HBV-specific T-cell function is nonenhanced by tenofovir-induced decline in HBV viremia or HBsAg titer in chronic hepatitis B

Abstract

Background: Chronic hepatitis B is associated with virus-specific and global T-cell dysfunction. We hypothesized that therapeutic reduction in serum HBV DNA, ALT, and HBsAg would restore HBV-specific T-cell function and modify T-cell regulatory phenotype, with associated posttreatment ALT flare.

Methods: HBV-specific T-cell lymphoproliferative responses and global T-cell phenotype were prospectively examined at baseline, weeks 24, 48, 192, 216, and 240 in 34 adults with immune-active chronic hepatitis B treated with 192 weeks of tenofovir alone (n=21) or combined with pegylated interferon (PegIFN) in the first 24 weeks (n=13). HBV-specific T-cell IFNγ responses at weeks 0, 24, and 48 were examined by ELISpot assay ex vivo in 24 patients. Posttreatment flare was defined by serum ALT >5 times the upper limit of normal.

Results: Tenofovir therapy did not promote sustained induction of HBV-specific T-cell proliferative responses, regardless of PegIFN therapy or decreased serum HBsAg, HBV DNA, or ALT levels. Instead, HBV-specific T-cell IFNγ responses declined significantly by 48 weeks of therapy (p=0.008). Posttreatment ALT flare was associated with higher baseline %PD1+/CD8 (p=0.019), %PD1+/CD4 (p=0.039), and %CTLA4+/CD4 (p=0.003) T cells compared to non-flares, but without associated HBsAg loss or increased HBV-specific T-cell responsiveness.

Conclusion: HBV-specific T-cell function was not restored after 192 weeks of tenofovir therapy and did not correlate with HBsAg levels before, during, or after therapy. Baseline global T-cell regulatory phenotype was a predictor for ALT flare post-therapy without associated HBsAg decline. These findings support the need for more novel immune-modulatory approaches to enhance HBV-specific T-cell responsiveness.

Keywords: antiviral therapy; functional cure; hepatitis flare; immune pathogenesis; interferon.

Graphical abstract

FIGURE 1

HBV-specific LPR responsiveness is not enhanced during antiviral therapy. (A) Lymphoproliferative response (LPR) to HBV (combined HBV S, PreS, Core, PreC, Reverse Transcriptase) and Flu peptide pools at baseline and on antiviral therapy at weeks 24, 48, and 192 time points. (B) HBV-specific LPR responsiveness in subgroups based on antiviral regimen (TDF, TDF+PegIFN) and HBeAg status (HBeAg+ and HBeAg−). LPR responses are shown in top panel as percentages of LPR responders in bar graphs and in bottom panel as stimulation index in box and whisker plots with vertical error bars (25% and 75% interquartile ranges), median values (horizontal line within the box), mean values (“x” within the box) and outliers (individual circles). p values comparing the 4 time points were calculated using generalized linear models with a logistic link, fit with a generalized estimating equation approach. %HBV-specific LPR responders include participants with a positive response to at least 1 HBV peptide pool (SI ≥3.0). Sum HBV-specific LPR (SI) represents the sum of SI to all HBV peptide pools. Abbreviations: LPR, lymphoproliferative response; PegIFN, pegylated interferon; SI, stimulation index; TDF, tenofovir.

FIGURE 2

HBV-specific LPR does not increase in magnitude during antiviral therapy nor correlate with circulating levels of HBsAg or HBeAg. (A) HBsAg titer is correlated with HBV-specific LPR (top HBV S, bottom sum HBV). (B) HBeAg titer is correlated with HBV-specific LPR (top HBV C, bottom sum HBV). (C) HBsAg titers (orange line) are superimposed with HBV S-specific lymphoproliferative response (LPR) in stimulation index (SI) in bar graphs. Crimson bars indicate significant HBV-specific LPR with SI ≥3.0. Gray bars indicate no significant HBV S-specific LPR with SI <3.0. The top panel shows 9 subjects with >1 log decline in HBsAg titers during antiviral therapy. The bottom panel shows 8 subjects with <1 log decline in HBsAg titers during antiviral therapy. The green asterisk indicates time points without available immune assays. Time points on antiviral therapy are shaded in green (TDF alone) or pink (TDF/PegIFN). *Below the bar graphs indicate time points without assays. Abbreviations: LPR, lymphoproliferative response; PegIFN, pegylated interferon; TDF, tenofovir.

FIGURE 3

HBV-specific T-cell cytokine responses are not enhanced by antiviral therapy. (A) Bar graphs showing median IFNg responses to HBV or Flu peptide pools in IFNg ELISpot measured in 24 patients at weeks 0, 24, and 48 on antiviral therapy. Error bars indicate interquartile ranges (75%, 25%), with p values comparing baseline response to those in weeks 24 or 48 by Wilcoxon signed-rank test. (B) Bar graphs show HBV-specific IFNg responses in spot-forming units (SFU) per million PBMCs measured by ELISpot at weeks 0, 24, and 48 for each of the 24 HBRN participants. Responses to each HBV peptide pool are represented by different colors, with the scale as shown for the y-axis on the left. Line graphs are overlaid to show titers of HBsAg (orange line graph) and HBV DNA (blue line graph), with the scale in log IU/ml indicated on the y-axis on the right. Red asterisk (*) marks 4 participants with >1 log decline in HBsAg titer by week 48, including 1Te+ (2.3 log), 12Te+ (2.1 log), 3Te− (1.6 log), and 4TPe+ (1.4 log). Sum HBV-specific IFNg+ SFU/M decreased by week 48 compared to baseline in 17/24 (71%) but not in 7/24 (29%). (C) Scatter plots comparing HBV S-specific and sum HBV IFNg responses (y-axis) with serum levels of HBsAg and ALT (x-axis). (D) Scatter plots comparing HBV Core-specific and sum HBV-specific IFNg responses (y-axis) with serum levels of HBeAg and ALT (x-axis). Relevant Spearman correlation (r _s) and associated p values are provided. Abbreviations: PBMC, peripheral blood mononuclear cells; SFU, spot-forming unit.

FIGURE 4

Circulating T-cell frequency and regulatory phenotype before, during, and after antiviral therapy. (A) Clinical and virological correlations with circulating T-cell frequency and phenotype at baseline: heatmap shows the Spearman correlation coefficient (r _s) on the left panel and associated p values on the right panel. (B) Scatter plots with %CD127+/CD8 T cells on the x-axis with y-axis showing ALT, HBsAg, and HBV DNA levels at baseline, on-therapy (weeks 24, 48, 192), and off-therapy (weeks 216 and 240 for those who stopped antiviral therapy), with associated Spearman correlation coefficient and p values shown. (C/D) Median frequencies for CD4 T cells (C) and CD8 T cells (D) with their expression of PD1, CTLA4, FoxP3, and CD127 are shown as box and whisker plots with vertical error bars (25% and 75% interquartile ranges), median values (horizontal line within the box), mean values (“x” within the box) and outliers (individual circles). p values with horizontal black bars compare time points at baseline (week 0) and on antiviral therapy (weeks 24, 48, and 192). p values with horizontal gray bars compare time points at end of treatment (week 192) and after treatment (weeks 216 and 240) for all (n=34) with further subgroup based on TDF alone (n=21) and TDF+PegIFN combination therapy (n=13). Baseline and on-treatment results were derived from all subjects, whereas post-treatment values were derived from 16 subjects who stopped antiviral therapy at week 192 (7 on TDF, 9 on TDF+PegIFN). For each immune variable, a generalized estimating equation (GEE) was used to assess the differences over time. Time was treated as a categorical covariate, with a working exchangeable correlation to account for the repeated measures from the same subject. The model was fitted separately for the treatment phase (weeks 0, 24, 48, 192) and the withdrawal phase (weeks 192, 216, 240). Pairwise comparisons were done using least-squares means and were not adjusted for multiple comparisons. The analyses were repeated within each treatment arm. Red asterisk (*) indicates p value <0.05 compared to baseline time point. Blue asterisk (*) indicates p value <0.05 compared to week 192 (end of treatment) time point. Abbreviations: PegIFN, pegylated interferon; TDF, tenofovir.

FIGURE 5

HBV-specific T-cell response and phenotype in patients with ALT flare post-treatment cessation. (A) Bar graphs comparing T-cell proliferative and IFNg responses to HBV Core (top) and HBV S (bottom) in median stimulation indices (SI) and IFNg+ SFU/million PBMC, with error bars for 25% and 75% IQRs. LPR response 10 ALT flare and 8 Non-Flare participants. IFNg responses were examined in 9 ALT flare and 7 Non-Flare participants. (B) Scatter plots comparing percentages of T-cells expressing PD1, CTLA4, CD127, or FoxP3 from patients with post-treatment ALT Flares (red X) and Non-Flare (blue X) with sample sizes for Flare (F) and Non-Flare (NF) as follows: baseline (10F, 8NF), 24 weeks (9F, 6NF), 48 weeks (9F, 7NF), 192 weeks (9F, 7NF), 216 weeks (6F, 7NF), and 240 weeks (6F, 5NF for CD8; 4F, 6NF for CD4—NS or not sufficient for statistical test). p values by Mann–Whitney U are shown above relevant time points. Abbreviations: LPR, lymphoproliferative response; PBMC, peripheral blood mononuclear cells; PegIFN, pegylated interferon; SFU, spot-forming unit; TDF, tenofovir.