T-Cell Recognition of Mycobacterial GroES Peptides in Thai Leprosy Patients and Contacts

Abstract

We report here the mapping of T-cell-stimulatory determinants of the GroES 10-kDa heat shock protein homologues from Mycobacterium leprae and Mycobacterium tuberculosis, which are known as major immunogens in mycobacterial infections. Peripheral blood mononuclear cells (PBMC) from treated tuberculoid leprosy or lepromatous leprosy patients and from healthy household or hospital staff contacts of the patients were cultured with 20 16-mer peptides covering the entire sequences of both M. leprae and M. tuberculosis GroES. The total number of recognized peptides was found to be the largest in family contacts, while responder frequencies to the individual tested peptides varied (5 to 80%) with specificity between the patient and contact groups. Proliferative responses to some peptides showed positive or negative associations of low statistical significance with DR and DQ alleles, though responses to most GroES peptides were genetically permissive. Notably, the sequence of the 25–40 peptide of M. leprae, but not that of M. tuberculosis, was more frequently stimulatory in tuberculoid leprosy patients than in either group of sensitized healthy contacts. This peptide bound to a number of HLA-DR molecules, of which HLA-DRB5*0101 had the strongest affinity. The epitope core binding to this allele was localized to the 29-to-37 sequence, and its key residue was localized to the M. leprae-specific glutamic acid at position 32. This epitope may be of interest for the development of a blood test- or skin test-based diagnostic reagent for tuberculoid leprosy, subject to further clinical evaluation in untreated patients.

Leprosy research has predominantly been concerned with the dichotomy between immunological events in tuberculoid leprosy and lepromatous leprosy. Thus, distinct cytokine-secreting profiles have been reported for cloned CD4 and CD8 T cells (30), and lepromatous leprosy patients were found to respond to some purified antigens while remaining anergic to whole mycobacterial extracts (21, 26, 34). Analysis of the specificity of “split anergy” at the level of individual antigenic determinants showed skewing of T-cell responses toward Mycobacterium tuberculosis-specific epitopes, accompanied by relative anergy to the cross-reactive common mycobacterial peptides (13). However, consistent differences in either the specificities or phenotypes of T cells between the sensitized healthy contacts and tuberculoid leprosy patients have so far not been revealed.

Previous studies of HLA associations reported the association of tuberculoid leprosy with DR2 in India (18, 37), Japan (11, 23), Thailand (31), and Korea (16); with DR3 in Venezuela and Surinam (36); or with DQ1 (11, 23, 31). On the basis of molecular modelling, the presence of arginine (R) and absence of negatively charged amino acids at positions 13 or 70–71 in pocket 4 of the DRB1 alleles (e.g., DR15) has been associated with susceptibility to tuberculoid leprosy (40), but the molecular source of the corresponding peptide specificity has not been identified.

Proteins with a molecular mass of 10 to 12 kDa from Mycobacterium leprae (10) and M. tuberculosis (22) were originally found to carry separate species-specific epitopes identified with monoclonal antibodies. Their sequence analysis showed 90% identity between M. leprae and M. tuberculosis and 44% homology with the GroES heat shock protein of Escherichia coli (1, 20, 33), and the protein was localized to the cell wall and cytosol of M. leprae (29). The protein induced strong DTH and Th1 cytokine production, and limiting cell dilution analysis showed a high frequency of responding T cells from blood or from lepromin-induced Mitsuda reactions (17, 19, 35), but it also reacted with human “suppressor” T-cell clones (27) and was reported either to suppress (32) or to induce (9) DTH responses in lepromatous leprosy patients.

This study was performed at the level of individual peptide determinants, with additional emphasis on the analysis of HLA class II genetic associations. Using a comprehensive set of GroES peptides of the M. leprae and M. tuberculosis sequences, the aim of this study has been to identify any differences in the proliferation of blood mononuclear cells between paucibacillary or multibacillary leprosy patients and healthy leprosy contacts or medical staff in Thailand. Possible HLA associations were ascertained on the basis of typing for HLA-DR and -DQ alleles and by peptide binding to purified DR molecules.

MATERIALS AND METHODS

Synthetic peptides.

Two sets of peptide 16-mers overlapping by 8 residues, covering the entire protein sequences of M. leprae and M. tuberculosis GroES, were designated L and T, respectively (Fig. 1). Synthesis was carried out on the Milligen 9050 peptide synthesizer, using the 9-fluorenylmethyloxycarbonyl α-amino protecting group NovaSyn PR-500 resin, preformed pentafluorophenyl esters, and counterion distribution monitoring. The cleaved peptides were side chain deprotected, purified by Sephadex G-15 gel permeation chromatography in 25% aqueous acetic acid, and lyophilized. Sequence integrity was verified by mass spectrometry, and homogeneity was verified by reverse-phase high-performance liquid chromatography. N-terminally biotinylated peptides were produced by shaking the fully side chain-protected peptide resin in a solution containing 1.5 molar equivalents of both biotin-amidocaproate-N-hydroxysuccinimide ester and diisopropylethylamine in dimethyl sulfoxide (50 ml per mmol) for 2 h. A series of 15-mer peptides, overlapping by 14 amino acid residues (21 to 40) or with substitutions at positions 32 or 33, synthesized by simultaneous multiple-pin synthesis technology, were obtained from Cambridge Research Biochemicals Ltd. (Northwich, United Kingdom). These peptides were cleaved from the pins by exposure to 0.05 M HEPES buffer (pH 7.8).

FIG. 1.

Position of synthetic peptides derived from the sequences of M. leprae (LP) and M. tuberculosis (TB) GroES. Dashes represent identical residues; boxes represent sequences of 16-mer peptides.

Subjects.

Blood samples were obtained with informed consent from different groups of donors, all of Thai ethnic origin. The subject groups were as follows: HM, 16 healthy medical staff (average age, 35 years; 12 males and 4 females) who worked in the hospital or leprosarium area; HF, 12 healthy leprosy family contacts (average age, 29 years; 2 males and 10 females) who were family members of leprosy patients and lived in the leprosarium area; PB, 18 paucibacillary leprosy patients (average age, 46 years; 10 males and 8 females) who were selected from Phrapradaeng Hospital, Raj-Pracha-Samasai Institute and the Skin Clinics, in Bangkok, Thailand; and MB, 20 multibacillary leprosy patients (average age, 43 years; 16 males and 4 females) from the same medical institutions.

Patients were diagnosed and classified by standard clinical, bacteriological, and histological parameters. The majority had completed chemotherapy, but a minority were still being treated, while none had a previous history of tuberculosis. All patients and healthy subjects were tested at the time of blood donation by intradermal injection of 10 μg of leprosin (Rees antigen; 100 μg/ml) (World Health Organization Bank, National Institute for Medical Research, London, United Kingdom) and 10 U of tuberculin (purified protein derivative [PPD]; 100 U/ml) (Science Division, Thai Red Cross Society) and were graded as positive when they produced 5- (leprosin) and 10 (PPD)-mm-diameter skin indurations 48 h later. All donors included in the study were responsive to both leprosin and PPD, whereas MB patients responded to PPD but not to leprosin. Of the 38 MB patients originally considered, 7 were excluded on the grounds of skin test anergy and 11 were excluded due to failure to respond to any of the tested peptides.

Lymphocyte proliferation assay.

Peripheral blood mononuclear cells (PBMC) were isolated from fresh defibrinated whole blood by centrifugation on Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) density gradients. PBMC were resuspended in RPMI 1640 medium (Gibco, Paisley, Scotland, United Kingdom) supplemented with 5% AB⁺ heat-inactivated human serum, 2 mM l-glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin sulfate/ml. The cells (2 × 10⁵/well) were dispensed into quadruplicate wells of 96-well round-bottom plates (Nunc, Roskilde, Denmark) in the presence of appropriate antigens or mitogen in a total volume of 200 μl. The plates were incubated in a 5% CO₂, humidified atmosphere at 37°C for 7 days (5 days for concanavalin A [ConA]). [³H]thymidine (18.5 kBq/well) (Amersham International, Amersham, United Kingdom) was added for the last 16 to 18 h of incubation. The cells were harvested onto glass fiber filters, and radioactive incorporation was determined by liquid scintillation counting. The results were expressed as mean counts per minute of quadruplicate cultures or as the stimulation index (SI). Peptides were used at a final concentration of 50 μg/ml, and the recombinant proteins were used at 5 μg/ml. PPD and ConA were used at 100 U/ml and 5 μg/ml, respectively. Positive proliferative responses were defined by at least a 2.5-fold increase in counts per minute in cultures with antigen compared to the value in cultures containing medium alone (SI ≥ 2.5). This cutoff point is above the range of stimulation of PPD-negative control subjects (5). The functional viability of PBMC was confirmed by pronounced proliferative responses to the T-cell mitogen ConA in all subjects tested.

DNA extraction.

Two methods were used for DNA extraction, giving the same results. Whole-blood samples were mixed with cell membrane-lysing buffer (0.32 M sucrose, 1% Triton X-100) and centrifuged, and the pellet was resuspended in 63 μl of 20% sodium dodecyl sulfate and 50 μl of proteinase K (10 mg/ml) and incubated overnight at 37°C. Protein was extracted once with phenol and once with chloroform-isoamyl alcohol (24:1). Alternatively (25), PBMC were treated with proteinase K with agitation and spun. After addition of ethanol to the supernatant, the precipitate was dissolved in 100 μl of H₂O and adjusted to a 0.05- to 0.5-mg/ml DNA concentration.

HLA-DRB and -DQB typing.

HLA class II (DR and DQ) phenotypes were determined by the restriction fragment length polymorphism typing system for healthy Thai medical staff, healthy leprosy family contacts, and paucibacillary leprosy patients or by PCR sequence-specific priming for multibacillary leprosy patients. For restriction fragment length polymorphism typing (2), 10 μg of DNA was digested with Taq I endonuclease, and the resultant fragments were separated by submerged agarose gel electrophoresis and subsequently transferred to an immobilizing membrane by Southern blotting. These were then hybridized to either 570-bp DRB-specific or 630-bp DQB-specific ³²P-labelled cDNA probes. For PCR sequence-specific priming (3, 25), 32 separate PCRs were performed, each with unique primer pairs or sets. All reaction mixtures contained internal amplification control primers amplifying a 760-bp conserved sequence from the human growth hormone gene. Amplifications were performed with the Perkin-Elmer 9600 GeneAmp thermal cycler.

HLA-DR peptide binding assay.

The methods used for the HLA-DR peptide binding assay were described in detail elsewhere (12). Briefly, the reaction mixture contained the purified HLA-DR molecules (0.5 μM), a biotinylated (bio) reference peptide (1.6 μM bioCLIP/104–119 for DR17 and DR9 or 1.8 μM bioHA [influenza virus hemagglutinin]/306–318 for DR1, DRB5*0101 [38], DR4, DR11, DR14, DR7, and DR8), and competitor peptide (0.1 to 1,000 μM) in a total volume of 30 μl of binding buffer (0.1 M sodium citrate [pH 6.0] plus 0.5% Nonidet P-40). After incubation at room temperature for 48 h, HLA-DR–peptide complexes were diluted with 100 μl of blocking buffer (5% skim milk, 0.1% Tween 20 in phosphate-buffered saline [PBST]), divided into triplicate sets, and transferred to a 96-well microtiter plate, precoated with 1 μg of L243 monomorphic anti-HLA-DR monoclonal antibody per well. After 2 h of incubation with shaking at 4°C, the plates were washed with PBST and the quantity of bound biotinylated peptide was determined colorimetrically with streptavidin-peroxidase and tetramethyl benzidine (Sigma Chemical Co., Poole, United Kingdom). Absorbances in the presence of the reference peptide, without added competitor, were in the range of 0.2 to 0.5, while the background absorbances were <0.02. The values shown represent means from repeated assays, each based on the absorbances of triplicate wells and with <20% variations about their means. The values are presented as 50% inhibitory peptide concentrations (IC₅₀s) calculated from regression analysis of absorbance values. The binding of biotinylated HA/306–318 and biotinylated CLIP/104–119 was efficiently inhibited by their homologous unlabelled peptides, with an IC₅₀ in the range of 3 to 5 mM.

Statistical analysis.

The frequencies of HLA alleles or peptide responders were compared among the four tested groups by using Fisher’s exact test. For the comparison of allele frequencies with the previously published reference values, the χ² test was used. Association of peptide responsiveness, irrespective of group, with HLA alleles was evaluated by the Mann-Whitney U test. SIs of proliferative responses to particular peptides were not normally distributed, and therefore comparisons between the subject groups were determined by the Kruskal-Wallis test. Responses to M. leprae and M. tuberculosis peptides within the same subject group were compared by the Wilcoxon matched-pairs signed-rank test. The relationship between the magnitudes of responses to two peptides was assessed by the Spearman rank correlation coefficient. The statistical analysis was performed with the Stata (College Station, Texas) package (Stata Statistical Software, release 4.0 [1995]).

RESULTS

Recognition of peptides in patient and control groups.

The proliferation results in the four tested groups (HM, HF, PB, and MB) are presented as frequencies in responsiveness with a SI of ≥2.5 (Fig. 2). The following differences emerged between groups in regard to the recognition of L or T sequence peptides. (i) Peptide 9-24L was recognized more frequently in HF (63%) than in HM (29%) subjects, whereas both groups had frequent responses to the corresponding T peptide. Lower responsiveness to both peptides was found in PB and MB patients. (ii) Peptide 25–40 (p25–40) was recognized with bias for L by the majority of PB patients (P = 0.008 versus other groups) but with bias for T in HF contacts (P = 0.04). (iii) p57–72 was recognized with significant bias for L by the majority (83.3%) of HF contacts (P = 0.01). (iv) p65–80 was recognized with bias for L in 50% of MB patients, though the differences from the other groups were not statistically significant. These observations were corroborated by the values of median SIs (results not shown). In particular, SI values for responses p25–40 were significantly biased toward L in PB patients (P = 0.001 versus other groups) and toward T in HF contacts (P = 0.04). SI values for responses to p17–32 were also biased significantly toward T in HF contacts (P = 0.01). Furthermore, SI values in the HF group were increased for responses to p57–72L (P = 0.03).

FIG. 2.

Frequency of responders to mycobacterial GroES peptides. PBMC (2 × 10⁵) were cultured with synthetic peptides (50 μg/ml) derived from the GroES sequence of M. leprae (dark shaded bars) or M. tuberculosis (white bars) or identical for the two proteins (solid bars) (see the peptide positions in Fig. 1). Responders were classified on the basis of SIs of ≥2.5. ∗, significant differences from the other subject groups as determined by Fisher’s exact test.

There was pronounced individual variation in the total number of recognized peptides. This variation between 1 and 15 peptides showed significant differences among the four groups tested (Table 1). Evaluation on an arbitrary quantitative basis (1 to 4, 5 to 10 or >10 peptides) showed that multipeptide recognition was most abundant in the HF group, represented by 42% responders to >10 peptides and none to <5 peptides. In contrast, oligopeptide recognition was observed in HM subjects, reflected by 50% responding to only 1 to 4 peptides but none to >10 peptides, and in both PB and MB patients, represented by 33 to 40% recognition of 1 to 4 peptides and only 5 to 11% recognition of >10 peptides. These numerical differences in the total numbers of recognized peptides were not significantly different for the L and T sequence peptides.

TABLE 1.

Group differences in the total number of peptides recognized by an individual

Subject groupa (no.)	No. (%) of responders to total no. of stimulatory peptidesb of:
	1–4c	5–10d	>10e
HM (16)	8 (50)	8 (50)	0 (0)
HF (12)	0 (0)	7 (58)	5 (42)
PB (18)	6 (33)	10 (56)	2 (11)
MB (20)	8 (40)	11 (55)	1 (5)

^aHM, healthy medical staff; HF, healthy leprosy family contacts; PB, paucibacillary leprosy patients; MB, multibacillary leprosy patients. 

^bSI ≥2.5. 

^cProportion from 20 peptides tested. Fisher’s exact test, 0.02. HLA association, P = 0.03 (DR4 and DQ8). 

^dFisher’s exact test, 0.98. 

^eFisher’s exact test, 0.007. HLA association, P = 0.009 (DQ9). 

Correlations in recognition of individual peptides were analyzed for responses to the L and T peptides within each homologous peptide pair. In the case of p25–40, the frequency of responders to pL only was significantly increased in PB (P = 0.01) whereas pT responders were more frequent in the HF group (P = 0.03). Evaluation on the basis of pL/pT ratios of SI values showed significant differences between PB and HM (P = 0.003), HF (0.01), and MB (P = 0.03) groups. The Spearman rank correlation test was used to compare the SIs of all pairs of peptides in all subjects tested, irrespective of their clinical grouping (data not shown). In respect to pairs of peptides of homologous L or T sequence, this evaluation showed significant correlation (P < 0.001) for p1–16, p9–24, p17–32, p57–72, and p65–80, suggesting T-cell cross-reactivity, whereas proliferation to four peptide pairs located in the central part of the protein sequence (p25–40, p33–48, p41–56, and p49–64) without such correlation indicated species-specific recognition.

Frequency of HLA-DR and -DQ alleles and their association with anti-peptide proliferative responses.

HLA-DR and -DQ alleles were determined in 53 of the total of 66 individuals investigated, and the distribution of HLA-DR and -DQ alleles in the whole tested population (data not shown) corresponded to values in a recently reported extensive study of the Thai population from the same geographical area (4). Considering the small numbers (10 to 18) of subjects per group, evaluation of the allele frequencies by Fisher’s exact test showed significant association (P = 0.015) for DQ1, but only without correction for multiple comparisons. This allele occurred in 28 of 32 (87.5%) PB and MB patients, compared to 11 of 21 (52.3%) HM and HF healthy subjects. Analysis of DR15 subtypes showed that the DRB1*1501/DRB1*1502 proportion was 4:1 in the PB group and 1:3 in the MB group, thus apparently opposite to the previously reported association of DRB1*1502 with tuberculoid leprosy (18). We also failed to find in the PB group a significant increase of alleles (DRB1*15, -*16, -*14, -*09, and -*10) which are arginine positive at positions 13, 70, or 71 (40).

In order to increase the test sample size, a composite analysis of HLA associations with specificity of peptide recognition was carried out for subjects from all tested groups. Despite such pooling of data, evaluation by the Mann-Whitney U test following correction for multiple comparisons failed to yield any significant associations. Nevertheless, the results obtained without such correction are presented for those alleles which occurred in at least eight subjects (Table 2). The positive associations at a level of significance where P was <0.01 were found only for p65–80L in respect to DR15 and for p33–48L in respect to DQ5. Responses to both these peptides were also associated with the DR15-DQ5 linkage group, which is common in the Thai population. In contrast, negative associations of significance (P < 0.01) were observed for the proliferative response to p41–56L, p65–80L, and p81–100 in the context of DR4 and for p9–24L and p41–56L in the context of DQ8. These negative associations with DR4 and DQ8 are apparently related to the DR4-DQ8 linkage and also reached significance (P = 0.03) in relation to the general oligopeptide response pattern (Table 1). However, association of the multipeptide response pattern with DQ9 (P = 0.009) appears to be of broad specificity, since the positive association with p25–40L and p57–72T (Table 2) reached only a level of significance where P was <0.05.

TABLE 2.

Association of proliferative responses to GroES peptides with HLA-DR and -DQ alleles^a

Allele	No. of subjects	Peptide(s) with positive association	Peptide(s) with negative association
DR2	30	65–80L	None
DR15	24	33–48L, 65–80L*	None
DR4	10	None	41–56L*, 65–80L*, 81–100*
DR5	17	41–56T	65–80L, 65–80T
DR12	12	41–56T	33–48L, 65–80T
DQ1	39	33–48L, 65–80L	None
DQ5	32	33–48L*	None
DQ3	17	None	41–56T
DQ7	21	None	65–80T, 81–100
DQ8	8	None	9–24L*, 41–56L*, 65–80L, 81–100
DQ9	9	25–40L, 57–72T	None
DR15–DQ5	19	33–48L*, 57–72L, 65–80L	None
DR4–DQ8	8	None	9–24L*, 41–56L*, 65–80L, 81–100

^aHLA types from 53 subjects irrespective of group (i.e., 12 HM, 9 FM, 14 PB, and 18 MB) were analyzed by the Mann-Whitney U test. The significance was at a P value of <0.05 or <0.01 (*) for alleles present in at least eight individuals. L, M. leprae-derived sequence; T, M. tuberculosis-derived sequence. 

Binding of GroES peptides to HLA-DR molecules.

The tested peptides showed pronounced differences in their binding profiles (Fig. 3). Binding with high affinity (IC₅₀ < 10 μM) to almost all HLA-DR alleles was observed for p81–100 and p9–24L, and high-affinity binding to most DR alleles was observed with p9–24T and p65–80L. Other peptides bound a variable number of DR molecules with different affinities. Pronounced differences in DR binding between peptides of L or T specificity were detected for p25–40L, most prominently in respect to DRB5*0101. This peptide failed to inhibit the binding of the biotinylated reference peptide HLA-A3/p153–168 to DRB1*15 (results not shown). A striking difference between the homologous T and L p9–24 peptides was observed in respect to binding to DR17. Two L and T peptide pairs, p33–48 and p49–64, altogether failed to inhibit the binding of biotinylated reference peptides to almost all DR alleles (IC₅₀ > 1,000 μM), although these peptides did stimulate PBMC proliferation in up to 50% of the subjects, at least in the PB and HF groups, respectively.

FIG. 3.

Binding of peptides to HLA-DR molecules. The box shadings indicate the different test peptide concentrations (micromolar) producing 50% inhibition of binding (IC₅₀) of biotinylated reference peptides (HA/306–318 or CLIP/104–119) to purified HLA-DR molecules in the following ranges: <10 (black), 10 to 100 (horizontal lines), 100 to 1,000 (light shading), or >1,000 (white).

We took special interest in the analysis of p25–40L in view of the pronounced differences in proliferation between the HF and PB groups. The epitope core and its critical residues were localized on the basis of binding to DR15 (DRB5*0101), which exhibited the highest binding affinity to this peptide. On the basis of analysis of peptides overlapping by a single residue, the epitope binding core could be clearly localized to the nonapeptide p29–37, with the sequence VIPENAKEK (Fig. 4A). We then investigated the binding properties of core peptides with single-residue substitution of the two L versus T species-specific amino acids, E32D and N33T. The results (Fig. 4B) showed that binding was abrogated by E→D substitution at position 32 but not by N→T substitution at position 33.

FIG. 4.

Binding affinities of peptides related to the 25-to-40 sequence to DR15. The IC₅₀ values with biotinylated HA/306–318 are shown. (A) Identification of epitope core location by “window pepscan” with 15-mer overlapping peptides. (B) Identification of key residues by reciprocal substitution of the species-specific residues 32 and 33. The arrows indicate IC₅₀ values of >1,000 μM.

DISCUSSION

The antigenicity of a broad range of peptides of the GroES homologues of M. leprae and M. tuberculosis has been demonstrated on the basis of proliferative responses of human PBMC. This finding in a high proportion of individuals, combined with a lack of clear restriction by one or few HLA class II genes, suggests that most GroES peptides, like those from other mycobacterial antigens, are recognized in a genetically permissive way (39). In contrast, we recently reported prominent H2^d restriction of the proliferative response of primed spleen cells to GroES whole protein and peptides in mice of both BALB and B10 backgrounds (6). Demonstration in these murine studies of T-cell recognition of whole GroES and of the synthetic peptides suggests that the latter reflected peptide determinants processed by murine antigen-presenting cells as well as by naturally infected human phagocytes.

The pronounced peptide responsiveness in HF subjects compared to that in HM subjects could be explained by the exposure of the former subjects to infection from childhood and for a longer period. The differential recognition of peptides of M. leprae or M. tuberculosis sequence could be due to the different specificities of their epitope cores. However, in the case of some peptides, namely, those differing in only one terminal amino acid residue, such as p33–48 and p41–56, the differences appear to be in the epitope flanking region, which could engage distinct antigen-processing enzymes in antigen-presenting cells. The observed oligopeptide recognition pattern in MB patients is consistent with the known tendency of patients with lepromatous leprosy toward anergy (14, 24) and with the reduced number of recognized peptides of the 16-kDa antigen in tuberculosis patients (7). The observed stimulation of PBMC proliferation by peptides (i.e., p33–48 and p49–64) which were not active in the DR binding test could possibly be due to peptide presentation by non-DR (e.g., DQ or DP) molecules or, in the case of DR15, by the DRB1 molecule. Alternatively, these synthetic peptides may have undergone cellular processing different in some way from that of the whole protein for the generation of the immunogenic epitopes.

Analysis of the disease associations of proliferative responses revealed that p25–40T and p57–72L were recognized in the HF group at a significantly higher frequency than in the PB and MB groups. This finding may indicate that the corresponding T cells are beneficial for resistance to leprosy. A striking change in the species specificity of T-cell recognition is represented by the significantly increased recognition of p25–40L by PB patients. This outcome is attributable to the M. leprae-specific glutamic acid at position 32, which was identified as the epitope core key residue. This finding corroborates the results from murine studies, which demonstrated the key residue functions of E32 and N33 in the context of H2A^d and H2E^d (6). However, the remaining eight amino acid differences between M. leprae and M. tuberculosis located in other peptides did not result in demonstrable leprosy-associated T-cell responses. We also considered the possibility that the disease status of PB patients could have influenced responsiveness to the 25–40 peptide. However, we could not establish an association of the peptide responses with the leprosy type (TT or BT), the duration of the disease since diagnosis, or the length of chemotherapy.

HLA-DR restriction is amply demonstrable for cloned T cells (28) but is much less apparent for proliferation in bulk PBMC cultures, due to the multiplicity of HLA class II molecules with epitope binding functions, represented by different “supertypic” DR specificities (e.g., B1, B3, or B5), different isotypes (DR, DQ, and DP), and heterozygosity at the several loci. Despite these complexities, at least some HLA-DR and -DQ associations of low significance were observed for responses to p33–48 and p49–64. T-cell recognition of peptides which did not show demonstrable binding to DR molecules could have occurred in the context of DQ or DP, which can be selectively functional for certain epitope specificities of the hsp65 antigen of M. leprae (8, 28).

p25–40L attracted special attention in view of its pronounced selective recognition in the PB group compared to that of the homologous peptide of the M. tuberculosis sequence, and also as a result of extensive experimental analysis of this determinant in the mouse model (6). On the basis of peptide-DR binding analysis, the highest affinity of this peptide was to DRB5*0101 and the epitope core was located between residues 29 and 37 with a sequence of VIPENAKEK. The DRB5*0101 binding specificity is consistent with computer-based modelling (12), whereby K₃₇ (P9) conforms to the distinct role of the positively charged C-terminal anchor residue for binding to DRB5*01, whereas the lack of an aromatic residue at P4 is detrimental to binding to DRB1*1501. The lack of DRB5*01 binding by the 25–40 peptide with M. tuberculosis sequence containing merely a conservative change of the negatively charged glutamic acid to aspartic acid in position 32 could be due to its shorter side chain. However, presentation of this epitope may have occurred also in the context of other DR, DQ, or DP molecules.

Since the conclusion of this study (5), these results have assumed further importance in the light of a publication by Kim et al. (15), who obtained corroborating and complementary data on the recognition of M. leprae GroES peptides by T cells from 12 tuberculoid leprosy patients (15). In that study, an epitope with core residues 28 to 39 was described as the single most prominent determinant associated with HLA-DR2. However, in our analysis of a larger Thai population, responsiveness to the corresponding peptide, p25–40L, was genetically permissive, while several other GroES peptides were also found to be immunogenic. Our DR peptide binding data slightly diverge on the exact location of this epitope core (positions 29 to 37 versus 28 to 39) and on the lack of a key residue role for the threonine at position 33, but they agree on the binding of this epitope to the HLA-DRB5*01 molecule and not to the HLA-DRB1*15 molecule.

In conclusion, the observed differences between the epitope specificities of the T-cell repertoires to mycobacterial GroES antigens in healthy subjects and leprosy patients seems to justify further analysis, preferably in larger and genetically restricted study groups, including patients with recently diagnosed active leprosy. In particular, the high frequency of recognition of M. leprae p25–40 in paucibacillary leprosy patients, with specificity in respect to the homologous M. tuberculosis peptide, is of interest for the possible use of this epitope in the form of a blood T-cell-testing or skin-testing reagent for the early diagnosis of, or even predicting of susceptibility to, paucibacillary leprosy.