1. Abstract PROTEOMIC FINGERPRINTING IN HCV MONO- AND HIV-1/HCV CO-INFECTION REVEALS PLASMA BIOMARKERS PROGNOSTIC OF FIBROSIS Carlos Enrique Melendez-Pena* 1, Cynthia Santamaria 1, Brian Conway 2, Curtis Cooper 3, Bianca Segatto 1, Brian Ward 1, Momar Ndao 1, Marina Klein 1, and Canadian Co-infection Cohort Study group 1 McGill University Health Center, Montréal, Canada~ 2 Anesthesiology, Pharmacology and Therapeutics University of British Columbia, Vancouver, Canada~ and 3 Division of Infectious Diseases University of Ottawa at The Ottawa Hospital, Ottawa, Canada Background: Reliable biomarkers that can distinguish fibrosis states are essential for studying the natural history and for clinical management of HCV and HCV/HIV co-infection as serial liver biopsies are impractical on a population scale. We evaluated plasma proteome profiles in HCV mono and HIV/HCV co-infected patients using Surface-enhanced laser desorption (SELDI)-time-of-flight (TOF) mass spectrometry (MS), to identify novel plasma biomarkers capable of identifying different stages of fibrosis. Methods: 151 individuals were prospectively recruited (68 HCV mono-infected and 84 co-infected). Fibrosis stage (Batt and Ludwig 0-1, 2, 3, 4) was determined by liver biopsy and only plasma collected within one year of biopsy were profiled. Plasma were fractionated, randomly applied to ProteinChip arrays (IMAC, CM10 and H50) and spectra were generated at low and high laser intensities. Results: The majority of individuals were Caucasian men between the ages of 20 to 62 years old. For HIV co- infected, the median CD4 cell count was 640 (136-900) cell/mm 3, HIV RNA was 49 (24- 29 973) copies/ml and 89.7 % were receiving antiretroviral therapy. There was no statistically significant difference between the HCV mono and HIV/HCV co-infected groups in regards to their age, duration of HCV infection, alcohol use or smoking. After manual peak re-labeling, 16 biomarkers achieved a p-value 0.75 or 0.05). MALDI-TOF MS/MS and immunologic methods were used to profile 9 of these biomarker peaks. Thus far Serotransferin, Ig Heavy Chain, Haptoglobin have been identified by MALDI. The remaining proteins require confirmation but the leading candidates are: Apolipoprotein A1 and AII, Fibronectin, POTEE, Histidine-rich glycoprotein and complement C3. Conclusion: We obtained several potential biomarkers peaks useful for staging liver fibrosis in HIV- HCV co-infected persons. Identification of such protein biomarkers may reduce the need for liver biopsy, facilitate follow-up and the timing of HCV treatment, and help in the understanding of the impact of HIV and its treatment on liver disease in the setting of HCV infection. Figure 1: Decision tree to differentiate non fibrotic individuals(F1) from individuals with significant fibrosis in HCV mono-infection. Biomarker pattern based on CART analysis was used to generate candidate diagnostic algorithms. In this algorithm, the intensities of the 4.1-, 24.8-, 100- and 133-Kda biomarkers establish the splitting rules. This decision tree wasn’t able to correctly diagnose the HIV/HCV co-infection samples (71,69% sensitivity and 46.6 % specificity). Table 2: SELDI-TOF Spectra in HCV mono-infection. Mass (m/z), mean signal intensities, and AUC for selected individual differentially expressed peptides/proteins between fibrosis 0-1 (F1), fibrosis 2 (F2), fibrosis 3 (F3) and ESLD or fibrosis 4 (F4) patients. Table 3: SELDI-TOF Spectra in HIV/HCV co-infection. Mass (m/z), mean signal intensities, and AUC for selected individual differentially expressed peptides/proteins between fibrosis 0-1 (F1), fibrosis 2 (F2), fibrosis 3 (F3) and ESLD or fibrosis 4 (F4) patients. Biomarkers with * are also found in HCV mono-infection. Table 4: Sample identification of selected candidate biomarkers. Plasma fractionation were run on SDS-Page. Selected bands were cut with their negative counterparts. These biomarkers were identified using matrix-assisted laser desorption/ionisation-time-of-flight (MALDI-TOF) mass spectrometry. Identification was based on peptide present in the positive sample and absent in negative sample.  Using SELDI-TOF we observe 14 peaks for mono-infection and 16 for co-infection in which 5 are for both.  Using the BPS software we obtained a decision tree that helped us to distinguish between healthy patients and individuals with significant fibrosis with high insensitivity and good specificity..  We weren’t able to obtain a decision tree that was able to distinguish between the 4 stages of fibrosis.  The decision tree for mono-infection is not valid for co-infection and vice-versa.  Some of the biomarkers that we identified are already published in the literature. We are currently trying to identify the remaining biomarkers and confirm them by immune assays.  Study subjects had evidence of replicating HCV (RT-PCR RNA+). HCV mono-infected individuals were HIV sero-negative (determine by a negative ELISA). HIV/HCV co-infected individuals were HIV sero- positive (determine by positive ELISA with confirmatory Western blot).  All binding and washing steps were performed using a Biomek2000 robot (Beckman Coulter) extended by an integrated microplates shaker (MicroMix 5; Diagnostic Products Company).  The plasma samples collected within one year of biopsy, were fractionated by PH, using a ProteinChip serum fractionation kit (Bio-Rad), fraction 1,3 and 6 were analysed.  Arrays were analyzed in a ProteinChip biology system reader (series 4000) equipped with an autoloader using ProteinChip software, version 3.5 (Bio-Rad). Study population (n=151) HCV mono (n=68) Co-infection (n=83) P value Age**49.59± 0.94145.39± 0.879<0.01 HCV*20.75 ± 1.81216.53± 1.128<0.05 Sex* men69%82% <0.05 women31%16% transgender0%2% Ethnicity caucasian82% 1.0000 other18% Alcohol88%95%0.17 Last 6 month**38%63%<0.01 Smoking92%81%0.06 Last 6 month57%69%0.16 Table 1: Characteristics according viral co-infection status. m/z (/1,000) Fractions and Chemistries P value ( F1 vs F3-4) AUC for ROC Curve (fold F1/F3-4) Mean signal intensity ± SE P value F1 (n=20)F2 (n=20)F3 (n=20)F4 (n=8) 2.2F6 CM100.002,0.73(1.65)6.45±2.547.19±2.689.05±3.0112.20±2.610.008 4.6F3 IMAC300.006,0.31(-2.04)4.05±2.012.68±1.642.44±1.561.09±1.360.010 6.4F3 H500.007,0.27(-2.14)6.42±2.533.75±1.943.97±1.991.61±1.740.002 8.6F3 IMAC,F6 CM100.007,0.30(-1.71)3.07±1.752.97±1.722.04±1.431.51±1.010.006 9.4 F6 H50, F3 IMAC30- CM10 0.002,0.29(-2.55)11.33±3.3710.12±3.185.84±2.422.50±1.680.002 12.4F1 CM100.009,0.28(-2.13)0.29±0.540.46±0.670.11±0.330.17±0.460.021 13.8F6 H500.004,0.30(0.75)1.80±1.341.39±1.181.45±1.201.14±0.960.004 14.4F1 H500.006,0.28(-1.60)1.42±1.191.01±1.001.01±1.000.66±0.830.007 22.8*F6-F3 CM100.010,0.67(1.65)0.60±0.780.82±0.910.74±0.861.28±1.110.014 24.2*F1 CM100.018,0.71(1.74)0.58±0.760.85±0.920.87±0.931.13±0.850.071 33.3*F3 CM10-IMAC300.004,0.71(1.32)3.21±1.793.91±1.984.21±2.054.54±1.200.015 66.4*F3 IMAC300.003,0.71(1.16)23.88±4.8924.60±4.9626.75±5.1729.25±1.920.002 78.8F6 IMAC300.010,0.69(1.54)0.45±0.670.66±0.810.61±0.780.76±0.640.030 133*F3 IMAC300.006,0.71(1.14)2.32±1.522.38±1.542.69±1.642.67±0.740.015 162.3F6 IMAC300.004,0.71(1.82)0.03±0.180.05±0.220.05±0.210.07±0.200.005 177.9F1 IMAC300.009,0.72(2.45)0.01±0.070.02±0.150.01±0.110.01±0.140.100 SensitivitySpecificity 81.25%90,00% m/z (/1,000) Fractions and Chemistries P value ( F1 vs F3-4) AUC for ROC Curve (fold F1/F3-4) Mean signal intensity ± SE P value F1 (n=20)F2 (n=20)F3 (n=20)F4 (n=8) 2.5F1 IMAC300.003,0.76(2.80)2.37±1.542.76±1.667.17±2.685.30±2.510.008 4.1 F1 CM10, F3-F6 H50 0.001,0.91(2.34)3.21±1.795.90±2.437.10±2.667.50±1.320.005 4.5 F1,F3 CM10- IMAC30 0.001,0.22(-2.40)10.10±3.186.04±2.464.45±2.113.58±1.570.014 9.2 F3 IMAC30, F3 CM10 0.002,0.23(-2.33)11.72±3.427.92±2.815.14±2.274.78±2.550.017 18.4F1 CM100.011,0.30(-5.86)0.54±0.740.04±0.210.12±0.340.01±0.090.052 22.8F3 CM100.016,0.70(1.41)0.76±0.870.83±0.910.97±0.991.30±0.800.043 24.2F1 CM100.048,0.68(2.47)0.51±0.710.73±0.861.03±1.021.98±0.990.013 27.6F1 CM100.013,0.30(-2.31)0.12±0.350.04±0.210.05±0.230.05±0.260.025 33.3F3 IMAC300.013,0.70(1.15)5.31±2.305.35±2.316.02±2.456.23±1.040.067 46.8F1 CM100.004,0.74(1.75)0.41±0.640.49±0.700.64±0.800.94±0.630.022 66.4F3 IMAC300.001,0.75(1.16)24.62±4.9626.05±5.1028.36±5.3329.34±1.960.015 84.6F6 IMAC300.003,0.76(-1.63)0.10±0.320.18±0.420.18±0.420.14±0.230.003 100.2F3 IMAC300.004,0.77(1.51)0.16±0.410.21±0.460.24±0.490.26±0.290.036 133F3 IMAC300.000,0.79(1.21)2.34±1.532.68±1.642.79±1.672.91±0.510.003 Figure 2: Decision tree to differentiate non fibrotic individuals(F1) from individuals with significant fibrosis (F2 and above) in HIV/HCV mono-infection. Biomarker pattern based on CART analysis was used to generate candidate diagnostic algorithms. In this algorithm, the intensities of the 8,2-, 8.8-, 13.8- and 22.8-Kda biomarkers establish the splitting rules. This decision tree wasn’t able to correctly diagnose the HCV mono-infection samples (50,00% sensitivity and 55.00 % specificity). SensitivitySpecificity 90.57%73,30% m/z (/1,000) Protein Identity Predicted Mass (Da) 12.4Ig kappa chain C region11,609 18.4Haptoglobin alpha chain15,920 22.8Apolipoprotein- A130,778 46.8Haptoglobin45,205 84.6Plasminogen90,569 78.8Serotransferrin77,064 Funded by: We acknowledge: Dr Ward’s and Ndao’s Lab members, the participants of HIV-HCV Canadian Cohort (CTN 222), the Co-Investigators, Jeff Cohen, Brian Conway, Pierre Côté, Joseph Cox, David Haase, Shariq Haider, Marianne Harris, Julio Montaner, Neora Pick, Anita Rachlis, Danielle Rouleau, Roger Sandre, Mark Tyndall, David Wong, Marie- Louise Vachon and the study coordinators/ nurses. Visit out website: www.cocostudy.ca Methods Results Methods Acknowledgment