Abstract
The aim of this study was to investigate the underlying mechanisms of the genetic association between certain HLA-DRB1* alleles and the immune response to HBsAg vaccination. Therefore, HBsAg peptide binding to HLA-DR molecules was measured in vitro by peptide binding ELISAs. Additionally, HBsAg-specific T cell reaction and cytokine profile of immune response were analysed ex vivo in ELISPOT assays and DR-restriction of T-cell proliferative responses was investigated with HBsAg specific T cell clones. In addition, we compared HBsAg specific T cell responses of 24 monozygotic and 3 dizygotic twin pairs after HBsAg vaccination. Our results showed that the peptide binding assays did not reflect antigen presentation in vivo. DR alleles associated with vaccination failure like DRB1*0301 and 0701 efficiently presented HBsAg peptides. In 11 of 24 investigated monozygotic twin pairs we observed pronounced differences in the recognition of HBsAg peptides. This study indicates that HLA–DR associations with HBsAg vaccination response are not caused by differences in peptide binding or by a shift in the Th1/Th2 profile. Our findings strongly argue for differences in the T cell recognition of peptide/MHC complexes as the critical event in T cell responsiveness to HBsAg.
Keywords: HLA-DR, HBsAg, peptide presentation, T cell response
Introduction
Hepatitis B vaccination programs are currently carried out in many countries worldwide and have been highly effective in decreasing the numbers of new HBV infections [1]. However, approximately 5% of healthy Caucasion adults fail to develop anti-HBs antibodies after a regular vaccine course [2,3]. In a twin vaccination study more than 60% of the observed variability in anti-HBs responses were explained by genetic factors [4]. Most of the genes involved in the regulation of HBsAg responsiveness are still unknown. Results from the twin vaccination study have suggested that approximately 40% of the genetic contribution is determined by genes within the major histocompatibility complex (MHC) on chromosome 6 [4]. Several case control studies that have compared frequencies of HLA-class II antigens among HBsAg responders and nonresponders have shown that the HLA-DRB1* alleles 3 and 7 confer a high risk for vaccination failure [2,5,6], whereas the alleles 1, 11, 13 and 15 are correlated with a good immune response to HBsAg [6–8]. Several groups have demonstrated that vaccine antibody low or nonresponders fail to develop HBsAg specific Th cell activation in vitro [9,10]. MHC class II molecules mediate Th cell activation by binding of antigenic peptides. These MHC/peptide complexes are presented on the surface of antigen presenting cells (APC) to Th cells. T cell nonresponse can result from the failure of a given MHC class II molecule to bind a T cell epitope from the antigen (i.e. presentation defect) or from the lack of specific T cells recognizing a particular MHC/peptide complex (i.e. a hole in the T cell repertoire). Both defects have been proposed to explain nonresponse to HBsAg vaccination in humans [7,9,10]. In the murine system nonresponse apparently is due to a presentation defect of the H-2 s and H-2f haplotypes. In humans systematic studies investigating the interaction of HBsAg peptides with HLA-DR molecules are lacking.
The immunogenetics of HBsAg responsiveness are a valuable model for immune responsiveness to viral proteins. Understanding the mechanisms of HBsAg presentation and recognition will help to investigate more complex disease associations.
We carried out competitive peptide binding ELISAs to determine in vitro the binding of overlapping peptides covering the hepatitis B surface protein to the different HLA-DR molecules. To confirm the results of these tests at the cellular level we analysed T cell responses of HBsAg vaccinated probands to seven of these HBsAg peptides in ELISPOT assays. In participants of a large twin vaccination trial we investigated HBsAg peptide recognition by peripheral blood mononuclear cells (PBMC) in mono- and di-zygotic twins. Taken together our results provide further evidence that recognition of peptide/MHC complexes rather than binding of antigenic peptides to MHC class II molecules is important for immune responsiveness to HBsAg.
Materials and methods
Probands
Details of the twin vaccination trial have been described elsewhere [4]. The study was conducted with written informed consent of all probands involved and was performed according to the declaration of Helsinki and ICH Guidelines of Good Clinical Practice. Local ethics committee approval was obtained. Mono- and dizygotic twins simultaneously received three doses of a combined HAV/HBV vaccine (Twinrix, GlaxoSmithKline, Munich, Germany) in the deltoid muscle according to a standard HBsAg vaccination protocol (0-, 1-, and 6-month schedule). Peripheral blood was drawn 4 weeks after the third vaccination, PBMC were isolated by standard density gradient centrifugation over Ficoll-Hypaque (1077 g/ml; Biochrom Ag Seromed, Berlin, Germany) and HBsAg antibody titre was determined. HLA-DRB1*-typing was done by nested PCR amplification with sequence-specific primers [11].
Competitive peptide binding ELISA
74 overlapping 15-mer peptides covering the whole HBV envelope protein (HBV subtype adw) were synthesized on an Advanced ChemTech 396 Peptide Synthesizer (Advanced ChemTech, Louisville, USA). Peptides were analysed by HPLC and only peptides with an average purity >70% (45 peptides) were tested using high flux DR/peptide binding assay. For small HBsAg position 86–102, 132–167, and 191–202 no peptides were tested due to difficulties with peptide synthesis. HLA-DR molecules were isolated from homozygous EBV-transformed B cell lines (HOM-2 for DRB*0101; WT49 for DRB1*0301, EKR for DRB1*0701, BM9 for DRB1*0801, SWEIG for DRB1*1101, WT-47 for DRB1*1302 and PGF for DRB1*1501) as described [12]. For the peptide binding ELISAs [13] purified DR-molecules were incubated with a biotinylated CLIP peptide (amino acid position 89–101) as competitor and the HBsAg peptide of interest in varying concentrations (0·001–100 µm) over night. The peptide/DR complexes were then transferred in plates coated with an anti-DR antibody (L234; ATCC, Rochester, USA). After 2 h at room temperature, plates were washed and a streptavidin-coupled alkaline phosphatase and 4-nitrophenylphosphate were added. Optical density was measured in a photometer and IC50 (inhibitory concentration: concentration of the specific peptide inhibiting CLIP peptide binding by 50%) values were calculated. Peptides with IC50 ≤ IC50 of the CLIP peptide were considered to be high affinity binders.
Isolation of CD4+ T cells
CD4+ T cells were separated from PBMC by positive isolation with MACS cell isolation kits (Miltenyi Biotec, Gladbach, Germany) according to the manufacturers instructions. After incubation with MACS CD4 immuno-magnetic beads cells were applied to a MACS separation column in a magnetic field, washed and – after removal of the magnetic field – CD4+ T cells were eluted from the column. T cells were mixed with an equal volume of 50% FCS (PAA Systems, Coelbe, Germany), 30% RPMI 1640 (Biochrom Ag Seromed, Berlin, Germany), and 20% DMSO (Sigma-Aldrich, München, Germany), and were frozen and stored in liquid nitrogen until use.
Culture of mature dendritic cells
Dendritic cells were generated from myeloid precursor cells of the peripheral blood according to standard procedures [14]. 1 × 107 PBMC/well were incubated for 30 min in PBS in 6-well plates (Costar, Germany). Plates were washed to remove non adherent cells. The remaining cells were cultured for seven days in X-Vivo 15 (Bio Whittaker, Maryland, USA) supplemented with 800 U/ml GM-CSF (Leucomax; Sandoz, Basel, Suisse) and 1·000 U/ml IL-4 (Strathmann Biotech, Hamburg, Germany) at 37°C, 5% CO2 and were fed every 2–3 days. After 6 days, cells were pulsed with antigen: tetanus toxoid (10 µg/ml, Behring Werke, Marburg, Germany) as positive control or one of the seven selected HBsAg peptides (25 µg/ml; Neosystem, Strasbourg, France), and one well was was not pulsed with antigen to serve as negative control. After 7 days, IL-1β (10 ng/ml), IL-6 (1·000 U/ml; both Strathmann Biotech, Hamburg, Germany), TNF-α (10 ng/ml; R & D Systems, Wiesbaden, Germany) and prostaglandin E2 (1 µg/ml, Sigma-Aldrich, München, Germany) were added for 48 h to induce final differentiation. Mature dendritic cells were used for ELISPOT assays without freezing.
INF-γ- and IL-10-ELISPOT assays
HBsAg peptide specific CD4+ T cell responses and their cytokine profiles were analysed by enzyme linked immunospot assays (ELISPOTs) for INF-γ (Th1) and IL-10 (Th2). 96-well plates with polyvinylidene-diflouride membrane bottom (Millipore MAIP, Eschborn, Germany) were coated overnight at 4°C with a primary antibody to INF-γ (8 µg/ml; clone MAB1-D1K) or IL-10 (12 µg/ml; clone JES3–9D7; both Mabtech, Nacka, Sweden). Frozen T cells were thawn, counted and 105 living cells (as verified by eosin staining) were used for the assay. After washing the ELISPOT plate with PBS, 105 CD4+ T cells and 104 peptide-loaded, mature dendritic cells per well were seeded out in duplicates or triplicates and incubated at 37°C, 5% CO2 in RPMI 1640 (Biochrom Ag Seromed, Berlin, Germany). As negative control 105 CD4+ T cells were incubated with 104 non peptide-loaded, mature dendritic cells. After 48 h cells were discarded, plates were intensively washed with PBS/0,05% Tween20, and 1 µg/ml of a secondary, biotin-conjugated antibody to INF-γ (clone 7B6-1) or IL-10 (clone JES3–12G8; both Mabtech, Nacka, Sweden) was added for 90 min. After washing, plates were incubated with a streptavidin/alkaline phosphatase (1 : 500 diluted Extravidin; Sigma-Aldrich, München, Germany) for 30 min. For development of the enzyme reaction, plates were washed and freshly filtered carbazole (Sigma-Aldrich, München, Germany)/H2O2 (Merck, Darmstadt, Germany)-solution was added for at least 1 h. Reaction was stopped under rinsing water and the number of spots (corresponding to the number of antigen specific cytokine producing T cells) was counted under a stereo binocular microscope (SV6; Zeiss, Oberkochen, Germany). The threshold for antigen recognition was set at twice the number of spots of the negative controls (dendritic cells and T cells without antigen).
Establishment of T cell lines and T cell clones
Antigen specific T cell lines were established by incubation of PBMC with the respective HBsAg peptide. IL-2 (20 U/ml, Roche Diagnostics, Mannheim, Germany) was added at day 3, 6, and 10. T cell lines were then tested for antigen specificity in proliferation assays. For cloning, T cell lines were diluted and seeded out in 60-well microtitre plates (Terasaki plates, Nunc, Wiesbaden, Germany) in concentrations of 0·5 T cells per well. Proliferating T cells were expanded in RPMI 1640 + 10% human AB serum + 20 U/ml IL-2.
Proliferation assays with T cell lines or T cell clones
To test peptide specificity of a given T cell line or clone 104 T cells and 3 × 104 Mitomycin C (Sigma-Aldrich, München, Germany)-treated autologous APC were incubated with either medium as negative control, 1 µg/ml PHA (Sigma-Aldrich, München, Germany) as positive control or 25 µg/ml peptide antigen in duplicates or triplicates. After 2 days tests were pulsed with 0·25 µCi/well 3H-thymidine/well (Amersham Pharmacia Biotech, Braunschweig, Germany) and harvested after 3 days. Antigen reaction was considered as positive with a stimulation index >3 (SI = cpm (counts per minute) antigen/cpm medium). For demonstration of DR-restriction the same tests were carried out with addition of 1/8 volume of an HLA-DR antibody (HB55, from cell culture supernatants). The antigen presenting DR molecule in heterozygous probands was determined by incubation with two different sorts of nonautologous, partially DR-matched APC (each with one irrelevant DR molecule and the other matching to one of the DR types of the proband).
Results
Binding of HBsAg peptides to DR molecules
The binding properties of different HLA-DR molecules for peptides derived from the small hepatitis B surface antigen were analysed by in vitro peptide binding ELISAs. Interestingly, there were no high affinity epitopes for the DRB1*03 molecule within the HBsAg region of the protein. In contrast, numerous epitopes for the DR7 molecule could be identified within the HBsAg region suggesting a different reason for the nonresponse correlated with this allele. Positions of the identified epitopes within the small HBsAg for the relevant DR molecules are given in Table 1.
Table 1.
IC50 values (inhibitory concentration: concentration of the specific peptide inhibiting CLIP peptide binding by 50%) determined in the competitive peptide binding ELISAs for some peptides of the small HBsAg that showed affinity for one or several DR molecules. IC50 values for the CLIP peptide self competition are given in the last row. Peptides with an IC50 ≤ IC50 of the CLIP peptide were considered to be high affinity binders (given in bold).
| Peptide sequence | Aa position | DRB1* 0101 | DRB1* 0301 | DRB1* 0701 | DRB1* 0801 | DRB1* 1101 | DRB1* 1501 | DRB1* 1302 |
|---|---|---|---|---|---|---|---|---|
| ENITSGFLGPLLVLQ | ″2–16 | 22 | 100 | 10 | 100 | 70 | 0·3 | 100 |
| AGFFLLTRILTIPQS | 17–31 | ″1 | 60 | 0·001 | ″0·03 | ″0·016 | 2 | ″0·3 |
| LTRILTIPQSLDSWW | 22–36 | 50 | 23 | 0·1 | ″9 | 16 | 1·2 | 45 |
| TIPQSLDSWWTSLNF | 27–41 | 100 | 100 | 4·2 | 100 | 100 | 6 | 100 |
| LDSWWTSLNFLGGSP | 32–46 | 60 | 100 | 0·15 | 100 | 40 | 18 | 100 |
| TSLNFLGGSPVCLGQ | 37–51 | ″0·003 | 100 | 1·5 | 100 | 100 | 1 | 100 |
| PICPGYRWMCLRRFI | 67–81 | ″7 | 30 | 3·8 | ″0·06 | ″0·014 | 0·5 | ″0·1 |
| YRWMCLRRFIIFLFI | 72–86 | 50 | 100 | 2·2 | ″4·5 | ″2 | 10 | ″4 |
| GMLPVCPLIPGSTTT | 102–116 | 100 | 100 | 1·2 | 100 | 100 | 25 | 100 |
| SVRFSWLSLLVPFVQ | 167–181 | ″3 | 40 | 0·1 | 60 | 15 | 0·5 | 70 |
| WLSLLVPFVQWFVGL | 172–186 | 65 | 100 | 2·2 | 100 | 100 | 1 | 100 |
| VPFVQWFVGLSPTVW | 177–191 | ″2·5 | 100 | 0·16 | ″2·5 | ″2·5 | 0·2 | 14 |
| GPSLYSIVSPFIPLL | 202–216 | ″5 | ″7 | 0·03 | ″3·5 | 16 | 0·1 | 100 |
| SIVSPFIPLLPIFFC | 207–221 | ″5·5 | 100 | 0·08 | ″1·8 | ″5 | 0·5 | 30 |
| CLIP (89–101) | ″0·7 | ″0·05 | 0·1 | ″1·5 | ″2·5 | 0·06 | ″2·5 |
T cell responses to HBsAg derived peptides
The peptide recognition patterns seen ex vivo in the ELISPOT assays did not confirm the peptide binding determined for the different DR molecules by the in vitro peptide binding assays. The seven tested HBsAg peptides including DR restrictions for five of them as observed in the peptide binding ELISAs are given in Table 2. CD4+ T cells of probands with given DR alleles did not recognize all peptides that were bound in the peptide binding ELISAs by the respective DR molecules and even responded sometimes to peptides that showed low or no affinity. Such contradictions are marked in grey for the examples shown in Table 3.
Table 2.
List of the seven HBsAg peptides used in the ELISPOT and proliferation assays. Positions in the HBsAg and the DR-restriction detected in the peptide binding ELISAs, as far as tested, are shown.
| Aa position in small HBsAg | DR-restriction | |
|---|---|---|
| Peptide 1 | 17–31 | DR7, 8, 11, 13 |
| Peptide 2 | 37–51 | DR1 |
| Peptide 3 | 67–81 | DR8, 11, 13 |
| Peptide 4 | 177–191 | DR11 |
| Peptide 5 | 202–216 | DR7 |
| Peptide 6 | 140–155 | not tested |
| Peptide 7 | 165–179 | * |
165–172 predicted to bind to DR11 and DR14 by [15].
Table 3.
Examples of INF-γ- and IL-10-ELISPOT assays with seven HBsAg peptides (given in Table 2) for monozygotic (MZ) and dizygotic (DZ) twin pairs. Capital letters in the second column correspond to proband initials. Results are given as the calculated mean values of spot forming cell numbers per well (SFC/well) from duplicate or triplicate tests. Also DR3-homozygous probands (last three rows) recognized several HBsAg peptides (these and the further contradictions to the peptide binding ELISAs are marked in italics). Antigen specific cytokine productions = twice the medium value are displayed in bold letters. Several MZ and DR-identical DZ pairs showed deviant peptide recognition. INF-γ production in general was much higher and more abundant than that of IL-10.
| IFN-γ-ELISPOT (SFC/well) | IL-10-ELISPOT (SFC/well) | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Zygositiy | Proband | Anti-HBs | DRB1* | ||||||||||||||||
| med | p1 | p2 | p3 | p4 | p5 | p6 | p7 | med | p1 | p2 | p3 | p4 | p5 | p6 | p7 | ||||
| MZ | HI | 408 | 7/7 | 8 | 55 | 28 | 32 | 27 | 8 | 19 | 9 | 9 | 4 | ||||||
| MZ | HK | ″5·492 | 7/7 | 8 | 8 | 6 | 13 | 18 | 4 | 3 | 8 | 7 | 6 | ||||||
| MZ | WW | >20·000 | 13/11 | 5 | 11 | 15 | 22 | 38 | 9 | 26 | 17 | 25 | 31 | ||||||
| MZ | WG | ″1·524 | 13/11 | 17 | 30 | 39 | 29 | 61 | 12 | 20 | 14 | 13 | 22 | ||||||
| DZ | LF | ″1·886 | 11/3 | 13 | 38 | 30 | 25 | 72 | 1 | 1 | 1 | 1 | 1 | ||||||
| DZ | LK | ″3·672 | 11/3 | 16 | 40 | 35 | 22 | 70 | 2 | 1 | 1 | 1 | 2 | ||||||
| MZ | CS | >20·000 | 11/11 | 2 | 35 | 5 | 16 | 9 | 3 | 9 | 5 | 4 | 3 | ||||||
| MZ | FG | >20·000 | 11/11 | 11 | 29 | 9 | 20 | 9 | 3 | 2 | 3 | 3 | 1 | ||||||
| MZ | SA | >20·000 | 11/13 | 16 | 85 | 42 | 38 | 18 | 6 | 21 | 15 | 11 | 13 | ||||||
| MZ | SH | >20·000 | 11/13 | 6 | 13 | 4 | 32 | 14 | 11 | 15 | 21 | 10 | 20 | ||||||
| MZ | WM | >20·000 | 3/7 | 1 | 5 | 3 | 20 | 5 | |||||||||||
| MZ | WS | ″2·560 | 3/7 | 11 | 26 | 9 | 31 | 5 | |||||||||||
| MZ | BD | >20·000 | 3/8 | 6 | 2 | 2 | 15 | 17 | |||||||||||
| MZ | BT | >20·000 | 3/8 | 6 | 19 | 5 | 7 | 24 | |||||||||||
| MZ | PJ | >20·000 | 1/4 | 3 | 9 | 5 | 2 | ||||||||||||
| MZ | PA | 11·246 | 1/4 | 3 | 19 | 13 | 4 | ||||||||||||
| MZ | RF | ″1·366 | 3/3 | 26 | 31 | 28 | 36 | 72 | 8 | 10 | 7 | 11 | 10 | ||||||
| MZ | RS | 15·412 | 3/3 | 56 | 64 | 84 | 78 | 119 | 11 | 13 | 9 | 13 | 18 | ||||||
| SC | 1·266 | 3/3 | 7 | 11 | 34 | 16 | 16 | ||||||||||||
These contradictions were not caused by peptide presentation in the context of one of the other class II molecules (HLA-DP or -DQ) as proliferation of peptide specific T cell clones or lines could be completely blocked by addition of an HLA-DR antibody. Two representative examples are shown in Fig. 1. The figure also proves the presentation of peptide 2 by DR11 and DR13, which both were not predicted to bind this peptide.
Fig. 1.
Example for determination of HLA-DR restriction of HBsAg specific T cell clones from a DR11/13 positive proband. Clones were DR-restricted because proliferation could be blocked by an DR-specific antibody (DR-Ab, shown in the second lanes). Incubation with two types of non autologous APC each matching to only one of the DR alleles of the proband (given in lanes 3 and 4) identified the presenting DR molecule. Although peptide 2 was not predicted to bind to DR11 and DR13 these examples demonstrate that it is recognized by the tested T cell clones in the context of both molecules. The culture on the right probably contained a mixture of two different T cell clones. (SI = stimulation index).
Twenty-nine T cell clones of three probands heterozygous for the DR13 allele were established. Twenty-three of these 29 clones were DR13-restricted.
Moreover, all individuals recognized at least one of the tested HBsAg peptides irrespective of their DR type with precursor frequencies of HBsAg specific T cells between 1/1·700 and 1/33·000. These frequencies were calculated from the number of spots counted in the ELISPOT assays in which 105 T cells were used per test. There was no correlation between certain types of DR alleles and the number or strength of recognized peptides. The same was true for the strength of anti-HBs antibody production. Even individuals homozygous for DR3, which is associated with HBsAg nonresponse and did not bind any HBsAg epitope in the peptide binding ELISAs, recognized several of the peptides (some examples are given in the last three rows of Table 3). To demonstrate the DR restriction of these T cell responses T cell lines of a DR3-homozygous proband were established. Proliferation of these T cell lines specific for peptide 1 and 5 could be blocked by addition of an HLA-DR antibody (for representative examples see Fig. 2), giving evidence for the DR3-restriction of the T cell response.
Fig. 2.
Representative proliferation assays with T cell lines of a DR3-homozygous proband: Results demonstrate a DR3-restricted T cell response to two different HBsAg peptides (p1 and p5) as proliferation can be partially or totally inhibited by an HLA-DR antibody (DR-Ab). SI, stimulation index; med, medium.
T cell response to HBsAg is of the Th1 type
In the case of peptide recognition INF-γ was the predominantly produced cytokine. IL-10 production was less abundant and weaker than the INF-γ responses. In contrast to INF-γ, IL-10 was never produced alone, it always occurred in addition to INF-γ secretion which was always stronger than that of IL-10. Some representative examples are given in Table 3. This indicates that T cell response to HBsAg is mainly of the Th1 type.
Antigen recognition in monozygotic twin pairs
We analysed HBsAg peptide recognition for 54 probands of the twin vaccination study. Within the monozygotic and DR-identical twin pairs a surprising phenomenon occurred: Some twin partners recognized peptides not seen by the other twin. In 11 of 24 monozygotic and all of 3 dizygotic DR-identical twin pairs tested such a divergent peptide recognition pattern was observed. Examples are listed in Table 3.
Discussion
We have recently shown that immune responsiveness to HBsAg is largely genetically determined (60%). Variability in MHC encoded genes accounts for approximately 40% of this genetic determination [4]. Vaccination failure has been linked to the presence of certain HLA-class II alleles. In particular, DRB1*0301, DRB1*0701 and DQB1*0201 have been associated to vaccination failure [2,5,6]. Other antigens such as DRB1*0101, *1301, *1501 and DPB1*0401 appear to promote vaccine response [6–8]. However, since the initial description the functional background of these associations has remained unclear. As allelic differences between MHC class II molecules are clustered around the antigen binding groove the most obvious explanation for the observed HLA–class II associations would be variable efficiency in the antigen presentation between DRB1* antigens. In accordance with this hypothesis, we observed remarkable differences in peptide binding in the in vitro peptide binding assays. Peptides previously described as containing immunodominant epitopes (peptides 17–31, 167–181, and 207–221 [15]) showed in our assay high binding capacity to several HLA-DR alleles. In addition, high binding affinities observed for particular DRB1* antigens were in accordance with published restrictions (e.g. DR11 for peptide 17–31, DR7 and DR8 for peptide 207–221) [15]. However, in many cases binding assay results did not reflect peptide recognition patterns seen in proliferation assays and in the more sensitive ex vivo ELISPOT assays. In some instances negative controls (T cells and DC not loaded with peptide) showed a few positive spots. In these cases, the often used presentation of results as the difference of absolute spot counts between the control and the antigen tested would not give a conclusive insight into the antigen reaction. For example, a difference of 6 spots translates to a 500% increase if control value is 1 but to only 20% if the control value is 30 spots. In consequence we expressed our results in relative terms and considered antigen reactions with equal or more than twice as many spots than the negative control to be significant.
CD4+ T cells of probands often recognized peptides predicted not to bind or to bind with very low binding affinity to their respective DR molecules. In addition, several high affinity binding peptides were not recognized in individuals carrying the corresponding DRB1* antigens (examples for such contradictions are marked in grey in Table 3). Moreover, irrespective of their DR type all probands recognized at least one of the tested peptides with a stimulation index ≥2. We established 29 HBsAg specific DR-restricted T cell clones in 3 individuals (Fig. 1). In experiments with nonautologous, partially DR-matched antigen presenting cells T cell recognition of peptides not predicted to bind to particular DR molecules was observed. Even persons homozygous for DRB1*03 recognized two of the peptides (Fig. 2). Former studies have shown HBs-peptide presentation by HLA-DP- and -DQ-antigens [16]. This cannot be ruled out in our investigation since HLA-DR-antibodies did not completely block antigen recognition in all cases. In conclusion, high peptide binding affinity does not reliably predict immune dominant HBsAg peptides. Differences in HBsAg peptide binding affinity between the HLA-DR alleles do not explain the known HLA–DR association of HBsAg vaccination failure.
The use of the sensitive ELISPOT assays allowed the determination of the number of peptide specific T cells in the peripheral blood four to eight weeks after the 3rd vaccination. IFN-γ producing T cells were found with frequencies of 1/1·700–1/33·000 (estimated from the number of spot forming cells out of the 105 T cells used in the tests). Other authors using proliferative T cell responses as read out and limiting dilution techniques reported frequencies ranging between 1/12·000 in good responders and 1/50·000 in nonresponders [15,17].
Since all participants in this investigation were either mono- or dizygotic twins, we had the unique opportunity to compare cellular immune responses to HBsAg peptides in genetically identical individuals. Surprisingly, in almost half of the monozygotic twins, twin partners recognized peptides not recognized by the other twin (divergent peptide recognition was observed in 11 of 24 monozygotic twin pairs). Considering the high number of specific T cells in the peripheral blood, we regard a mere sample error as an unlikely explanation since the ELISPOT assay turned out to be highly sensitive tool. Peptide specific T cells that were detected in one twin with a very high frequency were absent in the other. These findings strongly suggest, that there are important postgenetic factors influencing T cell responsiveness. These could be differences in the peripheral T cell repertoire since rearrangement of the T cell receptor elements is a somatic event. Furthermore, environmental factors like infections may contribute to shaping the T cell repertoire thereby modifying the identical genetic background. Unfortunately, no DRB1*03 or *07 homozygous nonresponders were present in our study, but the efficient HBs antigen presentation of DRB1*03 or *07 positive responders and low responders make it unprobable that a problem with antigen presentation could be the reason for nonresponse. Off course, we cannot rule out the possible existence of clonally expanded T cells suppressing or even killing antigen-specific T cell like described earlier [18–20].
Our findings are further supported by observations in HBsAg nonresponders. Two groups have independently shown that antigen presenting cells from HBsAg nonresponders are able to present HBsAg peptides to DR-matched T cells of responders but in contrast there was no activation of nonresponder T cells either by autologous or by responder APC, so that the defect in nonresponse to HBsAg must be on the side of the T cells and not of the APC [9,10].
Although the MHC contributes 40% of the genetic effect to HBsAg responsiveness, there is strong evidence that other genes than HLA class II antigens are also important. Observations in mice and in humans strongly suggest that complete and partial deficiencies for complement factor C4A, which is encoded telomeric to the class II region, is responsible for nonresponsiveness particularly in HLA-DRB1*0301 individuals [21–23]. Most of these individuals carry a deletion of the C4A gene on the HLA A1/B8/DRB1*0301 haplotype.
In conclusion, our findings demonstrate that HBsAg peptide presentation on DR antigens cannot reliably be predicted by binding affinity. Peptide recognition crucially depends on the presence of specific T cells whose specificity is influenced by random somatic events like T cell receptor rearrangements and past infections.
Acknowledgments
This study was supported by a grant of the Deutsche Forschungsgemeinschaft (SFB490, A3) and by GlaxoSmithKline (Munich, Germany).
References
- 1.Whittle H, Jaffar S, Wansbrough M, Mendy M, Dumpis U, Collinson A, Hall A. Observational study of vaccine efficacy 14 years after trial of hepatitis B vaccination in Gambian children. Br Med J. 2002;325:569. doi: 10.1136/bmj.325.7364.569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Craven DE, Awdeh ZL, Kunches LM, et al. Nonresponsiveness to hepatitis B vaccine in health care workers. Results of revaccination and genetic typings. Ann Intern Med. 1986;105:356–60. doi: 10.7326/0003-4819-105-3-356. [DOI] [PubMed] [Google Scholar]
- 3.Alper CA. The human immune response to hepatitis B surface antigen. Exp Clin Immunogenet. 1995;12:171–81. doi: 10.1159/000424869. [DOI] [PubMed] [Google Scholar]
- 4.Hohler T, Reuss E, Evers N, et al. Differential genetic determination of immune responsiveness to hepatitis B surface antigen and to hepatitis A virus: a vaccination study in twins. Lancet. 2002;360:991–5. doi: 10.1016/S0140-6736(02)11083-X. [DOI] [PubMed] [Google Scholar]
- 5.Alper CA, Kruskall MS, Marcus-Bagley D, et al. Genetic prediction of nonresponse to hepatitis B vaccine. N Engl J Med. 1989;321:708–12. doi: 10.1056/NEJM198909143211103. [DOI] [PubMed] [Google Scholar]
- 6.Desombere I, Willems A, Leroux-Roels G. Response to hepatitis B vaccine: multiple HLA genes are involved. Tissue Antigens. 1998;51:593–604. doi: 10.1111/j.1399-0039.1998.tb03001.x. [DOI] [PubMed] [Google Scholar]
- 7.Hohler T, Meyer CU, Notghi A, et al. The influence of major histocompatibility complex class II genes and T- cell Vbeta repertoire on response to immunization with HBsAg. Hum Immunol. 1998;59:212–8. doi: 10.1016/s0198-8859(98)00014-7. [DOI] [PubMed] [Google Scholar]
- 8.Lango-Warensjo A, Cardell K, Lindblom B. Haplotypes comprising subtypes of the DQB1*06 allele direct the antibody response after immunisation with hepatitis B surface antigen. Tissue Antigens. 1998;52:374–80. doi: 10.1111/j.1399-0039.1998.tb03058.x. [DOI] [PubMed] [Google Scholar]
- 9.Desombere I, Hauser P, Rossau R, Paradijs J, Leroux-Roels G. Nonresponders to hepatitis B vaccine can present envelope particles to T lymphocytes. J Immunol. 1995;154:520–9. [PubMed] [Google Scholar]
- 10.Salazar M, Deulofeut H, Granja C, et al. Normal HBsAg presentation and T-cell defect in the immune response of nonresponders. Immunogenetics. 1995;41:366–74. doi: 10.1007/BF00163994. [DOI] [PubMed] [Google Scholar]
- 11.Bein G, Glaser R, Kirchner H. Rapid HLA-DRB1 genotyping by nested PCR amplification. Tissue Antigens. 1992;39:68–73. doi: 10.1111/j.1399-0039.1992.tb01909.x. [DOI] [PubMed] [Google Scholar]
- 12.Sinigaglia F, Romagnoli P, Guttinger M, Takacs B, Pink JR. Selection of T cell epitopes and vaccine engineering. Meth Enzymol. 1991;203:370–86. doi: 10.1016/0076-6879(91)03021-8. [DOI] [PubMed] [Google Scholar]
- 13.Hammer J. New methods to predict MHC-binding sequences within protein antigens. Curr Opin Immunol. 1995;7:263–9. doi: 10.1016/0952-7915(95)80013-1. [DOI] [PubMed] [Google Scholar]
- 14.Jonuleit H, Kuhn U, Muller G, Steinbrink K, Paragnik L, Schmitt E, Knop J, Enk AH. Pro-inflammatory cytokines and prostaglandins induce maturation of potent immunostimulatory dendritic cells under fetal calf serum-free conditions. Eur J Immunol. 1997;27:3135–42. doi: 10.1002/eji.1830271209. [DOI] [PubMed] [Google Scholar]
- 15.Honorati MC, Dolzani P, Mariani E, Piacentini A, Lisignoli G, Ferrari C, Facchini A. Epitope specificity of Th0/Th2 CD4+ T-lymphocyte clones induced by vaccination with rHBsAg vaccine. Gastroenterology. 1997;112:2017–27. doi: 10.1053/gast.1997.v112.pm9178695. [DOI] [PubMed] [Google Scholar]
- 16.Celis E, Ou D, Otvos L., Jr Recognition of hepatitis B surface antigen by human T lymphocytes. Proliferative and cytotoxic responses to a major antigenic determinant defined by synthetic peptides. J Immunol. 1988;140:1808–15. [PubMed] [Google Scholar]
- 17.Desombere I, Gijbels Y, Verwulgen A, Leroux-Roels G. Characterization of the T cell recognition of hepatitis B surface antigen (HBsAg) by good and poor responders to hepatitis B vaccines. Clin Exp Immunol. 2000;122:390–9. doi: 10.1046/j.1365-2249.2000.01383.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Watanabe H, Okumura M, Hirayama K, Sasazuki T. HLA-Bw54-DR4-DRw53-DQw4 haplotype controls nonresponsiveness to hepatitis-B surface antigen via CD8-positive suppressor T cells. Tissue Antigens. 1990;36:69–74. doi: 10.1111/j.1399-0039.1990.tb01802.x. [DOI] [PubMed] [Google Scholar]
- 19.Soroosh P, Shokri F, Azizi M, Jeddi-Tehrani M. Analysis of T-cell receptor beta chain variable gene segment usage in healthy adult responders and nonresponders to recombinant hepatitis B vaccine. Scand J Immunol. 2003;57:423–31. doi: 10.1046/j.1365-3083.2003.01256.x. [DOI] [PubMed] [Google Scholar]
- 20.Barnaba V, Franco A, Alberti A, Benvenuto R, Balsano F. Selective killing of hepatitis B envelope antigen-specific B cells by class I-restricted, exogenous antigen-specific T lymphocytes. Nature. 1990;345:258–60. doi: 10.1038/345258a0. [DOI] [PubMed] [Google Scholar]
- 21.De Silvestri A, Pasi A, Martinetti M, Belloni C, Tinelli C, Rondini G, Salvaneschi L, Cuccia M. Family study of non-responsiveness to hepatitis B vaccine confirms the importance of HLA class III C4A locus. Genes Immun. 2001;2:367–72. doi: 10.1038/sj.gene.6363792. [DOI] [PubMed] [Google Scholar]
- 22.Hohler T, Stradmann-Bellinghausen B, Starke R, Sanger R, Victor A, Rittner C, Schneider PM. C4A deficiency and nonresponse to hepatitis B vaccination. J Hepatol. 2002;37:387–92. doi: 10.1016/s0168-8278(02)00205-2. [DOI] [PubMed] [Google Scholar]
- 23.Milich DR. Influence of C4A deficiency on nonresponse to HBsAg vaccination: a new immune response gene. J Hepatol. 2002;37:396–9. doi: 10.1016/s0168-8278(02)00237-4. [DOI] [PubMed] [Google Scholar]