HLA-B*2705 is strongly associated with ankylosing spondylitis (AS) and reactive arthritis. In contrast, B*2709 has been reported to be more weakly or not associated to AS. These two molecules differ by a single amino acid change: aspartic acid in B*2705 or histidine in B*2709 at position 116. In this study, we analyzed the degree of T cell epitope sharing between the two subtypes. Ten allospecific T cell clones raised against B*2705, 10 clones raised against B*2703 but cross-reactive with B*2705, and 10 clones raised against B*2709 were examined for their capacity to lyse B*2705 and B*2709 target cells. The anti-B*2705 and anti-B*2703 CTL were peptide dependent as demonstrated by their failure to lyse TAP-deficient B*2705-T2 transfectant cells. Eight of the anti-B*2705 and five of the anti-B*2703 CTL clones lysed B*2709 targets. The degree of cross-reaction between B*2705 and B*2709 was donor dependent. In addition, the effect of the B*2709 mutation (D116H) on allorecognition was smaller than the effect of the other naturally occurring subtype change at this position, D116Y. These results demonstrate that B*2705 and B*2709 are the antigenically closest HLA-B27 subtypes. Because allospecific T cell recognition is peptide dependent, our results imply that the B*2705- and B*2709-bound peptide repertoires are largely overlapping. Thus, to the extent to which linkage of HLA-B27 with AS is related to the peptide-presenting properties of this molecule, our results would imply that peptides within a relatively small fraction of the HLA-B27-bound peptide repertoire influence susceptibility to this disease.

The HLA class I proteins constitutively bind and present endogenous peptides at the cell surface, which are recognized by CTL. The complexity of alloreactive T cell responses results from involvement of many alloantigen-bound peptides in allorecognition. For this reason, the clonal diversity of alloreactive CTL and their patterns of cross-reaction are indicative of the diversity of HLA-bound peptides and the relationships between the peptide repertoires of cross-reactive class I molecules.

HLA-B27 is strongly associated with ankylosing spondylitis (AS)3 (1, 2) and reactive arthritis (3), and there is strong evidence that it is directly involved in the pathogenesis of these diseases (4). A number of HLA-B27 subtypes (B*2701–B*2715) have been identified, which differ among each other by one or few amino acid residues. These changes are mainly located in the peptide-binding site of the molecule and affect peptide specificity and T cell recognition. B*2705, which is widespread in human populations, B*2702, which is restricted to Caucasians, B*2704, which is restricted to and frequent in Orientals, and B*2707, found in India and Southeast Asia, are associated to AS (5). Three independent studies (6, 7, 8) have reported lack of association of B*2706 to AS, a subtype found mainly in Southeast Asia, in the same populations in which B*2705 or B*2704 was associated to this disease. A few B*2706 AS patients were reported from China, where this subtype is found at very low frequency (5), suggesting that lack of association of B*2706 with AS is not absolute and might be modulated to some extent by genetic or environmental factors. B*2709 is relatively frequent in Sardinia. In one study conducted in the Sardinian population, B*2709 was found in 10 of 40 healthy HLA-B27-positive individuals and in none of 36 HLA-B27+ AS patients (9). This study strongly suggested that B*2709 is not or is more weakly associated to AS than other B27 subtypes in this population. Recently, a B*2709 individual with sacroiliitis and oligoarthritis was found in continental Italy, where B*2709 is present at very low frequency (50), suggesting that, as for B*2706, lack of association of B*2709 with AS is not absolute.

Among the various hypotheses to explain the mechanism of HLA-B27 association to disease, one proposes that HLA-B27 is the restriction element for an “arthritogenic” peptide that would be the target Ag of CTL activated on external challenge, such as a bacterial infection (10). CTL-mediated damage in relevant tissues would be a primary pathogenic event that, together with additional processes susceptible to modulation by additional genetic factors, would finally lead to disease. The arthritogenic peptide model has not been proved and is only suggested by circumstantial evidence, and other pathogenetic mechanisms remain open (reviewed in 4). However, in contrast to other hypotheses, this one does not invoke unknown features for HLA-B27, other than its normal Ag-presenting function. In addition, recent findings in transgenic rats indicate that a major alteration of the HLA-B27-bound peptide repertoire prevents arthritis in these animals (11). According to this model, any disease-associated subtypes should be able to present the arthritogenic peptide(s) to CTL but B*2706 and B*2709 would be unable or less capable to do so. Thus, the peptide binding and antigenic differences between subtypes differentially associated to AS may be relevant to our understanding of the relationship between HLA-B27 and spondyloarthropathy.

Much of the polymorphism among HLA-B27 subtypes is clustered in the C/F pocket and determines the specificity for C-terminal peptide residues (12). Multiple studies have analyzed subtype-bound peptide repertoires (13, 14, 15, 16, 17, 18, 19, 20, 21), but the capacity of CTL to recognize a given peptide when presented by different subtypes has not been analyzed in depth. Few viral peptides are known to be recognized by the same T cells in the context of different subtypes (22, 23, 24, 25), and in only one case was cross-reactivity of an alloreactive CTL clone with various B27 subtypes shown to be mediated by recognition of the same peptide (26). However, allospecific T cell epitopes shared between B*2705 and other subtypes are not rare, suggesting that HLA-B27 subtypes behave synonymously in presentation of multiple peptides to CTL. Of the subtypes tested in a previous study (B*2701 to B*2706), B*2703 showed the highest cross-reaction with B*2705, with about 50% of anti-B*2705 CTL clones recognizing B*2703 (27).

The aim of our study was to assess T cell epitope sharing between B*2705 and B*2709, because they differ only by the D116H substitution (28) but are differentially associated with AS in one population (9). Differences in the recognition of a viral peptide in the context of both subtypes by at least a B*2709-restricted T cell line have been reported (29), but our goal was to determine whether such differences are common among T cells recognizing these two subtypes. Thus, we analyzed the degree of clonal cross-reaction between B*2705 and B*2709 in alloreactive responses. Because allospecific CTL recognize class I-bound peptides, the cross-reaction between B*2705 and B*2709 is an indication of the capacity of these subtypes to interchangeably present peptides to CTL.

The human class I-deficient HMy2.C1R (C1R) cell line and their transfectants expressing B*2705 and the mutant D116Y have been previously described (30, 31). Full length cDNA encoding B*2709 (28) (a kind gift of Dr. Rosa Sorrentino, University of L’Aquila, Italy) cloned into RSV5neo expression vector was transfected into C1R cells as previously described (30). In addition, genomic B*2709 DNA was obtained by site-directed mutagenesis of the B*2705 gene, cloned into the pUC18 vector and equally transfected into C1R cells. T2, a TAP-deficient human cell line, transfected with B*2705 was a gift of Dr. David Yu (University of California, Los Angeles, CA). When incubated at 26°C, this cell line expresses class I molecules presumably devoid of peptides. All these cells were cultured in DMEM supplemented with 5% heat-inactivated FCS (both from Life Technologies Laboratories, Paisley, U.K.). The expression levels of HLA-B27 on C1R transfectants were very high and identical for B*2705 and B*2709 (cDNA) and somewhat higher (166%) for the genomic B*2709 transfectant. They were tested by flow cytometry with the mAb ME1 (IgG1 specific for HLA-B27, -B7 and -B22) (32), as previously described (33).

Three sets of alloreactive CTL clones were used. A first set (Table I) consisted of CTL generated against B*2705 from three unrelated donors: DL (HLA-A29, 31; B39, 44; DR2, 7); GM (HLA-A1, 24; B7, 8; DR1, 3); and SR (HLA-A3, 29; B7, 44; DR2, 7). A second set (Table II) consisted of CTL generated against B*2703 from donors DL and GM, but cross-reactive with B*2705. The generation, culture conditions, and fine specificity of both sets of CTL have been previously described (27). A third set consisted of CTL clones generated against B*2709 as follows. About 106 PBMC from healthy HLA-B27-negative donor MM (HLA-A3, 11; B7; DR2, 7) were stimulated with 105 irradiated (80 Gy) allogeneic HLA-B*2709-positive lymphoblastoid cell line (LCL) Ci (HLA-A2, 10; B16, 27) and 106 irradiated (50 Gy) autologous PBMC in Iscove’s modified DMEM with Glutamax (Life Technologies Laboratories, Paisley, U.K.) supplemented with human AB sera (15%), gentamicin (50 mg/ml), streptomycin (0.1 mg/ml), and penicillin (100 U/ml). At weekly intervals, MLR cultures were washed and counted, and 3 × 105 cells were seeded in a 24-well plate and restimulated with the same mixture of feeder cells and stimulator LCL as before in the presence of 20 U/ml rIL-2 (Hoffman LaRoche, Palo Alto, CA). Clones were obtained by limiting dilution. Cells were seeded in 96-well plates containing the same irradiated autologous PBMC:B*2709 LCL ratio as in the MLR. Positive wells were considered clones when they grew at dilutions below the limit of clonality determined by Poisson statistical analysis. Statistical clones were screened for HLA-B27 reactivity using B*2709-C1R target cells. Their specificity was confirmed by testing the capacity of the ME1 mAb to inhibit lysis of these target cells and their failure to kill untransfected C1R cells. The 64.8P CTL clone, used as control (Table III), has been previously described (31, 34).

Table I.

Cytotoxicity of anti-B*2705 CTL clones against C1R transfectant cells expressing HLA-B*2705 or B*2709a

DonorCTLE:T RatioC1RB*2705-C1RB*2709-C1RbB27 Subtypes Recognizedc
SR 37SLG 1:1 7 (5) 42 (5) 61 (5) 5, 1, 2, 3, 4 
  2:1 8 (2) 56 (2) 68 (2)  
GM 31GRK 1.5:1 5 (2) 39 (2) 45 (2) 5, 3 
SR 27S69 1:1 3 (7) 28 (7) 34 (7) 5, 2, 3 
  2:1 2 (5) 60 (5) 63 (5)  
GM 20.8GRK 2:1 7 (6) 29 (6) 32 (6) 5, 1, 2, 3, 6 
  4:1 17 (2) 56 (2) 56 (2)  
GM 58GRK 1:1 1 (6) 37 (6) 29 (6) 5, 1, 4 
  2:1 3 (11) 64 (11) 67 (11)  
GM 100GRK 1.5:1 3 (3) 45 (3) 36 (3) 5, 1, 2, 4, 6 
GM 37GRK 0.5:1 3 (6) 21 (6) 8 (6) 5, 1, 2, 3, 4, 6 
  1:1 3 (3) 38 (3) 17 (3)  
  3:1 4 (6) 66 (6) 51 (6)  
SR 33S69 1:1 1 (4) 30 (4) 13 (4) 5, 1, 3 
  2:1 8 (4) 62 (4) 33 (4)  
DL 12.8DM5 2:1 5 (3) 36 (3) 4 (3) 
  4:1 12 (3) 52 (3) 18 (3)  
DL 102DRF 1:1 2 (5) 25 (5) 0 (5) 
  2:1 2 (4) 48 (4) 1 (4)  
DonorCTLE:T RatioC1RB*2705-C1RB*2709-C1RbB27 Subtypes Recognizedc
SR 37SLG 1:1 7 (5) 42 (5) 61 (5) 5, 1, 2, 3, 4 
  2:1 8 (2) 56 (2) 68 (2)  
GM 31GRK 1.5:1 5 (2) 39 (2) 45 (2) 5, 3 
SR 27S69 1:1 3 (7) 28 (7) 34 (7) 5, 2, 3 
  2:1 2 (5) 60 (5) 63 (5)  
GM 20.8GRK 2:1 7 (6) 29 (6) 32 (6) 5, 1, 2, 3, 6 
  4:1 17 (2) 56 (2) 56 (2)  
GM 58GRK 1:1 1 (6) 37 (6) 29 (6) 5, 1, 4 
  2:1 3 (11) 64 (11) 67 (11)  
GM 100GRK 1.5:1 3 (3) 45 (3) 36 (3) 5, 1, 2, 4, 6 
GM 37GRK 0.5:1 3 (6) 21 (6) 8 (6) 5, 1, 2, 3, 4, 6 
  1:1 3 (3) 38 (3) 17 (3)  
  3:1 4 (6) 66 (6) 51 (6)  
SR 33S69 1:1 1 (4) 30 (4) 13 (4) 5, 1, 3 
  2:1 8 (4) 62 (4) 33 (4)  
DL 12.8DM5 2:1 5 (3) 36 (3) 4 (3) 
  4:1 12 (3) 52 (3) 18 (3)  
DL 102DRF 1:1 2 (5) 25 (5) 0 (5) 
  2:1 2 (4) 48 (4) 1 (4)  
a

Results are expressed as percentage of specific 51Cr release and are means from the number of experiments in parentheses. Reactivity with HLA-B27 subtypes is from Ref. 27 .

b

C1R cells transfected with B*2709 cDNA.

c

These anti-B*2705 CTL clones show 9 of 18 previously observed reaction patterns with B*2701-B*2706, accounting for 68% of all the anti-B*2705 clones analyzed previously (27 ).

Table II.

Cytotoxicity of anti-B*2703 CTL clones cross-reactive with B*2705 against C1R transfectants expressing HLA-B*2705 or B*2709a

DonorCTLE:T RatioC1RB*2705-C1RB*2709-C1RbB27 Subtypes Recognizedc
GM 42GCP 1:1 2 (2) 74 (2) 77 (2) 3, 5, 2 
  2:1 5 (4) 89 (4) 77 (4)  
DL 98DCJ 2:1 1 (6) 78 (6) 72 (6) 3, 5, 1, 2 
DL 81DCJ 2:1 1 (6) 74 (6) 64 (6) 3, 5 
GM 63GLM 2:1 1 (2) 45 (2) 27 (2) 3, 5 
  4:1 1 (4) 64 (4) 24 (4)  
GM 62GCP 2:1 1 (6) 73 (6) 41 (6) 3, 5 
DL 3DLH 2:1 0 (4) 38 (4) 3 (4) 3, 5, 4 
GM 29GCP 1:1 0 (2) 43 (2) 0 (2) 3, 5 
  2:1 3 (4) 69 (4) 5 (4)  
DL 18DLH 1:1 2 (4) 17 (4) 0 (4) 3, 5 
  2:1 2 (3) 23 (3) 0 (3)  
DL 36DLH 1:1 3 (5) 35 (5) 2 (5) 3, 5 
  2:1 2 (6) 59 (6) 1 (6)  
GM 54GCP 1:1 3 (4) 62 (4) 0 (4) 3, 5 
  2:1 0 (2) 83 (2) 1 (2)  
DonorCTLE:T RatioC1RB*2705-C1RB*2709-C1RbB27 Subtypes Recognizedc
GM 42GCP 1:1 2 (2) 74 (2) 77 (2) 3, 5, 2 
  2:1 5 (4) 89 (4) 77 (4)  
DL 98DCJ 2:1 1 (6) 78 (6) 72 (6) 3, 5, 1, 2 
DL 81DCJ 2:1 1 (6) 74 (6) 64 (6) 3, 5 
GM 63GLM 2:1 1 (2) 45 (2) 27 (2) 3, 5 
  4:1 1 (4) 64 (4) 24 (4)  
GM 62GCP 2:1 1 (6) 73 (6) 41 (6) 3, 5 
DL 3DLH 2:1 0 (4) 38 (4) 3 (4) 3, 5, 4 
GM 29GCP 1:1 0 (2) 43 (2) 0 (2) 3, 5 
  2:1 3 (4) 69 (4) 5 (4)  
DL 18DLH 1:1 2 (4) 17 (4) 0 (4) 3, 5 
  2:1 2 (3) 23 (3) 0 (3)  
DL 36DLH 1:1 3 (5) 35 (5) 2 (5) 3, 5 
  2:1 2 (6) 59 (6) 1 (6)  
GM 54GCP 1:1 3 (4) 62 (4) 0 (4) 3, 5 
  2:1 0 (2) 83 (2) 1 (2)  
a

Results are expressed as percentage of specific 51Cr release and are means from the number of experiments in parentheses. Reactivity with HLA-B27 subtypes is from Ref 27 .

b

C1R cells transfected with B*2709 cDNA.

c

These anti-B*2703 CTL used exhibited 4 of the 9 reaction patterns observed among this type of clones with B*2701-B*2706, which account for 69% of the clones analyzed previously (27 ).

Table III.

Cytotoxicity of anti-B*2705 and anti-B*2703 CTL clones cross-reactive with B*2705 against T2 transfectant expressing B*2705a

CTLC1RB*2705-C1RB*2705-T2
Anti-B*2705    
12.8DM5 39 
20.8GRK 73 
27S69 75 
31GRK 61 
33S69 72 
37GRK 86 
37SLG 45 11 
58GRK 75 
100GRK 57 
102DRF 45 
Anti-B*2703    
3DLH 71 
18DLH 62 
29GCP 69 ND 
36DLH 79 
42GCP 51 10 
54GCP 51 
62GCP 31 
63GLM 62 
81DCJ 74 ND 
98DCJ 81 
Anti-B*2704    
64.8P 51 48 
CTLC1RB*2705-C1RB*2705-T2
Anti-B*2705    
12.8DM5 39 
20.8GRK 73 
27S69 75 
31GRK 61 
33S69 72 
37GRK 86 
37SLG 45 11 
58GRK 75 
100GRK 57 
102DRF 45 
Anti-B*2703    
3DLH 71 
18DLH 62 
29GCP 69 ND 
36DLH 79 
42GCP 51 10 
54GCP 51 
62GCP 31 
63GLM 62 
81DCJ 74 ND 
98DCJ 81 
Anti-B*2704    
64.8P 51 48 
a

Data are expressed as percent specific 51Cr release and are means of 3 to 5 experiments carried out at an E:T ratio of 2:1. B*2705-T2 target cells were previously incubated at 26°C for 18–24 h. The CTL clone 64.8P (31, 34) was used as a positive control of lysis of the B*2705-T2 target cells.

A standard 4-h 51Cr release assay was conducted as previously described (35), with minor modifications. For T2 transfectants, the same assay was used, but the cells were incubated for 18–24 h at 26°C before using them as target cells.

A total of 10 anti-B*2705 CTL clones (Table I) from three unrelated donors (DL, GM and SR) and 10 anti-B*2703 CTL clones (Table II) from DL and GM were tested for recognition of B*2709. The fine specificity of these CTL with other HLA-B27 subtypes (B*2701-B*2706) and mutants was previously established. Based on the diverse reaction patterns with B*2701-B*2706 (Tables I and II), the anti-B*2705 and anti-B*2703 CTL used in this study are a fair sample of the clonotypic diversity observed in these alloreactive responses (27). All these CTL clones were peptide dependent, as demonstrated by their failure to lyse the T2-B*2705 transfectant cell line (Table III). The natural peptide epitope has recently been reported (26) for one of these clones (CTL 27S69). In addition, distinct HPLC fractions from the B*2705-bound peptide pool sensitized B*2705-T2 targets for lysis by CTL 20.8GRK, 37GRK, 33S69, and 36DLH (our unpublished observations).

The alloreactive CTL clones raised against B*2705 were examined for their ability to lyse C1R transfectant targets expressing B*2709 (Table I). All clones lysed B*2705-C1R targets and failed to lyse or lysed marginally (CTL, 20.8GRK and 12.8DM5) C1R cells transfected only with pSV2neo.

Of the 10 CTL clones analyzed, 8 cross-reacted with B*2709; 6 of them reacted equally with B*2705 and B*2709, and 2 (37GRK and 33S69) recognized B*2709 less efficiently (50% or less relative lysis). These results demonstrate that a large majority (80%) of the epitopes recognized by anti-B*2705 CTL are also present in B*2709, and they indicate that B*2709 is the antigenically closest subtype to B*2705.

In a previous study (27), we showed that most (78%) anti-B*2703 CTL clones cross-reacted with B*2705, whereas <20% of these clones cross-reacted with any of the B*2701, B*2702, B*2704, or B*2706 subtypes. Now 10 of the anti-B*2703 CTL cross-reactive with B*2705 from donors DL and GM were tested for recognition of B*2709 on C1R cells (Table II). Five clones failed to lyse B*2709 target cells. Of those cross-reacting with this subtype, three (42GCP, 98DCJ, and 81DCJ) recognized equally B*2705 and B*2709, and two (63GLM and 62GCP) lysed more efficiently B*2705 than B*2709 targets. These results indicate that 50% of the B*2703 epitopes shared with B*2705 are maintained in B*2709. Thus, B*2709 is antigenically closer to B*2703 than other subtypes except B*2705.

Although the number of CTL clones tested from each individual was small, a clear difference in cross-reactivity with B*2709 was observed among donors (Table IV). Only 2 of 7 CTL clones from DL (29%) recognized B*2709, whereas 8 of 10 clones from GM (80%), and the 3 clones from SR cross-reacted with this subtype. This difference was due mainly to the anti-B*2705 CTL; none of the two anti-B*2705 clones from DL and all those from GM (5 clones) or SR (3 clones) recognized B*2709. In contrast, the fraction of anti-B*2703 CTL clones from DL (2 of 5) or GM (3 of 5) cross-reacting with B*2709 was similar.

Table IV.

Cross-reaction of anti-HLA-B27 CTL clones from individual donors with B*2705 and B*2709a

DonorSubtypeCTL
Anti-B*2705Anti-B*2703Total (%)
DL B*2705 7 (100%) 
 B*2709 2 (29%) 
GM B*2705 10 (100%) 
 B*2709 8 (80%) 
SR B*2705 3 (100%) 
 B*2709 3 (100%) 
All donors B*2705 10 10 20 (100%) 
 B*2709 13 (65%) 
DonorSubtypeCTL
Anti-B*2705Anti-B*2703Total (%)
DL B*2705 7 (100%) 
 B*2709 2 (29%) 
GM B*2705 10 (100%) 
 B*2709 8 (80%) 
SR B*2705 3 (100%) 
 B*2709 3 (100%) 
All donors B*2705 10 10 20 (100%) 
 B*2709 13 (65%) 
a

Values represent the number and percentage of CTL clones recognizing each subtype.

These results are in agreement with previous studies (27, 36) wherein DL was a “narrow” responder with a tendency to respond against “private” allospecific epitopes, whereas GM and SR were “broad” responders, generating more cross-reactive responses. Our results indicate that the alloantigenic properties of B*2709 differ among individuals and must be analyzed with different donors to exclude individual bias.

Three different residues are found at position 116 among HLA-B27 subtypes: aspartic acid in B*2705, B*2703, and various other subtypes; tyrosine in B*2706, B*2707, and B*2711; and histidine in B*2709. In a previous report from our laboratory (36), the Tyr116 mutation abrogated recognition by 54% of anti-B*2705 and 53% of anti-B*2703 CTL clones. This mutant was not or marginally recognized (lysis relative to B*2705 <30%) by 12 of 19 (63%) CTL clones used in our current study (Table V). In contrast, only 7 of the 20 anti-B*2705 plus anti-B*2703 clones (35%) failed to cross-react with B*2709 (His116). This result indicates that the H116 change has an smaller effect on allorecognition than Tyr116.

Table V.

Reaction patterns of anti-B*2705 and anti-B*2703 CTL clones with C1R transfectants expressing B*2709 and the Y116 mutanta

GroupaCross-reaction withCTLRelative LysisNo. of CTL Clones
B*2709Y116B*2709Y116bAnti-B*2705Anti-B*2703Total
− − 12.8DM5 23    
   102DRF    
   3DLH    
   18DLH 2 /9 (22%) 5 /10 (50%) 7 /19 (37%) 
   29GCP    
   36DLH    
   54GCP    
         
II − 37GRK 50 2 /9 (22%)  2 /19 (11%) 
   33S69 45    
         
III ++ − 27S69 109    
   81DCJ 86 13 2 /9 (22%) 1 /10 (10%) 3 /19 (15%) 
   100GRK 80 20    
         
IV ++ 37SLG 138 56 1 /9 (11%)  1 /19 (5%) 
         
++ 62GCP 56 142  2 /10 (20%) 2 /19 (11%) 
   63GLM 45 165    
         
VI ++ ++ 20.8GRK 108 102    
   58GRK 95 83 2 /9 (22%) 2 /10 (20%) 4 /19 (21%) 
   42GCP 92 93    
   98DCJ 92 140    
GroupaCross-reaction withCTLRelative LysisNo. of CTL Clones
B*2709Y116B*2709Y116bAnti-B*2705Anti-B*2703Total
− − 12.8DM5 23    
   102DRF    
   3DLH    
   18DLH 2 /9 (22%) 5 /10 (50%) 7 /19 (37%) 
   29GCP    
   36DLH    
   54GCP    
         
II − 37GRK 50 2 /9 (22%)  2 /19 (11%) 
   33S69 45    
         
III ++ − 27S69 109    
   81DCJ 86 13 2 /9 (22%) 1 /10 (10%) 3 /19 (15%) 
   100GRK 80 20    
         
IV ++ 37SLG 138 56 1 /9 (11%)  1 /19 (5%) 
         
++ 62GCP 56 142  2 /10 (20%) 2 /19 (11%) 
   63GLM 45 165    
         
VI ++ ++ 20.8GRK 108 102    
   58GRK 95 83 2 /9 (22%) 2 /10 (20%) 4 /19 (21%) 
   42GCP 92 93    
   98DCJ 92 140    
a

The symbols used in groups indicate: −, 0–30%; +, 31–60%; and ++, >60% lysis, relative to the specific lysis of B*2705-C1R. Results are expressed as means of the relative lysis obtained at the various E:T ratios tested (Tables I and II).

b

Data with the Tyr116 mutant are from Refs. 36 and 37 , except for CTL 33S69 and 18DLH, and are included here for comparison.

On the basis of their reactivity with the two mutants, the CTL clones could be classified in six reaction patterns (Table V). All clones not recognizing B*2709 also failed to recognize Tyr116 (reaction pattern I). This group represented 37% of the CTL clones tested. Those CTL cross-reacting with B*2709 (64%) showed different reaction patterns. Two clones were partially affected by His116 and failed to recognize Tyr116 (reaction pattern II). Three clones were unaffected by His116 but did not or marginally recognized Tyr116 (reaction pattern III). One clone was partially affected by Tyr116 and not by His116 (reaction pattern IV) and two clones showed the reverse situation (reaction pattern V). Finally, 4 CTL clones were not affected by either change (reaction pattern VI). Thus, the two mutations exhibited differential effects on 8 (42%) of 19 CTL clones compared (reaction patterns II–V). In 6 of these 8 clones (reaction patterns II–IV) the effect of Tyr116 on CTL recognition was more drastic than for H116.

Taken together, these results indicate that the two natural mutations occurring at position 116 among subtypes have distinct effects on the antigenic properties of HLA-B27, and that D116Y is more disruptive than D116H.

Ten statistical clones raised from an HLA-B27-negative donor against B*2709 were tested for recognition of B*2705 (Table VI). Nine of the clones (90%) lysed B*2705-C1R transfectant cells, 8 of them with an efficiency comparable with but somewhat higher than the lysis of B*2709 targets and 1 (CTL 2B212) with decreased efficiency (38% relative lysis). Only 1 of the 10 clones tested failed to recognize B*2705. These results indicate that a large majority, but not all, of the allospecific epitopes of B*2709 are also present in B*2705.

Table VI.

Cytotoxicity of anti-B*2709 CTL clones against C1R transfectant cells expressing B*2709 or B*2705a

CTLC1RB*2709-C1RbB*2705-C1R
3F212 56 (2) 81 (2) 
2E1006 48 (2) 74 (2) 
3B806 57 (2) 68 (2) 
3D312 76 (1) 90 (1) 
3G812 66 (1) 74 (1) 
3G1012 63 (2) 78 (2) 
3G612 59 (1) 71 (1) 
3B1103 27 (1) 37 (1) 
2B212 34 (1) 13 (1) 
2E703 59 (1) 6 (1) 
CTLC1RB*2709-C1RbB*2705-C1R
3F212 56 (2) 81 (2) 
2E1006 48 (2) 74 (2) 
3B806 57 (2) 68 (2) 
3D312 76 (1) 90 (1) 
3G812 66 (1) 74 (1) 
3G1012 63 (2) 78 (2) 
3G612 59 (1) 71 (1) 
3B1103 27 (1) 37 (1) 
2B212 34 (1) 13 (1) 
2E703 59 (1) 6 (1) 
a

Results are expressed as percentage of specific 51Cr release. Data are means from the number of experiments in parentheses. The E:T ratio used in all experiments was 1:1.

b

C1R cells transfected with B*2709 genomic DNA.

The results in this study demonstrate that B*2709, although distinguishable from B*2705 by allospecific CTL, has the highest antigenic similarity with this subtype among all the HLA-B27 subtypes tested thus far.

The overwhelming majority of allospecific CTL, including those in this study are peptide dependent, and it is likely that CTL cross-reaction between subtypes generally reflects recognition of the same peptide in the different contexts (26). Thus, our results strongly suggest that many of the peptides that are relevant for allospecific T cell recognition are presented by both B*2705 and B*2709 to CTL.

The D116H change between B*2709 and B*2705 influences the nature of the C-terminal (PΩ) amino acid side chain of bound peptides (20). Whereas B*2705 binds in vivo peptides with basic, aliphatic, and aromatic PΩ residues (38), known B*2709-bound peptides have only nonpolar aliphatic (L, V, I, M) and aromatic (F) PΩ residues (20). This is probably due to loss of the Asp116 charge and to the larger size of His116, which would hinder acceptance of very bulky PΩ side chains. This means a significant restriction in the B*2709-bound peptide repertoire, relative to B*2705. In a recent compilation of natural B*2705 ligands, 27 of 46 self peptides (59%) showed basic or Tyr C-terminal residues, not reported among B*2709 ligands, and only 41% of the B*2705 ligands showed C-terminal motifs common to those of B*2709 (39). The disparity between the apparent degree of peptide sharing and the higher degree of CTL cross-reaction has three possible explanations. 1) many cross-reactive CTL would recognize different peptides in the context of B*2705 and B*2709. This type of cross-reaction, although observed between structurally distant class I allotypes (40, 41, 42), might be less frequent for closely related subtypes and was not the case for the only CTL clone in this study for which the peptide epitope is known (Ref. 26 and our unpublished observations). 2) Ligands with nonpolar PΩ residues would be more immunogenic than those with C-terminal basic or tyrosine residues. This is also unlikely because B*2705-restricted viral epitopes have both aliphatic and basic PΩ residues (13, 22, 43, 44, 45, 46). 3) B*2709 could present ligands with basic and/or Tyr PΩ residues in low amounts, detectable by CTL but not easily by biochemical methods. In support of this possibility, a B*2705 ligand with C-terminal tyrosine (RRFFPYYVY) (26) was found in much smaller amounts in the B*2709-bound peptide pool (our unpublished observations). Thus, a comparison of peptide motifs is insufficient to define the functional and antigenic similarity of structurally close class I proteins.

The lower cross-reactivity of anti-B*2703 CTL with B*2709, relative to anti-B*2705 CTL, can be explained by two related reasons. First, B*2703 probably binds and presents to CTL mainly a subset of the B*2705-bound peptides (27, 36) consisting of those with the canonic B*2705 motifs at P2 and PΩ, plus a basic P1 residue (16, 17). Because B*2709 has a more restricted PΩ motif than B*2705, it probably shares less natural ligands with B*2703 than with B*2705. In addition, the added effect of the two changes between B*2703 and B*2709 (H59Y and D116H) on the conformation of some peptide epitopes might be larger than the single H59Y change between B*2703 and B*2705.

It is interesting that whereas the D116H change occurs alone in B*2709, the more disruptive D116Y change always occurs together with additional changes in HLA-B27 subtypes. At least in some cases, these additional changes, such as Y114D in B*2706, have compensatory effects on allorecognition (36). The significance of this observation is unclear, but it is compatible with the idea that HLA-B27 has evolved to diversify the structure of the C/F pocket, but avoiding too large disruptive effects at position 116, a critical one for this pocket. This would have been achieved either by introducing a single mutation with relatively moderate effects, as in B*2709, or by introducing additional compensatory changes.

Although, obviously, alloreactive CTL are not related to B27-mediated spondyloarthropathy, the findings in this study might be relevant to the understanding of the reported differential association of B*2705 and B*2709 to AS (9). They suggest that both subtypes may interchangeably act as restriction elements for many peptides. If the pathogenetic role of HLA-B27 is related to its normal Ag-presenting function (10), one would expect that those subtypes associated to AS, such as B*2705, B*2702, or B*2704, are able to present the relevant peptide(s), whereas subtypes apparently not or less associated to this disease, such as B*2709, are not. However, there is no correlation between the Ag-presenting similarities among B27 subtypes, as assessed with allospecific CTL, and their reported association to AS. For instance, whereas 80% of our anti-B*2705 CTL cross-react with B*2709, only 12% cross-react with B*2704 (27). The fact that B*2709 has lost a relatively small amount of B*2705-allospecific T cell epitopes considerably narrows down the number of B27-bound peptides, that are antigenic in the B*2705 but not in the B*2709 context. Presumably, an “arthritogenic” peptide might have this feature. Therefore, our results open the way to further characterization of peptides putatively involved in HLA-B27-mediated disease, focusing on a limited fraction of the B*2705-bound repertoire.

The relevance of our results to the understanding of the pathogenic role of HLA-B27 obviously depends on the strength of the evidence supporting that B*2709 is more weakly associated to AS than B*2705. Therefore, this issue deserves to be discussed here.

Because B*2709 has been found with a frequency high enough for statistical analysis only in Sardinia, the idea of its differential association to AS is based on findings in this population (9), described in the Introduction. In contrast to B*2706, in which an initial study reporting lack of association of this subtype in Thailand (6) was subsequently confirmed in two additional populations of Southeast Asia (7, 8), no statistical studies in other populations have thus far been possible for B*2709. With this limitation in mind, we must nevertheless point out that the critical point in the Sardinian study, as well as in those dealing with B*2706, is not only the absence of B*2709+ patients, but the fact that other HLA-B27 subtypes are normally associated to AS in the same population. This makes it unlikely that population-specific genetic or environmental factors account for the observed differential association, because such factors would have to act differently on individuals carrying different subtypes in the same population. In addition, although populations differ among each other genetically, environmental factors influencing HLA-B27-associated disease are probably ubiquitous (47). The situation with B*2709 and B*2706 is totally different from the reported lack of association of B*2703 with AS in West Africa (48); unknown genetic factors seem to protect from AS both B*2703 and B*2705 individuals in this population, where AS is extremely rare regardless of the subtype (49).

The occasional B*2709 or B*2706 individuals with AS reported from populations where these subtypes are in very low frequency indicates that lack of association with AS is not absolute. However, this is not in conflict with differential subtype associations established by statistical analysis within a population. It is well known that a low percentage of HLA-B27 negative individuals also develop AS. In conclusion, although the negative association of B*2709 with AS should be confirmed in other populations, the findings in the only population where it is found with significant frequency are not easily explained by genetic or environmental factors alien to HLA-B27 itself.

We thank Dr. J. L. Vicario (Centro de Transfusiones de la Comunidad Autónoma de Madrid) for cell typing and blood reagents, Dr. Rosa Sorrentino (University of L’Aquila, Italy) for providing the B*2709 cDNA and the Ci LCL, and Dr. David Yu (UCLA) for the B*2705-T2 transfectant cell line.

1

This work was supported by Grants SAF97/0182 from the Plan Nacional de I+D, and PM95-002 from the Spanish Ministry of Education. We thank the Fundación Ramón Areces for an institutional grant to the Centro de Biología Molecular Severo Ochoa.

3

Abbreviations used in this paper: C1R, Hmy2.C1R; AS, ankylosing spondylitis; PΩ, C-terminal peptide position; LCL, lymphoblastoid cell line.

1
Brewerton, D. A., M. Caffrey, F. D. Hart, D. C. O. James, A. Nichols, R. D. Sturrock.
1973
. Ankylosing spondylitis and HL-A27.
Lancet
1
:
904
2
Schlosstein, L., P. I. Terasaki, R. Bluestone, C. M. Pearson.
1973
. High association of an HL-A antigen, W27, with ankylosing spondylitis.
N. Engl. J. Med.
288
:
704
3
Kingsley, G., J. Sieper.
1993
. Current perspectives in reactive arthritis.
Immunol. Today
14
:
387
4
Lopez de Castro, J. A..
1998
. The pathogenetic role of HLA-B27 in chronic arthritis.
Curr. Opin. Immunol.
10
:
59
5
Gonzalez-Roces, S., M. V. Alvarez, S. Gonzalez, A. Dieye, H. Makni, D. G. Woodfield, L. Housan, V. Konenkov, M. C. Abbadi, N. Grunnet, E. Coto, C. Lopez-Larrea.
1997
. HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis.
Tissue Antigens
49
:
116
6
Lopez-Larrea, C., K. Sujirachato, N. K. Mehra, P. Chiewsilp, D. Isarangkura, U. Kanga, O. Dominguez, E. Coto, M. Pena, F. Setien.
1995
. HLA-B27 subtypes in Asian patients with ankylosing spondylitis: evidence for new associations.
Tissue Antigens
45
:
169
7
Nasution, A. R., A. Mardjuadi, S. Kunmartini, N. G. Suryadhana, B. Setyohadi, D. Sudarsono, N. M. Lardy, T. E. Feltkamp.
1997
. HLA-B27 subtypes positively and negatively associated with spondyloarthropathy.
J. Rheumatol.
24
:
1111
8
Ren, E. C., W. H. Koh, D. Sim, M. L. Boey, G. B. Wee, S. H. Chan.
1997
. Possible protective role of HLA-B*2706 for ankylosing spondylitis.
Tissue Antigens
49
:
67
9
D’Amato, M., M. T. Fiorillo, C. Carcassi, A. Mathieu, A. Zuccarelli, P. P. Bitti, R. Tosi, R. Sorrentino.
1995
. Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis.
Eur. J. Immunol.
25
:
3199
10
Benjamin, R., P. Parham.
1990
. Guilt by association: HLA-B27 and ankylosing spondylitis.
Immunol. Today
11
:
137
11
Zhou, M., A. Sayad, W. A. Simmons, R. C. Jones, S. D. Maika, N. Satumtira, M. L. Dorris, S. J. Gaskell, R. S. Bordoli, R. B. Sartor, C. A. Slaughter, J. A. Richardson, R. E. Hammer, J. D. Taurog.
1998
. The specificity of peptides bound to human histocompatibility leukocyte antigen (HLA)-B27 influences the prevalence of arthritis in HLA-B27 transgenic rats.
J. Exp. Med.
188
:
877
12
Lopez de Castro, J. A..
1995
. Structural polymorphism and function of HLA-B27.
Curr. Opin. Rheumatol.
7
:
270
13
Jardetzky, T. S., W. S. Lane, R. A. Robinson, D. R. Madden, D. C. Wiley.
1991
. Identification of self peptides bound to purified HLA-B27.
Nature
353
:
326
14
Rotzschke, O., K. Falk, S. Stevanovic, V. Gnau, G. Jung, H. G. Rammensee.
1994
. Dominant aromatic/aliphatic C-terminal anchor in HLA-B*2702 and B*2705 peptide motifs.
Immunogenetics
39
:
74
15
Garcia, F., B. Galocha, J. A. Villadangos, J. R. Lamas, J. P. Albar, A. Marina, J. A. Lopez de Castro.
1997
. HLA-B27 (B*2701) specificity for peptides lacking Arg2 is determined by polymorphism outside the B pocket.
Tissue Antigens
49
:
580
16
Boisgerault, F., V. Tieng, M. C. Stolzenberg, N. Dulphy, I. Khalil, R. Tamouza, D. Charron, A. Toubert.
1996
. Differences in endogenous peptides presented by HLA-B*2705 and B*2703 allelic variants: implications for susceptibility to spondylarthropathies.
J. Clin. Invest.
98
:
2764
17
Griffin, T. A., J. Yuan, T. Friede, S. Stevanovic, K. Ariyoshi, S. L. Rowland-Jones, H. G. Rammensee, R. A. Colbert.
1997
. Naturally occurring A pocket polymorphism in HLA-B*2703 increases the dependence on an accessory anchor residue at P1 for optimal binding of nonamer peptides.
J. Immunol.
159
:
4887
18
Garcia, F., A. Marina, J. A. Lopez de Castro.
1997
. Lack of carboxyl-terminal tyrosine distinguishes the B*2706-bound peptide repertoire from those of B*2704 and other HLA-B27 subtypes associated with ankylosing spondylitis.
Tissue Antigens
49
:
215
19
Tieng, V., N. Dulphy, F. Boisgerault, R. Tamouza, D. Charron, A. Toubert.
1997
. HLA-B*2707 peptide motif: Tyr C-terminal anchor is not shared by all disease-associated subtypes.
Immunogenetics
47
:
103
20
Fiorillo, M. T., L. Meadows, M. Damato, E. Appella, R. Sorrentino.
1997
. Susceptibility to ankylosing spondylitis correlates with the C-terminal residue of peptides presented by various HLA-B27 subtypes.
Eur. J. Immunol.
27
:
368
21
Garcia, F., D. Rognan, J. R. Lamas, A. Marina, J. A. Lopez de Castro.
1998
. An HLA-B27 polymorphism (B*2710) that is critical for T-cell recognition has limited effects on peptide specificity.
Tissue Antigens
51
:
1
22
Brooks, J. M., R. J. Murray, W. A. Thomas, M. G. Kurilla, A. B. Rickinson.
1993
. Different HLA-B27 subtypes present the same immunodominant Epstein-Barr virus peptide.
J. Exp. Med.
178
:
879
23
Brooks, J. M., R. A. Colbert, J. P. Mear, A. M. Leese, A. B. Rickinson.
1998
. HLA-B27 subtype polymorphism and CTL epitope choice: studies with EBV peptides link immunogenicity with stability of the B27: peptide complex.
J. Immunol.
161
:
5252
24
Lamas, J. R., J. M. Brooks, B. Galocha, A. B. Rickinson, J. A. Lopez de Castro.
1998
. Relationship between peptide binding and T cell epitope selection: a study with subtypes of HLA-B27.
Int. Immunol.
10
:
259
25
Rowland-Jones, S., R. A. Colbert, T. Dong, S. McAdam, M. Brown, K. Ariyoshi, S. Sabally, H. Whittle, A. McMichael.
1998
. Distinct recognition of closely-related HIV-1 and HIV-2 cytotoxic T- cell epitopes presented by HLA-B*2703 and B*2705.
AIDS
12
:
1391
26
Paradela, A., M. Garcia-Peydro, J. Vazquez, D. Rognan, J. A. Lopez de Castro.
1998
. The same natural ligand is involved in allorecognition of multiple HLA-B27 subtypes by a single T cell clone: role of peptide and the MHC molecule in alloreactivity.
J. Immunol.
161
:
5481
27
Lopez, D., R. Garcia Hoyo, J. A. Lopez de Castro.
1994
. Clonal analysis of alloreactive T cell responses against the closely related B*2705 and B*2703 subtypes: implications for HLA- B27 association to spondyloarthropathy.
J. Immunol.
152
:
5557
28
Del Porto, P., M. D’Amato, M. T. Fiorillo, L. Tuosto, E. Piccolella, R. Sorrentino.
1994
. Identification of a novel HLA-B27 subtype by restriction analysis of a cytotoxic γδ T cell clone.
J. Immunol.
153
:
3093
29
Fiorillo, M. T., G. Greco, M. Maragno, I. Potolicchio, A. Monizio, M. L. Dupuis, R. Sorrentino.
1998
. The naturally occurring polymorphism Asp116→His116, differentiating the ankylosing spondylitis-associated HLA-B*2705 from the non-associated HLA-B*2709 subtype, influences peptide-specific CD8 T cell recognition.
Eur. J. Immunol.
28
:
2508
30
Calvo, V., S. Rojo, D. Lopez, B. Galocha, J. A. Lopez de Castro.
1990
. Structure and diversity of HLA-B27-specific T cell epitopes: analysis with site-directed mutants mimicking HLA-B27 subtype polymorphism.
J. Immunol.
144
:
4038
31
Villadangos, J. A., B. Galocha, J. A. Lopez de Castro.
1994
. Unusual topology of an HLA-B27 allospecific T cell epitope lacking peptide specificity.
J. Immunol.
152
:
2317
32
Ellis, S. A., C. Taylor, A. McMichael.
1982
. Recognition of HLA-B27 and related antigen by a monoclonal antibody.
Hum. Immunol.
5
:
49
33
Villadangos, J. A., B. Galocha, F. Garcia, J. P. Albar, J. A. Lopez de Castro.
1995
. Modulation of peptide binding by HLA-B27 polymorphism in pockets A and B, and peptide specificity of B*2703.
Eur. J. Immunol.
25
:
2370
34
Jaraquemada, D., B. Galocha, P. Aparicio, S. Rojo, V. Calvo, J. A. Lopez de Castro.
1988
. Modulation on immunogenicity by HLA-B27 subtype polymorphism.
Eur. J. Immunol.
18
:
1945
35
Aparicio, P., D. Jaraquemada, J. A. Lopez de Castro.
1987
. Alloreactive cytolytic T cell clones with dual recognition of HLA-B27 and HLA-DR2 antigens: selective involvement of CD8 in their class I-directed cytotoxicity.
J. Exp. Med.
165
:
428
36
Villadangos, J. A., B. Galocha, R. Garcia-Hoyo, D. Lopez, F. Garcia, J. A. Lopez de Castro.
1994
. Structure of HLA-B27-specific T cell epitopes. Antigen presentation in B*2703 is limited mostly to a subset of the antigenic determinants on B*2705.
Eur. J. Immunol.
24
:
2548
37
Villadangos, J. A. 1994. Papel de residuos conservados y polimórficos de HLA-B27 en el reconocimiento por células T y en la unión de péptidos. Doctoral dissertation, Universidad Autónoma de Madrid, Spain.
38
Rammensee, H. G., T. Friede, S. Stevanovic.
1995
. MHC ligands and peptide motifs: first listing.
Immunogenetics
41
:
178
39
Lamas, J. R., A. Paradela, F. Roncal, and J. A. López de Castro. 1999. The peptide specificity of HLA-B27 subtypes differentially associated to ankylosing spondylitis is modulated at multiple anchor residues. Arthritis Rheum. In press.
40
Tallquist, M. D., L. R. Pease.
1995
. Alloreactive 2C T cells recognize a self peptide in the context of the mutant Kbm3 molecule.
J. Immunol.
155
:
2419
41
Tallquist, M. D., T. J. Yun, L. R. Pease.
1996
. A single T cell receptor recognizes structurally distinct MHC/peptide complexes with high specificity.
J. Exp. Med.
184
:
1017
42
Zhao, R., D. J. Loftus, E. Appella, E. J. Collins.
1999
. Structural evidence of T cell xeno-reactivity in the absence of molecular mimicry.
J. Exp. Med.
189
:
359
43
Huet, S., D. F. Nixon, J. B. Rothbard, A. Townsend, S. A. Ellis, A. J. McMichael.
1990
. Structural homologies between two HLA B27-restricted peptides suggest residues important for interaction with HLA B27.
Int. Immunol.
2
:
311
44
Van Binnendijk, R. S., J. P. Versteeg van Oosten, M. C. Poelen, H. F. Brugghe, P. Hoogerhout, A. D. Osterhaus, F. G. UytdeHaag.
1993
. Human HLA class I- and HLA class II-restricted cloned cytotoxic T lymphocytes identify a cluster of epitopes on the measles virus fusion protein.
J. Virol.
67
:
2276
45
Brander, C., G. Corradin, T. Hasler, W. J. Pichler.
1996
. Peptide immunization in humans: a combined CD8+/CD4+ T cell-targeted vaccine restimulates the memory CD4 T cell response but fails to induce cytotoxic T lymphocytes (CTL).
Clin. Exp. Immunol.
105
:
18
46
Goulder, P. J., R. E. Phillips, R. A. Colbert, S. McAdam, G. Ogg, M. A. Nowak, P. Giangrande, G. Luzzi, B. Morgan, A. Edwards, A. J. McMichael, S. Rowland-Jones.
1997
. Late escape from an immunodominant cytotoxic T-lymphocyte response associated with progression to AIDS.
Nat. Med.
3
:
212
47
Brown, M. A., L. G. Kennedy, A. J. MacGregor, C. Darke, E. Duncan, J. L. Shatford, A. Taylor, A. Calin, P. Wordsworth.
1997
. Susceptibility to ankylosing spondylitis in twins. The role of genes, HLA, and the environment.
Arthritis Rheum.
40
:
1823
48
Hill, A. V., C. E. Allsopp, D. Kwiatkowski, M. N. Anstey, Greenwood B. M., A. J. McMichael.
1991
. HLA class I typing by PCR: HLA-B27 and African B27 subtype.
Lancet
342
:
646
49
Brown, M., A. Jepson, A. Young, H. Whittle, B. Greenwood, P. Wordsworth.
1997
. Spondyloarthritis in West Africa: evidence for a non-B27 protective effect.
Ann. Rheum. Dis.
56
:
68
50
Marchionni, L., V. Modena, R. Roggero, I. Ariata Maiocco, G. Mazzola, C. Lomater, M. Berrino, F. Btancatello, and E. S. Curtoni. 1999. The polymorphism of HLA-B27 and the seronegative spondyloarthropathies in the Italian population [abstract]. In 13th European Histocompatibility Conference, Crete, 1999. European Federation for Immunogenetics, p.36.