Major histocompatibility complex phenotypes influence serum testosterone concentration.

OBJECTIVES: (a) To confirm our earlier observation that the phenotype HLA-DR4,7 occurs with higher frequency in male patients with rheumatoid arthritis (RA) than in female patients. (b) To test the hypothesis that DR7 is associated with low normal serum testosterone (Te) levels in healthy males; this might explain the increased frequency of DR4,7 in male patients since there appears to be a relationship between low serum Te and RA. (c) To characterize the association between HLA alleles and serum Te concentration in healthy males. METHODS: An additional 82 Newfoundland (NF) RA patients were HLA-DR typed and, combined with our earlier data and data from the 11th International Histocompatibility Workshop, gave HLA-DR and sex information on 373 RA patients. Ninety-four healthy NF males were typed for HLA, the microsatellite marker TNFa (located close to the tumour necrosis factor alpha gene) and complement factor B (BF). An additional 38 males were included, selected partly based on their HLA-B type. RESULTS: We confirmed our earlier finding of a higher frequency of HLA-DR4,7 in male RA patients compared with female RA patients (P<0.01). Contrary to our expectations we found that DR7 was associated with higher than mean values of Te as were B5, B27, DR1, TNFa7 and BF F positivity. Conversely, low Te concentrations were found in men with B15, DR2, DR5, TNFa5 and who were BF F negative. In 28 male 'early-onset' RA patients we did not find an increased frequency of HLA alleles associated with low Te levels as compared with the frequency in 41 'late-onset' patients, suggesting that if low Te level is a risk factor and is present before onset of RA then the level cannot be explained by an association between Te level and major histocompatibility complex (MHC) phenotype. CONCLUSION: This study indicates that a man's MHC phenotype may influence his serum Te concentration, but the relationship of this, if any, to the pathogenesis of RA remains an area of speculation.

Evidence showing that the 1105 and 1106 isotypic residues of the fourth component of human complement, C4A, are not involved in amide bond formation.

Human C4A and C4B have different functions that may stem from their ability to bind hydroxyl or free amino groups on complement activating surfaces. Previous studies suggest that C4B binds to hydroxyl or amino groups whereas C4A binds to free amino groups on acceptor molecules. Comparison of the derived amino acid sequences of C4A and C4B has shown that differences exist between them at positions 1101, 1102, 1105 and 1106. These residues appear to be involved in the binding specificity of C4B. Less is known about the corresponding residues of C4A. It has been suggested that the aspartic acid of C4A at position 1106 is involved in amide bond formation by serving as a catalytic residue for the reaction or by promoting an increased interaction with amino nucleophilic groups. To examine the functional role of residues 1101-1106, we studied the effects of the C4A site-specific antipeptide mAb, AII-1 in assays dependent on the covalent binding properties of C4A; the C4 mediated inhibition of hemolysis and the C4 mediated inhibition of immune precipitation. This study shows that mAb AII-1 has no effect on C4-mediated hemolysis or its ability to inhibit the rate of immune precipitate formation. The lack of interference by AII-1 in these assays could not be explained by low affinity interaction between antibody and C4A showing that mAb AII-1 does not affect the covalent binding activity of C4A. Furthermore, results from epitope mapping studies show that AII-1 binds to Leu1105 and Asp1106 suggesting that these residues are not critical for amide bond formation by C4A.

Amino acid residues 1101-1105 of the isotypic region of human C4B is important to the covalent binding activity of complement component C4.

The C4A and C4B isotypes of human C4 show certain functional differences that stem from their relative preference for transacylation to amino (-NH2) vs hydroxyl (-OH) nucleophiles, respectively, on complement-activating surfaces. Comparison of amino acid sequences of the alpha-chain fragment of C4, C4d, has shown C4A- and C4B-specific sequences at residues 1101-1106 are the only consistent structural difference between isotype, i.e., Pro, Cys, Pro, Val, Leu, Asp in C4A and Leu, Ser, Pro, Val Ile, His in C4B. These residues may be responsible either in part or entirely for properties associated with isotype. To examine the functional role of residues 1101-1106 in C4B-mediated hemolysis, whole serum or immunopurified human C4 with allotypes, A3B1, A3, B2B1, or B1 were preincubated in the presence or absence of an antipeptide mAb (BII-1) specific for amino acid residues 1101-1105 of C4B. Sensitized sheep E and C4-deficient guinea pig serum was then added and lysis measured by absorbance at 415 nm. Our results show lysis of antibody-sensitized sheep E is inhibited by antibody and C4B2B1, C4B1, or C4A3B1 but not antibody and C4A3. The interference of hemolysis by BII-1 could not be explained by inhibition of activation of C4B or inhibition of C3 or C5 convertase activity. Furthermore, results from uptake experiments show that BII-1 interferes with the covalent binding activity of C4B, indicating residues 1101-1105 play a role in the covalent binding reaction of C4B to the target E-antibody complex.

Monoclonal antipeptide antibodies against amino acid residues 1101-1106 of human C4 distinguish C4A from C4B.

Comparison of amino acid sequences of the alpha-chain fragment of human C4, C4d, has shown C4A- and C4B-specific sequences at residues 1101-1106 in which the aspartic acid-histidine substitution at position 1106 may be related to the amide and ester bond forming properties of these molecules. Peptides containing twelve amino acid residues of the C4A- or C4B-specific sequences were synthesized and injected into female Balb/c mice. Serum from 2 mice, one immunized with the C4A-specific peptide and the other with the C4B-specific peptide, gave strong isotype-specific responses in an enzyme-linked immunosorbent assay against affinity-purified C4A3 and C4B2B1. Spleen cells from these mice were fused with the mouse myeloma SP2/0-Ag 14, and two cloned cell lines, AII-1 and BII-1, were established from hybrids. Enzyme-linked immunosorbent assay and western blotting of monoclonal antibodies AII-1 and BII-1 show that the former reacts with the C4A but not with the C4B alpha-chain and the latter with C4B but not with the C4A alpha-chain. Furthermore, immunoblotting of C4 allelic variants showed that AII-1 reacted with all C4A allotypes tested, including A6, A4, A3 and A2, whereas BII-1 reacted with all C4B allotypes tested, including B5, B3, B2, and B1.

The fluid-phase binding of human C4 and its genetic variants, C4A3 and C4B1, to immunoglobulins.

Covalent binding of the fourth complement protein, C4, to immune complexes is an important first step in the complement mediated processing of the complexes. Many of the initial encounters between the proteins of the complement system and antigen and antibody occur in solution, and prior to this report, studies of the interactions between them have focused on complement binding to preformed immune precipitates that most likely are not found in vivo. We have characterized the covalent binding of C4b to immunoglobulin molecules in a fluid-phase system consisting only of antibody in solution and purified C4 and C1s. We demonstrate that human C4b binds to IgG in the fluid phase, that its covalent binding is predominantly to the heavy chain of IgG, and that the covalent linkage is by either amide or acyl ester bonds. In addition, we compare the covalent binding efficiencies of two genetic variants of C4, C4A3 and C4B1, to IgG. C4A3 binds 3-4 times more IgG than C4B1 over a range of C4 concentrations, and C4A3 has a higher binding efficiency than C4B1 for IgM, IgA, IgG2a and F(ab')2 as well as for a protein antigen, BSA. Furthermore, we found that whereas C4A3 is bound to immunoglobulins in the fluid-phase predominantly by amide linkage, C4B1 is bound by either amide or acyl ester bonds. The results presented here suggest that the covalent binding efficiency of C4A3 and C4B1 to IgG is similar to that reported for their covalent binding to small molecules.

Major histocompatibility complex (MHC) factors predisposing to and protecting against Graves' eye disease.

We investigated the influence of gene mapping within the major histocompatibility complex on the susceptibility to Graves' eye disease. We studied 133 randomly selected patients with Graves' disease, many of whom had eye disease. HLA B8 and DR3 carried the greatest risk for disease but the difference between the two patient groups was not statistically significant. An earlier finding that Hungarian patients with a subset of B8, (B8 + DR7 +) had a greatly enhanced risk for eye disease was confirmed in Newfoundland patients. HLA B8 and DR7 are probably carried on different homologous chromosomes and their interaction enhances eye disease. HLA-DR4 was negatively correlated with eye disease. In particular, a subset of DR4 (B35 + DR4 +) appears to protect against eye disease. We have also derived the haplotypes of 22 probands, half of whom had eye disease. The haplotype data emphasized the high frequency of HLA A1 B8 DR3 C4A*QO and C4B*1 in both patient groups, 15% of the haplotypes in the group with eye disease and 25% in that without eye disease. Forty-one percent of haplotypes in the eye disease group and 32% in the no eye disease group were either C4A*QO or C4B*QO. In one proband with eye disease B8 and DR7 were carried on separate chromosomes. The phenotype DR4, C4A*3 C4B*1 was found in 3/20 haplotypes of patients without eye disease but in 0/20 of patients with eye disease. This finding is in keeping with the increased frequency of the DR4 C4A*3 C4B*1 in the patient group with no eye disease when 94 patients were phenotyped.(ABSTRACT TRUNCATED AT 250 WORDS)

C4-mediated inhibition of immune precipitation and differences in inhibitory action of genetic variants, C4A3 and C4B1.

The inhibition of immune precipitation is mediated by the classical complement pathway. We report here that the rate of precipitate formation depends on the genetic form of human C4 present during immune precipitation. C4A3 is more effective than C4B1 in its capacity to inhibit the rate of immune precipitate formation in serum and in serum-free reaction mixtures containing C1 and C4. Immune precipitates form within seconds after antigen is mixed with antibody, and the activation of the classical pathway is known to occur within seconds after C1 binds to antibody molecules. The covalent deposition of C4b on immune complexes is an essential step in the inhibition of immune precipitate formation, and if any of the reactions that lead to covalent C4b deposition become limiting, the rate of immune precipitation could exceed the complement system's inhibitory capacity. Hence, the inhibition of this rate may be an important function underlying the complement-mediated processing of immune complexes, and a decreased ability of the complement system to mediate this process in the presence of C4B1, in contrast to C4A3, could explain, at least in part, the association between the C4A-null phenotype and autoimmune diseases such as systemic lupus erythematosus.

Deletion of C4A genes in patients with systemic lupus erythematosus.

To define the relationship between inheritance of major histocompatibility complex (MHC) alleles and susceptibility to the development of systemic lupus erythematosus (SLE), we examined the MHC class I, II, and III phenotypes of white SLE patients and characterized the structures of their class III MHC genes, using Southern blotting. Nine of 88 SLE patients (10.2%) were C4A null. As detected by Southern blot analysis, the C4A gene was deleted from both chromosomes in 8 of the 9 C4A-null patients. Deletions affecting only 1 chromosome (heterozygous) were detected in the remaining C4A-null patient and in 34.5% of SLE patients who were not C4A deficient (compared with 12.5% of controls; P less than 0.05). These results indicate that deletion of the C4A gene is a common genetic marker for SLE. Deletions of C4A were observed most commonly as part of the HLA-B8;DR3 extended haplotype, although deletions were also detected in different HLA haplotypes. Because of the critical role of C4A in the processing of immune complexes, deficiency of C4A may, itself, confer susceptibility to the development of SLE.

Familial systemic lupus erythematosus: evidence for separate loci controlling C4 deficiency and formation of antibodies to DNA, nRNP, Ro and La.

A family has been identified in the Canadian province of Prince Edward Island in which 2 sisters have systemic lupus erythematosus in a sibship of 14. Studies are reported on 11 of the siblings and 16 other family members. The affected siblings, and 4 other members of their sibship, are halfnull homozygotes for the C4A component of complement. We studied the distribution in family members of antibodies to ss and dsDNA, and to Ro(SSA), La(SSB), Sm and nRNP. Eight of 11 members of the affected sibship are antibody producers, compared to only 3 of 13 members of the parental generation. Our study provides further evidence for an association between null genes for C4A and familial lupus, and suggests, in an unusually large kindred, that several other genetic factors are involved in the production of antinuclear antibodies.

Class III alleles and high-risk MHC haplotypes in type I diabetes mellitus, Graves' disease and Hashimoto's thyroiditis.

By typing a large quantity of family-based material for HLA-B, HLA-DR, C4, C2 and factor B, we were able to derive four-gene complement haplotypes (C4A, C4B, C2, BF) and six-gene MHC haplotypes (HLA-B, complement, HLA-DR). Fourteen six-gene MHC haplotypes showed linkage disequilibrium but exact frequencies could not be determined because it was not always possible to assign null C4 alleles in families where null genes were not clearly seen to segregate. Comparison of unrelated type I diabetes, Graves' disease and Hashimoto's thyroiditis patients with healthy unrelated controls revealed the following MHC allele associations: C4B*3, HLA-DR3 and HLA-DR4 with type I diabetes; BF*F1 and HLA-DR3 with Graves' disease; HLA-DR4 with Hashimoto's thyroiditis. By typing families of type I diabetes and Graves' disease patients we were able to derive two high-risk DR3+ MHC haplotypes for both type I diabetes and Graves' disease. These are HLA-B8 C4A*Q0 C4B*1 BF*S HLA-DR3 and HLA-B18 C4A*3 C4B*Q0 BF*F1 HLA-DR3, and these haplotypes account for most of the associations between these diseases and HLA-DR3. The MHC haplotype HLA-B15 C4A*3 C4B*3 BF*S HLA-DR4 also carries high risk for type I diabetes in this group of patients. Our data suggest that other DR4+ haplotypes, probably containing C4A*3 C4B*1, carry increased risk for type I diabetes whereas haplotypes containing DR4 and C4 C4A*3 C4B*Q0 do not. Our phenotype data suggest that DR4 in Hashimoto's thyroiditis is frequently associated with HLA-B44, C4A*3, C4B*1 and BF*S.

C4B3 allotype with a novel Ch phenotype.

The fourth component of complement (C4) has two classes of protein, C4A and C4B, both of which have many allelic forms. The serological determinants Rodgers (Rg1, Rg2) and Chido (Ch1, Ch2, Ch3) are generally associated with C4A and C4B, respectively. The C4B3 allotype has been detected in a single Canadian family that expresses a novel Ch phenotype, Ch:-1, 2, -3. There was no information for the Rg determinants, as the C4A*2B*3 haplotype would normally express Rg on the C4A protein. Other C4B3 allotypes in informative families have different Ch phenotypes, and the relationships of these within extended major histocompatibility complex haplotypes are discussed in this paper.

Study of alleles of the second complement component (C2) on Canadian HLA haplotypes.

Typing for genetic variation in the second complement component C2 was performed on sera from HLA haplotyped Canadian families. Part of the data has been studied and analysed as a population; in addition there is a further random collection of haplotypes bearing the C2*2 allele. In the population data there were 444 separate haplotypes from unrelated parents or other founders: 4.7% of the haplotypes carried the uncommon allele C2*2; one haplotype carried the rare C2*3. Study of C2 2-1 heterozygotes in the population data revealed 59 haplotypes which carried the common C2*1 allele and one which carried a deficiency allele C2*0. The remaining haplotypes carried either C2*1 or else an undetectable C2*0 allele. In the entire data there were 281 meioses informative for C2. The only recombinant between HLA-B and C2 showed the C2 locus to be on the DR side of the B locus. Strong allelic association between C2 *2 and Bw22 and less strong association between C2 *2 and B15 suggested the possibility of two ancestral C2 mutants. Examination of other markers on these and subsequently collected haplotypes do not conflict with this idea since the B15 haplotypes mostly carry C4A *4, C4B *2 whilst the Bw22 haplotypes mostly carry C4A *4, C4B *4. The alternative idea, that there was one original mutant which crossed over from a B15 to a Bw22 haplotype or vice versa is not excluded, however. Since approximate equilibrium has been reached between Bw22 and the HLA-A locus alleles on these C2 *2 bearing haplotypes, we conclude that this mutation is at least 5000 years old. Other haplotypes carrying C2 *2 are assumed to be ancestral recombinants; if this is true, the C2 locus map position between HLA-B and HLA-DR is confirmed. Study of C2 mutation may provide a model for understanding the genetics of some disease susceptibility genes in the HLA region.

HLA-B27 haplotypes in family studies of ankylosing spondylitis.

Although HLA-B27 carries a high relative risk for development of ankylosing spondylitis (AS), most B27 positive individuals do not have spondylitis. One interpretation of this observation is that there may be 2 types of B27, 1 which carries the risk factor and 1 which does not. If this were the case, then with the help of markers closely linked to HLA-B, it might be possible to detect differences between the B27 haplotypes in AS patients and those in healthy probands. We studied 197 members of 18 families with known AS and 110 members of 19 families in which HLA-B27 was present without any known inflammatory spinal disease. HLA antigens A, B, and C and alleles of complement components C2, C4, and Factor B and glyoxalase-1 were determined in all cases. Detailed haplotypes were assigned and their associations with development of the disease were examined. We were unable to identify any distinct HLA-B27 haplotype associated with AS; 2 common haplotypes and several miscellaneous ones were found in both groups. Thinking that the development of AS might be influenced by the other, non-B27 haplotype, we analyzed this and found that there was no detectable influence. The data do not contradict the notion that B27 in whites is a single entity and is itself the susceptibility factor predisposing to the development of ankylosing spondylitis.