Hepatitis B Virus--Specific and Global T-Cell Dysfunction in Chronic Hepatitis B

Abstract

Background & aims: T cells play a critical role in viral infection. We examined whether T-cell effector and regulatory responses can define clinical stages of chronic hepatitis B (CHB).

Methods: We enrolled 200 adults with CHB who participated in the National Institutes of Health-supported Hepatitis B Research Network from 2011 through 2013 and 20 uninfected individuals (controls). Peripheral blood lymphocytes from these subjects were analyzed for T-cell responses (proliferation and production of interferon gamma and interleukin 10) to overlapping hepatitis B virus (HBV) peptides (preS, S, preC, core, and reverse transcriptase), influenza matrix peptides, and lipopolysaccharide. T-cell expression of regulatory markers FOXP3, programmed death-1, and cytotoxic T lymphocyte-associated antigen-4 was examined by flow cytometry. Immune measures were compared with clinical parameters, including physician-defined immune-active, immune-tolerant, or inactive CHB phenotypes, in a blinded fashion.

Results: Compared with controls, patients with CHB had weak T-cell proliferative, interferon gamma, and interleukin 10 responses to HBV, with increased frequency of circulating FOXP3(+)CD127(-) regulatory T cells and CD4(+) T-cell expression of programmed death-1 and cytotoxic T lymphocyte-associated antigen-4. T-cell measures did not clearly distinguish between clinical CHB phenotypes, although the HBV core-specific T-cell response was weaker in hepatitis B e antigen (HBeAg)(+) than HBeAg(-) patients (percent responders: 3% vs 23%; P = .00008). Although in vitro blockade of programmed death-1 or cytotoxic T lymphocyte-associated antigen-4 increased T-cell responses to HBV, the effect was weaker in HBeAg(+) than HBeAg(-) patients. Furthermore, T-cell responses to influenza and lipopolysaccharide were weaker in CHB patients than controls.

Conclusions: HBV persists with virus-specific and global T-cell dysfunction mediated by multiple regulatory mechanisms, including circulating HBeAg, but without distinct T-cell-based immune signatures for clinical phenotypes. These findings suggest additional T-cell-independent or regulatory mechanisms of CHB pathogenesis that warrant further investigation.

Keywords: HBRN; IFN; IL10; LPS.

Figure 1. Effector T-cell responses to HBV and control conditions in subjects with chronic hepatitis B and uninfected HBV S vaccine recipients

(A) Comparison of lymphoproliferative and IFNγ responses in CHB patients (gray bars) and uninfected vaccinees (white bars). The antiviral effector T-cell responses from 200 chronic hepatitis B (CHB) participants and 17 vaccinees were examined by stimulating PBMC with HBV and flu matrix peptide pools, lipopolysaccharide (LPS) and phytohemagglutinin (PHA) in lymphoproliferation (LPR) assay with ³H-thymidine uptake and in IFNγ Elispot assay. LPR response is shown as median stimulation index (SI) and % positive LPR responder based on SI cutoff of 3.0. IFNγ response in Elispot assays is shown as median IFNγ spot forming units (SFU) per million PBMC (SFU/M) and % positive IFNγ responders with a cutoff of 62 SFU/M PBMC (2 standard deviations above background in media control wells for all subjects). (B) Sum HBV-specific LPR response (combined SI for all 6 HBV peptide pools) is correlated to LPR responses to Flu and LPS. (C) Sum HBV-specific IFNγ response (combined IFNγ responses for all 6 HBV peptide pools) is compared with IFNγ responses to Flu and LPS. Statistical significance was determined by Wilcoxon’s tests (A, top panel), Fisher’s Exact or Chi-Square (A, bottom panel), and Spearman rank correlation test for correlation coefficient and p-values (B and C). P-values <0.05 are considered significant and highlighted in red font.

Figure 2. Comparison of effector T-cell responses between baseline CHB phenotype groups

(A) T-cell responses to each HBV peptide, Flu matrix peptides and LPS are shown as %LPR or %IFNγ responders in 4 CHB phenotype groups: black bar (Immune Tolerant: IT, n=21), dark gray (HBeAg+ immune active: IA+, n=58), light gray (HBeAg− immune active: IA−, n=66), white bar (Inactive Carriers: IC, n=45); (B) %LPR and IFNγ responders to at least one HBV peptide pool in 4 CHB phenotype groups. Statistical significance was determined by non-parametric Chi-square test, p-values <0.05 are considered significant and highlighted in red font.

Figure 3. HBV-specific IL-10+ regulatory T-cell response is not induced in chronic hepatitis B

(A) Median IL-10+ SFU/million PBMC (upper) and %IL-10 responders (lower) in PBMC stimulated with 6 HBV peptide pools, Flu matrix peptides and LPS in 200 CHB patients. The cutoff for a positive IL-10 response for each peptide pool was defined as 122 SFU/million PBMC (2 standard deviations of background response in media control wells for all subjects). (B) Comparison of median IL-10+ SFU/million PBMC between CHB phenotype groups. (C) Scatterplot correlating Sum HBV-specific IL-10+ SFU/M (sum of IL-10 response to all 6 HBV peptide pools) and sum HBV-specific LPR (top) or IFNγ (bottom) responses. Statistical significance was determined by non-parametric Kruskal-Wallis test (B) and Spearman rank correlation test (C). P-values <0.05 are considered significant and highlighted in red font.

Figure 4. Positive correlations between different T-cell regulatory pathways in CHB without correlation with serum ALT or HBV DNA levels

Scatterplots with trend lines (in red) correlate immune regulatory parameters with (A) each other and (B) with clinical parameters (log serum ALT and HBV DNA). Statistical significance was determined by non-parametric Spearman rank correlation test. P-values below 0.05 are considered significant and highlighted in red font.

Figure 5. Increased HBV-specific and non-HBV-specific T-cell proliferation with PD-1 or CTLA-4 blockade in vitro in PBMC from CHB patients

(A) Bar graph shows average background ³H thymidine uptake in log₁₀ counts per minutes (cpm) with standard deviations for PBMC cultured in 5 replicates in media alone with and without αPDL1 or αCTLA4 blocking antibodies in the absence of HBV or Flu peptides. (B) LPR response by SI in PBMC stimulated for 7 days with HBV or Flu peptides, without or with αPDL1 or αCTLA4. 19 CHB and 14 uninfected control subjects were examined. Changes from a previous non-response to a newly positive response with SI>3 are highlighted in red lines. (C) Comparison of LPR responses by SI to HBV S, Core and RT2 peptide in 11 HBeAg+ and 8 HBeAg− CHB patients cultured with media (left panel), with αPDL1 (middle panel) or with αCTLA4 (right panel). Significant differences are detected between HBeAg+ and HBeAg− group aPDL1 or aCTLA4 that were not apparent in media control stimulations without blockade. Statistical significance between the two groups was determined by Wilcoxon signed-rank test, with p-values below 0.05 considered significant and highlighted in red font.