Long-term drug survival of biological agents in patients with rheumatoid arthritis in clinical practice.

OBJECTIVES: To assess and compare the long-term drug survival (time to drug discontinuation) of biological agents (BA) in patients with rheumatoid arthritis (RA) in clinical practice. Factors associated with discontinuation of BAs were also investigated. METHOD: We conducted an observational longitudinal study of RA patients taking BAs from 1999 to 2013. The primary endpoint was BA discontinuation due to: adverse drug reactions (ADRs), inefficacy, and other causes. Incidence rates of discontinuation (IRs) per 100 patient-years were estimated using survival techniques. Comparisons between BA discontinuation rates and other associated factors were made using Cox regression models. RESULTS: We included 851 courses of BA therapy (1869 patient-years). Adalimumab (33%) was the BA most frequently used, followed by etanercept (24.4%), infliximab, and rituximab. Treatment was suspended in 558 cases [IR 29.8, 95% confidence interval (CI) 27-32]. In the first year of therapy 68% continued on BAs, and after 10 years the retention rate did not exceed 10%. The IR due to inefficacy was 12.1 (95% CI 10.6-13.8) and the IR of ADRs was 13.6 (95% CI 12-15). The unadjusted IR was higher for rituximab than for tumour necrosis factor (TNF) antagonists. In multivariate analysis, infliximab was the BA with the highest risk of discontinuation, compared to adalimumab. Calendar period, taking subsequent courses of BAs, concomitant therapy, and specific comorbidities were also independent factors associated with discontinuation. CONCLUSIONS: After several years of BA treatment in clinical practice, the survival rate was low, mainly as a result of ADRs and inefficacy. We also found differences between the discontinuation rates of BAs and other clinical factors that modify their survival.

Expression of human endogenous retrovirus HERV-K18 is associated with clinical severity in osteoarthritis patients.

OBJECTIVES: The aim of this study was to evaluate the involvement of human endogenous retrovirus K18 (HERV-K18) in osteoarthritis (OA), by genotyping the HERV-K18 env locus in OA patients and controls, and analysing HERV-K18 RNA expression and its association with OA risk and clinical variables. METHOD: We recruited 558 patients with symptomatic OA and 600 controls. We performed the genotyping by TaqMan assays and the analysis of expression by quantitative real-time polymerase chain reaction (qRT-PCR). Scores on the Western Ontario and McMasters Universities Osteoarthritis Index (WOMAC), the Lequesne index, and the Stanford Health Assessment Questionnaire (HAQ) were analysed with regard to the expression levels of HERV-K18. RESULTS: The 18.3 haplotype tended towards an association with OA risk and concordantly this haplotype was associated with a higher HERV-K18 expression (p = 0.05). We found statistically significant differences when we compared the scores on the WOMAC, the Lequesne index for knee and hip, and the HAQ between OA patients with higher expression [normalization ratio (NR) > 10] and OA patients without HERV-K18 expression (p = 0.0003, 0.0005, 0.002, and 0.05, respectively), and also when the comparison was made between OA patients with higher expression (NR > 10) and OA patients with low expression of HERV-K18 (NR = 1) for the WOMAC and the Lequesne index for knee and hip (p = 0.002, 0.013, and 0.006, respectively). CONCLUSIONS: We found an association between health status measurement systems and severity index for OA and the levels of expression of HERV-K18. These results suggest the possible involvement of HERV-K18 in the aetiopathogenesis of the disease.

Association study of ghrelin receptor gene polymorphisms in rheumatoid arthritis.

OBJECTIVES: Ghrelin is a newly characterised growth hormone (GH) releasing peptide widely distributed that may play an important role in the regulation of metabolic balance in inflammatory diseases such as rheumatoid arthritis (RA) by decreasing the pro-inflammatory Th1 responses. In this study we investigated the possible contribution of several polymorphisms in the functional Ghrelin receptor to RA susceptibility. METHODS: A screening of 3 single nucleotide polymorphisms (SNPs) was performed in a total of 950 RA patients and 990 healthy controls of Spanish Caucasian origin. Genotyping of all 3 SNPs was performed by real-time polymerase chain reaction technology, using the TaqMan 5'-allele discrimination assay. RESULTS: We observed no statistically significant deviation between RA patients and controls for the GHSR SNPs analysed. In addition, we performed a haplotype analysis that did not reveal an association with RA susceptibility. The stratification analysis for the presence of shared epitope (SE), rheumatoid factor (RF) or antibodies anti cyclic citrullinated peptide (anti-CCP) did not detect significant association of the GHSR polymorphisms with RA. CONCLUSIONS: These findings suggest that the GHSR gene polymorphisms do not appear to play a major role in RA genetic predisposition in our population.

Chromosomal region 16p13: further evidence of increased predisposition to immune diseases.

OBJECTIVE: Genome-wide studies have identified the chromosomal region 16p13 in the susceptibility to type 1 diabetes (T1D) and multiple sclerosis (MS). This region includes the CLEC16A/KIAA0350 gene and an adjacent gene, MHC2TA (MHC class II transactivator), previously associated with susceptibility to MS and rheumatoid arthritis (RA). The role of CLEC16A polymorphisms in the pathogenesis of T1D, MS and RA and its relationship with the association reported with a MHC2TA haplotype were investigated. METHODS: CLEC16A (rs2903692/rs6498169/rs11074956) polymorphisms were analysed in 435 patients with MS, 316 with T1D and 600 with RA and in 550 ethnically matched controls. The MHC2TA rs3087456G/rs4774C risk haplotype was studied in an independent RA cohort. RESULTS: rs2903692 conferred a protective effect on patients with T1D, MS and RA. The described association of rs6498169 with MS was replicated in MS and RA cohorts. The effect of the MHC2TA rs3087456G/rs4774C haplotype on RA susceptibility was confirmed, and the haplotype was found to be in negative linkage disequilibrium with the CLEC16A rs2903692A/rs6498169A haplotype. CONCLUSIONS: Associations of CLEC16A polymorphisms with T1D and MS were successfully replicated in a Spanish population. A novel association of rs6498169 with a predisposition to RA was described which is consistent with previous MHC2TA results. These data provide evidence for the influence of variants within this chromosomal region on the development of complex diseases.

Investigation of CD69 as a new candidate gene for rheumatoid arthritis.

The aim of this study was to investigate the CD69 gene as a new functional candidate gene for rheumatoid arthritis (RA) genetic predisposition. A case-control association study including 933 RA patients and 800 healthy individuals was conducted. Five haplotype-tagging single nucleotide polymorphisms (SNPs) (rs929615, rs3176806, rs4763299, rs11052877, and rs3176789) covering the CD69 gene coding, 5' and 3' untranslated regions were selected as CD69 genetic markers and genotyped using a Taqman 5' allelic discrimination assay. No statistically significant differences were observed in the single marker association study with regard to either genotypic or allelic frequencies when considering the rs929615, rs3176806, rs4763299, rs11052877, and rs3176789 CD69 SNPs independently. According to these findings, no statistically significant skewing was observed between the RA patients and the controls in the distribution of CD69 haplotypes. In summary, our results do not support a major role for the CD69 gene polymorphisms in RA genetic predisposition in our population.

Potential relationship between herpes viruses and rheumatoid arthritis: analysis with quantitative real time polymerase chain reaction.

OBJECTIVE: To determine whether the human herpes viruses, cytomegalovirus (CMV), Epstein-Barr virus (EBV), and human herpesvirus 6 (HHV-6), are detectable in serum and peripheral blood mononuclear cells (PBMCs) of patients with rheumatoid arthritis (RA). METHODS: 133 PBMC samples (61 RA, 72 healthy donors) and 136 serum samples (59 RA, 77 healthy donors) were analysed by quantitative real time polymerase chain reaction for DNA prevalence and viral load of HHV-6, EBV, and CMV. RESULTS: For PBMC samples significant differences were found for EBV in DNA prevalence (56% in RA v 33% in controls, p = 0.009) and viral load (copies/microg DNA 0-592.3 for RA v 0-40.4 for controls, p = 0.001). For serum samples a significant difference was found for HHV-6 DNA prevalence (10% in RA v 0% in controls, p = 0.006) and viral load (copies/microg DNA 0-529.1 for RA v 0 for controls, p = 0.007). CONCLUSIONS: Herpes viruses may have a role in RA, although alternative explanations are possible: (a) defects in cellular immunity in patients with RA may result in a relatively high viral load; (b) patients with RA may be more prone to infection/reactivation. The usefulness of monitoring the DNA viral load in patients with RA is questioned by these data.

Human parvovirus B19, varicella zoster virus, and human herpes virus 6 in temporal artery biopsy specimens of patients with giant cell arteritis: analysis with quantitative real time polymerase chain reaction.

OBJECTIVE: To evaluate the role of parvovirus B19 (B19), varicella zoster virus (VZV), and human herpes virus 6 (HHV-6) in the aetiopathology of giant cell arteritis (GCA). METHODS: Temporal artery biopsy specimens from 57 patients with GCA and 56 controls were investigated. DNA was obtained by biopsy, and quantitative real time polymerase chain reaction assay performed to establish the prevalence and viral load of B19, VZV, and HHV-6. Amplification of the human beta-globin gene was used as internal positive control. RESULTS: (a) B19 was detected in 31/57 (54%) patients (median viral load 45.2 (25th-75th centiles 0-180.2) copies/microg DNA) v 21/56 (38%) controls (median viral load 0 (0-66.7) copies/microg of DNA; p = 0.07 for DNA prevalence, p = 0.007 for viral load. Among 31 B19 positive samples, 21 (68%) patients with biopsy proven GCA had >10(2) B19 copies/microg of DNA v 5/21 (24%) controls; p = 0.001. (b) No significant difference was found for VZV (p = 0.94 for DNA prevalence; p = 0.76 for viral load) and HHV-6 (p = 0.89 for DNA prevalence; p = 0.64 for viral load) in the GCA group compared with controls. CONCLUSION: B19 may have a role in the aetiopathology of GCA, particularly in those patients with high viral load; no evidence was found for VZV and HHV-6.

Peptide presentation to an alloreactive CTL clone is modulated through multiple mechanisms involving polymorphic and conserved residues in HLA-B27.

This study addressed the mechanisms by which HLA class I polymorphism modulates allorecognition. CTL 27S69 is an alloreactive clone raised against HLA-B*2705, with a known peptide epitope. This CTL cross-reacts with B*2702, which differs from B*2705 in the D77N, T80I, and L81A changes, but not with B*2701, which has D74Y, D77N, and L81A changes. To explain this differential recognition, B*2705 mutants mimicking subtype changes were used. The A81 mutant was not recognized, despite binding the natural epitope in vivo, suggesting that, when bound to this mutant, this peptide adopts an inappropriate conformation. The N77 and I80 mutations restored recognition in the N77A81 or I80A81 mutants. These compensatory effects explain the cross-reaction with B*2702. The Y74 and the Y74N77 mutants were weakly recognized or not recognized by CTL 27S69. This correlated with the absence or marginal presence of the peptide epitope in the Y74N77-bound pool. As with B*2701, exogenous addition of the peptide epitope sensitized Y74 and Y74N77 targets for lysis, indicating that failure to cross-react with B*2701 or these mutants was due to poor binding of the peptide in vivo and not to inappropriate presentation. The abrogating effect of Y74 was critically dependent upon the K70 residue, conserved among subtypes, as demonstrated with mutants at this position. Thus, HLA polymorphism affects allorecognition by modulating peptide binding or the conformation of bound peptides. Compensatory mutations and indirect effects of a polymorphic residue on residues conserved play a critical role.

Modulation at multiple anchor positions of the peptide specificity of HLA-B27 subtypes differentially associated with ankylosing spondylitis.

OBJECTIVE: To investigate the rules governing peptide binding to HLA-B*2705, and to B*2704 and B*2706, which are 2 subtypes differentially associated with ankylosing spondylitis. METHODS: Poly-Ala analogs carrying the HLA-B27 motif Arg-2, and substitutions at anchor positions P1, P3, or Pomega, were used to determine a binding score for each residue at each position. Binding was assessed in a quantitative epitope stabilization assay, where the cell surface expression of HLA-B27 was measured by flow cytometry as a function of peptide concentration. RESULTS: Peptide anchor residues contributed additively to B*2705 binding. About 15% of the natural B*2705 ligands used a deficient P3 or Pomega anchor, but never both, indicating that detrimental anchoring at one of these positions is always compensated by a good anchor at the other one. About 50% of the B*2705 ligands used suboptimal P1 residues. However, this was compensated with optimal P3 and/or Pomega anchoring. Peptides that were longer than decamers used good anchor residues at the 3 positions, suggesting more stringent binding requirements. B*2704 and B*2706 differed in their residue specificity at P1, P3, and Pomega. The rules derived for B*2705 also applied to the known ligands of these 2 subtypes. CONCLUSION: The B*2705, B*2704, and B*2706 peptide repertoires are limited by the allowed residue combinations described in this study. The differential association of B*2704 and B*2706 with spondylarthropathy correlates with differences in their peptide specificity at multiple anchor positions. However, it is now possible to predict the peptide features that determine this differential binding to both subtypes.

Nonapeptide analogues containing (R)-3-hydroxybutanoate and beta-homoalanine oligomers: synthesis and binding affinity to a class I major histocompatibility complex protein.

Crystal structures of antigenic peptides bound to class I MHC proteins suggest that chemical modifications of the central part of the bound peptide should not alter binding affinity to the MHC restriction protein but could perturb the T-cell response to the parent epitope. In our effort in designing nonpeptidic high-affinity ligands for class I MHC proteins, oligomers of (R)-3-hydroxybutanoate and(or) beta-homoalanine have been substituted for the central part of a HLA-B27-restricted T-cell epitope of viral origin. The affinity of six modified peptides to the B2705 allele was determined by an in vitro stabilization assay. Four out of the six designed analogues presented an affinity similar to that of the parent peptide. Two compounds, sharing the same stereochemistry (R,R,S,S) at the four stereogenic centers of the nonpeptidic spacer, bound to B2705 with a 5-6-fold decreased affinity. Although the chiral spacers do not strongly interact with the protein active site, there are configurations which are not accepted by the MHC binding groove, probably because of improper orientation of some lateral substituents in the bound state and different conformational behavior in the free state. However we demonstrate that beta-amino acids can be incorporated in the sequence of viral T-cell epitopes without impairing MHC binding. The presented structure-activity relationships open the door to the rational design of peptide-based vaccines and of nonnatural T-cell receptor antagonists aimed at blocking peptide-specific T-cell responses in MHC-associated autoimmune diseases.

Structure-based design of nonnatural ligands for the HLA-B27 protein.

X-ray studies as well as structure-activity relationships indicate that the central part of class I MHC-binding nonapeptides represents the main interaction site for a T cell receptor. In order to rationally manipulate T cell epitopes, several nonpeptidic spacer have been designed from the X-ray structure of a MHC-peptide complex and substituted for the T cell receptor-binding part of several antigenic peptides. The binding of the modified epitopes to the HLA-B*2705 protein was studied by an in vitro stabilisation assay and the thermal stability of all complexes examined by circular dichroism spectroscopy. Depending on their chemical nature and length, the introduced spacers may be classified into two categories. Monofunctional spacers (11-amino undecanoate, (R)-3-hydroxybutyrate trimer) simply link two anchoring peptide positions (P3 and P9) but loosely contact the MHC binding groove, and thus decrease more or less the affinity of the altered epitopes to HLA-B*2705. Bifunctional spacers ((R)-3-hydroxybutyrate and beta-homoalanine combinations) not only bridges the two distant anchoring amino acids but also strongly interact with the binding cleft and lead to an increase in binding to the MHC protein. The presented modified ligands constitute interesting tools for perturbing the T cell response to the parent antigenic peptide.

Substituting nonpeptidic spacers for the T cell receptor-binding part of class I major histocompatibility complex-binding peptides.

X-ray diffraction studies as well as structure-activity relationships indicate that the central part of class I major histocompatibility complex (MHC)-binding nonapeptides represents the main interaction site for a T cell receptor. In order to rationally manipulate T cell epitopes, three nonpeptidic spacers have been designed from the x-ray structure of a MHC-peptide complex and substituted for the T cell receptor-binding part of several antigenic peptides. The binding of the modified epitopes to the human leukocyte antigen-B*2705 protein was studied by an in vitro stabilization assay, and the thermal stability of all complexes was examined by circular dichroism spectroscopy. Depending on their chemical nature and length, the introduced spacers may be classified into two categories. Monofunctional spacers (11-amino undecanoate, (R)-3-hydroxybutyrate trimer) simply link two anchoring peptide positions (P3 and P9) but loosely contact the MHC binding groove and thus decrease more or less the affinity of the altered epitopes to human leukocyte antigen-B*2705. A bifunctional spacer ((R)-3-hydroxybutyrate tetramer) not only bridges the two distant anchoring amino acids but also strongly interacts with the binding cleft and leads to a 5-fold increase in binding to the MHC protein. To our knowledge, this is the first report of a nonpeptidic modification of T-cell receptor binding residues that significantly enhances the binding of altered peptide ligands to their host MHC protein. The presented modified ligands constitute interesting tools for perturbing the T cell response to the parent antigenic peptide.

Effects of a negative visual hypnotic hallucination on ERPs and reaction times.

The study of changes in ERPs provoked by negative hypnotic hallucinations has thus far yielded contradictory results. Most previous studies have failed to separate specific changes in the processing of the hallucinated stimuli from non-specific changes in arousal due to hypnosis. The present study addresses this issue with the combination of two experimental effects in the task designed to test the hallucination: the noise-compatibility effect and the Simon effect. A choice reaction task was used in which targets appearing at either side of a screen were defined by color and were accompanied by irrelevant noise, and a negative hallucination suggestion was given for noise stimuli. Four highly hypnotizable subjects performed the task on two separate occasions, with and without the suggestion. Increases in reaction times and P300 latencies were found as a function of noise and spatial stimulus-response (SR) incompatibility (Simon effect). Suggestion decreased the response times for all types of trials non-significantly, and it decreased significantly those of left-hand responses. On the other hand, suggestion reduced the increase in P300 latencies in noise-incompatible trials, but did not influence the Simon effect. This result indicates a specific effect of suggestion in the processing of hallucinated stimuli, which is consistent with the hallucinatory experience reported by the subjects.

Relationship between peptide binding and T cell epitope selection: a study with subtypes of HLA-B27.

The effect of HLA-B27 polymorphism on antigen presentation was analysed by comparing the binding of three Epstein-Barr virus-derived peptide epitopes to HLA-B27 subtypes with their immunogenicity and antigenicity in the context of these subtypes. The effect of altering the major anchor residue Arg2 on binding or on recognition by peptide-specific cytotoxic T lymphocytes (CTL) was also examined. The three peptides bound significantly to all the B*2701-B*2706 subtypes. This did not correlate with the peptides being immunogenic or recognized by specific CTL in the context of only particular subtypes. In addition, of the three viral epitopes tested, those that were immunogenic in B*2702- or B*2705-restricted responses bound to these subtypes less efficiently than one peptide that was immunogenic only in the B*2704 context. Thus, among several potentially immunogenic peptides from the same virus, the antiviral response is not necessarily directed against the one that binds best to the restricting subtype. These results indicate that HLA-B27 polymorphism influences antigen presentation in ways other than simply peptide affinity. Synthetic analogues lacking the canonical Arg2 motif of HLA-B27-bound peptides, even when binding much worse to the restricting subtype, were recognized equally by CTL specific for the parental peptide. This indicates that Arg2 is not required to maintain the structure of the epitope. The implications of these results for pathogenetic models of HLA-B27-associated disease are discussed.

An HLA-B27 polymorphism (B*2710) that is critical for T-cell recognition has limited effects on peptide specificity.

The B*2710 subtype differs from the HLA-B27 prototype (B*2705) only by having Glu instead of Val at position 152, in the alpha2 helix of the peptide-binding site. In spite of its structural similarity most alloreactive CTL raised against B*2705 fail to cross-react with B*2710. Indeed, of the residues that are polymorphic among HLA-B27 subtypes, the Val>Glu152 change has the greatest influence on HLA-B27 T-cell antigenicity. The molecular basis for this antigenic disparity was analyzed in this study. Sequence analysis indicated that B*2710-bound peptides have very similar motifs to B*2705-bound ones both at the main and auxiliary anchor positions. In addition, most of the individual ligands sequenced from B*2710 were previously found in B*2705. Together these results indicate that both subtypes have largely overlapping peptide repertoires. Molecular dynamics simulations of a common ligand in complex with either B*2710 or B*2705 failed to detect significant conformational changes in the peptidic main chain or in solvent accessibility of the side chains. In addition, modeling of the Val>Glu152 change into the MHC-peptide-TCR structure suggested a direct role of residue 152 in interaction with the TCR. Thus, the large differences in T-cell recognition between B*2710 and B*2705 are not explained by an effect of the Glu152 change on peptide specificity or conformation, but by different direct interactions with the TCR.

Fine specificity of antigen binding to two class I major histocompatibility proteins (B*2705 and B*2703) differing in a single amino acid residue.

Starting from the X-ray structure of a class I major histocompatibility complex (MHC)-encoded protein (HLA-B*2705), a naturally presented self-nonapeptide and two synthetic analogues were simulated in the binding groove of two human leukocyte antigen (HLA) alleles (B*2703 and B*2705) differing in a single amino acid residue. After 200 ps molecular dynamics simulations of the solvated HLA-peptide pairs, some molecular properties of the complexes (distances between ligand and protein center of masses, atomic fluctuations, buried versus accessible surface areas, hydrogen-bond frequencies) allow a clear discrimination of potent from weak MHC binders. The binding specificity of the three nonapeptides for the two HLA alleles could be explained by the disruption of one hydrogen-bonding network in the binding pocket of the HLA-B*2705 protein where the single mutation occurs. Rearrangements of interactions in the B pocket, which binds the side chain of peptide residue 2, and a weakening of interactions involving the C-terminal end of the peptide also took place. In addition, extension of the peptide backbone using a beta-Ala analogue did not abolish binding to any of the two HLA-B27 subtypes, but increased the selectivity for B*2703, as expected from the larger peptide binding groove in this subtype. A better understanding of the atomic details involved in peptide selection by closely related HLA alleles is of crucial importance for unraveling the molecular features linking particular HLA alleles to autoimmune diseases, and for the identification of antigenic peptides triggering such pathologies.

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.

Binding of peptides naturally presented by HLA-B27 to the differentially disease-associated B*2704 and B*2706 subtypes, and to mutants mimicking their polymorphism.

B*2704 and B*2706 are closely related HLA-B27 subtypes of which the former but not the latter is associated to ankylosing spondylitis. Their peptide specificity relative to other disease-associated subtypes was analyzed by testing binding of self-peptides naturally presented by B*2705 or B*2702, and synthetic analogs, to B*2704, B*2706, and site-specific mutants mimicking their changes. Peptides with basic, aliphatic or aromatic C-terminal residues bound to B*2705 with similar affinity. In B*2704 C-terminal aliphatic/ aromatic residues were preferred. B*2706 discriminated drastically between polar and nonpolar C-terminal residues, showing strong preference for Leu and Phe, and less than B*2704 for basic and Tyr residues. Loss of single acidic charges (D > S77, D > Y116) increased preference for C-terminal Leu and Phe, but allowed efficient binding of peptides with basic residues or Tyr. Their gain (V > E152, H > D114) maintained wide C-terminal specificity, but severely impaired binding, presumably by disrupting interactions with internal peptide residues. This was compensated by Y116 in the double D114Y116 mutant. The specificity of B*2704 and B*2706 was explained only partially by the separate effects of single mutations, indicating that novel properties arise from concomitant changes at various positions. For instance, specificity of B*2706 for nonpolar C-terminal residues required simultaneous removal of Asp77 and Asp116. B*2706 differed from B*2705, B*2702, and B*2704 in its lower suitability for C-terminal Tyr, suggesting that this feature might be relevant for HLA-B27 association to spondyloarthropathy.