Viral sequence determines HLA-E-restricted T cell recognition of hepatitis B surface antigen

Abstract

The non-polymorphic HLA-E molecule offers opportunities for new universal immunotherapeutic approaches to chronic infectious diseases. Chronic Hepatitis B virus (HBV) infection is driven in part by T cell dysfunction due to elevated levels of the HBV envelope (Env) protein hepatitis B surface antigen (HBsAg). Here we report the characterization of three genotypic variants of an HLA-E-binding HBsAg peptide, Env_371-379, identified through bioinformatic predictions and verified by biochemical and cellular assays. Using a soluble affinity-enhanced T cell receptor (TCR) (a09b08)-anti-CD3 bispecific molecule to probe HLA-E presentation of the Env_371-379 peptides, we demonstrate that only the most stable Env_371-379 variant, L6I, elicits functional responses to a09b08-anti-CD3-redirected polyclonal T cells co-cultured with targets expressing endogenous HBsAg. Furthermore, HLA-E-Env_371-379 L6I-specific CD8⁺ T cells are detectable in HBV-naïve donors and people with chronic HBV after in vitro priming. In conclusion, we provide evidence for HLA-E-mediated HBV Env peptide presentation, and highlight the effect of viral mutations on the stability and targetability of pHLA-E molecules.

Fig. 1. Identification and validation of HBV peptides binding HLA-E.

a Thermal melting point (Tm) assessed by thermal shift assay screen of HBV peptides (Env_371-379 peptide, magenta triangle) and Mtb RLPA (positive control, gray circle). Peptides were added to refolded and purified HIV Gag_275-283-HLA-E*01:03 complexes (in PBS at 0.25 mg/mL) at a 60:1 molar ratio. Data were presented as mean values ± SD of triplicates. b HLA-E binding of HBV Env_371-379 variants, Mtb RLPA, and signal peptides (A80, B7, and Cw3) was measured by HLA-E upregulation at the cell surface of K562 cells transduced with single chain β₂m-HLA-E*01:01 or β₂m-HLA-E*01:03. Plotted are flow cytometry histograms after staining with 3D12 antibody (HLA-E antibody). Histograms are representative of two independent experiments, each conducted in triplicates (full dataset and gating strategy available in Supplementary Fig. 1). c Summary of t_1/2 and Tm of HBV Env_371-379 variants and signal peptides (Cw3 and A1) pHLA-E*01:03 complexes as determined by SPR and thermal shift assays in triplicate. All SPR experiments were performed at 25 °C. Data were presented as mean values ± SD of triplicates. Source data are provided as a Source Data file.

Fig. 2. Binding kinetics of wildtype and a09b08 ImmTAVs to HBV Env_371-379 pHLA-E*01:03 complexes.

a Binding curves for wildtype ImmTAV and all three HBV Env_371-379 pHLA-E*01:03 complexes. Binding was determined over a range of analyte concentrations from 37 nM to 37.6 µM. Insets: calculation of steady-state affinity, data were presented as mean values ± SD. b Binding kinetics for a09b08 ImmTAV and all three HBV Env_371-379 pHLA-E*01:03 complexes. Graphs show the mean of the raw data (blue) and the 1:1 fit (dotted red line). For the binding kinetic graphs, the ImmTAV molecule was flown over the chip as the analyte, at concentrations ranging from 0.313 to 5 nM. Kinetic constants were determined using a 1:1 Langmuir model. c Summary of K_D values and t_1/2 values of the wild-type and a09b08 ImmTAV. All experiments were performed at 25 °C in triplicate. Data were presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 3. Structural overview of the a09b08 TCR in complex with three HBV Env_371-379 pHLA-E complexes.

a Cartoon overview of the three TCR-pHLA complexes aligned on HLA-E*01:03 (TCR alpha chain in gray, TCR beta chain in green, HLA-E in wheat, B2M is brown, peptides: Env_371-379 in magenta, Env_371-379 (S3N) in red and Env_371-379 (L6I) in blue). b Close-up view of the three TCR-peptide interactions. TCR residues within 4 Å of the peptide are shown as sticks. Dotted lines indicate polar contacts. c Overlay of the three peptides, aligned on HLA-E, showing highly similar conformation in the complexes. d a09b08 TCR contacts to pHLA-E mapped onto the truncated TCR sequence. A wild-type TCR sequence is provided to show the mutations introduced by affinity maturation. CDR residues are underlined and residues within 4.1 Å of the peptide are highlighted in bold with peptide positions indicated below. Residues highlighted in blue and orange indicate positions that are within 4.1 Å of the HLA-E helix 1 and HLA-E helix 2, respectively. e a09b08 TCR contacts to peptide mapped onto the peptide sequences. Peptide residues within 4.1 Å of the TCR alpha chain are highlighted gray, and those of the TCR beta chain are colored green.

Fig. 4. The a09b08 ImmTAV molecule elicits T cell responses against cells displaying HBV Env_371-379 pHLA-E complexes.

a ELISPOT assay measuring dose-dependent IFN-γ release induced by a09b08 ImmTAV in co-cultures of PBMC from 3 HBV-naïve donors and THP-1-E*01:03 cells pulsed with Env_371-379 (purple line), Env_371-379 (S3N) (red line) and Env_371-379 (L6I) (blue line) peptides (10 μg/mL). b IFN-γ ELISPOT assays showing titratable activation of PBMC from 3 HBV-naïve donors by a09b08 ImmTAV (1 nM) in the presence of THP-1-E*01:03 cells pulsed with the indicated amounts of peptide. Controls (a, b clear symbols) include PBMC + target cells (no ImmTAV; unpulsed or pulsed with 10 μg/mL peptide), and PBMC + ImmTAV + target cells (unpulsed). c ELISPOT assay measuring dose-dependent IFN-γ release from HBV-naïve donor PBMC in the presence of HepG2 targets stably transfected with minigenes encoding the indicated peptides Env_371-379, Env_371-379 (S3N), and Env_371-379 (L6I). Untransfected HepG2 (no minigene) and PBMC + minigene target cells (no ImmTAV) were included as controls (clear symbols). d Cumulative IFN-γ responses induced by a09b08 ImmTAV in PBMC cocultured with HepG2 Env_371-379 (L6I) (blue) or Env_371-379 (S3N) (red) minigene targets in the presence or absence of blocking mAbs against HLA-E or HLA-A*02:01 (10 µg/mL); ****p = <0.0001, two-way ANOVA. Results (a–d) are representative of one of three PBMC donors tested in triplicate. Data were presented as mean values ± SD. Average EC₅₀ are indicated in the inset at the top left corner of the figures (a–c). All donor EC₅₀ values and averages are displayed in Supplementary Table 5. Source data are provided as a Source Data file.

Fig. 5. The a09b08 ImmTAV elicits antigen-dependent T cell killing of target cells.

Apoptosis (measured by IncuCyte assay, see methods) of Env_371-379 (L6I) Ag⁺ PLC/PRF/5 target cells wildtype (a) or HLA-E*01:03 lentiviral transduced (b), cocultured with HBV-naïve donor PBMC and a titration of a09b08 ImmTAV. Dashed vertical lines indicate timepoints from which target cell apoptosis begins (at 1 nM ImmTAV). Results (a, b) are representative of one of three PBMC donors, each tested in triplicate. Data were presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 6. The a09b08 ImmTAV activates T cells to eliminate HBV-transfected HepG2 cells.

a Surface expression levels of HLA-E on HepG2 cells were analysed by flow cytometry. HepG2 wildtype and HLA-E*01:03 overexpressing cells were transfected with a replication-competent HBV plasmid encoding genotype A2 (Env_371-379 (L6I)). b 24 h post-transfection, some wells of HepG2 wildtype cells were stimulated with IFN-γ (1 ng/mL) for 16 h. Cells were washed and cocultured with pan T cells from three donors (1:1 ratio) with or without a09b08 ImmTAV, HBV HLA-A*02:01 ImmTAV or Mtb RLPA HLA-E ImmTAB (negative control) at 1 and 10 nM. Culture supernatants were harvested on day 4 and day 6 according to the schematics shown (modified from ref. ). Panel b is released under a Creative Commons Attribution-Non Commercial 4.0 International license (https://creativecommons.org/licenses/by-nc/4.0/deed.en). Levels of IFN-γ (c) and granzyme B (d) in the culture supernatants at day 4 were quantified using MSD. Data represents the mean ± SEM of triplicates (n = 3). HBeAg (e, f) and HBsAg (g, h) levels in the culture supernatant at day 4 and day 6 were quantified using ELISA. Data were represented as mean ± SEM of duplicates from three donors. Significant differences are *p < 0.05, **p < 0.001, ****p < 0.0001 by ANOVA followed by Tukey’s post hoc test. Flow cytometry gating strategy and additional datasets are shown in Supplementary Figs. 11, 12, 13. Source data are provided as a Source Data file.

Fig. 7. Detection of HBV antigen-specific T cells by pHLA-E multimers.

a T cells from people with CHB (n = 10) and HBV-naïve donors (n = 5) were expanded in vitro by stimulation with streptavidin magnetic beads conjugated to Env_371-379 (L6I)-pHLA-E and anti-CD28 for two weeks and stained with dextramer-E and antibody cocktail. Dot plots show the percentage of CD8⁺ dextramer-E⁺ cells (dextramer positive cells were gated as live CD3⁺CD4^-CD8⁺ singlet cells). b Quantification of % of CD8⁺ dextramer-E positive cells. Statistical analyses were performed using two-tailed unpaired t-test in GraphPad Prism software (version 10.0.2), data represented as dot plot ± SEM (n = 5 for HBV-naïve donors (in black); n = 10 for people with CHB (in purple)). Source data are provided as a Source Data file.