You are on page 1of 23

ANESTHETIC PHARMACOLOGY INTERNATIONAL SOCIETY FOR ANAESTHETIC PHARMACOLOGY

SECTION EDITOR
JAMES G. BOVILL
REVIEW ARTICLE

Inhaled Anesthetics and Immobility: Mechanisms, Mysteries,


and Minimum Alveolar Anesthetic Concentration
James M. Sonner, MD*, Joseph F. Antognini, MD†, Robert C. Dutton, MD*, Pamela Flood, MD‡,
Andrew T. Gray, MD, PhD*, R. Adron Harris, PhD§, Gregg E. Homanics, PhD储,
Joan Kendig, PhD¶, Beverley Orser, MD#, Douglas E. Raines, MD**, James Trudell, PhD¶,
Bryce Vissel, PhD††, and Edmond I Eger II, MD*
*Department of Anesthesia and Perioperative Care, University of California, San Francisco, California; †Department of
Anesthesiology, University of California, Davis, California; ‡Columbia University, New York, New York; §University of
Texas, Austin, Texas; 储University of Pittsburgh, Pittsburgh, Pennsylvania; ¶Stanford University, Palo Alto, California;
#University of Toronto, Toronto, Canada; **Department of Anaesthesia, Harvard Medical School, Cambridge,
Massachusetts; and ††Garvan Institute of Medical Research, Darlinghurst, Australia

Studies using molecular modeling, genetic engineer- present likely candidates for mediation. However, in
ing, neurophysiology/pharmacology, and whole ani- vivo studies to date suggest that several channels or re-
mals have advanced our understanding of where and ceptors may not be mediators (e.g., ␥-aminobutyric acid
how inhaled anesthetics act to produce immobility A, acetylcholine, potassium, 5-hydroxytryptamine-3,
(minimum alveolar anesthetic concentration; MAC) by opioids, and ␣2-adrenergic), whereas other receptors/
actions on the spinal cord. Numerous ligand- and channels (e.g., glycine, N-methyl-d-aspartate, and so-
voltage-gated channels might plausibly mediate MAC, dium) remain credible candidates.
and specific animo acid sites in certain receptors (Anesth Analg 2003;97:718 –40)

A
fter decades of ignorance, we now know how one characteristic of inhaled anesthetics: their capacity
some anesthetics (e.g., etomidate) act (1). This to cause immobility. The standard measure of anes-
knowledge resulted from information revealed thetic potency, MAC (the minimum alveolar concen-
by multiple approaches to studies of anesthetic mech- tration of an inhaled anesthetic required to suppress
anisms, including studies in whole animals using old- movement in response to noxious stimulation in 50%
fashioned pharmacologic techniques, in vivo and in of subjects), reflects this capacity (2). MAC is an anes-
vitro neurophysiological studies, and the application thetic 50% effective dose (ED50).
of genetic engineering. Genetic engineering allows the At the turn of the last century, Overton (3) and
manipulation of receptors and their study in vitro. It Meyer (4) reported that anesthetic potency correlated
allows the creation of animals (usually mice) with with lipophilicity, suggesting the importance of a
mutated receptors that do not respond to anesthetics lipid-like phase to anesthetic action. For conventional
and that, thereby, serve as tools to test the importance and many experimental inhaled anesthetics (5–9), the
of a given receptor to a particular anesthetic effect. correlation is remarkable (Fig. 1). Over the ensuing 80
This review discusses how these approaches have in- years, a few theories of narcosis evolved from the
creased understanding of the mechanisms by which a focus on lipids, particularly the membrane bilayer, but
particular class of anesthetics, the inhaled anesthetics, none of these theories withstood experimental scru-
acts. The review focuses on the mechanistic bases for tiny. For example, inhaled anesthetics increase mem-
brane disorder (increase fluidity) (10 –14), and this was
Dr. Eger is a paid consultant to Baxter Healthcare Corp.
proposed as a basis for anesthesia. However, increases
This work was supported by National Institute of General Med- in body temperature also increase this disorder, but
ical Sciences Grant 1P01GM47818. increases in body temperature increase rather than
Accepted for publication May 21, 2003. decrease MAC (15). The discovery of inhaled com-
Address correspondence and reprint requests to James M. Sonner,
MD, Department of Anesthesia, S-455, University of California, San pounds with potencies that do not correlate with li-
Francisco, CA 94143-0464. Address e-mail to sonnerj@anesthesia. pophilicity called into question the Meyer-Overton
ucsf.edu. hypothesis: alcohols are more potent than predicted
DOI: 10.1213/01.ANE.0000081063.76651.33 from their lipophilicity, whereas some compounds are

©2003 by the International Anesthesia Research Society


718 Anesth Analg 2003;97:718–40 0003-2999/03
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 719
2003;97:718 –40 MECHANISMS UNDERLYING MAC

release of neurotransmitters such as glutamate (24 –


26), but not necessarily all neurotransmitters (27). Be-
cause such studies identify candidate receptors, they
are essential for determining the plausibility of spe-
cific receptors as mediators of immobility, but such
identification does not prove relevance.
Antognini and Schwartz (28), Antognini et al. (29),
Borges and Antognini (30), Rampil et al. (31), and
Rampil (32) provided another key finding: that in-
haled anesthetic-induced immobility largely results
from an action on the spinal cord rather than an action
on higher centers. Their results informed many sub-
sequent studies, several of which are described in this
review.
Finally, the increasing capacity to study and manip-
ulate receptors through genetic engineering provides
a powerful tool for studies of anesthetic mechanisms.
Results of such work complement work from tradi-
tional pharmacology. For example, as will be dis-
cussed, the nonimmobilizer F6 blocks acetylcholine
Figure 1. The potencies (minimum alveolar anesthetic concentra- receptors in vitro, and thus the acetylcholine receptor
tion; MAC) of conventional [nitrous oxide (7), halothane (6), isoflu- is an unlikely target for the immobilizing effect of
rane (6), sevoflurane (6), desflurane (6), and cyclopropane (6)] and inhaled anesthetics. Consistent with this finding, in
several unconventional [xenon (9), 1A (1-chloro-1,2,2-
trifluorocyclobutane) (5), benzene (8), and hexafluorobenzene (8)] vivo blockade of acetylcholine receptors does not affect
inhaled anesthetics in rats correlate with their affinity to a lipid MAC.
phase, as defined by the oil/gas partition coefficient. A caveat:
although this relationship applies to all conventional inhaled anes-
thetics, several compounds (e.g., nonimmobilizers) do not conform
to the relationship. The Spinal Cord as the Primary Site
Mediating Immobilization
less potent (transitional compounds) (5). For example, Several studies demonstrate that the spinal cord me-
ethanol has an oil/gas partition coefficient of 108 (16), diates most of the ability of inhaled anesthetics to
similar to the value of 98 for isoflurane (17). However, produce immobility. The observation a decade ago
the MAC of isoflurane in rats is 1.5% (18,19), more that immobility during noxious stimulation does not
than 10 times larger than the MAC of 0.1% for ethanol correlate with electroencephalographic (EEG) activity
(16). At an extreme are nonimmobilizer compounds, prompted the hypothesis that cortical electrical activ-
compounds that have no anesthetic effect, alone or in ity does not control motor responses to noxious stim-
combination with known anesthetics, despite a li- ulation (33). In rats, precollicular decerebration (31)
pophilicity that predicts an anesthetizing capacity (5). (Fig. 2) or complete section of the upper thoracic spi-
For example, 1,2-dichlorohexafluorobutane (also nal cord (32) minimally affects isoflurane’s capacity to
called F6 or 2N) has an oil/gas partition coefficient of suppress movement. In vivo electrophysiologic exam-
44 (5), similar to the value of 47 found for sevoflurane ination of the spinal cord suggests that anesthetics can
(20), which has a MAC of 2.4% (21,22). However, F6 is suppress both sensory (34) and motor (35) neuron
not anesthetic at any concentration, including concen- activity. In rats (36) and humans (37), suppression of
trations exceeding 4% (5). movement in response to noxious stimulation corre-
The seminal work of Franks and Lieb (23) shifted lates with suppression of motor neuron excitability (as
attention from lipids to proteins (specifically, ion determined by recurrent impulses known as F waves).
channels) as anesthetic targets. Investigations soon In goats, MAC for isoflurane delivered to the whole
produced a plethora of plausible anesthetic targets, for body is 1.2%, but delivery to only the brain increases
the most part ligand- and voltage-gated channels. MAC to nearly 3% (28). The separation with halothane
Clinical concentrations of inhaled anesthetics could is still greater (29). Control halothane MAC is 0.9%,
enhance the effect of inhibitory neurotransmitter- but delivery confined to the head requires 3.4% to
gated or voltage-gated (notably, potassium) channels abolish movement. Some goats move during EEG si-
(i.e., increase the sensitivity of the channels to putative lence and move at the largest cerebral halothane con-
transmitters) or block the effect of excitatory centrations tested. These results indicate that inhaled
neurotransmitter-gated or voltage-gated (notably, so- anesthetics act primarily on the spinal cord to produce
dium) channels. They also could block the evoked immobility and that only a minor component of this
720 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

application of GABA or glycine. This action reduces


the excitability of neurons and might contribute to
anesthetic depressant actions. If this is the case, then
blockade of these channels should oppose the depres-
sant effects of anesthetics. In intact spinal cord, block-
ade of GABAA and glycine inhibitory chloride chan-
nels does, in fact, decrease the depressant effects of
anesthetics (41). However, whether blockade is only
partial reduction or is complete cannot be determined
experimentally in this preparation because of uncoor-
dinated fluctuations in the baseline brought about by
the convulsant effects of these blocking drugs.
Acetylcholine receptors are also affected by volatile
anesthetics and alcohols, but in opposite directions.
Although acetylcholine receptors can profoundly af-
Figure 2. Decerebration does not change the minimum alveolar
anesthetic concentration (MAC) of isoflurane in rats (31). These data fect spinal neuronal transmission, block of nicotinic or
and similar data gathered in Rampil’s laboratory (32) and Antogni- muscarinic receptors does not modify anesthetic po-
ni’s laboratory (28,30) demonstrate that the spinal cord mediates tency, either in vivo (42,43) or in vitro (44). Thus, in-
much or most of the capacity of inhaled anesthetics to produce
immobility. That is, effects on the spinal cord determine MAC. hibitory chloride channels, but not acetylcholine re-
ceptors, may be implicated in anesthetic-induced
immobility.
immobility results from cerebral effects. The common Spinal cord slice preparations allow isolation of
results from different species and experimental meth- postsynaptic actions on motor neurons from actions on
ods increase the compelling nature of these primary afferents and interneurons. In slices, ethanol
conclusions. (44) and volatile anesthetics (45) depress currents evoked
But where within the cord do inhaled anesthetics act by glutamate application, proving that these agents can
to produce immobility? One observation suggests a directly depress motor neuron excitability. Furthermore,
potential complexity: if coordinated movement occurs ethanol (44) and volatile anesthetics (45) depress both
during anesthesia, it often extends beyond the area ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
stimulated. For example, stimulation of the tail of a rat (AMPA) and N-methyl-d-aspartic acid (NMDA) recep-
provokes movement of the lower and the upper ex- tor-mediated currents by actions independent of
tremities. This example suggests initial signal process- GABAA and glycine receptors, suggesting that anes-
ing by sacral-coccygeal components, transmission to thetic reduction of motor output can result from depres-
lumbar and cervical components by ascending projec-
sion of excitation as well as enhancement of inhibition.
tion tracts, and transmission through cervical compo-
Anesthetic effects on the cord are diverse and com-
nents systems to effect movement. The next section
plex, and the ultimate effect on MAC is incompletely
shows that the specifics of where and how (as well as
understood. Spontaneous miniature current frequency
not where and not how), including possible putative
(presynaptically determined), amplitude, and kinetics
receptors, are now partially known.
(postsynaptically determined) reveal pre- and post-
synaptic anesthetic actions on motor neurons. Halo-
thane, isoflurane, and enflurane prolong glycinergic
What Has Been Learned from In Vitro current duration but do not increase amplitude (46), in
Spinal Cord Studies? contrast to nonneuronal expression systems, in which
Because motor neurons integrate all input and signal both amplitude and duration are increased. The three
muscles to move, anesthetic effects on motor neurons anesthetics also increase the frequency of spontaneous
may be important to the production of immobility. transmitter release, but only when sodium channels
Other spinal cord components (primary afferent ter- are blocked (46). With intact sodium channels, halo-
minals and interneurons) likely also play a role. The thane and isoflurane do not affect frequency, and en-
effect of anesthetics on motor neurons has been exam- flurane decreases frequency. Changes in charge trans-
ined in the intact isolated neonatal rat spinal cord. fer parallel these effects: increases in presynaptically
Volatile anesthetics (38,39) and ethanol (40) depress mediated frequency plus postsynaptically mediated
motor neuron output evoked by dorsal root prolongation increase the total charge transfer, but
stimulation. only during sodium channel blockade. That is, in nor-
Volatile anesthetics enhance the currents through mal intact preparations, anesthetics do not produce an
gamma-aminobutyric acid (GABA)A and glycine in- absolute increase in inhibition. Thus, for glycinergic
hibitory chloride channels that are produced by the transmission, anesthetics affect the timing rather than
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 721
2003;97:718 –40 MECHANISMS UNDERLYING MAC

the magnitude of inhibition, increasing the period Table 1. Neurotransmitter-Gated Ion Channels as Possible
during which inhibition is effective. Targets for Anesthetics
When glutamate AMPA currents are compared Effect of 1–2 MAC on
with glycinergic currents, a contrasting picture expressed receptora
emerges. Enflurane decreases miniature current fre-
Channel Volatile Gaseous Nonimmobilizer Mutationsb
quency regardless of sodium channel blockade (47)
and depresses current amplitude without changing Inhibitory
kinetics. Thus, volatile anesthetics tend to increase GABAA 1 0 0 Yes
glycinergic inhibition by two effects: a presynaptic Glycine 1 1 0 Yes
enhancement of transmitter release and a postsynaptic Excitatory
NMDA 2/0 2 0/2 No
prolongation of current duration. A depressant action AMPA 2/0 2 0 No
on release, probably via sodium channels, counterbal- Kainate 1 2 0 Yes
ances these inhibitory actions. Enflurane’s actions on Nicotinic 22 2 2 Yes
glutamate transmission are purely depressant, both 5-HT3 1 ND 0 Yes
pre- and postsynaptically. Immobility may result from MAC ⫽ minimum alveolar anesthetic concentration; GABA ⫽
a shift toward inhibition in the balance between exci- ␥-aminobutyric acid; NMDA ⫽ N-methyl-d-aspartate; AMPA ⫽ ␣-amino-3-
tation and inhibition, rather than from an absolute hydroxy-5-methyl-4-isoxazolepropionic acid; 5-HT ⫽ 5-hydroxytryptamine.
a
Indicates reliable effects of drug concentrations corresponding to one or
increase in inhibition. two times MAC of volatile inhaled anesthetics (primarily isoflurane, enflu-
In summary, results from studies in the isolated rane, sevoflurane, 1-chloro-1,2,2-trifluorocyclobutane [F3], or halothane), gas-
eous inhaled anesthetics (xenon, nitrous oxide, cyclopropane, or butane), or a
spinal cord suggest that actions involving both excita- nonimmobilizer (primarily 1,2-dichlorohexafluorocyclobutane [F6]). Up ar-
tory (AMPA and NMDA) and inhibitory (glycine and rows indicate enhancement; down arrows indicate block. 0 ⫽ no effect; ND ⫽
no data.
GABAA) transmission might produce immobility. Ac- b
Indicates whether mutation of amino acids in the receptor have been
tions on sodium channels and on other channels, in- studied and shown to alter the action of volatile anesthetics on the receptor (in
vitro studies) (e.g., Mihic J, personal communication, and Ref. 51).
cluding those that determine resting membrane poten-
tial, also may be important. Cholinergic receptors play
little or no role. transmission of specific ions (e.g., chloride ions), with
resulting changes in membrane potential. Ionotropic
receptors are composed of several subunits (e.g., a
pentameric receptor), and each subunit consists of
What Do In Vitro Studies of Isolated four transmembrane segments. In contrast, metabo-
Inhibitory and Excitatory Receptors tropic receptors are monomeric receptors consisting of
seven transmembrane segments. Binding of neuro-
Reveal? transmitter (e.g., acetylcholine) to metabotropic recep-
Which multiple neuronal signaling systems govern tors causes activation of guanosine triphosphate-
crucial components of anesthesia, such as immobility? binding proteins (G-proteins) associated with the
Drugs that produce anesthesia by acting on a single receptor, and these G-proteins act as second messen-
signaling system offer powerful clues to the site of gers to activate other signaling molecules, such as
anesthetic action. Propofol, alfaxalone, and etomidate protein kinases, or potassium or calcium channels.
primarily produce anesthesia by enhancing GABAA Note that the properties of a receptor may be deter-
receptor function (i.e., by increasing the effect of a mined by the particular subunits from which it is
given concentration of GABA) (48). Ketamine may composed (i.e., not all receptors of a given class [e.g.,
produce anesthesia primarily by inhibiting NMDA GABAA receptors] are the same).
receptor function, although it also acts on acetylcho- Ionotropic GABA receptors include GABAA and
line receptors (49). However, blockade of NMDA GABAC classes. (A metabotropic GABAB receptor cou-
alone does not appear to cause immobility (50), so pled through guanosine triphosphate-binding proteins
ketamine probably causes immobility by more than is not a ligand-gated ion channel.) GABAA receptors
simply blocking NMDA receptors. Emerging, detailed mediate much of the inhibitory neurotransmission in the
molecular analyses of anesthetic actions on GABAA central nervous system (CNS) (52). GABAA and GABAC
and glycine receptors and the potential importance of receptors are formed from combinations of ␣, ␤, ␥, and ␦
NMDA receptors prompt the following discussion fo- subunits, each of which may have multiple types (␣1– 6,
cused on these three ion channels (Table 1). ␤1–3, and ␥1–3). Although ␲ and ⑀ subunits also exist,
Neurotransmitters signal through two families of their function remains obscure. GABAC receptors likely
receptors: ionotropic and metabotropic. Ionotropic re- form from ␳1–3 subunits. Homomeric (all subunits the
ceptors are also known as ligand-gated ion channels same) ␳1 receptors are functional (53,54). Most CNS
because the neurotransmitter (e.g., GABA) binds di- GABAA receptors have ␣, ␤, and ␥ subunits. GABA
rectly to the ion channel protein, and this interaction binding to extracellular sites on the receptor opens a
causes opening (gating) of the ion channel, allowing chloride channel through the receptor. A response to
722 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

GABA requires both ␣ and ␤ subunits; benzodiazepine direct evidence for binding of anesthetics to these
enhancement of GABA function requires ␣, ␤, and ␥ amino acids in glycine and GABAA receptors (74).
subunits (55). Propanethiol and the related sulfhydryl-specific
Two types of subunits (␣ and ␤) form glycine recep- reagent, propyl methanethiosulfonate (PMTS),
tors. Only one ␤ subunit has been identified, and ␣1 is covalently bind to the mutated cysteines and perma-
the major ␣ subunit in adults. Homomeric ␣1 or ␣2 nently enhance receptor function (i.e., they produce
receptors are functional (56 –58). As discussed below, irreversible anesthetic-like effects.) Competition ex-
homomeric (e.g., ␳1 or glycine ␣1) receptors expressed periments performed before and after irreversible
in Xenopus oocytes have advanced studies of molecu- binding of PMTS to the cysteine mutations in the
lar sites of anesthetic action. GABAA receptor suggested competition for a single
Several observations suggest the potential impor- binding site among PMTS and the inhaled anesthetics
tance of these receptors to the immobility produced by or octanol, but not between PMTS and alfaxalone. This
inhaled anesthetics. Neuronal and several nonneuro- is consistent with the binding of inhaled anesthetics
nal expression systems reveal that relevant (1 MAC) and n-alcohols to the same site but the binding of
concentrations of volatile anesthetics and n-alcohols alfaxalone to a different site (75). Similar conclusions
potentiate GABAA and glycine receptor function apply to the glycine receptor.
(52,59 – 63). Most IV anesthetics also potentiate As noted above, the NMDA receptor also holds
GABAA receptor function, and some potentiate gly- promise as a site of anesthetic action. NMDA receptors
cine receptor function (61– 63). Nonimmobilizing con- constitute one glutamate receptor subtype, and they
geners of several fluorinated anesthetics do not affect require the coagonists glycine and glutamate for acti-
these receptors (52,59 – 61,64). vation (76). The subunits NR1, NR2 (A-D), and NR3
Comparison of the effects of anesthetics versus non- (A-B) form NMDA receptors (77–79). The NR1 subunit
immobilizers can be used as one test of the relevance of confers essential function to the NMDA receptor and
a particular ion channel. Although the anesthetic is expressed throughout the CNS. In adult mouse
1-chloro-1,2,2-trifluorocyclobutane (called 1A or F3) and
brain, NR1, NR2A, and NR2B subunits are widely
the structurally similar nonimmobilizer 2N, or F6, have
expressed, whereas NR2C occurs predominantly in
lipid solubilities consistent with an anesthetic capacity,
the cerebellum, and small levels of NR2D exist only in
only F3 affects the function of GABAA, glycine, AMPA,
the thalamus, brainstem, and spinal cord (80). Most
kainate, and 5-hydroxytryptamine (5-HT)3 receptors
NR3B subunits are on motor neurons (79), whereas
(59), and thus these receptors pass the test. The nonim-
NR3A has wide distribution (81,82).
mobilizer 2,3-dichlorooctafluorobutane (F8) also does
The Xenopus oocyte expression system allows study
not affect the function of several ligand-gated ion chan-
nels. In contrast, both F3 and F6 inhibit neuronal nico- of receptors with pharmacologic properties that may
tinic receptors and several metabotropic receptors (5- mimic those of native neuronal receptors (59). NMDA
HT2C, muscarinic-1, mGluR5) (65,66), and thus these receptors expressed from cloned receptors also pro-
receptors do not pass the test. vide evidence that inhaled anesthetics depress NMDA
How do inhaled anesthetics change the function of receptors. Clinically relevant concentrations of isoflu-
inhibitory neurotransmitter receptors? The differential rane, sevoflurane, and desflurane inhibit recombinant
actions of anesthetics on two homologous ligand- NR1/NR2A and NR1/NR2B NMDA receptors in a
gated ion channels allow an exploration of the molec- reversible, dose-dependent, and voltage-insensitive
ular sites of anesthetic action. In vitro, homomeric manner (83). Similarly, enflurane, urethane, nitrous
glycine ␣1 and GABA ␳1 receptors respond oppositely oxide, xenon, cyclopropane, and butane inhibit
to volatile anesthetics, which enhance function in the NMDA-stimulated currents in oocytes expressing
former (61) and inhibit function in the latter (67). A NMDA receptors (84 – 87). Thus, considerable evi-
genetic approach (one that combines various parts of dence shows that a wide range of volatile anesthetics
homomeric glycine ␣1 and GABA ␳1 receptors to pro- inhibit NMDA receptor function.
duce new receptors) allows identification of amino These results indicate that inhaled anesthetic actions
acids in transmembrane segment 2 and in transmem- on ligand-gated ion channels may explain the capacity
brane segment 3 of glycine and GABAA receptors that of inhaled anesthetics to produce immobility; they
permit ethanol enhancement of receptor function (68). point to candidate proteins that might mediate immo-
These amino acids lie near the extracellular surface of bility. Whether the associated receptors contribute to
the membrane. Results from independent studies sug- anesthetic-induced immobility can be determined
gest the importance of a site near the extracellular only from studies in intact organisms. A definitive test
surface (69 –72). of the importance of anesthetic influence on these
The possibility that propanethiol, an anesthetic in channels could, in principle, be achieved by designing
animals (73), could irreversibly react with cysteines mutant receptors that provide normal synaptic trans-
engineered at those critical positions provided more mission but resist any anesthetic effect and expressing
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 723
2003;97:718 –40 MECHANISMS UNDERLYING MAC

these receptors in animals. The molecular understand- these limitations. This complex approach requires
ing needed to produce these mutants is emerging for crossing two different lines of genetically engineered
the GABAA and glycine receptors but is lacking for mice. Making mutations in specific cells or neuronal
most other ligand-gated ion channels. As discussed pathways after key developmental events have oc-
elsewhere in this review, the introduction of drug- curred minimizes compensation and simplifies the as-
resistant receptors into animals (knockin mice) has signment of phenotype to neuroanatomic substrates.
provided remarkable insights regarding the actions of Although this approach has successfully dissected as-
benzodiazepines (88,89) and etomidate and propofol pects of learning and memory (92), it has not been
(90) and will likely prove important for understanding applied to studies of anesthetic mechanisms.
inhaled anesthetic action. The gene knockin animal differs from a knockout
animal (in which a gene of interest is inactivated) in
that a knockin animal expresses a mutation in the gene
of interest. For example, the gene might now produce
Genetic Engineering: Definitions and a protein changed by a single amino acid. Because the
Implications for Studies of Anesthetic normal endogenous promoter controls expression of
the mutant gene, a knockin animal expresses the mu-
Mechanisms tant gene in the same amount, at the same time, and in
Present technologies allow the creation of animals the same tissues as expressed by the normal gene. This
with genetic modifications that allow tests of specific approach allowed dissection of the contribution of
receptor targets of anesthetics: transgenic, knockout individual GABAA-receptor subunits to behavioral re-
(global or conditional), and knockin animals (Table 2). sponses induced by benzodiazepines (93). Most im-
Transgenic animals overexpress or misexpress an portantly, it allowed the construction of mice resistant
added gene (referred to as a transgene). They are cre- to anesthesia from propofol and etomidate (1). It also
ated by microinjection of DNA into the pronucleus of may allow the dissection of inhaled anesthetic mech-
a recently fertilized embryo. Several disadvantages anisms. Single amino acid substitutions in individual
limit the usefulness of this older, fast, widely available GABAA-receptor subunit genes can change the re-
technique. Because the added gene inserts randomly sponse of the receptors in expression systems to alco-
into the genome, expression can be variable, and, hol and volatile anesthetics without changing the re-
more importantly, the transgene must be dominant sponse to GABA (68), and mice harboring these
because the animals also have the endogenous gene. mutations may allow for unambiguous testing of the
Mating transgenic animals to animals lacking the gene contribution of specific receptor subunits to MAC.
of interest (not always available) can circumvent this Such genetically engineered mice are currently being
problem, but this is time consuming. tested (94). However, the inhaled anesthetics may
Creation of animals with precise genetic alterations present a challenge not offered by anesthetics such as
requires more sophisticated techniques such as gene propofol or etomidate. Probably only one receptor
targeting and embryonic stem cell technologies (91). (GABAA) mediates the effects of propofol and etomi-
With this approach, virtually any new endogenous date, but multiple receptors may mediate the effects of
gene can be created. Mutations rendering specific en- inhaled anesthetics. Thus, results from a knockin in-
dogenous genes nonfunctional are “gene knockouts” volving a single receptor may elucidate only one as-
(typically for the animal’s lifetime in all cells, a “global pect of inhaled anesthetic actions.
knockout”). Although expensive and time consuming,
this now-routine technology has produced more than
1000 global gene knockouts, many available from mu- What Do In Vivo Studies Reveal
tant mouse repositories (see http://www.jax.org/ Concerning the Ion Channels Mediating
imr/index.html). Although global knockouts have ad-
vanced an understanding of anesthetic mechanisms MAC?
(90), two problems can confound results from global Pharmacologic approaches can complement the previ-
knockout studies (41). First, gene inactivation may ously described genetic approaches to reveal the rele-
induce compensation by altering the expression of vance of candidate receptors to MAC. Both assume
other genes, and this result functionally compensates that altering the function of a relevant receptor will
for the loss of the targeted gene. Second, it is nearly change MAC. The genetic approach determines the
impossible to assign a phenotype such as MAC to a effect of receptor mutations (e.g., elimination or mod-
particular neural site because the knockout affects all ification of a specific receptor) on MAC. The pharma-
neurons. cologic approach determines the effect of agonists or
Conditional knockouts in which specific endoge- antagonists of specific ion channels on MAC.
nous genes are inactivated in specific cells or tissues Genetic manipulations and pharmacologic treat-
and/or at specific animal ages circumvent some of ments can change MAC directly or indirectly. For
724 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

Table 2. Genetically Engineered Animals


Animal type Purpose Advantages Disadvantages
Transgenic Overexpress or Relatively fast; least technically Random insertion, variable
misexpress gene demanding expression, transgene must
have dominant effect or be bred
onto knockout background
Global gene Inactivate gene in all cells Now routine; ⬎1000 strains Slow and expensive, technically
knockout of body at all times available demanding, developmental
compensation, nearly
impossible to attribute specific
phenotype to specific cell type,
circuit, or brain region
Conditional Inactivate gene in specific Can limit knockout to specific Slow and expensive, technically
knockout cells and tissues and/ tissue and/or time, minimizes demanding, not presently
or at specific times compensation, many strains possible to target every tissue
currently under development, or time point of interest
can mix and match mouse
strains to achieve desired
knockout
Gene knockin Normally express mutant Compensation absent or Slow and expensive, technically
gene from endogenous minimized demanding
control elements

example, consider the action of two drugs on heart IV. If it does not, then the antagonist effect is probably
rate. Epinephrine acts directly on the sympathetic supraspinal and not directly relevant to MAC.
nerve terminals to increase heart rate. Atropine also However, perhaps an antagonist affects MAC, and
increases heart rate but does so indirectly by blocking the systemic dose required for a given change exceeds
the action of acetylcholine on the parasympathetic the intrathecal dose. To be relevant, the receptor must
nervous system. Similar considerations apply to MAC. pass one more test. To be a relevant receptor (one
Although opioids decrease MAC, inhaled anesthetics directly acted on to produce the anesthetic effect), the
do not act via opioid receptors (see below). Opioid injection of an antagonist should change MAC in pro-
receptors can play an indirect (modulating) role but portion to the in vitro capacity of each anesthetic to
not a direct (mediating) role. This critical distinction affect the receptor. The power of this approach is
will become apparent in the discussion of the role of suggested in the “proof-of-principle” experiment
the GABAA receptor in MAC. A direct effect reflects a shown in Figure 4. In rats, the noncompetitive GABAA
mediating role for an ion channel (the terms “direct antagonist picrotoxin increases the immobilizing ED50
effect” and “mediating role” are interchangeable; see of the non-GABAergic anesthetic ketamine to a ceiling
Fig. 3). of approximately 60% (reflecting an indirect antago-
As described previously, the most powerful genetic nism of the effect of naturally occurring tonic GABA
approach to this issue develops mice with subtle site- release) but increases the ED50 of the GABAergic an-
directed mutations that retain the normal function of esthetic propofol by approximately 400% and shows
the receptor but abolish the capacity of anesthetics to no ceiling effect (primarily reflecting a direct antago-
affect normal function. For a mechanistically relevant nism) (49,95). In this study, isoflurane acts like ket-
receptor, such an alteration should produce an appar- amine, the non-GABAergic anesthetic. Gabazine (a
ently normal animal that, compared with mice with- competitive antagonist at GABAA receptors) has the
out the mutation, moves in response to noxious stim- same effect as picrotoxin.
ulation at inhaled anesthetic concentrations markedly The genetic and pharmacologic approaches are
exceeding those that normally produce immobility. complementary. When different mutations and
Pharmacologic approaches can also be used to dis- drug treatments show the same pattern (e.g., of no
tinguish between direct and indirect effects of receptor discernible effect), the consistency of findings
systems on MAC. If the administration of a receptor strengthens the conclusions regarding the relevance
antagonist produces no measurable effect on MAC, of the target receptor. One caveat applies to both
then that receptor neither directly nor indirectly af- genetic and pharmacologic tests of the relevance of
fects MAC and is not relevant. But suppose an antag- a specific ion channel: if two systems must be
onist has an effect? Because the spinal cord mediates blocked concurrently to produce immobility, then
MAC, an intrathecal (spinal) dose of antagonist should blocking or knocking out just one may not provide
change MAC more than the same or larger dose given an adequate test.
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 725
2003;97:718 –40 MECHANISMS UNDERLYING MAC

Figure 4. In rats, IV infusion of the noncompetitive ␥-aminobutyric


acid (GABA)A antagonist picrotoxin increases the immobilizing 50%
effective dose (ED50) of the non-GABAergic anesthetic ketamine to
a ceiling of approximately 60%. Because ketamine is not known to
materially influence GABAA receptors, this effect of picrotoxin
probably reflects an indirect antagonism of the effect of naturally
occurring tonic GABA release. Picrotoxin similarly increases the
immobilizing ED50 (minimum alveolar anesthetic concentration;
MAC) of isoflurane. In contrast, picrotoxin increases the ED50 of the
GABAergic anesthetic propofol by approximately 400% and shows
no ceiling effect (primarily reflecting a direct antagonism) (49,95).
The reader will note that 0.5 MAC isoflurane was added in both the
ketamine and the propofol studies. This was done because the
administration of picrotoxin in the absence of isoflurane was lethal
in animals anesthetized with ketamine. The administration of isoflu-
Figure 3. This drawing illustrates the difference between the site (or rane prevented death. Similar results were obtained when the
sites) that directly mediates the actions of inhaled anesthetics versus isoflurane concentration was 0.25 MAC.
sites that may influence inhaled anesthetic potency indirectly by
modulating the function of the true (directly affected) site (or sites)
of action. Modulation may occur through an action on the mediat-
ing site or by affecting the input to or output from that site. For in halothane MAC (100). Spasmodic mice have de-
example, a tonic release of ␥-aminobutyric acid (GABA) from an creased sensitivity to glycine because of a missense
indirect site might affect the measured minimum alveolar anesthetic mutation in the glycine receptor ␣ subunit (101), but
concentration (MAC) of all inhaled anesthetics by inhibiting activity
in the anesthetic site of action. Blocking this input would increase
they show no change in MAC to enflurane or halo-
MAC, but this increase would not mean that GABAA receptors thane (100). Interpretation of these results with mutant
mediate inhaled anesthetic actions. mice may be compromised by the development of
compensatory changes (see above).
Inhibitory Ionotropic Ligand-Gated and GABAA. Although they mediate the immobility pro-
Voltage-Gated Channels duced by injectable anesthetics (propofol and etomi-
date), GABAA receptors may not mediate the immobility
Glycine. Glycine receptors are major mediators of produced by inhaled anesthetics. The proof-of-principle
inhibitory neurotransmission in the spinal cord. Their experiment illustrated in Figure 4 indicates that isoflu-
spinal localization and their potentiation by clinically rane resembles the non-GABAergic anesthetic ketamine,
important volatile anesthetics make them prime can- but not the GABAergic anesthetic propofol. This sug-
didates as mediators of MAC. Evidence in animals gests that isoflurane’s immobilizing effect is not directly
supports this conjecture. IV and intrathecal adminis- mediated by GABAA receptors.
tration of strychnine, a glycine receptor antagonist, Similar conclusions follow from results of studies
increases MAC (96,97). The MAC increase for cyclo- applying intrathecal picrotoxin to rats inhaling isoflu-
propane, isoflurane, and halothane produced by intra- rane versus xenon or cyclopropane, anesthetics with
thecal strychnine administration correlates with the in minimal effects on GABAA receptors (Zhang Y, un-
vitro enhancing effect of these anesthetics on the gly- published data) (Fig. 5). If inhaled anesthetic enhance-
cine receptors (97) (Fig. 5) (98), a result expected if the ment of the response of GABAA receptors directly
enhancing action on the glycine receptor mediates mediates the capacity of inhaled anesthetics to sup-
part of the immobilizing effect of these anesthetics. press movement in response to noxious stimuli, then
Mice with mutations in the glycine receptor show a blockade of GABAA receptors should increase the
complex and inconsistent pattern of glycine effects. MAC of isoflurane more than the MAC of xenon or
Spastic mice have decreased expression of the glycine cyclopropane. However, MAC increases equally for
receptor because of abnormal RNA splicing of the ␤ all three anesthetics, indicating that GABAA receptors
subunit (99). Compared with controls, spastic mice are no more important to the immobilizing action of
have a 30% increase in enflurane MAC but no increase isoflurane than for anesthetics with minimal effects on
726 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

Figure 6. In rats, blockade of the 5-hydroxytryptamine-3 receptor by


systemic (114) (Fig. 6) or intrathecal (Flood, unpublished data)
administration of ondansetron does not affect the minimum alveo-
lar anesthetic concentration (MAC) of isoflurane, even at doses of
ondansetron modestly below the lethal dose. These findings indi-
cate that this receptor does not directly mediate the immobility
produced by inhaled anesthetics.
Figure 5. Administration of strychnine, a glycine receptor antago-
nist, increases minimum alveolar anesthetic concentration (MAC)
for rats (96). The MAC increase for cyclopropane, isoflurane, and The MAC for a series of fluorinated alkanols (103)
halothane produced by intrathecal strychnine administration corre- bears no consistent relationship to their capacity to
lates with the in vitro capacity of these anesthetics (at a concentra- enhance the effect of GABA on GABAA receptors
tion equal to MAC) to enhance (increase) the response of glycine
receptors to glycine (97). Each point is the measured increase in
(104). Enflurane and halothane enhance GABA-
MAC of a rat at the corresponding enhanced receptor activity mediated chloride conductance in rat hippocampal
calculated from pooled in vitro results (85,97). In contrast, although neurons (105), but a given MAC multiple of enflurane
administration of picrotoxin, a ␥-aminobutyric acid (GABA)A an- has twice the effect of halothane (i.e., the result is not
tagonist, also increases MAC for cyclopropane, xenon, and isoflu-
rane, the increase in MAC does not correlate with the in vitro quantitatively consistent across anesthetics.)
enhancing effect of these anesthetics (98) (data for xenon and halo- In summary, anesthetizing concentrations of many
thane were not available, respectively, for glycine and GABAA inhaled anesthetics enhance the in vitro action of
receptors).
GABA as much as do etomidate and propofol, anes-
thetics thought to produce immobility solely by en-
GABAA receptors. It may be argued that the in vitro hancing the action of GABA on GABAA receptors.
data for the effective concentration of anesthetics may Also, studies of ␤3 knockout and knockin mice indi-
be influenced by many variables (e.g., agonist concen- cate that GABAA receptors may have a mediating role,
tration, use of peak current, or current decay) and that albeit a minor one, in immobility. These findings sug-
the studies of Zhang selected inappropriate in vitro gest that enhancement of GABA actions should un-
concentrations. However, that argument ignores the derlie at least part of the immobility produced by
finding that the dose-response relationship in Figure 5 inhaled anesthetics. However, drugs that block
is flat (i.e., it almost does not matter where the in vitro GABAA receptor function in vivo do not markedly
data lie on the abscissa). increase MAC for inhaled anesthetics; the increase
Other evidence supports the conclusion that the parallels that for ketamine, an anesthetic not thought
enhancing effect of volatile anesthetics on GABAA to act by enhancement of GABAA receptors; the in-
receptors minimally influences MAC. Knockout of the crease in MAC is the same for inhaled anesthetics that
␤3 subunit of GABAA increases enflurane MAC by do and do not enhance the effect of GABA in vitro;
26% but increases halothane MAC by only 9% (102), and, similarly, the capacity of fluorinated alcohols to
and even these small increases may be attributable to enhance the effect of GABA does not predict their
compensation for lack of the subunit (41). Such results potency. Thus, in vivo results suggest a minor or no
for inhaled anesthetics quantitatively agree with those role for GABAA receptors as mediators of immobility,
found by Jurd et al. (1) for ␤3 N265M knockin mice whereas in vitro receptor studies and parallels with
unresponsive to the anesthetic effects of propofol: en- anesthetics such as propofol support a role. Although
flurane MAC increased by 15% and halothane MAC GABAA receptors are unlikely mediator candidates, a
by 21%. The mutation knocked into the mice had conclusive determination may require the develop-
normal in vitro sensitivity to enflurane and halothane. ment of knockin mice with GABAA receptors that
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 727
2003;97:718 –40 MECHANISMS UNDERLYING MAC

respond normally to GABA but are not potentiated by


inhaled anesthetics.
Potassium Channels. Because potassium channels
are numerous and diverse and because an increase in
potassium conductance can decrease the excitability of
the nervous system, potassium channels are plausible
candidates as mediators of the immobility produced
by inhaled anesthetics. Franks and Lieb (106) found
that inhaled anesthetics potentiate potassium leak
channels, and therefore reduce neuronal excitability,
in the pond snail genus Lymnaea. Mammalian potas-
sium channel subunits of the KCNK (K is potassium,
CN is channel, and K is the subfamily of potassium
channels) subfamily of leak channels (channels whose
currents are not strongly gated by voltage) have prop-
erties similar to the properties of the Lymnaea channel, Figure 7. Administration of nicotinic (mecamylamine) or muscarinic
including activation by inhaled anesthetics (107–110). (atropine or scopolamine) antagonists to rats does not decrease the
Although the KCNK subfamily provides a plausible isoflurane concentrations required to produce immobility (42,43).
These and other data indicate that cholinergic inhibition by volatile
anesthetic target, are there any holes in this leak compounds does not mediate anesthetic-induced immobility. MAC
theory? ⫽ minimum alveolar anesthetic concentration.
No present evidence demonstrates that potassium
channels directly mediate immobility. Only one (neg-
one a ligand-gated ion channel (nicotinic or nicotinic
ative) study of inhaled anesthetic potency in potas-
acetylcholine receptor [nAChR]) and the other a
sium channel knockout mice has been reported (111).
G-protein-linked (muscarinic) receptor. Nicotinic re-
One study of the nonspecific KCNK activator riluzole
ceptors expressed in neurons have been suggested as
supports the hypothesis that, although this drug has
putative mediators of general anesthetic actions be-
anesthetic properties, this action occurs in the brain
cause small concentrations of volatile anesthetics and
and not the spinal cord (Gray A, Zhang Y, unpub-
ketamine inhibit many such receptors and because the
lished data). That is, IV and intrathecal infusions of
(⫹) isomer of isoflurane is approximately 50% more
riluzole that do not produce permanent injury or
potent than the (⫺) isomer at inhibiting MAC in rats
death do not decrease isoflurane MAC, indicating that
(115), and the same ratio of potency has been observed
KCNK channels do not mediate MAC. In summary,
at molluscan nAChRs (116).
present evidence does not indicate that a specific po-
If antagonism of nicotinic receptors by volatile an-
tassium channel (or channels) mediates the immobil-
esthetics decreases motor responses to noxious stim-
ity produced by inhaled anesthetics, but the large
ulation, the administration of a nicotinic antagonist
number and diversity of potassium channels pre-
(mecamylamine) to rats should, but does not (Fig. 7),
cludes a conclusion that they play no role.
decrease the volatile anesthetic concentrations re-
quired to produce immobility (42,43). Similarly, the
Excitatory Ionotropic Ligand-Gated and administration of the classic agonist nicotine (at con-
Voltage-Gated Channels centrations known to have behavioral affects attrib-
uted to nicotinic modulation) (117) does not alter the
5-HT3 Receptors. Although most 5-HT receptors
MAC of isoflurane (42). Finally, nonimmobilizers (in-
couple to second-messenger systems, potentially lead-
haled compounds predicted from their hydrophobic-
ing to various downstream effects, the 5-HT3 receptor
ity to have anesthetic activity that do not cause immo-
is a cationic ionophore with considerable sequence
bility alone or add to the anesthesia provided by
homology to GABAA, glycine, and nicotinic acetylcho-
known anesthetics) (5) inhibit nAChRs (118). Such
line receptors. Thus, not surprisingly, some inhaled
observations suggest that nAChRs do not mediate
anesthetics potentiate the ionophoric response of the
inhaled anesthetic-induced immobility.
5-HT3 receptor (112,113). This in vitro proexcitatory
effect of some anesthetics suggests that the 5-HT3 re- Glutamate (NMDA, AMPA, and Kainate)
ceptor cannot mediate immobility. Consistent with
this conclusion, in rats, blockade of the 5-HT3 receptor NMDA Receptors. As noted above, glutamate is
by systemic (114) (Fig. 6) or intrathecal (Flood, unpub- the major mammalian excitatory neurotransmitter in
lished data) administration of ondansetron does not the CNS. Glutamate receptors include metabotropic
affect the MAC of isoflurane. G-protein-coupled receptors and three ligand-gated
Neuronal Nicotinic Acetylcholine Receptors. There ion channel classes: NMDA, AMPA, and kainate re-
are two types of acetylcholine receptors in the CNS: ceptors (119,120).
728 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

Numerous results suggest the potential importance


of NMDA receptors as mediators of the immobilizing
capacity of inhaled anesthetics. Diethyl ether, chloro-
form, methoxyflurane, halothane, enflurane, and
isoflurane decrease glutamate-stimulated binding of
MK-801 to the NMDA receptor (MK-801 [also called
dizocilpine] is a standard blocker of NMDA recep-
tors). Application of glutamate normally increases
binding of MK-801, which thereby serves as an indi-
cation of NMDA receptor activity. Thus, a decrease of
binding produced by the volatile anesthetics indicates
decreased NMDA receptor activity (121). Inhaled an-
esthetics inhibit NMDA receptors in receptor expres-
sion systems (87). Nitrous oxide and xenon have
greater in vitro effects than potent inhaled anesthetics
on NMDA receptors (122), and their actions on
NMDA receptors and/or sodium channels may con-
tribute more to their immobilizing capacity than is the
case with anesthetics such as isoflurane. As discussed Figure 8. Isoflurane concentrations (mean ⫾ sd) required to produce
above, both enflurane (45) and ethanol (44) can de- immobility in response to noxious stimulation (intermittent 50-V,
0.5-ms square-wave pulses) increase with decreasing interstimulus
crease NMDA-mediated glutamate currents in motor interval (ISI) (127). The concentration required to suppress move-
neurons in the spinal cord slice, an effect independent ment at an ISI of 0.1 s does not differ from the minimum alveolar
of effects on GABAA and glycine receptors. anesthetic concentration (MAC). MK-801 (an N-methyl-d-aspartate
antagonist that impairs temporal summation) administration added
Masaki et al. (123) applied the NMDA antagonist D to isoflurane administration prevents shorter ISIs from provoking
(⫺)-2-amino-5-phosphonopentanoic acid (D-AP5) movement at larger isoflurane concentrations.
intracerebroventricularly and noted a sevoflurane
MAC-sparing effect, suggesting a supraspinal effect.
noxious stimulation (MAC) depends on the inter-
Ishizaki et al. (124,125) found a maximal decrease of
stimulus interval, suggesting the persistence of tem-
30% in isoflurane MAC in rats given intrathecal bolus
poral summation and transmission via NMDA path-
doses of the NMDA antagonists AP5, MK-801, CPP,
ways (Fig. 8). Strengthening this interpretation, if
and 7CKA. McFarlane et al. (126) reported an 80%
temporal summation persists, then the administration
decrease in MAC of halothane from IV application of of the NMDA blocker MK-801 should abolish summa-
the competitive NMDA receptor antagonist CGS tion, and that is what appears to occur (Fig. 8). Also
19755. consistent with the interpretation that isoflurane does
MK-801 administration to rats decreases isoflurane not block temporal summation, electrophysiologic
MAC, and the decrease primarily correlates with MK- studies of neuronal windup show that temporal sum-
801 concentrations in the spinal cord, rather than mation can occur during anesthesia (129). However,
whole brain or cerebral cortex concentrations, a find- another interpretation is possible: perhaps isoflurane
ing consistent with mediation of the decreased MAC causes suppression of NMDA receptor transmission,
by the cord (50). Do inhaled anesthetics produce ef- and blockade with MK-801 simply substitutes for the
fects similar to MK-801: do they block NMDA recep- blockade produced by isoflurane.
tors so as to produce immobility (i.e., do NMDA re- In contrast to the findings with isoflurane, temporal
ceptors in the spinal cord directly mediate MAC)? summation is muted during anesthesia with xenon
Temporal summation is the cumulative effect of a (Dutton R, unpublished data). The greater effect of
succession of repeated stimuli presented at sufficiently xenon is consistent with its potent in vitro capacity to
close intervals (127). Phrased another way, a shorten- block NMDA receptors (87,122).
ing of the intervals between stimuli increases the col- In summary, evidence suggesting the importance of
lective stimulation and is more likely to provoke a NMDA receptors to MAC include the observations
response. In vitro, blocking NMDA receptors blocks that (1) in vitro, inhaled anesthetics can block NMDA
temporal summation, indicating that NMDA recep- receptors; (2) blockade of NMDA receptors (e.g., with
tors lie in the pathway that mediates temporal sum- MK-801) decreases MAC; and (3) this decrease in
mation (128). These observations suggest that persis- MAC correlates with the concentration of MK-801 in
tence of temporal summation during anesthetic the lower part of the spinal cord. However, studies
administration reflects intact NMDA receptor func- indicating that temporal summation persists during
tion. Studies with isoflurane suggest that approxi- isoflurane anesthesia and that MK-801 blocks such
mately 40% of the generation of movement evoked by summation argue against a role for NMDA receptors
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 729
2003;97:718 –40 MECHANISMS UNDERLYING MAC

as mediators of isoflurane MAC. Conversely, the con- Thus, studies using GluR2 knockout mice support
trasting data for xenon and isoflurane suggest the the postulate that different neuronal circuits mediate
possibility that NMDA receptors might mediate a por- MAC versus antinociception and loss of the righting
tion of the capacity of some anesthetics (xenon), but reflex (133). Motor neurons in the ventral spinal cord
not others (isoflurane), to produce immobility. None (i.e., that might mediate MAC) lack GluR2 subunits
of these results excludes a role for NMDA receptors in (136 –138), but neurons that mediate the righting reflex
the mediation of other aspects of anesthesia, including and nociception do express GluR2-containing AMPA
the capacity of inhaled anesthetics to impair learning receptors. Thus, knockout of the GluR2-containing re-
and memory (i.e., produce amnesia). ceptors should not change MAC but might change the
AMPA Receptors. AMPA receptors mediate the fast righting reflex and antinociception. Furthermore, neu-
(initial) component of excitatory postsynaptic trans- ronal circuits in the spinal cord that mediate MAC do
mission and provide plausible targets for volatile an- not appear to require the GluR2 subunit. In summary,
esthetics. Such fast transmission contrasts with the although in vivo data suggest that ablation of the
slower (later) excitatory transmission mediated by GluR2 subunit does not affect MAC, this cannot ex-
NMDA receptors. Competitive antagonists for AMPA clude the possibility that other subunits of AMPA
receptors, administered IV (130) or intrathecally (124), receptors contribute to MAC.
markedly decrease the concentration of halothane that Kainate Receptors. Combinations of GluR5–7, KA1,
causes immobility, findings consistent with the medi- and KA2 subunits form the kainate subtype of iono-
ation of anesthetic-induced immobility by AMPA re- tropic glutamate receptors (139). In vitro, inhaled an-
ceptors. However, although showing that movement esthetics enhance currents mediated by kainate recep-
requires AMPA receptor function, these results do not tors containing GluR6 (51). However, GluR6 knockout
prove or disprove that anesthetics impair AMPA re- mice have a normal isoflurane MAC (Sonner, unpub-
ceptor function. Indeed, one study found that intra- lished data). These findings are difficult to interpret
thecal antagonists fail to affect isoflurane MAC (131), because kainate receptors can be assembled from
but this study used smaller doses of antagonist than in other kainate subunits, even in the absence of the
the study that did show a decrease. GluR6 subunit. Perhaps other forms of GluR6 muta-
In vitro electrophysiologic studies demonstrate that tion, such as GluR6 editing mutant mice (140), will
volatile anesthetics inhibit AMPA-mediated excitatory provide insights that a knockout may not. Presently,
postsynaptic currents, that enflurane inhibits the the importance of kainate receptors to MAC remains
postsynaptic action of glutamate on AMPA receptors speculative.
in mouse spinal cord (45), and that halothane similarly Sodium Channels. A lingering dogma asserts that
affects the hippocampus at larger MAC values (25). clinical concentrations of inhaled anesthetics do not
Also, clinically relevant concentrations of volatile an- block axonal conduction and, thus, voltage-dependent
esthetics partially inhibit native and recombinant sodium channels. However, more recent data show
AMPA receptors activated by exogenous agonists that anesthetics can inhibit presynaptic terminal re-
(45,87,132–134). Whether minor inhibition of AMPA lease of neurotransmitters, particularly glutamate
receptors materially influences neuronal network ac- (141). Inhibition of these presynaptic sodium channels
tivity remains unclear. could produce this action. A tantalizing finding is that
Genetically modified mice provide insights into the systemic injection of lidocaine decreases MAC in ani-
importance of inhibition of AMPA receptors to volatile mals (142,143) and humans (144). The effect has a floor
anesthetic actions. GluR1– 4 subunit combinations form (a maximum decrease of 40%–50%) (143) and thus
AMPA receptors. Mice lacking the GluR2 subunit have cannot simply result from blockade of all nerve
inconsistent phenotypic responses to inhaled anesthet- conduction.
ics. They show a greater sensitivity to halothane, isoflu- Molecular cloning provides a basis for reconciling
rane, and sevoflurane than wild-type littermates for loss old ideas about axon conduction with anesthetic inhi-
of the righting reflex and antinociception but no differ- bition of sodium channel function. Cloning reveals
ence in MAC (133). In electrophysiologic studies, clinical nine different sodium channel ␣ subunits and three ␤
concentrations of isoflurane and halothane minimally subunits differentially localized in the brain and the
inhibit AMPA receptors (both GluR2-containing and periphery. Different members of this family may vary
-deficient) (133). Therefore, blockade of AMPA receptors in their anesthetic sensitivity. Several volatile anes-
cannot underlie enhanced sensitivity (for righting reflex thetics, including 1A, or F3, may inhibit synaptosomal
and antinociception) in GluR2 knockout mice. Other ev- sodium channels (141,145–156), but the nonimmobi-
idence suggests that decreased excitatory neurotrans- lizer 2N, or F6, does not inhibit (157). Thus, sodium
mission due to altered AMPA receptor kinetics renders channels pass this test of anesthetic relevance. Studies
GluR2-deficient neurons more sensitive to anesthetics of volatile anesthetic action on cloned sodium chan-
acting on other target receptors (135). nels show somewhat divergent results that, in part,
730 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

may be due to differences in the origin (neural versus


nonneural origin of the channel) (148,158,159).
Thus, consistent with the notion that sodium chan-
nels directly mediate anesthesia, several inhaled anes-
thetics, but not the nonimmobilizer F6, inhibit sodium
channels. Key questions remain: which sodium chan-
nel subunits exhibit sensitivity or resistance to anes-
thetics, and what is the neuroanatomical localization
of the sensitive channels? Finally, what is the mecha-
nism of channel inhibition— do anesthetics act directly
on the protein or act through protein kinase C (PKC)
or other modulators?

Metabotropic Receptors
Much of the preceding discussion focused on ion
channels because so much is known about inhaled
anesthetic effects on these receptors. In contrast to ion
channels, metabotropic receptors act through second
messengers released by an intracellular component
such as the G-protein coupled to the receptor (160).
Often, these effects are prolonged, lasting seconds to
minutes. They influence the electrophysiologic prop- Figure 9. In rats, the administration of enormous doses of naloxone,
a drug that blocks the effect of opioids, does not increase (or
erties of a neuron, including membrane resting poten- decrease) the minimum alveolar anesthetic concentration (MAC) of
tials (160). As noted below, changes in some of these halothane (164). If opioid receptors mediated the effect of inhaled
receptors (e.g., 5-HT2, opioid, and ␣2 adrenoreceptors) anesthetics, administration of naloxone would be expected to in-
can significantly decrease MAC. However, of these crease MAC. Parallel findings of lack of an effect of naloxone have
been found for nitrous oxide (165).
examples, only the 5-HT2 receptor might directly me-
diate the capacity of inhaled anesthetics to produce
immobility. something (analgesia) that inhaled anesthetics do not.
Neuronal Muscarinic Acetylcholine Receptors. Isoflu- Finally, the administration of enormous doses of nal-
rane inhibits muscarinic receptors through the activa- oxone, a drug that blocks the effect of opioids, has no
tion of PKC (117). If muscarinic antagonism by volatile effect on the MAC of halothane (Fig. 9) (164) or nitrous
anesthetics decreases motor responses to noxious oxide (165) in rodents.
stimulation, muscarinic antagonists (e.g., atropine or ␣2 Adrenoreceptors. Several findings suggest the pos-
scopolamine) given to rats should, but do not (Fig. 7), sibility that ␣2 adrenoreceptors might mediate the im-
decrease the volatile anesthetic concentrations re- mobility produced by inhaled anesthetics. The adminis-
quired to produce immobility (43), and intrathecal tration of ␣2-adrenoreceptor agonists decreases the MAC
administration of atropine does not alter the MAC of of inhaled anesthetics in rats (166,167) and dogs
isoflurane (43). Even the combination of muscarinic (168,169). The ␣2-adrenoreceptor agonist clonidine de-
and nicotinic blocking drugs does not change MAC creases anesthetic requirements in humans (170–172).
(Fig. 7). Thus, the hypothesis that cholinergic inhibi- ␣2-Adrenoreceptor agonists suppress nociceptive neuro-
tion by volatile compounds mediates anesthetic- transmission in the neonatal rat spinal cord, probably by
induced immobility cannot be correct. depressing substance P and glutamate-mediated path-
Opioid Receptors. Analgesia constitutes a nominal, ways (39,173). Spinal ␣2 adrenoreceptors may mediate a
albeit sometimes contested, feature of the anesthesia portion of the antinociceptive effect of medullary and
provided by inhaled anesthetics. Thus, it would be spinal ␦2-opioid receptors (174). Intrathecal or epidural
reasonable to suppose that inhaled anesthetics en- injection of dexmedetomidine in the dog has a far more
hance the release of endogenous opioids and that this powerful antinociceptive effect than that achieved with
release contributes to the anesthetic state. However, intracisternal or IV injection (175). In rats, intrathecal
inhaled anesthetics do not appear to increase endog- injection of the ␣2-adrenoreceptor agonist dexmedetomi-
enous opioids in cerebrospinal fluid (161), and they do dine produces a dose-dependent inhibition of nocicep-
not suppress autonomic or ventilatory responses to tive C and innocuous A ␤ responses of dorsal horn
surgical stimulation at concentrations that suppress neurons to transcutaneous electrical stimulation (176).
movement (162). Also, small doses of opioids mark- However, depletion of ␣ 2 adrenoreceptors by
edly decrease inhaled anesthetic concentrations that N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline
prevent movement (163). That is, the opioids supply does not change halothane MAC in rats (167). Also,
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 731
2003;97:718 –40 MECHANISMS UNDERLYING MAC

are consistent with the idea, but do not prove, that


5-HT2A receptors may directly mediate a small portion
of the capacity of inhaled anesthetics to produce im-
mobility. However, in vitro studies find that the 5-HT2
receptor is equally affected by the nonimmobilizer F6
and by halothane at concentrations predicted to equal
1 MAC (65).
Other Metabotropic Receptors. Little is known con-
cerning the capacity of other metabotropic receptors
(e.g., GABAB and neurokinin receptors) (191) to me-
diate immobility. GABAB receptor activation can hy-
perpolarize dorsal horn neurons, produce antinoci-
ception, and decrease motor neuron activity (192–195).
Although such effects suggest that enhanced GABAB
activity might decrease MAC, the only reported effect
is that a GABAB antagonist did not significantly alter
a measure related to MAC (196). However, the dose of
antagonist was carefully chosen to avoid touch-
evoked allodynia in the awake rat and thus may have
Figure 10. Administration of yohimbine 0.8 –1.0 mg/kg and atipa- been too small to provide an adequate test. The
mezole (␣-adrenoreceptor-blocking drugs) increases the minimum neurokinin-1 receptor and its chief ligand, substance
alveolar anesthetic concentration (MAC) of isoflurane in rats by P, mediate nociception. Genetic deletion of this recep-
approximately 10%. However, still greater doses do not increase tor impairs windup of dorsal horn neurons (197).
MAC (177). The small (10%) increase in MAC probably results from
suppression of the effect of normal tonic stimulation of ␣2 adreno- Windup may be defined as an increase in the number
receptors. The absence of a material increase in MAC argues against of action potentials elicited by each stimulus as a train
a role of ␣-adrenoreceptors as mediators of MAC. of stimuli progresses. Impairment of windup may re-
sult from suppression of processes involved in tempo-
the ␣-adrenoreceptor-blocking drug tolazoline at a ral summation, an effect that could also decrease MAC
single dose of 5 mg/kg does not affect halothane MAC (127,198,199). Metabotropic glutamate receptors are
in dogs (168). Yohimbine 0.8 –1.0 mg/kg and atipam- less likely to be involved in immobility because the
ezole (␣-adrenoreceptor-blocking drugs) increase the Class I receptors (types 1 and 5), the forms most
MAC of isoflurane in rats by approximately 10%, but frequently found in the spinal cord, are equally af-
still larger doses do not increase MAC (177) (Fig. 10). fected by the nonimmobilizer F6, whereas anesthetics
The small (10%) increase in MAC probably results only indirectly affect type 5 and do not affect type 1
from suppression of the effect of normal tonic stimu- (66,191).
lation of ␣2 adrenoreceptors. Thus, ␣-adrenoreceptors
do not mediate the capacity of inhaled anesthetics to
produce immobility in the face of noxious stimulation. The Place of Molecular Modeling in
5-HT2A Receptors. Inhaled anesthetics can block the
in vitro effect of 5-HT on 5-HT2A receptors (65), doing Future Studies
so at concentrations of approximately 1 MAC. 5-HT2A How do anesthetics interact with ion channels? Mo-
receptors may participate in nociceptive processes, as lecular modeling contributes to theories of anesthesia
shown in several studies using a specific blocker of at three atomic levels: modeling of site-directed mu-
5-HT2A receptors: ketanserin (178 –186). Blockade with tations, molecular dynamics simulations of anesthetics
ketanserin decreases evidence of nociception. Al- in membranes, and modeling of the structure of
though many of these studies focused on the supraspi- ligand-gated ion channels.
nal actions of 5-HT and ketanserin, some results also Single amino acid mutations in ligand-gated ion
suggested a direct spinal effect (182,185,187). channels can profoundly affect anesthetic potency
Although Dringenberg (188) found no anesthetic (68,70,74,200). Although the lack of radiograph struc-
effect after intraperitoneal administration of ketan- tures of ion channels presently thwarts visualization
serin 700 ␮g/kg in sheep, Doherty et al. (189) found of anesthetic-induced structural changes, three-
that the antagonist R51703 decreased halothane MAC dimensional models of single amino acid mutations in
in dogs. Similarly, Zhang et al. (190) found maximum putative anesthetic targets allow a mechanistic inter-
decreases in isoflurane MAC in rats of approximately pretation of the effects of these changes and suggest
60%. Larger doses were lethal. Intrathecal administra- new mutations (70,201–205). These models allow vi-
tion of ketanserin produced smaller (20%–25%), but sualization of complex interactions, such as those of
significant, maximum decreases in MAC. These data widely separated double mutations (68,70,205), and
732 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

within pure compounds, whereas molecular dynamics


simulations reveal anesthetic molecule locations
within the heterogeneous mixtures of protein, lipid,
and water that characterize nerve membranes (215).
Molecular dynamics studies demonstrate that anes-
thetic molecules distribute throughout the phospho-
lipid bilayer with a preferential localization near phos-
pholipid head groups (216,217).
Molecular modeling, combined with x-ray crystal-
Figure 11. A molecular model of a transmembrane (TM) segment of
a ␥-aminobutyric acid (GABA)A ␣1 receptor was built by threading
lography and nuclear magnetic resonance (NMR)
the primary sequence of a GABA receptor onto a template of a structure determination, can reveal atomic details of
four-helical bundle found in the three-dimensional structure of ion channel structure and function. The structure of a
cytochrome oxidase (208). The five subunits each are composed of bacterial mechanosensitive channel provides a tem-
four ␣ helices. Four subunit helices are rendered as red cylinders
connected by green interhelical loops. Helices show discontinuities plate for pentameric ion channels (209) that underlies
at proline residues. The fifth subunit shows the amino acid back- molecular modeling predictions of the mechanism of
bone trace as a yellow ribbon and shows the three amino acid channel opening and open state structure (218). By
residues most important for modulation of anesthetic potency: L232,
S270, and A291 (rendered with a space-filling surface in which the
using NMR structures of helices in solution (219,220),
carbon, oxygen, and hydrogen atoms are colored green, red, and several studies modeled the structure of the five heli-
white, respectively). (a) The transmembrane segment of the receptor ces surrounding the pore in the acetylcholine receptor
from the side in the plane of the membrane and (b) the ion pore of (220,221). The radiograph structure of the potassium
the receptor looking down the fivefold symmetry axis from the
extracellular side are shown. channel (222) permits detailed molecular models of
the ion selectivity filter (223,224).
Convergence of the three applications described
thereby suggest far more efficient tests than the pro- above may indicate whether single amino acid muta-
duction of random double mutations. tions change alcohol-/anesthetic-binding sites or
Secondary structure prediction algorithms indicate whether they alter protein stability in a way that al-
that bundles of four ␣ helices form the subunits of the lows anesthetic molecules to act remotely by nonspe-
transmembrane segment of homopentameric GABAA cific mechanisms. At present, considerable evidence
receptor ␣1 and glycine receptor ␣1 ion channels (206). indicates that site-directed mutations modify the vol-
The ␣ helical nature of transmembrane segment 2 in ume and polarity of specific anesthetic binding sites.
these ligand-gated ion channels is consistent with ob- Thus, at the level of atomic detail, the binding site in a
servations of electron density in cryoelectron micro- GABAA ␣1␤2␥2 receptor will have a different volume
graphs of acetylcholine receptors by Unwin et al. and polarity than a similar site in a GABAA ␣1␤3␥2
(207), with the crystal structures of a four-␣ helical receptor. Given sufficient information about the phar-
bundle found in a bacterial cytochrome (208) and of a macophores that determine anesthetic binding in a
bacterial mechanosensitive ion channel (209). Such particular cavity, it may be possible to design anes-
considerations led to models of these channels (and thetic molecules that bind to the former site, but not to
others in the “Cys-loop” superfamily) (201,204,208).
the latter. Such a “designer” molecule could offer in-
Cavities in the central axis of these bundles share
creased selectivity of regions of the nervous system
many properties, including volume and polarity, with
affected by anesthetics. For example, it might be pos-
binding sites in four-helical bundles engineered by
sible to target receptors present in the spinal cord, but
Johansson et al. (210) and Tanner et al. (211). The latter
not in the hippocampus.
binding sites bind halothane with Kd values close to
MAC. Combining five of the four ␣-helix subunits
allows formation of a homopentameric ion channel
(Fig. 11). The second transmembrane segments of each
of the five ␣ helices form a central pore with a right- The 5-Angstrom Hypothesis
hand supertwist and a funnel shape that is most nar- Results of studies of partially fluorinated series of
row at the intracellular face (212). n-alkanes and alcohols suggest that inhaled anesthet-
Molecular dynamics simulations of anesthetic mol- ics may produce immobility by concurrent actions on
ecules in proteins and phospholipid bilayers may re- two sites separated by 5 angstroms (225). Five ang-
veal how and where anesthetics act because such sim- stroms (fortuitously?) approximates the distance that
ulations can provide insights into anesthetic effects on separates adjacent loops of an ␣ helix (226), or the
the forward and backward rate constants that control distance between some sites on adjacent transmem-
the equilibrium between resting, open, and desensi- brane segments, or the effective pore diameter of gly-
tized states of ion channels (213). Partition coefficient cine and GABAA receptors (227). The ultimate signif-
measurements (214) provide anesthetic distributions icance of this finding remains unclear.
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 733
2003;97:718 –40 MECHANISMS UNDERLYING MAC

Nonspecific Mechanisms binding (232). Also, a redistribution of lateral pressures


within the membrane in response to a solution of volatile
The preceding observations could be interpreted to compounds within the membrane can predict the low
indicate that actions on ion channels cannot entirely potency of some compounds that disobey the Meyer-
explain the capacity of inhaled anesthetics to produce Overton rule (229). Reexamination of observations cited
immobility—although changes in transmission gov- as evidence against indirect mechanisms does not ex-
erned by such channels ultimately must underlie this clude Cantor’s theory. For example, his theory does not
capacity. Several lines of evidence support this minor- predict that increasing temperature should cause anes-
ity view. In vitro studies of specific receptors usually thesia (229), and small stereoselective effects of anesthet-
demonstrate wide variations in the potencies of dif- ics may be seen with either a protein- or membrane-
ferent anesthetics: in vitro potencies do not always based mechanism because both proteins and lipids have
correlate with MAC (e.g., Fig. 5), and, for a given chiral centers.
receptor, different anesthetics may produce opposing Using molecular dynamics simulation, Tang and Xu
effects (stimulation by some and depression by others) (233) found that halothane probably does not act on a
(113). Present data support a role for glycine and a few hydrophobic protein pocket. They suggested that
other receptors, but for the most part, such roles are global, rather than local, effects underlie the actions of
limited. No inhaled anesthetic action on a single re- halothane. Thus, a few investigators still believe that
ceptor can explain immobility, and immobility as a variants of Meyer’s (4) and Overton’s (3) hypothesis
result of concurrent actions on many receptors is un- may be correct.
likely (228), particularly because several proposed
candidates now seem less likely (e.g., acetylcholine
and GABAA).
The importance of effects on receptors versus some Conclusions
nonspecific effect may be gauged from Figure 1. Af- Our view and knowledge of the mechanisms by which
finity to a lipid phase is a measure of a nonspecific inhaled anesthetics act to produce immobility have
effect. If receptors play a role as mediators, then dif- expanded in the past two decades. Actions on one or
ferences in receptor effects (as determined in vitro) a few spinal cord receptors may underlie immobility.
might be reflected in deviations from the correlation Afferent (sensory) pathways, efferent (motor) path-
with lipophilicity. However, the data illustrated in ways, or interneurons might mediate immobility. Be-
Figure 1 do not demonstrate marked deviations. The cause they provide the final common pathway for
values for cyclopropane and halothane lie equally pain-evoked movement, depression of motor neuron
close to the line, but these anesthetics differ consider- receptors (by pre- and postsynaptic effects on excita-
ably in their in vitro enhancement of glycine receptors. tory and inhibitory neurotransmission) has received
Similarly, benzene and hexafluorobenzene differ in particular study.
their in vitro capacities to block NMDA receptors Although results from in vitro studies show that
(Raines D, unpublished data) but are approximately relevant concentrations of inhaled anesthetics can
equidistant from the line of correlation. modulate ion channels in ways that could cause im-
If immobility does not result from direct actions on mobility, only a few plausible candidates may be rel-
one or many channels, how might it be produced? Have evant (Table 3). These include some neurotransmitter-
we prematurely discarded nonspecific mechanisms, par- gated ion channels. Molecular, neurophysiologic, and
ticularly those involving the membrane bilayer? Cantor behavioral investigations support a role for glycine
(229 –231) suggested that there are enormous counterbal- receptors (e.g., Fig. 5), but glycinergic effects do not
ancing pressures (averaging approximately 300 atm) explain most of the immobilizing action of any inhaled
within the hydrophobic core of the lipid bilayer of the anesthetic. The putative binding site underlying the
plasma membrane. These pressures compensate for the enhancement of glycine receptors by inhaled anesthet-
attractive interfacial tensions at the aqueous interfaces of ics lies near the extracellular membrane interface.
the membrane. Variations in the distribution of such In contrast to evidence supporting glycine receptors
pressures, as would accompany the incorporation of gas- as direct mediators of MAC, other members of the
eous or vapor molecules, are predicted to strongly influ- ligand-gated ion channel superfamily (GABAA, neu-
ence the activity of transmembrane proteins (e.g., ion ronal nACh, and 5-HT3 receptors) probably do not
channels). The changes in activity might produce anes- mediate the immobility produced by inhaled anesthet-
thesia or other CNS effects, such as the convulsions seen ics, despite compelling evidence that GABAA recep-
with nonimmobilizers. Although Cantor’s predictions tors are responsible for the immobilizing effects of
have not been adequately tested experimentally, there is some injected anesthetics. This does not exclude im-
some accord with experimental results. The theories pre- portant roles for these receptors for other aspects of
dict sigmoidal ion channel dose-response curves, sug- anesthesia. Some metabotropic receptors that do not
gesting marked allosteric effects in the absence of protein appear to be important to the immobilizing action of
734 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

Table 3. Relevance of Specific Receptors as Direct Mediators of MAC


Receptor Rationale
Relevant
Glycine receptors Spinal cord distribution; anesthetics prolong glycinergic duration of miniature
postsymptic currents at clinically relevant concentrations; effect of antagonists
on MAC correlates with in vitro effect
Possibly relevant
5-HT2A receptors Intrathecal ketanserin decreases MAC by 20%–25%, but no in vitro data support a
direct effect
Sodium channels Systemic lidocaine decreases MAC by 40%–50%; anesthetics can inhibit sodium
channels at relevant concentrations, but nonimmobilizers do not; blockade of
sodium channels may decrease glutamate release from NMDA nerve terminals
NMDA receptors In vitro, several inhaled anesthetics can block NMDA receptors at clinically
relevant concentrations; NMDA receptor blockade decreases MAC, and the
decrease correlates with blocker concentration in the spinal cord; inhaled
anesthetic blockade of sodium channels of nerve terminals may decrease
glutamate release and thus decrease NMDA receptor activity; we speculate that
enhancement of glycine receptors on NMDA-containing nerve terminals may
increase glutamate release and thus decrease NMDA receptor activity; temporal
summation appears to persist during isoflurane but not during xenon
anesthesia; and NMDA blockade eliminates temporal summation for isoflurane
Possibly irrelevant
Potassium channels One knockout does not change MAC; intrathecal riluzole does not decrease MAC
at nontoxic infusions or decreases MAC by the same amount as IV infusions
AMPA and kainate receptors Although AMPA/kainate receptor blockade decreases MAC and anesthetics
inhibit AMPA-/kainate-mediated excitatory postsynaptic currents (but only a
little), GluR2 null mutant mice have normal MAC values
Probably irrelevant
GABAA receptors Increase in MAC from intrathecal administration of GABAA antagonists does not
correlate with in vitro effect; inconsistent relationship of MAC to in vitro effects;
IV administration of antagonists increases MAC as much as ketamine ED50
Opioid receptors Naloxone does not change MAC
␣2 adrenoreceptors Systemic and/or intrathecal blockers and depletors do not appreciably increase
MAC
5HT3 receptors Ondansetron does not affect MAC; some in vitro anesthetic effects are excitatory
Acetylcholine receptors Blockers of nicotinic and muscarinic receptors and combinations of the blockers
do not change MAC; the nonimmobilizer F6 inhibits nicotinic cholinergic
receptors
MAC ⫽ minimum alveolar anesthetic concentration; GABA ⫽ ␥-aminobutyric acid; NMDA ⫽ N-methyl-d-aspartate; AMPA ⫽ ␣-amino-3-hydroxy-5-methyl-
4-isoxazolepropionic acid; 5-HT ⫽ 5-hydroxytryptamine; ED50 ⫽ 50% effective dose; F6 ⫽ 1,2-dichlorohexafluorobutane.

inhaled anesthetics include muscarinic acetylcholine mechanism, although negative studies in one knock-
receptors, opioid receptors, and ␣2-adrenergic recep- out animal and with a nonspecific KCNK channel
tors. 5-HT2A receptors may contribute a small direct activator call into question the relevance of these chan-
effect, although no direct effects of anesthetics on nels. However, characterization of anesthetic effects
5-HT2A receptors have been demonstrated. on K channels has been performed for only a few
NMDA receptors may be candidate targets of vola- members of this large family.
tile anesthetic action. Kainate receptors probably are Combined pharmacologic and genetic approaches al-
not targets because in vitro effects require large anes- low a determination of the likely relevance of specific ion
thetic concentrations. The same may be said of AMPA channels as mediators of MAC. Several receptors and
receptors. The lack of effect of a knockout of an AMPA ion channels previously thought to be likely candidates
receptor GluR2 subunit on MAC adds to this conclu- (acetylcholine, GABAA, 5-HT3, ␣2-adrenergic, and
sion, but, of itself, such a lack may not be sufficiently opioid receptors and potassium channels) now seem
defining, because ventral horn (but not dorsal horn) less likely (but some may be crucial to other aspects of
neurons lack GluR2. anesthesia, such as amnesia); glycine and NMDA recep-
Blockade of voltage-gated sodium channels by in- tors and sodium channels seem more likely candidates.
haled anesthetics may produce presynaptic effects Developments in the next few years should lead to a
that underlie MAC. Enhancement of potassium chan- definitive understanding of how inhaled anesthetics pro-
nel function also has not been excluded as a possible duce immobility.
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 735
2003;97:718 –40 MECHANISMS UNDERLYING MAC

23. Franks NP, Lieb WR. Do general anaesthetics act by competi-


References tive binding to specific receptors? Nature 1984;310:599 – 601.
1. Jurd R, Arras M, Lambert S, et al. General anesthetic actions in 24. Perouansky M, Kirson ED, Yaari Y. Presynaptic and postsyn-
vivo strongly attenuated by a point mutation in the GABA(A) aptic actions of halothane at glutamatergic synapses in the
receptor beta3 subunit. FASEB J 2003;17:250 –2. mouse hippocampus. Br J Pharmacol 1996;124:1607–14.
2. Merkel G, Eger EI II. A comparative study of halothane and 25. Perouansky M, Kirson ED, Yaari Y. Mechanism of action of
halopropane anesthesia: including a method for determining volatile anesthetics: effects of halothane on glutamate receptors
equipotency. Anesthesiology 1963;24:346 –57. in vitro. Toxicol Lett 1998;100 –1:65–9.
3. Overton E. Studien über die Narkose, Zugleich ein Beitrag zur 26. MacIver MB. General anesthetic actions on transmission at
allgemeinen Pharmakologie. Jena, Germany: Gustav Fischer, glutamate and GABA synapses. In: Biebuyck JF, Lynch C III,
1901:1–195. Maze M, et al., eds. Anesthesia: biological foundations. New
4. Meyer HH. Theorie der Alkoholnarkose. Arch Exp Pathol York: Lippincott-Raven, 1997:277– 86.
Pharmakol 1899;42:109 –18. 27. Pashkov VN, Westphalen RI, Hemmings HC Jr. General anes-
5. Koblin DD, Chortkoff BS, Laster MJ, et al. Polyhalogenated and thetics do not affect release of the neuropeptide cholecystoki-
perfluorinated compounds that disobey the Meyer-Overton nin from isolated rat cortical nerve terminals. Anesthesiology
hypothesis. Anesth Analg 1994;79:1043– 8. 2002;97:1500 – 6.
6. Taheri S, Halsey MJ, Liu J, et al. What solvent best represents 28. Antognini JF, Schwartz K. Exaggerated anesthetic require-
the site of action of inhaled anesthetics in humans, rats and ments in the preferentially anesthetized brain. Anesthesiology
dogs? Anesth Analg 1991;72:627–34. 1993;79:1244 –9.
7. Gonsowski CT, Eger EI II. Nitrous oxide minimum alveolar 29. Antognini JF, Carstens E, Atherley R. Does the immobilizing
anesthetic concentration in rats is greater than previously re- effect of thiopental in brain exceed that of halothane? Anesthe-
ported. Anesth Analg 1994;79:710 –2. siology 2002;96:980 – 6.
8. Fang Z, Sonner J, Laster MJ, et al. Anesthetic and convulsant 30. Borges M, Antognini JF. Does the brain influence somatic
properties of aromatic compounds and cycloalkanes: implica- responses to noxious stimuli during isoflurane anesthesia? An-
tions for mechanisms of narcosis. Anesth Analg 1996;83: esthesiology 1994;81:1511–5.
1097–104. 31. Rampil IJ, Mason P, Singh H. Anesthetic potency (MAC) is
9. Koblin DD, Fang Z, Eger EI II, et al. Minimum alveolar con- independent of forebrain structures in the rat. Anesthesiology
centrations (MACs) of noble gases, nitrogen, and sulfur 1993;78:707–12.
hexafluoride in rats: helium and neon as nonimmobilizers 32. Rampil IJ. Anesthetic potency is not altered after hypothermic
(nonanesthetics). Anesth Analg 1998;87:419 –24. spinal cord transection in rats. Anesthesiology 1994;80:606 –10.
10. Trudell JR. A unitary theory of anesthesia based on lateral
33. Rampil IJ, Laster MJ. No correlation between quantitative elec-
phase separations in nerve membranes. Anesthesiology 1977;
troencephalographic measurements and movement response
46:5–10.
to noxious stimuli during isoflurane anesthesia in rats. Anes-
11. Rosenberg PH, Jansson SE, Gripenberg J. Effects of halothane,
thesiology 1992;77:920 –5.
thiopental, and lidocaine on fluidity of synaptic plasma mem-
34. Jinks SL, Martin JT, Carstens E, et al. Peri-MAC depression of
branes and artificial phospholipid membranes. Anesthesiology
a nociceptive withdrawal reflex is accompanied by reduced
1977;46:322– 6.
dorsal horn activity with halothane but not isoflurane. Anes-
12. Ueda I, Shieh DD, Eyring H. Anesthetic interaction with a
thesiology 2003;98:1128 –38.
model cell membrane: expansion, phase transition, and melt-
35. King BS, Rampil IJ. Anesthetic depression of spinal motor
ing of the lecithin monolayer. Anesthesiology 1974;41:217–25.
neurons may contribute to lack of movement in response to
13. Trudell JR, Hubbell WL, Cohen EN. The effect of two inhala-
tion anesthetics on the order of spin-labeled phospholipid noxious stimuli. Anesthesiology 1994;81:1484 –92.
vesicles. Biochim Biophys Acta 1973;291:321–7. 36. Rampil IJ, King BS. Volatile anesthetics depress spinal motor
14. Miller KW, Pang K-YY. General anaesthetics can selectively neurons. Anesthesiology 1996;85:129 –34.
perturb lipid bilayer membranes. Nature 1976;263:253–5. 37. Zhou HH, Jin TT, Qin B, Turndorf H. Suppression of spinal
15. Eger EI II, Saidman LJ, Brandstater B. Temperature depen- cord motoneuron excitability correlates with surgical immobil-
dence of halothane and cyclopropane anesthesia in dogs: cor- ity during isoflurane anesthesia. Anesthesiology 1998;88:
relation with some theories of anesthetic action. Anesthesiol- 955– 61.
ogy 1965;26:764 –70. 38. Kendig JJ, Kodde A, Gibbs LM, et al. Correlates of anesthetic
16. Fang ZX, Ionescu P, Chortkoff BS, et al. Anesthetic potencies of properties in isolated spinal cord: cyclobutanes. Eur J Pharma-
n-alkanols: results of additivity studies suggest a mechanism col 1994;264:427–36.
of action similar to that for conventional inhaled anesthetics. 39. Savola MK, Woodley SJ, Maze M, Kendig JJ. Isoflurane and an
Anesth Analg 1997;84:1042– 8. alpha2-adrenoceptor agonist suppress nociceptive neurotrans-
17. Cromwell TH, Eger EI II, Stevens WC, Dolan WM. Forane mission in neonatal rat spinal cord. Anesthesiology 1991;75:
uptake, excretion and blood solubility in man. Anesthesiology 489 –98.
1971;35:401– 8. 40. Wong SME, Fong E, Tauck DL, Kendig JJ. Ethanol as a general
18. White PF, Johnston RR, Eger EI II. Determination of anesthetic anesthetic: actions in spinal cord. Eur J Pharmacol 1997;329:
requirement in rats. Anesthesiology 1974;40:52–7. 121–7.
19. Vitez TS, White PF, Eger EI II. Effects of hypothermia on 41. Wong SME, Cheng G, Homanics G, Kendig JJ. Enflurane ac-
halothane MAC and isoflurane MAC in the rat. Anesthesiology tions on spinal cords from mice that lack the beta3 subunit of
1974;41:80 –1. the GABA(A) receptor. Anesthesiology 2001;95:154 – 64.
20. Strum DP, Eger EI II. Partition coefficients for sevoflurane in 42. Flood P, Sonner JM, Gong D, Coates KM. Heteromeric nicotinic
human blood, saline, and olive oil. Anesth Analg 1987;66: inhibition by isoflurane does not mediate MAC or loss of
654 – 6. righting reflex. Anesthesiology 2002;97:902–5.
21. Kashimoto S, Furuya A, Nonaka A, et al. The minimum alve- 43. Eger EI II, Zhang Y, Laster MJ, et al. Acetylcholine receptors do
olar concentration of sevoflurane in rats. Eur J Anaesthesiol not mediate the immobilization produced by inhaled anesthet-
1997;14:359 – 61. ics. Anesth Analg 2002;94:1500 – 4.
22. Crawford M, Lerman J, Saldivia V, Carmichael F. Hemody- 44. Wang MY, Rampil IJ, Kendig JJ. Ethanol directly depresses
namic and organ blood flow responses to halothane and AMPA and NMDA glutamate currents in spinal cord motor
sevoflurane anesthesia during spontaneous ventilation. Anesth neurons independent of actions on GABAA or glycine recep-
Analg 1992;75:1000 – 6. tors. J Pharmacol Exp Ther 1999;290:362–7.
736 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

45. Cheng G, Kendig JJ. Enflurane directly depresses glutamate 67. Mihic SJ, Harris RA. Inhibition of r1 receptor GABAergic cur-
AMPA and NMDA currents in mouse spinal cord motor neu- rents by alcohol and volatile anesthetics. J Pharmacol Exp Ther
rons independent of actions on GABAA or glycine receptors. 1996;277:411– 6.
Anesthesiology 2000;93:1075– 84. 68. Mihic SJ, Ye Q, Wick MJ, et al. Sites of alcohol and volatile
46. Cheng G, Kendig JJ. Pre- and postsynaptic volatile anaesthetic anaesthetic action on GABAA and glycine receptors. Nature
actions on glycinergic transmission to spinal cord motor neu- 1997;389:385–9.
rons. Br J Pharmacol 2002;136:673– 84. 69. Jenkins A, Andreasen A, Trudell JR, Harrison NL. Tryptophan
47. Cheng G, Kendig JJ. Enflurane decreases glutamate neuro- scanning mutagenesis in TM4 of the GABAa receptor alpha1
transmission to spinal cord motor neurons by both pre- and subunit: implications for modulation by inhaled anesthetics
postsynaptic actions. Anesth Analg 2003;96:1354 –9. and ion channel structure. Neuropharmacology 2002;43:
48. Krasowski M, Jenkins A, Flood P, et al. General anesthetic 669 –78.
potencies of a series of propofol analogs correlate with potency 70. Jenkins A, Greenblatt EP, Faulkner HJ, et al. Evidence for a
for potentiation of gamma-aminobutyric acid (GABA) current common binding cavity for three general anesthetics within the
at the GABA(A) receptor but not with lipid solubility. J Phar- GABAA receptor. J Neurosci 2001;21:RC136.
macol Exp Ther 2001;297:338 –51. 71. Pohorille A, Cieplak P, Wilson MA. Interactions of anesthetics
49. Flood P, Krasowski MD. Intravenous anesthetics differentially with the membrane-water interface. Chem Phys 1996;204:
modulate ligand-gated ion channels. Anesthesiology 2000;92:
337– 45.
1418 –25.
72. Pohorille A, Wilson MA, Cieplak P. Interaction of alcohols and
50. Stabernack C, Sonner JM, Laster M, et al. Spinal NMDA recep-
anesthetics with the water-hexane interface: a molecular dy-
tors may contribute to the immobilizing action of isoflurane.
namics study. Prog Colloid Polym Sci 1997;103:29 – 40.
Anesth Analg 2003;96:102–7.
51. Minami K, Wick MJ, Stern-Bach Y, et al. Sites of volatile anes- 73. Mascia MP, Gong D, Eger EI II, Harris RA. The anesthetic
thetic action on kainate (glutamate receptor 6) receptors. J Biol potency of propanol and butanol versus propanethiol and
Chem 1998;273:8248 –55. butanethiol in alpha1 wild type and alpha1(S267Q) glycine
52. Tanelian DL, Kosek P, Mody I, MacIver MB. The role of the receptors. Anesth Analg 2000;91:1289 –93.
GABAA receptor/chloride channel complex in anesthesia. An- 74. Mascia MP, Trudell JR, Harris RA. Specific binding sites for
esthesiology 1993;78:757–76. alcohols and anesthetics on ligand-gated ion channels. Proc
53. Whiting PJ, McKernan RM, Wafford KA. Structure and phar- Natl Acad Sci U S A 2000;97:9305–10.
macology of vertebrate GABAA receptor subtypes. In: Bradley 75. Sanna E, Murgia A, Casula A, Biggio G. Differential subunit
RJ, Harris RA, eds. International review of neurobiology. San dependence of the actions of the general anesthetics alphaxa-
Diego: Academic Press, 1995:95. lone and etomidate at gamma-aminobutyric acid type A recep-
54. McKernan RM, Whiting PJ. Which GABAa-receptor subtypes tors expressed in Xenopus laevis oocytes. Mol Pharmacol 1997;
really occur in the brain? Trends Neurosci 1996;19:139 – 43. 51:484 –90.
55. Sieghart W. Structure and pharmacology of gamma- 76. Johnson JW, Ascher P. Glycine potentiates the NMDA re-
aminobutyric acid A receptor subtypes. Pharmacol Rev 1995; sponse in cultured mouse brain neurons. Nature 1987;325:
47:181–234. 529 –31.
56. Rajendra S, Schofield PR. Molecular mechanisms of inherited 77. Meguro H, Mori H, Araki K, et al. Functional characterization
startle syndromes. Trends Neurosci 1995;18:80 –2. of a heteromeric NMDA receptor channel expressed from
57. Betz H. Glycine receptors: heterogenous and widespread in the cloned cDNAs. Nature 1992;357:70 – 4.
mammalian brain. Trends Neurosci 1991;14:458 – 61. 78. Monyer H, Sprengel R, Schoepfer R, et al. Heteromeric NMDA
58. Betz H. Ligand-gated ion channels in the brain: the amino acid receptors: molecular and functional distinction of subtypes.
receptor superfamily. Neuron 1990;5:383–92. Science 1992;256:1217–21.
59. Harris RA, Mihic SJ, Dildy-Mayfield JE, Machu TK. Actions of 79. Chatterton J, Awobuluyi M, Premkumar LS, et al. Excitatory
anesthetics on ligand-gated ion channels: role of receptor sub- glycine receptors containing the NR3 family of NMDA recep-
unit composition. FASEB J 1995;9:1454 – 62. tor subunits. Nature 2002;415:793– 8.
60. Downie DL, Hall AC, Lieb WR, Franks NP. Effects of inhala- 80. Mori H, Mishina M. Structure and function of the NMDA
tional general anaesthetics on native glycine receptors in rat receptor channel. Neuropharmacology 1995;34:1219 –37.
medullary neurones and recombinant glycine receptors in Xe- 81. Ciabarra AM, Sullivan JM, Gahn LG, et al. Cloning and char-
nopus oocytes. Br J Pharmacol 1996;118:493–502. acterization of chi-1: a developmentally regulated member of a
61. Mascia MP, Machu TK, Harris RA. Enhancement of homo- novel class of the ionotropic glutamate receptor family. J Neu-
meric glycine receptor function by long-chain alcohols and rosci 1995;15:6498 –508.
anaesthetics. Br J Pharmacol 1996;119:1331– 6. 82. Sucher NJ, Akbarian S, Chi CL, et al. Developmental and
62. Harrison NL, Kugler JL, Jones MV, et al. Positive modulation
regional expression pattern of a novel NMDA receptor-like
of human gamma-aminobutyric acid type A and glycine re-
subunit (NMDAR-L) in the rodent brain. J Neurosci 1995;15:
ceptors by the inhalation anesthetic isoflurane. Mol Pharmacol
6509 –20.
1993;44:628 –32.
83. Hollmann MW, Liu HT, Hoenemann CW, et al. Modulation of
63. Zimmerman SA, Jones MV, Harrison NL. Potentiation of
gamma-aminobutyric acid A receptor Cl⫺ current correlates NMDA receptor function by ketamine and magnesium. II.
with in vivo anesthetic potency. J Pharmacol Exp Ther 1994; Interactions with volatile anesthetics. Anesth Analg 2001;92:
270:987–91. 1182–91.
64. Mihic SJ, McQuilkin SJ, Eger EI II, et al. Potentiation of gamma- 84. Lin LH, Chen LL, Harris RA. Enflurane inhibits NMDA,
aminobutyric acid type A receptor-mediated chloride currents AMPA, and kainate-induced currents in Xenopus oocytes ex-
by novel halogenated compounds correlates with their abilities pressing mouse and human brain mRNA. FASEB J 1993;7:
to induce general anesthesia. Mol Pharmacol 1994;46:851–7. 479 – 85.
65. Minami K, Minami M, Harris RA. Inhibition of 85. Hara K, Eger EI II, Laster MJ, Harris RA. Nonhalogenated
5-hydroxytryptamine type 2A receptor-induced currents by alkanes cyclopropane and butane affect neurotransmitter-
n-alcohols and anesthetics. J Pharmacol Exp Ther 1997;281: gated ion channel and G-protein-coupled receptors: differen-
1136 – 43. tial actions on GABA(A) and glycine receptors. Anesthesiology
66. Minami K, Gereau RW, Minami M, et al. Effects of ethanol and 2002;97:1512–20.
anesthetics on type 1 and 5 metabotropic glutamate receptors 86. Hara K, Harris RA. The anesthetic mechanism of urethane: the
expressed in Xenopus laevis oocytes. Mol Pharmacol 1998;53: effects on neurotransmitter-gated ion channels. Anesth Analg
148 –56. 2002;94:313– 8.
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 737
2003;97:718 –40 MECHANISMS UNDERLYING MAC

87. Yamakura T, Harris RA. Effects of gaseous anesthetics nitrous 109. Sirois JE, Lynch C III, Bayliss DA. Convergent and reciprocal
oxide and xenon on ligand-gated ion channels: comparison modulation of a leak K(⫹) current and I(h) by an inhalational
with isoflurane and ethanol. Anesthesiology 2000;93:1095–101. anaesthetic and neurotransmitters in rat brainstem motoneu-
88. McKernan RM, Rosahl TW, Reynolds DS, et al. Sedative but rones. J Physiol 2002;541:717–29.
not anxiolytic properties of benzodiazepines are mediated by 110. Yost CS. Tandem pore domain K channels: an important site of
the GABA(A) receptor alpha1 subtype. Nat Neurosci 2000;3: volatile anesthetic action? Curr Drug Targets 2000;1:207–17.
587–92. 111. Gerstin KM, Gong DH, Abdallah M, et al. Mutation of KCNK5
89. Rudolph U, Crestani F, Benke D, et al. Benzodiazepine actions or Kir3.2 potassium channels in mice does not change mini-
mediated by specific gamma-aminobutyric acid(A) receptor mum alveolar anesthetic concentration. Anesth Analg 2003;96:
subtypes. Nature 1999;401:796 – 800. 1345–9.
90. Jurd R, Arras M, Lambert S, et al. General anesthetic actions in 112. Machu TK, Harris RA. Alcohols and anesthetics enhance the
vivo strongly attenuated by a point mutation in the GABAA function of 5-hydroxytryptamine3 receptors expressed in Xe-
receptor beta3 subunit. FASEB J 2003;17:250 –2. nopus laevis oocytes. J Pharmacol Exp Ther 1994;271:898 –905.
91. Homanics GE. Knockout and knockin mice. In: Liu Y, Lovinger 113. Suzuki T, Koyama H, Sugimoto M, et al. The diverse actions
DM, eds. Methods for alcohol-related neuroscience research. of volatile and gaseous anesthetics on human-cloned
Boca Raton, FL: CRC Press, 2002:31– 60. 5-hydroxytryptamine3 receptors expressed in Xenopus oo-
cytes. Anesthesiology 2002;96:699 –704.
92. Tsein JZ, Huerta PT, Tonegawa S. The essential role of hip-
114. Rampil IJ, Laster MJ, Eger EI II. Ondansetron does not alter
pocampal CA1 NMDA receptor-dependent synaptic plasticity
isoflurane MAC in rats. Anesthesiology 2001;95:562– 4.
in spatial memory. Cell 1996;87:1327–38.
115. Lysko GS, Robinson JL, Casto R, Ferrone RA. The stereospe-
93. Rudolph U, Crestani F, Mohler H. GABA(A) receptor
cific effects of isoflurane isomers in vivo. Eur J Pharmacol
subtypes: dissecting their pharmacological functions. Trends 1994;263:25–9.
Pharmacol Sci 2001;22:188 –94. 116. Franks NP, Lieb WR. Stereospecific effects of inhalational gen-
94. Harrison NL, Homanics GE. Production and characterization eral anesthetic optical isomers on nerve ion channels. Science
of knockin mice containing a serine to histidine mutation at 1991;254:427–30.
position 270 of the alpha 1 subunit of the GABAa receptor 117. Do SH, Kamatachi GL, Durieux ME. The effects of isoflurane
[abstract]. Alcohol Clin Exp Res 2001;25:A38. on native and chimeric muscarinic acetylcholine receptors: the
95. Sonner JM, Zhang Y, Stabernack C, et al. GABA(A) receptor role of protein kinase C. Anesth Analg 2001;93:375– 81.
blockade antagonizes the immobilizing action of propofol but 118. Raines DE, Claycomb RJ, Forman SA. Nonhalogenated anes-
not ketamine or isoflurane in a dose-related manner. Anesth thetic alkanes and perhalogenated nonimmobilizing alkanes
Analg 2003;96:706 –12. inhibit alpha4beta2 neuronal nicotinic acetylcholine receptors.
96. Zhang Y, Wu S, Eger EI II, Sonner JM. Neither GABA(A) nor Anesth Analg 2002;95:573–7.
strychnine-sensitive glycine receptors are the sole mediators of 119. Boulter J, Hollmann M, O’Shea-Greenfield A, et al. Molecular
MAC for isoflurane. Anesth Analg 2001;92:123–7. cloning and functional expression of glutamate receptor sub-
97. Zhang Y, Laster MJ, Hara K, et al. Glycine receptors may unit genes. Science 1990;31:1033–7.
mediate part of the immobility produced by inhaled anesthet- 120. Hollmann M, Heinemann S. Cloned glutamate receptors.
ics. Anesth Analg 2003;96:97–101. Annu Rev Neurosci 1994;17:31–108.
98. Zhang Y, Sonner JM, Eger EI II, et al. GABAA receptors do not 121. Martin DC, Plagenhoef M, Abraham J, et al. Volatile anesthet-
mediate the immobility produced by inhaled anesthetics. ics and glutamate activation of N-methyl-d-aspartate recep-
Anesth Analg. In press. tors. Biochem Pharmacol 1995;49:809 –17.
99. Kingsmore S, Giros B, Suh D, et al. Glycine receptor beta- 122. de Sousa LM, Dickinson R, Lieb WR, Franks NP. Contrasting
subunit gene mutation in spastic mouse associated with synaptic actions of the inhalational general anesthetics isoflu-
LINE-1 element insertion. Nat Genet 1994;7:136 – 41. rane and xenon. Anesthesiology 2000;92:1055– 66.
100. Quinlan JJ, Ferguson C, Jester K, et al. Mice with glycine 123. Masaki E, Yamazaki K, Ohno Y, et al. The anesthetic interac-
receptor subunit mutations are both sensitive and resistant to tion between adenosine triphosphate and N-methyl-d-
volatile anesthetics. Anesth Analg 2002;95:578 – 82. aspartate receptor antagonists in the rat. Anesth Analg 2001;
101. Ryan SG, Buckwalter MS, Lynch JW, et al. A missense muta- 92:134 –9.
tion in the gene encoding the alpha 1 subunit of the inhibitory 124. Ishizaki K, Yoon DM, Yoshida N, et al. Intrathecal administra-
glycine receptor in the spasmodic mouse. Nat Genet 1994;7: tion of N-methyl-d-aspartate receptor antagonist reduces the
131–5. minimum alveolar anaesthetic concentration of isoflurane in
102. Quinlan JJ, Homanics GE, Firestone LL. Anesthesia sensitivity rats. Br J Anaesth 1995;75:636 – 8.
in mice that lack the beta3 subunit of the gamma-aminobutyric 125. Ishizaki K, Yoshida N, Yoon DM, et al. Intrathecally adminis-
tered NMDA receptor antagonists reduce the MAC of isoflu-
acid type A receptor. Anesthesiology 1998;88:775– 80.
rane in rats. Can J Anaesth 1996;43:724 –30.
103. Eger EI II, Ionescu P, Laster MJ, et al. MAC of fluorinated
126. McFarlane C, Warner DS, Dexter F. Interactions between
alkanols in rats: relevance to theories of narcosis. Anesth Analg
NMDA and AMPA glutamate receptor antagonists during
1999;88:867–76.
halothane anesthesia in the rat. Neuropharmacology 1995;34:
104. Ueno S, Trudell JR, Eger EI II, Harris RA. Actions of fluori- 659 – 63.
nated alkanols on GABAA receptors: relevance to theories of 127. Dutton RC, Zhang Y, Stabernack CR, et al. Temporal summa-
narcosis. Anesth Analg 1999;88:877– 83. tion governs part of the minimum alveolar concentration of
105. Jones MV, Brooks PA, Harrison NL. Enhancement of gamma- isoflurane anesthesia. Anesthesiology 2003;98:1372–7.
aminobutyric acid-activated Cl⫺ currents in cultured rat hip- 128. Woolf CJ, Thompson SWN. The induction and maintenance
pocampal neurones by three volatile anaesthetics. J Physiol of central sensitization is dependent on N-methyl-d-aspartic
(Lond) 1992;449:279 –93. acid receptor activation: implications for the treatment
106. Franks NP, Lieb WR. Volatile general anaesthetics activate a of post-injury pain hypersensitivity states. Pain 1991;44:
novel neuronal K⫹ current. Nature 1988;333:662– 4. 293–9.
107. Gray AT, Winegar BD, Leonoudakis DJ, et al. TOK1 is a vola- 129. Dickenson AH, Sullivan AF. Electrophysiologic studies on the
tile anesthetic stimulated K⫹ channel. Anesthesiology 1998;88: effects of intrathecal morphine on nociceptive neurons in the
1076 – 84. rat dorsal horn. Pain 1986;24:211–22.
108. Patel AJ, Honore E, Lesage F, et al. Inhalational anesthetics 130. McFarlane C, Warner DS, Todd MM, Nordholm L. AMPA
activate two-pore-domain background K⫹ channels. Nat Neu- receptor competitive antagonism reduces halothane MAC in
rosci 1999;2:422– 6. rats. Anesthesiology 1992;77:1165–70.
738 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

131. Ishizaki K, Sasaki M, Karasawa S, et al. Intrathecal co- 151. Weigt H, Kwok WM, Rehmert GC, et al. Modulation of cardiac
administration of NMDA antagonist and NK-1 antagonist re- sodium current by alpha1-stimulation and volatile anesthetics.
duces MAC of isoflurane in rats. Acta Anaesthesiol Scand Anesthesiology 1997;87:1507–16.
1999;43:753–9. 152. Weigt HU, Rehmert GC, Bosnjak ZJ, Kwok WM. Conforma-
132. Carla V, Moroni F. General anaesthetics inhibit the responses tional state-dependent effects of halothane on cardiac Na⫹
induced by glutamate receptor agonists in the mouse cortex. current. Anesthesiology 1997;87:1494 –506.
Neurosci Lett 1992;146:21– 4. 153. Weigt HU, Kwok WM, Rehmert GC, et al. Voltage-dependent
133. Joo DT, Gong D, Sonner JM, et al. Blockade of AMPA receptors effects of volatile anesthetics on cardiac sodium current.
and volatile anesthetics: reduced anesthetic requirements in Anesth Analg 1997;84:285–93.
GluR2 null mutant mice for loss of the righting reflex and 154. Ries CR, Puil E. Mechanism of anesthesia revealed by shunting
antinociception but not minimum alveolar concentration. An- actions of isoflurane on thalamocortical neurons. J Neuro-
esthesiology 2001;94:478 – 88. physiol 1999;81:1795– 801.
134. Nishikawa K, MacIver MB. Excitatory synaptic transmission 155. Schwarz DW, Tennigkeit F, Adam T, et al. Membrane proper-
mediated by NMDA receptors is more sensitive to isoflurane ties that shape the auditory code in three nuclei of the central
than are non-NMDA receptor-mediated responses. Anesthesi- nervous system. J Otolaryngol 1998;27:311–7.
ology 2000;92:228 –36. 156. Rehberg B, Xiao YH, Duch DS. Central nervous system sodium
135. Orser BA, Joo DT, Zhu G, et al. Increased desensitization of channels are significantly suppressed at clinical concentrations
AMPA receptors may cause hypersensitivity to anesthetics in of volatile anesthetics. Anesthesiology 1996;84:1223–33.
GluT2 null mutant mice. In: Urban BW, Barann M, eds. Mo- 157. Ratnakumari L, Vysotskaya TN, Duch DS, Hemmings HC Jr.
lecular and basic mechanisms of anesthetics. Lengerich, Differential effects of anesthetic and nonanesthetic cyclobu-
Germany: Pabst Science Publishers. In press. tanes on neuronal voltage-gated sodium channels. Anesthesi-
136. Bar-Peled O, O’Brien RJ, Morrison JH, Rothstein JD. Cultured ology 2000;92:529 – 41.
motor neurons possess calcium-permeable AMPA/kainate re- 158. Mounsey JP, Patel MK, Mistry D, et al. Protein kinase C co-
ceptors. Neuroreport 1999;10:855–9. expression and the effects of halothane on rat skeletal muscle
137. Van Damme P, Van Den Bosch L, Van Houtte E, et al. GluR2- sodium channels. Br J Pharmacol 1999;128:989 –98.
dependent properties of AMPA receptors determine the selec- 159. Patel MK, Mistry D, John JE III, Mounsey JP. Sodium channel
tive vulnerability of motor neurons to excitotoxicity. J Neuro- isoform-specific effects of halothane: protein kinase C co-
physiol 2002;88:1279 – 87. expression and slow inactivation gating. Br J Pharmacol 2000;
138. Petralia RS, Wang YX, Mayat E, Wenthold RJ. Glutamate re- 130:1785–92.
ceptor subunit 2-selective antibody shows a differential distri- 160. Siegelbaum SA, Schwartz JH, Kandel ER. Modulation of syn-
bution of calcium-impermeable AMPA receptors among pop- aptic transmission: second messengers. In: Kandel ER,
ulations of neurons. J Comp Neurol 1997;385:456 –76. Schwartz JH, Jessell TM, eds. Principles of neural science. 4th
139. Dingledine R, Borges K, Bowie D, Traynelis SF. The glutamate ed. New York: McGraw-Hill, 2000.
receptor ion channels. Pharmacol Rev 1999;51:7– 61. 161. Way WL, Hosobuchi Y, Johnson BH, Eger EI II. Anesthesia
140. Vissel B, Royle GA, Christie BR, et al. The role of RNA editing does not increase opioid peptides in cerebrospinal fluid of
of kainate receptors in synaptic plasticity and seizures. Neuron humans. Anesthesiology 1984;60:43–5.
2001;29:217–27. 162. France CJ, Plumer HM, Eger EI II, Wahrenbrock EA. Ventila-
141. Lingamaneni R, Birch ML, Hemmings HC Jr. Widespread in- tory effects of isoflurane (Forane) or halothane when combined
hibition of sodium channel-dependent glutamate release from with morphine, nitrous oxide and surgery. Br J Anaesth 1974;
isolated nerve terminals by isoflurane and propofol. Anesthe- 46:117–20.
siology 2001;95:1460 – 6. 163. Katoh T, Ikeda K. The effects of fentanyl on sevoflurane re-
142. Himes RS Jr, DiFazio CA, Burney RG. Effects of lidocaine on quirements for loss of consciousness and skin incision. Anes-
the anesthetic requirements for nitrous oxide and halothane. thesiology 1998;88:18 –24.
Anesthesiology 1977;47:437– 40. 164. Harper MH, Winter PM, Johnson BH, Eger EI II. Naloxone
143. DiFazio CA, Neiderlehner JR, Burney RG. The anesthetic po- does not antagonize general anesthesia in the rat. Anesthesi-
tency of lidocaine in the rat. Anesth Analg 1976;55:818 –21. ology 1978;49:3–5.
144. Himes RS Jr, Munson ES, Embro WJ. Enflurane requirement 165. Smith RA, Wilson M, Miller KW. Naloxone has no effect on
and ventilatory response to carbon dioxide during lidocaine nitrous oxide anesthesia. Anesthesiology 1978;49:6 – 8.
infusion in dogs. Anesthesiology 1979;51:131– 4. 166. Segal IS, Vickery RG, Walton JK, et al. Dexmedetomidine di-
145. Weigt HU, Kwok WM, Rehmert GC, Bosnjak ZJ. Modulation of minishes halothane anesthetic requirements in rats through a
the cardiac sodium current by inhalational anesthetics in the postsynaptic alpha-2 adrenergic receptor. Anesthesiology
absence and presence of beta-stimulation. Anesthesiology 1988;69:818 –23.
1998;88:114 –24. 167. Rabin BC, Reid K, Guo TZ, et al. Sympatholytic and minimum
146. Irnaten M, Wang J, Chang KS, et al. Ketamine inhibits sodium anesthetic concentration-sparing responses are preserved in
currents in identified cardiac parasympathetic neurons in nu- rats rendered tolerant to the hypnotic and analgesic action of
cleus ambiguus. Anesthesiology 2002;96:659 – 66. dexmedetomidine, a selective alpha-2 adrenergic agonist. An-
147. Asai T, Kusudo K, Ikeda H, et al. Effect of halothane on esthesiology 1996;85:565–73.
neuronal excitation in the superficial dorsal horn of rat spinal 168. Bloor BC, Flacke WE. Reduction in halothane anesthetic re-
cord slices: evidence for a presynaptic action. Eur J Neurosci quirement by clonidine, an alpha-adrenergic agonist. Anesth
2002;15:1278 –90. Analg 1982;61:741–5.
148. Stadnicka A, Kwok WM, Hartmann HA, Bosnjak ZJ. Effects of 169. Vickery RG, Maze M. Action of the stereoisomers of medeto-
halothane and isoflurane on fast and slow inactivation of hu- midine in halothane-anesthetized dogs. Acta Vet Scand Suppl
man heart hH1a sodium channels. Anesthesiology 1999;90: 1989;85:71– 6.
1671– 83. 170. Ghinone M, Calvillo O, Quintin L. Anesthesia and
149. Scholz A, Appel N, Vogel W. Two types of TTX-resistant and hypertension: the effect of clonidine on perioperative hemody-
one TTX-sensitive Na⫹ channel in rat dorsal root ganglion namics and isoflurane requirements. Anesthesiology 1987;67:
neurons and their blockade by halothane. Eur J Neurosci 1998; 3–10.
10:2547–56. 171. Lawrence CJ, De Lange S. Effects of a single pre-operative
150. Ratnakumari L, Hemmings HC Jr. Inhibition of presynaptic dexmedetomidine dose on isoflurane requirements and peri-
sodium channels by halothane. Anesthesiology 1998;88: operative haemodynamic stability. Anaesthesia 1997;52:
1043–54. 736 – 44.
ANESTH ANALG REVIEW ARTICLE SONNER ET AL. 739
2003;97:718 –40 MECHANISMS UNDERLYING MAC

172. Aantaa R, Jaakola ML, Kallio A, Kanto J. Reduction of the 195. Albright AL, Cervi A, Singletary J. Intrathecal baclofen for
minimum alveolar concentration of isoflurane by dexmedeto- spasticity in cerebral palsy. JAMA 1991;265:1418 –22.
midine. Anesthesiology 1997;86:1055– 60. 196. Mason P, Owens CA, Hammond DL. Antagonism of the an-
173. Kendig JJ, Savola MKT, Woodley SJ, Maze M. Alpha-2 adre- tinocifensive action of halothane by intrathecal administration
noceptors inhibit a nociceptive response in neonatal rat spinal of GABAA receptor antagonists. Anesthesiology 1996;84:
cord. Eur J Pharmacol 1991;192:293–300. 1205–14.
174. Grabow TS, Hurley RW, Banfor PN, Hammond DL. Supraspi- 197. Weng HR, Mansikka H, Winchurch R, et al. Sensory processing
nal and spinal delta(2) opioid receptor-mediated antinocicep- in the deep spinal dorsal horn of neurokinin-1 receptor knock-
tive synergy is mediated by spinal alpha(2) adrenoceptors. out mice. Anesthesiology 2001;94:1105–12.
Pain 1999;83:47–55. 198. Sivilotti LG, Thompson SW, Woolf SJ. Rate of rise of the
175. Sabbe MB, Penning JP, Ozaki GT, Yaksh TL. Spinal and sys- cumulative depolarization evoked by repetitive stimulation of
temic action of the alpha 2 receptor agonist dexmedetomidine small-caliber afferents is a predictor of action potential windup
in dogs: antinociception and carbon dioxide response. Anes- in rat spinal neurons in vitro. J Neurophysiol 1993;69:1621–31.
thesiology 1994;80:1057–72. 199. Ishizaki K, Karasawa S, Takahashi K, et al. Intrathecal
176. Sullivan AF, Kalso EA, McQuay HJ, Dickenson AH. The an- neurokinin-1 receptor antagonist reduces isoflurane MAC in
tinociceptive actions of dexmedetomidine on dorsal horn neu- rats. Can J Anaesth 1997;44:543–9.
ronal responses in the anaesthetized rat. Eur J Pharmacol 1992; 200. Beckstead MJ, Phelan R, Trudell JR, et al. Anesthetic and
215:127–33. ethanol effects on spontaneously opening glycine receptor
177. Eger EI II, Xing Y, Laster MJ, Sonner JM. Alpha-2 adrenore- channels. J Neurochem 2002;82:1343–51.
ceptors probably do not mediate the immobility produced by 201. Yamakura T, Bertaccini E, Trudell JR, Harris RA. Anesthetics
inhaled anesthetics. Anesth Analg 2003;96:1661– 4. and ion channels: molecular models and sites of anesthetic
178. Vonvoigtlander P, Lewis RA, Neff GL. Kappa opioid analgesia action. Annu Rev Pharmacol Toxicol 2001;41:23–51.
is dependent on serotonergic mechanisms. J Pharmacol Exp 202. Scheller M, Forman SA. The gamma subunit determines
Ther 1984;231:270 – 4. whether anesthetic-induced leftward shift is altered by a mu-
179. Kellstein DE, Malseed RT, Goldstein FJ. Opioid-monamine tation at alpha1S270 in alpha1beta2gamma2L GABA(A) recep-
interactions in spinal antinociception: evidence for serotonin tors. Anesthesiology 2001;95:123–31.
but not norepinephrine reciprocity. Pain 1988;34:85–92. 203. Raines DE, Claycomb RJ, Scheller M, Forman A. Nonhaloge-
180. Paul D, Mana MJ, Pfaus JG, Pinel JP. Attenuation of morphine nated alkane anesthetics fail to potentiate agonist actions on
analgesia by the S2 antagonists, pirenperone and ketanserin. two ligand-gated channels. Anesthesiology 2001;95:470 –7.
Pharmacol Biochem Behav 1988;31:641–7. 204. Trudell JR. Unique assignment of inter-subunit association in
181. Kiefel JM, Paul D, Bodnar RJ. Reduction in opioid and non- GABAa alpha1 beta3 gamma2 receptors determined by molec-
opioid forms of swim analgesia by 5-HT2 receptor antagonists. ular modeling. Biochim Biophys Acta 2002;1565:91– 6.
Brain Res 1989;500:231– 40.
205. Kash TL, Jenkins A, Trudell JR, Harrison NL. Electrostatic
182. Giordano J. Analgesic profile of centrally administered
interactions couple agonist binding to channel gating in the
2-methylserotonin against acute pain in rats. Eur J Pharmacol
GABAA receptor. Nature 2003;421:421–5.
1991;199:233– 6.
206. Bertaccini E, Trudell JR. Predicting the transmembrane second-
183. Coimbra NC, Brandao ML. Effects of 5-HT2 receptors blockade
ary structure of ligand-gated ion channels. Protein Eng 2002;
on fear-induced analgesia elicited by electrical stimulation of
15:443–53.
the deep layers of the superior colliculus and dorsal periaque-
207. Unwin N, Miyazawa A, Li J, Fujiyoshi Y. Activation of the
ductal gray. Behav Brain Res 1997;87:97–103.
nicotinic acetylcholine receptor involves a switch in conforma-
184. Yamazake N, Umeno H, Kuraishi Y. Involvement of brain
tion of the alpha subunits. J Mol Biol 2002;319:1165–76.
serotonergic terminals in the antinociceptive action of periph-
erally applied calcitonin. Jpn J Pharmacol 1999;81:367–74. 208. Trudell JR, Bertaccini E. Molecular modelling of specific and
185. Bardin L, Lavarenne J, Eschalier A. Serotonin receptor sub- non-specific anaesthetic interactions. Br J Anaesth 2002;89:
types involved in the spinal antinociceptive effect of 5-HT in 32– 40.
rats. Pain 2000;86:11– 8. 209. Chang G, Spencer RH, Lee AT, et al. Structure of the MscL
186. Hentall ID, Andresen MJ, Taguchi K. Serotonergic, cholinergic homolog from mycobacterium tuberculosis: a gated mechano-
and nociceptive inhibition or excitation of raphe magnus neu- sensitive ion channel. Science 1998;282:2220 – 6.
rons in barbiturate-anesthetized rats. Neuroscience 1993;52: 210. Johansson JS, Scharf D, Davies LA, et al. A designed four-
303–10. alpha-helix bundle that binds the volatile general anesthetic
187. Zhang L. Effects of 5-hydroxytryptamine on cat spinal mo- halothane with high affinity. Biophys J 2000;78:982–93.
toneurons. Can J Physiol Pharmacol 1991;69:154 – 63. 211. Tanner JW, Johansson JS, Liebman PA, Eckenhoff RG. Predict-
188. Dringenberg HC. Serotonergic receptor antagonists alter re- ability of weak binding from X-ray crystallography: inhaled
sponses to general anaesthetics in rats. Br J Anaesth 2000;85: anesthetics and myoglobin. Biochemistry 2001;40:5075– 80.
904 – 6. 212. Wilson GG, Pascual JM, Brooijmans N, et al. The intrinsic
189. Doherty TJ, McDonell WN, Dyson DH, et al. The effect of a electrostatic potential and the intermediate ring of charge in
5-hydroxytryptamine antagonist (R51703) on halothane MAC the acetylcholine receptor channel. J Gen Physiol 2000;115:
in the dog. J Vet Pharmacol Ther 1995;18:153–5. 93–106.
190. Zhang Y, Laster MJ, Eger EI II, et al. Blockade of 5-HT2A 213. Scheller M, Forman SA. Coupled and uncoupled gating and
receptors may mediate or modulate part of the immobility desensitization effects by pore domain mutations in GABA(A)
produced by inhaled anesthetics. Anesth Analg. In press. receptors. J Neurosci 2002;22:8411–21.
191. Coggeshall RE, Carlton SM. Receptor localization in the mam- 214. Ionescu P, Eger EI II, Trudell J. Direct determination of oil/
malian dorsal horn and primary afferent neurons. Brain Res saline partition coefficients. Anesth Analg 1994;79:1056 – 8.
Brain Res Rev 1997;24:28 – 66. 215. Karplus M, McCammon JA. Molecular dynamics simulations
192. Knangrga I, Jiang MC, Randic M. Actions of (⫺)-baclofen on of biomolecules. Nat Struct Biol 2002;9:646 –52.
rat dorsal horn neurons. Brain Res 1991;562:265–75. 216. Koubi L, Tarek M, Bandyopadhyay S, et al. Effects of
193. Dirig DM, Yaksh TL. Intrathecal baclofen and muscimol, but the nonimmobilizer hexafluroethane on the model mem-
not midazolam, are antinociceptive using the rat-formalin brane dimyristoylphosphatidylcholine. Anesthesiology 2002;
model. J Pharmacol Exp Ther 1995;275:219 –27. 97:848 –55.
194. Dickenson AH, Brewer CM, Hayes NA. Effects of topical ba- 217. Johansson JS, Scharf D. Towards an understanding of how
clofen on C fiber-evoked neuronal activity in the rat dorsal general anesthetics alter central nervous system protein func-
horn. Neurosci Biobehav Rev 1985;14:557– 62. tion. Reg Anesth Pain Med 2001;26:267–70.
740 REVIEW ARTICLE SONNER ET AL. ANESTH ANALG
MECHANISMS UNDERLYING MAC 2003;97:718 –40

218. Sukharev S, Durell SR, Guy HR. Structural models of the MscL 227. Bormann J, Hamill OP, Sakmann B. Mechanism of anion per-
gating mechanism. Biophys J 2001;81:917–36. meation through channels gated by glycine and gamma-
219. Tang P, Mandal PK, Xu Y. NMR structures of the second aminobutyric acid in mouse cultured spinal neurons. J Physiol
transmembrane domain of the human glycine receptor al- 1987;385:243– 86.
pha(1) subunit: model of pore architecture and channel gating. 228. Eger EI II, Fisher DM, Dilger JP, et al. Relevant concentrations
Biophys J 2002;83:252– 62. of inhaled anesthetics for in vitro studies of anesthetic mech-
220. Law RJ, Forrest LR, Ranatunga KM, et al. Structure and dy- anisms. Anesthesiology 2001;94:915–21.
namics of the pore-lining helix of the nicotinic receptor: MD 229. Cantor RS. Breaking the Meyer-Overton rule: predicted effects
simulations in water, lipid bilayers, and transbilayer bundles. of varying stiffness and interfacial activity on the intrinsic
Proteins 2000;39:47–55.
potency of anesthetics. Biophys J 2001;80:2284 –97.
221. Opella SJ, Marassi FM, Gesell JJ, et al. Structures of the M2
230. Cantor RS. The lateral pressure profile in membranes: a phys-
channel-lining segments from nicotinic acetylcholine and
NMDA receptors by NMR spectroscopy. Nat Struct Biol 1999; ical mechanism of general anesthesia. Biochemistry 1997;36:
6:374 –9. 2339 – 44.
222. Doyle DA, Cabral JM, Pfuetzner RA, et al. The structure of the 231. Cantor RS. Lateral pressures in cell membranes: a mechanism
potassium channel: molecular basis of K⫹ conduction and for modulation of protein function. J Phys Chem B 1997;101:
selectivity. Science 1998;280:69 –77. 1723–5.
223. Roux B, MacKinnon R. The cavity and pore helices in the KcsA 232. Cantor RS. Solute modulation of conformational equilibria in
K⫹ channel: electrostatic stabilization of monovalent cations. intrinsic membrane proteins: apparent “cooperativity” without
Science 1999;285:100 –2. binding. Biophys J 1999;77:2643–7.
224. Aqvist J, Luzhkov V. Ion permeation mechanism of the potas- 233. Tang P, Xu Y. Large-scale molecular dynamics simulations of
sium channel. Nature 2000;404:881– 4. general anesthetic effects on the ion channel in the fully hy-
225. Eger EI II, Halsey MJ, Harris RA, et al. Hypothesis: volatile drated membrane: the implication of molecular mechanisms of
anesthetics produce immobility by acting on two sites approx- general anesthesia. Proc Natl Acad Sci U S A 2002;99:16035– 40.
imately five carbons apart. Anesth Analg 1999;88:1395– 400.
226. Pauling L. The nature of the chemical bond. 3rd ed. Ithaca, NY:
Cornell University Press, 1960:499.

You might also like