Total IgG and IgG subclass specific antibody responses to insulin in diabetic patients

Abnormal antibody responses to insulin in diabetic patients have been associated with syndromes of insulin hypersensitivity and abnormal insulin pharmacokinetics. In this study, we evaluated total and IgG subclass antibody responses to insulin in 70 diabetic subjects on insulin distributed into five clinical groups, and in two control groups using ELISAs with CDC/WHO recommended monoclonal antibodies. As expected, levels of total IgG insulin antibody were greater in diabetic patients treated with insulin than in the control group of diabetic patients on oral agents or nondiabetic controls. Insulin antibody responses of the IgG2 subclass were negligible to absent in all groups. Adult diabetic patients on insulin without complications and those with insulin associated anaphylaxis had mean values of IgG1, IgG3, and IgG4 insulin antibodies no different from those of controls. Patients with local hypersensitivity had elevated IgG1 responses. Type I diabetic patients had elevated IgG3 responses. A group of Type II diabetic patients selected for high levels of total IgG insulin antibodies had elevated levels of IgG1, IgG3, and IgG4 antibody responses. Thus, the IgG subclass response to insulin primarily involves IgG subclasses 1, 3, and 4 and varies with the type of diabetes and complications of insulin therapy.

Increasing incidence of hypoglycemic coma in children with IDDM

OBJECTIVE

To examine the incidence of hypoglycemic coma in children with insulin-dependent diabetes mellitus (IDDM) over 8 yr from 1981 to 1988 and to investigate the importance of residual beta-cell function of HbA1 levels and other variables as risk factors for hypoglycemic coma.

RESEARCH DESIGN AND METHODS

The study consisted of 155 children with IDDM aged less than 16 yr at study entry. Mean age at onset of diabetes was 7.9 yr (range 1.1-15.6 yr). We made a prospective assessment of hypoglycemic coma episodes, with a standardized questionnaire, over a total observation time of 816.6 person-yr. Three monthly clinical and laboratory examinations, which included determinations of C-peptide and HbA1 levels, were conducted. We compared children with hypoglycemic coma (cases) with children without hypoglycemic coma (controls) in a case-control analysis matched for diabetes duration. Yearly incidence of hypoglycemic coma, calculated as the number of subjects having an attack in 1 yr divided by the cumulative number of person-years for that year, was measured. Univariate and multivariate odds ratios were calculated from logistic regression.

RESULTS

Over the first 4 yr, the average yearly incidence was 4.4/100 person-yr compared with 7.4/100 person-yr during the later 4 yr (P less than 0.0001). This tendency was accompanied by intensification of insulin treatment with an increase in the mean number of daily injections and a decrease in mean HbA1 levels. In the case-control analysis, absent residual beta-cell function was the most important risk factor for hypoglycemic coma (adjusted odds ratio 7.8, 95% confidence intervals 2.0-31.2), followed by near-normal HbA1 levels (adjusted odds ratio 4.5, 95% confidence intervals 1.9-10.5).

CONCLUSIONS

In this group of children, improvement of glycemic control apparently led to an increase in the incidence of severe hypoglycemia. In children with recurrent hypoglycemic coma and undetectable C-peptide levels, it may be safer to aim for somewhat less tight glycemic control.

Proinsulin autoantibodies are more closely associated with type 1 (insulin-dependent) diabetes mellitus than insulin autoantibodies

The disease association of autoantibodies to proinsulin and insulin was compared in patients with Type 1 (insulin-dependent) diabetes mellitus and first-degree relatives. Following the recommendation of the Fourth International Workshop on the Standardization of insulin autoantibodies, autoantibodies were determined by fluid-phase radioimmunoassay using equimolar concentrations of mono-125I-A14-insulin or -proinsulin to detect insulin or proinsulin autoantibodies, respectively. A higher prevalence of proinsulin autoantibodies vs insulin autoantibodies was found in 97 patients with Type 1 diabetes prior to insulin treatment (34.0% vs 22.7%, p less than 0.05) and in 16 islet cell antibody-positive relatives (43.8% vs 31.3%, NS). There was only one serum positive for insulin and proinsulin autoantibodies in 110 islet cell antibody-negative first degree relatives (0.9%). None of 88 normal sera contained proinsulin autoantibodies or insulin autoantibodies. There was a close correlation of proinsulin autoantibody and insulin autoantibody titres in individual sera (r = 0.95, p less than 0.01) due to crossreaction of all insulin autoantibodies with proinsulin. However, some proinsulin autoantibodies did not crossreact with insulin. Background binding in normal sera was lower for proinsulin autoantibodies. We conclude that proinsulin autoantibodies have a higher association to acute Type 1 diabetes than insulin autoantibodies.

Insulin autoantibodies as determined by competitive radiobinding assay are positively correlated with impaired beta-cell function--the Ulm-Frankfurt Population Study

Out of a random population of 4208 non-diabetic pupils without a family history of Type I diabetes 44 (1.05%) individuals had islet cell antibody (ICA) levels greater or equal to 5 Juvenile Diabetes Foundation (JDF) units. 39 of these ICA-positives could be repeatedly tested for circulating insulin autoantibodies (CIAA) using a competitive radiobinding assay. The results were compared with the insulin responses in the intravenous glucose tolerance tests (IVGTT) and with HLA types. Six pupils were positive for CIAA. All of them had complement-fixing ICA, and 5 of them were HLA-DR4 positive. Three of the 6 showed a first-phase insulin response below the first percentile of normal controls. Our data indicate that in population-based studies CIAA can be considered as a high risk marker for impaired beta-cell function in non-diabetic ICA-positive individuals.

Nocturnal electroencephalogram registrations in type 1 (insulin-dependent) diabetic patients with hypoglycaemia

Eight Type 1 (insulin-dependent) diabetic patients with no diabetic complications were studied overnight for two consecutive and one subsequent night with continuous monitoring of electroencephalogram and serial hormone measurements. The aims were: 1) to evaluate the influence of spontaneous and insulin-induced hypoglycaemia on nocturnal electroencephalogram sleep-patterns and, 2) to evaluate counter-regulatory hormone responses. Spontaneous hypoglycaemia occurred on six nights (38%) with blood glucose concentrations less than 3.0 mmol/l and on four nights less than 2.0 mmol/l. All the patients experienced insulin-induced hypoglycaemia with a blood glucose nadir of 1.6 (range 1.4-1.9) mmol/l. The electroencephalogram was analysed by a new method developed for this purpose in contrast to the traditional definition of delta-, theta-, alpha- and beta-activity. The blood glucose concentration could be correlated to the rank of individual electroencephalogram-patterns during the whole night, and specific hypoglycaemic amplitude-frequency patterns could be assigned. Three of the eight patients showed electroencephalogram changes at blood glucose levels below 2.0 (1.6-2.0) mmol/l. The electroencephalogram classes representing hypoglycaemic activity had peak frequencies at 4 and 6 Hz, respectively, clearly different from the patients' delta- and theta-activity. The changes were not identical in each patient, however, they were reproducible in each patient. The changes were found equally in all regions of the brain. The three patients with electroencephalogram changes during nocturnal hypoglycaemia could only be separated from the other five patients by their impaired glucagon responses. Against this background the possibility of protection by glucagon, against neurophysiologic changes in the brain during hypoglycaemia may be considered.

Risk for developing type 1 (insulin-dependent) diabetes mellitus and the presence of islet 64K antibodies

First-degree relatives of Type 1 (insulin-dependent) diabetic patients are at increased risk for developing clinical diabetes. The presence of islet cell or insulin autoantibodies further identifies relatives at greater risk, but not all immunologic-marker-positive relatives progress to disease. Beta-cell dysfunction, however, seems to be more prevalent than clinical Type 1 diabetes, since stable subclinical pancreatic Beta-cell dysfunction may occur. Antibodies against a Mr 64,000 (64K) islet Beta-cell protein, identified as glutamic acid decarboxylase, have been reported both at and several years prior to the clinical onset of Type 1 diabetes. We measured 64K antibodies in first-degree relatives with varying degrees of Beta-cell dysfunction and risk for subsequent Type 1 diabetes to determine whether 64K antibodies improve the predictive power of islet cell antibodies and/or insulin autoantibodies. In the Seattle Family Study first-degree relatives of Type 1 diabetic patients are followed prospectively using detailed Beta-cell function tests, insulin sensitivity, quantitative evaluation of islet cell antibodies and fluid phase assay insulin autoantibodies. 64K antibodies were measured using dog islets. Relatives were selected, based on Beta-cell function to represent individuals at high (n = 6) and low (n = 30) risk for subsequent Type 1 diabetes. The 30 low-risk individuals followed-up for 78 months, had stable Beta-cell function, and six (20%) were negative for all autoantibodies, ten (33%) were positive for insulin autoantibodies, 16 (53%) were islet cell antibody positive while six (20%) were positive for 64K antibodies.(ABSTRACT TRUNCATED AT 250 WORDS)

Presence of insulin autoantibodies as regular feature of nondiabetic repertoire of immunity

With an ultrasensitive noncompetitive enzyme-linked immunosorbent assay (ELISA), we tested the hypothesis that the presence of insulin autoantibodies in nondiabetic individuals is a normal event. Plasma and peripheral blood mononuclear cells were obtained from 50 nondiabetic whites for determination of insulin autoantibodies by ELISA and radioimmunoassay (anti-insulin IgG [AI-IgG] and 125I-labeled insulin bound [%]), islet cell antibodies, anti-nuclear antibodies and rheumatoid factor, and HLA class II-type antigens (DR, DRw, and DQ). The range of 125I-insulin binding was significantly less than was seen in pretreatment sera from individuals with diabetes (from -0.4 to 0.4% vs. -0.8 to 7.7%, respectively, P = 0.001). Eighty-eight percent of these nondiabetic individuals had significant levels of AI-IgG with preferential binding to human insulin. The geometric mean of AI-IgG concentrations in individuals with significant levels was 180 pM. Binding to human insulin was seen in 88%, to pork insulin in 42%, and to beef insulin in 24% of individuals (P less than 0.001 overall; P less than 0.05 where more bound to pork than beef insulin). Binding of AI-IgG to human insulin-coated plates was substantially inhibited by preincubation with human insulin (median inhibition 57.6%) with little if any inhibition by glucagon, C-peptide, albumin, or IgG. Four individuals had highly specific human AI-IgG as shown by immunoaffinity studies. AI-IgGs were significantly higher in individuals with the HLA haplotype DR4,DRw53,DQ3 and lower in individuals with DR5,DRw52,DQ1 (P = 0.03 for both).(ABSTRACT TRUNCATED AT 250 WORDS)

HLA-associated insulin autoantibody formation in newly diagnosed type I diabetic patients

To assess a possible HLA association with anti-insulin autoantibodies (IAAs) in human insulin-dependent (type I) diabetes, 51 newly diagnosed type I diabetic patients (mean age 22 +/- 8 yr) were typed for HLA-DR and HLA-DQ and studied for IAAs before exogenous insulin therapy with a competitive radioimmunoassay (normal range less than or equal to 49 nU/ml). The level of IAAs in 16 patients exceeded our upper limit of normal, and 18 had high-titer islet cell antibodies (ICAs; greater than or equal to 40 Juvenile Diabetes Foundation U). A striking association with HLA-DR4 (DQw3) in both the prevalence and the level of IAAs was found (IAA positivity in patients with DR4/4 vs. DR4 heterozygous vs. non-DR4: 90 vs. 29%, corrected [c] P less than 0.01, vs. 5%, Pc less than 0.0001; IAA positivity in patients with DR4 vs. non-DR4: 50 vs. 5%, Pc less than 0.005; IAA level in patients with DR4/4 vs. DR4 heterozygous vs. non-DR4: 111 vs. 17 nU/ml, Pc less than 0.01, vs. 20 nU/ml, Pc less than 0.0001; IAA level in patients with DR4 vs. non-DR4: 45 vs. 20 nU/ml, Pc less than 0.01). In contrast, none of the DR3+ subjects had IAAs above normal range, except in conjunction with DR4 (DR3 vs. non-DR3: 12 vs. 42%, Pc less than 0.05). However, there was no significant relationship between DR3 and IAAs after correcting for the number of DR4 alleles. No relationship was seen between age of onset, IAA level, and HLA typing in our population, and no relationship was found between ICA positivity and HLA antigens.(ABSTRACT TRUNCATED AT 250 WORDS)

Gestational diabetes mellitus: evidence for autoimmunity against the pancreatic beta cells

Diabetes mellitus is a frequent transient or rare permanent complication of pregnancy. The role of autoimmune phenomena in this gestational form of diabetes is incompletely understood. We have examined sera from 312 pregnant women who had abnormal glucose tolerance (based on a screening examination during the second trimester) for the presence of islet cell surface antibodies or insulin autoantibodies. Fifty-eight of these women were lost to follow-up. Of the remaining subjects, 144 (57.1%) had gestational diabetes diagnosed by formal glucose tolerance testing and the others (42.9%) were normal. Sixty percent of the women with gestational diabetes eventually required insulin to control their blood glucose during pregnancy. One serum from the non-diabetic women was positive for insulin antibodies (0.9%); 8 of the sera from the patients with gestational diabetes were positive (5.6%). Subsequent analysis revealed that all nine of the women whose sera were positive for insulin autoantibodies had been treated with insulin previously. Islet cell surface antibodies were strongly correlated with gestational diabetes. Forty-five of 144 gestational diabetic sera were positive (31.3%) whereas only 9 of 108 suspect control sera (8.3%) and 7 of 60 unknown sera (11.7%) were positive. These data suggest that a high percentage of pregnant women who screen positive for glucose intolerance have serological evidence of an autoimmune response against the pancreatic islets, in spite of the state of relative immune tolerance during pregnancy. These data suggest that autoimmune phenomena may play a role in gestational diabetes and that the presence of islet cell antibodies can predict insulin-requiring gestational diabetes.

Specific association of HLA-DR4 with increased prevalence and level of insulin autoantibodies in first-degree relatives of patients with type I diabetes

First-degree relatives of patients with insulin-dependent (type I) diabetes (n = 264 from 106 families) were evaluated with HLA typing and determination of competitive insulin autoantibodies (CIAAs) and islet cell autoantibodies (ICAs). The levels of CIAAs in 30 relatives exceeded our upper limit of normal (greater than or equal to 39 nU/ml), and 30 had high-titer ICAs (greater than or equal to 40 Juvenile Diabetes Foundation units [JDF U]). Eleven of the HLA-typed relatives developed diabetes during follow-up. Twenty-three percent (28 of 123) of the relatives with at least one HLA-DR4 allele were CIAA+ (CIAA greater than or equal to 39 nU/ml) versus 4% (6 of 141) among DR4- relatives (P less than 0.0001). Twenty-one of 22 of the highest CIAA values were all in the DR4+ group (DR4+ vs. DR4-, P = 0.003, Wilcoxon's rank-sum test). HLA-DR3 did not correlate with the level of CIAAs, and neither DR3 nor DR4 correlated with titer of ICAs measured in JDF U. We conclude that, in first-degree relatives of patients with type I diabetes, there is a striking association with HLA-DR4 in both the prevalence of relatives exceeding the normal CIAA range and in the level of CIAAs. These data suggest that a gene on HLA-DR4 haplotypes contributes to the level of anti-insulin autoimmunity, and we hypothesize that DR4-associated diabetes susceptibility, distinct from DR3-associated susceptibility, may be secondary to this influence.

T-lymphocyte changes linked to autoantibodies. Association of insulin autoantibodies with CD4+CD45R+ lymphocyte subpopulation in prediabetic subjects

The onset of insulin-dependent (type I) diabetes is predictable before hyperglycemia by the presence of islet cell autoantibodies (ICAs) and competitive insulin autoantibodies (CIAAs). CIAA+ICA+ first-degree relatives of individuals with type I diabetes have increased numbers of CD4 cells bearing the CD45R antigen and reciprocal depressions of the CD4 cells bearing the CD29 determinant. In addition, depressed CD4/CD8 ratios are present. In this study, we investigated the correlation between autoantibody levels and T-lymphocyte changes in the prediabetic state. The data demonstrate a clear linear relationship between rising CIAA levels, a marker of disease rate, and rising elevations in the CD4+CD45R+/CD4+CD29+ ratio in 37 CIAA+ICA+ and CIAA+ICA- relatives (r = 0.93). In marked contrast, the degree of CD4/CD8 depression found in individuals with prediabetes or long-term diabetes failed to correlate with either CIAA (r = 0.32) or ICA (r = 0.29) levels. The investigation of T-lymphocyte changes in siblings of individuals with type I diabetes with different stable autoantibody patterns (CIAAs and/or ICAs), and thus varying risks for diabetes, revealed differences in the prediabetic groups. Fifteen CIAA+ICA- relatives with high CIAA levels (greater than 80 nU/ml) had high CD4+CD45R+/CD4+CD29+ ratios (P = 0.03) and depressed CD4/CD8 ratios (P = 0.008). In contrast, CIAA+ICA- relatives with low CIAA levels (39-80 nU/ml), and thus low risk of diabetes, had no alteration in their CD4/CD8 ratio (P = 0.75) or CD4+CD45R+/CD4+CD29+ ratio (P = 0.33). Nineteen CIAA-ICA+ siblings with a predicted intermediate risk for diabetes showed heterogeneity in the presence of T-lymphocyte abnormalities.(ABSTRACT TRUNCATED AT 250 WORDS)

A twelve month double-blind randomized study of the efficacy and immunogenicity of human and porcine insulins in non-insulin-dependent diabetics

This study aims to compare the effectiveness and immunogenicity of semisynthetic human insulin (NOVO) and highly purified (monocomponent) porcine insulin over a 12 month period in 16 non-insulin-dependent diabetic subjects not previously exposed to insulin. Sixteen patients were randomly allocated for treatment with either human (n = 9) or monocomponent porcine (n = 7) insulin in a double-blind trial. Both groups were identical with respect to age, sex and measures of metabolic control. Significant insulin antibody was detected in seven of the nine patients (78%) 3 months after the commencement of human insulin therapy whereas it was detected in all patients (100%) in the group treated with monocomponent porcine insulin as early as 2 months after insulin therapy. Besides the delayed rise of insulin antibodies during the first 3 months of human insulin therapy, it tended to have a lower mean insulin antibody titer, though statistically insignificant, at the end of the study. No adverse reaction to either type of insulins was noted. It is concluded that both semisynthetic human and monocomponent porcine insulin were safe and effective. Although human insulin showed a slightly lower immunogenicity than monocomponent porcine insulin of the same formulation and purities, it was not non-immunogenic. Hence, there is no reason to treat all insulin-requiring diabetic subjects with human insulin except those who have developed insulin allergy or those at risk or with a history of allergy.

Insulin prevents adoptive cell transfer of diabetes in the autoimmune non-obese diabetic mouse

Early intensive insulin treatment is thought to improve subsequent Beta-cell function in Type 1 (insulin-dependent) diabetic patients. Prophylactic insulin administration also reduced diabetes incidence in diabetes-prone animals. To study the mechanisms by which these effects occur, we tested the ability of insulin therapy in the model of non-obese-diabetic mice, to prevent the penetration of committed T cells into the islets and subsequent Beta-cell destruction. Sublethally irradiated non-obese-diabetic males of 8 weeks of age were adoptively transferred with splenocytes from diabetic donors and treated with the maximum tolerable dosage of fast-acting insulin (0.5 U, twice daily) until 30 days after cell transfer. Diabetes incidence was compared to control animals injected with the same concentration of insulin diluent. After one month of treatment, the cumulative diabetes frequency was significantly less within the insulin-treated group (4 of 15, 26.6%) than in the control group (15 of 18, 83.3%; p less than 0.01). Pancreatic histological analysis of insulin-treated animals revealed a lower severity of insulitis and Beta-cell necrosis and a higher percentage of normal islets (46.6 +/- 10% vs 2.3 +/- 2%, p less than 0.01), including five (33%) mice with no lesions. Immunoperoxydase staining of pancreatic sections indicated similar insulin and ganglioside staining of Beta cells from insulin-treated mice and control animals. Insulin-treated mice had comparable pancreatic insulin content to normal mice. Flow cytometry analysis of spleen cell populations indicated that insulin increased the number of Thy1,2+ and Lyt-2+ T cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Continuous subcutaneous insulin infusion (CSII) and insulin antibodies in rabbits

Using prepubertal male New Zealand White rabbits, continuous subcutaneous insulin infusion (CSII), delivered by either an external or an implantable infusion device, resulted in significantly higher insulin antibody (I-Ab) production than bolus injection (BII). We tested the influence during CSII of (1) the insulin species, (2) the insulin diluent, (3) the materials of which the infusion devices were made and (4) the incubation of insulin in a syringe on the backs of rabbits ('sham-infusion'), with the following results: (1) beef and sulphated beef insulins produced high levels of I-Abs, while porcine and human insulins produced moderate levels; (2) with all insulins used, 0.9% NaCl and 0.9% NaCl with 24-26 mmol NaHCO3 added, produced high levels of I-Ab. A buffer containing 0.7% NaCl, 0.136% sodium acetate trihydrate and 0.1% methyl-p-hydroxybenzoate and a buffer containing 16 mg/ml glycerol and 2 mg/ml phenol, produced highly significantly lower I-Abs (P less than 0.001); (3) insulin glass syringes produced much lower I-Ab levels than in standard polypropylene syringes and (4) polypropylene syringes in a 'sham-infusion' technique, resulted in intermediate levels of insulin antibodies [(P less than 0.02) vs CSII; (P less than 0.005) vs BII]. Our data suggest that insulin immunogenicity is influenced by all four factors tested. We suggest that benefits of CSII therapy may be attenuated unless a best possible control of these factors is achieved.

[Comparative multicenter study of bovine insulin with forms more purified of swine and human insulin in the treatment of type 1 diabetes mellitus]

A multicentric double-blind randomized study was organized to investigate the relationship between insulin antibodies and metabolic control in type I diabetics who changed from bovine insulin to monopic porcine and monocomponent human insulin. Twenty eight type I diabetic patients treated with bovine insulin (proinsulin less than 3,000 ppm) were selected. In a 6 month study, 9 patients maintained bovine insulin, 9 changed to monopic porcine insulin (proinsulin less than 10 ppm) and 10 to human insulin (proinsulin less than 1 ppm). The insulin were a gift from Biobras laboratory. The insulin antibodies (IA) were measured by an ELISA method and the metabolic control assessed by fasting blood sugar (FBS), 24 hour glucosuria and glycated protein (GP) measured by affinity chromatography method. After switching insulin therapy, IA decreased with porcine and human insulin, but increased with bovine insulin. Concerning metabolic control, only an increase of FBS with human insulin was found. In the beginning of study, there was negative correlation between IA and 24h glucosuria (rs = -0.509; p = 0.006). In conclusion, there was no improvement of metabolic control, in spite of a decrease of IA in type I diabetics treated during 6 months with less immunogenic insulin preparations.

[Insulin antibodies: methodology and clinical implications]

Anti-insulin antibodies detection is based on the demonstration of a specific and saturable binding to insulin either radiolabelled with 125 I (Radiobinding assay or RBA) or coated on a solid phase (Enzyme linked immunosorbent assay or EIA). The 2 assays are remarkably different by their sensitivity to the affinity of the antigen antibody reaction. In addition, RBA may be biased by the presence of the iodine atom on the radioiodinated insulin whereas, at least on theoretical grounds. EIA could be biased because of denaturation or non availability of some epitopes when insulin is coated. Anti-insulin antibodies may be induced by insulin therapy. When they "spontaneously" appear, they are called autoantibodies. Insulin autoantibodies may be detected in the normal population, in type 1 diabetic patients before any administration of exogenous insulin and in patients suffering from the autoimmune hypoglycemic syndrome. In some patients, this syndrome may be associated with administration of a thiol containing drug. In some cases, insulin antibodies may appear several years after a transient insulin therapy, possibly as a consequence of a disturbance of the immunologic memory. The properties of antibodies and autoantibodies (concentration, affinity, number and nature of epitopes, heavy and light chain composition and ability to form aggregates) are relatively characteristic of the disease with which they are associated and determine their potential effects on insulin bioavailability and plasma glucose homeostasis.

Autoantibodies and genetic factors associated with the development of type 1 (insulin-dependent) diabetes mellitus in first degree relatives of diabetic patients

Factors associated with diabetes onset were analysed for their predictive value in 708 first-degree relatives of Type 1 (insulin-dependent) diabetic patients including 374 parents and 308 siblings of Type 1 diabetic patients. Relatives were prospectively followed for 2,304 subject years with blood samples for specific autoantibody evaluation. Islet cell cytoplasmic autoantibody titres were quantified in Juvenile Diabetes Foundation units with a threshold of positivity of 5 units. Insulin autoantibodies were determined using Tyr-A14 iodinated human insulin. HLA typing was performed in 92% of the relatives. During the time of study, 17 of 646 (2.6%) relatives showed islet cell antibodies. During follow-up, eight relatives developed diabetes, including six with high islet cell antibody titre. Taking titres above 20 units increased the positive predictive value from 35% to 75% whereas the presence of insulin autoantibodies did not increase the positive predictive value for the disease. Analysis of metabolic profiles months before the onset of diabetes by either oral or intravenous glucose loads, indicated a considerable level of heterogeneity with relatives with a high islet cell antibody titre who rapidly developed insulin-dependent diabetes, whereas other remained insulin-independent during the same observation period despite comparable titres. This study clearly indicates that initial islet cell antibody titre is not sufficient to predict individual outcome. Follow-up samples are clearly needed to monitor progression of the disease. Few relatives with persistent immunologic positivity progress to clinical Type 1 diabetes, suggesting that non-progressive and sub-clinical Beta-cell dysfunction is common.(ABSTRACT TRUNCATED AT 250 WORDS)

Solid-phase enzyme-immunoassay of anti-insulin antibodies: effect of labeling site in insulin and of labeled number of horseradish peroxidase on the assay sensitivity

Three horseradish peroxidase (HRP)-labeled porcine insulins which have definite labeling site(s) were compared regarding sensitivity in a solid-phase enzyme-immunoassay (EIA) of anti-insulin antibodies. The standard curves obtained with LysB29-HRP-insulin and GlyA1-HRP-insulin were steeper than that with GlyA1,LysB29-diHRP-insulin for both polyclonal and monoclonal antibodies. Thus, the mono-HRP-labeled insulins can afford higher sensitivities in the EIA. The importance of the HRP-labeling site in insulin and of the number of labeled HRP was first demonstrated by using HRP-labeled insulins having definite labeling site(s).

Factors predicting residual beta-cell function in the first year after diagnosis of childhood type 1 diabetes

Twenty-five children aged 2-14 years (mean age 8.39 +/- 0.78 years) were studied prospectively during the first year after the diagnosis of type 1 diabetes. Of their clinical and metabolic features at diagnosis, only age showed a significant independent relationship with endogenous C-peptide production during the first year. Age was correlated with higher values for basal and stimulated plasma C-peptide at 7-14 days after diagnosis, at 6 months and at 12 months. At diagnosis, age was also associated with a higher value for HbA1c and a lower prevalence of insulin antibodies. C-peptide production peaked at 3 months and thereafter declined. Mean HbA1c and insulin requirement were both minimal at 6 months. At diagnosis, there were significant inverse relationships between basal C-peptide production and both insulin dose and HbA1c and between stimulated C-peptide production and HbA1c. Basal and stimulated C-peptide production were inversely related to insulin dose at 6 and 12 months. Stimulated C-peptide was higher at 12 months in children retaining islet cell antibodies. These findings confirm the importance of age as a predictor of residual beta-cell function in type 1 diabetes and indicate that older children present clinically following a slower course of beta cell destruction.

Relating homology between the Epstein-Barr virus BOLF1 molecule and HLA-DQw8 beta chain to recent onset type 1 (insulin-dependent) diabetes mellitus

A role for the Epstein-Barr virus in initiating Type 1 (insulin-dependent) diabetes mellitus has been proposed since Epstein-Barr virus BOLF1 (497-513) AVTPL RIFIVPPAAEY has an 11 amino acid identity with HLA-DQw8 beta (49-60) AVTPL GPPAAEY. Rabbit antisera to the BOLF1 (496-515) peptide crossreacted with the homologous DQw8 beta (44-63) peptide but not with the related DQw7 beta (44-63) peptide, which differed from the DQw8 peptide only in an ALA to ASP substitution in position 57. Antisera to DQw8 beta (49-60) reacted with the DQw8 beta (44-63) peptide and BOLF1 (496-515), but not with DQw7 beta (44-63). The antiserum to the BOLF1 peptide bound to denatured class II major histocompatibility complex beta chains from Epstein-Barr virus-transformed DQw8-positive lymphocytes in an immunoblotting analysis. Epstein-Barr virus antibodies were detected at equal frequencies and similar titres in sera of 30 patients with Type 1 diabetes (16 of 30; 63%) and in sera of 20 non-diabetic control subjects (13 of 20; 65%). Sera from diabetic patients did not bind to DQw8 beta (44-63) or BOLF1 (496-515) peptides. From these data we conclude that there is no simple relationship between serological evidence of Epstein-Barr virus infection and crossreactions between homologous Epstein-Barr virus and class II major histocompatibility complex peptides.