DQ Molecules Research Articles

Peripheral T Cell Response to A-Gliadin in Celiac Disease: Differential Processing and Presentation Capacities of Epstein-Barr-Transformed B Cells and Fibroblasts

Celiac disease (CD) is a small intestinal disorder characterized by the malabsorption of most nutrients. Disease pathogenesis appears to be associated with immunemediated pathology. Susceptibility is associated with genes coding for DQw2 class II molecules. In the present report we investigated T cell responses to A-gliadin (AGL), a major α-gliadin component known to activate disease. Gliadin-specific lines were generated from a CD patient and a normal donor. Three major points were revealed by the analysis of these T cells: (1) On the basis of mapping experiments using Epstein-Barr virus (EBV) lines and DR-transfected fibroblasts and DR-, DP-, and DQ-specific monoclonal antibodies (mAb), all responses appeared to be DR-restricted. Thus, in contrast to the strong association of disease susceptibility with DQ molecules, no DQ-restricted, gliadin-specific response was detectable. (2) Fine specificity analysis, using a panel of synthetic peptides spanning the entire α-gliadin component molecule, revealed that the clones derived from the normal donor were DR53-restricted and AGL 21-40-specific, while clones derived from the CD patient were DR7-restricted and peptide 1-20-specific. (3) Both whole AGL and AGL 1-20 were presented to the patient derived clones with much higher efficiency by DDR-transfected fibroblasts than by EBV lines. These data suggested that fibroblasts processed this determinant efficiently, while EBV lines were unable to do so. Indeed, analysis of a panel of truncated AGL 1-20 analogs revealed that peptide AGL 1-8, which contained the minimal T cell epitope, was presented with equal efficiency by fixed or irradiated EBV and irradiated DR7-transfected fibroblasts.

Role of human major histocompatibility complex DQ molecules in superantigenicity of streptococcus-derived protein

Antigenicity of peptic extract from type 12 group A streptococci (PEAST12) for T cells was examined in major histocompatibility complex (MHC) class II transgenic mice. PEAST12 was mitogenic for murine T cells when antigen-presenting cells were obtained from human MHC (HLA)-DQ4 alpha beta transgenic mice or from DQ6 alpha beta transgenic mice but was not mitogenic in DR alpha transgenic, DR51 alpha beta transgenic, E alpha transgenic, or nontransgenic mice. In addition, PEAST12 showed mitogenicity for murine T cells in DQ4 alpha singly transgenic mice but not in DQ4 beta singly transgenic mice. T-cell stimulation by PEAST12 was unrestricted by but dependent on the expression of HLA-DQ molecules on antigen-presenting cells, and PEAST12 selectively activated T-cell receptor V beta 11-, V beta 15-, and V beta 18-positive T cells in mice. We propose that PEAST12 contains a superantigen which binds preferentially to the alpha-chain of HLA-DQ molecules. The well-known phenomenon that peptic extracts from group A streptococci are mitogenic in humans but not in mice is likely due to structural differences in MHC class II molecules between these two species of mammals.

Pool sequencing of natural HLA-DR, DQ, and DP ligands reveals detailed peptide motifs, constraints of processing, and general rules.

We have approached the problem of MHC class II ligand motifs by pool sequencing natural peptides eluted from HLA-DR, DQ, and DP molecules. The results indicate surprisingly clear patterns, although not quite as clear as with natural class I ligands. The most striking feature is a highly dominant Proline at position 2. We interpret this to be a consequence of aminopeptidase N-like activity in processing. Another general aspect is the existence of three to four hydrophobic or aromatic anchors, whereby the first and the last are separated by five to eight residues. The peptide motifs for HLA-DR1, DR5, DQ7, and DPw4 are allele-specific and differ by spacing and occupancy of anchors. The anchors tend to be flanked by clusters of charged residues, and small residues, especially Ala, are frequent in the motif centers. These detailed motifs allow one to interpret most previous (DR-) motifs as fitting one or more of the anchors or conserved clusters. The relative motif symmetry suggests the possibility of bidirectional binding of peptides in the class II groove.

Binding of peptides to HLA-DQ molecules: peptide binding properties of the disease-associated HLA-DQ(alpha 1*0501, beta 1*0201) molecule.

Peptide binding to DQ molecules has not previously been described. Here we report a biochemical peptide-binding assay specific for the DQ2 [i.e. DQ(alpha 1*0501, beta 1*0201)] molecule. This molecule was chosen since it shows a strong association to diseases such as celiac disease and insulin-dependent diabetes mellitus. Initially we radiolabelled some selected peptides and tested them for binding to affinity-purified DQ2 molecules. One of the peptides, a Mycobacterium bovis (MB) 65 kDa 243-255Y peptide, displayed a good signal-to-noise ratio and was thus chosen as an indicator peptide in the DQ2 binding assay. The MB 65 kDa 243-255Y peptide bound to DQ2 in a strictly pH-dependent fashion, with optimal binding around pH 5 and only weak binding at pH 7.4. The association of the MB 65 kDa 243-255Y peptide to DQ2 was slow, but once formed, the peptide-HLA complexes were very stable. The binding of peptides to DQ2 was specific, as shown in inhibition experiments with a panel of 47 peptides, differing in length, sequence, and origin. The binding of peptides to DR3 was tested in a similar assay with a Mycobacterium tuberculosis 65 kDa 3-13 peptide as the binding indicator. DQ2 and DR3 molecules bound to different sets of peptides. However, the peptide binding to DQ2 and DR3 showed, in general, similar characteristics with respect to pH dependence and kinetic parameters, indicating that the overall rules for peptide binding to DQ molecules are the same as those previously shown for human DR and murine I-A and I-E molecules.

気管支喘息の発症に関する研究

Heredity has long been suspected to play an important role in the occurrence of bronchial asthma. Recently familial and twin sibling studies demonstrated the role of genetic factors in asthma. Therefore, we measured HLA antigens of 53 asthmatics and analyzed the relationship to clinical parameters to elucidate the role of HLA in the occurrence of bronchial asthma.Four major findings were obtained. 1) Study of the HLA system in asthmatics showed a significantly higher frequency of DR53 and DQ3 and lower frequency of DQ1, DQ6 (1) and DQ7 (3). This suggests that DR molecules demonstrate positive associations and DQ molecules negative associations with bronchial asthma. 2) DQ1 was associated with IgE production in asthmatics. 3) B39 (16) positively associated with RAST score of Candida. 4) The study of late onset asthmatics, intractable asthmatics and the high responder group of in vitro lymphocyte reaction to Candida allergen showed a cooperatively and significantly higher frequency of DR53 (P<0.05).These findings suggest that DR53 has a positive association with onset of late onset intractable asthma.

Uncoordinate induction and differential regulation of HLA class-I and class-II expression by gamma-interferon in differentiating human neuroblastoma cells.

Recombinant gamma-interferon (IFN-gamma) has recently been shown to be one of the most effective inducers of neuroblastoma (NB) cell differentiation. Since increasing evidence indicates that expression of MHC class-I and class-II antigens by tumour cells is important for immunorecognition and cell targeting, we tested whether induction of NB cell differentiation by IFN-gamma is followed by expression of HLA class-I and class-II molecules. LAN-5 human NB cell line completely lacks HLA class-I antigens. Their expression was induced in a dose-dependent manner by IFN-gamma. HLA class-II molecules are also absent on LAN-5 cells, but only DP antigens were dose-dependently induced by IFN-gamma, while DR and DQ molecules were unaffected by the treatment. To confirm and extend the immunological data to all the class-II molecules, we performed Northern blot analysis, observing that DP alpha mRNA was induced in a dose- and time-dependent manner. DO beta and DZ alpha genes were also induced peaking after 3 days of IFN-gamma treatment. DR beta and DQ beta genes, which were not induced by IFN-gamma, gave a normal pattern of enzyme restriction by Southern blot. To get an insight into the regulation of HLA class-II gene expression in the neuronal model, we measured the decline of the steady-state HLA class-II mRNA. DO beta mRNA rapidly returned to baseline level after removing IFN-gamma, while the decay rates of DP alpha and DZ alpha mRNA were very slow. This might indicate different regulation at the post-transcriptional level for DO beta mRNA with respect to DP alpha and DZ alpha mRNA. To strengthen these findings we evaluated the half-lives of the mRNA after IFN-gamma induction by means of actinomycin D treatment. HLA-DO beta mRNA had a shorter half-life, while DZ alpha and DP alpha had a longer decay rate. Finally, we report that treatment of LAN-5 cells with cycloheximide did not alter the rate of transcription of the HLA-DP alpha gene, suggesting that no protein factor(s) is/are needed to maintain DP alpha gene expression.

Patterns of major histocompatibility complex class II expression on T cell subsets in different immunological compartments. 1. Expression on resting T cells.

In this study we have investigated the expression of major histocompatibility complex (MHC) class II molecules on T cells from various lymphoid compartments in the sheep. Monoclonal antibodies which react specifically with sheep MHC class II molecules homologous to the human DQ and DR molecules have been characterized. These antibodies have been used, together with the monoclonal antibodies specific for sheep CD4-, CD8- and T19-positive T cells, to quantitate DQ and DR expression on T cell subsets in adult and fetal peripheral blood, afferent lymph, lymph node and efferent lymph. The results show that expression of class II by T cells depends on the age of the animal and the physiological location of the T cell. In fetal blood there is no expression of class II on CD8+ or T19+ cells and very low expression on CD4+ T cells. In adult peripheral blood and efferent lymph a significant proportion of cells express DR but not DQ. A very different situation is found in afferent lymph and the peripheral lymph node: in afferent lymph the majority of T cells in all three subsets express both DQ and DR molecules; in the lymph node over 50% of T cells express DR and 30% are DQ+. These results suggest that within all T cell subsets there is a progression from DQ-DR- to DQ-DR+ and DQ+DR+ which correlates with physiological stages of T cell differentiation in vivo.

Fine restriction analysis and inhibition of antigen recognition in HLA-DQ-restricted T cells by major histocompatibility complex blockers and T cell receptor antagonists.

The role of polymorphic residues of the beta chain of human histocompatibility leukocyte antigen-DQw5/w6 in antigen presentation to a hepatitis B surface antigen-specific T cell clone was studied. The results obtained demonstrate that the residue situated at position 57 of the beta chain (a valine) is critical for presentation of antigen by antigen-presenting cells to the DQ-restricted T cell clone. Experiments were also done to study the feasibility of peptide blocking of antigen recognition by DQ-restricted T cells. The results indicate that peptides known to associate with DQ molecules are capable of blocking the presentation of antigen to the DQ-restricted T cell clone, presumably by competing with antigen for binding to major histocompatibility complex (MHC) molecules. Moreover, truncations of the stimulatory antigenic peptide resulted in the production of T cell receptor antagonists, which inhibited the response of the T cells to antigen at 10-100-fold lower concentrations than conventional MHC blockers. The role of DQ-restricted T cell responses and peptide blocking approaches in autoimmunity are discussed.

Polymorphism of the 5' flanking region of the HLA-DQA1 gene in coeliac disease.

Coeliac disease (CD) is associated with particular HLA genotypes. The susceptibility gene (or genes) has been mapped to the class II region, most probably to the DQ loci. Polymorphism of the upstream promoter region of the DQA1 gene (QAP) has been recently reported. At least ten variants or QAP alleles have been found, some of which are present in the cis-acting regulatory sequences. Allelic differences in DQ molecule expression may play a role in susceptibility to CD. We investigated the QAP polymorphism in 102 CD patients and 142 unrelated healthy controls of Czech origin using polymerase chain reaction amplification (PCR) of genomic DNA and oligonucleotide probes. We found a significant frequency increase of the alleles QAP 4.1 (RR = 10.3, p.c. = 10(-6) and QAP 2.1 (RR = 2.4, p.c. = 0.017) in patients over controls. An increased susceptibility is provided by the presence of both alleles, as is shown by the higher proportion of QAP 4.1, 2.1 heterozygotes among patients than expected from the Hardy-Weinberg equilibrium and by the comparison of the odds ratios for these alleles. There is a strong linkage disequilibrium between the QAP alleles and the DQA1, DQB1, and DRB1 loci. Two haplotypes carrying the QAP alleles whose frequency is increased are predominant in this group of CD patients: DQB1*0201, DQA1*0501, QAP4.1, DRB1*0301 and DQB1*0201, DQA1*0201, QAP 2.1, DRB1* 0701. Thus, the QAP variants are increased as part of these haplotypes and we cannot discriminate if they are responsible for the primary association.

Early events in immune evasion by the lentivirus maedi-visna occurring within infected lymphoid tissue.

Infections caused by lentiviruses, including human immunodeficiency virus, are characterized by slowly progressive disease in the presence of a virus-specific immune response. The earliest events in the virus-host interaction are likely to be important in determining disease establishment and progression, and the kinetics of these early events following lentiviral infection are described here. Lymphatic cannulation in the sheep has been used to monitor both the virus and the immune response in efferent lymph after infection of the node with maedi-visna virus (MVV). Viral replication and dissemination could be detected and consisted of a wave of MVV-infected cells leaving the node around 9 to 18 days postinfection. No cell-free virus was recovered despite the fact that soluble MVV p25 was detected in lymph plasma. The maximum frequency of MVV-infected cells was only 11 in 10(6) but over the first 20 days of infection amounted to greater than 10(4) virus-infected cells leaving the node. There was a profound increase in the output of activated lymphoblast from the lymph nodes of infected sheep, characterized by an increased percentage of CD8+ lymphoblasts. All of the CD8+ lymphoblasts at the peak of the response expressed both major histocompatibility complex class II DR and DQ molecules but not interleukin-2 receptor (CD25). The in vitro proliferative response of efferent lymph cells existing the node after challenge with MVV to both recombinant human interleukin-2 and the mitogen concanavalin A was decreased between days 8 and 16 postinfection, and a specific proliferative response to MVV was not detected until after day 15. Despite the high level of CD8+ lymphoblasts in efferent lymph, direct MVV-specific cytotoxic activity was demonstrated in only one of the five MVV-challenged sheep. MVV-specific antibody responses, including neutralization and MVV p25 immune complexes in efferent lymph, were detectable during the major period of virus dissemination. The relationship of these findings to the evasion of the host's acute immune response by MVV is discussed.

MHC class II molecules deliver costimulatory signals in human T cells through a functional linkage with IL-2-receptors.

MHC class II-positive T cells are found in tissues involved in autoimmune and infectious disorders. Because stimulation of class II molecules by mAb or bacterial superantigens induces protein tyrosine phosphorylation through activation of PTK3 in T cells, we hypothesized that class II signals play a regulatory function in T cell activation. Here, we show that cross-linking HLA-DR and -DP but not -DQ molecules by immobilized mAb enhanced proliferative T cell responses to IL-2. In contrast, class II stimulation had no effect on IL-4-induced proliferation. The costimulatory effect was most pronounced at low concentrations of IL-2, was blocked by IL-2R mAb, and was at least partly mediated through an up-regulation of IL-2 high affinity receptors. As expected, activation of IL-2R by IL-2 induced tyrosine phosphorylation of several proteins including p56lck, and class II cross-linking by mAb induced tyrosine phosphorylation of specific substrates including PLC-gamma 1. Combined stimulation of IL-2R and class II molecules had an additive effect on tyrosine phosphorylation. Pretreatment of T cells with a protein tyrosine kinase inhibitor, herbimycin A, inhibited IL-2 and class II-induced proliferation suggesting that class II costimulation of IL-2 responses may involve activation of tyrosine kinases. Taken together, the present data suggest that extensive cross-linking of class II molecules delivers costimulatory signals that enhance IL-2 sensitivity in human T cells.

Particular HLA-DQ molecules play a dominant role in determining susceptibility or resistance to Type 1 (insulin-dependent) diabetes mellitus

Genes in the HLA complex are by far the most important in determining genetic predisposition or resistance to Type 1 (insulin-dependent) diabetes mellitus. In this review evidence is presented that the HLA genes mainly involved are those encoding some particular HLA-DQ molecules. Both among Black, Caucasian and Japanese subjects particular cis or trans encoded DQ molecules are significantly associated with susceptibility, while others are associated with resistance. A varying degree of susceptibility or resistance seems to be conferred by these DQ molecules, where those determining resistance are dominant over those determining susceptibility. The degree of genetic predisposition to develop Type 1 diabetes carried by an individual would therefore be the result of his or her particular combination of DQ molecules. A primary association to particular DQ molecules explains previously found associations to other HLA complex genes by linkage disequilibrium. Some mechanisms by which particular DQ molecules may determine susceptibility or resistance are also discussed. Potential islet beta-cell reactive CD4+ T-cells may escape negative selection (deletion) in the thymus, but normally become anergized or remain ignorant extra-thymically. However, under particular circumstances they may be triggered. The DQ molecules associated with Type 1 diabetes susceptibility may preferentially bind and present triggering and/or beta-cell derived peptides to such T cells, causing beta-cell destruction. The finding that particular DQ molecules determine susceptibility may lead to new methods of preventing development of Type 1 diabetes in susceptible individuals.

T cell epitope specificity in human allergic and nonallergic subjects to bee venom phospholipase A2.

Phospholipase A2 (PLA) is a biochemically fully defined glycoprotein, representing the main allergen of bee venom. We have established CD4+ T cell clones specific to PLA from subjects allergic and nonallergic to bee sting. By screening the epitope specificity of these clones with 62 synthetic overlapping dodecapeptides representing the PLA molecule, two immunogenic epitopes, PLA81-92 and PLA113-124, were identified. Additional screening, using longer peptides of up to 18 residues, revealed a third epitope at position PLA 45-62. These epitopes are recognized by specific T cell clones in association with HLA-DP and -DQ molecules, although HLA-DR-associated responses to PLA exist. Primary in vitro proliferation of unfractionated PBMC from bee sting allergic, hyposensitized, or hyperimmune subjects to PLA-derived peptides revealed the same immunogenic sites as found in the T cell clones. Among the different groups of individuals, proliferative responses to the PLA molecule and fragments thereof were generally higher in allergic patients than in nonallergic subjects. Thus, at least three linear epitopes are involved in T cell recognition of PLA, independently whether or not subjects are allergic.

Use of transfectants to characterize a monoclonal antibody recognizing a monomorphic DR β‐chain epitope shared by some DQ and DP molecules

Use of transfectants to characterize a monoclonal antibody recognizing a monomorphic DR β‐chain epitope shared by some DQ and DP molecules

Molecular analysis of HLA class II genes in primary sjo¨gren's syndrome: A study of Israeli Jewish and Greek Non-Jewish patients

Abstract In an attempt to define the role of HLA class II genes in predisposition to primary Sjo¨rgen's syndrome, patients of two different ethnic groups (Israeli Jews and Greeks of non-Jewish origin) suffering from this disorder were studied. Oligonucleotide genotyping revealed the majority in both groups to carry either DRB1 *1101 or DRB1*1104, alleles tha re in linkage disequilibrium with DQB1*0301 and DQA1*0501. The high frequency of the two alleles in these SS patients is in contrast with the accepted association of primary SS with HLA-DR3 in Italian and American individuals. Molecular analysis of DQB1 and DQA1 alleles found in American Caucasian and American black SS (or SLE) patients demonstrated high frequencies of DQB1*0201 and DQA1*0501. The fact that the majority of SS patients, across racial and ethnic boundaries, carry a common allele, DQA1*0501, implies its involvement in the predisposition t to primary SS. Based on sequence analysis and the computer imaging of the HLA class II molecule structure, a hypothetical model for the role of the DQ molecule in promoting primary SS us proposed.

T cell clones specific for p21 ras‐derived peptides: Characterization of their fine specificity and HLA restriction

Peptides corresponding to the mutated regions of the oncoprotein p21 ras are immunogenic and capable of eliciting HLA class II-restricted T cell responses. Here we report studies on the fine specificity of four T lymphocyte clones (TLC) from a single donor, using various truncated peptides derived from the residues 6-19 of p21 ras and a panel of well-characterized HLA homozygous cells as antigen-presenting cells. Putative minimum peptides of nine or ten amino acids could be defined for each TLC. Two of the TLC recognized peptides presented by DR2, and the two others recognized peptides presented by DQ6. Some notable differences in the requirement for certain amino acids were seen between the DR- and DQ-restricted TLC. Thus, Ser at residue 17 was required for stimulation of the DQ6-but not the DR2-restricted TLC. Val at residue 8 was essential for stimulation of all TLC, whereas one of the DR2-restricted TLC also required Val at residue 7. Some peptides which were nonstimulatory were still capable of binding to DQ6 molecules in peptide competition experiments. The results may be of importance for potential immunotherapy of cancer where transforming ras oncoproteins are involved.

Distinct associations of HLA Class II genes with inflammatory bowel disease

Background: There are relatively few studies of HLA class II association either with Crohn's disease (CD) or ulcerative colitis (UC). The few available association studies have been carried out by serological techniques, and the results from these studies are inconclusive. Methods: The association between HLA class II genes was studied using molecular genotyping in combination with allele-specific oligonucleotide hybridization by polymerase chain reactions. Results: In UC (n = 74), we observed a positive association with the HLA DR2 allele (P = 0.008) and negative associations with the DR4 (P = 0.018) and DRw6 (P = 0.028) when compared with ethnically matched controls (n = 77). No associations were observed with any DQ alleles. In contrast, in CD (n = 95) we observed a positive association with the combination of DR1 and DQw5 alleles (P = 0.021). Furthermore, stratifying DR1 and DQw5 alleles indicated that neither allele was independently associated with CD, suggesting that the association was with the haplotype rather than either of the alleles individually. A suballele of DQw5, DQB1 ∗0501, contributed this haplotypic association (P = 0.012). Conclusions: DR and DQ molecules firmly separate UC and CD on genetic grounds, suggesting that the contribution of the HLA class II genes to the disease susceptibility is quite different for the two disorders.

Defective HLA class II expression in monocytes of type 1 diabetic patients. The Childhood Diabetes in Finland Study Group.

Activation of T-helper cells is modulated by the intensity of HLA class II expression on antigen-presenting cells. We evaluated whether any abnormalities could be found in the expression of HLA-DR and -DQ molecules on monocytes in type 1 diabetic subjects. DR and DQ molecules were induced by human recombinant interferon-gamma on cultured peripheral blood monocytes obtained from children with type 1 diabetes (N = 28), their siblings (N = 18) and unrelated healthy controls (N = 21). The response in DQ induction varied considerably between different individuals, but the average responsiveness was significantly lower in patients compared to siblings and unrelated controls. In addition to the diabetic subjects deficient DQ induction was also observed in three siblings. One of them had high levels of islet cell antibodies and presented with diabetes 6 months later, and another had active rheumatoid arthritis. The response in DR induction was also slightly lower in patients than in siblings, but did not differ from that in unrelated controls. The results suggest abnormalities in the regulation of HLA class II expression in type 1 diabetic subjects possibly reflecting the ongoing autoimmune process.

Relative abilities of distinct isotypes of human major histocompatibility complex class II molecules to bind streptococcal pyrogenic exotoxin types A and B.

The relative ability of distinct isotypes of human leukocyte antigen class II molecules to bind streptococcal pyrogenic exotoxins A and B (SPE A and SPE B, respectively) was investigated by a direct-binding assay with 125I-labeled toxin for SPE A and by a functional assay system measuring the accessory cell activity of human leukocyte antigen class II transfectants in toxin-induced T-cell activation for SPE A and SPE B. SPE A binding was observed in L cells transfected with DQw1 genes. By contrast, it was not detected in L cells transfected with DR2, DR4, DPw4 or DP(Cp63) genes. All the transfectants supported SPE-induced interleukin-2 production by human T cells except the DP transfectants for SPE B. Levels of accessory cell activity were low in the DP transfectants induced by stimulation with SPE A and in the DR and DP transfectants induced by SPE B. The results indicate that SPE A and SPE B bind well to DQ molecules, less well to DR molecules, and very weakly to DP molecules.

Studies of RFLP-inferred HLA-DR-DQ haplotypes in Danish women with recurrent fetal losses.

Results of HLA-DR and -DQ typing by RFLP in 152 unrelated women with at least 3 unexplained fetal losses were compared with those of 210 normal controls. The overall distribution of DR-DQ phenotypes did not differ significantly between patients and controls. In a subgroup of 59 patients having had at least 4 fetal losses, a significantly higher number of patients than controls were DRw17,DQw2-positive (RR = 2.9, pcorrected less than 0.02). DR-DQ haplotypes which carry DQB1 alleles encoding an amino acid other than aspartic acid at codon 57 in the second exon (non-Asp57 haplotypes) have been reported to confer susceptibility to several immunologically-mediated diseases. We found that the total frequency of non-Asp57 haplotypes (other than DR7,DQw2 which has unique features) was significantly increased in patients (50.0%) compared with controls (39.3%) (p less than 0.005). In the total group of patients and in the group with at least 4 fetal losses, 22.4% and 32.2% respectively were homozygous non-Asp57/non-Asp57 compared with 13.3% among controls (RR = 1.9, p less than 0.025 and RR = 3.1, p less than 0.001, respectively). The results suggest that DRw17,DQw2 confers susceptibility to suffer recurrent fetal losses. However, it is possible that the HLA-associated susceptibility may be primarily conferred by DQ molecules lacking aspartic acid at codon 57 in the DQ beta chain.