Impact of Interferon Lambda 4 Genotype on Interferon-Stimulated Gene Expression During Direct-Acting Antiviral Therapy for Hepatitis C

Abstract

New directly acting antivirals (DAAs) provide very high cure rates in most patients infected by hepatitis C virus (HCV). However, some patient groups have been relatively harder to treat, including those with cirrhosis or infected with HCV genotype 3. In the recent BOSON trial, genotype 3, patients with cirrhosis receiving a 16-week course of sofosbuvir and ribavirin had a sustained virological response (SVR) rate of around 50%. In patients with cirrhosis, interferon lambda 4 (IFNL4) CC genotype was significantly associated with SVR. This genotype was also associated with a lower interferon-stimulated gene (ISG) signature in peripheral blood and in liver at baseline. Unexpectedly, patients with the CC genotype showed a dynamic increase in ISG expression between weeks 4 and 16 of DAA therapy, whereas the reverse was true for non-CC patients. Conclusion: These data provide an important dynamic link between host genotype and phenotype in HCV therapy also potentially relevant to naturally acquired infection. (Hepatology 2018; 00:000-000).

Figure 1

IFNL4 genotypes and clinical outcomes in the BOSON study. (A) OR graph showing the impact of IFNL4 genotype on SVR during therapy in chronic HCV patients from the BOSON study (DAA only arms). (B) Summary data from gene expression studies of peripheral blood from patients with cirrhosis taken at baseline. Venn diagram shows differentially regulated genes in IFNL4 CC and non‐CC patients (adjusted P value <0.05) when compared to common referenced healthy control group (n = 8).

Figure 2

IFNL4 genotypes and gene expression in peripheral blood during chronic HCV infection. (A,B) Volcano plot showing change in overall gene expression in peripheral blood between non‐CC patients and healthy controls (A) and direct comparison between non‐CC versus CC patients (B) Log2 fold change is plotted along the X‐axis and the P value along the y‐axis. The dotted red line shows the threshold of significance (P < 0.05). Genes labeled in the figure are those that were validated by real‐time PCR. (C,D) GSEA analysis showing enrichment of IFNα/β gene set in non‐CC versus control (C) and a similar comparison of genes comparing non‐CC versus CC patients. (E) IFNL4 genotypes and pathway analyses during chronic HCV infection: over‐representation analysis of reactome pathways on the list of up‐regulated genes between non‐CC versus CC patient groups. The Y‐axis shows the top six pathways (significance is color coded) X‐axis shows the proportion of gene in each pathway compared to the total list of genes. The size of the circles represent the gene counts in the pathway.

Figure 3

IFNL4 genotypes and gene expression in liver of HCV patients. (A). Volcano plot showing differentially modulated genes in liver by microarray analysis comparing non‐CC versus CC patients. X‐axis shows log2 fold change and Y‐axis the –log 10 P value. The dotted red line shows the threshold of significance (P < 0.05). Genes illustrated in the figure are those that were found in the blood in Fig. 1C/D when comparing the non‐CC patients versus healthy control and the non‐CC versus CC groups. (B) Reactome pathways enrichment analysis on differentially up‐regulated list of genes obtained by microarray analysis of liver tissue between non‐CC versus CC patients. The X‐axis shows the gene counts in each pathway and the Y‐axis shows the eight most significant pathways sorted by significance. Abbreviations: MDA5, melanoma differentiation‐associated gene 5; RIG‐I, retinoic acid‐inducible gene I.

Figure 4

IFNL4 genotypes and RNAseq analysis in liver of HCV patients. (A) Reactome pathways enrichment analysis on differentially up‐regulated list of genes obtained by RNAseq analysis between non‐CC versus CC patients. Criteria for inclusion were a minimum of 1 lfc (log 2 fold change) in expression and P value <0.05. The size of circles reflects the gene count in the pathway, the X‐axis represent the gene ratio (N genes in pathway/n genes in the entire list of genes), and the color of the circles reflects the adjusted P value. (B) IFNL4 genotypes and cytoband analysis in liver of HCV patients: expression graph of chromosomally sorted genes in cytoband 19q13.2. Lines across the graph represent average gene expression in each IFNL4 genotype (TT in red, black dotted lines CT, and black solid line represent expression in CC groups). IFNL3/4 gene is not shown because the genes included in the graph are those that have an average gene expression of log2 cpm >2. (C,D) GSEA showing enrichment of IFNα and IFNL Huh‐7 signature gene (uniquely up‐regulated genes when Huh7 cells are stimulated by IFNα or IFNL3) on non‐CC versus CC gene list in liver. Abbreviations: MDA5, melanoma differentiation‐associated gene 5; RIG‐I, retinoic acid‐inducible gene I.

Figure 5

qPCR analysis of peripheral blood gene expression at baseline according to IFNL4 genotype. A set of genes was quantified by PCR from RNA derived from peripheral blood using the relative quantification method with genes normalized against a housekeeping gene (GAPDH). Results from eight ISG gene expressions are shown in a total of 45 patients from the same cohort, of which 27 were CC and 18 patients were non‐CC genotype. Results from qPCR analysis were fed into GraphPad Prism (7.0; GraphPad Software Inc., La Jolla, CA) to generate the graphs. P values were generated using the statistical software within Prism using log values of fold changes and SEM is shown. Abbreviation: GAPDH, glyceraldehyde 3‐phosphate dehydrogenase.

Figure 6

Dynamic changes in peripheral blood gene expression during therapy. (A) Expression of an average of a subset of genes (RSAD2, IFIT1, IFIT3, IFI44L, MX1, CXCL10, ISG15, and HERC5) over time, that is, BL, 4 weeks, 16 weeks, and at the end of study in CC versus non‐CC patient group. Similar results were obtained when all genes of the IFNα/β pathway were analyzed. (B) Gene expression data on the average of 27 genes shown in Table 1 over time as described above. (C) Representative example of gene expression in gene IFIT3 showing the increased expression in the CC patient group at week 16 of treatment compared to week 4. Significance was determined using a paired t test. Abbreviations: BL, baseline; EOS, end of study.

Figure 7

(A) Dynamic changes in reactome pathways during therapy. Over‐representation analysis of reactome pathway genes in CC patients at 16 versus 4 weeks. The up‐regulated list of genes was generated using paired samples at 16 and 4 weeks for each patient and analyzed using paired t test. The Y‐axis shows the eight most significant over‐represented reactome pathways (significance color coded) and X‐axis shows gene counts in each pathway. (B,C) Up‐regulation of IFN‐regulated genes according to IFNL4 genotypes: GSEA analysis showing enrichment of the IFNα/β gene set between 16 versus 4 weeks in CC and non‐CC patient blood samples. The dotted vertical line shows the enrichment “core.” Genes were enriched for the IFNα/β pathway in the CC patient group at 16 weeks. However, in the non‐CC patient group, such an enrichment was not observed.