Hepatitis C Virus Treatment with Direct Acting Antivirals Induces Rapid Changes in the Hepatic Proteome

Abstract

Treatment of chronic hepatitis C virus with direct acting antivirals usually eradicates infection, but liver fibrosis does not resolve concurrently. In patients who develop cirrhosis prior to hepatitis C virus treatment, hepatic decompensation and hepatocellular carcinoma can still occur after viral elimination due to residual fibrosis. We hypothesized the liver proteome would exhibit meaningful changes in inflammatory and fibrinogenic pathways change upon hepatitis C virus eradication, which could impact subsequent fibrosis regression. We analyzed the liver proteome and phosphoproteome of paired liver biopsies obtained from 8 hepatitis C virus-infected patients before or immediately after treatment with direct acting antivirals. Proteins in interferon signaling and antiviral pathways decreased concurrent with hepatitis C virus treatment, consistent with prior transcriptomic analyses. Expression of extracellular matrix proteins associated with liver fibrosis did not change with treatment, but the phosphorylation pattern of proteins present within signaling pathways implicated in hepatic fibrinogenesis, including the ERK1/2 pathway, were altered concurrent with hepatitis C virus treatment. Hepatitis C virus treatment leads to reduced expression of hepatic proteins involved in interferon and antiviral signaling. Additionally, changes in fibrosis signaling pathways are detectable before alteration in extracellular matrix proteins, identifying a putative chronology for the dynamic processes involved in fibrosis reversal.

Introduction

Hepatitis C virus (HCV) causes a chronic infection in 81 million people worldwide, including over 2 million people in the United States [1, 2]. HCV infection results in liver inflammation and fibrinogenesis due to hepatic stellate cell and myofibroblast activation, which can progress to cirrhosis and hepatocellular carcinoma (HCC) [3, 4]. Treatment with direct acting antiviral (DAA) therapy typically results in a sustained virologic response (SVR), synonymous with cure, but fibrotic and neoplastic complications of residual liver scarring can still occur after SVR since fibrosis does not immediately resolve upon viral elimination [2, 5, 6].

While liver fibrosis can improve in the years following SVR, the mechanisms dictating why individuals differ in their extent of fibrosis regression are poorly understood [7–11]. The sustained risk of developing liver cancer in patients with hepatic cirrhosis correlates with epigenetic changes associated with HCC risk that persists after SVR [12–16]. Inhibiting HCV replication with DAAs causes a rapid decline in interferon signaling and inflammation, indicating that immune function is at least partially restored with treatment [17–19]; however, persistent elevation of CD4+ regulatory T-cells in liver and blood post-SVR suggest immune function does not completely normalize [20, 21]. Therefore, evaluating how the liver changes during HCV treatment and how these changes impact fibrosis and immunity over time may identify new targets for therapeutic modulation to facilitate fibrosis regression.

In prior work, we identified transcriptional changes in the liver that occur during HCV treatment with DAAs, including a global down-regulation of endogenous interferon signaling pathways [17, 22]. Previous work suggested the interferon-associated proteome in the liver may not change in parallel with the interferon-associated transcriptome, suggesting HCV could impact translation [23, 24]. Our long term goal is to understand clinically significant changes in the hepatic proteome that can predict the extent of fibrosis regression and which may thus be susceptibility to therapeutic modulation. In the current study, we analyzed immediate changes in the whole liver proteome and phosphoproteome using paired liver biopsies from 8 HCV-infected patients collected before or immediately after treatment with DAAs. We sought to determine the extent of acute changes in proteins associated with immune function, interferon signaling, and fibrogenesis after HCV suppression.

Methods

Subject characteristics

Paired liver biopsies obtained before and immediately after completing HCV treatment (within 1–5 days) were collected from 8 subjects with genotype 1 HCV infection treated with DAAs for either 6 or 12 weeks. Subjects received sofosbuvir combined with ledipasvir for 12 weeks (n=3), or sofosbuvir and ledipasvir combined with either a non-nucleoside analog (radalbuvir, n=2) or an NS3/4A protease inhibitor (vedroprevir, n=3) for 6 weeks [25]. Seven subjects achieved SVR, and 1 subject who received sofosbuvir, ledipasvir, and radalbuvir relapsed after treatment, with relapse occurring 4 weeks after the post-treatment liver biopsy was obtained. Liver fibrosis and inflammation were assessed by a pathologist using Ishak and Histologic Activity Index (HAI) scores [26]. As previously reported, all subjects provided written informed consent which included permission to use samples and data collected in the trial for future studies [25]. This study was approved by the Institutional Review Board at the Medical University of South Carolina and conformed to the ethical guidelines of the Declaration of Helsinki.

Sample collection

Liver biopsies were snap frozen in cryovials by immersion in liquid nitrogen immediately after collection and were stored at −80 °C until processing (the overall work flow is shown in Supplemental Figure 1). At processing, cryovials were passively warmed for 1 minute at room temperature and biopsy cores were transferred with a pipette tip to a chilled Eppendorf tube containing 100 ul of homogenization buffer (0.2 M EPPS, 8 M MS grade urea; PhosSTOP™ (Sigma-Aldrich); Complete mini-protease inhibitor cocktail (Roche); Universal Nuclease (Pierce)). Samples were manually homogenized with a pestle followed by 3x1-minute sonification in a chilled water bath and were manually inspected to assure adequate disruption of the tissue. Samples were centrifuged at 10,000xg for 10 minutes at 4°C and the supernatant was transferred to a new cryovial tube. Protein concentration was determined by BCA assay (ThermoScientific). An aliquot of each sample was loaded on a 1D gel and stained with Coomassie Blue to evaluate potential blood contamination and to assess lysate integrity. Samples were then snap frozen in liquid nitrogen and stored at −80°C before shipment on dry ice to the Thermo Fisher Center for Multiplexed Proteomics at Harvard Medical School for TMT labeling and LC-MS3 analysis.

Sample preparation, liquid chromatography-MS3 spectrometry (LC-MS/MS), and database searching

A detailed description of how samples were processed and analyzed for multiplexed quantitation can be found in the Supplemental Methods [27]. In brief, sample processing steps included protein precipitation, sequential protein digestion using LysC and trypsin, peptide labeling with TMT10plex reagents (Thermo Scientific), and peptide fractionation to perform global proteomics and phosphopeptides were enriched from the re-combined fractions by IMAC enrichment. Multiplexed quantitative mass spectrometry data for protein fractions were collected on an Orbitrap Fusion mass spectrometer using a RTS-MS3 method [28]. Enriched phosphopeptides were analyzed by an Orbitrap Lumos mass spectrometer using an LC-MS3 method [29]. MS/MS data were searched against a Uniprot human database (downloaded in 2014) with common contaminants and reversed sequences using the target-decoy search strategy with SEQUEST [30]. Further data processing steps included controlling peptide and protein level false discovery rates, assembling proteins from peptides, protein quantification from peptides, and phosphosite localization and quantification.

Statistical and bioinformatic analysis

Prior to statistical analysis, the top 20 proteins with abundance profiles similar to hemoglobin, indicative of erythrocyte contamination, were removed (Supplemental Figure 2). To identify significantly regulated proteins and phosphorylation sites, paired two-sided Student’s t-tests were performed using a p-value <0.05 or a permutation-based false discovery rate (FDR) <0.05 with 250 randomizations and S0 fold change parameter of 0.1 using Perseus v.1.6.14.0 (Max Plank Institute) [31, 32]. Enrichment analysis was performed to test for enrichment of annotations associated with proteins (p-value <0.05) as compared to all observed proteins [31]. Annotations including GO biological process (GOBP), molecular function (GOMF), cellular component (GOCC), and Reactome pathways were downloaded from the Uniprot human proteome and Reactome databases (October 2020) and uploaded into Perseus. Phosphorylation site-specific annotations were downloaded from the PhosphoSitePlus database (September 2020) [33]. Known sites of phosphorylation were annotated with residue-level regulatory function, disease correlations, protein interactions and function, and probed for the enrichment of kinase consensus motifs and kinase-substrate relations. For protein and phosphosite level enrichment analysis, the log2 transformed intensities of proteins or phosphorylation sites with a p-value <0.05 were normalized by z-score and hierarchical clustering was performed using Pearson correlation, average linkage, and preprocessing with k-means. Clusters of proteins, visualized in heatmaps, were extracted and enrichment of GO terms was evaluated using a Fisher exact test with a Benjamini-Hochberg FDR<0.02 threshold.

To probe for pathways that are regulated by increasing and decreasing relative phosphorylation at specific residues, we used PTM-Signature Enrichment Analysis (PTM-SEA) to query the PTM signatures database (PTMsigDB) v1.9.0 using the seven amino acid sequence flanking the sites of phosphorylation [34]. The sign of the normalized enrichment score (NES) calculated for each signature corresponds to the direction of change in the pre- or post-treatment conditions. P-values for each signature were derived from 1,000 random permutations and further adjusted for multiple hypothesis testing using a Benjamini-Hochberg approach. The p-values reflect enrichment of signatures in the sample sets as compared to the PTM signatures database.

Results

Liver tissue samples obtained before or immediately after completing HCV treatment were analyzed by mass spectrometry to identify relative changes in protein abundance and phosphorylation status that occurred as a result of HCV treatment. The patient cohort had differing levels of liver fibrosis and received DAA treatment for 6 or 12 weeks (Table 1). Quantitative proteomics and small scale phospho-proteomics were performed by labeling samples with tandem mass tags (TMT, ThermoScientific) followed by LC-MS3 analysis [28, 35] (workflow shown in Supplemental Figure 1). A comparison of protein abundance before and after treatment of the 5,361 identified proteins (Supplemental File 1) revealed 191 proteins that reached a significance threshold of an adjusted p-value (FDR<0.05) and minimum log2 fold change (S0, 0.1) (Figure 1A, volcano plot). Hierarchical clustering of differentially expressed proteins pre- and post-treatment is shown in Figure 1B (protein list in Supplemental File 2). Annotation enrichment analysis of the differentially expressed proteins was dominated by Gene Ontology terms and Reactome pathways reflecting a reduced response to viral infection, reduced interferon signaling, and reduced inflammation over the course of HCV treatment (Figure 1C, Supplemental File 3). These data are comparable to findings from transcriptional analyses of pre- and post-treatment liver biopsies after DAA treatment of HCV [17, 22]. Of the 191 differentially expressed proteins, 159 were identified as products of interferon stimulated genes (ISGs) by the Interferome database [36]. Proteins associated with immune cell activation and antigen presentation decreased with treatment (e.g. NFKB, HLA-DR). In addition, proteins associated with lipid metabolism that are known ISGs, including APOE, APOH, APOL2, and APOL3, changed with treatment, consistent with previously observed changes during HCV treatment in hepatic lipid gene expression and serum apolipoprotein levels [37, 38].

Table 1:

Patient Characteristics

ID	Gender	BMI > 30	Age	Race	HCV Genotype	IFNL4 Genotype	Baseline Viral Load (IU/ml)	Ishak (Pre/Post)	HAI (Pre/Post)	Regimen	Outcome	AST (Pre/Post)	Steatosis (Pre/Post)
1	Male	Yes	52	AA	1a	ΔG/ΔG	1,856,814	4/4	7/10	C	SVR	152/22	4/1
2	Male	No	47	AA	1a	ΔG/TT	3,130,991	6/6	7/5	A	SVR	77/23	1/0
3	Male	No	62	AA	1a	ΔG/TT	2,807,294	6/6	7/2	A	SVR	178/48	Tr/Tr
4	Male	Yes	59	AA	1b	TT/TT	21,502,372	2/0	10/3	A	SVR	40/32	Tr/0
5	Male	No	52	White	1a	TT/TT	1,981,381	2/2	11/5	C	SVR	68/15	Tr/Tr
6	Female	No	55	AA	1b	ΔG/TT	14,157	4/4	8/5	B	SVR	32/32	Tr/Tr
7	Male	No	61	AA	1a	ΔG/TT	1,922,287	4/3	11/10	B	Relapse	41/18	Tr/Tr
8	Female	Yes	56	AA	1a	ΔG/TT	352,592	0/0	4/5	C	SVR	23/14	Tr/Tr

BMI= body mass index, AA=African-American, IFNL4 genotype = rs368234815 genotype, Ishak: Biopsy determined Ishak Fibrosis score, HAI: Biopsy determined Histologic Activity Index score, Regimen: A (sofosbuvir + ledipasvir for 12 weeks), B (sofosbuvir + ledipasvir + radalbuvir for 6 weeks), C (sofosbuvir + ledipasvir + vedroprevir for 6 weeks), Outcome: SVR (sustained virologic response) or relapse, AST: Aspartate aminotransferase, Steatosis: Biopsy determined steatosis score, Tr=trace.

Figure 1: Proteomic changes in the liver during DAA therapy of HCV.

(A) Volcano plot of statistical significance plotted against the log₂ fold change in protein intensity pre- and post-HCV treatment (n=8 pairs). Significantly regulated proteins (red) were determined using a paired t-test with an adjusted p-value <0.05 and a minimum fold change threshold (S0 parameter of 0.1 in Perseus). Additional proteins reaching a threshold of p-value <0.05 are shown in black, while non-significant proteins are shown in grey. Protein lists are provided in Supplemental Files 1, 2 and 4. (B) Hierarchical clustering (Pearson correlation) of z-score normalized, log₂ transformed protein intensities from each patient pre- and post-treatment revealed two predominant clusters. (C) Annotation enrichment score of selected GO-terms (black) and Reactome pathways (red) of significantly regulated proteins (adjusted p-value<0.05) plotted as a function of log10 Benjamini-Hochberg adjusted p-value demonstrates the reduction in the endogenous hepatic response to viral infection following treatment. Enrichment analyses using 2 thresholds of significance are provided in Supplemental Files 3 and 5.

Expansion of the annotation enrichment analysis to include 878 proteins with an unadjusted p-value of <0.05 (Supplemental File 4) identified a decrease in tRNA ligases in the cytosolic tRNA aminoacylation pathway, with tryptophanyl-tRNA synthetase (WRS), implicated in the innate immune response to viral infection [39], exhibiting the largest and most significant change (Log₂FC 0.81, adj. p-value 0.003). Proteins associated with the translational machinery necessary for viral replication were also reduced (Supplemental File 5). In contrast, the abundance of proteins in pathways related to collagens, ECM proteins, and the molecular and structural pathogenesis of fibrosis did not change significantly by the end of HCV treatment (Figure 2).

Figure 2: No change in proteomic pathways associated with fibrosis.

No significant difference is observed in expression of proteins associated with collagen, the extracellular matrix, and TGFB signaling. Shown are Volcano plots with relevant proteins identified within the dataset that are present in the respective shown pathways. Within each pathway, proteins changing significantly are shown in red and proteins without significant change are shown in blue.

We next examined phosphoproteomic changes, hypothesizing that signaling pathways associated with fibrosis progression and regression may be identifiable early after HCV treatment before proteomic changes. This small-scale analysis identified 2,519 phosphopeptides (Supplemental File 6) in the eight pairs of samples with 467 regulated sites (p-value <0.05) (Supplemental File 7). Phosphorylation was detected on 87 kinases and 19 phosphatases or subunits thereof (Supplemental File 8). To identify site-and direction-specific phosphoproteomic profiles associated with pathways, kinase activation, or drug perturbations, the entire data set of sequence motifs (phosphosite +/−7 residues) was probed against the post-translational modification (PTM) signatures database (PTMsigDB) using PTM signature enrichment analysis (PTM-SEA) (Figure 3A, Supplemental File 9). Matches with significantly regulated phosphorylation sites (Benjamini-Hochberg adj. p-value<0.05) predicted activation of cAMP-dependent PKA (PRACA) and PKC (PRKACA) following treatment (Figure 3B). Consistent with this observation, the abundance and phosphorylation of PKA at an activating site of phosphorylation (PRKCA_226) was elevated during the time course of therapy (Figure 3B). Treatment-induced changes in site-specific phosphorylation also matched the PTM signature associated with perturbation with phorbol ester (Figure 3A), consistent with PKC activation.

Figure 3: Post-translational modification enrichment analysis identifies changes in kinases associated with fibrosis.

(A) Shown are predicted changes in kinases and “perturbations” based on analysis of phosphorylation sites throughout treatment as analyzed by PTM-SEA. Blue represents lower relative activity and red represents higher relative activity when comparing pre- to post-treatment liver biopsies. Phosphorylation sites, kinases, and enrichment analysis are provided in Supplemental Files 6–9. (B) Volcano plot showing the increase in substrates with motifs recognized by basophilic kinases including PKA and PKC (blue) and decrease in substrates with motifs recognized by proline directed kinases including ERK1/2 (green). Motifs are shown on the above horizontal axis. Sequence logos demonstrate enriched kinase substrate consensus motifs pre- and post-treatment.

PTM-SEA also predicted reduced activity of proline-directed kinases and down-regulation of ERK2 (MAPK1) mediated signaling (Figure 3A, Supplemental File 7). A heatmap of putative ERK1/2 substrates decreasing with treatment is shown in Figure 4A (Supplemental File 10). Phosphorylation of ERK2 at activating sites T185 and Y187 was detected but was highly variable among patient samples. The majority of the ERK1/ERK2 substrates had reduced abundance in post-treatment relative to pre-treatment samples (Figure 4B, Supplemental File 10).

Figure 4: Change in phosphorylation of ERK substrates.

(A) Heatmap of regulated phosphorylation sites with consensus motifs for proline directed kinases, including potential phosphorylation by ERK1/2. Putative ERK1/2 substrates are provided in Supplemental File 10. (B) Volcano plot showing gene name with regulated sites of phosphorylation observed pre- and post-treatment. Sites highlighted in blue are putative substrates of ERK1/2. Sites highlighted in red are other substrates with an adjusted p-value <0.05 and a minimum fold change of 0.1 (S0 parameter).

Finally, changes in key regulatory phosphorylation sites suggested a reduction in apoptosis and cell death may occur concurrent with HCV eradication. This included reduced abundance and phosphorylation of ISGs, including STAT1 S727 phosphorylation, a site known to induce transcription and apoptosis [40, 41] and increased inhibitory phosphorylation at RIPK1_S320, a kinase implicated in liver fibrosis and TNFA signaling [42, 43].

Discussion

We performed a comprehensive proteomic analysis of paired pre- and post-treatment liver biopsies from 8 HCV patients treated with DAAs for 6 -12 weeks. Interferon-stimulated and antiviral proteins were the primary analytes rapidly down-regulated during treatment. These results are consistent with findings from transcriptomic studies in the liver [17, 19, 22] and argue translation of ISGs is intact in HCV-infected liver, contrary to prior observations in cell lines [23, 24]. In contrast, we did not identify a proteomic signature suggestive of fibrosis regression when examining known pathways associated with collagen and the ECM (Figure 2). This result may be unsurprising given that post-treatment liver biopsies were obtained within 1-5 days after completing treatment and fibrosis resolution after SVR can take years and does not occur in all people [7–10]. Interestingly, examination of the phosphoproteome identified prominent changes in ERK1/2 signaling pathways suggesting HCV treatment reduced ERK1/2 activity. Since the pathogenesis of hepatic fibrosis is associated with increased ERK1/2 activity [3, 44], this observation suggests HCV suppression and elimination alone may immediately alter fibrogenic pathways in the liver, which could facilitate fibrosis resolution over time.

ERK1/2 signaling associates with activation of hepatic stellate cells and myofibroblasts that function as the key source of ECM proteins and fibrogenesis during HCV infection [3]. The reduction observed in this pathway is potentially an important finding, as while there does not appear to be an immediate change in expression of proteins associated with fibrosis pathways (Figure 2), the rapid down-regulation of signaling within fibrinogenic pathway highlights a putative mechanism for fibrosis reversal that occurs in many patients after HCV elimination [11]. This is remarkable and suggests the liver’s fibrinogenic signaling potential is rapidly reduced, even when these pathways have been active for years in HCV patients. Additionally, this implies that these fibrinogenic signaling pathways are dynamic. One potential limitation of this study is that since we examined whole liver, and DAA treatment likely alters the composition of immune cells in the liver [45–47], it is possible that some contribution of the change in ERK signaling was caused by immune cells. However, we suspect that hepatic stellate cells make a critical contribution to the change in ERK signaling that we identified given that they proliferate aggressively during liver injury [3, 44, 48] and thus may contribute meaningfully to the proteomic signal.

A number of additional results in this study correlate well with previous observations. The reduction in NFKB and HLA-DR (Figure 1) correlates with the observed reduction in inflammatory liver cells that occurs due to DAA therapy that likely reflects hepatic egress [45–47]. In addition, changes in proteins associated with lipid metabolism are consistent with prior transcriptional work from liver biopsies [37] and in vitro analyses examining HCV infection in hepatocyte-like cell lines [24, 49, 50]. Although interferon stimulated proteins decreased with treatment, we were unable to distinguish whether these changes were more related to type-I or type-III IFN signaling given the overlap in signatures for these two pathways. Several of the negative regulators of IFN signaling that can differentially impact these signaling pathways (e.g., SOCS1, SOCS3, USP18) were not quantified in this analysis and therefore, we cannot draw any conclusions as to their involvement.

We observed no overlap in the protein signatures previously associated with rapid post-liver transplant fibrosis [51] and minimal overlap in protein signatures associated with fatty acid metabolism, PPARA, and IL-6/STAT3 pathways or hepatic stellate signature associated with fibrosis progression [23, 52]. These differences may relate to the nature of the samples analyzed and/or the early time point post-treatment that was analyzed since fibrosis resolution occurs at timepoints later than what was sampled.

The study’s limitations include the heterogeneity of the patient cohort with respect to baseline liver fibrosis, the DAA regimen received, patient race, baseline HCV viral load, and IFNL4 genotype (Table 1), all variables which could have a significant impact on the host proteome. There were insufficient samples to stratify the results by fibrosis stage or any other host or viral variable. One patient experienced treatment relapse after their post-treatment liver biopsy had been obtained, and thus, residual HCV could have influenced the results, but exclusion of data from this subject did not alter the results. A lack of ability to attribute protein signatures to specific cell-types specificity is intrinsic to analyses of whole tissue biopsies and may have diluted the proteomic signal from stellate and immune cells. Variability in liver biopsies between pre- and post-treatment samples and between subjects likely also influenced the results. Finally, because the biopsies analyzed were collected immediately after HCV treatment and longitudinal samples were not available, we are unable to correlate the findings reported here with biopsy-documented fibrosis regression.

In conclusion, this analysis identified immediate changes in the liver proteome and phosphoproteome that occur following HCV replication suppression with DAAs. Identification of changes in interferon stimulated proteins and ERK1/2 pathways provides important insight into the molecular events that may portend future fibrosis regression and that merit further evaluation.

Supplementary Material

fS1

Supplemental Figure 1: Experimental workflow.

fS2

Supplemental Figure 2: Contaminant erythrocyte proteins with ratio profiles similar to hemoglobin subunits. These proteins were removed prior to statistical and bioinformatic analysis.

fS3

Supplemental Figure 3: Histograms illustrating the distribution of log2 transformed intensity ratios of proteins pre- and post-treatment for each subject.

Table 2:

Annotation enrichment of GO terms and Reactome pathways for proteins changing during DAA therapy for HCV.

Annotation Category	Category Name	Enrichment factor	P value
Decreased with Treatment
Gene ontology (biological process)	Type I Interferon Signaling Pathway [GO:0060337]	8.1375	9.35E-21
Gene ontology (biological process)	Defense Response to Virus [GO:0051607]	4.706	7.27E-19
Gene ontology (biological process)	Negative Regulation of Viral Genome Replication [GO:0045071]	5.859	1.37E-09
Gene ontology (biological process)	Interferon-gamma-mediated Signaling Pathway [GO:0060333]	4.5069	8.21E-09
Gene ontology (molecular function)	Translation Initiation Factor Activity [GO:0003743]	3.906	1.20E-07
Reactome Pathway	Interferon Gamma Signaling	6.1031	4.99E-07
Reactome Pathway	ISG15 Antiviral Mechanism	3.3734	7.62E-07
Gene ontology (biological process)	Response to Virus [GO:0009615]	3.568	2.84E-07
Gene ontology (cellular component)	Eukaryotic 48S Preinitiation Complex [GO:0033290]	6.2775	1.37E-06
Gene ontology (cellular component)	Eukaryotic 43S Preinitiation Complex [GO:0016282]	5.859	3.09E-06
Gene ontology (cellular component)	Cytoplasmic Stress Granule [GO:0010494]	3.3201	3.66E-06
Reactome Pathway	Cytosolic tRNA Aminoacylation	6.51	3.73E-06
Reactome Pathway	Interferon Alpha/Beta Signaling	8.37	7.08E-06
Gene ontology (cellular component)	Eukaryotic Translation Initiation Factor3 Complex [GO:0005852]	5.58	1.83E-05
Gene ontology (molecular function)	RNA Cap Binding [GO:0000339]	7.595	3.32E-06
Reactome Pathway	ER-Phagosome Pathway	4.4386	2.13E-05
Gene ontology (molecular function)	Calcium Ion Binding [GO:0005509]	0.20238	6.48E-06
Gene ontology (molecular function)	tRNA Binding [GO:0000049]	3.255	1.75E-05
Increased with Treatment
Gene ontology (cellular component)	Extracellular Exosome [GO:0070062]	1.5043	1.15E-07
Gene ontology (molecular function)	RNA Binding [GO:0003723]	0.42416	3.88E-08
Gene ontology (molecular function)	Alcohol Dehydrogenase Activity, Zinc-dependent [GO:0004024]	13.579	4.77E-06
Gene ontology (biological process)	NAD Biosynthetic Process [GO:0009435]	10.863	3.57E-06
Gene ontology (biological process)	Cellular Aldehyde Metabolic Process [GO:0006081]	10.863	3.57E-06
Gene ontology (molecular function)	Retinol Dehydrogenase Activity [GO:0004745]	8.8881	1.73E-05

Enrichment of Gene Ontology terms and Reactome Pathways using proteins with a p-value <0.05 (n=878) as compared to all quantified proteins. Shown are the annotation category, category name, enrichment factor, and p-value. Only pathways with a Benjamini-Hochberg FDR <0.02 are shown.

Significance statement:

How the liver heals after HCV has been cured with antiviral therapy is poorly understood. Using liver biopsies obtained before and after HCV treatment, we measured how liver proteins and their phosphorylation, or activation status, change during treatment. We found that proteins that promote antiviral defense decreased with treatment. In addition, we found that the phosphorylation status of proteins that contribute to liver fibrosis changed during treatment, suggesting treating HCV immediately lowers the activity of pathways that can cause progressive liver fibrosis.