HLA-B27: a registry of constitutive peptide ligands

The very strong association of human leukocyte antigen (HLA)-B27 with spondyloarthritis might be related to its peptide-presenting properties. The natural polymorphism of this molecule influences both peptide specificity and disease susceptibility. In this study, we present a comprehensive compilation of known natural ligands of HLA-B27 arising from endogenous proteins of human cells, together with a statistical assessment of residue usage among constitutive peptide repertoires of multiple HLA-B27 subtypes. This analysis provides evidence that every peptide position, including "non-anchor" ones, may be subjected to selection on the basis of its contribution to HLA-B27 binding and also allows a quantization of residue preferences at known anchor positions. The present registry is intended as a basis on which to build up reliable criteria to assess the effect of HLA-B27 polymorphism on peptide presentation, for T-cell epitope predictions, and for molecular mimicry studies.     

The pathogenetic role of HLA-B27 and its subtypes

The strong association between HLA-B27 and ankylosing spondylitis has been known for more than 33 years, but the enigma of the pathogenetic role of the gene and its product has not yet been solved. Ongoing studies have produced evidence supporting different theories to explain this association, and structural and functional studies of HLA-B27 allele products at molecular level have provided information of broad and multidisciplinary value and disclosed new avenues leading to autoimmunity and immune disregulation.     

HLA-B27亚型及其与强直性脊柱炎关系的研究进展

强直性脊柱炎(AS)是与mA关联最强的疾病.HLA-B27由22个以上同种异型基因型(亚型B*2701～B*22)组成,不同亚型核苷酸序列之间只存在个别位点的差异,其亚型具有分布不同的种族和人种流行情况,以B*2705分布最广.近年来建立了大量的AS动物模型,人类B27转基因鼠实验证实B27分子是AS的原发关联成分,这种带有B27等位基因的实验动物可发生类似人类AS疾病.目前倾向于以关节源性肽假说来解释HLA-B27在AS发病中的作用.     

Cross-reactivity between an HLA-B27-derived peptide and a retinal autoantigen peptide: a clue to major histocompatibility complex association with autoimmune disease

Statistical correlations between the expression of various HLA antigens and certain autoimmune diseases have been observed for both HLA class I and II antigens. Autoimmune diseases like spondyloarthropathies and anterior uveitis are associated with HLA-B27, but uveitis in Behçet's disease with HLA-B51. We describe a peptide from disease-associated HLA class I antigens sharing sequence homologies with a highly uveitogenic epitope from the retinal autoantigen S-antigen. S-antigen induces autoimmune uveitis in the animal model and is a major autoantigen in human disease. The HLA peptide induced uveitis in the Lewis rat and, moreover, suppressed S-antigen-induced disease when administered orally. Patients' PBL cross-reacted with the HLA- and corresponding retinal peptide, explaining the organ specificity of the disease.     

HLA-B27 and HLA-A2 subtypes: structure, evolution and function

Beyond the resolution of tissue typing serology, HLA class I antigens display a certain level of structural microheterogeneity, that allows their subdivision into subtypes. The structure of these subtypes shows that multiple mechanisms operate in the generation of HLA polymorphism and suggests possible evolutionary pathways for subtype diversification. In addition, subtype polymorphism critically affects cellular allorecognition and antigen presentation to self-restricted T cells. These properties are used to define the structure and diversity of T-cell epitopes. In this review, José López de Castro discusses the nature and evolution of this polymorphism and its modulation of antigen recognition by cytolytic T lymphocytes.     

HLA binding and T cell recognition of a tissue transglutaminase-modified gliadin epitope.

DQ2 confers susceptibility to celiac disease (CD) and intestinal CD4(+) T cells of DQ2(+) CD patients preferentially recognize deamidated gliadin peptides. This modification can be mediated by tissue transglutaminase (tTG). We have investigated what role the tTG-modified residues play in DQ2 binding and T cell presentation using a model gamma-gliadin peptide (residues 134 - 153). Treatment of this peptide with tTG resulted in deamidation of Gln residues at positions 140, 148 and 150. Two of these residues act as DQ2 anchors at position P7 (148) and P9 (150) and increased the affinity of the modified peptide for DQ2 50-fold. Testing of a mutant DQ2 molecule demonstrated that the Lys residue at beta71 of DQ2 is important for binding of the deamidated peptide. A variant DQ2 molecule (with the same beta-chain but different alpha-chain) that does not confer susceptibility to CD was capable of presenting the gliadin peptide, but not pepsin/trypsin-digested gliadin, equally well to a T cell. This suggests that processing events might be involved in the preferential presentation of the gliadin peptide by the DQ2 molecule. Substitution of Gln with Glu in some positions not targeted by tTG, but in positions likely to be deamidated via non-enzymatic mechanisms, disrupted T cell recognition. This provides additional evidence that tTG is responsible for modification of gliadin in vivo.     

A common epitope between HLA-B27, -B13 and -B37 alloantigens defined by a monoclonal antibody

An anti-HLA-B27 monoclonal antibody produced by the hybridoma technique is described. This BD.7 reagent is a cytotoxic IgM antibody. Its reactivity was studied by lymphocytotoxicity tests, indirect immunofluorescence tests and biochemical analysis against an extensive panel of peripheral blood mononuclear cells. All HLA-B27 positive samples, either from normal subjects or from patients with Ankylosing Spondylitis, were recognized by this reagent. Moreover, a cross-reaction was observed with HLA-B13 cells, and a new unexpected reaction with all HLA-B37 cell suspensions. The interest of such a reagent is discussed.     

Frequency of HLA-B27 subtypes in a Danish population and in Danish patients with ankylosing spondylitis

Polymerase chain reaction in combination with sequence-specific oligonucleotide probes were used to analyze nine HLA-B27 subtypes among 51 healthy I HLA-B27 positive Danish blood donors and 30 Danish HLA-B27 positive patients with ankylosing spondylitis (AS). In the group of healthy Danes we found two subtypes, B*2705 (90.2%) and B*2702 (9.8%), however, among the AS patients only the B*2705 subtype was detected. We did not find a significant evidence for associations between AS and a particular HLA-B27 subtype in a Danish population.     

Different linkage disequilibria of HLA-B27 subtypes and HLA-C locus alleles

Subtypes of HLA-B27 have been identified by cellular, serological and biochemical techniques. Comparison of the various B27 subtype designations showed the existence of seven B27 subtypes. The new WHO nomenclature (1987) of the B27 subtypes is included. We further report on different linkage disequilibria (ld) of the B27 subtypes. In Caucasoids, the prevalent subtype B27.5 is in ld with Cw1 and Cw2, whereas B27.2 is linked only with Cw2. In Orientals, the most frequent subtypes B27.4 and B27.6 usually occur with Cw3 or Cw blank; B27.5 mostly occurs with Cw2, and B27.2 is almost absent.     

TCRB Junctional regions from HLA-B27-restricted T cells and HLA-B27 binding peptides display conserved hydropathy profiles in the absence of primary sequence homology

Analysis of formal amino acid sequence identity between differentTCRB chain (TCRB) hypervariable regions (CDR3) is commonly usedto localize relevant sites of TCR antigen interaction or toyield indirect information on unknown corresponding antigens.However, this analysis sometimes fails to demonstrate expectedconcordances, e.g. between CDR3 from T cell clones of identicalreactivity. Since this may be due to ignorance of physico-chemicaiparameters, we have now used hydropathy profile analysis asan additional method to examine TCRB-CDR3 and putative peptideantigens. Superimposed hydropathy plots (SHOP) of 20 TCRB-CDR3from HLA-B27-restricted autoreactive and Yersinia enterocoliticaspecificsynovial cytotoxic T lymphocytes (CTL) isolated from patientswith reactive arthritis (ReA) revealed restricted distributionof polar amino acids resulting in characteristically differentSHOP profiles between the two CTL groups. Similarly, Yersinia-derivedand self nonapeptides known to bind HLA-B27 differed in SHOPprofiles. To validate the method we have extended SHOP analysisto published TCRB sequence data from additional HLA-B27- andHLA-A2-restricted CTL. Limited variability of hydropathy wasobserved in TCRB-CDR3 from peptide-specific CTL but not in TCRBfrom HLAB27- alloreactive CTL or non-HLA-B27-restricted controlCTL. We here demonstrate that SHOP may improve TCR-CDR3 sequenceanalysis by detection of structural constraints which remaincryptic by conventional sequence analysis. Our data suggestthat electrostatic properties rather than rigid sequence motifsdetermine T cell specificities.     

Modulation of peptide binding by HLA-B27 polymorphism in pockets A and B,and peptide specificity of B*2703

The results in this study address three aspects of peptide binding to the disease-associated antigen HLA-B27 and its modulation by polymorphism: the contribution of major anchor residues 2 and 9, the role of pocket B polymorphism in modulating peptide specificity, and the binding properties of B*2703, a subtype not found to be associated with spondyloarthropathy. Synthetic analogs of peptides naturally presented by B*2705 were used to demonstrate that residue 2 is essential, since Ala2 analogs bound marginally to B*2705, but the specificity of B*2705 for Arg2 is not absolute, and show that the contribution of basic residue 9 to binding was significant, but less than Arg2. The effect of single mutations in the B pocket was to decrease or - with the Glu > Met-45 mutation - totally shift pocket B specificity for Arg2 towards other residues at this position. This was shown by quantitating the relative binding of Gln2 and Ala2 analogs, and by pool-sequencing of the peptides bound in vivo to these mutants. Peptides naturally presented by B*2705 apparently bound with a lower affinity to pocket A variants with altered hydrogen bonding to the peptide N terminus, including B*2703. Binding of peptide analogs with changes at positions 2 or 9 suggested that in B*2703 pocket A, interactions are weaker and pocket B interactions are stronger than in B*2705. This can be explained by the effect of the unique His59 change in B*2703 in both pockets. Thus, B*2703 is probably the HLA-B27 subtype with the most stringent specificity for the Arg2 peptide motif.     

A ribosomal protein of Yersinia pseudotuberculosis having partial epitope identity with HLA-B27.

The phenotype HLA-B27 is common in patients who develop reactive arthritis after having an infection. One hypothesis concerning the pathogenesis of reactive arthritis is that molecular mimicry between HLA-B27 and certain bacterial components might be involved. It is known that an infection with Yersinia is commonly associated with reactive arthritis in B27 positive patients. Therefore, we were interested to investigate whether cross-reactivity between Yersinia and HLA-B27 exists. A gene library of Yersinia pseudotuberculosis was created in the plasmid vector pUC13. One of the resulting clones contained a gene encoding an intracytoplasmic protein that seems to have partial epitope identity with HLA-B27. It reacted in western blot. ELISA and immunoprecipitation with three different HLA-B27 specific monoclonal and polyclonal antibodies of the IgG and IgM class. However DNA-sequencing of the cloned Yersinia gene and the predicted amino acid sequence revealed only a very remote similarity with HLA-B27 in the primary structure. Instead, an extremely high degree of similarity with the ribosomal protein L4 of the S10 operon of Escherichia coli was identified indicating that the protein encoded by the cloned Y. pseudotuberculosis gene is a corresponding ribosomal protein.     

In vivo persistence of a HIV-1-encoded HLA-B27-restricted cytotoxic T lymphocyte epitope despite specific in vitro reactivity.

A large number of human immunodeficiency virus type 1 (HIV-1) specific HLA-restricted cytotoxic T cell (CTL) epitopes have been mapped, including an HLA-B27-restricted immunodominant epitope within p25gag. Accordingly, this segment of the HIV-1 provirus was amplified by the polymerase chain reaction from DNA derived from fresh uncultured peripheral blood mononuclear cells (PBMC) of four HLA-B27 HIV-1-infected individuals. In all cases the majority of infected PBMC bore sequences encoding the HLA-B27-restricted peptide. CTL escape mutants had not accumulated in vivo 8 and 14 months later despite demonstrable CTL activity in vitro. These data emphasize the importance of silently infected lymphocytes in evading immune surveillance.     

Association of ankylosing spondylitis with HLA-B27 and ERAP1: Pathogenic role of antigenic peptide

Ankylosing spondylitis (AS) is a form of seronegative inflammatory arthritis whose strong genetic association with the human leucocyte antigen (HLA)-B27 has been known for almost 4 decades. However, its mechanism remains poorly understood. Recently, with the development of genetics, further more genes have been robustly associated with the disease. Genome-wide association studies identified the association between AS and ERAP1 (endoplasmic reticulum associated aminopeptidase 1). And ERAP1 has shown the potential in trimming antigenic peptides to optimal length for binding to HLA-B27 in the ER (endoplasmic reticulum). However, the length of the peptides are strictly restricted in the process of peptide transporting, processing and presentation. A hypothesis is proposed that the abnormal mechanism of AS may related to the trimming of N-terminal sequences from antigenic precursors in the ER and the length of the antigenic peptides that are presented to the T-cell receptors.     

Functional degeneracy of residues in a T cell peptide epitope contributes to its recognition by different T cell hybridomas.

Synthetic antigen Poly EYK(EYA)5 induces T cells of narrowly defined fine specificity as represented by the two I-Ad-restricted T cell hybridomas, A.1.1 and B.1.1. Both these hybridomas recognize the minimum 15-amino-acid peptide sequence EYK(EYA)4. We have characterized the residues involved in the recognition of EYK(EYA)4 peptide by these hybridomas with synthetic peptides and discovered a distinct functional hierarchy for the residues in the sequence. Even with the repeating tripeptide (EYA)5, which is recognized by B.1.1 cells, the residues that are essential cluster near the middle of the sequence but not near the N- or C-terminal region. Different MHC binding and TCR contacting residues were found for each of the hybridomas. The results suggest that different T cells either recognize different parts of the peptide MHC complex or that the peptide binds to MHC in multiple conformations. This was supported by the fact that Poly EYK(EYA)5 is alpha-helical but the peptides used here showed only a slight propensity to adopt this structure and it did not correlate with their functional activity. We also found that (EYA)5 does not compete with EYK(EYA)4 in the stimulation of A.1.1 cells despite its obvious capacity to interact with I-Ad when it stimulates B.1.1 cells. This may be because these peptides have a low affinity for Ia and therefore only appropriate TCR interactions would stabilize the antigen-Ia complex. In conclusion, antigen-MHC-TCR interaction appears to be a dynamic process which allows recognition of different residues of a T cell determinant by different T cells.     

T cell epitope selection: dominance may be determined by both affinity for major histocompatibility complex and stoichiometry of epitope.

The majority of T cell hybridomas produced in the BALB/c mouse in response to immunization with lambda repressor cI recognize a peptide fragment comprising of residues 12 to 26 (P12-26). Some other parts of the cI (P1-14, P33-48 and P73-88) are defective in generating T cell responses in the BALB/c mouse. P73-88 may be converted into a T cell determinant if a few more amino acid residues are included (P67-88). Together with P46-67 and P80-102, most peptides derived from cI were capable of eliciting T cell responses by themselves in BALB/c mouse. The mechanisms underlying the selection of P12-26 over the other epitopes when lambda repressor was used as immunogen were examined. The dominant response to P12-26 was attenuated by tolerizing with intravenous administration of P12-26. Under such treatment the T cell response to P12-26 was reduced by 80% but there was no enhancement on the responses toward other epitopes. The selection of P12-26 is, thus, unlikely to be due to a competition at the T cell level. It was also found that the dominance of P12-26 was not simply due to a higher affinity of P12-26 for major histocompatibility complex molecules. For example P12-26 binds better to I-Ad molecule than P80-102, but co-injection with equimole of P12-26 only slightly inhibited P80-102-induced T cell response. Instead, it required a few molar excess of P12-26 to effectively block the association of P80-102 with I-Ad molecules and to inhibit the T cell immunity to P80-102. Since epitopes such as P46-67, P67-88 and P80-102 were generated from lambda repressor cI at a lower molar basis than that of P12-26, it is suggested that the dominance of P12-26 was probably generated by such stoichiometry difference, in addition to the higher affinity of P12-26 for I-Ad molecules.     

Cytokine profile and T cell adhesiveness to endothelial selectins: in vivo induction by a myasthenogenic T cell epitope and immunomodulation by a dual altered peptide ligand

Myasthenia gravis (MG) is a T cell-regulated antibody-mediated autoimmune disease. Immunization with two myasthenogenic peptides, p195-212 and p259-271, that are sequences of the human acetylcholine receptor alpha subunit was shown to induce experimental autoimmune MG (EAMG)-associated immune responses. A peptide composed of the two altered peptide ligands (APL) of the myasthenogenic peptides (designated as dual APL) inhibited, in vitro and in vivo, those responses. The objectives of this study were to examine (i) whether in vivo T cell activation by p259-271 affects the cytokine profile and the T cell migration ability, and (ii) whether the latter are immunomodulated by in vivo administration of the dual APL. Our results showed that immunization of mice with p259-271 enriched the population of lymph node and spleen cells with subsets of T cells with strong adhesiveness towards E- and P-selectins. This enrichment was associated with an acquisition of a T(h)1-type cytokine profile. Treatment of the immunized mice with the dual APL interfered with both the migratory potential of the autoreactive T cells, and the production of the T(h)1-type cytokines IL-2 and IFN-gamma (known to play a pathogenic role in MG and EAMG). T cells derived from APL-treated mice acquired a T(h)3-type cytokine profile, characterized by the secretion of the immunosuppresive cytokine transforming growth factor-ss. Thus, our results suggest that T cell selectin ligands and T cell-derived cytokines are involved in the induction and immunomodulation of EAMG- and MG-associated T cell responses.     

Cytotoxic T cell reactivity and HLA-B35 binding of the variant Plasmodium falciparum circumsporozoite protein CD8+ CTL epitope in naturally exposed Kenyan adults

In this study, we have investigated the extent of natural polymorphism in the CD8⁺ cytotoxic T lymphocyte (CTL) determinant (amino acids 368–390) of circumsporozoite (CS) protein of Plasmodium falciparum field isolates from a holoendemic region of Kenya, and determined how this variation affects the CTL reactivities in clinically immune adults and binding specificities to human histocompatibility leukocyte antigen (HLA)-B35. Among the eight variant sequences that were found in this region, four were new and not seen in parasites from other geographical regions. When synthetic peptides corresponding to the eight variants were used to test the presence of CTL response in different donors, a different spectrum of CTL reactivity to these variants was noticed. While CTL from some donors recognized the P1 sequence (the most prevalent type of sequence) but not P8 (another major variant), other donors showed a reverse pattern of reactivity. Although none of the donors was able to recognize all the variants, CTL responses to all the eight variant sequences were found in this population. An octamer peptide with P1 sequence KPKDELDY in this polymorphic determinant was known to bind HLA-B35. When we tested the effect of natural variation in this octamer sequence on HLA-B35 binding, it became evident that SP13 with D → N substitution retained its binding specificity to HLA-B35. On the other hand, the SP12 octamer sequence which had two substitutions did not bind HLA-B35. The most interesting finding was the observation that a D → G substitution at position 374 rescued the binding ability of SP14, which otherwise could not bind to this HLA molecule due to E → Q amino acid substitution at position 372. To our knowledge, this is the first demonstration showing that a natural polymorphism can rescue the binding specificity to an HLA-class I molecule that was lost due to another natural amino acid substitution. Altogether, these results demonstrate that natural polymorphism in the CS protein affects both the CTL reactivity and the ability to bind to HLA-B35.     

A retro-inverso analog mimicks the cognate peptide epitope of a CD4+ T cell clone

Synthetic analogs of peptide epitopes may activate specific T helper cells, antagonize their antigen receptors, or block recognition by competing for major histocompatibility complex (MHC) class II binding sites. Rationally designed peptides may therefore prove useful as vaccines and for treatment of autoimmune diseases and allergies mediated by CD4⁺ T cells. However, their susceptibility to proteolytic degradation limits the applicability of conventional peptides in vivo. By contrast, retro-inverso analogs, in which a native sequence is substituted with D -amino acids linked with a reversed backbone, resist proteolysis and still maintain the side chain topology of the corresponding natural peptide. We report here that an end group-modified retro-inverso analog of the IgG2a^b heavy chain allopeptide determinant γ2a^b 435–447 was recognized by an I-A^d-restricted, γ2a^b 435–447-reactive T cell clone. The pseudopeptide elicited near-maximal interleukin-2 responses, although 300-fold higher concentrations were needed than the native determinant. The weaker antigenicity of the retro-inverso analog could be fully accounted for by an impaired I-A^d binding capacity, which might reflect reduced ability of the distorted main chain to form hydrogen bonds with I-A^d. Glycine substitution at the residue corresponding to the first primary anchor (P1) of the native peptide abrogated I-A^d binding and antigenicity of the retro-inverso analog. Thus, the pseudopeptide resembled the native determinant with respect to orientation in the class II binding site, configuration of the epitopic side chains, and the constraints that governed the interactions between a major anchoring side chain and I-A^d. In conclusion, proteolytically resistant compounds with predefined capacity to interact with MHC class II allelic products and T cell antigen receptors may be designed by retro-inverso modification of native determinants.