Michael J. Dobersen

Cytotoxic Autoantibodies to Beta Cells in the Serum of Patients with Insulin-Dependent Diabetes Mellitus

We studied serum from 36 patients with insulin-dependent diabetes mellitus (IDDM) for the capacity to lyse beta cells. Immunofluorescence revealed an islet-cell cytoplasmic antibody (ICA) in 20 patients with IDDM and an islet-cell-surface antibody (ICSA) in 23. Neither ICA nor ICSA was found in any of 21 normal controls or 15 patients with non-insulin-dependent diabetes. In the presence of complement. ICSA-positive serum caused significant lysis as measured by release of 51Cr (50.1 +/- 8.8 per cent) from cultured rat islet cells, but ICSA-negative serum did not (17.7 +/- 7.3 per cent) (P < 0.001). Proof that ICSA-positive serum was lytic for beta cells was obtained by a double-fluorescence technique that identified lysed cells by their capacity to take up ethidium bromide and beta cells by their staining with fluorescein-conjugated antibody to insulin. These findings suggest that cytotoxic ICSA contributes to the pathogenesis of IDDM, but the mere presence of ICSA does not appear to be sufficient to produce diabetes; family studies showed that one fourth of the serum samples from nondiabetic first-degree relatives of diabetic probands were ICSA-positive and cytotoxic for beta cells.

Congenital rubella syndrome as a model for Type 1 (insulin-dependent) diabetes mellitus: increased prevalence of islet cell surface antibodies

SummaryAn increased prevalence of Type 1 (insulin-dependent) diabetes has been reported in patients with congenital rubella. Rubella virus multiplies in the pancreas, and we have hypothesized that studies of children with congenital rubella would be of great importance in following the development of Type 1 diabetes in a defined, susceptible population. Two hundred and forty-one children with congenital rubella (mean age 17.4±0.3 years; 65% black and hispanic) have been evaluated, 30 of whom already have diabetes and 17 of whom have borderline glucose tolerance. In these latter two groups, HLA-DR3 is significantly increased and HLA-DR2 significantly decreased. Pancreatic islet cell cytotoxic surface antibodies are found in 20% of the total congenital rubella population, including in more than 50% in the time period before they develop diabetes and are not related to any specific HLA type. In addition, anti-microsomal and anti-thyroglobulin antibodies are found in 34% of this population. The data demonstrate that Type 1 diabetes developing in congenital rubella patients has the genetic and immunological features of classical Type 1 diabetes, namely the presence of HLA-DR3, the absence of HLA-DR2, islet cell surface antibodies before decompensation and an increased prevalence of anti-thyroid antibodies. Patients with non-diabetic congenital rubella represent an easily identifiable group in whom other immunological factors associated with Type 1 diabetes can be elucidated and possibly modified.

A novel procedure for labeling immunoglobulins by conjugation to oligosaccharide moieties

A novel method is described for the biotinylation of immunoglobulins. The procedure relies on the generation of reactive aldehydes on the carbohydrate moieties of the immunoglobulin by oxidation with sodium periodate and subsequent reaction with biotin hydrazide. The method is simple and specific and results in stable conjugates retaining full immunologic activity. It has been applied successfully to a number of mouse monoclonal antibodies of both IgG and IgM classes, and to human IgM preparations. The procedure may also be applied to conjugation of immunoglobulins with fluorescent dyes.

The species distribution of nervous system antigens that react with anti-myelin-associated glycoprotein antibodies

The reactivity of monoclonal and polyclonal antibodies directed against human central nervous system (CNS) myelin-associated glycoprotein (MAG) was investigated in a number of animal species. The antibodies included mouse monoclonal antibodies obtained by immunization with human MAG; HNK-1, a mouse monoclonal antibody raised against a human lymphoblastoma and used to identify a subset of lymphocytes with natural killer function; human IgM paraproteins associated with neuropathy; and polyclonal antibodies obtained from rabbits immunized with rat or human MAG. Following polyacrylamide gel electrophoresis of CNS and peripheral nervous system (PNS) tissue from human, bovine, cat, rabbit, guinea pig, rat, mouse, frog, gold fish and chicken, proteins were electrophoretically transferred onto nitrocellulose. The immune-staining of electroblots showed distinct interspecies variation in the reactivity of the antibodies with MAG. In addition, the species distribution of several low molecular weight glycoproteins present in PNS tissue that cross-react with anti-MAG antibodies was determined. These low molecular weight antigens are not present in CNS homogenates or in purified human CNS myelin. It was also shown that IgM from a patient with peripheral neuropathy and paraproteinemia associated with anti-MAG antibodies recognized these low molecular weight antigens. The results suggest that IgM paraproteins, HNK-1 and some mouse monoclonal antibodies react with carbohydrate determinants shared by MAG and several lower molecular weight glycoproteins present only in human, bovine, cat and chicken PNS. Rabbit polyclonal anti-rat MAG antisera and mouse monoclonal antibodies reacting with peptide epitopes of MAG are much more specific for detecting MAG than antibodies reacting with carbohydrate epitopes of human MAG. The results are discussed in relation to human demyelinating peripheral neuropathy associated with IgM paraproteinemia.

Generation and characterization of mouse monoclonal antibodies to the myelin-associated glycoprotein (MAG)

A panel of mouse monoclonal antibodies to rat and human myelin-associated glycoprotein (MAG) was developed. Normal mice were unresponsive to rat MAG, and successful immunization with rat MAG was only achieved in autoimmune NZB mice. By contrast, all strains of mice were responsive to human MAG. The monoclonal antibodies developed differ with respect to immunoglobulin type, their specificity for human and/or rat MAG, and their recognition of protein or carbohydrate epitopes in MAG. In general, the antibodies that react with the protein backbone recognize both rat and human MAG, whereas a large number of the monoclonal antibodies recognize a carbohydrate determinant in human MAG that is not in rat MAG. Immunocytochemical staining of adult human spinal cord with the monoclonal antibodies resulted in periaxonal staining of myelin sheaths, similar to that produced by well-defined, rabbit, polyclonal anti-MAG serum. In addition, the antibodies recognizing, carbohydrate determinants in human MAG strongly stained oligodendrocyte cytoplasm. These monoclonal antibodies will be of value for the further chemical and biological characterization of MAG.

A human lymphocyte antigen is shared with a group of glycoproteins in peripheral nerve

The monoclonal antibody HNK-1 binds to a carbohydrate determinant in the myelin-associated glycoprotein (MAG) and other glycoproteins of human peripheral nerve. Some glycoproteins of lower Mr than the major P0 glycoprotein of myelin appear to bind more antibody than MAG. These glycoproteins electrophorese in the Mr range of 20,000 to 26,000 and are present in the purified myelin fraction. The results indicate that an antigen on the surface of a subset of lymphocytes is shared with a group of glycoproteins in human peripheral nerve. The antigen appears to be similar to that recognized by IgM paraproteins associated with a type of neuropathy.

Preferential Lysis of Pancreatic B-Cells by Islet Cell Surface Antibodies

Sera from patients with insulin-dependent diabetes mellitus (IDDM) containing islet cell surface antibodies (ICSA) were studied for their capacity to lyse cultured rat islet cells. The uptake of ethidium bromide was used to identify lysed cells and immunofluorescent staining with antisera to insulin, glucagon, somatostatin, or pancreatic polypeptide was used to identify the different islet cell types (B-, A-, D-, and PP-cells, respectively). Our experiments showed that in the presence of complement, sera containing ICSA lysed 81% of the B-cells, but 10% or less of the A-, D-, and PP-cells. Normal control sera resulted in lysis of less than 4% of each of the islet cell types. The demonstration that ICSA are preferentially lytic for B-cells may be important in defining the role of these autoantibodies in the pathogenesis of IDDM, particularly since the B-cell mass in diabetics is markedly reduced relative to the other islet cell types.

HLA Antigens, Cytoplasmic Islet Cell Antibodies, and Carbohydrate Tolerance in Families of Children with Insulin-dependent Diabetes Mellitus

Cytoplasmic pancreatic islet cell antibodies were found in 21﹪ of 244 unaffected first degree relatives of type I diabetic patients. Twenty-five percent of HLA-identical, 35﹪ of HLA-haploidentical, 16﹪ of HLA-nonidentical siblings, and 14﹪ of parents were ICA-positive. In the HLA-identical sibs, irrespective of ICA, and in the 18 ICA-positive parents but not the other groups, increased plasma glucose levels were observed after the administration of glucose. In most children, these were associated with reduced insulin levels, while in the adults elevated insulin responses were noted. In 48﹪ of the ICA-positive children and 84﹪ of the ICA-positive parents, other evidence of “autoimmunity” was obtained either by history or by testing for specific autoantibodies. Two of the originally unaffected HLA-identical and ICA-positive siblings developed diabetes during the course of the study. These findings, plus previously reported data in families with two diabetic sibs demonstrating that the empiric risk for developing IDDM is of the order of 30﹪ for HLA-identical sibs but < 5﹪ for those that are HLA-haploidentical, suggest that HLA-identity may be a useful predictor of potential type I diabetes. The presence of ICA may, at times, portend the need for future antidiabetic therapy but prospective studies must be continued to fully elucidate this relationship.

Incorporation of 5-Substituted Analogs of Deoxycytidine into DNA of Herpes Simplex Virus-Infected or -Transformed Cells Without Deamination to the Thymidine Analog

The incorporation into DNA of 5-bromocytosine and 5-iodocytosine, derived from their respective administered deoxyribonucleoside analogs, has been demonstrated in studies with cells infected with herpes simplex virus types 1 and 2 (HSV-1 and HSV-2) and in cells transformed with the thymidine kinase gene of HSV-1. No significant incorporation of iodocytosine or iodouracil occurred in the DNA of uninfected or nontransformed cells when the deaminating enzymes were inhibited, in accord with past studies in our laboratory with 5-bromodeoxycytidine and tetrahydrouridine. When 2′-deoxytetrahydrouridine, a potent inhibitor of cytidine deaminase and dCMP deaminase, was utilized, all the counts in DNA that were derived from [125I]iododeoxycytidine appeared as iodocytosine in HSV-infected cells. In the absence of a deaminase inhibitor, 32 to 45% of the counts associated with DNA pyrimidines appeared as iodocytosine, and 55 to 68% appeared as iodouracil in HSV-infected cells. Substantial incorporation of iodocytosine (16%) occurred in cells transformed with the HSV thymidine kinase gene, suggesting the importance of the specificity of cellular nucleoside kinases and the activity of the deaminases in presenting unmodified bases to an undiscriminating polymerase. Incorporation into DNA of bromocytosine derived from [3H]bromodeoxycytidine was demonstrated in HSV-2 infected cells; very little incorporation of bromocytosine compared with bromouracil could be demonstrated in these cells in the absence of inhibition of the deaminases (19% of the total counts associated with pyrimidines with deaminase inhibition and 1.5% without). Limited studies with 5-methyl[5-3H]deoxycytidine indicated essentially no (or very little) incorporation of this analog as such in the DNA of HSV-1- and HSV-2-infected and -transformed cells. This suggests an exclusion or repair mechanism preventing inappropriate methylcytosine incorporation in DNA. The addition of nucleoside and deoxyribonucleoside deaminase inhibitors, which leads to the incorporation of 5-halogenated analogs of deoxycytidine into DNA as such, does not impair their antiviral activity. We infer from studies with 4-N-alkyl (ethyl and isopropyl)-substituted analogs of iododeoxycytidine that they are incorporated as such into DNA without deamination and effectively inhibit the virus at concentrations that are marginally toxic. Among the several reasons presented for the heightened potential efficacy of analogs of deoxycytidine compared with those of deoxyuridine is that the former, as analogs of 5-methyldeoxycytidine, may impair viral replication by perturbing processes involving methylation and changes in the methylation of deoxycytidine in DNA which appear to be important for the process of HSV maturation. In addition, this capacity to perturb methylation may, in turn, be the key to their potential as agents affecting entry into or emergence from latency, a process in which dramatic changes in the postpolymer 5-methylation of deoxycytidine occur in the DNA of herpesviruses.

Mouse monoclonal and rabbit polyclonal antibodies prepared to human myelin-associated glycoprotein also react with glycosphingolipids of peripheral nerve

A panel of mouse monoclonal antibodies and rabbit polyclonal antisera that were raised to myelin-associated glycoprotein (MAG) were screened for reactivity with acidic glycolipids from brain and peripheral nerve by enzyme-linked immunosorbent assay (ELISA) and/or a thin-layer chromatogram overlay technique. Seven out of 7 mouse monoclonal antibodies that recognize carbohydrate epitopes in human MAG also reacted with acidic glycolipids from human and cat peripheral nerve, while monoclonal antibodies that react with polypeptide epitopes on MAG did not react with these glycolipids. Rabbit anti-human MAG antisera also strongly reacted with the glycolipids from peripheral nerve, while rabbit antisera raised to rat MAG did not. None of the antibodies reacted with similar glycolipid fractions prepared from adult human brain. Overlay of thin-layer chromatograms revealed that all the mouse and rabbit antibodies showing reactivity with peripheral nerve glycolipids were binding to the same two sphingoglycolipids that react with human anti-MAG IgM paraproteins in neuropathy and with HNK-1 (anti-Leu-7), a mouse IgM monoclonal antibody that identifies a subset of human lymphocytes with natural killer function. Thus, the carbohydrate epitope(s) in MAG which is shared with nerve acidic glycolipids appears to be highly immunogenic in mice and rabbits. Further, it is clear that the antibodies that react with the carbohydrate moieties of human MAG cannot be used as specific probes for this glycoprotein.