Search

Diabetes is associated with a high incidence of cardiovascular disease, which is the major cause of morbidity and mortality in this disease. There is considerable interest in defining factors responsible for the accelerated development of atherosclerosis in diabetes. There is no evidence to suggest that the inflammatory process in diabetic patients is different from those in non-diabetic individuals. The main difference may lie on factors able to trigger the inflammatory process. Diabetes is a major predisposing factor for the generation of modified proteins though advanced glycation and oxidation, two intimately interrelated processes. Advanced glycation end-products modified low density lipoprotein (AGE-LDL) and other AGE-modified proteins as well as oxidized LDL (oxLDL) are able to interact with a variety of cells and induce cell dysfunction and the release of pro-inflammatory mediators. But AGE-LDL and oxLDL are also immunogenic. Activated T lymphocytes reacting with peptides derived from oxidized LDL have been detected in atheromatous lesions. Their pro-inflammatory potential is directly linked to the release of interferon-gamma and other cytokines able to activate macrophages, smooth muscle cells, and endothelial cells. On the other hand, antibodies to oxidized and AGE-modified LDL have been isolated from diabetic patients and shown to belong predominantly to the IgG isotype, subclasses 1 and 3, which have well-defined proinflammatory properties. These autoantibodies to modified lipoproteins have sufficient affinity to form stable antigen-antibody complexes, which have been also shown to have pro-atheromatous and pro-inflammatory properties.

Modified lipoproteins, particularly different forms of oxidized LDL (ox-LDL), have been reported to elicit humoral immune responses both in experimental animals and humans. In diabetes, glycation and oxidation processes coexist and lead to the formation of glycoxidation products. Ox-LDL has been demonstrated in atheromatous lesions, anti-ox-LDL antibodies have been detected in circulation and in atheromatous plaques, and immune complexes (ICs) formed with LDL and anti-LDL (LDL-IC) have been isolated from the serum of patients with manifestations of atherosclerosis. In addition, in vitro formed LDL-ICs and ICs isolated from patients have been demonstrated to cause intracellular accumulation of cholesteryl esters (CEs) in human macrophages and fibroblasts. The accumulation of CEs in macrophages exposed to LDL-ICs is unique to this type of IC and is associated with paradoxical overexpression of LDL receptor and with increased synthesis and release of interleukin 1 beta and tumor necrosis factor (TNF) alpha. The overexpression of LDL receptors is higher in LDL-IC-stimulated macrophages that release markedly high amounts of TNF-alpha than in macrophages that release low amounts of TNF-alpha into the medium. The release of cytokines in the subendothelial space may have a significant role in promoting the interaction of endothelial cells with mononuclear cells, causing endothelial cell damage directly or indirectly, and also in inducing smooth muscle cell proliferation. Thus, in view of the above data, it can be concluded that humoral autoimmunity may play a significant role in the pathogenesis of atherosclerosis in diabetes.

We have studied soluble immune complexes (IC) in the sera of both insulin-dependent and non-insulin-dependent diabetics by a variety of non-specific techniques and also by a method that detects specifically insulin-anti-insulin IC. Our screening studies, detailed in the first section of this work, showed that insulin-anti-insulin IC appear not to be the only type of IC present in diabetics. Non-specific screening tests gave practically identical percentages of positive results in insulin-dependent and non-insulin-dependent diabetics. However, the agreement between different screening tests was poor. We propose the use of 'IC scores' (numerical expressions of the general trend of several screening tests performed with one given serum sample) for the statistical analysis of correlations between the presence of soluble IC and clinical evidence of diabetic microangiopathy. As expected, the use of 'scores' minimized false positive or negatives and considerably enhanced statistical correlations between levels of IC and proteinuria, nephropathy, retinopathy, peripheral neuropathy, and peripheral vasculopathy. The second section of this report describes our isolation studies, which provided definitive proof for the existence of soluble insulin-anti-insulin IC and allowed us to carry out the first successful studies of the biological properties of soluble IC purified from the sera of diabetic patients, as detailed in the third section of this report. Such IC-induced platelet aggregation and activation which in vivo could lead to the development of microvascular lesions could explain, at least in part, the abnormalities in platelet function seen in diabetics. Although the precise mechanisms by which soluble IC could induce pathological damage in diabetics have not been totally clarified, we have obtained sufficient evidence to prove that antigen-antibody complexes exist in diabetics, are associated with higher frequencies of complications, and have the capacity to interact with cells in a potentially pathogenic fashion.

A simplified approach for estimating free and total anti-insulin antibodies and soluble insulin-anti-insulin immune complexes in serum has been developed and evaluated. For determination of free anti-insulin antibodies, a binding ratio for 125I-labeled insulin at fivefold serum dilution is calculated; the binding ratios obtained are reproducible (run-to-run coefficient of variation, 7.3%) and correlate well with titration of anti-insulin antibodies by other techniques (correlation coefficient, 0.9327). Total anti-insulin antibody concentrations are determined by calculating the [125I]insulin binding ratio for a sample previously acidified, adsorbed with acid dextran-coated charcoal, and neutralized. The difference in binding ratios between two aliquots of the same serum, one diluted 10-fold without manipulation and the second studied at an identical dilution after acidification and removal of free insulin, is taken as an index of the presence of soluble insulin-anti-insulin immune complexes. Because the tests can be performed at a single dilution, both time and materials are conserved without apparent loss of discrimination between sera with high antibody titers and high concentrations of insulin-anti-insulin immune complexes and sera with negative titers or low concentrations.

Human platelets are known to have Fc receptors that are able to recognize soluble immune complexes and to respond to that stimulation by aggregating and releasing soluble factors. In diabetes, enhanced platelet aggregation has been proposed as one of the factors contributing to the development of microangiopathy. Soluble immune complexes isolated from seven diabetic patients were found to enhance ADP-induced platelet aggregation and release of ATP. This enhancement was proven not to be an artifact due to the isolation protocol, by comparison of purified immune complexes with nonspecific protein purified from normal sera by identical or slightly modified isolation protocols. Soluble immune complex appear to be the first well-characterized platelet aggregating factors form the sera of diabetic patients. The natural of the antigen involved in their formation does not appear relevant, since very similar results were obtained whether soluble immune complexes were purified from patients with insulin-anti-insulin complexes in their serum, or from those without such complexed but with positive results in nonspecific screening techniques.

A series of 88 diabetic patients were studied for the presence of soluble immune complexes, proteinuria, microangiopathy, and diabetic complications. Results of the five different assays for immune complexes were analyzed individually, and four combinations of the individual results (i.e. four different immune complex "scores") were also analyzed. The only assay which consistently discriminated between the different patient groups was the PEG-IgG test, in which a ratio between the amount of IgG precipitated with 3% PEG 6000 and the serum concentration of IgG is determined. In contrast, all four of the immune complex "scores" detected significant differences between patients with and without the clinical or biochemical parameter in question. One combination, designated as the "weighted and corrected IC score", gave a particularly high probability of detecting differences between groups. These results indicate that proper compilation of the results of a battery of immune complex screening assays can provide definite advantages over the results of individual tests for the investigation of correlations between the presence of soluble immune complexes and the course and pathology of various diseases.

A comparative study of four nonspecific screening techniques (direct nephelometry, PEG-C4, PEG-IgG, and radiolabeled Clq binding) for immune complexes (IC) and of a technique specific for the detection of insulin-anti-insulin IC was undertaken in four groups of patients with diagnosis of infectious endocarditis, systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), and diabetes. The highest frequency of positive results was given by the PEG-IgG test in RA, the Clq-binding test in SLE, the insulin-anti-insulin IC screening test in diabetes, and the PEG-IgG and Clq-binding tests in infectious endocarditis. Of the four nonspecific tests, the PEG-C4 assay appeared to be the least discriminative, since it failed to show significant differences between any group of patients and the group of controls. Direct nephelometry, PEG-IgG, and radiolabeled Clq binding gave consistently higher results in RA than in other diseases, and in this disease the rates of agreement between these tests were highly significant. Significant agreements between the rates of positivity of Clq binding and PEG-IgG tests were seen in all groups of patients studied. Spearman's analysis of rank showed the best correlations among tests based on similar principles (ie, PEG precipitation), and also a strong correlation between Clq binding and the PEG-IgG test in RA. The PEG-IgG test appears to be a reliable IC screening test for general use with the advantage of not involving radioisotopes. In regard to antigen-specific tests, although their specificity and sensitivity may be high, their results may show no correlation with nonspecific screening tests nor with the presence or absence of clinical or laboratory abnormalities suggestive of IC deposition, as exemplified by the insulin-anti-insulin IC screening test in diabetic patients.

During a prospective screening for proteinuria in diabetic patients, isolated Bence-Jones proteinuria was detected in 2 cases. The first patient, a 52-year-old black female, was seen for evaluation of a slow but progressive weight loss which was attributed to poor adjustment of insulin therapy. The patient gained weight after an increase of the daily insulin administration. She had plasmocytosis in a bone marrow aspirate, but no other evidence of myelomatosis. The second patient, a 59-year-old black male who was seen for routine evaluation of his diabetes, had no clinical or laboratory evidence of myelomatosis. Although precise definition of these cases as "benign" or "idiopathic" Bence-Jones proteinuria is impossible without prolonged follow-up, at the time of presentation they appeared to fit this classification. This observation is one further example that isolated Bence-Jones proteinuria may be seen without any evidence of malignant B-cell dyscrasia.

Plasma fron some diabetic patients, particularly those with advanced retinopathy and nephropathy, will potentiate ADP-induced platelet aggregation. Partial purification of plasma from a diabetic patient with nephropathy has yielded a fraction with this activity. A linea dose response curve relating plasma factor (6.25-50 ng protein) and platelet aggregation or ATP release at 4 minutes after adding ADP has been found. The effect is blocked by exposure of platelets to aspirin, prostaglandin or eicosapentaenoic acid. Soluble immune complexes are found in many diabetics and are known to be platelet-active. We have isolated immune complexes from sera of six diabetics and have shown enhancement of the second phase of platelet aggregation and of ATP release by these immune complexes from platelets sensitized to ADP. We conclude that plasma proteins from diabetics may accentuate ADP-induced platelet aggregation and ATP release, possibly by acting through prostaglandin pathways. Immune complexes appear to be one group of plasma proteins with such platelet active behavior. Further studies are indicated to characterize these plasma factors and to assess their in vivo significance regarding platelet function and vascular disease in diabetes mellitus.