PWE-049 Shotgun proteomics in the discovery of biomarkers of non-alcoholic fatty liver disease: an improved methodology

Abstract

<h3>Introduction</h3> Non-alcoholic fatty liver disease is a common condition affecting up to one third of the US population. The current gold standard diagnostic tool is liver biopsy, which has associated morbidity and mortality concerns, as well as sampling error and inter-observer variability problems. There is an urgent need for the discovery and validation of non-invasive markers of this disease. Many studies have taken a targeted approach, examining the utility of various biochemical and clinical markers thought to be important in the pathogenesis of this condition. Shotgun proteomics offers an untargeted tool for the discovery of protein markers in complex clinical samples. <h3>Methods</h3> Materials: 30 patients with biopsy proven NAFLD were recruited. Patients had negative viral hepatitis screens, negative autoimmune screens, normal iron stores, and were taking no hepatotoxic medications. Biochemical and anthropological measurements were also recorded. <h3>Method</h3> Serum samples were pooled into three groups (SS/NASH/NASH fibrosis, <i>F</i>_1,2, with n=7 per group). The pooled samples were immunodepleted of the top 91 most abundant proteins. Buffer exchange and lyophilisation was performed before samples were reduced alkylated then digested with trypsin. The resultant peptide mixtures were labelled with iTRAQ tags as per the manufacturers9 protocol (Applied Biosystems), and then pooled together. Sample fraction was performed using a reverse phase C18 column at high pH. A total of 88 fractions were collected and combined to 47 fractions based on the appearance of the chromatogram. The 47 fractions were analysed by LC-MS/MS (QTOF, Agilent), using an optimised LC and MS methodology. The resultant MS data were processed using MASCOT Distiller software and the IPI_human database searched for matches. Only proteins with 2 peptide identification and a MASCOT score indicative of identity were included for analysis. <h3>Results</h3> A total of 557 proteins were identified and included for analysis. Previous iTRAQ technical replicate data indicated an error range of 1.13 (±0.10)–1.37 (±0.12). Proteins with fold changes out with this range were considered significantly differentially expressed. At least 10 proteins had a fold change &gt;5 in the NASH early fibrosis group. Interestingly, several protein markers of NAFLD fibrosis previously reported in the literature were also found to be appropriately unregulated in this experiment, including α-2-macroglobulin. <h3>Conclusion</h3> Discussion iTRAQ technology is a powerful tool, allowing analysis of up to 8 different samples simultaneously. We have improved our method of sample preparation and processing to increase the number of protein identifications. The findings from this study require further validation.