Interferon lambda alleles predict innate antiviral immune responses and hepatitis C virus permissiveness

Abstract

Hepatitis C virus (HCV) infection can result in viral chronicity or clearance. Although host genetics and particularly genetic variation in the interferon lambda (IFNL) locus are associated with spontaneous HCV clearance and treatment success, the mechanisms guiding these clinical outcomes remain unknown. Using a laser capture microdissection-driven unbiased systems virology approach, we isolated and transcriptionally profiled HCV-infected and adjacent primary human hepatocytes (PHHs) approaching single-cell resolution. An innate antiviral immune signature dominated the transcriptional response but differed in magnitude and diversity between HCV-infected and adjacent cells. Molecular signatures associated with more effective antiviral control were determined by comparing donors with high and low infection frequencies. Cells from donors with clinically unfavorable IFNL genotypes were infected at a greater frequency and exhibited dampened antiviral and cell death responses. These data suggest that early virus-host interactions, particularly host genetics and induction of innate immunity, critically determine the outcome of HCV infection.

Figure 1. Isolation of HCV infected PHH via LCM

(A) The HCV-dependent fluorescence relocalization (HDFR) reporter. Phase contrast (left) and fluorescent (right) images of HFDR expressing PHH infected with HCV. Filled arrowhead = HFDR nuclear translocated HCV infected cells. Empty arrowhead = HDFR non-translocated adjacent cells. (B) Schematic of laser capture microdissection (LCM). The collection surface is sticky for the slide membrane facilitating capture upon cutting. (C) Schematic of the capture of Mock, HCV infected and adjacent cells. (D) Schematic of the workflow from LCM to array. (E) LCM of mock, HCV infected and adjacent cells chronicled in a step-wise fashion from left to right. (F) Quantitative RT-PCR for HCV genomes in LCM sample lysate. Each dot represents the average value per 10 cells. The number of donors (donor n) per time point (days post infection, dpi) is indicated in the gray box. Asterisks indicate statistical significance as determined by ANOVA (P< 0.001). See also Figures S1–4.

Figure 2. Systems virology reveals unique signatures associated with HCV infected and adjacent cells

(A) HCV infectious virus production in PHH (left). Cells were infected at a similar MOI. Titers in culture supernatants were determined by TCID50 (donor n: 1, 3dpi = 19, 7dpi = 17). Each timepoint represents the 24hr accumulation of virus. The dashed line indicates the limit of detection. HCV infection frequency (right). Each dot represents the average infection frequency per donor (donor n: 1, 3dpi = 22, 7dpi = 20). (B) The numbers of significantly regulated genes (ANOVA, Benjamin Hochberg MTC FDR 0.05, fold change cutoff of 2) in HCV infected and adjacent cells. Donor N per time point: 8 (1dpi), 3 (3dpi) and 4 (7dpi). (C) The numbers of antiviral response genes induced in HCV infected and adjacent cells. (D) Microarray data for IFNL1 (E) IL29 protein in culture supernatants from donors 4720, 5679, 5728, 5123, 5686, and 3868. Select cultures were treated with HCV replication inhibitor 2’CMA. (F) The temporal order of the antiviral response in HCV infected and adjacent cells (ADJ). The absence of a colored box indicates fold change did not exceed the 2-fold cutoff. (G) Example expression data on 1dpi. (H) Confirmation of example array data by qPCR of cDNA libraries. Asterisks indicated statistical significance as determined by ANOVA (P<0.001, panels A, B, D, and G, P<0.05 panel E) or t-test (panel H). Error bars in D, E, G and H represent the range. See also Figure S5.

Figure 3. Host responses associated with more effective viral control

(A) Infection frequency in high (donors 4728, HFTR, 5737) and low (donors 5123, 5780, 5763) infection frequency donors. Each dot represents the average infection frequency per donor. Asterisks indicate statistical significance as determined by t-test (P<0.005). (B) Significantly regulated genes in HCV infected and adjacent cells at 1dpi. (C) Antiviral response genes in HCV infected and adjacent cells at 1dpi. (D) Venn diagram comparing all significantly regulated genes or antiviral response genes for HCV infected and adjacent cells and heatmap of antiviral response genes. White boxes indicate the gene did not exceed the 2-fold change cutoff. (E) Pathway analysis in Ingenuity IPA for the biological function “RNA virus replication.” The –Log(P-value) (orange bars) is measures the statistical strength of the submitted gene list. The numbers of associated genes are listed above each bar. The Z-score (purple bars) predicts the functional effect of the gene list. Negative Z-scores indicate a reduction in function. ADJ refers to “adjacent cells” throughout.

Figure 4. IFNL3, IFNL4 and SLC45A2 genotype data for all donors

(A) Description of the major and minor alleles for interferon lambda 3 (IFNL3) and 4 (IFNL4). (B) IFNL3, IFNL4, and SLC45A2 genotype data for all donors and their relation to infection frequency. Polymorphisms in SLC45A2 (solute carrier family 45, member 2, SNP (rs16891982) can predict European or African ethnicity. Af = African. Eur = European. Infection frequency classification is based on the infection frequency for each donor on 1dpi and 3dpi. For example, “low donors” had the lowest infection frequency on 1dpi and/or 3dpi. See also Table S1.

Figure 5. Genetic variation in IFNL is associated with increased HCV infection frequency and distinct host responses

(A) Infection frequency for each IFNL3 (rs12979860) genotype. Each dot represents the average frequency per donor. Statistical significance was determined by ANOVA (P <0.05). (B) Significantly regulated genes on 1dpi. Donor N: IFNL3 C/C = 2, C/T = 4, T/T = 2. (C) Antiviral response genes on 1dpi. The percentage of the overall host response dedicated to the antiviral program is indicated in the white box. (D) Predictive biological functional analysis for “viral infection.” The –Log(P-value) (orange bars) measures statistical strength of each gene list. The numbers of associated genes are listed above each bar. The Z-score (purple bars) predicts the functional effect of the gene list. Negative Z-scores indicate a reduction in function. (E) Venn diagram of significantly regulated genes and antiviral response genes for IFNL3 C/C and T/T genotypes. (F) Heatmap displaying the antiviral gene expression for unique and shared genes from panel E. (G) Venn diagram comparing gene expression data for “low infection frequency”, C/C and T/T donors. (H) Expression heatmap for genes shared between low infection frequency donors and C/C or T/T donors. Cell death related genes (purple) and ISGs (gray) are marked in the boxes below (I) Microarray data for housekeeping and select cell death genes. Statistical significance was determined by ANOVA. (J) Relative expression of cell death related genes in mock cells from IFNL3 C/C donors (ANOVA). ADJ refers to “adjacent cells” throughout. See also Figure S6.

Figure 6. Stochasticity of the host response is associated with host genotype and infection frequency

Data for each quadruplicate LCM sample (a-d) per group (mock, HCV, adjacent) per donor is depicted in each panel (A-C). The IFNL3 and SLC45A2 genotype for each donor is indicated by the color of each header. Genotype abbreviations are described in Figure 4. The donors are ordered from left to right with decreasing infection frequency. (A) Quantitation of HCV genomes by qRT-PCR in LCM sample lysate from 10 cells. (B and C) Microarray (red circles, left y-axis) and qPCR confirmation of array data (black open circles, right y-axis) for (B) the house keeping gene, beta actin, and (C) three interferon stimulated genes IFIT1, CCL5 and BST2.