 10 to 15 mg/kg body weight of Ca2+ over 4 to 6 hours. Since the usual 10-ml vial of a 10% solution contains only 93 mg of Ca 2+, many vials are needed. The intramuscular route should not be employed, since abscess formation at the injection site may result. For control of milder hypocalcemic symptoms, oral medication suffices, frequently in combination with vitamin D or one of its active metabolites. Numerous oral Ca2+ salts are available. Average doses for hypocalcemic patients are calcium gluconate, 15 g daily in divided doses; calcium lactate, 4 g plus 8 g lactose, with each meal; calcium carbonate or calcium phosphate, 1 to 2 g with meals. Calcium carbonate and calcium acetate are used to restrict phosphate absorption in patients with chronic renal failure and oxalate absorption in patients with inflammatory bowel disease. Acute administration of calcium may be life-saving in patients with extreme hyperkalemia. Calcium gluconate (10 to 30 ml of a 10% solution) can reverse some of the cardiotoxic effects of hyperkalemia while other efforts are under way to lower plasma concentrations of K+ . Use of supplemental calcium in the prevention and treatment of osteoporosis is discussed below.

Hypercalcemic States Hypercalcemia occurs in many diverse clinical conditions and requires differential diagnosis and appropriate corrective measures. Ingestion of large quantities of a Ca2+ salt does not generally by itself cause hypercalcemia, an exception being patients with hypothyroidism, who absorb Ca 2+ with increased efficiency (Benker et al., 1988). Also, the uncommon hypercalcemic disorder called milkalkali syndrome is caused by concurrent ingestion of large quantities of milk and alkalinizing powders, in which setting renal Ca2+ excretion is impaired. In an outpatient setting, the most common cause of hypercalcemia is primary hyperparathyroidism (HPT), accompanied frequently by significant hypophosphatemia; the latter reflects diminished renal tubular phosphorus reabsorption due to hypersecretion of PTH. Some patients have renal calculi and peptic ulceration, and a few still show classical parathyroid skeletal disease. However, most patients today show few if any symptoms, and those that are present are often vague and nonspecific. Contemporary use of immunoradiometric (IRMA) assays for the intact PTH molecule obviates many of the difficulties with previous assays and is associated with a diagnostic accuracy of >90% (Endres et al., 1991). Familial benign hypercalcemia (or familial hypocalciuric hypercalcemia) is an inherited hypercalcemic disorder that generally is accompanied by extremely low urinary Ca2+ excretion. Hypercalcemia usually is mild, and circulating PTH is often normal to slightly elevated. The importance of making this diagnosis lies in the fact that patients mistakenly diagnosed to have primary HPT may be submitted to surgical exploration without discovery of an adenoma, and without therapeutic benefit. Patients do not experience long-term clinical consequences, except for homozygous infants, who may have severe, even lethal, hypercalcemia. Diagnosis is established by demonstrating hypercalcemia in first-degree family members. The molecular basis for familial benign hypercalcemia is a mutation in the Ca2+-sensing receptor (Pollak et al., 1993). Most hypercalcemia discovered in hospitals is associated with a systemic malignancy, either with or without bony metastasis. PTH-related protein (PTHrP) is a primitive, highly conserved protein that may be abnormally expressed in malignant tissue, particularly by squamous cell and other epithelial cancers. The presence of substantial sequence homology of the amino-terminal portion of PTHrP with the amino terminus of native PTH permits this molecule to interact with the PTH receptor in target tissues and underlies the hypercalcemia and hypophosphatemia that are seen in humoral hypercalcemia of malignancy (see Grill and Martin, 1994). Other tumors release cytokines or prostaglandins that stimulate bone resorption. Hypercalcemia associated with malignancy is generally more severe than in HPT (frequently >13 mg/dl) and may be associated with lethargy, weakness, nausea, vomiting, polydipsia, and polyuria. Assays for PTHrP may aid diagnosis. In some patients with lymphomas, hypercalcemia is due to overproduction of 1,25-dihydroxyvitamin D by the tumor cells. A similar mechanism explains the hypercalcemia that is seen occasionally in sarcoidosis and other granulomatous disorders. Vitamin D excess may cause hypercalcemia. In this case, sufficient 25-hydroxyvitamin D is present to stimulate intestinal Ca2+ hyperabsorption, leading to hypercalcemia and suppression of PTH and 1,25-dihydroxyvitamin D production. Thus, measurement of 25-hydroxyvitamin D is diagnostic. Occasional patients with hyperthyroidism show mild hypercalcemia, presumably due to a direct effect of thyroid hormone on bone turnover. Immobilization may lead to hypercalcemia in growing children and young adults, but is an unusual cause of hypercalcemia in older individuals unless bone turnover is already increased, as in Paget's disease or in hyperthyroidism. Hypercalcemia is sometimes noted in adrenocortical deficiency, as in Addison's disease, or following removal of a hyperfunctional adrenocortical tumor. Hypercalcemia occurs following renal transplantation, owing

to persistent hyperfunctioning parathyroid tissue that resulted from the previous renal failure. Differential diagnosis of the various causes of hypercalcemia may pose difficulties, but recent advances in serum tests for PTH, PTHrP, 25-hydroxy- and 1,25-dihydroxyvitamin D have facilitated accurate diagnosis in the great majority of cases. Hypercalcemia of any etiology can have dire consequences. The predominant and most devastating lesion usually occurs in the kidney, with reductions of renal function and nephrocalcinosis. Treatment of Hypercalcemia Hypercalcemia occasionally may be life threatening. Such patients frequently are severely dehydrated because hypercalcemia has compromised renal concentrating mechanisms. Thus, fluid resuscitation with large volumes of isotonic saline must be early and aggressive (6 to 8 liters/day). Agents that augment Ca2+ excretion, such as loop diuretics, may help to counteract the effect of plasma volume expansion by saline, but they are contraindicated by themselves, as they will aggravate dehydration and hypercalcemia. Corticosteroids administered at high doses (e.g., 40 to 80 mg/day of prednisone) may be useful in situations where hypercalcemia results from diseases such as sarcoidosis, lymphoma, or hypervitaminosis D. The response to steroid therapy is slow; 1 to 2 weeks may be required before plasma Ca2+ falls. Calcitonin (CALCIMAR, MIACALCIN) acts specifically on osteoclasts to inhibit bone resorption and may be useful in managing hypercalcemia. Reduction in Ca2+ may be rapid, although escape from the hormone regularly occurs in several days. The recommended starting dose is 4 units/kg body weight every 12 hours; if there is no response within one or two days, the dose may be increased to 8 units/kg every 12 hours. If the response after two more days still is unsatisfactory, the dose may be increased to a maximum of 8 units/kg every 6 hours. Plicamycin (mithramycin;MITHRACIN) is a cytotoxic antibiotic that also decreases plasma Ca2+ concentrations by inhibiting bone resorption. Reduction in plasma Ca2+ concentrations occurs within 24 to 48 hours when a relatively low dose of this agent is given (15 to 25 g/kg body weight) to minimize the high systemic toxicity of the drug. Intravenous bisphosphonates (etidronate, pamidronate) have proven very effective in the management of hypercalcemia. These agents act as potent inhibitors of osteoclastic bone resorption. Oral bisphosphonates have been relatively unsuccessful for treating hypercalcemia. For this purpose, pamidronate (AREDIA ) is given as an intravenous infusion of 60 to 90 mg over 4 to 24 hours. With bisphosphonates, resolution of hypercalcemia occurs over several days, and the effect usually persists for several weeks. Gallium nitrate (GANITE) is a potent inhibitor of bone resorption that was approved for treating malignancy-associated hypercalcemia, but its utility was limited by nephrotoxicity. Gallium nitrate is not currently available in the United States. Oral sodium phosphate lowers plasma Ca concentrations and may offer short-term calcemic control of some patients with HPT who are awaiting surgery. However, the risk of precipitating calcium phosphate salts in soft tissues throughout the body is of concern. In light of satisfactory responses to other agents, administration of intravenous sodium phosphate cannot be recommended as a means to treat hypercalcemia.
2+

Edetate disodium (disodium EDTA; ENDRATE, others) is a chelating agent that forms soluble complexes with Ca2+. It is mentioned here only for historical purposes and is not currently recommended for any therapeutic use involving Ca2+. Chelation in the blood rapidly lowers Ca2+, with a substantial risk of cardiac, renal, and neurological toxicity. EDTA (as the calcium, disodium salt) still may be used for chelation therapy of heavy metal toxicity (see Chapter 67: Heavy Metals and Heavy-Metal Antagonists). Phosphate In addition to its role as a dynamic constituent of intermediary and energy metabolism, phosphate is an essential component of all body tissues. More than 80% of total body phosphorus occurs in bone, and about 15% is in soft tissue. Phosphorus is a component of membrane phospholipids. It modifies tissue concentrations of Ca 2+ and plays a major role in renal H + excretion. Absorption, Distribution, and Excretion Phosphate is absorbed from, and to a limited extent secreted into, the gastrointestinal tract. Transport of phosphate from the gut lumen is an active, energy-dependent process that is modified by several factors including vitamin D, which stimulates absorption. Presence of large quantities of Ca2+ or Al 3+ may lead to formation of large amounts of insoluble phosphate and diminish net phosphate absorption. In adults, about two-thirds of ingested phosphate is absorbed and is almost entirely excreted into the urine. In growing children, phosphate balance is positive. Concentrations of phosphate in plasma are higher in children than in adults. This "hyperphosphatemia" decreases the affinity of hemoglobin for oxygen and is hypothesized to explain the physiological "anemia" of childhood (Card and Brain, 1973). Phosphate is present in plasma and extracellular fluid, in cell membranes and intracellular fluid, and in collagen and bone tissue. In extracellular fluid, the bulk of phosphate exists in inorganic form as the two constituents, NaH2PO4 and Na2HPO4; the ratio of disodium to monosodium phosphate is 4:1 at pH 7.40. This ratio varies with pH; however, due to its relatively low concentration, phosphate contributes little to the buffering capacity of extracellular fluid. The concentration of plasma inorganic phosphate varies with age (Greenberg et al., 1960) and inversely with the rate of renal hydroxylation of 25-hydroxycholecalciferol (see below). A reduction of plasma phosphate concentration permits the presence of more Ca2+ in the blood without mineral precipitation. Renal phosphate excretion has been extensively studied. More than 90% of plasma phosphate is filterable, of which 80% is actively reabsorbed. Most reabsorption occurs in the initial segment of the proximal tubule, with a lesser component in the pars recta. The extent of phosphate reabsorption at more distal sites remains controversial (see Yanagawa and Lee, 1992). There is little evidence for tubular phosphate secretion in the mammalian kidney. Phosphate excreted in the urine represents the difference between the amount filtered and that reabsorbed. Expansion of plasma volume increases urinary phosphate excretion (Steele, 1970). PTH increases urinary phosphate excretion by blocking reabsorption. Vitamin D and its metabolites directly stimulate proximal tubular phosphate reabsorption (Puschett et al., 1972). Role of Phosphate in the Acidification of the Urine Although the concentration of phosphate is low in the extracellular fluid, the anion is progressively concentrated in the renal tubule and represents the most abundant buffer system in the distal tubule. At this site, the secretion of H+ by the tubular cell in exchange for Na+ in the tubular urine converts disodium hydrogen phosphate to sodium dihydrogen phosphate. In this manner, large amounts of

acid can be excreted without lowering the pH of the urine to a degree that would block H+ transport by a high concentration gradient between the tubular cell and luminal fluid. Actions of the Phosphate Ion Once phosphate gains access to body fluids and tissues, it exerts little pharmacological effect. If the ion is introduced into the intestine, the absorbed phosphate is rapidly excreted. If large amounts are given by this route, much of it may escape absorption. Because this property leads to a cathartic action, phosphate salts are employed as mild laxatives. Inorganic phosphate poisoning following ingestion of laxatives that contain phosphate salts has been reported in adults and children (McConnell, 1971). Ingestion of large amounts of sodium dihydrogen phosphate lowers urinary pH. If excessive phosphate salts are introduced intravenously or orally, they may prove toxic by reducing the concentration of Ca2+ in the circulation and from the precipitation of calcium phosphate in soft tissues (Vernava et al., 1987). Phosphate Depletion Phosphate is a ubiquitous component of ordinary foods; thus simple dietary inadequacy is not likely to cause phosphate depletion. Sustained abuse of aluminum-containing antacids, however, can severely limit phosphate absorption and result in clinical phosphate depletion, manifest as malaise, muscle weakness, and osteomalacia. Familial hypophosphatemia is an X-linked trait due to defective intestinal and/or renal handling of phosphate that results in rickets and dwarfism. Hypophosphatemia can decrease markedly erythrocyte ATP and 2,3-diphosphoglycerate content. Acute hemolytic anemia and impaired tissue oxygenation can occur in severe hypophosphatemia (Jacob and Amsden, 1971), raising the possibility that cellular stores of other high-energy phosphates also may be depleted. Pathological Conditions Associated with Disturbed Phosphate Metabolism Rickets The consequences of vitamin D deficiency with regard to the metabolism of both phosphate and calcium are described below, as are other forms of rickets. Familial hypophosphatemia is due to defective phosphate absorption and/or excretion, as mentioned above. Osteomalacia Osteomalacia is characterized by undermineralized bone matrix. Osteomalacia may occur when sustained phosphate depletion is caused by inhibiting its absorption in the gut (as with aluminumcontaining antacids) or by renal hyperexcretion due to PTH. Primary or Secondary Hyperparathyroidism In these disorders, the increase in PTH secretion reduces renal tubular reabsorption of phosphate and decreases the plasma inorganic phosphate concentration. By contrast, in hypo- or pseudohypoparathyroidism, deficient PTH action on the renal tubule leads to a rise in plasma phosphate concentrations. Chronic Renal Failure In this condition, the retention of phosphate is primary and reflects the degree of renal compromise.

Reduction of Ca2+ by the increased phosphate level stimulates hypersecretion of PTH, but since renal function is grossly impaired, hyperphosphatemia persists. The continuing hyperphosphatemia may be modified by vigorous administration of aluminum hydroxide gel or calcium carbonate supplements. Therapeutic Uses The phosphates have limited therapeutic utility. Sodium phosphate has been employed to diminish hypercalcemia (see above). The phosphates have a role in management of the phosphate-depletion syndrome and in chronic management of patients with vitamin Dresistant hypophosphatemic osteomalacia or rickets. Phosphate salts are also effective cathartics (see Chapter 39: Agents Used for Diarrhea, Constipation, and Inflammatory Bowel Disease; Agents Used for Biliary and Pancreatic Disease). Parathyroid Hormone (PTH) History Credit for the discovery of the parathyroid gland usually is given to Sandstrom, who in 1880 published an anatomical report that attracted little attention. The glands were rediscovered a decade later by Gley, who determined the effects of their extirpation with the thyroid. Vassale and Generali then successfully removed only the parathyroids and noted that tetany, convulsions, and death quickly followed. MacCallum and Voegtlin (1909) first noted the effect of parathyroidectomy on the concentration of plasma Ca2+. The relation of low plasma Ca2+ to symptoms was quickly appreciated, and a comprehensive picture of parathyroid function began to form. Active glandular extracts alleviated hypocalcemic tetany in parathyroidectomized animals and raised the concentration of Ca2+ in the plasma of normal animals (Berman, 1924; Collip, 1925). For the first time, the relation of clinical abnormalities to parathyroid hyperfunction was appreciated. While American and British investigators used physiological approaches to explore the function of the parathyroid glands, German and Austrian pathologists related the skeletal changes of osteitis fibrosa cystica to the presence of parathyroid tumors. In a delightful historical review, Albright (1948) traced the manner in which these two diverse types of investigations finally arrived at the same conclusion. Chemistry Human, bovine, and porcine parathyroid hormones are all single polypeptide chains of 84 amino acids. Their molecular masses approximate 9500 daltons, and the entire amino acid sequence has been established for each. Biological activity is associated with the N-terminal portion of the peptide; residues 1 to 27 are required for optimal binding to the PTH receptor and hormone activity. Derivatives lacking the first or second residue bind to PTH receptors but are virtually inert (Aurbach, 1988). Bovine and porcine PTH differ by seven amino acid residues, and the aminoterminal segment of human PTH differs from its bovine and porcine equivalents by only four and three residues, respectively. The three hormones differ little in biological activity but are distinguishable immunologically. Synthesis, Secretion, and Immunoassay

PTH is synthesized in a prehormone form. The 115amino-acid translation product that is destined to become PTH is called preproparathyroid hormone. This single-chain peptide is rapidly converted to proparathyroid hormone by cleavage of 25 amino-terminal residues as the peptide is transferred to the intracisternal space of the endoplasmic reticulum. Proparathyroid hormone moves to the Golgi apparatus for conversion to PTH by cleavage of six amino acids. PTH resides within secretory granules until it is secreted into the circulation. Most of the PTH normally is degraded by proteolysis before it can be secreted. During periods of hypocalcemia, more PTH is secreted and less is hydrolyzed. This mechanism provides a supply of hormone that can be mobilized rapidly in response to acute need without the delay entailed by increased synthesis of protein. In prolonged hypocalcemia, PTH synthesis also increases, and the gland hypertrophies. Neither preproparathyroid hormone nor proparathyroid hormone appears in plasma. The synthesis and processing of PTH have been reviewed by Kronenberg et al. (1994). Intact PTH has a half-life in plasma of 2 to 5 minutes; removal by the liver and kidney accounts for about 90% of its clearance. Metabolism of PTH releases fragments that circulate in blood. Fragments also are released by proteolysis of PTH within the parathyroid gland. Although these are not biologically active, they react with antibodies prepared against the intact hormone. Nonetheless, satisfactory PTH immunoassays have been developed for clinical use. Immunoradiometric assays using two monoclonal antibodies, one directed toward the amino-terminal and one toward the carboxyl-terminal portion of the hormone, permit accurate and sensitive detection of the intact PTH. These assays have replaced standard radioimmunoassays for clinical diagnostic purposes (see Nussbaum and Potts, 1994). Physiological Functions The primary function of PTH is to elicit the adaptive changes that maintain a constant concentration of Ca2+ in the extracellular fluid. Processes that are regulated include intestinal Ca2+ absorption, mobilization of bone Ca2+, and excretion of calcium in urine, feces, sweat, and milk (Figure 623). The actions of PTH on its target tissues are mediated by a cell surface G proteincoupled receptor (see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect); its predicted amino acid sequence and seven transmembrane spanning topography have been revealed by molecular cloning (Jppner et al., 1991). Figure 623. Calcium Homeostasis and Its Regulation by Parathyroid Hormone (PTH) and 1,25-Dihydroxyvitamin D. PTH has stimulatory effects on bone and kidney, including the stimulation of 1 -hydroxylase activity in kidney mitochondria leading to the increased production of 1,25-dihydroxyvitamin D (calcitriol) from 25-hydroxycholecalciferol, the monohydroxylated vitamin D metabolite (see also Figure 626). Calcitriol is the biologically active metabolite of vitamin D. Solid lines indicate a positive effect; dashed lines refer to negative feedback. See text for further explanation.

Regulation of Secretion The concentration of Ca2+ in plasma is the most powerful factor that regulates parathyroid gland secretory activity. When the concentration of Ca2+ is low, PTH secretion increases, and hypertrophy and hyperplasia of the gland result if the hypocalcemia is sustained. If the concentration of Ca2+ is high, PTH secretion decreases. In vitro studies show that amino acid transport, nucleic acid and protein synthesis, cytoplasmic growth, and PTH secretion are stimulated by exposure to low concentrations of Ca2+ and suppressed by high concentrations over an extended period of time. Thus, Ca2+ per se appears to regulate parathyroid gland growth as well as hormone synthesis and secretion. Changes in circulating Ca2+ regulate PTH secretion via a plasma membraneassociated calcium sensor on the surface of parathyroid cells (Brown et al., 1993). Binding of Ca2+ by this sensor inhibits PTH secretion, whereas reduced sensor occupancy promotes hormone secretion. Hypercalcemia is associated with a reduction of intracellular cyclic AMP content and protein kinase C (PKC) activity, and reduced circulating Ca2+ leads to activation of PKC. However, the precise link between these changes and alterations in hormone secretion is incompletely resolved. Other agents that increase parathyroid cell cyclic AMP levels, such as -adrenergic receptor agonists and dopamine, also increase PTH secretion, but the magnitude of response is far less than that seen with hypocalcemia. The active vitamin D metabolite, 1,25-dihydroxy-vitamin D (calcitriol), directly suppresses PTH gene expression. There appears to be no relation between physiological concentrations of extracellular phosphate and PTH secretion, except insofar as changes in phosphate concentrations alter circulating Ca2+. Severe hypermagnesemia and hypomagnesemia each can inhibit the secretion of PTH (Rude et al., 1976). The extracellular concentration of Ca2+ is controlled by an elaborate feedback system, the afferent

limb of which is sensitive to the ambient activity of Ca2+ and the efferent limb of which releases PTH. The hormone acts on various peripheral target tissues to mobilize Ca2+ into the extracellular fluid and thus to restore the concentration to normal. Effects on Bone PTH action on bone increases the delivery of Ca2+ to the extracellular fluid by increasing overall bone resorption, a process that involves the release of organic as well as mineral matrix components. The skeletal target cell for PTH probably is the osteoblast. With the exception of avian cells, specific receptors for PTH have not been described in osteoclasts, nor does an increase in resorption follow incubation of PTH with osteoclasts that are layered onto devitalized bone. Hormone responsiveness does appear if the osteoclasts are cultured in conditioned medium from osteoblasts that first have been exposed to PTH, suggesting an important role for osteoblasts in PTH-dependent bone resorption (McSheehy and Chambers, 1986; Perry et al., 1987; Takahashi et al., 1988). PTH recruits osteoclast precursor cells into forming new bone remodeling units. Sustained increases in circulating PTH result in characteristic histological changes in bone, including an increase in the prevalence of osteoclastic resorption sites and in the proportion of bone surface that is covered with unmineralized matrix. Although excessive osteoid surfaces may indicate defective mineralization, they signify in this case an increase in bone-forming surface due to an overall increase in remodeling activity. Direct effects of PTH on individual, incubated osteoblasts are generally inhibitory and include reductions in the formation of type I collagen, alkaline phosphatase, and osteocalcin. However, the measurable response in vivo to PTH reflects not only hormone action on individual cells but also the increased total number of active osteoblasts due to initiation of new remodeling units. Thus, plasma levels of osteocalcin and alkaline phosphatase activity actually may be increased. No simple model fully explains the molecular basis of PTH actions on bone. PTH stimulates cyclic AMP production in osteoblasts, but there also is evidence for a role for intracellular Ca2+ in mediating some of PTH's actions. Effects on Kidney PTH acts on the kidney to enhance the efficiency of Ca2+ reabsorption, to inhibit tubular reabsorption of phosphate, and to stimulate conversion of vitamin D to its hormonal form, 1,25dihydroxyvitamin D, or calcitriol (see Figure 623 and below). As a result, filtered Ca2+ is avidly retained, and its concentration increases in plasma; phosphate is excreted, and its plasma concentration falls. Calcitriol, at the same time, is secreted into the circulation, interacts with specific, high-affinity receptors in the intestine, and contributes to the rise in plasma Ca2+ concentration by improving the efficiency of gut Ca2+ absorption. Calcium PTH increases tubular reabsorption of Ca2+ at a distal site (Agus et al., 1973). When the plasma concentration of Ca2+ is normal, parathyroidectomy decreases tubular reabsorption of Ca2+ and thereby increases Ca2+ excretion in the urine. When the plasma concentration falls below 7 mg/dl (1.75 mM), a decrease in Ca2+ excretion occurs, because the amount of Ca2+ filtered through the glomeruli is lowered to the point that the cation is almost completely reabsorbed despite the reduced tubular capacity. If PTH is administered to hypoparathyroid animals or human beings, tubular reabsorption of Ca2+ is increased and Ca2+ excretion decreases. This effect, along with mobilization

of calcium from bone and increased absorption from the gut, results in an increased concentration of Ca2+ in plasma. When the value rises above normal, the increased glomerular filtration of Ca2+ overwhelms the stimulatory effect of PTH on tubular reabsorption, and hypercalciuria ensues. Phosphate PTH increases the renal excretion of inorganic phosphate by decreasing its reabsorption. Patients with primary hyperparathyroidism typically show low values for tubular phosphate reabsorption. Cyclic AMP mediates the renal effects of PTH (see Aurbach, 1988). PTH-sensitive adenylyl cyclase is located in the renal cortex, and cyclic AMP synthesized in response to the hormone affects tubular transport mechanisms. A portion of the cyclic nucleotide synthesized at this site escapes into the urine, and its assay serves as a measure of parathyroid activity and responsiveness. Other Ions PTH reduces renal excretion of Mg . This effect reflects the net result of increased renal Mg reabsorption and increased mobilization of the ion from bone (MacIntyre et al., 1963). PTH + + 2 increases excretion of water, amino acids, citrate, K , bicarbonate, Na , Cl , and SO4 , whereas it + decreases the excretion of H . Although the effects of PTH on renal acidbase metabolism are similar to those of acetazolamide, they are independent of the carbonic anhydrase system. Calcitriol Synthesis The final step in activation of vitamin D to its hormonal form, calcitriol, occurs in the kidney tubular cell (see later section on vitamin D). The activity of the hydroxylase enzyme involved in this step is governed by three primary regulators: inorganic phosphate, PTH, and Ca2+. Reductions in circulating or tissue phosphate content rapidly increase calcitriol production, whereas hyperphosphatemia suppresses it. PTH is a powerful initiator of calcitriol production, whereas hypercalcemia suppresses it. Thus, when hypocalcemia causes a rise in PTH concentration, both the PTH-dependent lowering of circulating inorganic phosphate and a more direct effect of the hormone on the hydroxylase lead to increased circulating concentrations of calcitriol. Miscellaneous Effects PTH decreases the concentration of Ca2+ in milk and saliva. These effects are the opposite of those that would be expected from the concurrent changes in plasma Ca2+ concentration. It appears, therefore, that the hormone can conserve Ca2+ in the extracellular fluid also by reducing the rate of Ca2+ transport to milk and saliva. Integrated Regulation of Extracellular Ca2+ Concentration by PTH The response of parathyroid cells to even modest reductions in Ca2+ occurs within minutes. For minute-to-minute regulation of Ca 2+, adjustments in renal Ca2+ handling more than suffice to maintain plasma calcium homeostasis. With more prolonged hypocalcemic stress, activation of the renal 1a-hydroxylase system leads to increased secretion of calcitriol, which directly stimulates intestinal calcium absorption (see Figure 623). In addition, increased delivery of labile calcium from bone into the extracellular fluid is stimulated. With a prolonged and severe hypocalcemic challenge, activation of new bone remodeling units helps to restore circulating Ca2+ concentrations, albeit at the expense of skeletal integrity.
2+ 2+

When plasma Ca2+ activity rises, PTH secretion is suppressed and tubular Ca2+ reabsorption decreases. The reduction in circulating PTH promotes renal phosphate conservation, and both of these reduce calcitriol production, thereby decreasing intestinal Ca2+ absorption. Finally, bone remodeling is suppressed. Thus, one may construct a coherent model for calcium homeostasis based on the participation of two hormones, PTH and 1,25-dihydroxyvitamin D, and involving the hierarchical contributions of kidney, intestine, and bone (Figure 623). The importance in human beings of other hormones, such as calcitonin, to this scheme remains unsettled, but it is likely that these modulate the Ca2+-parathyroid-vitamin D axis rather than serving as primary regulators. Hypoparathyroidism Hypoparathyroidism is only one of the many causes of hypocalcemia (see above) and occurs rarely. The deficiency state most commonly follows operative procedures on either the thyroid or parathyroid glands. Less frequently, the disorder stems from a genetic or autoimmune cause. Pseudohypoparathyroidism (PHP) is a disorder manifest by biochemical effects of hypoparathyroidism, but with elevated circulating levels of PTH. In this condition, end-organ responsiveness to PTH is severely impaired, frequently due to mutations in the adenylyl cyclaseG protein complex (see Levine, 1999). In all varieties of hypoparathyroidism, hypocalcemia and its associated symptoms are encountered clinically. The earliest symptoms of hypocalcemia are paresthesias in the extremities. Mechanical stimulation of peripheral nerves during physical examination may produce contraction of the appropriate skeletal muscles (Chvostek's sign). These signs and symptoms may be followed by tetany, consisting of muscle spasms, particularly of the hands and feet, and laryngospasm. Eventually, generalized convulsions and other central nervous system manifestations occur. Smooth muscle also is affected. Hypocalcemia may be attended by spasm of the ciliary muscle, iris, esophagus, intestine, urinary bladder, and bronchi. Electrocardiogram changes and a marked tachycardia indicate cardiac involvement. Vascular spasm in fingers and toes also is commonly observed. In chronic hypoparathyroidism, ectodermal changesconsisting of loss of hair, grooved and brittle fingernails, defects of dental enamel, and cataractsare encountered; calcification in the basal ganglia may be seen on routine skull radiographs. Psychiatric symptoms such as emotional lability, anxiety, depression, and delusions often are present. Hypoparathyroidism is treated primarily with vitamin D (see below). Dietary supplementation with Ca2+ also may be necessary. Hyperparathyroidism Primary hyperparathyroidism (HPT) results from hypersecretion of PTH by one or more parathyroid glands. Plasma concentrations of Ca2+ occasionally may be normal in HPT, but they usually are elevated. Plasma inorganic phosphate concentrations usually are low normal to decreased. Urinary Ca2+ excretion generally is increased, reflecting the dominant effect of filtered load over the conserving effect of PTH on tubular Ca2+ reabsorption. However, for any given level of plasma Ca2+, urinary Ca 2+ excretion in HPT is not as high as it would be in nonparathyroid hypercalcemic states. Secondary hyperparathyroidism results as a compensation for reductions in circulating Ca2+ and is not associated with hypercalcemia. In these cases, the concentration of inorganic phosphate is particularly low (except when associated with renal failure), and the serum alkaline phosphatase activity is very high. Severe primary or secondary HPT may be associated with a skeletal disorder known as osteitis fibrosa cystica. However, most patients with primary HPT have few if any skeletal findings. These

are generally restricted to a modest reduction in overall bone mineral density, particularly at sites of cortical bone. By contrast, patients with primary HPT generally show reasonable conservation trabecular density (see Bilezikian et al., 1994). The diagnosis of HPT has been simplified by the introduction of specific immunoradiometric assays for the intact PTH molecule. The combination of hypercalcemia and an elevated intact PTH concentration is sufficient to establish the diagnosis of HPT with greater than 90% accuracy. Treatment of HPT In the hands of a skilled parathyroid surgeon, resection of a single adenoma (about 80% of cases) or of the hyperplastic glands (about 15% of cases) leads to cure of HPT. Transient postoperative hypocalcemia may reflect temporary disruption of blood supply to remaining parathyroid tissue or skeletal avidity for calcium. Permanent hypoparathyroidism is a serious but unusual complication of parathyroid surgery that requires lifelong treatment with vitamin D and supplemental calcium. Clinical Uses of PTH PTH has no FDA-approved therapeutic use, but it may be approved in the near future for treatment 2+ of osteoporosis. Although it was used in the past to elevate the concentration of Ca in plasma, this 2+ can be accomplished with greater safety and efficacy by administration of Ca and/or vitamin D. Daily administration of PTH or its analogs has been shown to be of possible value in the treatment of patients with osteoporosis. While still an experimental strategy, substantial gains in axial bone mass have been observed in osteoporotic subjects treated with daily PTH(134) (see"Osteoporosis," below). PTH(134) can be used diagnostically to distinguish between pseudohypoparathyroidism and hypoparathyroidism. Since the former disease features target-organ resistance to the hormone, patients with PHP fail to increase their cyclic AMP excretion in response to acute administration of the peptide. Although this test is useful to characterize specific abnormalities in patients or families with PHP, clinical diagnosis usually can be made by measuring the circulating concentration of intact PTH. Vitamin D Traditionally, vitamin D was assigned a passive role in calcium metabolism in that its presence in adequate concentrations was thought to permit efficient absorption of dietary calcium and to allow full expression of the actions of PTH. It is now known that vitamin D has a much more active role in calcium homeostasis. Even though it is termed "vitamin" D, it is a hormone that, together with PTH, is a major regulator of the concentration of Ca2+ in plasma. The following characteristics of vitamin D are consistent with its hormonal nature: it is synthesized in the skin and under ideal conditions probably is not required in the diet; it is transported in blood to distant sites in the body, where it is activated by a tightly regulated enzyme; its active form binds to specific receptors in target tissues, resulting ultimately in an increased concentration of plasma Ca2+. Moreover, it is now known that receptors for the activated form of vitamin D are expressed in many cells throughout the body, including hematopoietic cells, lymphocytes, epidermal cells, pancreatic islets, muscle, and neurons; these receptors mediate a variety of actions that are unrelated to Ca2+ homeostasis. History Vitamin D is the name applied to two related fat-soluble substances, cholecalciferol and ergocalciferol, that have in common the ability to prevent or cure rickets. Prior to the discovery of vitamin D, a high percentage of urban children living in the temperate zones developed rickets.

Some researchers believed that the disease was due to lack of fresh air and sunshine; others claimed a dietary factor caused the disease. Mellanby (1919) and Huldschinsky (1919) showed both notions to be correct; addition of cod liver oil to the diet or exposure to sunlight prevented or cured the disease. In 1924, it was found that ultraviolet irradiation of animal rations was as efficacious in curing rickets as was irradiation of the animal itself (Hess and Weinstock, 1924; Steenbock and Black, 1924). These observations led to the elucidation of the structures of chole- and ergocalciferol and eventually to the discovery that these compounds require further processing in the body to become active. The discovery of metabolic activation is primarily attributable to studies conducted in the laboratories of DeLuca in the United States and Kodicek in England (see Kodicek, 1974; DeLuca and Schnoes, 1976). Chemistry and Occurrence Ultraviolet irradiation of several animal and plant sterols results in their conversion to compounds with vitamin D activity. Cleavage of the carbon-to-carbon bond between C 9 and C 10 is the essential alteration produced by the photochemical process, but not all sterols that undergo this cleavage possess antirachitic activity. The principal provitamin found in animal tissues is 7dehydrocholesterol, which is synthesized in the skin. Exposure of the skin to sunlight converts 7dehydrocholesterol to cholecalciferol (vitamin D3) (see Figure 624). Holick and associates have found an intermediate in the photolysis reactionprevitamin D3, a 6,7-cis isomer that accumulates in the skin after exposure to ultraviolet radiation (see Holick, 1981). This isomer slowly converts spontaneously to vitamin D3 and may provide a sustained source of D 3 for some time thereafter. Figure 624. Photobiogenesis and Metabolic Pathways for Vitamin D Production and Metabolism. Circled letters and numbers denote specific enzymes: 7, 7dehydrocholesterol reductase; 25, vitamin D-25 hydroxylase; 1 , 25-OHD-1 hydroxylase; 24R, 25-OHD-24R hydroxylase. (Adapted from Holick, 1981, with permission.)

Ergosterol, which is present in plants, is the provitamin for vitamin D2 (ergocalciferol). Ergosterol and vitamin D2 differ from 7-dehydrocholesterol and vitamin D3, respectively, only by each having a double bond between C 22 and C 23 and a methyl group at C 24. Vitamin D2 is the active constituent in a number of commercial vitamin preparations as well as in irradiated bread and irradiated milk. The material historically designated vitamin D1 was later shown to be a mixture of antirachitic substances. In some species the antirachitic potencies of vitamin D 2 and vitamin D3 differ greatly from each other. In human beings there is no practical difference between the two, and in the following discussion vitamin D will be used as the collective term for the two vitamers (Figure 625). Figure 625. Structures of 7-Dehydrocholesterol, Ergosterol, Cholecalciferol, and Ergocalciferol. Numbering system for steroid molecules is shown.

Metabolic Activation Both dietary and intrinsically synthesized vitamin D require activation to become biologically active. The primary active metabolite of the vitamin is calcitriol (1,25-dihydroxy-vitamin D), the product of two successive hydroxylations of vitamin D. The pathway of activation is shown in Figure 626. This subject has been reviewed by Horst and Reinhardt (1997). Figure 626. Regulation of 1a-Hydroxylase Activity. 25-OHD, 25hydroxycholecalciferol; 1,25-(OH)2-D, calcitriol; PTH, parathyroid hormone.

25-Hydroxylation of Vitamin D The initial step in the activation of vitamin D occurs in the liver, and the product is 25hydroxycholecalciferol (25-OHD, or calcifediol). The hepatic enzyme system responsible for 25hydroxylation of vitamin D is associated with the microsomal and mitochondrial fractions of homogenates and requires NADPH and molecular oxygen. 1-Hydroxylation of 25-OHD After production in the liver, 25-OHD enters the circulation, where it is carried by vitamin D binding globulin. Final activation to calcitriol occurs primarily in the kidney but also takes place in other sites, including macrophages (Reichel et al., 1989). The kidney is the predominant source of circulating calcitriol. The enzyme system responsible for 1-hydroxylation of 25-OHD is associated with mitochondria in the proximal tubules. It is a mixed-function oxidase and requires molecular oxygen and NADPH as cofactors. Cytochrome P450, a flavoprotein, and ferredoxin are components of the enzyme complex. The 1 -hydroxylase is subject to tight regulatory controls that result in changes in calcitriol

secretion appropriate for optimal calcium homeostasis. Enzyme activity increases in dietary deficiency of vitamin D, calcium, and phosphate; it is stimulated by PTH, and probably also by prolactin and estrogen. Conversely, its activity is suppressed by high calcium, phosphate, and vitamin D intake. Regulation is both chronic, suggesting changes in enzyme protein synthesis, as well as acute (Figure 626). In the case of PTH, a rapid increase in calcitriol production is mediated by cyclic AMP, apparently through an indirect stimulation of a phosphoprotein phosphatase that acts on the ferredoxin component of the hydroxylase (Siegel et al., 1986). There is evidence that hypocalcemia can activate the hydroxylase directly in addition to affecting it indirectly by eliciting secretion of PTH. Hypophosphatemia greatly increases hydroxylase activity (Haussler and McCain, 1977; Fraser, 1980; Rosen and Chesney, 1983). Calcitriol exerts negative-feedback control of the enzyme that reflects a direct action on the kidney as well as inhibition of PTH production. The nature of the regulatory mechanisms of estrogens and prolactin on the 1 -hydroxylase is not known. Physiological Functions, Mechanism of Action, and Pharmacological Properties Vitamin D is best characterized as a positive regulator of Ca2+ homeostasis. Phosphate metabolism is affected by the vitamin in a manner parallel to that of Ca2+. Although regulation of Ca2+ homeostasis is considered to be its primary function, increasing evidence indicates that vitamin D is important in a number of other processes (see below). The mechanisms by which vitamin D acts to maintain normal concentrations of Ca 2+ and phosphate in plasma are to facilitate their absorption by the small intestine, to interact with PTH to enhance their mobilization from bone, and to decrease their excretion by the kidney. A direct role of the vitamin in bone mineralization has been difficult to validate; rather, the predominant view is that normal bone formation occurs when Ca2+ and phosphate concentrations in the plasma are adequate. However, it is now clear that vitamin D has both direct and indirect effects on the cells that are involved in bone remodeling. The mechanism of action of calcitriol resembles that of the steroid and thyroid hormones. Calcitriol binds to cytosolic receptors within target cells, and the receptorhormone complex interacts with DNA to modify gene transcription. The calcitriol receptor belongs to the same supergene family as the steroid and thyroid hormone receptors (see Evans, 1988; Pike, 1992; see also Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). Calcitriol also exerts effects that occur so rapidly that they are interpreted as being nongenomic actions (see Barsony and Marx, 1988). Intestinal Absorption of Calcium A defect in intestinal absorption of Ca2+ in vitamin Ddeficient rats was demonstrated more than 50 years ago. Treatment of such animals with the activated hormone leads within 2 to 4 hours to increased movement of Ca2+ from the mucosal to the serosal surface of the intestine. The complex mechanisms underlying this action are not completely understood (see Wasserman, 1997). One relatively early event is the induction of one of a family of small Ca2+-binding proteins (CaBP, or calbindin). Some investigators propose that CaBP acts to facilitate passage of Ca2+ through the brush border and its diffusion to the basolateral membrane of mucosal cells; others contend that the accumulation of CaBP correlates poorly with Ca2+ transport (Nemere and Norman, 1986 and 1988) and propose instead that calcitriol enhances the endocytotic uptake of Ca2+ from the intestinal lumen into vesicles within the mucosal cell brush border. These vesicles fuse with lysosomes, which deliver Ca2+ to the basolateral membrane for extrusion (see Cancela et al., 1988). The mechanisms by which calcitriol might promote such vesicle-mediated transport have not been defined. Extrusion of calcium from the intestinal cell is accomplished by a plasma membrane

calcium pump, the number of which is increased by calcitriol (see Wasserman, 1997). Although the time to onset of effects in vitamin Ddeficient animals suggests the involvement of genomic mechanisms, calcitriol also causes a rapid (within minutes), receptor-mediated stimulation of Ca 2+ transport in vitamin Dreplete animals (see Cancela et al., 1988). Mobilization of Bone Mineral Although vitamin Ddeficient animals show obvious deficits in bone mineral, there is little evidence that vitamin D directly promotes mineralization; thus, it is thought that normal mineral deposition is sustained by maintenance of optimal plasma concentrations of Ca2+ and phosphate through promoting their intestinal absorption (see Stern, 1980). Indeed, children with vitamin D resistant rickets type II have been treated successfully with intravenous infusions of Ca2+ and phosphate (see below). In contrast, physiological doses of vitamin D promote mobilization of Ca 2+ from bone, and large doses cause excessive bone turnover. Although calcitriol-induced bone resorption may be reduced in parathyroidectomized animals, the response is restored when hyperphosphatemia is corrected (see Stern, 1980). Thus, both PTH and calcitriol act independently to enhance bone resorption. The mechanisms by which calcitriol increases bone turnover have been partially defined and involve the interaction of multiple factors (see Haussler, 1986; Reichel et al., 1989). Mature osteoclasts themselves do not appear to be directly acted upon by calcitriol, nor do they apparently contain calcitriol receptors. Instead, calcitriol promotes the recruitment of osteoclast precursor cells to resorption sites as well as the development of differentiated functions that characterize mature osteoclasts (Mimura et al., 1994). Osteopetrosis is a disease characterized by deficient bone resorption, in which osteoclast responsiveness to calcitriol or other bone resorbing agents is profoundly impaired. The cells responsible for bone formation (osteoblasts) do contain calcitriol receptors, and calcitriol causes them to elaborate several proteins, including osteocalcin, a vitamin Kdependent protein that contains -carboxyglutamic acid residues, and interleukin-1, a lymphokine that promotes bone resorption (Spear et al., 1988). Renal Retention of Calcium and Phosphate The effects of vitamin D on renal handling of Ca2+ and phosphate are of uncertain importance. Vitamin D increases retention of Ca2+ independently of phosphate and probably enhances reabsorption of each by the proximal tubules. Other Effects of Calcitriol It is now evident that the effects of calcitriol extend beyond calcium homeostasis. Receptors for calcitriol are distributed widely throughout the body (see Pike, 1992). Calcitriol affects maturation and differentiation of mononuclear cells and influences cytokine production. Its effects on the immune system have been reviewed by Amento (1987). One focus of current research is the potential therapeutic application of calcitriol's ability to inhibit proliferation and to induce differentiation of malignant cells (see Van Leeuwen and Pols, 1997). The possibility of dissociating the hypercalcemic effect of calcitriol from its actions on cell differentiation has encouraged the search for analogs that might be useful in cancer therapy. Calcitriol inhibits epidermal proliferation and promotes epidermal differentiation, thus establishing a basis for evaluating it as a potential treatment of psoriasis vulgaris (see Kragballe, 1997). The relation of vitamin D to skeletal muscle function has been reviewed by Boland (1986), and

effects of vitamin D in the brain have been discussed by Carswell (1997). Signs and Symptoms of Deficiency Deficiency of vitamin D results in inadequate absorption of Ca2+ and phosphate. The consequent decrease in plasma Ca2+ stimulates PTH secretion, which acts to restore plasma Ca2+ at the expense of bone; plasma concentrations of phosphate remain subnormal because of the phosphaturic effect of increased circulating PTH. In children, the result is a failure to mineralize newly formed bone and cartilage matrix, causing the defect in growth known as rickets. As a consequence of inadequate calcification, bones of individuals with rickets are soft, and the stress of weight bearing gives rise to characteristic deformities. In adults, vitamin D deficiency results in osteomalacia, a disease characterized by generalized accumulation of undermineralized bone matrix. Severe osteomalacia may be associated with extreme bone pain and tenderness. Muscle weakness, particularly of large proximal muscles, is typical. Its basis is not fully understood but may reflect hypophosphatemia and inadequate vitamin D action on muscle. Gross deformity of bone occurs only in advanced stages of the disease. Circulating 25-OHD concentrations below 8 ng/ml are highly predictive of osteomalacia. Hypervitaminosis D The acute or long-term administration of excessive amounts of vitamin D or enhanced responsiveness to normal amounts of the vitamin leads to clinically manifest derangements in calcium metabolism. The responses to vitamin D reflect endogenous vitamin D production, tissue reactivity, and vitamin D intake. Some infants may be hyperreactive to small doses of vitamin D. In adults, hypervitaminosis D results from overtreatment of hypoparathyroidism and from faddist use of excessive doses. Toxicity in children also may occur following accidental ingestion of adult doses. The amount of vitamin D necessary to cause hypervitaminosis varies widely among individuals. As a rough approximation, continued ingestion of 50,000 units or more daily by a person with normal parathyroid function and sensitivity to vitamin D may result in poisoning. Hypervitaminosis D is particularly dangerous in patients who are receiving digitalis, because the toxic effects of the cardiac glycosides are enhanced by hypercalcemia (see Chapters 34: Pharmacological Treatment of Heart Failure and 35: Antiarrhythmic Drugs). Signs and Symptoms The initial signs and symptoms of vitamin D toxicity are those associated with hypercalcemia (see above). Hypercalcemia with hypervitaminosis D is due generally to very high circulating levels of 25-OHD, and plasma concentrations of PTH and calcitriol are typically but not uniformly suppressed. In children, a single episode of moderately severe hypercalcemia may arrest growth completely for 6 months or more, and the deficit in height may never be fully corrected. Vitamin D toxicity may be manifested in the fetus. There is a relationship between excess maternal vitamin D intake or extreme sensitivity and nonfamilial congenital supravalvular aortic stenosis. In infants, this anomaly is often associated with other stigmata of hypercalcemia. Maternal hypercalcemia also may result in suppression of parathyroid function in the newborn, with resultant

hypocalcemia, tetany, and seizures. Treatment Treatment of hypervitaminosis D consists of immediate withdrawal of the vitamin, a low-calcium diet, administration of glucocorticoids, and vigorous fluid support. With this regimen the plasma Ca2+ falls to normal and Ca2+ in soft tissue tends to be mobilized. Conspicuous improvement in renal function occurs unless renal damage has been severe. Absorption, Fate, and Excretion Vitamin D usually is given by mouth, and intestinal absorption is adequate under most conditions. Both vitamins D2 and D3 are absorbed from the small intestine, although vitamin D3 may be absorbed more efficiently. The exact portion of the gut that is most effective in vitamin D absorption reflects the vehicle in which the vitamin is dissolved. Most of the vitamin appears first within chylomicrons in lymph. Bile is essential for adequate absorption of vitamin D; deoxycholic acid is the major constituent of bile in this regard. Thus, hepatic or biliary dysfunction seriously impairs vitamin D absorption. Absorbed vitamin D circulates in the blood in association with vitamin Dbinding protein, a specific -globulin. The vitamin disappears from plasma with a half-life of 19 to 25 hours but is stored in fat depots for prolonged periods. As discussed above, the liver is the site of conversion of vitamin D to 25-OHD, which circulates with the same binding protein. In fact, 25-OHD has a higher affinity for the protein than does the parent compound. The 25-hydroxy derivative has a biological half-life of 19 days and constitutes the major circulating form of vitamin D. Normal steady-state concentrations of 25-OHD in human beings are 15 to 50 ng/ml, although concentrations below 25 ng/ml may be associated with increased circulating PTH and greater bone turnover. The plasma half-life of calcitriol is estimated to be between 3 and 5 days in human beings, and 40% of an administered dose is excreted within 10 days (Mawer et al., 1976). Calcitriol is hydroxylated to 1,24,25-(OH)3D by a renal hydroxylase that is induced by calcitriol and suppressed by those factors that stimulate the 25-OHD-1 -hydroxylase. This enzyme also hydroxylates 25-OHD to form 24,25-(OH)2D. Both 24-hydroxylated compounds are less active than calcitriol and presumably represent metabolites destined for excretion. Side chain oxidation of calcitriol also occurs. The primary route of excretion of vitamin D is the bile; only a small percentage of an administered dose is found in urine. Vitamin D and its metabolites undergo extensive enterohepatic recirculation, and patients who have undergone intestinal bypass surgery or who otherwise have severe shortening or inflammation of the small intestine fail to reabsorb vitamin D sufficiently to maintain normal vitamin D nutriture. An important interaction has been demonstrated between vitamin D and phenytoin or phenobarbital. Rickets and osteomalacia have been reported in patients receiving chronic anticonvulsant therapy. More often, the drugs induce a state of high-turnover osteoporosis secondary to a decrease in intestinal Ca2+ absorption (Weinstein et al., 1984). Plasma concentrations of 25-OHD are decreased in patients receiving these drugs, and it was proposed that phenytoin and phenobarbital accelerate the metabolism of vitamin D to inactive products (Hahn et al., 1972). However, concentrations of calcitriol in plasma remain normal in most patients receiving anticonvulsant therapy (Jubiz et al., 1977). The drugs also accelerate hepatic metabolism of vitamin K and reduce the synthesis of

vitamin Kdependent proteins, such as osteocalcin. Human Requirements and Unitage An exhaustive and critical summary of the prophylactic requirements for vitamin D has been compiled by the Committee on Nutrition of the American Academy of Pediatrics (see Committee on Nutrition, 1963). Many years have elapsed since 1919, when Mellanby demonstrated the efficacy of cod liver oil in the prevention of rickets, a disease that has become a clinical rarity in the United States. Although sunlight provides adequate prophylaxis in the equatorial belt, in temperate climates insufficient cutaneous solar radiation in winter may necessitate dietary vitamin D supplementation. Previously, the recommended allowance of vitamin D could be achieved only by adding oral vitamin D supplements to a normal diet. Since the advent of the addition of the vitamin to foodstuffs (especially milk, milk products, cereals, and candy), individuals of all ages receive variable and even excessive vitamin D without its special addition to the diet. Thus, supplemental requirements vary not only with age, pregnancy, and lactation but also with diet quality. Serious toxicity may result from excessive ingestion of vitamin D, and even as little as 1800 USP units (see equivalency below) per day in infants may inhibit growth. Therefore, any recommendation for vitamin D supplementation must be made only after careful scrutiny of the diet. In both premature and normal infants, a total of 400 units per day of vitamin D ensures full antirachitic prophylaxis and optimal growth regardless of how it is obtained. During adolescence and beyond, this amount is probably also sufficient. There is some evidence that vitamin D requirements increase during pregnancy and lactation, although daily intake of 400 units is sufficient in these conditions as well (see Table XIII1). The USP unit is identical with the international unit (IU) and is equivalent to the specific biological activity of 0.025 g of vitamin D3 (i.e., 1 mg equals 40,000 units). Bioassay procedures used in the past depended upon alleviation of the rachitic state. They are still in use for experimental purposes. Modified Forms of Vitamin D Several derivatives of vitamin D are of considerable experimental and therapeutic interest. Dihydrotachysterol (DHT) is a vitamin D analog that may be regarded as a reduction product of vitamin D2 (and is sometimes referred to as DHT2); its structural formula is as

follows: DHT is about 1/450 as active as vitamin D in the antirachitic assay, but at high doses it is much more effective than vitamin D in mobilizing bone mineral. The latter effect is the basis for the use of DHT to maintain normal concentrations of Ca2+ in plasma in hypoparathyroidism. DHT undergoes 25-hydroxylation to yield 25-hydroxydihydrotachysterol (25-OHDHT), which appears to be the active form in both intestine and bone. 25-OHDHT is active in nephrectomized rats, indicating that it does not require 1-hydroxylation in the kidney. A comparison of the structures of DHT and 1,25-(OH)2D shows that ring A of DHT is rotated so as to place its 3hydroxyl group in approximately the same geometrical position as the 1-hydroxyl group of 1,25(OH)2D. It seems reasonable, therefore, to assume that 25-OHDHT could interact with receptor sites for 1,25-(OH)2D without undergoing 1-hydroxylation. Thus, DHT bypasses the renal mechanisms of metabolic control. 1 -Hydroxycholecalciferol (1-OHD3) is a synthetic derivative of vitamin D3 that is hydroxylated in the 1 position. It is readily hydroxylated in the 25 position by the hepatic microsomal system to form 1,25-(OH)2D and therefore was introduced as a substitute for the latter compound. In chick assays for stimulation of intestinal absorption of Ca2+ and bone mineralization, it is equal in activity to calcitriol. Because it does not require renal hydroxylation, it has been used to treat renal osteodystrophy. This drug is available in the United States for experimental purposes. Analogs of Calcitriol Calcipotriol (calcipotriene) contains a 2223 double bond, a 24(S)-hydroxy functional group, and carbons 2527 incorporated into a cyclopropane ring. This compound has receptor affinity similar to that of calcitriol, but it is less than 1% as active as calcitriol in regulating calcium metabolism. Calcipotriol has been studied extensively as a treatment for psoriasis (see Chapter 65: Dermatological Pharmacology), and a topical preparation (DOVONEX) is available for that purpose. In clinical trials, topical calcipotriol has been found to be an effective and safe treatment, slightly more effective than glucocorticoids. The mode of action of calcipotriol in psoriasis is not known. Paricalcitol (ZEMPLAR) is a synthetic calcitriol derivative that reduces PTH production without producing hypercalcemia, except in overdoses. It has been approved by the FDA for treatment of secondary hyperparathyroidism in patients with chronic renal failure. 22-Oxacalcitriol also is a potent suppressor of PTH gene expression and shows very limited activity on intestine and bone. It is, therefore, an attractive compound for use in patients with

overproduction of PTH in chronic renal failure or even with primary hyperparathyroidism (Finch et al., 1993). Therapeutic Uses Many preparations containing vitamin D are marketed. Ergocalciferol (calciferol;DRISDOL) is pure vitamin D2. It is available for oral, intramuscular, or intravenous administration. Dihydrotachysterol (DHT;HYTAKEROL) is the pure crystalline compound obtained by reduction of vitamin D2 and is available for oral administration. Calcifediol (25-hydroxycholecalciferol;CALDEROL) also is available for oral use. Calcitriol (1,25-dihydroxycholecalciferol; CALCIJEX, ROCALTROL) is available for oral administration or injection. The major therapeutic uses of vitamin D may be divided into four categories: (1) prophylaxis and cure of nutritional rickets; (2) treatment of metabolic rickets and osteomalacia, particularly in the setting of chronic renal failure; (3) treatment of hypoparathyroidism; and (4) prevention and treatment of osteoporosis. Nutritional Rickets Nutritional rickets results from inadequate exposure to sunlight or deficiency of dietary vitamin D. The condition is extremely rare in the United States and other countries where food fortification with the vitamin is practiced. Infants and children receiving adequate amounts of vitamin D fortified food do not require additional vitamin D; however, breast-fed infants or those fed unfortified formula should receive 400 units of vitamin D daily as a supplement. The usual practice is to administer vitamin A in combination with vitamin D. A number of well-balanced vitamin A and D preparations are available for this purpose. Premature infants are especially susceptible to rickets and may require supplemental vitamin D, since the fetus acquires more than 85% of its calcium stores during the third trimester. The curative dose of vitamin D for the treatment of fully developed rickets is larger than the prophylactic dose. One thousand units daily will normalize Ca2+ and phosphate concentrations in plasma in approximately 10 days, and radiographic evidence of healing is seen within about 3 weeks. However, a daily dose of 3000 to 4000 units often is prescribed for more rapid healing; this is of particular importance in severe cases of thoracic rickets when respiration is embarrassed. Certain conditions are known to lead to poor absorption of vitamin D. If these are untreated by vitamin supplementation, a frank deficiency may develop. Vitamin D may be of definite prophylactic value in such disorders as diarrhea, steatorrhea, biliary obstruction, and any other abnormality in gastrointestinal function in which absorption is appreciably diminished. Parenteral administration may be used in such cases. Metabolic Rickets and Osteomalacia This group of disorders is characterized by abnormalities in the synthesis of or the response to calcitriol. Hypophosphatemic vitamin Dresistant rickets, in its most characteristic form, is an X-linked disorder of calcium and phosphate metabolism (XLH). Although calcitriol levels are normal, they would be predicted to be higher for the degree of hypophosphatemia that is observed. Patients experience clinical improvement when treated with large doses of vitamin D, usually in combination with inorganic phosphates. However, even with vitamin D treatment, calcitriol

concentrations may remain lower than expected. A specific mutation giving rise to the most common form of XLH has been described (HYP Consortium, 1995). The affected protein, called PEX, is a neutral endoprotease. The specific substrate for this enzyme has not been clarified, but it is considered likely to be involved in renal phosphorus transport. Closely related syndromes to XLH have been described, including hereditary hypophosphatemic rickets with hypercalciuria (HHRH) and autosomal dominant hypophosphatemic rickets. The precise mechanisms for transmission and pathophysiology of these variant conditions also are unknown (see Econs and Drezner, 1992). Vitamin Ddependent rickets is an autosomal recessive disease caused by an inborn error of vitamin D metabolism involving defective conversion of 25OHD to calcitriol. The condition responds to physiological doses of calcitriol (Fraser et al., 1973). Hereditary 1,25-dihydroxyvitamin D resistance (also called vitamin Ddependent rickets type II) is an autosomal recessive disorder that is characterized by hypocalcemia, osteomalacia and rickets, and complete alopecia. Studies of skin fibroblasts from these patients have identified mutations in the calcitriol receptor that lead either to defective hormone binding or defective binding of the hormonereceptor complex to DNA. The latter mutations result from single amino acid substitutions on the zinc-finger portion of the DNA-binding domain of the vitamin D receptor (see Pike, 1992). Affected children are completely unresponsive to massive doses of vitamin D and calcitriol, and they may require prolonged treatment with parenteral Ca2+. Some remission in symptoms has been observed during adolescence, but the basis for improvement is not known. Renal osteodystrophy (renal rickets) is associated with chronic renal failure and is characterized by decreased conversion of 25-OHD to calcitriol. Phosphate retention decreases plasma Ca 2+ concentrations, leading to secondary hyperparathyroidism. In addition, calcitriol deficiency impairs intestinal Ca2+ absorption and mobilization from bone. Hypocalcemia commonly results (although in some patients, prolonged and severe hyperparathyroidism eventually may lead to hypercalcemia). Aluminum deposition in bone also may play a role in the genesis of the skeletal disease. Pathologically, lesions are typical of hyperparathyroidism (osteitis fibrosa), deficiency of vitamin D (osteomalacia), or a mixture of both. In patients with chronic renal failure who are not receiving dialysis, emphasis has been on treatment of hyperphosphatemia with phosphate binders and calcium supplementation; these goals can be accomplished by the oral administration of calcium carbonate, combined with dietary phosphate restriction (Coburn and Salusky, 1989). Use of vitamin D analogs in predialysis patients is experimental, but it is clearly beneficial for patients who are undergoing dialysis. Administration of calcitriol raises the concentration of Ca2+ in plasma, lowers the concentration of PTH, and helps to maintain bone mineralization and growth in children (Berl et al., 1978; Chesney et al., 1978). Intravenous calcitriol may be effective in patients who are refractory to oral therapy (Andress et al., 1989). DHT and 1-OHD3 also can be used effectively, since renal hydroxylation is not required for their activity. Although 25-OHD also may be effective, high doses must be used. Hypoparathyroidism Hypoparathyroidism is characterized by hypocalcemia and hyperphosphatemia (see above). DHT has long been used to treat this condition, since it has a faster onset, shorter duration of action, and a greater effect on bone mobilization than does vitamin D. Calcitriol is effective in the management of hypoparathyroidism and at least certain forms of pseudohypoparathyroidism in which endogenous levels of calcitriol are abnormally low. However, most hypoparathyroid patients respond to any form of vitamin D. Calcitriol may be the agent of choice for temporary treatment of

hypocalcemia while waiting for a slower-acting form of vitamin D to become effective. Miscellaneous Uses of Vitamin D These include treatment of hypophosphatemia seen in the Fanconi syndrome. The use of large doses of vitamin D (over 10,000 units/day) in patients with osteoporosis is not of value and can be dangerous. However, administration of 400 to 800 units/day of vitamin D to frail, elderly men and women has been shown to suppress bone remodeling, protect bone mass, and reduce fracture incidence (see later section on osteoporosis). Clinical trials suggest that calcitriol may become an important agent for the treatment of psoriasis (see Holick, 1993; Kragballe, 1992). As such nontraditional uses of vitamin D are discovered, it will become important to develop noncalcemic analogs of calcitriol that achieve effects on cellular differentiation without the risk of hypercalcemia. Calcitonin History and Source A hypocalcemic hormone, the effects of which are generally opposite to those of PTH, was discovered and named calcitonin by Copp in 1962 (see Copp, 1964). It was demonstrated as a result of perfusion of canine parathyroid and thyroid glands with hypercalcemic blood, which caused an acute and transient hypocalcemic effect occurring significantly earlier than the hypocalcemia of total parathyroidectomy. Copp concluded that the parathyroid glands secreted calcitonin in response to hypercalcemia and in this way normalized plasma Ca2+ concentrations. Munson and colleagues (Hirsch et al., 1963) noted that parathyroidectomy in rats performed by cauterization caused more severe hypocalcemia than did thyroparathyroidectomy and suspected the existence of a hypocalcemic principle in the thyroid gland. They found that extracts of thyroid produced hypocalcemia, and named this factor thyrocalcitonin. It is now known that these two factors are the same and that the hormone does originate from the thyroid; however, calcitonin (CT) is the name that is generally used. The parafollicular C cells from the thyroid, which are embryologically derived from neural crest ectoderm, are the site of production and secretion of CT. In nonmammalian vertebrates, CT is found in ultimobranchial bodies, which are separate organs from the thyroid gland. In human beings, C cells are widely distributed in the thyroid, parathyroid, and thymus. Chemistry and Immunoreactivity CT is a single-chain peptide of 32 amino acid residues. Eight of these residues are invariant, including a carboxyl-terminal prolinamide and a disulfide bridge between cysteines at positions 1 and 7. Both these structural features are essential for biological activity. The residues in the middle portion of the molecule (positions 10 to 27) are variable and appear to influence potency and/or duration of action. CTs derived from the ultimobranchial bodies of salmon and eel are more potent than mammalian thyroidal CTs both in vivo and in vitro, and they differ from the human hormone by 13 and 16 amino acid residues, respectively. Therapeutically, salmon CT appears to be more potent than human CT, in part because it is cleared more slowly from the circulation. Human CT is processed from a propeptide of 135 amino acid residues; two additional peptides are generated, but their biological significance is unknown. The calcitonin gene contains six exons; calcitonin itself is encoded by exon 4. In C cells, messenger RNA is processed such that exons 1 to 4 are represented in the final transcript. In neural tissue, the sequence corresponding to exon 4 is

removed, and the sequences for exons 1 to 3, 5, and 6 are included. Following translation and proteolytic cleavage of a precursor molecule, a mature peptide of 37 amino acids is generated, the calcitonin generelated peptide (CGRP). CGRP mimics some effects of CT in some species but causes PTH-like effects in others and acts on receptors distinct from those that mediate the actions of CT. Since little or no CGRP is produced by the thyroid C cells, it is unlikely to function in calcium homeostasis. CGRP and its binding sites are widely distributed in the central nervous system (CNS), where it is believed to serve as a neurotransmitter. CGRP is found in many bipolar neurons in sensory ganglia and produces marked vasodilation. The structure and synthesis of CT and CGRP have been reviewed by MacIntyre and coworkers (1987). Multiple forms of CT are found in plasma, including high-molecular-weight aggregates or crosslinked products. This fact has impeded development of useful immunoassays for CT. Assays for the intact monomeric peptide have been introduced (Body and Heath, 1983). Regulation of Secretion Biosynthesis and secretion of CT are regulated by the concentration of Ca2+ in plasma. When plasma Ca2+ is high, CT secretion increases; when plasma Ca 2+ is low, CT secretion is low or undetectable. Normal circulating CT concentrations in human beings are less than 10 pg/ml (Body and Heath, 1983). Mean concentrations of CT in women are lower than those in men, as are responses to the secretagogues pentagastrin and Ca2+. The circulating half-life of CT is about 10 minutes. CT secretion can be stimulated by a number of agents, including catecholamines, glucagon, gastrin, and cholecystokinin, but is there little evidence for a physiological role for secretion in response to these stimuli. It is not even known whether or not CT plays a significant role in calcium homeostasis in human beings. Thyroidectomized patients with no detectable CT have normal calcium metabolism and bone mass. High concentrations (50 to 5000 times normal) of CT occur in plasma, urine, and tumor tissue of patients with medullary carcinoma of the thyroid. The tumor cells originate from the thyroid parafollicular cells, and the disease represents a true CT-excess syndrome. Measurement of the response of plasma CT to infusions of calcium gluconate and pentagastrin is the standard procedure to detect the condition (Wells et al., 1978). Because one form of this disease is inherited as a dominant trait [multiple endocrine neoplasia type II (MEN II)], relatives of patients should be examined repeatedly from early childhood (Tashjian et al., 1974). Localization of the mutation for MEN II to the RET protooncogene offers hope that genetic screening will supplant calcium/pentagastrin tests (Donis-Keller et al., 1993; Carlson et al., 1994). Mechanism of Action The hypocalcemic and hypophosphatemic effects of calcitonin are caused predominantly by direct inhibition of osteoclastic bone resorption (see MacIntyre et al., 1987). Although CT inhibits the effects of PTH on osteolysis, it does not act as a global inhibitor of PTH. It does not block activation of bone cell adenylyl cyclase by PTH and does not inhibit PTH-induced uptake of Ca2+ into bone. The actions of calcitonin are not blocked by inhibitors of RNA and protein synthesis. CT interacts directly with receptors on osteoclasts to produce a rapid and profound decrease in ruffled border surface area, thereby diminishing resorptive activity. Depressed bone resorption leads to reduced urinary excretion of Ca2+, Mg2+, and hydroxyproline. Plasma phosphate concentrations also are lowered, due also to increased urinary phosphate excretion. Direct renal effects of CT vary with species. In human beings, CT promotes excretion of

Ca2+, phosphate, and Na +. At least some of the actions of CT on kidney and bone are mediated by cyclic AMP (Murad et al., 1970; Heersche et al., 1974). Bioassay of CT preparations is performed by assessing their ability to lower the plasma Ca2+ concentration in the rat. Salmon and eel CTs are more potent than are human and porcine CTs (see above). Therapeutic Uses CT lowers Ca2+ and phosphate concentrations in patients with hypercalcemia, the effect of a single dose lasting 6 to 10 hours. This effect results from decreased bone resorption and is greater in patients in whom bone turnover rates are high. Although CT is effective in the initial treatment of hypercalcemia, escape from the response is observed after a few days. Use of this agent does not substitute for aggressive fluid resuscitation, and the response to other agents, such as bisphosphonates, may be more satisfactory (see Hypercalcemic States). CT is effective in disorders of increased skeletal remodeling, such as Paget's disease, and in some patients with osteoporosis (see below). In Paget's disease, chronic use of CT produces long-term reduction in symptoms and in serum alkaline phosphatase activity. Development of antibodies to CT does occur with long-term therapy, but this is not necessarily associated with clinical resistance. After initial therapy at 100 units/day, favorable results usually are obtained when dosage is reduced to 50 units three times a week. Side effects of CT include nausea, hand swelling, urticaria, and, rarely, intestinal cramping. Side effects appear to occur with equal frequency with human and salmon CT. Salmon CT is approved for clinical use as CALCIMAR or MIACALCIN . The latter is now available as a nasal spray, introduced for once-daily treatment of postmenopausal osteoporosis (see below). Subcutaneous or intramuscular doses of from 100 units up to 8 units/kg every 12 hours have been used to treat hypercalcemia. An initial dose of 100 units per day is used for Paget's disease, with reduction to 50 units three times a week once a response has occurred. Bisphosphonates Bisphosphonate is the name given to a group of drugs characterized by geminal bisphosphonate bond (Figure 627). These compounds, when added to appropriate solutions and suspensions of calcium phosphate, slow the formation and dissolution of hydroxyapatite crystals. The first bisphosphonate to be developed for clinical use was etidronate (Figure 627), the most potent mineralization inhibitor of this group. Subsequent clinical experience has shown that inhibition of mineralization actually constitutes a disadvantage, leading over time to osteomalacia. Thus, secondand third-generation bisphosphonates have been developed that minimize this action. The clinical utility of bisphosphonates resides in their ability to inhibit bone resorption. The mechanism by which this antiresorptive effect occurs is not completely known, but it is thought that the bisphosphonate becomes incorporated into bone matrix and is imBIBed by osteoclasts during resorption. Bisphosphonates affect osteoclasts by at least two different mechanisms. Some bisphosphonates, such as etidronate, clodronate, and tiludronate, are metabolized into an ATP analog (AppCCl2p) that accumulates within and impairs the function and viability of cells (Frith et al., 1996). By contrast, potent aminobisphosphonates, such as alendronate and ibandronate, are not metabolized but directly inhibit multiple steps in the pathway from mevalonate to cholesterol and isoprenoid lipids, such as geranylgeranyl diphosphate, that are required for the prenylation of various proteins that are important for osteoclast function (Luckman et al., 1998).

Figure 627. Structures of Bisphosphonates.

Several bisphosphonates currently are available in the United States. Etidronate sodium (DIDRONEL) is used for treatment of Paget's disease of bone and may be used parenterally to treat hypercalcemia. As etidronate is the only bisphosphonate to inhibit mineralization, it will likely be completely replaced by newer members of this class. Pamidronate (AREDIA) (Figure 627) is approved for management of hypercalcemia but has been found to be effective in other skeletal disorders. Pamidronate is available in the United States only for parenteral administration. For treatment of hypercalcemia, pamidronate may be given as an intravenous infusion of 60 to 90 mg over 4 to 24 hours. Several newer bisphosphonates have been approved for treatment of Paget's disease. These include tiludronate (SKELID ), alendronate (FOSAMAX), and risedronate (ACTONEL). All oral bisphosphonates are absorbed very poorly from the intestine. Thus, it is important to administer these drugs following an overnight fast and at least 30 minutes before breakfast. They should be taken only with a full glass of water. Therapeutic Uses Paget's Disease Paget's disease is a skeletal condition of single or multiple foci of disordered bone remodeling. Pagetic lesions are characterized by many abnormal multinucleated osteoclasts in association with a disordered "mosaic" pattern of bone formation. Pagetic bone is thickened and has abnormal microarchitecture. The alterations in bone structure may produce secondary problems, such as

deafness, spinal cord compression, high-output cardiac failure, and pain. Malignant degeneration to osteogenic sarcoma is a rare but lethal complication of Paget's disease. Bisphosphonates and CT decrease the elevated biochemical markers of bone turnover, such as plasma alkaline phosphatase activity and urinary excretion of hydroxyproline. Typically an initial course of bisphosphonate is given once daily for 6 months. With treatment, most patients experience a decrease in bone pain over several weeks. Such treatment may induce long-lasting remission. If symptoms recur, additional courses of therapy can be effective. When etidronate is given at higher doses (10 to 20 mg/kg per day) or continuously for longer than 6 months, there is a substantial risk for osteomalacia. At lower doses (5 to 7.5 mg/kg per day) focal osteomalacia has been observed occasionally. Defective mineralization has not been observed with other bisphosphonates or with CT. Choice of optimal therapy for Paget's disease varies among patients. Bisphosphonates have the advantage of oral administration, lower cost, lack of antigenicity, and generally being less prone to side effects compared to CT. However, CT is highly reliable and may have a distinct skeletal analgesic property. Some evidence suggests that control of Paget's disease may be more effective when bisphosphonate and CT are used in combination (O'Donoghue and Hosking, 1987). Mithramycin has been used occasionally in difficult cases of Paget's disease. Therapeutic utility of this agent is limited by a high potential for toxicity, and it is not generally recommended. Hypercalcemia Etidronate and pamidronate have been used successfully in the management of malignancyassociated hypercalcemia. Etidronate has been used in the hope that its antimineralizing effect would benefit patients with heterotopic formation of bone or myositis ossificans. Results have not been impressive. Postmenopausal Osteoporosis Much interest currently is focused on the role of bisphosphonates in treatment of osteoporosis (see"Osteoporosis," below). Recent clinical trials show that treatment is associated with increases in bone mineral density and protection against fracture (see below). Fluoride Fluoride is of interest because of its toxic properties and its effect on dentition and bone. Fluoride is distributed widely in nature, and soils of different regions of the world vary greatly in their fluoride content. Sources of atmospheric fluoride include the burning of soft coal and the manufacturing of superphosphate, aluminum, steel, lead, copper, and nickel. Human beings obtain fluoride in particular from the ingestion of plants and water. Absorption, Distribution, and Excretion Fluorides are absorbed from the intestine, lungs, and skin. The intestine is the major site of absorption. The degree of absorption of a fluoride compound is best correlated with its solubility. The relatively soluble compounds, such as sodium fluoride, are almost completely absorbed, whereas relatively insoluble compounds, such as cryolite (Na3AlF 6) and the fluoride found in bone meal (fluoroapatite), are poorly absorbed. The second most common route of absorption is via the lungs. Inhalation of fluoride present in dusts and gases constitutes the major route of industrial exposure.

Fluoride has been detected in all organs and tissues, and it is concentrated in bone, thyroid, aorta, and perhaps kidney. Fluoride is primarily deposited in bone and teeth, and the skeletal burden is related to intake and age. Storage in bone reflects skeletal turnover, growing bone showing greater deposition than bone in mature animals. The major route of fluoride excretion is via the kidneys; however, small amounts of fluoride appear in sweat, milk, and intestinal secretions. When sweating is excessive, the fraction of total fluoride excretion in sweat can reach nearly one-half. About 90% of the fluoride filtered by the glomerulus is reabsorbed by the renal tubules. Pharmacological Actions The pharmacological actions of fluoride, with the possible exception of its effects on bone and teeth, can be classified as toxic. Fluoride is an inhibitor of several enzyme systems and diminishes tissue respiration and anaerobic glycolysis. Fluoride also is a useful anticoagulant in vitro because it binds Ca2+. It also inhibits erythrocyte glycolysis. For this reason, fluoride is added to specimen tubes for blood glucose determinations. Fluoride is a mitogen for osteoblasts and stimulates bone formation (Baylink et al., 1970). Thus, fluoride has been an attractive agent for potential use in osteoporosis. Many, but not all, patients treated with fluoride salts show substantial increase in trabecular bone mass, whereas cortical bone responds less well. It remains to be established whether or not increases in axial bone mineral promote bone strength and protect against fracture (see"Osteoporosis," below). The radionuclide 18 F has been used in skeletal imaging (Jones et al., 1973). Acute Poisoning Acute fluoride poisoning is not rare. It usually results from accidental ingestion of fluoridecontaining insecticides or rodenticides. Initial symptoms (salivation, nausea, abdominal pain, vomiting, and diarrhea) are secondary to the local action of fluoride on the intestinal mucosa. Systemic symptoms are varied and severe. There is 2+ increased irritability of the nervous system, consistent with the Ca -binding effect of fluoride, hypocalcemia, and hypoglycemia. The blood pressure falls, presumably owing to central vasomotor depression as well as direct cardiotoxicity. Respiration is first stimulated and later depressed. Death usually results from respiratory paralysis or cardiac failure. The lethal dose of sodium fluoride for human beings is about 5 g, although there is considerable variation. Treatment includes the intravenous administration of glucose in saline and gastric lavage with lime water (0.15% calcium hydroxide solution) or other Ca2+ salts to precipitate the fluoride. Calcium gluconate is given intravenously for tetany; urine volume is kept high with vigorous fluid resuscitation. Chronic Poisoning In human beings, the major manifestations of chronic ingestion of excessive fluoride are osteosclerosis and mottled enamel. Osteosclerosis is characterized by increased bone density secondary both to elevated osteoblastic activity and to the replacement of hydroxyapatite by the denser fluoroapatite. The degree of skeletal involvement varies from changes that are barely detectable radiologically to marked cortical thickening of long bones, numerous exostoses scattered throughout the skeleton, and calcification of ligaments, tendons, and muscle attachments. In its severest form it is a disabling and crippling disease.

Mottled enamel, or dental fluorosis, is a well-recognized entity that was first described more than 60 years ago. The gross changes in very mild mottling consist of small, opaque, paper-white areas scattered irregularly over the tooth surface. In severe cases, discrete or confluent, deep brown- to black-stained pits give the tooth a corroded appearance. Mottled enamel results from a partial failure of the enamel-forming cells to elaborate and lay down enamel. It is a nonspecific response to a variety of stimuli, one of which is the ingestion of excessive amounts of fluoride. Since mottled enamel is a developmental injury, ingestion of fluoride following the eruption of teeth has no effect. Mottling is one of the first visible signs of excessive fluoride intake during childhood. Continuous use of water containing about 1.0 ppm of fluoride may result in very mild mottling in 10% of children; at 4.0 to 6.0 ppm the incidence approaches 100%, with marked increase in severity. Severe dental fluorosis formerly occurred with high frequency in regions of the world (e.g., Pompeii, Italy, and Pike's Peak, Colorado) where local water supplies had a very high fluoride content. The water supply for some regions of the arid American Southwest contains very high concentrations of fluoride, and skeletal fluorosis is common in animals that graze in these areas. Federal regulations currently require lowering the fluoride content of the water supply or providing an alternative source of acceptable drinking water for affected communities. Sustained consumption of water with a fluoride content of 4 mg/liter has been shown to be associated with deficits in cortical bone mass and increased rates of bone loss over time (Sowers et al., 1991). Fluoride and Dental Caries Experiments in controlling the fluoride content of water took an unexpected and significant turn when it was observed that children born at Bauxite, Arkansas, after a new water supply had been obtained, showed a much higher incidence of caries than those who had been exposed to the former fluoride-containing water. This led to extensive studies by the United States Public Health Service to ascertain whether water fluoridation could be a practical measure to reduce the incidence of tooth decay. It has now been established definitely, on the basis of many large-scale studies, that regulation of water fluoride content to 1.0 ppm is a safe and practical public health measure that substantially reduces the incidence of caries in permanent teeth. There are partial benefits for children who begin drinking fluoridated water at any age; however, optimal benefits are obtained at ages before permanent teeth erupt. Topical application of fluoride solutions by dental personnel appears to be particularly effective on newly erupted teeth and can reduce the incidence of caries by 30% to 40%. Prescription of dietary fluoride supplements should be considered for children under the age of 12 years whose drinking water contains less than 0.7 ppm fluoride. Conflicting results have been reported from studies of fluoride-containing toothpastes. Adequate incorporation of fluoride into teeth causes the outer layers of enamel to be harder and more resistant to de-mineralization. Deposition of fluoride appears to be an anion-exchange process with hydroxyl or citrate ions. Fluoride occupies the anionic spaces in the enamel apatite crystal surface. The mechanism of caries prevention by fluoride is not completely understood. There is no convincing evidence that fluoride from any source reduces the development of caries after the permanent teeth are completely formed (usually about age 14). The fluoride salts usually employed in dentifrices are sodium fluoride and stannous fluoride. Sodium fluoride also is available in a variety of preparations for oral and topical use, including tablets, drops, rinses, and gels. Sodium fluoride, sodium fluorosilicate (Na2SiF6), and cryolite are

the salts commonly used in insecticides. Since its inception, regulation of the fluoride concentration of community water supplies periodically has encountered vocal opposition from a number of groups. The nature of such opposition has ranged from strictly political rhetoric to allegations about putative adverse health consequences of fluoridated water. Careful examination of these issues in studies sponsored by the National Cancer Institute and by the United States Public Health Service indicate that mortality from cancer and all-cause mortality do not differ significantly between communities with fluoridated and nonfluoridated water (Hoover et al., 1976; Erickson, 1978). Osteoporosis Osteoporosis is a condition of low bone mass and microarchitectural disruption that results in fractures with minimal trauma. Characteristic sites of fracture include vertebral bodies, the distal radius, and the proximal femur, but osteoporotic individuals have generalized skeletal fragility, and fractures at other sites, such as ribs and long bones, also are common. Osteoporosis is a major and growing public health problem for older women and men in western society. It is described generally as primary or secondary. Secondary osteoporosis is due to systemic illness or medications such as glucocorticoids or phenytoin. The most successful approach to secondary osteoporosis is prompt resolution of the underlying cause. However, mechanisms of secondary osteoporosis all can be related in terms of disordered bone remodeling, so that the same therapeutic strategies may apply to these conditions as are appropriate for primary osteoporosis. In 1948, Albright and Reifenstein concluded that primary osteoporosis was composed of two separate entities: one related to menopausal estrogen loss, and the other to aging. Support for this concept has been published by Riggs and associates (1982), who proposed that primary osteoporosis represents two fundamentally different conditions: type I osteoporosis, loss of trabecular bone due to estrogen lack at menopause, and type II osteoporosis, loss of cortical and trabecular bone in men and women, due to long-term remodeling inefficiency, dietary inadequacy, and activation of the parathyroid axis with age. Compelling evidence has not been presented that these two entities are truly distinct, and the model fails to account for decreased bone mass resulting from inadequate skeletal acquisition during growth. Although many osteoporotic women undoubtedly have experienced excessive menopausal bone loss, it may be more appropriate to consider osteoporosis as the result of multiple physical, hormonal, and nutritional factors acting alone or in concert. Skeletal Organization Because bone turnover rates differ from one portion of the skeleton to the next, it is useful to consider the appendicular, or peripheral, skeleton separate from the axial, or central, skeleton. Appendicular bones make up 80% of whole-body bone mass and are composed predominantly of compact cortical bone. Axial bones, such as the spine and pelvis, contain substantial amounts of trabecular bone within a thin cortex. Trabecular bone consists of highly connected bony plates that resemble honeycomb. The intertrabecular interstices contain bone marrow and fat. For several reasons, alterations in bone turnover are observed first and most extensively in axial bone rather than in the appendicular skeleton. These include the facts that bone remodeling takes place on bone surfaces, that there is a higher surface density in trabecular bone compared to cortical bone, and that marrow precursor cells that ultimately participate in bone turnover lie in close proximity to trabecular surfaces.

Bone Mass Bone mineral density (BMD) and fracture risk in later years reflect the maximal bone mineral content at skeletal maturity (peak bone mass) and the subsequent rate of bone loss. Major increases in bone mass, accounting for about 60% of final adult levels, occur during adolescence, mainly during years of highest growth velocity. Bone acquisition is almost complete by age 17 years in girls and by 20 years in boys. Inheritance accounts for much of the variance in bone acquisition; other factors include circulating estrogen and androgens, physical activity, and dietary calcium. Bone is lost during adult life. Radiographic measurements of metacarpal bone by Garn and colleagues (1966) described a characteristic trajectory of bone mass throughout life, by which bone mass levels off during the third decade, remains stable until age 50, and then progressively declines. Similar trajectories occur for men, women, and all ethnic groups. The fundamental accuracy of this model has been amply confirmed for cortical bone, although trabecular bone loss probably begins prior to age 50 at some sites. In women, loss of estrogen at menopause accelerates the rate of bone loss for several years. The primary regulators of adult bone mass include physical activity, reproductive endocrine status, and calcium intake. Optimal maintenance of BMD requires sufficiency in all three areas, and deficiency in one area is not compensated by excessive attention to another. For example, amenorrheic athletes lose bone despite frequent high-intensity exercise (Marcus et al., 1985). Prevention and Treatment of Osteoporosis A rational strategy to prevent osteoporosis follows from the above considerations. Regular physical activity of reasonable intensity is endorsed at all ages. For children and adolescents, adequate dietary calcium is important if peak bone mass is to reach the level appropriate for genetic endowment. Attention to nutritional status may be required in the seventh decade and beyond, taking the form of increased dietary calcium or of calcium and/or vitamin D supplements. For women at menopause, timely administration of estrogen is the most powerful intervention to preserve bone and protect against fracture. Indeed, at any age, prevention or correction of hypogonadism is an important consideration. With appropriate lifelong attention to these preventive factors, important reductions in fracture risk can be achieved. Pharmacological agents used to manage osteoporosis act by decreasing the rate of bone resorption, thereby slowing the rate of bone loss, or by promoting bone formation. The only drugs currently approved in the United States for use in osteoporosis are those that decrease resorption. Since bone remodeling is a coupled process, antiresorptive drugs ultimately decrease the rate of bone formation. Thus, antiresorptive therapy cannot lead to substantial gains in BMD. Increases in BMD that typically are seen during the first years of antiresorptive therapy represent a constriction of the remodeling space to a new steady-state level, after which BMD reaches a plateau. One consequence of this phenomenon is that therapeutic trials in osteoporosis must be of sufficient duration to determine whether an increase in BMD represents anything more than a simple reduction in remodeling space. At least 2 years are required for this purpose. Antiresorptive Agents Calcium The physiological roles of Ca2+ and its use in the treatment of hypocalcemic disorders are discussed above. The rationale for using supplemental calcium to protect bone varies with time of life. For

preteens and adolescents, adequate substrate calcium is required for bone accretion. Controlled trials indicate that supplemental calcium promotes adolescent bone acquisition (Johnston et al., 1992; Lloyd et al., 1993), but its impact on peak bone mass is not known. Higher calcium intake during the third decade of life is positively related to the final phase of bone acquisition (Recker et al., 1992). There is controversy about the role of calcium during the early years after menopause, when the primary basis for bone loss is estrogen withdrawal. Although little effect of calcium on trabecular bone has been reported, reduction in cortical bone loss with calcium supplementation has been observed, even in populations characterized by high dietary calcium (Riis et al., 1987). In elderly subjects, supplemental calcium suppresses bone turnover, improves BMD, and decreases the incidence of fracture (Chapuy et al., 1992; Recker et al., 1996; Dawson-Hughes, et al., 1997). Patients who are unable or unwilling to increase dietary calcium via dietary means alone may choose from many palatable, low-cost calcium preparations. Numerous Ca2+ salts are available for human use, the most frequently prescribed being the carbonate. Other salts available include calcium lactate, gluconate, phosphate, and citrate as well as hydroxyapatite. Lead contamination of some lots of powdered bone diminishes the acceptability of this product. Calcium citrate may be more efficiently absorbed than other salts. However, absorption efficiency for most commonly prescribed calcium products is reasonable, and for many patients cost and palatability outweigh modest differences in efficacy. Traditional dosing of calcium is about 1000 mg/day, nearly the amount present in a quart of milk. Added to the 500 to 600 mg of dietary calcium that typifies the diet of elderly men and women, this provides a total daily intake of about 1500 mg. More than this amount may be necessary to overcome endogenous intestinal calcium losses, but daily intakes of 2000 mg or more frequently are reported to be constipating. Calcium supplements are most often taken with meals. Vitamin D and Its Analogs The physiological role of vitamin D and its metabolites is discussed above, as are their uses in the treatment of hypocalcemic disorders, rickets, and osteomalacia. Modest supplementation with vitamin D (400 to 800 IU per day) may improve intestinal Ca2+ absorption, suppress bone remodeling, and improve BMD in individuals with marginal or deficient vitamin D status. Supplemental vitamin D has been shown to reduce fracture incidence in two European trials (Chapuy et al., 1992; Heikinheimo et al., 1992). The use of calcitriol to treat osteoporosis is distinct from assuring vitamin D nutritional adequacy. Here, the rationale is directly to suppress parathyroid function and reduce bone turnover. Calcitriol and another polar vitamin D metabolite, 1 hydroxycholecalciferol, are used frequently in Japan and other countries (Fujita, 1992; Tilyard et al., 1992), but experience in the United States has been mixed. Higher doses of calcitriol appear to be more likely to improve BMD, but at the risk of hypercalciuria and hypercalcemia, so that close scrutiny of patients and dose modification are required. Restriction of dietary calcium may reduce toxicity during calcitriol therapy (Gallagher and Goldgar, 1990). A low incidence of hypercalciuric and hypercalcemic complications of therapy in Japan may reflect relatively poor calcium intakes in that country. Polar metabolites of vitamin D are promising for future study, but their toxicity makes it premature to endorse them for widespread use. Estrogen Overwhelming evidence confirms a major role for menopausal estrogen replacement in the conservation of bone and protection against osteoporotic fracture (Lindsay et al., 1976; Horsman et al., 1977; Recker et al., 1977; Hutchinson et al., 1979; Weiss et al., 1980). Studies indicate that 17 -estradiol acts on osteoblasts to decrease production of interleukin-6 and to upregulate the production of osteoprotegerin, thereby interfering with recruitment of osteoclast precursors

(Girasole et al., 1992). As sole therapy, the minimum effective dose of estrogen for skeletal protection is 0.625 mg/day of conjugated equine estrogens (PREMARIN ) or its equivalent. Both oral and transcutaneous estrogen decrease bone turnover and conserve bone. Cessation of estrogen eventually results in bone loss, so therapy should be long term. Standard practice recommends cyclic or continuous administration of progestational drugs to women with an intact uterus. The C-21 progestins (e.g., medroxyprogesterone acetate) do not interfere with the skeletal effects of estrogen, whereas androgenic progestins (e.g., norethisterone) may actually increase BMD and provide added skeletal benefit when added to estrogen (Christiansen and Riis, 1990). For women without a uterus, estrogen therapy is continuous and does not require adding a progestin. The optimal time to institute estrogen replacement is early menopause, when bone turnover accelerates. However, even for women beyond age 65, beneficial skeletal effects of estrogen are observed. Many older women will not accept cyclic bleeding or other anticipated side effects of estrogen. Thus, initiation of estrogen therapy to elderly women must be individualized. For other discussion of estrogens and progestins see Chapter 58: Estrogens and Progestins. Selective Estradiol Receptor Modulators (SERMS) Considerable work has been undertaken to develop estrogenic compounds with tissue-selective activities. One of these, raloxifene (EVISTA), acts as an estrogen agonist on bone and liver, is inactive on the uterus, and acts as an antiestrogen on the breast (see Chapter 58: Estrogens and Progestins). In postmenopausal women, raloxifene stabilizes and modestly increases BMD and has been shown to reduce the risk of vertebral compression fracture (Delmas et al., 1997; Ettinger et al., 1999). Raloxifene is approved for both prevention and treatment of osteoporosis. Calcitonin (CT) The physiological role and therapeutic use of CT for hypercalcemia and Paget's disease are discussed above. As a powerful inhibitor of osteoclastic bone resorption, CT produces modest increases in bone mass in patients with osteoporosis (Gruber et al., 1984; Civitelli et al., 1988; Mazzuoli et al., 1986). Increases are most impressive in patients with high intrinsic rates of bone turnover (Civitelli et al., 1988), approaching 10% to 15% before reaching a plateau. These represent simple reductions in the remodeling space. Recent experience with calcitonin nasal spray (MIACALCIN ), 200 units/day, indicates that this agent reduces the incidence of vertebral compression fracture by about 40% in women with osteoporosis. Bisphosphonates The use of these antiresorptive agents to treat hypercalcemia and Paget's disease is discussed above. Bisphosphonates have emerged as the most effective drugs currently approved for prevention and treatment of osteoporosis. Although osteomalacia is a worrisome side effect of etidronate, newer bisphosphonates have sufficient potency to suppress bone resorption at doses that do not inhibit mineralization. The first of these to be developed for osteoporosis was alendronate (FOSAMAX). In a 3-year clinical trial involving postmenopausal women with low BMD and prevalent vertebral fractures, alendronate improved BMD and reduced fracture incidence (Black et al., 1996). Women assigned to alendronate showed approximately 50% fewer vertebral and nonvertebral fractures, including hip fracture, than women taking placebo. In a companion study, women who had low BMD but no prevalent vertebral fractures on entry experienced a significant reduction in vertebral fracture incidence (Cummings et al., 1998). Alendronate conserves BMD in recently menopausal

women (Hosking et al., 1998) as well as in men and also improves BMD in patients receiving glucocorticoids (Saag et al., 1998). Alendronate is approved for prevention and treatment of postmenopausal osteoporosis and for treatment of glucocorticoid-associated osteoporosis. The approved prevention dose is 5 mg/day, and the treatment dose is 10 mg/day. Although alendronate was well tolerated in clinical trials, some patients experience symptoms of esophagitis. Symptoms often abate when patients fastidiously take the medication with water and remain upright. Where symptoms persist despite these precautions, use of a proton pump inhibitor (see Chapter 37: Agents Used for Control of Gastric Acidity and Treatment of Peptic Ulcers and Gastroesophageal Reflux Disease) at bedtime may be helpful. Alendronate may be better tolerated on a 40-mg once weekly schedule, and efficacy appears not to diminish. For patients with severe esophageal distress despite these countermeasures, intravenous pamidronate (AREDIA ) offers skeletal protection without causing adverse gastrointestinal side effects. For treatment of osteoporosis, pamidronate is given as a 3-hour infusion, 30 mg every 3 months. It is well tolerated. Mild fever and aches may attend the first infusion of pamidronate, but these symptoms are shortlived and generally do not recur with subsequent administration. Risedronate (ACTONEL), 5 mg/day, improves BMD and reduces vertebral fracture incidence in postmenopausal women (HARRISet al., 1999). Risedronate will likely be approved for this indication in the near future. Ibandronate is another potent bisphosphonate currently under development. Thiazide Diuretics Although not strictly antiresorptive, thiazides reduce urinary Ca2+ excretion and constrain bone loss in patients with hypercalciuria. Whether they will prove to be useful in patients who are not hypercalciuric is not clear, but data suggest that they reduce hip fracture risk. Hydrochlorothiazide, 25 mg once or twice daily, may achieve substantial reductions in calciuria. Effective doses of thiazides for reducing urinary Ca2+ excretion generally are lower than those necessary for control of blood pressure. For a more detailed discussion of thiazide diuretics, see Chapter 29: Diuretics. Bone-Forming Agents Fluoride The skeletal consequences of excessive fluoride and the role of water fluoridation to prevent dental caries are discussed above. Sodium fluoride increases bone volume, an effect due specifically to increased osteoblastic activity (Baylink et al., 1970; Briancon and Meunier, 1981). In doses of 30 to 60 mg/day, fluoride increases trabecular BMD in many, but not all patients. A controlled clinical trial (Riggs et al., 1990) reported no protective effect against vertebral compression fractures although fluoride increased lumbar spine density. This same trial showed a significant increase in the occurrence of peripheral fractures and stress fractures in the fluoride group. This study has been criticized for the high dose of fluoride that was used (75 mg/day), and another study (Mamelle et al., 1988) found that 30 to 50 mg/day of sodium fluoride decreased fracture risk. A recent trial of sustained-release fluoride, which is associated with lower blood fluoride levels, has shown favorable results on fracture incidence (Pak et al., 1994). However, the results of Riggs et al. (1990) clearly show that increased bone mass is not synonymous with increased bone strength and that, if any dose of fluoride proves useful, there will be a narrow therapeutic window. Androgen Testosterone replacement therapy increases BMD in hypogonadal men. Androgens also improve BMD in osteoporotic women, but therapy is limited by virilizing side effects. Nandrolone

decanoate (50 mg by injection every three weeks) increases peripheral and axial BMD without bothersome side effects in osteoporotic women. The androgenic progestin norethisterone acetate acts synergistically with estrogen to increase BMD in osteoporotic women (Christiansen and Riis, 1990). Adequate fracture data are not yet available to permit a conclusion on the clinical utility of this approach (see Chapter 59: Androgens for further discussion of androgens.) Parathyroid Hormone (PTH) The deleterious skeletal effects of PTH in patients with severe hyperparathyroidism are described above. An anabolic effect of intermittent administration of PTH on trabecular bone has been shown. Consequently, several laboratories have examined the effects of PTH on BMD in patients with osteoporosis. In these studies, the synthetic analog hPTH(134) increased axial bone mineral, although effects on cortical bone were disappointing. Coadministration of hPTH(134) with estrogen or synthetic androgen led to impressive gains in axial mineral without loss of cortical bone (Lindsay et al., 1997). PTH has been shown to induce substantial gains in BMD in patients with glucocorticoid-associated osteoporosis (Lane et al., 1998). Phase III controlled clinical trials of PTH and its analogs are currently in progress.