Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Int J Neuropsychopharmacol. Author manuscript; available in PMC 2012 August 1.
Published in final edited form as:
PMCID: PMC3107900

Plasma proteomic alterations in non-human primates and humans after chronic alcohol self-administration


Objective diagnostics of excessive alcohol use are valuable tools in the identification and monitoring of subjects with alcohol use disorders. A number of potential biomarkers of alcohol intake have been proposed, but none have reached widespread clinical usage, often due to limited diagnostic sensitivity and specificity. In order to identify novel potential biomarkers, we performed proteomic biomarker target discovery in plasma samples from non-human primates that chronically self-administer high levels of ethanol. 2-dimensional in-gel electrophoresis (2D-DIGE) was used to quantify plasma proteins from within subject samples collected before exposure to ethanol and after three months of excessive ethanol self-administration. Highly abundant plasma proteins were depleted from plasma samples to increase proteomic coverage. Altered plasma levels of SAA4, RBP, ITIH4, clusterin, and fibronectin, identified by 2D-DIGE analysis, were confirmed in unmanipulated, whole plasma from these animals by immunoblotting. Examination of these target plasma proteins in human subjects with excessive alcohol consumption (and control subjects) revealed increased levels of SAA4 and clusterin and decreased levels of fibronectin compared to controls. These proteins not only serve as targets for further development as biomarker candidates or components of biomarker panels, but also add to the growing understanding of dysregulated immune function and lipoprotein metabolism with chronic, excessive alcohol consumption.

Keywords: Alcohol abuse, alcoholism, biomarker, diagnostic, plasma, proteomic


The personal and economic costs of alcohol abuse and alcoholism are well-documented (Harwood, 2000). Unfortunately, treatment for and monitoring of harmful alcohol consumption is hindered by the lack of widely accepted, highly accurate diagnostics of ethanol intake. Sustained efforts have been invested to create and refine interview formats that can accurately quantify alcohol intake, such as AUDIT-C (Alcohol Use Disorders Identification Test Consumption) (Bradley et al., 2007; Bush et al., 1998), or CAGE (Ewing, 1984), but these approaches have limitations (Kip et al., 2008). This is especially true in cases where individuals are motivated to deny or minimize the magnitude of their drinking behavior in order to mitigate personal, professional, or legal ramifications of alcohol abuse (Pernanen, 1974). Reports have also documented variations in self-report accuracy between populations (Dhalla and Kopec, 2007; Frank et al., 2008;) and depending on the manner and setting of the interview (Steinweg and Worth, 1993). One difficulty in quantifying the validity of alcohol self-reports is that, unlike other abused substances, alcohol abuse lacks a clinical test to corroborate self-report data (Brener et al., 2003).

Biological tests–biomarkers–of alcohol intake for use in diagnosing and monitoring alcohol consumption in at-risk populations and during treatment offer the potential for more accurate indices of alcohol intake (Freeman and Vrana, 2010; Hannuksela et al., 2007). A number of accurate methods for determining acute blood alcohol content (BAC) through breath and blood now exist. What has remained elusive are diagnostics that can retrospectively examine alcohol intake over a longer window of time and provide insight into long-term drinking behavior.

A number of biochemical markers of alcoholism have been proposed (Helander, 2003), but these assays have limited accuracy and sensitivity (Alte et al., 2003; Anton et al., 2002; Conigrave et al., 2002). A number of potential plasma-, blood-, urine-, and hair-based biomarkers have been identified and extensively reviewed elsewhere (Conigrave et al., 2002, 2003; Das et al., 2008; Hannuksela et al., 2007; Helander, 2003). Proposed biomarkers of alcohol intake include: gamma-glutamyltransferase (GGT) (Taracha et al., 2001), mean corpuscular volume (MCV) (Hock et al., 2005), aspartate aminotransferase (AST; also known as SGOT, serum glutamate oxaloacetate transaminase) alanine aminotransferase (ALT) (Niemela, 2007), sialylation of apolipoprotein J (SIJ) (Ghosh et al., 2001), carbohydrate-deficient transferrin (CDT) (Golka and Wiese, 2004; Koch et al., 2004), ethyl glucuronide (EtG) (Kissack et al., 2008), and 5-hydroxytryptophol (5HTOL) (Helander and Eriksson, 2002). Indeed, while not perfect, CDT has been approved by the FDA as a biomarker of heavy alcohol consumption (in combination with self-reports or other clinical evidence). Limitations in the sensitivity and specificity of these biomarkers have led to efforts seeking to combine multiple analytes into biomarker panels (Anton et al., 2002; Korzec et al., 2005, 2009) and discovery efforts to identify new potential biomarkers through genomic and proteomic screening.

Recently, we reported a panel of 17 plasma proteins that classify non-drinking, moderate drinking, and excessive drinking with high accuracy in a non-human primate self-administration model of alcohol consumption (Freeman et al., 2010). These plasma proteins were identified through a directed proteomic approach in which 90 known plasma cytokines, growth factors, and other proteins were examined by immunological techniques. To expand upon these findings, the current study performed open discovery analysis of the plasma proteome using longitudinal samples collected from non-human primates self-administering ethanol. The purpose of this study was to identify additional plasma proteins that could be combined with those we have previously determined.

Non-human primates share similar alcohol-related pathology to humans (Grant and Bennett, 2003), but can be studied in a controlled environment and with the capacity for longitudinal within-subject examination. Monkeys also display alcohol absorption and metabolism pharmacokinetics similar to human beings (Green et al., 1999) and display commonalities with changes in human plasma proteins (Freeman et al. 2006a, 2010). These qualities make non-human primates a valuable model system for biomarker discovery. This study used a non-human primate model of chronic ethanol exposure in which animals self-administer large quantities of ethanol chronically over months or years (Grant et al., 2008a, 2008b; Ivester et al., 2007; Vivian et al., 2001).

In this study, we performed open proteomic discovery on plasma proteins to identify biomarker targets for development in biomarker panels as diagnostics of alcohol intake and to illuminate plasma protein changes that may reflect alcohol-induced pathology. Proteomic discovery used a within-subject design and non-human primate ethanol self-administration model. Validation experiments were performed in the same animals using independent techniques. Moreover, biomarker targets were also assessed in human alcoholics and healthy controls to evaluate the potential for clinical translation of these biomarker targets.


Non-human primates

Ten male cynomolgus monkeys (macaca fascicularis) were part of a 21-month experimental time line (Fig 1A). For the first year (Naïve samples), monkeys aged 5–6 years (average weight 4.5 kg), were acclimated to the study environment and operant instrumentation, and trained to present their leg for venipuncture from the saphenous and/or femoral veins without the use of an anesthetic agent. Plasma samples were collected for endocrine tests, monitoring of blood alcohol levels, and for biomarker discovery and validation. Monkeys were induced to consume liquids under a schedule of food pellet deliveries (i.e., schedule-induced polydipsia (Falk, 1961)) as described previously for these animals (Grant et al., 2008a). Induction conditions did not require food deprivation, just scheduling access, and were not associated with weight loss. Ethanol was presented in the form of 4% w/v ethanol (unflavored). Following one month of 0.5 g/kg/day ethanol, the animals were escalated to drink 1.0 g/kg/day for 30 consecutive days, and finally, 1.5 g/kg/day for 30 consecutive days. Animals were then given unlimited access (22 hours per day) to ethanol and water for six months (Drinking samples) (Grant et al., 2008a; Vivian et al., 2001). Plasma samples from Naïve and Drinking time points were collected for proteomic discovery and validation experiments. The study was conducted in accordance with the Wake Forest University Animal Care and Use Committee and the guidelines for the care and use of laboratory animal resources

Fig 1
Animal treatment paradigm and ethanol intakes. (A) Time course of non-human primate alcohol self-administration and plasma sampling points. Following one year of acclimatization (naïve period), monkeys were induced to drinking under schedule-induced ...

Hu-14 protein depletion

We have previously described methods for depletion of abundant proteins from non-human primate samples to improve the sensitivity of proteomic experiments (Freeman et al., 2006b). In the present study, an affinity column designed to capture the 14 most abundant plasma proteins (albumin, IgG, antitrypsin, IgA, transferrin, haptoglobin, fibrinogen, alpha2-macroglobulin, alpha 1-acid glycoprotein, IgM, apolipoprotein AI, apolipoprotein AII, complement C3, and transthyretin) (Hu-14, Agilent) was used (Zolotarjova et al., 2008). To ensure required protein yields, samples were depleted in duplicate. Plasma samples were passed through a 0.22μm spin filter and 40μL of plasma sample was injected onto a 4.6mm internal diameter Hu-14 column at room temperature (Aktaprime Plus, GE Healthcare). The flow-through fraction was collected and, after the addition of high salt buffer, the bound proteins were collected. Fractions were monitored in-line by UV absorbance. Samples were concentrated before further use with a 5 kDa MW cut-off spin concentrator (Vivascience).


Quantitative two-dimensional in-gel electrophoresis (2D-DIGE) was performed as detailed previously (Umstead et al., 2010; VanGuilder et al., 2010a). Protein from depleted plasma was purified of contaminants by precipitation (2-D Cleanup, GE Healthcare), resuspended in sample buffer (Tris-HCl, 2M thiourea, 7M urea, 2.0% CHAPS, 1.0% ASB-14, pH 8.5), and quantified by 2D-Quant protein assay (GE Healthcare). Individual samples (n=10/group) were adjusted to pH 8.0-9.0 and 50 μg of each sample was labeled with appropriate cyanine (Cy) dyes (GE Healthcare). A normalization pool containing equal protein from each sample, labeled with Cy2, was included on each gel to standardize quantitation. Unlabeled normalization pool protein (250μg) was used for preparative/picking gels for mass spectrometry (MS) identification of proteins. Labeled samples were combined (one Naïve control, one Drinking, and one normalization pool aliquot per analytical gel) for a total of ten gels, mixed with equal volumes of 2X sample buffer (7M urea, 2M thiourea, 2.0% CHAPS, 2.0% pH 3-10NL pharmalyte, 1.0% DeStreak reagent), and brought to a volume of 450μL with DeStreak rehydration buffer containing 0.5% 3–10 NL pharmalyte (GE Healthcare). Samples were resolved by isoelectric point on 24-cm, pH 3–10 nonlinear strip gels (GE Healthcare). Strip gels were then equilibrated to SDS and reduced and alkylated. SDS-PAGE second dimension separation was performed on 10–14% polyacrylamide gradient gels poured using an automated pump system (a2DE optimizer, NextGen Sciences, Ann Arbor MI). Picking gels were fixed and post-stained with Deep Purple (GE Healthcare). Gels were imaged (Typhoon 9410 scanner, GE Healthcare) with identical, optimized PMT voltages used for all gels for each channel. Automated analysis of gel images was performed using DeCyder 6.5 software (GE Healthcare) to detect spots and calculate relative expression values. Only spots confidently matched across at least 8 of 10 quantitative gels were included in statistical analyses. Differential expression was determined by two-tailed paired t-test with false discovery rate correction (p<0.05).

MALDI-ToF/ToF mass spectrometry

MALDI-ToF/ToF mass spectrometry was performed as described previously (VanGuilder et al., 2010a). Spots were excised using a robotic Ettan Spot Picker (GE Healthcare), trypsin digested, and desalted (ZipTip, Millipore, Bedford, MA). Samples were spotted onto a MALDI target followed by 0.8μL of ACH-cinnamic acid. MS Measurements were taken in the positive ion mode between 800–4000 m/z with a signal-to-noise filter of 10, mass exclusion tolerance of 0.2Da, and a peak density filter of 50 peaks per 200Da (Applied Biosystems 4800). The 10 most intense ions with a minimum signal to noise of 75 that were not included on an exclusion list containing trypsin autolysis, matrix, and tryptic peptides of human keratin were subjected to MS/MS. MS/MS was performed without collision-induced decay in a mass range from 60Da to 20Da below the precursor mass with a fragment tolerance of 0.2Da for +1 charged ions. Protein identifications were made using GPS explorer v3.6 with Mascot v2.0.00 and the primate taxonomy of the NCBI database downloaded on Feb 16 2008 (107,758 entries). Identifications required a confidence interval of greater than 99%. When multiple isoforms of a protein were identified in the same MALDI target spot, the isoform with the highest confidence interval was reported which either: 1) had an identified peptide exclusive to that isoform; or 2) had the theoretical molecular weight and pI that matched the gel spot location.

LC-MS/MS mass spectrometry

Tryptic peptides extracted from ten gel plugs, matching to the ten largest magnitude and statistically-significant changes, were analyzed by on-line nano LC/MS/MS and ThermoFisher LTQ Orbitrap XL (NextGen Sciences, Ann Arbor, MI) as previously described (VanGuilder et al., 2010a). This additional analysis was performed to maximize identification success for these protein spots of interest. 30 μL of tryptic peptide solution was loaded on a 75 mm C12 vented column at a flow-rate of 10 μL/min and eluted at 300 nL/min with the following one hour gradient (time in minutes:% elution buffer): 0.1 min:3%, 30 min:23%, 38 min:32%, 42 min:50%, 44 min:95%, 47 min:95%, 47.5 min:1.0%, 55 min: 1.0%. MS/MS ion searches were conducted with the following specifications: monoisotopic mass values; peptide mass accuracy of ±2.0Da; fragment mass tolerance of ±0.5Da; one missed trypsin cleavage; fixed carbamidomethyl (C); variable modifications: oxidation (M), acetyl (N-term), pyro-glu (N-term-Q). MASCOT-generated data files were processed using the Scaffold algorithm. Parameters for LTQ-based protein identification required a minimum of three peptide matches per protein with minimum probabilities of 95% at the protein level and >50% at the corresponding peptide level.

Human subjects

To assess biomarker targets in human samples, cross-sectional samples were collected under Institutional Review Board approved protocol with informed consent at the Yale School of Medicine NIAAA Center from medically healthy (including screening for Hepatitis C) control subjects and patients who met DSM-IV diagnostic criteria for alcohol dependence (American Psychiatric Association, 2000). Both groups may have met diagnostic criteria for nicotine dependence, but they were excluded if structured diagnostic interview (American Psychiatric Association, 2000; First et al., 2010; Spitzer and Williams, 2002) found evidence of Axis I psychiatric diagnoses or substance abuse dependencies. Both groups also were free of psychotropic substances for at least four weeks prior to study entry (excluding nicotine, ethanol, and caffeine). Clinical history was supported by urine toxicology screening. Time-line-follow-back was used to estimate the quantity of alcohol consumed in the 30 days prior to sample collection (Sobell and Sobell, 1992).


Whole, unmanipulated plasma samples from monkeys and human subjects were used to eliminate the potential of any artifacts introduced by affinity depletion. Paired samples from one monkey were not included in immunoblot experiments as the samples for this animal and time point were exhausted in the 2D-DIGE study. Samples were diluted 1:10 with protein lysis buffer (100mM NaCl, 20mM HEPES, 1mM EDTA, 1mM dithiothreitol, 1.0% Tween20, 1mM Na3VO4, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail Tablet for every 10mL lysis buffer), and filtered with 0.22μm spin filters to remove particulate matter. Protein concentrations were determined by BCA protein assay (Pierce), and all samples were adjusted to equal protein concentration with additional lysis buffer. Equal amounts of protein from all samples were resolved on Criterion 4-20% acrylamide gradient Tris-HCl gels (Bio-Rad) and transferred to HyBond PVDF membranes (GE Healthcare). After blocking in 5% fat free milk (mouse primary antibodies) or 3% BSA (goat primary antibody) in phosphate-buffered saline with 1% Tween-20, membranes were probed with primary antibodies (mouse anti-RBP, Santa Cruz [#sc46688]; goat anti-ITIH4, Santa Cruz [#sc21987]; mouse anti-albumin, Sigma [#A6684]; mouse anti-fibronectin, Sigma [#F0916]; mouse anti-SAA4, Sigma [#sab1400251]; mouse anti-clusterin, Millipore [#05-354]) diluted in new blocking buffer, by overnight incubation at 4°C. After washing, blots were incubated with species-appropriate HRP-conjugated secondary antibodies, developed with ECL substrate (GE Healthcare), imaged on x-ray film, and digitized at a resolution of 600 dpi for subsequent quantitation. Images were quantified using ImageQuant TL 2007 software with rolling ball background subtraction.


Immunoblotting data from longitudinal non-human primate samples were analyzed with a paired two-tailed t-test, or a Wilcoxon signed rank test for non-normally distributed data. Human subject samples were analyzed by unpaired two-tailed t-test, or a Mann-Whitney rank sum test for non-normally distributed data for cross-sectional human samples (α<0.05).


Non-human primate ethanol self-administration

For non-human primate within-subject biomarker target discovery and confirmation experiments, alcohol Naïve samples and samples collected after 3 months of excessive alcohol consumption (Drinking) were used (Fig 1a). At collection of Drinking samples, monkeys had 22/hr/day access to ethanol. Cumulative intake over the experimental timeline at the time of plasma collection averaged 495g/kg (Fig 1b). In the two weeks prior to sampling, average daily ethanol consumption was 2.89g/kg (Fig 1c)–a value equivalent to approximately 12 drink equivalents per day.

Abundant plasma protein depletion

Approximately 95–98% of plasma protein content is comprised of only a few highly abundant (12–14) proteins that can mask detection and quantitation of less abundant and potentially informative proteins (Anderson and Anderson, 2002). We have previously described depletion of abundant proteins by affinity chromatography in non-human primate samples (Freeman et al., 2006b). To characterize the consistency and specificity of the 14 protein affinity depletion column used in this study with non-human primate samples, a series of quality control experiments were performed. In-line monitoring of plasma depletion demonstrated consistent flow-through and protein binding patterns across all 40 depletions (Fig 2a). Quantitation of the flow through and bound fractions demonstrated 95% of plasma protein bound to the column with 5% of the protein found in the flow-through. No differences in flow-through fraction yields were evident between Naïve and Drinking samples. Quality control assessment of equal amounts of protein from undepleted whole plasma, bound fraction, and flow through separated by SDS-PAGE demonstrated an expected change in the pattern of proteins observed with the most prominent bands in the whole plasma present in the bound fraction but not in the flow-through (Fig 2b). Immunoblotting for proteins to be removed by the depletion column (albumin and α2 macroglobulin) and for a negative control that should not be removed by the column (ceruloplasmin) revealed depletion of intended proteins, but no non-specific depletion (Fig 2c). Two-dimensional electrophoresis of equal protein masses of whole plasma, bound fraction, and flow-through visualized the removal of abundant proteins and the detection of additional proteins apparent in the flow-through that were masked in the whole plasma (Fig 2d).

Fig 2
Reproducibility of plasma depletions. To improve the sensitivity for examining rarer proteins, abundant plasma proteins were removed by affinity chromatography. (A) Chromatograms from forty independent depletions were co-plotted onto the same graph. Minimal ...

2D-DIGE and Mass Spectrometry

Labeled proteins separated by isoelectric point and molecular weight produced consistent spot patterns on analytical and picking gels. A total of 722 protein spots were matched across 8 or more of the analytical gels. Paired, two-tailed t-tests determined that 190 protein spots (~25% of total matched spots) were differentially regulated in drinking vs. naïve subjects (paired t-test with false discovery rate multiple testing correction; p<0.05, 1.15-fold change cutoff). 106 protein spots were upregulated by 1.15- to 10-fold with drinking, while 84 protein spots were downregulated by 1.15- to 17-fold. 199 protein spots, representing both ethanol-regulated and stably-expressed species, that were matched to a picking gel were identified with >99% confidence by either MALDI- or LTQ-mass spectrometry (Supplementary Table S1). 46 differentially-regulated plasma proteins were confidently identified (e.g., clusterin, inter-alpha inhibitor H4, serum amyloid A4) (Fig 3). In agreement with our previous examination of abundant protein depletion in monkey plasma samples using an anti-human antigen affinity column (Freeman et al., 2006b), depletion efficacy was very high but not 100% as evidenced by identification of apoliprotein AI , C3, and alpha 1-acid glycoprotein by mass spectrometric analysis. This is most likely due to amino acid sequence differences in these proteins between human and monkey (Freeman et al., 2006b).

Fig 3
2D-DIGE results. Abundant protein depleted plasma was subjected to 2D-DIGE quantitative analysis to identify proteins with differing abundance between Naïve and Drinking group. (A) An example 2D gel image is shown. A total of 190 spots were significant ...

Immunoblot confirmation

To confirm protein expression changes with alcohol intake, selected proteins with the largest magnitude differences between Naïve and Drinking time points and identified by 2D-DIGE/MS were quantitated in whole monkey plasma by immunoblotting. Whole plasma was used in confirmation experiments to eliminate any quantitation bias introduced by plasma depletion and to provide a matrix (whole plasma) most similar to that which would be used clinically. Undepleted naïve and drinking samples were probed for SAA4, RBP4, ITIH4, clusterin, and fibronectin, with albumin included as an unchanged loading control. Total protein staining (Deep Purple and Ponceau) of gels also demonstrated highly consistent protein quantitation and loading (Supplementary Fig 1). In agreement with the proteomic quantiation, samples collected from drinking animals contained significantly more SAA4 (229±25.0% of naïve, t(18)=−11.11, p<0.001), RBP4 (120±9.3% of naïve, t(18)=−3.93, p<0.01), and ITIH4 (844±251.7% of naïve, z(18)=2.55, p<0.01) compared to their paired ethanol-naïve samples (Fig 4.0. Decreased expression of clusterin (82±16.9% of naïve, t(16)=2.55, p<0.05) and fibronectin (52±8.2% of naïve, t(18)=5.13, p<0.001) was also observed by immunoblotting (Fig. 4). Consistent changes were observed between paired Naïve and Drinking samples with some overlap between conditions (Supplementary Fig 2).

Fig 4
Orthogonal confirmation of 2D-DIGE findings. To confirm biomarker targets identified in the 2D-DIGE analysis, specific immunoblotting was performed using undepleted plasma samples. Significantly higher protein concentrations of SAA4, RBP, and ITIH4 were ...

Human subject validation

To test the potential translation of these findings to human subjects, plasma samples from healthy and heavily drinking patients were immunoblotted for protein targets identified in monkey samples. Plasma samples from eight alcohol dependent men (average age 41±7.5 years) and seven healthy male subjects (average age 35±9.8 years) were collected. In the month prior to sample collection, alcohol abusing subjects averaged 402±216 standard drinks consumed, as estimated by time-line-follow-back. Healthy subjects reported 4±4 standard drinks consumed in the previous month. Decreased fibronectin (75±5.7% of healthy, t(15)=3.42, p<0.01) and increased SAA4 (204±38.4% of naïve, U(15)=10, p<0.05) were confirmed in humans (Fig. 5). Clusterin levels in alcohol abusing patients were 121±5.2% of healthy controls, t(15)=−2.35, p<0.05). Despite the significant differences between group means, some overlap in protein abundance was observed between Healthy and Drinking individuals (Supplementary Fig 3). No differences in ITIH4 and RBP4 protein abundance were observed in human subjects (data not shown).

Fig 5
Assessment of biomarker targets in human subjects. To translate the findings of the differential plasma protein abundance in the non-human primate to humans, biomarker targets were assessed by immunoblotting in alcoholics and healthy controls. * p<0.05, ...


The data from this study demonstrate novel changes in serum amyloid A4, retinol binding protein, inter-alpha inhibitor H4, clusterin, and fibronectin plasma protein abundance with chronic, excessive alcohol self-administration in a non-human primate model. In a translational assessment, decreased plasma levels of fibronectin and increased levels of serum amyloid A4 were also evident in human subjects with excessive alcohol intake compared to controls. These data also provide evidence for the utility of combining plasma protein depletion technology and 2D-DIGE quantitation to identify biomarker candidates. In agreement with our recent demonstration that directed antibody-based (Luminex) and open discovery gel-based (2D-DIGE) proteomic technologies provide complementary rather than redundant data (VanGuilder et al., 2010b) the proteins identified and confirmed in this study are novel and were not detected in our previous biomarker identification efforts (Freeman et al., 2010). These proteins not only provide insight into the health effects of chronic, excessive alcohol consumption, but also may serve as biomarkers of alcohol intake.

Serum amyloid A4 (SAA4) is expressed in a number of tissues, with the liver being a primary organ source. Similar elevations of plasma SAA4 were observed in monkeys and humans (200–250% of respective controls). SAA4 is an acute phase response protein and a major component of high density lipoprotein (HDL) (Bausserman et al., 1980). Circulating SAA4 levels are induced by a number of inflammatory factors (Akira et al., 1990). Increased circulating levels of serum amyloid A have been reported in mice after high dose, but not low dose, ethanol administration (Pruett and Pruett, 2006). In fact, this induction was highest at 24 hours after ethanol administration, providing both temporal and dose effect relationships that may aid in explaining the often contradictory evidence for ethanol as both an immunosuppressant and pro-inflammatory factor. Combined with the data from this study demonstrating elevated SAA4 levels in non-human primates and humans with high levels of ethanol consumption, this provides evidence of commonalities in the plasma proteomic response to ethanol across species.

Fibronectin is a ubiquitously expressed acute phase response glycoprotein with a wide range of functions from cell adhesion to blood coagulation. Alterations in fibronectin tissue levels with alcohol have been reported with a heterogeneous response depending on the tissue and dose of alcohol. Increased myocardial levels have been reported in rodents after low dose alcohol administration (Vendemiale et al., 2001), while this protein is reported to increase in the lung with alcohol abuse (Burnham et al., 2007). Moreover, increased hepatic levels may be an indicator of future cirrhosis in alcohol abusers (Junge et al., 1988). However, there are no reports of circulating plasma fibronectin levels with excessive alcohol consumption. As fibronectin is produced in most major tissue types, including muscle and adipose, the decreased levels observed here may reflect global changes in a variety of organ systems.

The function of clusterin, also known as apolipoprotein J, is not known; however, sialylation levels of clusterin have been previously proposed as a biomarker of alcohol intake (Ghosh et al., 2001). In this study, total levels of clusterin protein, but not protein modifications, were examined. Decreased levels of clusterin were evident in non-human primate samples, but were only statistically significant through within-subject, paired analysis. On the other hand, increased (20%) levels of total clusterin protein were evident in human subjects with excessive alcohol intake. The differences observed between the non-human primate and human data potentially arise from the difference of within-subject verses cross sectional measurements and will certainly warrant further investigation in future studies.

Inter-alpha inhibitor H4 (ITIH4) is a serine protease inhibitor produced in the liver and has characteristics of an acute phase response protein (Pineiro et al., 1999). The increase in ITIH4 plasma levels illuminated by 2D-DIGE discovery was confirmed in non-human primate samples by immunoblotting, but did not translate to human subjects. This is not unexpected given that the human samples did not originate from a within-subject design. It is possible that some alcohol-responsive changes will become apparent only for within-subject assessments where individual differences in basal levels of expression can mask the effects of alcohol. While this reduces utility as a potential cross-sectional biomarker, applicability as part of within-subject biomarker panel remains. Additionally, it also contributes to the body of evidence suggesting changes in acute phase responses following ethanol self-administration.

Retinol binding protein 4 (RBP4) is the plasma isoform of the retinol binding protein family secreted by the liver and adipose tissue with the primary function of transporting retinol (vitamin A alcohol). Vitamin deficiencies in alcoholic patients is well described and is most likely due to both poor diet and impaired liver function (Hoyumpa, 1986). While there is growing evidence for a role of RBP4 expression in obesity-induced diabetes (Yang et al., 2005), no previous data regarding circulating levels of RBP4 with excessive alcohol intake is present in the literature. However, a relationship between vitamin homeostasis and regulation of lipid levels with excessive alcohol intake has been observed in this non-human primate model and may integrate the roles of vitamin status, lipoprotein levels, and inflammatory processes in alcohol related pathology (Lebold et al., 2010). Like ITIH4, the changes in RBP4 levels observed in monkeys were not evident in the human subjects examined. Evaluation of RBP4, and the other plasma proteins, will need to be performed in larger subject populations to extend these findings.

This study expands upon our previous focused proteomic approach to biomarker discovery (Freeman et al., 2010). The proteins examined with the 2D-DIGE technology did not overlap with those we have examined by antibody-based approaches, demonstrating a complementarity of these different proteomic approaches similar to that we have observed previously (VanGuilder et al., 2010b). When the results of the present study are combined with previous data on plasma proteomic changes from this model (Freeman et al., 2010), disruption of inflammatory homeostasis is evident, in agreement with a number of previous studies (Crews et al. 2006). Further study is needed in both animal models and simpler in vitro model systems to determine the interactions of classical inflammatory processes with lipoprotein metabolism, endocrine function, and nutritional status. Non-human primate self-administration models provide a valuable approach to define alterations in plasma proteins with alcohol abuse in a controlled setting and with a high relevance to the human condition (Grant and Bennett, 2003).

As the primary goal of the present study was to identify novel biomarker targets, the validation of SAA4 and fibronectin as altered in human subjects indicates that they may have utility as analytes in a biomarker panel of alcohol intake. With some overlap in the distribution of protein abundance between Healthy and Drinking samples, the highest specificity will most likely be achieved adding these candidates to panels of biomarkers rather than as independent metrics. Changes in plasma abundance for these and other previously described alcohol-responsive plasma proteins have been reported with other conditions and disease states. Through combining multiple biomarkers into panels, as have been previously proposed for alcohol intake biomarkers, may also aid to in achieving disease specificity (Anton et al., 2002; Freeman et al., 2010; Korzec et al., 2005, 2009). As well, determining alcohol-responsive changes in plasma proteins is important for understanding potential sources of false-positives in biomarker diagnostics of other diseases. Confounding variables such as inflammation and diabetes may impact the expression of these proteins, and must be considered when these biomarker candidates are implemented clinically as components of an alcohol consumption diagnostic test.

Further validation of these proteins as potential biomarkers will need to include, assessment in larger cross-sectional human studies, time-course examinations of abstinence from excessive alcohol consumption, and moderate drinking groups. Future development studies will aid in determining whether a longitudinal monitoring biomarker approach or cross-sectional diagnostic test achieves the highest sensitivity and specificity. Additionally, future studies with female subjects are required to examine commonalities and differences in the sex response to chronic excessive alcohol. Inclusion of the biomarker targets identified here with biomarker panel we have previously identified offers the potential to avoid issues of sensitivity and specificity associated with unitary biomarker diagnostics and ultimately increase the utility of these proteins as indices of alcohol use and abuse (Freeman and Vrana, 2010).

Supplementary Material

Supplemental Figure 1

Supplemental Figure 2

Supplemental Figure 3

Supplmental Figure Legends

Supplmental Table 1


This study was supported by grants from the US National Institutes of Health (AA016613 to K.E.V.; AA11997, AA13510 and AA13641 to K.A.G; and AA011321, AA14906, and AA012870, AA012870, AA011321 to JHK). JHK is also supported by the VA Alcohol Research Center. The authors wish to thank Pamela Noto for technical assistance on the proteomic studies and Steve Gonzales for assistance with the non-human primate studies.


Statement of Interest:

WMF, KEV, and KAG report a pending patent application on diagnostics of alcohol intake that includes some of the proteins described in this work. All other authors declare they have no competing financial interests. JHK is on the Board of Directors of the Lohocla Research Corporation and he has provided scientific consultation to Gilead Pharmaceuticals and Eli Lilly and Company in the area of alcoholism research. He has one patent and two pending patents related to pharmacotherapies for other psychiatric disorders. He also consults to companies related to other areas of psychiatric drug development.


  • Akira S, Hirano T, Taga T, Kishimoto T. Biology of multifunctional cytokines: IL 6 and related molecules (IL 1 and TNF) FASEB Journal. 1990;4:2860–2867. [PubMed]
  • Alte D, Ludemann J, Piek M, Adam C, et al. Distribution and dose response of laboratory markers to alcohol consumption in a general population: results of the study of health in Pomerania (SHIP) Journal of Studies on Alcohol and Drugs. 2003;64:75–82. [PubMed]
  • American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders DSM-IV-TR. Washington, D.C.: American Psychiatric Association; 2000.
  • Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Molecular & Cellular Proteomics. 2002;1:845–867. [PubMed]
  • Anton RF, Lieber C, Tabakoff B. Carbohydrate-deficient transferrin and gamma-glutamyltransferase for the detection and monitoring of alcohol use: results from a multisite study. Alcoholism: Clinical and Experimental Research. 2002;26:1215–1222. [PubMed]
  • Bausserman LL, Herbert PN, McAdam KP. Heterogeneity of human serum amyloid A proteins. Journal of Experimental Medicine. 1980;152:641–656. [PMC free article] [PubMed]
  • Bradley KA, DeBenedetti AF, Volk RJ, Williams EC, et al. AUDIT-C as a brief screen for alcohol misuse in primary care. Alcoholism: Clinical and Experimental Research. 2007;31:1208–1217. [PubMed]
  • Brener ND, Billy JO, Grady WR. Assessment of factors affecting the validity of self-reported health-risk behavior among adolescents: evidence from the scientific literature. Journal of Adolescent Health. 2003;33:436–457. [PubMed]
  • Burnham EL, Moss M, Ritzenthaler JD, Roman J. Increased fibronectin expression in lung in the setting of chronic alcohol abuse. Alcoholism: Clinical and Experimental Research. 2007;31:675–683. [PubMed]
  • Bush K, Kivlahan DR, McDonell MB, Fihn SD, et al. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Archives of Internal Medicine. 1998;158:1789–1795. [PubMed]
  • Conigrave KM, Davies P, Haber P, Whitfield JB. Traditional markers of excessive alcohol use. Addiction. 2003;98(Suppl 2):31–43. [PubMed]
  • Conigrave KM, Degenhardt LJ, Whitfield JB, Saunders JB, et al. CDT, GGT, and AST as markers of alcohol use: the WHO/ISBRA collaborative project. Alcoholism: Clinical and Experimental Research. 2002;26:332–339. [PubMed]
  • Crews FT, Bechara R, Brown LA, Guidot DM, et al. Cytokines and alcohol. Alcoholism: Clinical and Experimental Research. 2006;30:720–730. [PubMed]
  • Das SK, Dhanya L, Vasudevan DM. Biomarkers of alcoholism: an updated review. Scandinavian Journal of Clinical & Laboratory Investigation. 2008;68:81–92. [PubMed]
  • Dhalla S, Kopec JA. The CAGE questionnaire for alcohol misuse: a review of reliability and validity studies. Clinical and Investigative Medicine. 2007;30:33–41. [PubMed]
  • Ewing JA. Detecting alcoholism. The CAGE questionnaire. JAMA. 1984;252:1905–1907. [PubMed]
  • Falk D. Production of polydipsia in normal rats by an intermittent food schedule. Science. 1961;133:195–196. [PubMed]
  • First M, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV-TR Axis I Disorders (SCID-I/NP) New York: Biometrics Research Department, New York State Psychiatric Institute; 2010.
  • Frank D, DeBenedetti AF, Volk RJ, Williams EC, et al. Effectiveness of the AUDIT-C as a screening test for alcohol misuse in three race/ethnic groups. Journal of General Internal Medicine. 2008;23:781–787. [PMC free article] [PubMed]
  • Freeman WM, Gooch RS, Lull ME, Worst TJ, et al. Apo-aii is an elevated biomarker of chronic non-human primate ethanol self-administration. Alcohol and Alcoholism. 2006a;41 :300–305. [PubMed]
  • Freeman WM, Lull ME, Guilford MT, Vrana KE. Depletion of abundant proteins from non-human primate serum for biomarker studies. Proteomics. 2006b;6:3109–3113. [PubMed]
  • Freeman WM, Salzberg AC, Gonzales SW, Grant KA, et al. Classification of alcohol abuse by plasma protein biomarkers. Biological Psychiatry. 2010;68:219–222. [PMC free article] [PubMed]
  • Freeman WM, Vrana KE. Future prospects for biomarkers of alcohol consumption and alcohol-induced disorders. Alcoholism: Clinical and Experimental Research. 2010;34:946–954. [PMC free article] [PubMed]
  • Ghosh P, Hale EA, Lakshman MR. Plasma sialic-acid index of apolipoprotein J (SIJ): a new alcohol intake marker. Alcohol. 2001;25:173–179. [PubMed]
  • Golka K, Wiese A. Carbohydrate-deficient transferrin (CDT)--a biomarker for long-term alcohol consumption. Journal of Toxicology and Environmental Health Part B: Critical Reviews. 2004;7:319–337. [PubMed]
  • Grant KA, Bennett AJ. Advances in nonhuman primate alcohol abuse and alcoholism research. Pharmacology & Therapeutics. 2003;100:235–255. [PubMed]
  • Grant KA, Leng X, Green HL, Szeliga KT, et al. Drinking typography established by scheduled induction predicts chronic heavy drinking in a monkey model of ethanol self-administration. Alcoholism: Clinical and Experimental Research. 2008a;32:1824–1838. [PMC free article] [PubMed]
  • Grant KA, Stafford J, Thiede A, Kiley C, et al. Who's at risk? Population characterization of alcohol self-administration in nonhuman primates helps identify pathways to dependence. Alcohol Research and Health. 2008b:289–297. [PMC free article] [PubMed]
  • Green KL, Szeliga KT, Bowen CA, Kautz MA, et al. Comparison of ethanol metabolism in male and female cynomolgus macaques (Macaca fascicularis) Alcoholism: Clinical and Experimental Research. 1999;23:611–616. [PubMed]
  • Hannuksela ML, Liisanantti MK, Nissinen AE, Savolainen MJ. Biochemical markers of alcoholism. Clinical Chemistry and Laboratory Medicine. 2007;45:953–961. [PubMed]
  • Harwood H. Updating Estimates of the Economic Costs of Alcohol Abuse in the United States: Estimates, Update Methods, and Data. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism; 2000.
  • Helander A. Biological markers in alcoholism. Journal of Neural Transmission Supplement. 2003:15–32. [PubMed]
  • Helander A, Eriksson CJ. Laboratory tests for acute alcohol consumption: results of the WHO/ISBRA Study on State and Trait Markers of Alcohol Use and Dependence. Alcoholism: Clinical and Experimental Research. 2002;26:1070–1077. [PubMed]
  • Hock B, Schwarz M, Domke I, Grunert VP, et al. Validity of carbohydrate-deficient transferrin (%CDT), gamma-glutamyltransferase (gamma-GT) and mean corpuscular erythrocyte volume (MCV) as biomarkers for chronic alcohol abuse: a study in patients with alcohol dependence and liver disorders of non-alcoholic and alcoholic origin. Addiction. 2005;100:1477–1486. [PubMed]
  • Hoyumpa AM. Mechanisms of vitamin deficiencies in alcoholism. Alcoholism: Clinical and Experimental Research. 1986;10:573–581. [PubMed]
  • Ivester P, Roberts LJ, Young T, Stafforini D, et al. Ethanol self-administration and alterations in the livers of the cynomolgus monkey, Macaca fascicularis. Alcoholism: Clinical and Experimental Research. 2007;31:144–155. [PubMed]
  • Junge J, Bentsen KD, Christoffersen P, Orholm M, et al. Fibronectin as predictor of cirrhosis in men who abuse alcohol. British Medical Journal (Clinical Research Edition) 1988;296:1629–1630. [PMC free article] [PubMed]
  • Kip MJ, Spies CD, Neumann T, Nachbar Y, et al. The usefulness of direct ethanol metabolites in assessing alcohol intake in nonintoxicated male patients in an emergency room setting. Alcoholism: Clinical and Experimental Research. 2008;32:1284–1291. [PubMed]
  • Kissack JC, Bishop J, Roper AL. Ethylglucuronide as a biomarker for ethanol detection. Pharmacotherapy. 2008;28:769–781. [PubMed]
  • Koch H, Meerkerk GJ, Zaat JO, Ham MF, et al. Accuracy of carbohydrate-deficient transferrin in the detection of excessive alcohol consumption: a systematic review. Alcohol and Alcoholism. 2004;39:75–85. [PubMed]
  • Korzec A, de Bruijn C, van Lambalgen M. The Bayesian Alcoholism Test had better diagnostic properties for confirming diagnosis of hazardous and harmful alcohol use. Journal of Clinical Epidemiology. 2005;58:1024–1032. [PubMed]
  • Korzec S, Korzec A, Conigrave K, Gisolf J, et al. Validation of the Bayesian Alcoholism Test compared to single biomarkers in detecting harmful drinking. Alcohol and Alcoholism. 2009;44:398–402. [PMC free article] [PubMed]
  • Lebold KM, Grant KA, Freeman WM, Wiren KM, et al. Individual differences in hyperlipidemia and vitamin E status in response to chronic alcohol self-administratin in cynomolgous monkeys. Alcoholism: Clinical and Experimental Research. 2010 Epub ahead of print Nov 30. [PMC free article] [PubMed]
  • Niemela O. Biomarkers in alcoholism. Clinica Chimica Acta. 2007;377:39–49. [PubMed]
  • Pernanen K. Validity of survey data on alcohol use. In: Gibbons, et al., editors. Research Advances in Alcohol and Drug Problems. New York: 1974. pp. 355–374.
  • Pineiro M, Alava MA, Gonzalez-Ramon N, Osada J, et al. ITIH4 serum concentration increases during acute-phase processes in human patients and is up-regulated by interleukin-6 in hepatocarcinoma HepG2 cells. Biochemical and Biophysical Research Communications. 1999;263:224–229. [PubMed]
  • Pruett BS, Pruett SB. An explanation for the paradoxical induction and suppression of an acute phase response by ethanol. Alcohol. 2006;39:105–110. [PMC free article] [PubMed]
  • Sobell LC, Sobell MB. Timeline follow-back: a technique for assessing self-reported alcohol consumption. In: Litten R, Allen J, editors. Measuring Alcohol Consumption. New York: Humana Press; 1992. pp. 41–72.
  • Spitzer RL, Williams JBW. Structured Clinical Interview for the DSM-IV Axis I Disorders-Patient Edition (SCID-I/P, Version 2.0) New York: Biometrics Research Department, New York State Psychiatric Institute; 2002.
  • Steinweg DL, Worth H. Alcoholism: the keys to the CAGE. American Journal of Medicine. 1993;94:520–523. [PubMed]
  • Taracha E, Habrat B, Wozniak P, Walkowiak J, et al. The activity of beta-hexosaminidase (uHex) and gamma-glutamyl-transferase (uGGT) in urine as non-invasive markers of chronic alcohol abuse: I. Alcohol-dependent subjects. World Journal of Biological Psychiatry. 2001;2:184–189. [PubMed]
  • Umstead TM, Lu CJ, Freeman WM, Myers JL, et al. Dual-platform proteomics study of plasma biomarkers in pediatric patients undergoing cardiopulmonary bypass. Pediatric Research. 2010;67:641–649. [PubMed]
  • VanGuilder HD, Yan H, Farley JA, Sonntag WE, et al. Aging alters the expression of neurotransmission-regulating proteins in the hippocampal synaptoproteome. Journal of Neurochemistry. 2010a;113:1577–1588. [PMC free article] [PubMed]
  • VanGuilder HD, Bixler GV, Kutzler L, Brucklacher RM, et al. Multi-modal proteomic analysis of retinal protein expression alterations in a rat model of diabetic retinopathy. PLoS ONE. 2010b In press. [PMC free article] [PubMed]
  • Vendemiale G, Grattagliano I, Altomare E, Serviddio G, et al. Mitochondrial oxidative damage and myocardial fibrosis in rats chronically intoxicated with moderate doses of ethanol. Toxicology Letters. 2001;123:209–216. [PubMed]
  • Vivian JA, Green HL, Young JE, Majerksy LS, et al. Induction and maintenance of ethanol self-administration in cynomolgus monkeys (Macaca fascicularis): long-term characterization of sex and individual differences. Alcoholism: Clinical and Experimental Research. 2001;25:1087–1097. [PubMed]
  • Yang Q, Graham TE, Mody N, Preitner F, et al. Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature. 2005;436:356–362. [PubMed]
  • Zolotarjova N, Mrozinski P, Chen H, Martosella J. Combination of affinity depletion of abundant proteins and reversed-phase fractionation in proteomic analysis of human plasma/serum. Journal of Chromatography A. 2008;1189:332–338. [PubMed]