A diagnostic plasma protein panel for alcohol abuse

DESCRIPTION (provided by applicant): Alcohol abuse and alcoholism remain significant societal problems. Unfortunately, treatment for and monitoring of harmful alcohol consumption is hampered by the lack of a highly accurate diagnostic test of alcohol drinking behavior. Our long-term goal for this research project is to develop biomarkers of alcohol abuse for use in at- risk populations. In particular, we have identified highly sensitive and specific plasma protein biomarkers using a well-controlled nonhuman primate model of alcohol abuse. We propose, in this application to translate these findings from the nonhuman primate into potential diagnostic tools for monitoring alcohol abuse in humans. This will be undertaken in collaboration with three clinical sites around the world. Specifically, we will obtain de- identified clinical samples from he Coatesville Veterans Affairs Medical Center, the Yale University School of Medicine, and the Institute for Health and Welfare of Finland (Helsinki, Finland). Studies associated with this research program will be divided into two specific aims. Specific aim 1 will validate a novel plasma biomarker panel in recovering alcoholics. In this aim, our clinical collaborators (from Yale and the Coatesville VAMC) will provide anonymous, within-subject longitudinal samples from subjects undergoing withdrawal and early abstinence (21 to 28 days) from abusive drinking as well as from the general population. These well-annotated samples will be used to refine the list of 27 biomarkers to identify the set that is most diagnostic of the human drinking phenotype and withdrawal/ abstinence. These experiments will employ a multiplex quantitative solution-phase immunoassay designed for human analytes (Myriad RBM DiscoveryMAP v1.0). As with all biomarker discovery projects, our primary emphasis will be on identifying a plasma protein analyte panel with high sensitivity and specificity. Specific aim 2 will then refine the plasma biomarker panel to permit diagnosis of hazardous drinking in a cross-sectional population from multiple sites. Within- subject assessment of drinking behavior (Aim #1) is much easier because the test does not compare differing baseline physiologies. Requirements for general population screening are much more stringent and will be assessed in samples from three independent sites (Yale University, Coatesville PA Veterans Administration Medical Center and the Finland National Institute for Health and Welfare). These sites provide a rich number of samples and drinking behaviors from which to draw experimental cohorts. In summary, we believe this work is innovative in that it capitalizes on new concepts in biomarker development through the identification of novel targets. The specific aims will build on preceding successes in the nonhuman primate model by characterization of human samples. All of the proposed studies will involve correlation with normal clinical and psychosocial assessments of alcohol consumption (AUDIT self-reports; percentCDT, AST/ALT, GGT, etc.). In addition, because we are not using unitary biomarker measures, we will apply sophisticated biostatistical approaches (random forest, support vector machine, principle component analysis) to the data analysis.