Degenerate Immunogenicity of an HLA-A2-Restricted Hepatitis B Virus Nucleocapsid Cytotoxic T-Lymphocyte Epitope That Is Also Presented by HLA-B51

Abstract

The recent identification of hepatitis B virus (HBV) epitopes restricted by multiple HLA alleles has greatly expanded the epitope repertoire available for T-cell-mediated therapeutic vaccine development. The HLA-B51-restricted peptide HBc19-27 is particularly interesting because it is located entirely within the HLA-A2-restricted HBc18-27 epitope. Here we show that HLA-B51-restricted cytotoxic T lymphocytes specific for HBc19-27 from a patient with acute HBV infection were also able to lyse HLA-B51-positive target cells pulsed with HBc18-27 and to produce gamma interferon when stimulated by that peptide, implying that HBc18-27 can be presented by HLA-B51 as well as by HLA-A2. These results demonstrate the concept of degenerate immunogenicity across HLA class supertype boundaries in a human viral disease setting. In addition, they could facilitate the development of an epitope-based therapeutic vaccine to terminate chronic HBV infection that could provide a broad and diverse population coverage with a single peptide.

It has been shown previously that the peripheral blood cytotoxic T-lymphocyte (CTL) response to hepatitis B virus (HBV) is polyclonal and multispecific in patients with acute viral hepatitis (2, 7–9, 12). The CTL response persists indefinitely after recovery, and it is maintained by continued antigenic stimulation by residual virus that persists, apparently harmlessly, in healthy convalescent individuals (11). In contrast, the CTL response to HBV is relatively weak in patients with chronic HBV infection, except during spontaneous disease flares or interferon (IFN)-induced recovery, when it is readily detectable (13). These observations have generated interest in therapeutic approaches to stimulate the CTL response to HBV in chronically infected patients.

The recent discovery that multiple class I alleles can recognize common sequence motifs (supermotifs) led to the identification of groups of alleles (supertypes) that recognize similiar motifs (15–18). For example, the B7 supertype (e.g., B*0702, B*3501, B*5301, B*5401, and B*5101) recognizes 8- to 11-mer peptides with P at position 2 and either A, I, L, M, V, F, W, or Y at the carboxy terminus, while the HLA-A2 supertype (e.g., A*0201, A*0204, A*0205, A*0206, A*0207, and A*6802) binds peptides containing L, V, M, I, T, A, or Q at position 2 and L, I, V, M, A, or T at the carboxy terminus.

Highly conserved HBV peptides that bind with high affinity to several members of a given supertype have the potential of being degenerate immunogens. Degenerate immunogenicity is defined as the capacity of a single peptide to be immunogenic in the context of multiple HLA molecules (15). Recognition of the same peptide by completely unrelated class I molecules can also theoretically occur if different HLA-binding motifs are nested within the peptide. An interesting example of this recognition is provided by the HBV core (HBc) region from positions 18 to 27 (1), which contains two nested epitopes, each capable of binding to a different class I supertype. In this case the nested recognition involves the HLA-A2 supertype-restricted HBc18-27 epitope (FLPSDFFPSV) and the HBc19-27 epitope (LPSDFFPSV), which binds multiple HLA-B7 supertype alleles, especially HLA-B51, with high affinity (18).

The precise nature of the HBc18-27-derived peptide that is bound by HLA-B51 is unresolved. To address this question, we analyzed the class I-restricted CTL response to HBc18-27 and HBc19-27 in an acutely infected HLA-A2-negative, HLA-B51-positive patient during and after resolution of disease, and we asked if the HLA-B51-restricted HBc19-27-specific CTL from the patient were able to recognize the HLA-A2-restricted peptide HBc18-27 in the context of HLA-A2 and HLA-B51.

CTL response to an HLA-B51-restricted HBc epitope during acute HBV infection.

Peripheral blood mononuclear cells (PBMC) of an HLA-A2-negative, HLA-B51-positive patient with acute HBV infection who ultimately resolved the infection were stimulated for 2 weeks with peptide HBc19-27, which binds with high affinity (50% inhibitory concentration [IC₅₀] of 9 nM) to multiple HLA-B7 supertype alleles, including HLA-B51 (4, 18). For these experiments, the synthetic peptide was added at 10 μg/ml, and recombinant HBcAg was added at 1 μg/ml as a source of T-cell help during the first week of stimulation. On days 3 and 10, 1 ml of RPMI 1640 with 10% (vol/vol) human AB serum and recombinant interleukin-2 (Hoffmann-La Roche, Inc., Nutley, N.J.) at 20 U/ml (final concentration) was added to each well. On day 7, the cultures were restimulated with peptide (10 μg/ml), recombinant interleukin-2 (20 U/ml), and 10⁶ irradiated (3,000 rad) autologous feeder cells. On day 14, the cells were tested for the ability to lyse autologous or allogeneic HLA-matched or HLA-mismatched ⁵¹Cr-labeled Epstein-Barr virus (EBV)-transformed B-lymphoblastoid cell lines that had been incubated overnight with synthetic peptides at 10 μg/ml in a 4-h ⁵¹Cr release assay exactly as previously described (4). Specific cytolytic activity was easily detected using peptide-pulsed autologous target cells as well as peptide-loaded allogeneic target cells that were matched with the effector cells only at the HLA-B51 locus (Fig. 1). Next, the HLA-B51-restricted, HBc19-27-specific CTL lines were tested for the ability to recognize endogenously synthesized nucleocapsid antigen by using allogeneic HLA-B51-positive target cells that had been infected with vaccinia virus that directs the synthesis of the HBV core protein by the cell (14). A significant level of specific cytolytic activity was detected (Fig. 1), demonstrating that HBc19-27 is generated by the intracellular processing of endogenously synthesized HBV core and that these CTL are likely primed in vivo during HBV infection. We also tested the ability of HBc19-27-specific CTL lines to produce IFN-γ following activation using reagents obtained from PharMingen (San Diego, Calif.) and a staining procedure previously well described (10). As shown in Fig. 2B, nearly 20% of the CD8⁺ T cells in the peptide-stimulated cultures produced IFN-γ when stimulated with the highest concentration of HBc19-27, and the fraction of IFN-γ producing CD8⁺ T cells decreased in a peptide dose-dependent manner.

FIG. 1.

Recognition of endogenously synthesized HBcAg by HBc19-27-specific CTL lines and HLA restriction analysis. PBMC from the patient were stimulated for 2 weeks with HBc19-27 and then tested for CTL activity against either autologous, partially HLA matched or mismatched peptide HBc19-27-loaded EBV-transformed B-lymphoblastoid target cells or allogeneic HLA-B51-positive target cells that were infected at a multiplicity of infection of 10 with wild-type (wt) or recombinant vaccinia viruses that express HBc as previously described (6). The effects/target ratio used was 50:1. When allogeneic cells were used, HLA-B51 was the only shared allele and the target cells were not positive for other members of the B7 and A2 superfamily.

FIG. 2.

Presentation of peptides HBc18-27 and HBc19-27 to HBc19-27-specific CTL. PBMC were stimulated for 2 weeks with HBc19-27, and then their cytolytic activity was measured against HLA-B51-positive, HLA-A2-negative target cells pulsed with various concentrations of HBc19-27 and HBc18-27 (A). Similarly stimulated PBMC were analyzed for simultaneous cell surface CD8 expression and intracellular IFN-γ production after 5 h of stimulation with various concentrations of both peptides (B), and the number of IFN-γ-producing cells was compared to the number of cells observed in the absence of peptides.

HLA-B51-restricted, HBc19-27-specific CTL also recognize HBc18-27.

Since the HLA-B51-restricted HBc19-27 epitope is contained entirely within an HLA-A2-restricted HBc18-27 epitope that binds to HLA-A2 with very high affinity (IC₅₀, 2.5 nM) (16), we asked if both of these peptides could be recognized by HLA-B51-restricted HBc19-27-specific T cells. We did not expect the CTL to recognize HBc18-27 since this peptide does not bind to HLA-B51 (IC₅₀, >25,000 nM) in liquid-phase binding assays just as HBc19-27 does not bind to HLA-A2 (IC₅₀, >25,000 nM) (4, 16, 18). Interestingly, however, the HLA-B51-restricted HBc19-27-specific CTL lysed HBc18-27- and HBc19-27-pulsed HLA-B51-positive target cells (Fig. 2A) and produced IFN-γ (Fig. 2B) equally efficiently at all peptide concentrations. Furthermore, the lytic activity and IFN-γ production profiles were nearly identical at the different peptide concentrations tested. These results contrast with the failure of HBc18-27 to bind to HLA-B51 in a liquid-phase peptide binding assay, raising the possibility that HBc18-27 may have been contaminated with the shorter peptide HBc19-27 or that it may have been converted to that peptide in vitro. Contamination of HBc18-27 with HBc19-27 and vice versa was excluded by reverse-phase high-pressure liquid chromatography (HPLC) and electrospray spectrometry of the original peptide stocks. As shown in Fig. 3, peptide HBc19-27 displayed one major peak and three minor peaks by HPLC and a molecular weight of 1,007.0 by mass spectroscopy. In contrast, HBc18-27 displayed one major peak and two minor peaks, all of which had different elution times than the peaks observed for HBc19-27, and it displayed a molecular weight of 1,154.2 by mass spectroscopy. Furthermore, electrospray spectrometry showed that there was no overlap in the mass of the HPLC peaks from HBc19-27 (from 860.4 to 1,007.8) and HBc18-27 (from 1,154.2 to 1,210.8) (data not shown).

FIG. 3.

HPLC chromatogram and mass spectrogram of HBc18-27 and HBc19-27. Peptide HBc19-27 displayed one major peak and three minor peaks by reverse-phase HPLC and a molecular weight of 1,007.0, determined by mass spectroscopy. In contrast, HBc18-27 displayed one major peak and two minor peaks and a molecular weight of 1,154.2. The minor HPLC peaks in the two peptide preparations were also analyzed by mass electrospray spectrometry and shown to represent unrelated minor species (data not shown).

We then tested the possibility that HBc18-27 might be trimmed to HBc19-27 by serum or cellular proteases in the CTL and IFN-γ production assays. In these experiments, we compared the abilities of HLA-B51-positive antigen-presenting cells (APC) to present the two peptides to HLA-B51-restricted HBc19-27-specific CD8⁺ T cells in the presence or absence of fetal calf serum, in the presence or absence of protease inhibitors, and using paraformaldehyde-fixed or unfixed autologous or allogeneic HLA-B51-positive target cells as APC. Specifically, allogeneic HLA B51-positive target cells were loaded with 10 μg of HBc18-27 or HBc19-27 per ml or no peptide overnight and then added to the HLA-B51-restricted HBc19-27-specific CTLs after they were washed extensively (six times). Five hours later, IFN-γ production of the CTL was measured exactly as previously described (10). In some experiments, APC were fixed with 1% paraformaldehyde for 30 min and then cultured overnight with the peptides in the presence or absence of a cocktail containing 10 different protease inhibitors (Complete; Boehringer Mannheim, Mannheim, Germany) with a wide spectrum of protease inhibitory activity. Table 1 shows that HBc18-27 was presented as efficiently as HBc19-27 to HLA-B51-restricted HBc19-27 CTL in the absence of serum, in the presence of protease inhibitors, and when the APC were fixed or not fixed. Furthermore, both peptides were identical in the ability to trigger peptide-specific IFN-γ production whether they were added to fixed APC for 15 min or 24 h before exposure to the CTL (Fig. 4) and also when they were added to fixed or unfixed APC over a range of concentrations spanning 5 orders of magnitude (Fig. 5). While it is still possible that these precautions did not completely prevent conversion of HBc18-27 to HBc19-27, the virtual identity of the peptide-specific IFN-γ responses to HBc18-27 and HBc19-27 under these conditions (Table 1; Fig. 4 and 5) suggests that artifactual processing of the longer peptide was not responsible for its ability to stimulate the HLA-B51-restricted HBc19-27-specific CTL.

TABLE 1.

Ability of HLA-B51-positive APC to present HBc18-27 (10 μM) and HBc19-27 (10 μM) to HLA-B51-restricted HBc19-27 CTL in the presence or absence of fetal calf serum, in the presence or absence of protease inhibitors, and using paraformaldehyde-fixed or unfixed autologous or allogeneic HLA-B51-positive target cells as APC

Serum	APC	Protease inhibitor	% Peptide-specific IFN-γ+ CD8+ cells
			HBc18-27	HBc19-27
+	Fixed	+	23	22
+		−	24	26
−	Fixed	+	21	22
−		−	19	23
+	Unfixed	+	27	29
+		−	26	29
−	Unfixed	+	22	22
−		−	21	23

FIG. 4.

Presentation after different APC loading times with peptides. Fixed HLA-B51 positive autologous APC (EBV-transformed B-lymphoblastoid cells) were loaded with HBc18-27 or HBc19-27 and cultured in serum-free medium in the presence of protease inhibitors as described in Materials and Methods. After different time periods of peptide loading, the APC were extensively washed and then added to HBc19-27-specific, HLA-B51-restricted CTL; peptide-specific (pep. spec.) IFN-γ production was measured after a 5-h stimulation.

FIG. 5.

Peptide presentation of fixed and unfixed APC. Fixed or unfixed HLA-B51-positive APC were loaded with different concentrations of HBc18-27 or HBc19-27 for one h in serum-free medium, extensively washed, and then added to the HBc19-27-specific, HLA-B51-restricted CTL. Peptide-specific (pep. spec.) IFN-γ production was measured after a 5-h stimulation and compared to IFN-γ production in the absence of peptides.

The foregoing results suggest that the HBc18-27 peptide binds very efficiently to HLA-B51 at the cell surface despite its failure to do so in the liquid-phase binding assay used in this study. This result is not entirely unexpected since HBV polymerase 354–363-specific HLA-B51-restricted CTL are able to recognize the epitope despite its failure to bind the allele in a liquid-phase binding assay (18). This apparent contradiction might be explained if the conformation of HLA-B51 at the cell surface or the presence of other cell surface molecules allows it to bind the longer peptide. Alternatively, it is conceivable that the T-cell receptor might stabilize the T-cell receptor–peptide–major histocompatibility MHC complex in this case as has been suggested for a DRB*1302-restricted tetanus toxoid peptide, where analogues with very unstable binding were still able to stimulate T cells equally well (5). Whatever the explanation, these results suggest that degenerate immunogenicity (e.g., the capacity of a given peptide to be immunogenic in the context of multiple HLA molecules) can occur across HLA supertype boundaries, in this case the HLA-B7 and HLA-A2 supertypes. Since it has been recently suggested that the HBc18-27 epitope can stimulate HLA class II restricted T-cell responses (3), the breadth of its reactivity across the boundaries of class I and II alleles and HLA supertypes strongly suggests that HBc18-27 should be included in the future design of multiple-epitope-based therapeutic vaccines.