This study was approved by the University of Southern California Institutional Review Board (HS-18-00254). Written and oral consent was obtained from all enrolled subjects.

Blood samples were collected from subjects with untreated chronic HCV infection (cHCV, n=4), from subjects that spontaneously resolved HCV infection (rHCV, n=4), and from non-HCV-exposed and control subjects (CTR, n=3). We also included chronic HCV patients successfully treated for 12 weeks with direct-acting-antivirals (DAA) (n=3) that had blood collected before therapy (Pre-tx), at an early stage (2-4 weeks coinciding with non-detectable viral load) of DAA therapy (Early-tx), in a later phase (8 weeks) of therapy (Late-tx), and at 12 weeks after treatment conclusion (Post-tx). All patients experienced a sustained virologic response ( Supplementary Table 1 ). All HCV subjects were infected with HCV genotype 1. Peripheral blood mononuclear cells (PBMCs) were isolated from the blood samples, viably frozen at 20-50 million cells/mL, and stored in liquid nitrogen for subsequent use. All subjects were previously screened for immunogenic response to CMV and Flu epitopes, and the HCV cohort was also evaluated regarding the HCV epitopes. Only the positive subjects were included in this study.

PBMCs were thawed in RPMI media and centrifuged at 350xg for 10min at 4˚C. Cells were washed with Washing Buffer (PBS, pH 7.4 + 5% FBS), centrifuged, and the cell pellet was resuspended in Stain Buffer (PBS, pH 7.4 + 5% FBS + 0.1g/L Herring sperm DNA). Then, we added 0.2 µl 100µM d-Biotin (Avidity LLC, Aurora, CO) per included dextramer and added 2µl of desired dCODE™ Dextramer^®-PE reagents (Immudex, Denmark), including the negative control for allele and random dextramers ( Table 1 , dextramers #1-4 for cHCV and rHCV samples, and dextramer #1, #5-6 for DAA therapy samples). After 10 min of incubation in the dark, we added the antibody cocktail for CD8⁺ T cell staining (CD4-APC, CD3-FITC, CD8-V500, and CD56-V450) and incubated for 30 min at 4˚C in the dark. We sorted the CD3⁺ CD8⁺ Dextramer⁺ cells using BD FACSAria™ Fusion into 90% FBS solution at 4˚C ( Figure 1A , Supplementary Figure 1 ). Sorted cells were centrifuged and resuspended in 50µl of PBS with 0.04% BSA for scRNAseq partitioning.

Single-cell suspensions were processed using Chromium Next GEM Single Cell 5’ Kit v1.1 (10x Genomics, Pleasanton, CA) for samples from chronic and resolved HCV patients, and Chromium Next GEM Single Cell 5’ Kit v2 - dual index, for samples from cHCV patients treated with DAA therapy. Briefly, up to 16,500 cells were resuspended in the reaction mix and loaded into the Chromium Next GEM Chip G, followed by the loading of the Gel Beads and the partitioning oil. Chip G was loaded into the 10x Chromium controller for cell partitioning. After partitioning, 100µL of GEMs (Gel Bead-In EMulsion) were transferred to a new tube and placed at the C1000 Touch Thermal Cycler for the first phase of reverse transcription. All gene expression libraries and surface protein libraries were prepared and sequenced together to avoid batch effects. Single-cell RNAseq libraries were sequenced at the USC Norris Molecular Genomics Core and Keck Genomics Platform. Libraries were quantified using Kapa Quant Library qPCR (Roche Diagnostics Corporation, Indianapolis, IN), and quality was assessed using Agilent BioAnalyzer 2100. Prepared libraries were sequenced on Illumina Novaseq 6000 at 2x100 cycles. At least 40,000 reads were obtained per cell.

Raw data processing: Cell by gene count matrices were generated by processing FASTQ files using the Cell Ranger count (version 5.0.1, 10X Genomics) pipeline with the GRCh38 (version refdata-gex-GRCh38-2020-A, 10X Genomics) human genome reference for each sample. For CD8⁺ T cell phenotype classification, we used signatures stablished by Hao et al. (14) and Doering et al. (15). Table 2 shows the genes included for each phenotype signature.

Acute HCV infection can either progress to chronicity (cHCV) or spontaneously resolve (rHCV). Accumulating evidence suggests changes in the adaptive immune response against other viruses during and after HCV infection. To identify differences that influence the fate of HCV infection, we isolated CD8⁺ T cells from the blood of four subjects with untreated chronic HCV and four subjects that resolved HCV without treatment. Here, we combined barcoded dCODE™ Dextramers and single-cell RNAseq (scRNAseq) technology (10x Genomics) to isolate and profile CD8⁺ T cells specific for the HCV epitopes ATDALMTGY_{NS3:1435–1443} (A1₁₄₃₅) or ALYDVVTKL_{NS5:2594–2602} (A2₂₅₉₄). To evaluate the effects of current and past HCV infection on the immune response against other viruses, we isolated from the same subjects CD8⁺ T cells specific for CMV (A2 - NLVPMVATV_{pp65:495–504}) and Flu (A2 - GILGFVFTL_MP:58–66). Three non-HCV-exposed subjects were included as controls. The dataset we generated allowed us to determine biologically relevant differentially expressed genes (DEG) and pathways influenced by HCV infection in HCV-, CMV- and Flu-specific CD8⁺ T cells in chronic and resolved HCV infection ( Figure 1A ).

After quality control and filtering out doublets and other cells not annotated as CD8⁺ T cells or not positive for Dextramer surface staining, we profiled a total of 5,121 A1₁₄₃₅-specific; 1,268 A2₂₅₉₄-specific; 10,177 CMV-specific; and 10,137 Flu-specific CD8⁺ T cells. The t-Distributed Stochastic Neighbor Embedding (tSNE) plot shows that the CD8⁺ T cells clustered according to virus specificity ( Figure 1B ). No significant differences were observed in the frequencies of the HCV clusters among cHCV and rHCV subjects (data not shown).

We evaluated the differences between HCV-specific CD8⁺ T cells in cHCV versus rHCV. We detected a total of 142 DEGs, where 57 genes were upregulated in cHCV and 87 were upregulated in rHCV ( Figure 1C , Supplementary Table 2 ). The genes upregulated in cHCV enriched biological processes and pathways mainly related to cytotoxicity, such as cell death induced by granzymes, cytolysis, and cell killing, and other processes related to lymphocyte activation, such as IL-12 signaling and IFN- γ response. The genes upregulated in rHCV were related to cell viability (apoptosis, cell senescence, stress response) and cell differentiation and activation ( Figure 1D ). Our findings suggest that after spontaneous resolution, the cytotoxicity of HCV-specific CD8⁺ T cells is significantly reduced.

For this study, we selected subjects that expressed the human leukocyte antigen (HLA) alleles HLA-A*0101 and HLA-A*0201, and we isolated CD8⁺ T cells specific for two HCV epitopes: A1₁₄₃₅ and A2₂₅₉₄. We evaluated the transcriptome of these cells by comparing cHCV to rHCV groups. In cHCV, we found 167 DEG in A1₁₄₃₅-specific CD8⁺ T cells (62 up and 105 downregulated), and 114 DEG in A2₂₅₉₄-specific CD8⁺ T cells (45 up and 69 downregulated) ( Figure 2A , Supplementary Table 2 ). The DEGs of immunological relevance in A1₁₄₃₅- and A2₂₅₉₄-specific CD8⁺ T cells are shown in Figures 2B, C .

The genes of the A1₁₄₃₅-specific CD8⁺ T cells upregulated in cHCV were related to cytotoxicity ( Figure 2D ), while the genes upregulated rHCV were associated with cell differentiation and viability, such as apoptosis and senescence ( Figure 2E ). In contrast, the genes of the A2₂₅₉₄-specific CD8⁺ T cells upregulated in cHCV subjects were involved in antigen processing and presentation, and PD-1 and CD28 signaling ( Figure 2F ), whereas the genes upregulated in rHCV were related to cytotoxicity ( Figure 2G ).

We used the Ingenuity Pathway Analysis (IPA, Qiagen) platform to identify the most relevant pathways based on the balance between up and downregulated genes during chronic HCV infection ( Figure 2H ). We found that A1₁₄₃₅-specific CD8⁺ T cells in cHCV display a robust upregulation of cytotoxicity, apoptosis of liver cells, and chemotaxis of antigen presenting cells, with decreased TCR signaling and focal adhesion kinase (FAK) signaling. The A2₂₅₉₄-specific CD8⁺ T cells demonstrated strong TCR signaling and chemotaxis of myeloid cells. Both cell types showed decreased cell viability.

Among the most relevant genes encoding proteins with immunological function ( Figure 2I , Supplementary Table 3 ), IL7R and KLF6 are upregulated in rHCV in A1 and A2 cells, while IL32 and KLF2 are upregulated in rHCV in the A1 group. We also found CCL4 upregulated in both A1 and A2 cells in cHCV, while TGFB1 is upregulated in the A1 group, and XCL1 is upregulated in the A2 group. Interestingly, CXCR3 is upregulated in the A1 group in cHCV and upregulated in the A2 group in rHCV.

Due to the significant functional changes observed in HCV-specific CD8⁺ T cells from cHCV compared to rHCV subjects suggested by our dataset, we investigated the overall HCV-specific CD8⁺ T cell phenotypes. For that purpose, we used the annotation established by Hao et al. (14) to determine naïve, central memory, and effector memory CD8⁺ T cells, and the dataset published by Doering et al. (15) was used to annotate CD8⁺ exhausted T cells ( Table 2 ). The A1 and A2 HCV-specific CD8⁺ T cells displayed similar phenotype changes when comparing cHCV and rHCV ( Supplementary Figures 2A, B ). Resolved HCV subjects showed a higher proportion of HCV-specific naïve and central memory T cells, while effector memory T cells were predominant in cHCV. There was no significant difference in T exhausted HCV-specific T cells between cHCV and rHCV subjects.

Direct-acting antiviral (DAA) therapy is highly effective, with standard cure rates above 95% (22). We evaluated three subjects, positive for the HLA-A*0101 allele, that were successfully treated with DAA therapy for 12 weeks. We collected blood samples from these subjects before therapy started (Pre-tx), at 2-4 weeks of therapy corresponding with viral clearance (Early-tx), towards the end of the therapy (Late-tx), and 12 weeks after DAA therapy conclusion (Post-tx). We isolated A1₁₄₃₅-specific CD8⁺ T cells and evaluated them by scRNAseq. We profiled 105 cells in Pre-tx, 533 cells in Early-tx, 231 cells in Late-tx, and 394 cells in Post-tx. We identified the DEGs resulting from DAA therapy in comparison to the Pre-tx samples ( Figure 3A , Supplementary Table 2 ). In the initial stages, DAA therapy modified the expression of only 10 genes (2 up, and 8 downregulated). The most relevant downregulated genes were GZMM, GNLY, and CCL5 ( Figure 3B and Supplementary Tables 2 , 3 ). In later stages, we detected 44 DEGs (11 up and 33 downregulated genes), where the most relevant downregulated genes were CCL4, GZMH, GZMM, CX3CR1, CCL5, GNLY, GZMA, GZMB and PRF1 ( Figure 3C , Supplementary Tables 2 and 3 ). Most of these genes remained downregulated after completion of DAA therapy ( Figure 3D , Supplementary Table 2 ), which showed 56 DEGs (23 up and 33 downregulated).

Downregulation of cytotoxicity starts in the early stages of DAA therapy and is still detected 12 weeks after DAA therapy was completed ( Figures 3E–G ). In the late phases of DAA therapy, there is also downregulation of antigen processing and presentation, and after cure, we also detected inhibition of complement cascade. Twelve weeks after DAA-treatment, we detected upregulation of events related to RNA translation ( Figure 3H ). We also observed the expression of genes related to the effector functions of CD8⁺ T cells ( Figure 3I ). Although some of them were not part of the DEG lists, they showed decreased expression from the Pre-tx to Post-tx groups. Of note, IFNG showed a peak in the expression in the early phases of the therapy, while IL2 was not detected in the chronic subjects and started to be expressed at the beginning of the therapy, which suggests an activation of these cells in the early phases of therapy. Moreover, we observed a long-lasting gene expression inhibition induced by DAA therapy on 7 genes early during treatment (CCL5, MYL6, GZMM, NKG7, KLRD1, IGLV2-14, and GNLY) and on 11 genes late in treatment (HLA-DRB5, ANXA4, FCRL6, PRF1, GZMA, PLEK, ZEB2, GZMB, FGFBP2, GZMH, and CCL4) ( Supplementary Table 3 ). These 18 genes remained downregulated at 12 weeks after the conclusion of DAA therapy.

We compared the most relevant biological events detected in rHCV and DAA-mediated resolution of HCV. TCR signaling is only activated in A1 HCV-specific CD8⁺ T cells from rHCV, which also displays upregulation of cell viability. A1 HCV-specific CD8⁺ T cells from DAA-mediated resolution of HCV demonstrated upregulation of EIF2 signaling, which is related to cytoplasmic translation and elongation, and downregulation of degranulation, immune response, and synthesis of reactive oxygen species ( Figure 3J ). Altogether, our findings suggest that the DAA therapy dampens HCV-specific CD8⁺ T cell effector functions.

We also evaluated the phenotypes of CD8⁺ T cells during DAA therapy and as seen in rHCV, there was a small but statistically significant increase in the proportion of CD8⁺ naïve T cells in successfully treated subjects ( Figure 4A ). We did not detect changes in A1 HCV-specific proliferating or TCM phenotypes rates during DAA therapy ( Figures 4B, C ). However, similar to rHCV, we detected a significant reduction in the proportion of TEM during DAA therapy ( Figure 4D ). We also detected a slight decrease in CD8⁺ exhausted T cells during treatment ( Figure 4E ). Of note, we detected a decrease in the expression of HAVCR2 (Tim-3), EOMES, and LAG3 ( Figure 4F ), which characterize terminally exhausted T cells.

The impact of HCV infection on the immune defense against other viruses is not fully understood. Here we evaluated the immune status of CMV-and Flu-specific CD8⁺ T cells in cHCV and rHCV subjects in comparison to non-HCV-exposed controls (CTRs).

We found 44 DEGs in CMV-specific T cells from cHCV subjects (19 up and 25 downregulated), and 78 DEGs (49 up and 27 downregulated genes) in rHCV subjects ( Figure 5A , Supplementary Table 2 ). The most relevant differentially expressed genes are shown in Figures 5B, C . Additionally, we identified 105 DEGs (30 up and 75 downregulated) when comparing cHCV to rHCV directly ( Supplementary Figure 3A ). The 19 genes found upregulated in CMV-specific CD8⁺ T cells in cHCV (vs CTRs) enriched processes related to cell killing, positive regulation of immune response, and TNF-α signaling, suggesting a more pro-inflammatory profile ( Figure 5D ). The 49 genes upregulated in rHCV subjects (vs CTRs) enriched pathways controlling the generation of second messenger molecules, antigen processing and presentation, TNF-α signaling, TCR signaling, and cell differentiation ( Figure 5E ). We also detected an overall inhibition of oxidative phosphorylation in CMV-specific T cells in both cHCV and rHCV subjects when compared to controls ( Figure 5F , Supplementary Figure 3B ), with inhibition of several genes involved in the mitochondrial respiratory chain ( Supplementary Figure 3C ). Of note, several genes with immunological relevance were identified among the DEGs ( Figure 5G , Supplementary Table 3 ). We found GNLY and IFNG downregulated and XCL1, XCL2, KLRC2, and LGALS1 upregulated in cHCV. GZMB, GZMK, KLRB1, and KLRC4 were downregulated in rHCV, while KLF6 was upregulated. CD127 (IL7R) and FCGR3 (CD16) were downregulated in both cHCV and rHCV compared to controls.

We found 47 DEGs in Flu-specific CD8⁺ T cells from cHCV subjects (36 up and 11 downregulated), and 32 DEGs (27 up and 5 downregulated) from rHCV cases in comparison to Flu-specific CD8⁺ T cells isolated from non-HCV-exposed controls ( Figure 6A , Supplementary Table 2 ). In a direct comparison between cells from cHCV and rHCV, we detected 46 DEGs, where 23 genes were upregulated and 23 downregulated ( Supplementary Figure 4A ). The volcano plots in Figures 6B, C show the DEG distribution for cHCV and rHCV versus control. Overall, the upregulated genes of Flu-specific CD8⁺ T cells in cHCV were involved in cytotoxicity and interferon response ( Figure 6D ), while the genes upregulated in the rHCV subjects were associated with mTOR, TNF-α, and TGF- β signaling, and platelet activation and aggregation ( Figure 6E ). The downregulated genes found in Flu-specific cells from cHCV were related to oxidative phosphorylation, as observed in CMV-specific cells in cHCV ( Supplementary Figure 4B ). The main genes associated with this process were involved in the mitochondrial respiratory chain ( Supplementary Figure 4C ).

The differentially expressed genes with immunological relevance can be found in Figure 6F ( Supplementary Table 3 ). The gene expression of CCL4, GZMA, GZMB, GZMH, GZMK, ISG15, KLRB1, NGK7, and PRF1 was upregulated in Flu-specific CD8⁺ T cells from cHCV subjects, while the genes KLF6 and LGALS3 were downregulated. In Flu-specific cells isolated from rHCV subjects, the expression of LGALS3, TIMP1, and XCL1 was upregulated, while granzymes GZMH and GZMK were downregulated.

The expression of GNLY and LGALS1 was upregulated in cHCV and rHCV. Using IPA, we found that, besides cytotoxicity, Flu-specific CD8⁺ T cells from cHCV subjects also upregulated pathways related to T cell activation, reactive oxygen species (ROS) production, and apoptosis of leukocytes, whereas cells from rHCV subjects showed upregulation of pathways involved in cytokine storm and TCR signaling ( Figure 6G ).

Regarding the CD8⁺ T cell phenotype signature, we observed that CMV-specific cells from cHCV subjects have low rates for naïve and TCM cells, while displaying a higher rate for TEM ( Supplementary Figure 5A ). No differences were observed in the proportion of CD8⁺ exhausted T cells. Flu-specific cells from cHCV subjects also demonstrated low proportions of the naïve phenotype and an increased proportion of TEM. However, in contrast to CMV-specific cells, we found a significant increase in the frequency of CD8⁺ exhausted T cells that were specific to Flu during chronic HCV infection ( Supplementary Figure 5B ).