Novel Flow Cytometric Immunoassay for Detection of Proinsulin Autoantibodies in Diabetes Mellitus Employing a Recombinant Autoantigen Expressed in E. coli.

Introduction: Insulin and proinsulin autoantibodies (IAA/PAA) are usually the first markers to appear in patients with type 1 Diabetes Mellitus (T1DM) and their prevalence ranges from 10 to 60% in the child-adolescent population. The reference method for IAA/PAA detection is the Radioligand Binding Assay (RBA), a highly specific and sensitive technique, but expensive and polluting. The aim of this work was to develop a novel flow cytometric microsphere-based immunoassay (FloCMIA) for PAA detection, employing recombinant human proinsulin (PI), as an alternative method to RBA, less expensive and harmful to the environment. Materials and Methods: Human PI was expressed as Thioredoxin fusion protein (TrxPI) in E. coli and a fraction was biotinylated. A double paratope model was used in which samples were incubated with TrxPI-biotin and microspheres adsorbed with TrxPI. The immune complexes were revealed using Streptavidin-Phycoerythrin. The geometric mean of the signals was analyzed, and the results were expressed as Standard Deviation scores (SDs). Sera from 100 normal human control and from 111 type 1 diabetic patients were evaluated by FloCMIA. To correlate the novel assay with RBA, 51 diabetic patients were selected, spanning a wide range of PAA reactivity by RBA. Results: The study of ROC curves allowed choosing a cut-off value of 3.0 SDs and the AUC was 0.705, indicating that FloCMIA has fair ability to distinguish between samples from each group. A prevalence of 50% for PAA was obtained in the population of diabetic patients studied. The specificity was 96% and the analytical sensitivity (percentage of patients RBA positive, also positive by FloCMIA) was 69%. There was a substantial agreement between methods (kappa statistic=0.700). Conclusions: A novel immunoassay based on flow cytometry that uses easy-to produce recombinant PI was developed. This assay constitutes an innovative and cost-effective alternative to RBA for the determination of PAA in patients' sera. The method developed here, presents good performance and a wide dynamic range together with a small required sample volume. Furthermore, these results make it possible to develop multiplex immunoassays that allow the combined detection of autoantibodies present in T1DM and other related autoimmune diseases.

Islet-Derived CD4 T Cells Targeting Proinsulin in Human Autoimmune Diabetes.

Type 1 diabetes results from chronic autoimmune destruction of insulin-producing ß-cells within pancreatic islets. Although insulin is a critical self-antigen in animal models of autoimmune diabetes, due to extremely limited access to pancreas samples, little is known about human antigenic targets for islet-infiltrating T cells. Here we show that proinsulin peptides are targeted by islet-infiltrating T cells from patients with type 1 diabetes. We identified hundreds of T cells from inflamed pancreatic islets of three young organ donors with type 1 diabetes with a short disease duration with high-risk HLA genes using a direct T-cell receptor (TCR) sequencing approach without long-term cell culture. Among 85 selected CD4 TCRs tested for reactivity to preproinsulin peptides presented by diabetes-susceptible HLA-DQ and HLA-DR molecules, one T cell recognized C-peptide amino acids 19-35, and two clones from separate donors responded to insulin B-chain amino acids 9-23 (B:9-23), which are known to be a critical self-antigen-driving disease progress in animal models of autoimmune diabetes. These B:9-23-specific T cells from islets responded to whole proinsulin and islets, whereas previously identified B:9-23 responsive clones from peripheral blood did not, highlighting the importance of proinsulin-specific T cells in the islet microenvironment.

Surface plasmon resonance reveals a different pattern of proinsulin autoantibodies concentration and affinity in diabetic patients.

Type 1 diabetes mellitus (DM) is characterized by autoimmune aggression against pancreatic beta cells resulting in absolute deficiency of insulin secretion. The first detectable sign of emerging autoimmunity during the preclinical asymptomatic period is the appearance of diabetes-related autoantibodies. In children at risk for type 1 DM, high-affinity Insulin autoantibodies reactive to proinsulin, are associated with diabetes risk. Autoantibodies are usually measured by radioligand binding assay (RBA) that provides quasi-quantitative values reflecting potency (product between concentration and affinity) of specific autoantibodies. Aiming to improve the characterization of the specific humoral immune response, we selected surface plasmon resonance (SPR) as an alternative method to measure proinsulin autoantibodies (PAA). This novel technology has allowed real time detection of antibodies interaction and kinetic analysis. Herein, we have employed SPR to characterize the PAA present in sera from 28 childhood-onset (mean age 8.31±4.20) and 23 adult-onset diabetic patients (≥65 years old, BMI<30) in terms of concentration and affinity. When evaluating comparatively samples from both groups, childhood-onset diabetic patients presented lower PAA concentrations and higher affinities (median 67.12×10(-9) M and 3.50×10(7) M(-1), respectively) than the adults (median 167.4×10(-9) M and 0.84×10(7) M(-1), respectively). These results are consistent with those from the reference method RBA (Standard Deviation score median 9.49 for childhood-onset group and 5.04 for adult-onset group) where the binding can be directly related to the intrinsic affinity of the antibody, suggesting that there is a different etiopathogenic pathway between both types of clinical presentation of the disease. This technology has shown to be a useful tool for the characterization of PAAs parameters as an alternative to radioimmunoassay, with high versatility and reproducibility associated to low occupational and environmental risk. However, this technology is not eligible for routine marker screening, but this is a powerful technique for a fine description of the thermodynamic parameters of antigen-antibody interaction.

Single-tube test for autoantibodies to glutamic acid decarboxylase and proinsulin as first-line screening for autoimmunity in adult-onset diabetic patients.

The aim of this work was develop a new combined radioligand-binding assay (RBA-combi) for the rapid and simultaneous determination of two autoimmunity markers, GADA and PAA, known to be differentially distributed in young and in some adult diabetic patients. The methodology was applied to sera from 85 young type 1 and 98 adult-onset diabetic patients with different marker profiles and insulin requirements, and to 53 normal control sera. Among type 1 diabetes sera used as autoimmunity controls, 100% of those with at least one positive marker by single methods and 17.7% of those with double negative markers were positive by RBA-combi (RBA-combi+). Among sera from adult-onset diabetes, 100% of those PAA+ (GADA+ or GADA-), 92.3% of GADA+/PAA-, and 1.3% of GADA-/PAA- were RBA-combi+. In conclusion, the new RBA-combi allowed the simultaneous detection of GADA and PAA markers with acceptable performance. Moreover, 16 out of 18 (88.9%) of adult patients RBA-combi+ evolved to insulin requirement, suggesting that this test is a valuable tool for assessing autoimmune processes associated to future impairment of insulin secretion.

A radioligand-binding assay for detecting antibodies specific for proinsulin and insulin using 35S-proinsulin.

A new radioligand-binding assay (RBA) is described for the detection of insulin/proinsulin-specific antibodies using 35S-labeled proinsulin produced by a cell-free reticulocyte extract. Direct use of the crude expression product in the RBA was not feasible because the protein failed to fold properly (or had incorrectly paired disulphide bridges) and purification was hindered by interfering by-products. A refolding protocol and a chromatographic procedure were devised that readily allowed production of purified and immunochemically competent 35S-labeled proinsulin. The new RBA was compared with the reference test, in which the tracer was standard 125I-insulin. The analysis included sera from 41 diabetic patients and 25 healthy controls. Twenty-six (63.4%) and 29 (70.7%) patients scored positive by RBA using 35S-PI and 125I-insulin, respectively. The methods showed a satisfactory correlation with r(2)=0.77 and a slope not significantly different from unity (m=1.16+/-0.10; 95% confidence interval). Since the nuclide used in the assay is 35S, the procedure is compatible with standard assays for GADA and IA-2A, and thus may permit combined assays for the major early markers of autoimmune diabetes.

A specific and highly sensitive time-resolved fluoroimmunoassay for human proinsulin.

We describe a time-resolved fluoroimmunoassay specific for human proinsulin using a combination of two high-affinity monoclonal antibodies, one against insulin and the other specific for intact proinsulin and for split 65-66 and des 64-65 proinsulin forms. The assay employs only 200 microl of serum, with a detection limit of 0.1 pmol/l. The intra-assay variation coefficient was less than 3% between 3 and 1000 pmol/l. There was 0% cross-reaction with insulin, C-peptide, split 32-33 and des 31-32 proinsulin. Serum concentration of proinsulin was analyzed in 50 subjects during an oral glucose tolerance test (10 non-obese control, 10 obese controls, 10 subjects with impaired glucose tolerance, 10 patients with type II diabetes mellitus (DM) and fasting blood glucose (FBG) < 140 mg/dl, and 10 patients with type II DM and FBG > 150 mg/dl). Mean fasting serum proinsulin levels measured by this assay in non-obese controls (0.84 - 0.90 pmol/l; 0.1-2.4 pmol/l) were lower than the results reported by other investigators. There was an increase of proinsulin related to obesity and increased glucose levels, suggesting that proinsulin levels increase with insulin resistance.

A specific and highly sensitive time-resolved fluoroimmunoassay for human proinsulin

We describe a time-resolved fluoroimmunoassay specific for human proinsulin using a combination of two high-affinity monoclonal antibodies, one against insulin and the other specific for intact proinsulin and for split 65-66 and des 64-65 proinsulin forms. The assay employs only 200 micro liters of serum, with a detection limit of 0.1 pmol/l. The intra-assay variation coefficient was less than 3 percent between 3 and 1000 pmol/l. There was 0 percent cross-reaction with insulin, C-peptide, split 32-33 and des 31-32 proinsulin. Serum concentration of proinsulin was analyzed in 50 subjects during an oral glucose tolerance test (10 non-obese controls, 10 obese controls, 10 subjects with impaired glucose tolerance, 10 patients with type II diabetes meIlitus (DM) and fasting blood glucose (FBG) <140 mg/dl, and 10 patients with type II DM and FBG >150 mg/dl). Mean fasting serum proinsulin levels measured by this assay in non-obese controls (0.84 +/-0.90 pmol/l; 0.1-2.4 pmol/l) were lower than the results reported by her investigators. There was an increase of proinsulin related to obesity and increased glucose levels, suggesting that proinsulin levels increase with insulin resistance.

Associaçäo entre auto-anticorpos anti-pro-insulina e anti-insulina com diabetes mellitus do tipo I / The association between insulin and proinsulin autoantibodies with tipe I Diabetes Mellitus

O presente trabalho compara a incidência de auto-anticorpos anti-insulina (IAA) e anti-pró-insulina (PAA) em diabéticos do tipo I de início recente e em seus parentes de primeiro grau. Foram estudados 33 indivíduos normais (grupo I), 16 diabéticos do tipo I de início recente (grupo II) e 141 parentes em primeiro grau de diabéticos do tipo I (grupo III). Os IAA e PAA foram determinados pelo método de radioensaio, sendo considerados anormais níveis de IAA acima de 0,584 pmol/L e de PAA acima de 0,441 pmol/L. Näo foram observadas diferenças significantes quanto a idade e sexo entre os 3 grupos. Nos indivíduos normais os níveis de PAA foram significantemente menores do que os de IAA. Entre os diabéticos de início recente foi encontrada uma incidência de IAA de 37,5 por cento e de PAA de 25,0 por cento, näo ocorrendo, entretanto, diferenças significantes entre os níveis destes dois anticorpos. Entre os parentes em primeiro grau a incidência de IAA foi de 3,5 por cento e de PAA de 7,8 por cento, näo ocorrendo também diferenças entre os dois testes. Houve uma correlaçäo significante entre os níveis dos IAA e dos PAA no grupo de diabéticos de início recente (r=0,64; p < 0,05). Os IAA e PAA parecem estar dirigidos contra o mesmo determinante antigênico e, portanto, têm o mesmo valor preditivo para o DM tipo I

Insulin and proinsulin in normal and chemical diabetic children.

Plasma levels of total immunoreactive insulin and immunoreactive proinsulin were studied in 10 normal children and 15 children with chemical diabetes ranging in age from 5 to 13 years. Eleven of the children with chemical diabetes demonstrated significantly elevated TIRI during fasting and following glucose administration. There was a delay in the increment of plasma TIRI in four children with chemical diabetes, but otherwise their TIRI levels were normal. In these four children IRP was not significantly different from normal; however, in the remaining children with chemical diabetes, those with elevated TIRI, the IRP was elevated following glucose administration. Although the IRP was significantly elevated in the hyperinsulinemic group, the TIRI was also increased to such an extent that the glucose intolerance demonstrated in these patients could not be attributed to the elevated IRP.