Brain Research Reviews 48 (2005) 360 – 369 www.elsevier.



From proteomics to biomarker discovery in Alzheimer’s disease
Lap Hoa, Naresh Sharmaa, Laurel Blackmana, Eugene Festaa, Guru Reddyb, Giulio Maria Pasinettia,c,d,*
Neuroinflammation Research Laboratories of the Department of Psychiatry, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029, USA b Ciphergen Biosystem, Fremont, CA 94555, USA c Department of Neuroscience, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029, USA d Department of Geriatrics and Adult Development, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029, USA Accepted 9 December 2004 Available online 9 February 2005

Abstract Alzheimer’s disease (AD) is the most common form of dementia in the elderly. AD is an invariably fatal neurodegenerative disorder with no effective treatment or definitive antemortem diagnostic test. Little is known about the changes in the brain preceding or accompanying initiation of the disease. Understanding the biological processes, which occur during AD onset and/or progression, will improve the diagnosis and treatment of the disease. As we will discuss in this review article, using high-throughput cDNA microarray we identified candidate genes whose expression is altered in the brain of cases at risk for AD dementia. However, it is possible that the use of the cDNA microarray technology alone may underestimate post-transcriptional modifications and therefore provides only a partial view of the biological problem of interest. As such, the combination of cDNA and protein arrays may provide a more global picture of the biological processes being studied. Based on this hypothesis, we initiated a series of high-throughput proteomic studies and found that the expressions of proteins involved in synaptic plasticity are selectively altered in the brain of cases at high risk to develop AD dementia (mild cognitive impairment; MCI). This is consistent with our cDNA microarray evidence showing that the expression of a-type synapsins is selectively altered in the brain of MCI cases. Collectively, these studies support the feasibility and usefulness of high-throughput cDNA microarray and proteomics techniques to study the sequential changes of distinctive gene expression patterns in the brain as a function of the progression of AD dementia. D 2004 Elsevier B.V. All rights reserved.
Theme: Proteomics Topic: Alzheimer’s disease Keywords: Proteomics; Alzheimer’s disease; Dementia

Contents 1. 2. 3. 4. 5. The use of high-throughput technology in search of molecular indexes of early AD dementia . . Proteomics and the search for novel indexes of AD clinical progression . . . . . . . . . . . . . cDNA microarrays and the search for novel biological indexes of AD clinical progression . . . . Identification of novel candidate proteins whose expression in the brain is altered in MCI cases relative to normal neurological controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consistencies between cDNA and PowerBlot proteomic arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 361 362 362 363

* Corresponding author. Neuroinflammation Research Laboratories, Mount Sinai School of Medicine, Department of Psychiatry, One Gustave L. Levy Place, New York, NY 10029, USA. Fax: +1 212 876 9042. E-mail address: (G.M. Pasinetti). 0165-0173/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.brainresrev.2004.12.025

L. Ho et al. / Brain Research Reviews 48 (2005) 360–369



Brain vs. body fluids: the search for novel biomarkers of AD dementia using surface-enhanced laser desorption ionization (SELDI )–mass spectrometry (MS) ProteinChip technology . . . . . 7. SELDI–MS proteomics in the search for AD biomarkers in biological fluids . . . . . . . . . . 8. Biomarker discovery in Alzheimer’s disease . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. Ah content in biological fluids as a prognostic biomarker of AD . . . . . . . . . . . . 8.2. Protein bcombinationQ as a diagnostic biomarker in AD . . . . . . . . . . . . . . . . . 8.3. Cytosolic Ah deposition in eye lens as an AD diagnostic biomarker . . . . . . . . . . . 8.4. Ongoing investigation of antecedent biomarkers in AD. . . . . . . . . . . . . . . . . . 9. Future perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

364 365 366 366 366 367 367 367 367 367

1. The use of high-throughput technology in search of molecular indexes of early AD dementia Critical questions in understanding the classic neuropathological features of AD (amyloid plaques and neurofibrillary tangles (NFT)) are if their occurrence precedes, is in synchrony with, or follows the earliest and mildest signs of cognitive decline. Indeed, frequently used diagnostic approaches to AD reflect more the progression of neuropathology and minimize the cognitive symptoms of the disease, especially when the deficits are mild or moderate. For example, Haroutunian et al. [31] found that cortical AD neuropathology (e.g., neuritic plaque pathology) increases significantly when the very first signs of dementia become detectable. With the exception of the visual cortex (VC), relative to cognitively intact subjects, AD neuropathology is already detectable in cerebral cortex regions (e.g., superior temporal gyrus (STG), entorhinal cortex (EC), inferior temporal gyrus) and in the hippocampal formation of early AD dementia cases. In some of these brain regions, AD neuropathology continues systematically with increasing severity of dementia [2]. The similarity of neuritic plaque pathology in the different neocortical regions in the early stages argues against a clear focus of initial neocortical abnormalities with subsequent recruitment of other regions as a function of increasing dementia. Moreover, increases in neuritic amyloid plaque density are reflected by similar increases in the levels of h-amyloid (Ah)-peptide [34], adding biochemical precision to the conclusions drawn from the neuropathological studies [61]. However, while AD neuritic Ah plaque neuropathology might significantly influence the severity of dementia during early AD, other factors, such as NFT pathology, which appear at relatively later AD dementia stages [31], may also contribute to the dementia progression. This information is of high interest and suggests that specific neuropathological features may influence AD dementia at specific stages of the disease. However, little is known about the molecular mechanism underlying the clinical progression of AD. To yield information relevant to the understanding of the molecular mechanisms involved in the onset and progression of clinical AD dementia, we initiated a series of high-throughput array studies (cDNA and proteomics) exploring the sequence of changes in gene

expression in the brain of cases at high risk to develop AD (MCI) relative to normal neurological control cases, and eventually, during the conversion from MCI through the early stages of AD dementia.

2. Proteomics and the search for novel indexes of AD clinical progression Proteomics has emerged in the last few years as a multidisciplinary and technology-driven science that focuses on proteomes: the complex of proteins expressed in a biological system, their structures, interactions, and posttranslational modifications. In particular, proteomics examines changes in protein levels and other protein alterations that result from or foster specific diseases, or are induced by various external factors, such as toxic agents. The disciplines that study protein structure, function, and level are called structural, functional, and expression proteomics, respectively. Expression proteomics, although itself under development, is the most advanced and widely used proteomic technology today. When we speak of proteomics we usually mean expression proteomics, as is the case of changes in steady-state levels of brain proteins in certain conditions (e.g., in neurodegenerative disorders). Although the term proteomics is only a few years old, its roots go back to the 1980s when the usual methods of protein identification were immunoblotting and co-migration with known purified proteins in one-dimensional gel electrophoresis technique [65]. As we will discuss below, the development of proteomics can be attributed primarily to the refinements in mass spectrometry (MS), improvements in computer and software sciences, and the flood of data now available from genomic sequencing of many organisms. The future promises the release of much more information. As a consequence, several recent studies have used this new technology to investigate common health concerns, including but not limited to, aging [12,13], bacterial infections [68], cancer [6,11,17,20,30,39,42,50,55,56,58,60,63], osteoporosis [41], atherosclerosis [16,21,44], and neuropathic and inflammatory pain [2,27]. As detailed below, protein alterations in brain of early (frank) AD dementia or in subjects at high risk for developing dementia have been


L. Ho et al. / Brain Research Reviews 48 (2005) 360–369

investigated and the research continues to proliferate [8,10,14,40,53,57,66]. As further discussed in the following sections, using cDNA microarray [53], and most recently proteomic technologies, we identified a series of novel gene products (biomarkers) whose abnormal expression in the brain may play an important role in the progression, and possibly, the onset of AD dementia.

3. cDNA microarrays and the search for novel biological indexes of AD clinical progression In high-throughput cDNA microarray studies in our laboratory, we screened 6794 (Incyte VI human microchips) human genes in the brains of cases characterized by moderate dementia, assessed by clinical dementia rating (CDR) 2 [33,35]. We found 32 genes (25 known and 7 EST) with N1.8 altered expression in the superior temporal gyrus relative to cases characterized by normal cognitive status [33]. Among others, we found down-regulation of the synaptic vesicle protein synapsin IIa, which normally plays an important role in synaptic vesicle turnover and neurotransmitter release [29,32]. We decided to continue our investigation by assessing the changes in protein expression of the known splice variants I, II, and III of the a- and b-type synapsin isoforms [32] by Western blot analysis. We examined brain regions at high risk (EC) or relatively unaffected (VC) during AD or in normal neurological control cases. Interestingly, relative to age matched control cases characterized by normal cognitive status (CDR 0), a N2-fold decrease of the splice variants I, II, and III of the atype synapsin isoform was found in the EC of MCI cases (CDR 0.5) [33]. In contrast, no detectable change of the splice variant II of the b-type synapsin isoform or synaptophysin was found in the same EC region. The altered expressions of synapsin a-type isoforms in the EC of MCI cases was region specific since no detectable decreases were found in VC from the same cases [33]. These results provide novel molecular evidence for the importance of selective altered gene expression during the earliest detectable stage of AD dementia. In view of the evidence that synapsin family members are important in the regulation of neurotransmitter release, our cDNA microarray studies suggest that the altered regulation of selected synapsin-mediated neurotransmitter release may be involved in the early phase of AD cognitive decline. While previous evidence suggested that MCI is characterized by neuronal loss (e.g., in the EC) and qualified neuropathologically for diagnosis of frank AD [45], we note that the MCI cases used in our cDNA microarray studies did not meet neuropathological AD criteria [33]. Most importantly, we also found that the expression of the pre-synaptic plasma membrane protein synaptophysin was unaffected in the EC of MCI cases, suggesting that the decreased expression of the a-type synapsin occurred in absence of structural loss of synapses. Thus, as summarized in Scheme 1, our working
Scheme 1. Selective alteration of functional synaptic proteins, e.g., a-type synapsin in synapses that are morphologically present may be an underlying mechanism in MCI cognitive impairment.

hypothesis regarding potential mechanisms involved in the conversion from normal cognitive status to MCI and eventually to AD dementia is that a potential selective and specific malfunction of synapses that are morphologically intact (e.g., altered synaptic vesicle release-cycling) might underlie the initial cognitive decline in the MCI cases. However, we also note that a recent study indicated a minor decline of synaptophysin expression in the frontal cortex of cases characterized by MCI and also mild dementia (CDR 0.5–1) as compared to cognitively normal individuals [45]. While the mechanisms leading to the selective altered expression of a-type synapsin in the EC of MCI cases is unknown [33], independent studies in our lab found that excitotoxicity, but not h-amyloid (Ah1–42) toxicity, recapitulates the selective altered expression of a-type synapsin in cortico-hippocampal neurons, in vitro [47] (Ho and Pasinetti, unpublished observations). Finally, because recent evidence suggests that MCI clinically progresses to greater dementia severity at rates dependent on the level of cognitive impairment, and that MCI cases often have the neuropathologic features of AD [33], our studies provide a novel biological basis for a better understanding of the mechanisms underlying the conversion from MCI to frank AD dementia. Collectively, our cDNA microarray studies strengthen the utility of our systematic approach of exploring molecular events in the brain of MCI cases and the subsequent development of independent studies to further clarify underlying mechanisms of clinical progression from MCI to AD dementia.

4. Identification of novel candidate proteins whose expression in the brain is altered in MCI cases relative to normal neurological controls As discussed above, we note that cDNA microarray techniques primarily provide a snapshot of steady state mRNA levels at one specific time under certain conditions, potentially underestimating the occurrence of post-transcriptional and translational modifications of target gene

L. Ho et al. / Brain Research Reviews 48 (2005) 360–369


products of interest. While in independent cDNA array studies we continued the characterization of 25 known candidate gene products whose expression is selectively altered in the brain of early AD [53], we initiated PowerBlot proteomic (BD Transduction Laboratories, Lexington, KY) studies that employed Western blot techniques for largescale application and allowed the screening of 750 independent proteins in a single assay. In ongoing studies in the lab we identified 50 candidate protein species whose content is consistently and reproducibly altered (N2.0-fold) in the EC of MCI cases (n = 6) relative to cases normal cognitive controls (CDR 0) (n = 5). The expression level of these 50 candidate index proteins of early AD was further explored using miniaturized PowerBlot immunoarrays employing a customized cocktail of 50 antibodies raised against our candidate index protein species. As shown in Fig. 1, 23 differentially regulated proteins (N2-fold; P b 0.05) were identified and primarily related to clusters of specific biological variations. Five clusters were found to have patterns that correlated primarily with variations of proteins involved in the following: (1) synaptic functions/neurotransmitter related, (2) cytoskeleton/cell adhesion, (3) cell cycle, (4) apoptosis, and (5) transcription/ translation. To perform an independent validation of the PowerBlot data, we used traditional Western immunoblot techniques to evaluate the level of expression of individual MCI-related protein species across a larger cohort of MCI cases and cognitively normal controls (n = 7–8 per CDR group). Consistent with the PowerBlot evidence, we confirmed a N3-fold decrease in the expression of the synaptic protein tomosyn in the EC MCI cases (Fig. 2A) relative to normal cognitive control cases (Fig. 2B). Tomosyn plays a major role in the mechanisms associated with neurotransmitter vesicle pool refilling in a calcium-dependent manner [69]. This result strengthens the hypothesis that a functional alteration of synaptic functions may be at the basis of cognitive alterations in MCI.

5. Consistencies between cDNA and PowerBlot proteomic arrays Our cDNA microarray studies discussed above demonstrate that the expression of selected gene products whose activity play an important role in synaptic activities [38] are differentially regulated in the brain of MCI cases. In particular, as discussed in Pasinetti [52], we found that the expression of gene products belonging to synaptic functions/neurotransmitter related included synapsin II [29,32], N-ethylmaleimide-sensitive factor [71], tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activating protein [37] is altered in the brain of AD. Consistent with this evidence, the PowerBlot proteomic immunoblot

Fig. 1. Differential protein expression in the brain of cases characterized by MCI relative to normal cognitive controls. In this study, the PowerBlot proteomic technology was used to assess the expression profile of 50 selected proteins in the EC of MCI cases (CDR 0.5) compared to (CDR 0) cases. Differentially regulated proteins were identified based on consistency of up- or down-regulation by N2.0-fold in a three-step screening as discussed in the text. Identified proteins were clustered based on their known biologic functions as defined and established by the nomenclature committee of the International Union of Biochemistry and Molecular biology (IUBMB). Up and down arrows indicate increased/decreased protein, respectively. Synapsin IIa, which we found to be down-regulated in the EC of MCI cases brain by cDNA microarray, is not among the 750 proteins being analyzed by PowerBlot. Quantitative comparisons of specific protein species across multiple gels (e.g., CDR cases) were based on normalized expression data with internal h-actin immunoreactivity (not shown).

studies showed alterations in the expression of three additional proteins involved in synaptic vesicle/neurotransmitter related functions (tomosyn [46], Rabaptin-5 [49], and vti-1b [3]) in the brain of MCI cases relative to neurological control cases. Although the identification of altered expressed gene products in our cDNA and PowerBlot array studies may be influenced by a random bias in the selection of target cDNA or the protein assayed, including the possibility of lack of representation


L. Ho et al. / Brain Research Reviews 48 (2005) 360–369

Fig. 2. Tomosyn protein expression is down-regulated in the EC during conversion from normal cognitive status to MCI. (A) Representative high-throughput PowerBlot immunoblot of EC protein extracts from CDR 0 vs. MCI (CDR 0.5) cases (secondary screening with 750 antibodies, n = 5–6 per group). The PowerBlot proteomic array used classic Western blot detection system and allowed the screening of 750 independent proteins; proteins were resolved by molecular weight using SDS–polyacrylamide gel electrophoresis and then transferred from the slab gel onto a sheet of nylon membrane, visualized by chemiluminescence, and quantified by phosphoimaging. In our studies, 10% of the immunoreactive bands having the strongest or weakest signals were not considered for calculation. A total sum of signals was generated from the remaining immunoreactive bands and the baverage signal per bandQ was calculated by dividing the total sum of signal by the total number of bands. Thereafter, immunoreactivity for each band was normalized and expressed as a percentage of the baverage signal per band; red arrows points toward tomosyn immunoreactivity in both CDR 0 and CDR 0.5 (MCI) immunoblots. (B) Regulation of tomosyn protein expression in the EC of individual CDR 0 and MCI cases assessed by conventional Western blot techniques; tomosyn antibody (BD Transduction Laboratories is the same antibody used in PowerBlot screening; 1: 500 dilutions). Data represent tomosyn normalized to actin and expressed as percentage of CDR 0 values. Student unpaired t test: *P b 0.01 vs. CDR 0. (B inset) Representative tomosyn and h-actin immunoreactive bands.

of pertinent gene products in both arrays, the present studies demonstrate that the two technologies can complement each other in the search of molecular markers of early AD.

6. Brain vs. body fluids: the search for novel biomarkers of AD dementia using surface-enhanced laser desorption ionization (SELDI )–mass spectrometry (MS) ProteinChip technology As compared to the brain, body fluids have the advantage of being more easily accessible for the identification of potential biomarkers that might reflect AD dementia progression. Procuring serum/plasma (and to some extent, CSF) is a non-invasive procedure and is thus more likely to be of clinical use. However, the identification of a biomarker in these samples requires that the protein be secreted at high enough levels to be identified. Moreover, circulating factors may be secreted by multiple cell types from different organs and may therefore lack specificity for the pathophysiological event being studied in the brain. For example, a complicating factor for biomarker studies in serum/plasma is that levels of albumin, transferrin, and IgG are likely to be more abundant than the levels of the potential biomarker and may interfere with biomarker detection. While the brain may still represent a starting point in the search for novel biomarkers of early clinical AD dementia, ongoing SELDI–MS studies in the lab are exploring variation in protein expression patterns in cerebral spinal fluid (CSF) during the transition from

normal cognitive functioning to MCI through frank AD dementia. SELDI–MS technology is a system that enables rapid protein profiling, identification, and characterization from crude biological samples [25]. In particular, this system uses ProteinChip Arrays that contain chemically (cationic, anionic, hydrophobic, hydrophobic, etc.) or biochemically (e.g., antibody, receptor, DNA, etc.) treated surfaces [60] (Fig. 3). Crude biological extracts (e.g., brain tissue extract, CSF, serum/plasma) are applied to the ProteinChip Arrays, which selectively capture subclasses of proteins with specific physical or biochemical characteristics based on protein interactions with the ProteinChip Array surfaces, as shown for detection of Ah-peptides (Fig. 4). Using this technology, the molecular size and quantity of individual proteins absorbed on each chip are then directly assessed by a btime of flightQ mass spectrometer [60]. This generates a quantitative protein mass profile of the proteins bound to each of the ProteinChip Array surfaces. While control samples are always run in parallel with the experimental samples, the resulting profiles can be compared directly using a number of integrated software features to highlight relevant changes in the patterns of underlying protein expression. The SELDI technology is highly sensitive; the lower limit of detection is 10 fmol [54]. In addition, protein analysis using the SELDI technology is quantitative and highly reproducible. As such, SELDI technology presently available in our laboratories is more amenable to highthroughput proteomic procedures and presents a viable platform for the discovery and validation of potential biomarkers of early AD dementia. For protein identification

L. Ho et al. / Brain Research Reviews 48 (2005) 360–369


Fig. 3. Chemical and biological surfaces for SELDI–MS protein expression profiling and protein interaction assays using ProteinChips technology.

purposes, the same chip platform used for biomarker discovery and validation can easily be incorporated into strategies to facilitate rapid target protein purification. Since purified protein can be digested on-chip with proteases, the same chip platform will be useful for generating peptide maps that can be compared to a peptide database for protein identification. Recent evidence confirmed the utility of the SELDI technology in monitoring proteins and protein profiles for diagnostic applications including staging of disease and disease progression [4,24,54,69]. Most excitingly, a recent study successfully employed the SELDI–MS proteomic technology to identify biomarkers in the circulating cells that predicted HIV-1 associated cognitive impair-

ment with a sensitivity of 100% and a specificity of 75% [43].

7. SELDI–MS proteomics in the search for AD biomarkers in biological fluids The goal of biomarker discovery studies is to identify protein profiles, possibly in accessible biological fluids, which can ultimately be used to develop a rapid, sensitive, and specific high-throughput diagnostic assay. While we are aware that 2D electrophoresis can effectively resolve a large number of proteins, we are also aware of its limited

Fig. 4. Schematic representation of Ah detection system using PS10 antibody ProteinChips by SELDI–MS protein expression analysis.


L. Ho et al. / Brain Research Reviews 48 (2005) 360–369

reproducibility rendering it difficult for use in biomarker discovery and validation studies involving multiple cases and multiple disease stages. The most likely scenario for studies using 2D gel technology would require the identification of the protein, the production of a highquality antibody to the protein, and then the development of an ELISA or other sensitive immunoassay. This may be acceptable if a diagnostic test with 95% accuracy would be involved. However, because of the heterogeneity expected in our samples, e.g., MCI vs. normal cognitive cases, it is conceivable that multiple biomarkers will be required to accurately discriminate among cases characterized by different dementia stages and normal cognitive status. The advantage that SELDI ProteinChip technology has over 2D electrophoresis is that the same chip platform used to identify the biomarker can also be used to develop a rapid, sensitive, and high-throughput assay. Protein profiling using SELDI technology offers a novel means for identifying and eventually monitoring the onset and progression of AD clinical dementia. An extra advantage of using SELDI–MS technology is that it will not be essential to identify the proteins themselves in order to make a diagnosis. Moreover, as further discussed below, by utilizing bgroups of biomarkersQ we will be not constrained by the sensitivity or specificity of any single biomarker. Next we will discuss the feasibility of using SELDI–MS technology in the identification and characterization of potential prognostic, diagnostic, and possibly biomarkers antecedent of AD and/or MCI.

8.1. Ab content in biological fluids as a prognostic biomarker of AD Amyloid plaque deposits in the AD brain consist mainly of aggregates of Ah-peptide [19,26], a 4 kDa peptide derived from proteolytic processing of a large transmembrane amyloid precursor (APP). Two forms of Ah-peptide, 1–40 and 1–42, of which 1–42 is believed to be the more toxic [22,64], constitute the primary components of the amyloid plaques. Numerous N- and C-terminal truncated forms have also been identified in clinical samples from patients with AD [1]. Precise mapping of the patterns of these various species of Ah-peptides in tissue extracts and biological fluids would provide a classical example of a prognostic biomarker and eventually contribute to our understanding of AD pathogenesis. In the past, amyloidogenic Ah-peptides have been analyzed by ELISA, which involves indirect chemical or radioactive detection methods [15,23,36,51]. More recently, the SELDI ProteinChip Ah assay discussed above proved to be an advantageous approach for the simultaneous detection of multiple Ahpeptide forms. For example, a monoclonal antibody raised against the N-terminal amino acid sequence 1–10 is covalently coupled to a pre-activated ProteinChip array to create a biologically active surface [7,18]. Moreover, samples of CSF and/or crude cellular or tissue extract can be added to the ProteinChip array, which captures multiple Ah-peptide fragments that have a common N-terminal but differ at their C-terminals. This assay has been widely used in studies of AD pathogenesis [18,28,67,70], the role of hand g-secretase processing of APP [4,5,24], as well as in discovery studies exploring new drugs [5,18], and the potential development of a vaccine therapy for AD [67]. 8.2. Protein bcombinationQ as a diagnostic biomarker in AD A combined increase in CSF levels of the Ah1–42 peptide and tau proteins was recently proposed as a diagnostic indicator of AD [59]. However, this CSF characteristic does not have the necessary sensitivity and specificity for use as a routine clinical assay. As discussed above, a definitive diagnosis of AD can still only be made by a postmortem assessment of brain neuropathology. There is still demand for a simple biochemical test for early diagnosis of AD, a test that could distinguish AD or MCI patients from normal cognitive elderly subjects and distinguish between AD and other forms of dementia. A recent report is promising. For example, Carrette and colleagues [9], using a SELDI ProteinChip array technique, demonstrated a panel of five polypeptides with significantly altered expression levels (four were up-regulated and one was down-regulated) in CSF of nine AD patients as compared to CSF from 10 healthy control subjects; four of these biomarkers were further purified and identified by MS as cystatin-C, two h-2-microglobulin isoforms, a 4.8 kDa VGF polypeptide, and an unknown 7.7 kDa polypeptide. Using

8. Biomarker discovery in Alzheimer’s disease An ideal biological marker for AD may represent an essential mechanism of the disease, and its validity should be confirmed by the underlying neuropathology. The biomarker should be sensitive and specific for detecting AD, as well as reliable, accessible, and non-invasive (reviewed in Morris [48]). The development of reliable and valid biomarkers is important because AD is a highly prevalent disorders but remain under diagnosed and under treated. For example, prognostic biomarkers that serve as indicators of AD progression could be used as surrogate endpoints for therapeutic responses, possibly reducing the time, effort, and cost associated with drug trials that currently require large samples to be followed for long time periods to measure clinical and cognitive outcomes [26]. Moreover, diagnostic biomarkers would aid in AD recognition and enable patients to benefit from available symptomatic drug therapies. Perhaps, and most importantly, antecedent biomarkers could be used to identify those cases at high risk for developing AD (MCI). Not only could antecedent biomarkers inform research into the cause or causes of AD, they might also used to predict those individuals who might benefit most from the development of disease-modifying or preventive interventions.

L. Ho et al. / Brain Research Reviews 48 (2005) 360–369


the combination of these five markers, Carrette and colleagues correctly classified six of the nine AD patients and all 10 controls demonstrating a 100% specificity and 66% sensitivity in a small sample size [9]. The CSF Ah concentration ratio (Ah1–42/Ah1–40) has also been proposed as a diagnostic tool for distinguishing patients with AD from normal control subjects and subjects with non-AD forms of dementia [62]. This assay has a diagnostic specificity of 88% and a sensitivity of 59%. Without a doubt, the five-polypeptide panel test, with a greater specificity and sensitivity, might be more promising as a diagnostic tool. In the future, if this evidence can be validated on a larger set of clinical samples, it could be used as a definitive antecedent biomarker for AD, or perhaps for assessing the probability, severity, and progression of AD as a prognostic biomarker. 8.3. Cytosolic Ab deposition in eye lens as an AD diagnostic biomarker Until recently, AD-associated accumulation of Ah amyloid plaques had only been observed in the brain. However, Goldstein and colleagues [28] have discovered cytosolic Ah-peptides in lens fiber cells from AD patients. The investigators first detected a high-molecular-weight protein and a low-molecular-weight (around 4 kDa) peptide in protein extracts of human lens. The protein and peptide were identified by Western immunoblotting as full-length APP and Ah-peptide, respectively, and sequencing with electrospray ionization MS–MS tandem technology confirmed the identities. Tryptic analysis of the peptides of the 4 kDa band excised from the gel revealed a 12-residue sequence that is an internal fragment of the Ah-peptide, a sequence that is common to both the Ah1–40 and Ah1–42 peptides. The MS–MS tandem technology could therefore not distinguish between the two Ah-peptide forms. Goldstein et al. [28] then used the SELDI ProteinChip-based Ah assay to analyze the lens extracts and identified both forms of Ah-peptides, neither of which was detectable in lens extracts from control subjects. The study tentatively suggests that Ah can be detected in lens cells of AD cases where it might promote region-specific lens protein aggregation, extra-cerebral amyloid formation, and eventually cataracts. 8.4. Ongoing investigation of antecedent biomarkers in AD Based on the encouraging cDNA/PowerBlot highthroughput array data discussed above, we have initiated a series of new studies to evaluate the potential of SELDI– MS technology as a means to identify novel protein biomarkers that may be used as indices of onset and eventually progression of early AD cognitive decline. In ongoing studies using SELDI–MS microchips for the detection of negatively charged or copper-binding proteins (Fig. 4), we found a novel protein biomarker whose levels

are increased in the CSF of four early MCI cases relative to four normal cognitive control cases [35]. We are currently exploring the possibility of using this novel protein as a biomarker across the entire spectrum of clinical AD by assessing its regulation in CSF as a function of progression of clinical dementia (unpublished observations). Work is presently in progress in our laboratory to purify this protein and to determine its identity by amino acid sequencing as a means to understand its biological functions in normal and pathologic conditions.

9. Future perspectives It is clear that the combination of high throughput cDNA and protein microarrays will have a significant effect on neurobiological research and on efforts to discover the mechanisms involved in neuropsychiatric diseases. Although these approaches are powerful, we note that they should be used to complement other traditional methods, such as those used in molecular biology and genetics. Most importantly they should be used with great care and rigor to avoid misidentification of false positives. The relatively youthful stage of proteomic techniques has left the field without a bgold standardQ to guide investigators and enhance comparability of proteomics and biomarker discovery studies in the field of AD. A recent consensus panel (Antecedent Biomarkers in Alzheimer’s Disease, in Alzheimer’s Forum, 2003) suggested that an ideal biomarker for AD should detect a fundamental feature of AD neuropathology that needs to be validated in neuropathologically confirmed cases. This biomarker should have a diagnostic sensitivity N80% for detecting AD and a specificity of N80% for distinguishing other forms of dementia. Recommended steps to establish a biomarker include (1) confirmation by at least two independent studies conducted by qualified investigators with the results published in peer-reviewed journals, and (2) evidence supporting the utility of the biomarker in assessing a beneficial effect in AD diseasemodifying therapy. The work has only just begun.

Acknowledgment Supported by The Dana Foundation for Brain Research Initiatives to GMP.

[1] H. Akiyama, H. Kondo, H. Mori, F. Kametani, T. Nishimura, K. Ikeda, M. Kato, P.L. McGeer, The amino-terminally truncated forms of amyloid beta-protein in brain macrophages in the ischemic lesions of Alzheimer’s disease patients, Neurosci. Lett. 219 (1996) 115 – 118. [2] O. Alzate, S.R. Hussain, V.M. Goettl, A.K. Tewari, F. Madiai, R.L. Stephens Jr., K.V. Hackshaw, Proteomic identification of brainstem


L. Ho et al. / Brain Research Reviews 48 (2005) 360–369 cytosolic proteins in a neuropathic pain model, Brain Res., Mol. Brain Res. 128 (2004) 193 – 200. W. Antonin, D. Riedel, G.F. von Mollard, The SNARE Vti 1a-beta is localized to small synaptic vesicles and participates in a novel SNARE complex, J. Neurosci. 20 (2000) 5724 – 5732. B.M. Austen, E.R. Frears, H. Davies, The use of seldi proteinchip arrays to monitor production of Alzheimer’s betaamyloid in transfected cells, J. Pept. Sci. 6 (2000) 459 – 469. D. Beher, J.D. Wrigley, A.P. Owens, M.S. Shearman, Generation of Cterminally truncated amyloid-beta peptides is dependent on gammasecretase activity, J. Neurochem. 82 (2002) 563 – 575. F. Boraldi, L. Bini, S. Liberatori, A. Armini, V. Pallini, R. Tiozzo, I.P. Ronchetti, D. Quaglino, Normal human dermal fibroblasts: proteomic analysis of cell layer and culture medium, Electrophoresis 24 (2003) 1292 – 1310. L.E. Bradbury, J.F. LeBlanc, D.B. McCarthy, ProteinChip array-based amyloid beta assays, Methods Mol. Biol. 264 (2004) 245 – 257. D.A. Butterfield, Proteomics: a new approach to investigate oxidative stress in Alzheimer’s disease brain, Brain Res. 1000 (2004) 1 – 7. O. Carrette, I. Demalte, A. Scherl, O. Yalkinoglu, G. Corthals, P. Burkhard, D.F. Hochstrasser, J.C. Sanchez, A panel of cerebrospinal fluid potential biomarkers for the diagnosis of Alzheimer’s disease, Proteomics 3 (2003) 1486 – 1494. A. Castegna, V. Thongboonkerd, J.B. Klein, B. Lynn, W.R. Markesbery, D.A. Butterfield, Proteomic identification of nitrated proteins in Alzheimer’s disease brain, J. Neurochem. 85 (2003) 1394 – 1401. T.D. Chanin, D.T. Merrick, W.A. Franklin, F.R. Hirsch, Recent developments in biomarkers for the early detection of lung cancer: perspectives based on publications 2003 to present, Curr. Opin. Pulm. Med. 10 (2004) 242 – 247. W. Chen, J. Ji, X. Xu, S. He, B. Ru, Proteomic comparison between human young and old brains by two-dimensional gel electrophoresis and identification of proteins, Int. J. Dev. Neurosci. 21 (2003) 209 – 216. Y.M. Cho, S.H. Bae, B.K. Choi, S.Y. Cho, C.W. Song, J.K. Yoo, Y.K. Paik, Differential expression of the liver proteome in senescence accelerated mice, Proteomics 3 (2003) 1883 – 1894. L.H. Choe, M.J. Dutt, N. Relkin, K.H. Lee, Studies of potential cerebrospinal fluid molecular markers for Alzheimer’s disease, Electrophoresis 23 (2002) 2247 – 2251. M. Citron, D. Westaway, W. Xia, G. Carlson, T. Diehl, G. Levesque, K. Johnson-Wood, M. Lee, P. Seubert, A. Davis, D. Kholodenko, R. Motter, R. Sherrington, B. Perry, H. Yao, R. Strome, I. Lieberburg, J. Rommens, S. Kim, D. Schenk, P. Fraser, H.P. St. George, D.J. Selkoe, Mutant presenilins of Alzheimer’s disease increase production of 42residue amyloid beta-protein in both transfected cells and transgenic mice, Nat. Med. 3 (1997) 67 – 72. J.A. Coppinger, G. Cagney, S. Toomey, T. Kislinger, O. Belton, J.P. McRedmond, D.J. Cahill, A. Emili, D.J. Fitzgerald, P.B. Maguire, Characterization of the proteins released from activated platelets leads to localization of novel platelet proteins in human atherosclerotic lesions, Blood 103 (2004) 2096 – 2104. J.F. Cui, Y.K. Liu, B.S. Pan, H.Y. Song, Y. Zhang, R.X. Sun, J. Chen, J.T. Feng, Z.Y. Tang, Y.L. Yu, H.L. Shen, P.Y. Yang, Differential proteomic analysis of human hepatocellular carcinoma cell line metastasis-associated proteins, J. Cancer Res. Clin. Oncol. 130 (2004) 615 – 622. H. Davies, L. Lomas, B. Austen, Profiling of amyloid beta peptide variants using SELDI ProteinChip arrays, BioTechniques 27 (1999) 1258 – 1261. D.W. Dickson, The pathogenesis of senile plaques, J. Neuropathol. Exp. Neurol. 56 (1997) 321 – 339. B. Duggan, K. Williamson, Molecular markers for predicting recurrence, progression and outcomes of bladder cancer (do the poster boys need new posters?), Curr. Opin. Urol. 14 (2004) 277 – 286. M.C. Duran, S. Mas, J.L. Martin-Ventura, O. Meilhac, J.B. Michel, J. Gallego-Delgado, A. Lazaro, J. Tunon, J. Egido, F. Vivanco, Proteomic analysis of human vessels: application to atherosclerotic plaques, Proteomics 3 (2003) 973 – 978. O.M. El Agnaf, D.S. Mahil, B.P. Patel, B.M. Austen, Oligomerization and toxicity of beta-amyloid-42 implicated in Alzheimer’s disease, Biochem. Biophys. Res. Commun. 273 (2000) 1003 – 1007. G. Franz, R. Beer, A. Kampfl, K. Engelhardt, E. Schmutzhard, H. Ulmer, F. Deisenhammer, Amyloid beta 1–42 and tau in cerebrospinal fluid after severe traumatic brain injury, Neurology 60 (2003) 1457 – 1461. E.R. Frears, D.J. Stephens, C.E. Walters, H. Davies, B.M. Austen, The role of cholesterol in the biosynthesis of beta-amyloid, NeuroReport 10 (1999) 1699 – 1705. E.T. Fung, G.L. Wright Jr., E.A Dalmasso, Proteomic strategies for biomarker identification: progress and challenges, Curr. Opin. Mol. Ther. 2 (2000) 643 – 650. B.I. Giasson, V.M. Lee, J.O. Trojanowski, Interactions of amyloidogenic proteins, Neuromol. Med. 4 (2003) 49 – 58. C. Gineste, L. Ho, P. Pompl, M. Bianchi, G.M. Pasinetti, Highthroughput proteomics and protein biomarker discovery in an experimental model of inflammatory hyperalgesia: effects of nimesulide, Drugs 63 (Suppl. 1) (2003) 23 – 29. L.E. Goldstein, J.A. Muffat, R.A. Cherny, R.D. Moir, M.H. Ericsson, X. Huang, C. Mavros, J.A. Coccia, K.Y. Faget, K.A. Fitch, C.L. Masters, R.E. Tanzi, L.T. Chylack Jr., A.I. Bush, Cytosolic betaamyloid deposition and supranuclear cataracts in lenses from people with Alzheimer’s disease, Lancet 361 (2003) 1258 – 1265. P. Greengard, F. Valtorta, A.J. Czernik, F. Benfenati, Synaptic vesicle phosphoproteins and regulation of synaptic function, Science 259 (1993) 780 – 785. M.B. Gretzer, D.W. Chan, C.L. van Rootselaar, J.M. Rosenzweig, S. Dalrymple, L.A. Mangold, A.W. Partin, R.W. Veltri, Proteomic analysis of dunning prostate cancer cell lines with variable metastatic potential using SELDI–TOF, Prostate 60 (2004) 325 – 331. V. Haroutunian, D.P. Purohit, D.P. Perl, D. Marin, K. Khan, M. Lantz, K.L. Davis, R.C. Mohs, Neurofibrillary tangles in nondemented elderly subjects and mild Alzheimer disease, Arch. Neurol. 56 (1999) 713 – 718. S. Hilfiker, V.A. Pieribone, A.J. Czernik, H.T. Kao, G.J. Augustine, P. Greengard, Synapsins as regulators of neurotransmitter release, Philos. Trans. R. Soc. London, B Biol. Sci. 354 (1999) 269 – 279. L. Ho, Y. Guo, L. Spielman, O. Petrescu, V. Haroutunian, D. Purohit, A. Czernik, S. Yemul, P.S. Aisen, R. Mohs, G.M. Pasinetti, Altered expression of a-type but not b-type synapsin isoform in the brain of patients at high risk for Alzheimer’s disease assessed by DNA microarray technique, Neurosci. Lett. 298 (2001) 191 – 194. L. Ho, D. Purohit, V. Haroutunian, J.D. Luterman, F. Willis, J. Naslund, J.D. Buxbaum, R.C. Mohs, P.S. Aisen, G.M. Pasinetti, Neuronal cyclooxygenase 2 expression in the hippocampal formation as a function of the clinical progression of Alzheimer disease, Arch. Neurol. 58 (2001) 487 – 492. L. Ho, C. Gineste, P.N. Pompl, A. Dang, M. Schall, G.M. Pasinetti, Expression of Psoriasin and Cystain C in the CSF of Early Alzheimer’s Disease, 2nd Annual Meeting of the Society of Neuroscience Orlando, Fl (Abstr.) (2002). L. Ho, W. Qin, P.N. Pompl, Z. Xiang, J. Wang, Z. Zhao, Y. Peng, G. Cambareri, A. Rocher, C.V. Mobbs, P.R. Hof, G.M. Pasinetti, Diet-induced insulin resistance promotes amyloidosis in a transgenic mouse model of Alzheimer’s disease, FASEB J. 18 (2004) 902 – 904. S.E. Hufton, I.G. Jennings, R.G. Cotton, Structure and function of the aromatic amino acid hydroxylases, Biochem. J. 311 (Pt. 2) (1995) 353 – 366. T. Ichimura, T. Isobe, T. Okuyama, N. Takahashi, K. Araki, R. Kuwano, Y. Takahashi, Molecular cloning of cDNA coding for brainspecific 14-3-3 protein, a protein kinase-dependent activator of tyrosine and tryptophan hydroxylases, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 7084 – 7088.









[7] [8] [9]

[26] [27]




















[19] [20]



L. Ho et al. / Brain Research Reviews 48 (2005) 360–369 [39] C.H. Kim, D.K. Kim, S.J. Choi, K.H. Choi, K.S. Song, J. Chi, J.S. Koo, S.Y. Hwang, J.H. Yoon, J.K. Seong, Proteomic and transcriptomic analysis of interleukin-1beta treated lung carcinoma cell line, Proteomics 3 (2003) 2454 – 2471. [40] M.A. Korolainen, G. Goldsteins, I. Alafuzoff, J. Koistinaho, T. Pirttila, Proteomic analysis of protein oxidation in Alzheimer’s disease brain, Electrophoresis 23 (2002) 3428 – 3433. [41] K. Kubota, K. Wakabayashi, T. Matsuoka, Proteome analysis of secreted proteins during osteoclast differentiation using two different methods: two-dimensional electrophoresis and isotope-coded affinity tags analysis with two-dimensional chromatography, Proteomics 3 (2003) 616 – 626. [42] X. Liu, Y. Wu, Z.E. Zehner, C. Jackson-Cook, J.L. Ware, Proteomic analysis of the tumorigenic human prostate cell line M12 after microcell-mediated transfer of chromosome 19 demonstrates reduction of vimentin, Electrophoresis 24 (2003) 3445 – 3453. [43] X. Luo, K.A. Carlson, V. Wojna, R. Mayo, T.M. Biskup, J. Stoner, J. Anderson, H.E. Gendelman, L.M. Melendez, Macrophage proteomic fingerprinting predicts HIV-1-associated cognitive impairment, Neurology 60 (2003) 1931 – 1937. [44] W. Martinet, D.M. Schrijvers, G.R. De Meyer, A.G. Herman, M.M. Kockx, Western array analysis of human atherosclerotic plaques: downregulation of apoptosis-linked gene 2, Cardiovasc. Res. 60 (2003) 259 – 267. [45] E. Masliah, M. Mallory, M. Alford, R. DeTeresa, L.A. Hansen, D.W. McKeel Jr., J.C. Morris, Altered expression of synaptic proteins occurs early during progression of Alzheimer’s disease, Neurology 56 (2001) 127 – 129. [46] E.S. Masuda, B.C. Huang, J.M. Fisher, Y. Luo, R.H. Scheller, Tomosyn binds t-SNARE proteins via a VAMP-like coiled coil, Neuron 21 (1998) 479 – 480. [47] M. Mirjany, L. Ho, G.M. Pasinetti, Role of cyclooxygenase-2 in neuronal cell cycle activity and glutamate-mediated excitotoxicity, J. Pharmacol. Exp. Ther. 301 (2002) 494 – 500. [48] J.C. Morris, Dementia update 2003, Alzheimer Dis. Assoc. Disord. 17 (2003) 245 – 258. [49] R.L. Neve, R. Coopersmith, D.L. McPhie, C. Santeufemio, K.G. Pratt, C.J. Murphy, S.D. Lynn, The neuronal growth-associated protein GAP-43 interacts with rabaptin-5 and participates in endocytosis, J. Neurosci. 18 (1998) 7757 – 7767. [50] D.K. Ornstein, E.F. Petricoin III, Proteomics to diagnose human tumors and provide prognostic information, Oncology (Huntingt) 18 (2004) 521 – 529. [51] S. Parvathy, P. Davies, V. Haroutunian, D.P. Purohit, K.L. Davis, R.C. Mohs, H. Park, T.M. Moran, J.Y. Chan, J.D. Buxbaum, Correlation between Abetax-40-, Abetax-42-, and Abetax-43-containing amyloid plaques and cognitive decline, Arch. Neurol. 58 (2001) 2025 – 2032. [52] G.M. Pasinetti, Use of cDNA microarray in the search for molecular markers involved in the onset of Alzheimer’s disease dementia, J. Neurosci. Res. 65 (2001) 471 – 476. [53] G.M. Pasinetti, L. Ho, From cDNA microarrays to high-throughput proteomics. Implications in the search for preventive initiatives to slow the clinical progression of Alzheimer’s disease dementia, Restor. Neurol. Neurosci. 18 (2001) 137 – 142. [54] C.P. Paweletz, D.K. Ornstein, M.J. Roth, V.E. Bichsel, J.W. Gillespie, V.S. Calvert, C.D. Vocke, S.M. Hewitt, P.H. Duray, J. Herring, Q.H. Wang, N. Hu, W.M. Linehan, P.R. Taylor, L.A. Liotta, M.R. EmmertBuck, E.F. Petricoin III, Loss of annexin 1 correlates with early onset





[58] [59]


[61] [62] [63]









of tumorigenesis in esophageal and prostate carcinoma, Cancer Res. 60 (2000) 6293 – 6297. E.F. Petricoin, D.K. Ornstein, L.A. Liotta, Clinical proteomics: applications for prostate cancer biomarker discovery and detection, Urol. Oncol. 22 (2004) 322 – 328. E.M. Posadas, B. Davidson, E.C. Kohn, Proteomics and ovarian cancer: implications for diagnosis and treatment: a critical review of the recent literature, Curr. Opin. Oncol. 16 (2004) 478 – 484. M. Puchades, S.F. Hansson, C.L. Nilsson, N. Andreasen, K. Blennow, P. Davidsson, Proteomic studies of potential cerebrospinal fluid protein markers for Alzheimer’s disease, Brain Res., Mol. Brain Res. 118 (2003) 140 – 146. A.J. Rai, D.W. Chan, Cancer proteomics: serum diagnostics for tumor marker discovery, Ann. N. Y. Acad. Sci. 1022 (2004) 286 – 294. N.S. Schoonenboom, Y.A. Pijnenburg, C. Mulder, S.M. Rosso, E.J. Van Elk, G.J. Van Kamp, J.C. Van Swieten, P. Scheltens, Amyloid beta(1–42) and phosphorylated tau in CSF as markers for early-onset Alzheimer disease, Neurology 62 (2004) 1580 – 1584. B. Seliger, R. Lichtenfels, R. Kellner, Detection of renal cell carcinoma-associated markers via proteome- and other domeT-based analyses, Brief. Funct. Genomics Proteomics 2 (2003) 194 – 212. D.J. Selkoe, The origins of Alzheimer disease: a is for amyloid, JAMA 283 (2000) 1615 – 1617. M. Shoji, Cerebrospinal fluid Abeta40 and Abeta42: natural course and clinical usefulness, Front. Biosci. 7 (2002) d997 – d1006. R.I. Somiari, A. Sullivan, S. Russell, S. Somiari, H. Hu, R. Jordan, A. George, R. Katenhusen, A. Buchowiecka, C. Arciero, H. Brzeski, J. Hooke, C. Shriver, High-throughput proteomic analysis of human infiltrating ductal carcinoma of the breast, Proteomics 3 (2003) 1863 – 1873. E. Storey, R. Cappai, The amyloid precursor protein of Alzheimer’s disease and the Abeta peptide, Neuropathol. Appl. Neurobiol. 25 (1999) 81 – 97. R.N. Thrift, T.M. Forte, B.E. Cahoon, V.G. Shore, Characterization of lipoproteins produced by the human liver cell line, Hep G2, under defined conditions, J. Lipid Res. 27 (1986) 236 – 250. T. Tsuji, A. Shiozaki, R. Kohno, K. Yoshizato, S. Shimohama, Proteomic profiling and neurodegeneration in Alzheimer’s disease, Neurochem. Res. 27 (2002) 1245 – 1253. A.K. Vehmas, D.R. Borchelt, D.L. Price, D. McCarthy, M. WillsKarp, M.J. Peper, G. Rudow, J. Luyinbazi, L.T. Siew, J.C. Troncoso, Beta-Amyloid peptide vaccination results in marked changes in serum and brain Abeta levels in APPswe/PS1DeltaE9 mice, as detected by SELDI–TOF-based ProteinChip technology, DNA Cell Biol. 20 (2001) 713 – 721. A. Walduck, T. Rudel, T.F. Meyer, Proteomic and gene profiling approaches to study host responses to bacterial infection, Curr. Opin. Microbiol. 7 (2004) 33 – 38. G.L. Wright Jr., M.L. Beckett, K.R. Newhall, B.L. Adam, L.H. Cazares, S.L. Cartwright, Z. Xiao, L. Gong, P.F. Schellhammer, Identification of a superimmunoglobulin gene family member overexpressed in benign prostatic hyperplasia, Prostate 42 (2000) 230 – 238. Z. Xiang, L. Ho, S. Yemul, Z. Zhao, W. Qing, P. Pompl, K. Kelley, A. Dang, D. Teplow, G.M. Pasinetti, Cyclooxygenase-2 promotes amyloid plaque deposition in a mouse model of Alzheimer’s disease neuropathology, Gene Expression 10 (2002) 271 – 278. X. Zheng, J.A. Bobich, A sequential view of neurotransmitter release, Brain Res. Bull. 47 (1998) 117 – 128.

Sign up to vote on this title
UsefulNot useful