Sex-determining region Y-related protein SOX13 is a diabetes autoantigen expressed in pancreatic islets.

The SOX (sex-determining region [SRY]-type high mobility group [HMG] box) family of transcription factors play key roles in determining cell fate during organ development. In this study, we have identified a new human SOX gene, SOX13, as encoding the type 1 diabetes autoantigen, islet cell antigen 12 (ICA12). Sequence analysis showed that SOX13 belongs to the class D subgroup of SOX transcription factors, which contain a leucine zipper motif and a region rich in glutamine. SOX13 autoantibodies occurred at a significantly higher frequency among 188 people with type 1 diabetes (18%) than among 88 with type 2 diabetes (6%) or 175 healthy control subjects (4%). Deletion mapping of the antibody epitopes showed that the autoantibodies were primarily directed against an epitope requiring the majority of the protein. SOX13 RNA was detected in most human tissues, with the highest levels in the pancreas, placenta, and kidney. Immunohistochemistry on sections of human pancreas identified SOX13 in the islets of Langerhans, where staining was mostly cytoplasmic. In mouse pancreas, Sox13 was present in the nucleus and cytoplasm of beta-cells as well as other islet cell types. Recombinant SOX13 protein bound to the SOX consensus DNA motif AACAAT, and binding was inhibited by homodimer formation. These observations-along with the known molecular interactions of the closely related protein, rainbow trout Sox23-suggest that SOX13 may be activated for nuclear import and DNA binding through heterodimer formation. In conclusion, we have identified ICA12 as the putative transcription factor SOX13 and demonstrated an increased frequency of autoantibody reactivity in sera from type 1 diabetic subjects compared with type 2 diabetic and healthy control subjects.

Autoantigenic reactivity of diabetes sera with a hybrid glutamic acid decarboxylase GAD67-65 molecule GAD67(1-101)/GAD65(96-585).

Glutamic acid decarboxylase (GAD) is a major autoantigen in insulin-dependent diabetes mellitus (IDDM). Two GAD isoforms exist, GAD65 and GAD67, which differ mostly in the first 100 amino acids of the amino terminus. IDDM sera are predominantly reactive with GAD65 but autoepitopes have been localised only to regions of GAD65 highly homologous with GAD67. In this study we investigated the contribution of the amino terminus to the IDDM epitope on GAD65, in order to test whether this region of GAD could explain the difference in reactivity between GAD65 and GAD67. A recombinant hybrid GAD molecule consisting of amino acids 1-101 of GAD67 and 96-585 of GAD65 was constructed and a truncated GAD65 was also constructed consisting of amino acids 98-585 of GAD65. The reactivity with the hybrid GAD molecule, GAD65 and GAD67, and truncated GAD65 was examined by radioimmunoprecipitation using 50 IDDM sera with known reactivity to purified porcine brain GAD. Over 90% of the IDDM sera were reactive with the hybrid GAD molecule confirming that the amino terminus of GAD65 does not contribute to the autoepitope and that the IDDM epitope is localised to the middle and carboxyl terminal domains of GAD65. Furthermore, evidence is presented that autoantibodies to GAD65 in IDDM sera react with an epitope formed on a dimeric configuration of the molecule.

Antibodies to SOX13 (ICA12) are associated with type 1 diabetes.

SOX13 is an islet cell autoantigen (ICA12), identified by antibody screening of an islet cDNA library, using sera from patients with Type 1 diabetes. We ascertained the frequency of antibody reactivity to SOX13 and compared it with other Type 1 diabetes autoantibody reactivities. Antibodies were measured by radioimmunoprecipitation (RIP) using (35) S labelled SOX13 expressed in rabbit reticulocyte lysate. Sera from 109 subjects with Type 1 diabetes, 29 with Type 2 diabetes, 144 with other autoimmune diseases and from 201 controls were tested for anti-SOX13, and results were compared with the frequency of antibodies to glutamic acid decarboxylase (anti-GAD), islet cell antigen 512 (anti-ICA512) and islet cell cytoplasm (ICA). Anti-SOX13 were detected in 20 (18.3%) of 109 subjects with Type 1 diabetes, and more frequently in adults than in children (29% vs 10%). Anti-SOX13 usually occurred with anti-GAD but rarely with anti-ICA512. Seven sera positive for anti-SOX13 did not react with either GAD, ICA512 or islet cell cytoplasm indicating that anti-SOX13 represented a distinct population of antibodies. Reactivity to SOX13 represents a further autoantibody response in adults with Type 1 diabetes and may provide a useful disease marker in subjects in whom other autoantibody tests are negative.

Correlations between holo-transcobalamin II, holo-haptocorrin, and total B12 in serum samples from healthy subjects and patients.

AIMS: To study the correlations between total vitamin B12(B12), holo-haptocorrin, and holo-transcobalamin II (holo-TCII) concentrations in human sera; the association between reduced holo-TCII concentrations and macrocytosis attributable to B12 deficiency. METHODS: Serum samples from 38 healthy volunteers, 113 patients with normal total serum B12 concentrations and 93 patients with low total serum B12 were studied. Holo-TCII was removed from whole serum by adsorption with amorphous precipitated silica, and both whole serum and adsorbed serum were assayed for B12 using the Becton Dickinson vitamin B12 [57Co] radioassay kit. RESULTS: In all three groups of subjects studied there were strong correlations between the logarithms of the total serum B12 and the holo-haptocorrin concentrations with regression coefficients between 0.884 and 0.967. By contrast, the correlations between the logarithms of the total serum B12 and holo-TCII concentrations were weaker, especially in the patients with normal or low total serum B12, for whom the regression coefficients were 0.491 and 0.391, respectively. Analysis of the clinical records of a proportion of the patients studied indicated that there were many more patients with low holo-TCII concentrations than with haematological disturbances related to B12 deficiency. CONCLUSIONS: The total serum B12 concentration is a relatively poor indicator of holo-TCII concentrations and, therefore, of the ability of serum to deliver B12 to tissues. Additional information regarding B12 values can therefore be gleaned from measuring holo-TCII concentrations in the serum. Low holo-TCII concentrations, however, are an early sign of negative B12 balance and are frequently unassociated with haematological abnormalities caused by B12 deficiency.

Antibodies to diabetes-associated autoantigens in Indian patients with Type 1 diabetes: prevalence of anti-ICA512/IA2 and anti-SOX13.

We ascertained frequencies of autoantibodies to a suite of islet cell antigens including ICA512/IA2 and SOX13 in Asian Indians with Type 1 diabetes and in other forms of diabetes. Autoantibodies to ICA512/IA2 and SOX13 were tested by radioimmunoprecipitation assay, and results were amalgamated with previous data on antibodies to glutamic acid decarboxylase (GAD) and to islet cell cytoplasmic antigens (ICA). The frequency of anti-SOX13 was higher in Asian Indians than in Europids. Overall, the combined frequency for all autoantibodies to diabetes-associated antigens in Type 1 diabetes in Indians approached the frequency reported for Europids. There was an unexpectedly high frequency of autoantibody reactions to any one of the autoantigens tested (24%) in fibrocalculous pancreatic diabetes, however, individual autoantibody frequencies were relatively low. Our data indicate that, whatever the population studied, testing for multiple autoantigenic reactivities is more informative than more limited testing, and that there may be regional (presumably ethnically based) differences in levels of particular autoantibodies in cases of Type 1 diabetes.

Bone marrow cells from vitamin B12- and folate-deficient patients misincorporate uracil into DNA.

Bone marrow cells from 15 patients with normal deoxyuridine (dU) suppression test results, 3 healthy subjects, and 11 patients with megaloblastic anemia caused by vitamin B12 or folate deficiency were examined for misincorporation of uracil into DNA. Cells were incubated with [5-3H] uridine for 2 hours and their DNA extracted. The DNA was hydrolyzed to deoxyribonucleosides with DNase 1, phosphodiesterase and alkaline phosphatase, and any dU present was separated from other deoxyribonucleosides by Aminex A6 chromatography. The quantity of dU/mg DNA and the radioactivity in the dU peak/mg DNA were then calculated. The results clearly showed that there was markedly increased uracil misincorporation into the DNA of vitamin B12- or folate-deficient marrow cells. Misincorporation of uracil into DNA may be an important biochemical lesion underlying both the megaloblastic change and the ineffectiveness of hematopoiesis in vitamin B12 and folate deficiency.

Misincorporation of uracil into the DNA of folate- and B12-deficient HL60 cells.

HL60 cells were cultured for 10 days under various experimental conditions. They were then incubated with 1 mumol/l [5-3H] uridine for 2 hours and their DNA extracted. The DNA was hydrolysed to deoxyribonucleosides with phosphodiesterase and alkaline phosphatase and the hydrolysate subjected to Aminex A6 chromatography. The elution profiles showed that, when compared with control cells, DNA from cells grown in medium deficient in folate, B12 or both folate and B12 contained increased amounts of deoxyuridine (dU) and increased radioactivity in the dU peak. The data demonstrate that misincorporation of uracil into DNA occurs in a myeloid cell line cultured in growth medium deficient in folate, B12 or both folate and B12.

Phenobarbital: a non-genotoxic agent which induces the repair of O6-methylguanine from hepatic DNA.

Exposure to phenobarbital (PB) (0.05% in drinking water) markedly increased the rate of repair of O6-methylguanine (O6-MeG) from the hepatic DNA of rats given N-nitrosodimethylamine (2 mg/kg). No effect of comparable magnitude was seen for the repair of O4-methylthymine. During 21 weeks of exposure to PB the increased repair of O6-MeG exhibited a biphasic response and was maximal at approximately 3 weeks of treatment. Although this increased repair was readily observed when direct measurements were made of the loss of O6-MeG from hepatic DNA in vivo, no corresponding increased level of methyltransferase activity was detected in cell-free liver extracts, indicating that the methyltransferase protein was induced in a relatively limited population of cells. Immunohistochemical procedures have been used to demonstrate the formation of O6-MeG in, and its repair from, the DNA of hepatocytes in the centrilobular region of the liver lobule. Comparison with published data, for changes in the level of asialoglycoprotein receptors [Evarts et al. (1985) Carcinogenesis, 6, 1767-1773] and for the induction of cytochrome P450 [Schwartz et al. (1987) Carcinogenesis, 8, 1355-1357] in hepatocytes during PB administration, indicate that PB is acting at membrane sites in a relatively limited population of cells associated with the central vein. These observations show that the methyltransferase activity responsible for the repair of the major promutagenic base O6-MeG can be induced by a membrane active agent, without recourse to the genotoxic action of initiators and toxins, or the induction of restorative hyperplasia, previously employed for this purpose.

Promutagenic alkyl lesions are induced in the tissue DNA of animals treated with isoniazid.

Further studies have been carried out to determine mechanisms for the toxic and carcinogenic properties of isoniazid (INH) in laboratory animals. 1. Single doses of INH (1.1 mg per mouse), which if given continuously to Swiss mice result in a 50% incidence of lung tumours, lead to the formation of approximately 0.5 and approximately 0.3 mumol O6-methylguanine per mol/guanine in the DNA of liver and lung, respectively. 2. Repeated doses of INH result in a progressive decrease in the levels of O6-methylguanine in lung DNA and relatively constant levels in hepatic DNA. Treatment with equimolar doses of hydrazine result in higher levels of alkylation in the DNA of liver than of lung. 3. Comparable experiments in Wistar rats show that treatment with hydrazine is very much more effective than INH in inducing the alkylation of liver and lung DNA. 4. Immunocytochemical staining of cryostat sections of liver has been used to show that the formation of O6-methylguanine occurs mainly in the nuclei of hepatocytes. 5. These results demonstrate that treatment with INH leads to the alkylation of tissue DNA and suggest that this may arise via a hydrazine intermediate. The implications of the formation of highly promutagenic lesions in tissue DNA for INH induced toxicity and carcinogenicity are discussed.

Natural and disease associated autoantibodies to the autoantigen, dihydrolipoamide acetyltransferase, recognise different epitopes.

Naturally occurring autoantibodies are ubiquitous and may serve physiological functions. We examined the relationship of natural and disease-associated autoantibodies in the context of autoantibodies to dihydrolipoamide acetyltransferase, the 74 kDa E2 sub-unit of the mitochondrial pyruvate dehydrogenase complex (PDC-E2), characteristic of primary biliary cirrhosis (PBC). We tested for natural autoantibodies to PDC-E2 in normal sera, and compared epitopes recognised by natural and disease-associated autoantibodies. Methods included affinity purification of anti-PDC-E2 from normal and PBC sera, ELISA and immunoblotting, capacity of antibodies to inhibit the enzyme function of the pyruvate dehydrogenase complex (PDC), use of F(ab)2 fragments of anti-PDC-E2 in inhibition assays, and testing affinity purified anti-PDC-E2 on peptide fragments of PDC-E2. We found that natural auto-antibodies to PDC-E2 of IgG class were demonstrable in all healthy human sera (10/10). However, their reactivity differed from that of disease-associated autoantibodies, in that anti-PDC-E2 from normal serum failed to inhibit the catalytic activity of PDC; and F(ab)2 fragments from PBC sera potently blocked the binding of anti-PDC-E2 from PBC sera to PDC-E2, but not the binding of natural anti-PDC-E2 to PDC-E2. Immunoblotting on fragments of PDC-E2 using affinity-purified preparations from PBC sera and normal sera failed to provide evidence for gross differences in epitope reactivity. We conclude that normal human sera contain natural IgG autoantibodies to the immunodominant inner lipoyl domain of PDC-E2, as seen characteristically in PBC. However, there is evidence for differences in fine epitope recognition.

The peculiar autoimmunity of primary biliary cirrhosis.

Autoantibodies to mitochondria (AMA, anti-M2) are a serologic hallmark of primary biliary cirrhosis (PBC). These react with three structurally and functionally related multienzymic complexes, the 2-oxoacid dehydrogenase complexes, but chiefly with the E2 subunit of pyruvate dehydrogenase complex (PDC-E2). Their very dose (95%) and specific association with PBC underpins the autoimmune concept of pathogenesis of that disease, notwithstanding several non-congruent features. Detailed studies, including structural analysis of epitopes, do not disclose how these autoantibodies originate. Their ubiquity in PBC has overshadowed the existence of a second set of relatively PBC-specific autoantibodies to nuclear antigens for which reactants have been cloned and characterized. These include centromeric proteins; proteins of the nuclear pore complex; nuclear dot proteins, which include Sp-100 and the promyelocytic leukemia antigen; and a recently identified autoantigen, SOX13. Certain of these reactants are DNA-binding proteins with transcriptional regulatory activity. Thus serum from individuals with the same clinical syndrome can have autoimmune reactivity to disparate mitochondrial and nuclear constituents in different cellular compartments. Antibody probing of phage displayed random peptide libraries, together with epitope scanning using overlapping sequential octameric peptides from the PDC-E2 sequence, showed that the discontinuous motifs MH, FV(E) and SYP contributed to a predicted conformational antibody epitope in the inner lipoyl domain of PDC-E2.

Autoantibodies to glutamic acid decarboxylase in diabetic patients from a multi-ethnic Australian community: the Fremantle Diabetes Study.

AIMS: To investigate ethnic/racial differences in the prevalence of serum antibodies to glutamic acid decarboxylase (GADA) and ICA512/IA-2 in diabetic patients from a large, urban community. METHODS: A cross-sectional sample of 1,381 diabetic patients aged 11-98 years, representing 61% of those identified in a postcode-defined population base of 120,097 people were studied. Diabetes was classified on clinical grounds. Serum GADA and anti-ICA512/IA-2 were measured by radioimmunoprecipitation assay. RESULTS: Anglo-Celts formed 62% of the sample, southern Europeans 18%, other Europeans 8% and Asians 3%. GADA prevalence in Type 1 and Type 2 diabetes mellitus was 46.0% and 4.2%, respectively, amongst Anglo-Celts and 22.2% and 1.7% in southern Europeans. The prevalence of anti-ICA512/IA-2 in Type 1 diabetes was 17.4% and, in a sample of 233 patients with Type 2 diabetes, 0.8%. GADA-positive Type 2 patients had a lower body mass index and greater glycosylated haemoglobin, and were more likely to be taking insulin, than GADA-negative Type 2 diabetic subjects (P < 0.05), consistent with the phentoype of latent autoimmune diabetes of adults (LADA). In both Type 1 and Type 2 diabetes, there was a strong inverse association between GADA and serum triglycerides (P < 0.001). CONCLUSIONS: The relatively low GADA prevalence in Anglo-Celt patients with Type 1 diabetes is a feature of this community-based study and suggests that GADA levels do fall with time, given the older age of the sample and a relatively long period between diagnosis and sampling. Southern Europeans had an even lower GADA prevalence, regardless of diabetes type. Variations in GADA frequency in diabetic patients of differing European ethnicity has implications for clinical management and healthcare planning.

Autoantibodies to the islet cell antigen SOX-13 are associated with duration but not type of diabetes.

AIMS: The autoantigen SOX-13 of the SRY-related high mobility group box is a low-frequency reactant in sera from patients with Type 1 diabetes. We further investigated the potential diagnostic role of anti-SOX-13, and in particular its ability to distinguish Type 1 from Type 2 diabetes, in two large, well-characterized cohorts. METHODS: SOX-13 autoantibody status was ascertained using a radioimmunoprecipitation assay in (i) a random sample of 546 participants in an Australian community-based study (the Fremantle Diabetes Study; FDS) of whom 119 had Type 1 and 427 Type 2 diabetes, and (ii) a sample of 333 subjects with Type 2 diabetes from the United Kingdom Prospective Diabetes Study (UKPDS) stratified by age, anti-glutamic acid decarboxylase (GAD) and islet cell antibody (ICA) status, and requirement for insulin therapy within 6 years of diagnosis. RESULTS: The frequencies of anti-SOX-13 in the FDS subjects were 16.0% and 14.8% for Type 1 and Type 2 patients, respectively, and levels were similar. In the UKPDS subjects, the frequency was 4.5%. In a logistic regression model involving demographic, anthropometric and metabolic variables, only diabetes duration was significantly associated with anti-SOX-13 positivity, especially for duration > 5 years (P < 0.002). When the coexistence of autoantibodies was assessed in the two study samples, there were no significant associations between anti-SOX-13 and ICA, anti-GAD or ICA512/IA-2. CONCLUSIONS: Whilst the frequency of anti-SOX-13 may be increased in some populations of diabetic patients, this reactivity does not usefully distinguish Type 1 from Type 2 diabetes. However, the association with diabetes duration suggests that anti-SOX-13 may be a non-specific marker of tissue damage associated with chronic hyperglycaemia.