HLA-B27 subtypes: implications for the spondyloarthropathies

Spondyloarthropathies and arthritis post-infection: a historical perspective

The spondyloarthropathies are a group of conditions characterised by spinal joint pain and have related clinical, epidemiological and genetic-related features. Ankylosing spondylitis, reactive arthritis, the spinal form of psoriatic arthritis and Crohn's and colitis enteropathic arthritis are the major clinical entities of the spondyloarthropathies, and principally occur in HLA-B27 positive individuals. Ankylosing spondylitis is much more common in males than females. Patients are usually seronegative for rheumatoid factor, and extra-articular features including iridocyclitis, mucous membrane and skin lesions: aortitis, may occur in some patients. The reactive arthritis form classically occurs following an infection of the gastrointestinal or genitourinary tract. The Crohn's and colitis enteropathic arthritis forms often have an associated large joint asymmetrical arthritis. Also discussed are acute rheumatic fever and Lyme disease which are conditions where the individual develops arthritis after an infection.

Subtypes of HLA-B27: history and implications in the pathogenesis of ankylosing spondylitis

HLA-B27 represents a family of 38 closely related cell surface proteins (encoded by the alleles HLA-B*2701-39) called subtypes of HLA-B27, most of which have evolved from the ubiquitous HLA-B*2705 (specifically the B*27052 allele). HLA-B27 subtypes are largely characterized by nucleotide substitutions (mostly nonsynonymous) in exons 2 and 3 which encode alpha1 and alpha2 domains ofthe peptide binding groove respectively. Table 1 shows the description of sequences of HLA-B27 allele sequences. The subtypes could have arisen from B*2705 by point mutation (B*2703, B*2709, B*2704), gene conversion (B*2701, B*2702, B*2708) and reciprocal recombination (B*2707) B*2706 could have arisen by interlocus gene conversion. Studies from different parts of the world reveal differences in the population distribution.

The effect of an anti-HLA-B27 immune response on CTL recognition of Chlamydia

The interplay between triggering bacteria and HLA-B27 in the pathogenesis of the spondyloarthropathies remains one of the most active areas of investigation in the rheumatic diseases. This has proved difficult to study systematically in the clinical setting, and in this study we utilized a rat model to address the influence that B27-related immunity may have on the process of generating anti-Chlamydia immunity. When splenocytes from HLA-B27 DNA-immunized Lewis (LEW) animals received restimulation in vitro with Chlamydia-treated cells from B27-transgenic LEW rats, we observed that in addition to the expected CTL recognition of HLA-B27, there was also anti-Chlamydia CTL killing of Chlamydia-sensitized syngeneic fibroblast targets. This was not seen when responding cells in vitro were naive LEW splenocytes. To confirm the existence of CTLs recognizing both HLA-B27 and Chlamydia, LEW rats were immunized with B27-transgenic LEW cells, instead of the B27 DNA construct. Splenocytes from the immune rats were restimulated in vitro with Chlamydia-treated B27-transgenic LEW cells. In this instance, the CTLs retained the allele-specific recognition of HLA-B27, as well as recognition of Chlamydia-sensitized syngeneic fibroblasts. Thus, if there is prior expansion of an immune response against HLA-B27, then the resulting splenocytes demonstrate a reduced threshold for generating a primary anti-Chlamydia CTL response. These studies implicate a dynamic interrelationship between recognition of HLA-B27 and Chlamydia trachomatis. The results may have implications for deciphering the cellular basis of Chlamydia-induced reactive arthritis.

Susceptibility to ankylosing spondylitis is independent of the Bw4 and Bw6 epitopes of HLA-B27 alleles

We have characterized HLA-B27 alleles in a sample of the population from the Azores (n=46) with the aim of investigating the contribution of different subtypes to ankylosing spondylitis (AS). The study was carried out using PCR-SSOP and in some samples genomic sequencing was conducted. Some significant new finding have arisen from this study. First, B*2705,B*2702,B*2703,B*2707 and B*2708 alleles were found to be represented in this population. The polymorphism of B27 alleles found in a sample of the population from the Azores is higher than the Caucasian groups described. B*2703 and B*2707 have not previously been described to be represented in Caucasians and this could indicate admixtures with different populations of the world. In addition, the B*2708 allele was found to be associated with AS in a large family from the Azores. This association has not been previously reported in either ethnic group and needs to be confirmed in other population studies. This is of considerable interest since has only been described as a rare subtype underrepresented in the British population and has not been previously found to be associated with AS. B*2708 carries the sequence specifying the Bw6 epitope in contrast to most B27 alleles which carry a Bw4 sequence. Differences in this region (residues 77-83) can alter the F-pocket and affect T-cell recognition. The importance that these molecular changes can play in the pathogenesis of AS is discussed.

HLA-B*27 subtyping by PCR-RFLP in Spanish patients with ankylosing spondylitis

We describe the use of restriction analysis on PCR-amplified DNA for detecting all B*27 subtypes except B*2710 and B*2711 (i.e. from B*2701 to B*2709). After detecting B*27 by Sty I, double digestions consisting of Sty I plus another informative enzyme led to subtype assignment. We used mismatched primers to create restriction sites when necessary. The method avoids group-specific amplifications and other laborious optimization procedures. It was successfully tested on a panel of well characterized cell lines covering different B*27 subtypes. Then, we studied a group of 57 ankylosing spondylitis patients and 746 controls from the south of Spain. B*27 showed a very strong association with the disease (OR=211.27, P=10(-7)). B*2702 and B*2705 distribution in controls (20% and 77.1%, respectively) differed from previously reported data in the Spanish population. We unexpectedly found the B*2707 allele in our population (one control).

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

HLA-B27 molecules are interesting because of their strong association with ankylosing spondylitis (AS) and reactive arthritis (ReA). A pathogenetic role for these molecules has been postulated in presenting a putative "arthritogenic" peptide to CD8 T cells. The HLA-B*2709 subtype, although differing by a single amino acid (His116-->Asp116) from the widespread and strongly AS-associated subtype HLA-B*2705, is not found in patients. Since residue 116 interacts with the C terminus of the peptide, it is possible that the two subtypes differ in their antigen-presenting features. We show here that CD8 T cells can distinguish the two HLA-B27 subtypes when presenting a same epitope derived from Epstein-Barr virus-latent membrane protein 2. Moreover, alanine scanning mutagenesis analysis revealed that the peptide residues relevant for such recognition are different depending on whether HLA-B*2705 or -B*2709 molecules present the epitope. These results give support to the belief that functional differences determined by subtype-specific polymorphisms can have a pathogenetic relevance and open up a new scenario where subtle modifications within the peptide/HLA ligand might be responsible for the differential association between HLA-B27 subtypes and spondyloarthropathies.

HLA-B27 (B*2701) specificity for peptides lacking Arg2 is determined by polymorphism outside the B pocket

B*2701 differs from B*2705-by three amino acid changes: D-->Y74, D-->N77, L-->A81, and from B*2702 only by two: D-->Y74 and T-->I80. Tyr74 is located in the C/F cavity of the peptide-binding site, and is unique to B*2701 among HLA-B27 subtypes. Binding of natural B*2705 and B*2702 ligands to B*2701, and to mutants mimicking subtype changes, was analyzed. In addition, sequencing of the peptides bound in vivo by B*2701 and the Y74 mutant was carried out. The main distinctive feature of B*2701 was its presentation of peptides with Gln2. Synthetic analogs bound in vitro similarly as the corresponding ligands with Arg2. Moreover, both Gln2 and Arg2 were dominant upon pool sequencing of B*2701-bound peptides, and 2 of 8 natural ligands contained Gln2. Suitability of Gln2 was largely determined by the Y74 change, as indicated by: 1) binding of Gln2 analogs to this mutant, and 2) detection of Gln2 by pool sequencing of Y74-bound peptides. B*2701 bound peptides with C-terminal aromatic or Leu residues, and interacted with these motifs more strongly than B*2702. The Y74 mutation alone was not responsible for poor binding of peptides with C-terminal basic residues to B*2701, since they bound efficiently and at least one was presented in vivo by this mutant. Most peptides bound to the A81 mutant worse than to B*2705, but frequently better than to B*2701 or B*2702, suggesting that other subtype changes were compensatory. The peptide specificity of B*2701 suggests that this subtype may determine susceptibility to spondyloarthropathy.

HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis

HLA-B27 is strongly associated to ankylosing spondylitis (AS) and represents a family of eleven B27 alleles (B*2701-11). Our aim was to analyze the distribution of B27 subtypes by PCR/SSOP and genomic sequencing in a large group of populations (n = 17). 711 B27-positive samples from Caucasoid, Asian, African, Amerindian and Polynesian populations were selected to ascertain transracial gene mapping of the B27 subtypes. 476 of these were AS patients, chosen to investigate the contribution of B27 alleles to AS susceptibility. Some significant new findings have arisen from this study: 1) B*2705 was the predominant subtype in circumpolar and subarctic areas. B*2702 was found to be practically restricted to Caucasian populations, showing a higher frequency in Middle-East (Jews) and North Africa (Arabs/Berbers) groups. 2) B*2703 appears associated with AS in Western Africans. This is of remarkable interest since it was suggested that B*2703 would be negatively disease-associated. 3) Although B*2706 appears negatively associated with AS in Thais, we identified two patients from northern China carrying it. This may be a reflection of a disease heterogeneity and could indicate that more than one pathogenic agent can be involved in AS. B*2709 has been recently described as negatively associated with AS in Sardinians. The molecular changes His114Asp (B*2706) and Asp116His (B*2709) could modify the genetic susceptibility to AS.

HLA-B27 presents a peptide from a polymorphic region of its own molecule with homology to proteins from arthritogenic bacteria

A possible mechanism for the pathogenesis of HLA-B27-associated spondyloarthropathies is that peptides from arthritogenic bacteria with homology to endogenous self-peptides presented by HLA-B27, including those derived from HLA-B27 itself, could elicit an autoimmune T-cell response upon infection. We report here that an undecamer corresponding to the polymorphic region of HLA-B27 spanning residues 169-179 is presented in vivo by the B*2701, B*2704 and B*2706 subtypes, but was not detected in the B*2703-bound peptide pool. This peptide binds to B*2705 in vitro with sufficient affinity to allow its natural presentation by this subtype, but it binds with low affinity to B*2703. In spite of homology of this peptide to proteins from arthritogenic bacteria, its binding specificity does not correlate with current evidence concerning association of HLA-B27 subtypes to ankylosing spondylitis, suggesting that presentation of this peptide is not the critical feature that determines linkage of HLA-B27 to this disease.

Automated routine HLA-B27 typing by flow cytometry

An increasing demand for HLA-B27 typing as one of the tests used in the diagnosis of ankylosing spondylitis had led us to develop a rapid, automated, flow cytometric assay using whole blood and an HLA-B27 specific monoclonal antibody FD705. This article shows the data from 2093 samples tested during a 2 year period of routine HLA-B27 typing. 21.6% were clearly HLA-B27 positive whilst 73.2% were HLA-B27 negative, the remaining 5.2% required further testing before assignment of HLA-B27 status. Additional work was carried out on blood samples from individuals positive for the newly described subtype HLA-B2708.

Seronegative spondyloarthropathies and HLA-B27 subtyes: a study in Asian Indians

In this study, 60 HLA-B27+ve SSA patients and 17 healthy controls belonging to North India were analyzed to ascertain heterogeneity of the B27 molecule in this population. ID-IEF and PCR-SSOP technologies were used to analyze polymorphism in exon 2 and 3 of the HLA-B27 gene. Four different subtypes were encountered: B*2702,04,05 and 07. Other subtypes of B27 viz B*2701,03,06 and 08 were not encountered. B*2704 (common oriental subtype) and B*2705 (common Caucasian subtype) were the most common subtypes in the control and patient groups. B*2707 was less frequently encountered in both groups and B*2702 was found in only one AAU patient. B*2704 was the predominant subtype in the AS group (70.8%) compared to its frequency of 47% in healthy controls (RR = 2.73) while in the undiff SpA group, B*2705 occurred most frequently (73.1%, RR = 3.05). B27 subtypes segregated differently in males and females. 12 of the 17 male AS patients carried B*2704 as compared to 1 of 8 healthy males (X2 = 3.9, P < 0.05). On the other hand, in the undiff SpA, B*2705 was significantly raised in female patients (100%) as compared to healthy females (22.2%, X2 = 4.9, P < 0.05). Subtype distribution is indicative of racial admixture in the Asian Indian population.

Clinical features and associated systemic diseases of HLA-B27 uveitis

PURPOSE

To delineate the clinical features, course, complications, and associated systemic diseases in patients with HLA-B27-associated uveitis.

METHODS

We reviewed the records of 148 patients with HLA-B27-associated uveitis from two large uveitis practices.

RESULTS

There were 127 (86%) white and 21 (14%) nonwhite patients, and a male-to-female ratio of 1.5:1. The median age at onset of uveitis was 32 years; eight patients (5%) had their first attack after age 55 years. Acute anterior uveitis was noted in 129 patients (87%), and nonacute inflammation was noted in 19 (13%). Ocular involvement was categorized as unilateral or unilateral alternating in 138 patients (93%), but ten patients (7%) had bilateral, concurrent disease. The median duration of an attack was six weeks, and the median number of recurrences for patients with more than 12 months of follow-up was three. Cataracts were associated with posterior synechiae (P = .03), increased intraocular pressure (P = .003), and cystoid macular edema (P = .04). An HLA-B27-associated systemic disorder was present in 83 patients (58%), 30 of whom were women, and it was diagnosed in 43 of the 83 patients as a result of the ophthalmologic consultation. Thirty-four (30%) of 112 patients had a family history of a spondyloarthropathy.

CONCLUSIONS

Although HLA-B27-associated uveitis is usually described as a disease of young white men, women and nonwhites may also be affected. A subgroup of patients have severe disease and consequently more complications. Most patients have an associated systemic disease, including women, who appear to have more atypical spondyloarthropathies. The systemic diseases were frequently undiagnosed before the onset of the ocular disease and before the uveitis consultation.

Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis

Ankylosing spondylitis (AS) is an autoimmune disorder strongly associated with HLA-B27. A direct role of B27 molecules in the disease pathogenesis has been postulated, possibly by presenting to T cells an as-yet unidentified arthritogenic peptide that triggers the autoimmune response. There are nine HLA-B27 alleles differing from each other at one or more amino acid positions. It is important, for the identification of the arthritogenic peptide, to define which alleles, and therefore which polymorphic positions, predispose to the disease. Here, we report that HLA-B*2709 is not associated with AS, as it was not found in patients. HLA-B*2709 differs from the most frequent and disease-associated HLA-B*2705 allele for a single substitution (His vs. Asp) at position 116. Amino acid 116 is located at the bottom of the groove where the antigenic peptide sits, and it has been proven to influence the peptide-binding specificity of HLA class I molecules. The most likely interpretation of these data is that the differences in charge and size that accompany the His-to-Asp substitution exclude the acceptance of the arthritogenic peptide.

HLA-B27 subtypes in Asian patients with ankylosing spondylitis. Evidence for new associations

The aim of this study was to investigate the contribution of the different B27 subtypes to ankylosing spondylitis (AS) susceptibility. The polymerase chain reaction (PCR) in combination with the sequence-specific oligonucleotide probes (SSOs) was used to analyse the polymorphism in exon 2 and 3 of HLA-B27 in two Asian groups with different genetic HLA structures: Indian (I) and Thai (T) populations. The same number of AS patients (45) and healthy B27 positive donors (n = 17) from both populations were analysed in order to ascertain the B27 subtypes. Three different findings can be concluded from this study: 1) B*2707 has been found to be associated with AS in both populations. This association has not been previously reported in either ethnic group. 2) B*2704 is strongly associated with AS in the Thai patients (91% in AS vs. 47% in C; RR = 11.5; EF = 0.83). In contrast, B*2704 was found with similar frequency in Asian Indians AS patients and controls (41% in AS vs. 41% in C.). 3) B*2706 was found overrepresented in control populations and absent in AS patients (0% in AS vs. 47% in C.; pc < 10(-6)) showing the maximum value of protective fraction (PF = 1). The B*2706 negative association with AS has not been previously described in other ethnic groups and could indicate a protective effect of this subtype on AS susceptibility. The B*2706 allele has two changes relative to B*2704 at residue 114 (His to Asp) and 116 (Asp to Tyr) in the pockets D/E. The importance that these differences can play in the pathogenesis of AS are discussed.

Reiter's syndrome and reactive arthritis: a current view

This paper reviews advances in the understanding of the pathogenesis of reactive arthritis that have occurred over the last decade. Inflammatory aseptic joint disease has been linked with prior infection initiated by many different species of microorganisms. The presence of intra-articular bacterial antigens has now been firmly established with the demonstration of bacteria, bacterial fragments, DNA, RNA, and bacterial lipopolysaccharide in joints of patients with reactive arthritis. Chlamydia trachomatis, Salmonella enteritidis, and Shigella flexneri have all been detected in the joint by immunological techniques, although there is still some doubt as to the form in which they reach the joint and whether or not they persist. A number of phlogistic bacterial components could be acting as arthritogens. Negative joint culture results from patients with reactive arthritis make it unlikely that bacteria in the joint are viable, although chlamydial DNA has been shown in the joints of patients with sexually acquired reactive arthritis using the polymerase chain reaction. The use of antimicrobial therapy in the treatment of reactive arthritis is under review; data suggests that long-term antibiotic treatment warrants further study. The role of HLA-B27 in disease pathogenesis is discussed as are possible mechanisms of interplay between germ and gene. HLA-B27 might confer disease susceptibility by affecting immune mechanisms other than classical antigen presentation. The immunopathogenesis of joint inflammation in reactive arthritis is explored with reference to studies of humoral and cellular immune responses. Serological evidence to support the concept of molecular mimicry is far from conclusive; the results of relevant studies are summarized. Lymphocyte proliferation experiments suggest that antigen presenting cells play an important role. Finally, our views on reactive arthritis in the 1990s, and areas of new and potentially fruitful future research are presented.

Comparison of the antibody responses to the 77 Klebsiella capsular types in ankylosing spondylitis and various rheumatic diseases

The production of antibodies to Klebsiella capsular polysaccharides was measured in sera from either HLA-B27-positive (HLA-B27+) or HLA-B27-negative (HLA-B27-) patients with classical ankylosing spondylitis (n = 54). These sera were compared with sera from patients with various rheumatic diseases (n = 82) and HLA-B27+ or HLA-B27- healthy individuals (n = 85). All sera were analyzed by means of an enzyme-linked immunosorbent assay specific to each of the 77 Klebsiella serotypes. The sera from HLA-B27+ patients with ankylosing spondylitis showed a significantly higher antibody frequency to the capsular types K26, K36, and K50 than the sera from HLA-B27- ankylosing spondylitis patients, patients with psoriatic arthritis, systemic lupus erythematosus, rheumatoid arthritis, or reactive arthritis after Yersinia enterocolitica infection, or healthy controls (P < 0.02). The antibodies were of the immunoglobulin G type. No significant antibody response to the other 74 Klebsiella serotypes, noncapsulated mutants of K26, K36, and K50, or preparations of Citrobacter, Serratia, Hafnia, or Morganella spp. or Streptococcus pneumoniae could be detected. The results might suggest a specific association between these capsular types and HLA-B27+ ankylosing spondylitis and might imply their predominance in this disease.

The peptide binding specificity of HLA-B27 subtypes

Five HLA-B27 subtypes, B*2701, B*2703, B*2704, B*2705, and B*2706, were tested for direct binding with twenty-six synthetic nonapeptides carrying the primary anchor residue motifs (combination of amino residues at positions 2 and 9) relevant to B*2705. The peptide sequences were derived from human HSP89 alpha, P53 and MBP. The alpha chains were immunospecifically isolated from LH (B*2701), CH (B*2703), WE1 (B*2704), BTB (B*2705), and LIE (B*2706) cells and their peptide binding was measured by the HLA class I alpha chain refolding assay. The data obtained indicated that the B27 subtypes tested can bind a common set of peptides carrying several different anchor residue motifs. The motifs, R-K and R-R, reported for B*2705 and a new motif H-R were accepted by B*2703, B*2704, and B*2706, but not by B*2701. However, other motifs, including known B*2702 and/or B*2705 motifs, R-H, R-L, R-A, and R-F, and a new motif found here, R-G, were apparently accepted by all B27 subtypes tested. The observed cross-peptide binding in the B27 subgroup is compatible with the so-called arthritogenic peptide hypothesis in the pathogenesis of ankylosing spondylitis.

Automated HLA-B27 testing using the FACSPrep/FACScan system

Tissue typing can help in the diagnosis of the seronegative arthropathy ankylosing spondylitis. Using an automatic sample preparation system and flow cytometry (FACSPrep/FACScan) we have developed a test for HLA-B27 screening using whole blood which is both rapid, reproducible, and permits simultaneous screening of large numbers of samples. We have used both indirect (248 cases) and direct (126 cases) monoclonal antibody staining techniques. Results were assessed using median channel shift (CS) from the negative control and relative fluorescence intensity. There is known cross-reactivity between HLA-B27 and HLA-B7 with the monoclonal antibody (MoAb) HLA-ABC-m3. Using this antibody and indirect staining, HLA-B7 samples had a significantly lower CS value than HLA-B27 samples. In 60% of these cases there was a clear distinction between HLA-B27 and HLA-B7. All HLA-B27 and HLA-B7 negative samples had a CS of 0 (P < 0.001). Using direct dual staining with CD3 and HLA-B27, all HLA-B27 and HLA-B7 negative samples had a low CS value (15 maximum), HLA-B7 samples had an intermediate CS value (20-85), and HLA-B27 samples had the highest CS values (70 upward). This system permits large scale rapid negative screening of samples with the elimination of over 60% as negatives (CS < 15). Limitations as a definitive test for HLA-B27 are due to MoAb cross-reactivity with HLA-B7 which necessitates other confirmatory techniques. The substitution of the HLA-ABC-m3 with the recently available HLA-B27 specific MoAb FD705 would substantially increase the value of this technique for routine HLA-B27 typing.

Differences in peptide presentation between B27 subtypes: the importance of the P1 side chain in maintaining high affinity peptide binding to B*2703

Susceptibility to spondyloarthropathies is strongly associated with the MHC class I molecule HLA-B27, and is hypothesized to result from the presentation of arthritogenic peptides. Subtypes of B27 that differ structurally but are disease-associated ought to be capable of presenting such peptides, while nondisease-associated subtypes would not. We demonstrate that B*2703, the predominant West African B27 subtype that may not predispose to disease, is not recognized by most B*2705-alloreactive CTL, and does not efficiently present a known B*2705-restricted influenza A nucleoprotein (NP) peptide. We show inefficient presentation is due to a reduced binding affinity of B*2703 for the NP peptide. Furthermore, substituting Arg for the naturally occurring Ser at P1 of the NP peptide, restores high affinity binding and efficient presentation by B*2703. Our results suggest that B*2703 will bind and present efficiently only a subset of the peptides that bind to B*2705, in particular those with Arg or Lys at P1. The apparent lack of disease in individuals with B*2703 may be due to an inability to bind and present putative arthritogenic peptides.

Psoriatic arthritis. Genetics and HLA antigens

There is convincing evidence of a genetic basis for both psoriasis and psoriatic arthritis (PsA). Part of this genetic predisposition is due to genes within the major histocompatibility complex (MHC). In psoriasis, the primary association is with HLA-Cw6. Further work on specific nucleotide frequencies, especially those in the alpha 1 domain helix of the HLA-C molecule, will be of interest in determining whether a specific nucleotide frequency is present in all patients. The situation in PsA is considerably more complex. It is now established that there is an association between HLA-B27 and PsA, both in its peripheral arthropathy and in spinal disease in which radiological sacroiliitis is present. Spinal disease without radiological sacroiliitis is probably not associated with HLA-B27. There is some suggestion that HLA-B16 or its splits, HLA-B38 and HLA-B39, may also be associated with PsA, but there is considerable heterogeneity between the series, which prevents a firm conclusion being made. It is possible, but again not conclusive, that there is an association between HLA-DR4 and the symmetrical seronegative pattern of peripheral PsA. It is also likely that genes outwith the MHC predispose to psoriasis and PsA. It is further likely that a role will be found for environmental factors in both psoriasis and PsA. There is a tantalizing possibility of a complex interplay between a variety of environmental factors and genetic factors, both within and outwith the MHC, determining not only susceptibility but also the individual clinical pattern of disease. Further clarification of these possibilities is likely to depend primarily on understanding the role of genes within the MHC in predisposing to comparatively more homogeneous diseases, such as psoriasis and ankylosing spondylitis, before the mechanisms operating in PsA can be analysed and better understood.

The pathogenesis of HLA-B27 associated arthritis: lessons from the B27 crystal

The most remarkable association between a major histocompatibility complex antigen and disease susceptibility--HLA-B27 and seronegative spondyloarthropathies, particularly ankylosing spondylitis--was discovered 20 years ago. During the two intervening decades advances in basic immunology and molecular biology have not only revealed the biosynthesis and structure of HLA-B27 but also given clues to the basic function of this molecule, the presentation of allele-specific peptides to CD8+ cytotoxic T cells. The recently reported three-dimensional structure of HLA-B27 and the identification of self-peptides bound to this major histocompatibility complex class I antigen can be viewed as a landmark in the understanding of the pathogenic role of HLA-B27. Based on crystallographic evidence, a peptide-binding motif can be postulated that should allow identification of HLA-B27 complexed peptides which may trigger an immune reaction causing arthritis.

Chemical reactivity of an HLA-B27 thiol group

Techniques have been developed to measure the reactivity of free thiols in the HLA class I antigen-binding cleft. HLA-B27, which sequencing predicts has a free cysteine at position 67, reacts rapidly with the positively charged thiol reagent monobromotrimethyl-ammoniobimane bromide (qBBr) to give products which are identifiable by isoelectric focusing. HLA-B38, B39, B64 and B65, all of which have a similar Cys 67, react less strongly. Several other class I molecules, notably HLA-C antigens, are reactive in this system, and it may be capable of recognizing subtypes such as A*0207 which also carry free cysteine. The accessibility of thiol to qBBr depends both on the chemistry of the class I molecule and other factors in the cell. Two human cell lines which are known to carry identical B27 genes but do not present the same peptides, differ considerably in the accessibility of their B27 thiol. Evidence from mouse cells transfected with mutant B27 genes suggests that a unique lysine at position 70 in the wild-type molecule increases reactivity to thiol-reactive metabolites. The failure of B27 to give a complete reaction with qBBr in our model systems suggests that it can exist in more than one chemical form. This may leave the molecule susceptible to oxidation, causing errors in T cell recognition and an exaggerated inflammatory response.

Different HLA-B27 subtypes present the same immunodominant Epstein-Barr virus peptide

An immunological basis has been postulated for the strong association between at least five subtypes of the HLA-B27 allele (B27.01, .02, .04, .05, and .06) and ankylosing spondylitis, namely that cytotoxic T lymphocyte (CTL) responses are induced against an "arthritogenic" peptide that these different subtypes can all present. This requires a degree of overlap between the peptide binding repertoires of different B27 molecules. The present work, using CTL responses to Epstein-Barr virus (EBV) as a model system in which to identify B27-restricted epitopes, provides the first direct evidence that different disease-related alleles can present the same immunodominant peptide. We first noted that EBV-specific CTL clones, whether from B27.05-, B27.02-, or B27.04-positive donors, were largely subtype-specific in their restriction, recognizing only EBV-transformed B cell lines of the relevant B27 subtype. However, when tested against targets expressing individual EBV proteins from recombinant vaccinia virus vectors, all B27.05-restricted, all B27.02-restricted, and a proportion of B27.04-restricted clones were reactive to the same viral nuclear antigen, Epstein-Barr nuclear antigen (EBNA)3C. In subsequent peptide sensitization assays, all the EBNA3C-specific clones tested and also the EBNA3C-specific component within polyclonal CTL preparations from B27.05-, B27.02-, or B27.04-positive donors recognized the same immunodominant viral peptide RRIYDLIEL (EBNA3C residues 258-266). This sequence accords well with the proposed B27.05 peptide motif and clearly must be accommodated within the different peptide binding grooves of B27.05, B27.02, and B27.04 molecules. Clonal analysis revealed a second component of the B27.04-restricted response that was not shared with other subtypes. This was directed against an EBV latent membrane protein LMP2 epitope whose sequence RRRWRRLTV satisfies some but not all requirements of the B27.05 peptide motif. We conclude that there is indeed a degree of functional overlap between different B27 subtypes in their selection and presentation of CTL epitopes.

HLA-B27 subtypes in the spondarthropathies

The spondarthropathy (Sp)-associated HLA-B27 antigen includes at least seven subtypes, B*2701-07, of which 01, 02, 05 and 07 occur in Caucasians. This study examined the B27 subtype distribution in British patients with Sp. The 133 HLA-B27+ subjects comprised 94 European Caucasian Sp (58 ankylosing spondylitis (AS), 22 reactive arthritis (ReA; 11 sexually acquired (SARA), 11 enteric (EReA)), eight undifferentiated Sp (USp), and six pauciarticular juvenile-onset chronic arthritis (pJCA)) patients, and 34 healthy Caucasian controls, together with four Asian Indian and one Chinese. 35S-labelled B27 was immunoprecipitated with anti-B27 MoAbs, and subtyped according to isoelectric point (pI) following isoelectric focussing. The use of B27 MoAb permitted subtype assignment without full class I HLA typing. The vast majority (95%) were B*2705 (Caucasian controls 31/34; AS 55/58; ReA 21/22; USp 8/8, and pJCA 6/6; Indian control 1/1 and AS 2/3; Chinese pJCA 1/1), and the remainder B*2702. No B*2701 or 07 subjects were identified. AS occurs in both B*2702 and 05 subjects, and we extend this observation to small numbers of ReA and of Indian AS subjects. This implicates molecular features shared between B27 subtypes, rather than subtype-determining regions of the antigen, in Sp pathogenesis.