Current classification divides primary diabetes into two major categories: type I or insulin dependent diabetes mellitus (IDDM) (previously termed juvenile onset diabetes), characterized by absolute insulin deficiency and thus a dependence on insulin therapy for the preservation of life; and type 2 or non-insulin dependent diabetes mellitus (NIDDM) (previously termed maturity onset diabetes) in which insulin therapy is not generally required for the maintenance of life but may be necessary for the control of symptoms or for the correction of disordered metabolism.
Insulin dependent diabetes mellitus is characterized by a reduction in the mass of insulin-secreting beta cells in the pancreatic islets, resulting in absolute insulinopenia, which is responsible for the accompanying metabolic defects.
The precise mechanisms involved in the etiopathogenesis of type I or IDDM remain uncertain, but are the subject of much current investigation. In fact, IDDM probably comprises a group of clinically similar conditions that share beta cell failure as the most important pathogenetic factor. Clear distinction and definition of the various subtypes of IDDM is not yet possible. It is likely that most forms of IDDM are a consequence of an interaction between genetic and environmental factors. Current evidence suggests that the etiologic factors involved include: genetically determined diatheses which predispose the individual to IDDM (diabetogenic genes), islet cell damage triggered by environmental factors (viral infections or chemical toxins), and immunologically mediated islet beta cell destruction.
The familial predilection for the development of diabetes mellitus has been recognized for nearly 3,000 years. Throughout the centuries that followed and into modern times, the familial tendency of diabetes mellitus has been a recurrent theme in the writings of physicians studying the disease. A variety of hypotheses have been proposed to account for the familial tendency of diabetes, but a specific mode of genetic transmission has not been established. That a genetic component exists is supported by the much higher concordance rate for diabetes among monozygotic (identical) twins than among dizygotic (nonidentical) twins, a finding which is generally indicative of an inherited component.
Much recent attention has been focused on the possible association of specific alleles in the major histocompatibility locus (HLA) with IDDM as well as whether the HLA region either contains or is in linkage with a diabetogenic gene(s) on the same chromosome.
The HLA (human leukocyte antigen) system is the major histocompatibility complex (MHC) of man. The HLA gene complex is located on the short arm of chromosome 6 and is recognized by a number of closely linked but distinct loci. There are two broad classes of HLA loci. Class I loci include the HLA-A, -B, and -C loci, each of which is highly polymorphic. Class I antigens are glycoprotein products of class I loci expressed on the cell surfaces of all nucleated cells. Class I molecules play an important role in the perception of antigen by cytotoxic T cells. The cytotoxic cells and their target cells must share the same class I antigens for lysis to occur. Class II loci (or D region loci) include the DP, DQ, and DR subregions of HLAD. Each subregion is highly polymorphic. Class II antigens are normally expressed on the surface of only macrophages and B- lymphocytes, although aberrant expression is possible and may be important in disease pathogenesis. Class 11 molecules regulate the interaction of macrophages and B-lymphocytes with T-lymphocytes (specifically helper and suppressor T-lymphocytes). Thus, the class II genes have been called "immune response" genes.
All of the HLA loci have considerable polymorphism and exist in multiple allelic forais. Each individual inherits one maternal and one paternal HLA-haplotype (a set of A, B, C, DP, DQ, and DR antigens). The HLA gene complex occupies only 1.6 centimorgans of chromosome 6 so that inheritance of HLA haplotypes from each parent usually occurs en bloc without recombinations taking place (at least for the more commonly determined serologically defined antigens, ie, A, B, C, and DR). The prevalence of different HLA antigens varies considerably in different populations throughout the world. The genes are codominant, which means that both alleles are expressed at each HLA locus such that any given individual may express two specificities at each of the loci. The HLA region is characterized by a high degree of linkage disequilibrium, which is the tendency for some alleles at closely linked loci to occur together in the same haplotype more often than would be predicted on a statistical basis.
The relationship between diabetes mellitus and the HLA system has been the subject of intense investigation for over a decade.1'4 The HLA system has been found to have no role in NIDDM. On the other hand, the relationship between IDDM and the HLA system is complex.
Across populations, there is clearly an association between IDDM and certain HLA alleles. There is a strong association between IDDM and both DR3 and DR4. Associations between IDDM and Al, Bo, B15, Bl 8, B40 and Cw3 are thought to be secondary to the increases in DR3 and DR4 since these antigens all are in linkage disequilibrium either with DR3 or DR4. Thus, "axes of susceptibility," dictated by linkage disequilibrium, have been defined. For example, Al, B8, Cw 7, and DR3 would be one such axis, inherited as a cluster. Another would be the A2, B15, Cw3, DR4 axis. The strongest relationships have been with DR3 and DR4, and these are thought to be primary. In white populations, 92% to 98% of IDDM individuals have been found to be DR3 or DR4 positive. This strong association between IDDM and DR3 or DR4 implies a causal relationship between these factors and the disease. Yet, it is important to note that 50% to 60% of the nondiabetic white population carry DR3 or DR4, while IDDM afflicts only 0.3% to 0.5% of the population. Thus, the DR relationships may be necessary but not sufficient to define disease susceptibility. Alternatively, there may be only some DR3 or DR4 subtypes that carry IDDM susceptibility.
By contrast, there is a negative association between DR2 and IDDM (with an accompanying reduced frequency of B7, probably explainable by linkage disequilibrium with DR2). The presence of DR2 may confer some "protective" effect against islet beta cell damage and IDDM.
The relative risk for carriers of both DR3 and DR4 is disproportionately increased when compared with other DR3 and DR4 individuals. This suggests that there may be two genes that predispose an individual to the development of IDDM - one associated with DR3, the other associated with DR4. When present together the two genes act in concert to increase susceptibility.
The above discussion has focused primarily on population studies of the association between IDDM and specific alleles recognized by HLA. Within families, there is clearly linkage between IDDM and the HLA system, regardless of what HLA alleles are manifest in any given family.
Linkage analyses between IDDM and HLA within families are designed to assess the relationship between loci of potential diabetogenic genes and HLA loci as well as to gain insight into the mode of diabetogenic gene inheritance. As noted earlier, an HLA haplotype is a set of HLA-A, -B, -C, and -DR antigens. Individuals are HLA-identical if they share two haplotypes, haploidentical if they share one haplotype, and nonidentical if they share no haplotypes. Analyses have focused principally on families with several members affected by IDDM, particularly "multiplex" kindreds in whom two or more sibs have diabetes. At least 263 such sib pairs have been studied.4,5 If there were no relationship between diabetes and the HLA system, the expectation would be that 25% of diabetic sib pairs would be HLA- identical, 50% would be haploidentical, and 25% would share no haplotypes. This is the case for nondiabetic sib pairs in these families. This distribution is not seen for IDDM sib pairs, in whom 59% are HLA-identical, 37% are haploidentical, and 4.6% share no haplotypes. These data indicate that IDDM is influenced by one or more loci in genetic disequilibrium with the HLA region.
A number of genetic analyses have been applied to this type of data with conflicting claims. It is likely that there is heterogeneity within IDDM and that more than one gene is involved. Thus, the actual mode of inheritance of IDDM disease susceptibility remains poorly defined.
It seems likely that there are one or more diabetogenic genes linked with or part of the HLA complex. The diabetogenic genes seem to confer susceptibility to diabetes, with the additional requirement of noxious agents in the environment to express the genetic susceptibility and reveal the clinical syndrome. Specific HLA antigens (eg, DR3, DR4) may also interact with other etiologic factors to influence directly the emergence of IDDM in genetically predisposed individuals. Different HLA antigens may modify different genes or have differing mechanisms of influence.
It should be noted that there are limitations with this hypothesis. An all-or-none relationship between IDDM and the HLA system has not been defined. A diabetogenic gene has not been identified. This has led some to propose that HLA-DR quantitatively controls the production of some factor critical for IDDM development.4 This would account for variation in risk of IDDM dependent on HLA-DR type, without requiring that there be specific diabetogenic genes.
Others are diligently trying to resolve the IDDMHLA association at the DNA level, with the goal of identifying and characterizing the specific diabetogenic gene(s).6 DNA restriction enzymes and radiolabeled cDNA probes have been used to probe the HLA-D region. These studies suggest that there are subtypes of DR3 and DR4, only some of which may be associated with susceptibility to IDDM. The differences may be at the DQ subregion, rather than the DR subregion. Whether such putative diabetogenic genes are components of particular HLA-D region subtypes, or whether such diabetogenic genes are in linkage disequilibrium with particular HLA-D region subtypes remains to be established.
The identification of diabetogenic genes or gene products would permit better genetic counseling, amniocentesis and fetal diagnosis, identification of susceptible individuals, and ultimately, even the potential of genetic manipulation either to alter susceptibility or to replace gene function.
Although genetic factors appear necessary for IDDM to develop, they are not sufficient. For example, although identical twins have a higher concordance rate for IDDM than do nonidentical twins, the rate is only approximately 54.4%. 7 The relatively high rate of discordance in IDDM twins is important because differences between monozygotic twins must be environmental, whereas similarity may be genetic or environmental in origin. Thus, these studies imply that there is an important environmental contribution in the IDDM twins. What is inherited, therefore, is disease susceptibility. Since the HLA gene system controls immune responses, it appears that the individual with IDDM may inherit a susceptibility to abnormal immune responses. This may lead to increased vulnerability to certain pancreatotoxic viruses and chemicals, immune over-reactions and islet beta cell damage, or chronic autoimmunity to beta cells, with the final result being beta cell destruction and insulinopenia.
Although it has been argued that the substantial discordance of IDDM in identical twins mandates that environmental factors play a role in human IDDM, this view has been challenged by invoking the potential of genetic diversity between such "identical" twins. This argument eliminates the need to include environmental events in the disease sequence.8 According to this view, susceptibility to IDDM is genetically determined and inevitable in individuals with the genetic diathesis.
There is some evidence, mostly circumstantial and indirect, that viral infections serve as an environmental trigger leading to the development of at least some forms of IDDM.9,10 Many viruses have been implicated including: mumps, rubella, coxsackievirus B3 or B4, infectious mononucleosis, infectious hepatitis, cytomegalovirus, polio, influenza, and varicella. The role viruses play in IDDM is supported by the following studies and case reports.
There have been a number of case reports of diabetes emerging in temporal relationships with viral infections. Moreover, some epidemiologic studies have shown an association between IDDM and serologic evidence of previous viral infection, although other extensive surveys have failed to find such serologic evidence of viral infections in relationship to IDDM.
Congenital rubella infection has also been associated with the subsequent development of IDDM in childhood or young adult life. Fully 10% to 20% of individuals subjected to congenital rubella ultimately develop IDDM, and another 20% develop impaired glucose tolerance.11
Epidemiologic studies have revealed seasonal trends in the onset of diabetes, usually with increased prevalence of new cases in either the winter or late summerearly autumn months.12 The peak incidence of diabetes may coincide with the prevalence of common viral infections in a given community. It is not clear, however, whether viral infection is etiologically important or merely a coincidental event that brings IDDM to clinical recognition.
It has been possible to demonstrate that several viruses can infect and destroy human pancreatic beta cells in tissue culture.
There is evidence that a number of viruses can infect and damage islet beta cells in experimental animals, and lead to diabetes.9,10 Certain forms of virus-induced diabetes in mice are under genetic control. Some strains of mice are highly susceptible (83% occurrence rate) to the development of diabetes when exposed to virus, whereas other strains are highly resistant to the same virus (2% occurrence rate). These studies provide us with remarkable insight into the potential interaction of a genetic predisposition for diabetes and viral infection.
Pathologic studies of pancreases obtained at autopsy of patients who died (often of other causes) shortly after the onset of insu lin -dependent diabetes have revealed an inflammatory picture of lymphocytic infiltration surrounding or penetrating the islets and termed "insulitis" or "isletitis." Moreover, an autopsy study of children who died of various viral illnesses has demonstrated insulitis even in the absence of ante mortem hyperglycemia.13 This type of inflammatory lesion is consistent with either viral or immune reactions and is identical to the histologic picture seen in animals with diabetes caused by viral infection.
It has also been possible to recover virus from human pancreatic tissue. In fact, in one report the investigators succeeded not only in isolating a virus from the pancreas of a child who died several days after acute onset of diabetic ketoacidosis, but also in demonstrating that the virus isolated (probably a variant of Coxsackie B4) could produce diabetes in mice.14 It is not yet known whether this case constitutes a rare nonrepresentative situation or whether beta cytotropic viral infection is a common feature associated with the fulminant onset of IDDM.
The role of viruses in the etiology of IDDM remains to be clarified. Issues still to be resolved are: whether many different viral infections can serve as environmental triggers for IDDM; whether a single unknown diabetogenic virus is usually responsible; or whether viral infections merely serve to bring IDDM to clinical recognition, without playing any specific role in beta cell destruction.
Other Environmental Factors
A variety of chemical toxins also appear to have the potential of inducing islet beta cell damage.10·15 Some have the potential to cause permanent diabetes by destroying beta cells and resulting in insulin deficiency. The nitrosourea compounds are among these toxins. Nitrosourea compounds are ubiquitous in our environment and represent only one class of chemical compounds that may lead to IDDM. One such chemical toxin implicated in a number of cases of human diabetes is the rodent icide Vacor®.
The induction of diabetes by strep tozotoc in provides another instructive model. When relatively large doses are given, streptozotocin has a direct toxic effect on islets. When administered as repeated, relatively small "subdiabetogenic" doses to mice, diabetes also emerges with insulitis and beta cell degranulation. According to this model, subtle chemical injury results in activation of an immunologic response, with genetic factors (strain of mice) influencing the response.
Potential Therapeutic Developments
The identification of environmental triggers, in theory, could permit the development of vaccines, the use of interferon therapy or other antiviral drugs, and the development of antidotes for chemical toxins. Unfortunately, these developments are not likely to occur because of the large array of viruses and chemical toxins that have the potential for serving as contributing factors to IDDM.
One formulation is that cumulative environmental insults (both viral and chemical) may lead to progressive beta cell destruction in genetically susceptible individuals. (There is animal evidence in support of this view.16) Existing evidence suggests that such environmental exposure may be remote in time from the onset of IDDM (even in utero in the case of congenital rubella). If this is the case, environmental manipulation is even less likely to be helpful because the toxin or virus would no longer be present.
Abundant evidence suggests that islet beta cell destruction is immunologically mediated. It has been argued that this is "autoimmune" in nature, which might well be the case if environmental factors are not involved. Yet, it must be emphasized that immunemediated destruction does not necessarily imply spontaneous autoimmunity. Moreover, the immunological mechanisms involved in the pathogenetic pathway have not yet been clearly defined, even in animal models. It also should be noted that several different pathogenetic sequences may result in immune destruction of beta-cells. Thus, for example, the pathways involved in typical IDDM may be different than those involved when IDDM is but one component of a polyendocrine autoimmune syndrome. IDDM may constitute a group of clinically similar conditions which share beta cell failure as the ultimate pathologic feature. Clear distinction and definition of various subtypes of IDDM is not yet possible.
The immunological events that mediate beta cell destruction are a subject of intense investigation. That the immune system is involved is suggested by many types of evidence. The relation of the HLA system to IDDM, discussed in detail above, implies immune response regulation.
Important evidence of immune-mediated destruction is provided by monocytic and lymphocytic infiltration of the islets, ie, insulitis (isletitis).17,18 This lesion is consistent with an immune type reaction, or at least is similar to the lymphocytic infiltration encountered in other reputed autoimmune conditions including endocrinopathies. The majority of lymphocytes are of the T-cytotoxic/suppressor type.19 This pathologic evidence supports a cell-mediated basis for the pathogenesis of IDDM. In addition, in one study there was aberrant expression of DR (class II molecules) in islets containing insulin at the time of onset of IDDM. This aberrant expression appears whether or not insulitis is present in that islet.
The presence of circulating antibodies to islet cells and insulin at the time of diagnosis may be additional evidence of immune activation potentially mediating beta cell destruction.20,22
Several types of antibodies have been described. Islet cell cytoplasmic antibodies (ICA) bind to islet cell cytoplasmic antigens and are detected by immunofluorescence using cryostat sections of pancreas. ICAs are islet-specific, but not beta cell-specific, and are of the IgG and IgM classes. Islet cell surface antibodies (ICSA) bind to surface antigens on the plasma membrane, and are detected using cultured living cells. They react mainly with islet beta cells and can mediate complement dependent lysis (complement dependent, antibody-mediated cytotoxicity, or CAMC) of islet cells in vitro, and also inhibit insulin secretion in vitro. Complement-fixing islet cell antibodies (CF-ICA) have been described and may be more specific than conventional ICA, although it would seem that CF-ICA merely represents high titers of conventional ICA. There are also anti-insulin antibodies (insulin autoantibodies, or IAA) present at the onset of IDDM, prior to administration of exogenous insulin. The presence of IAA suggests that insulin itself may serve as a surface antigen stimulating the immune system (presumably in association with aberrantly expressed class II antigens), or that insulin release (presumably in a damaged form) as a consequence of beta cell destruction may induce antibody formation.
One of the various anti-islet antibodies is found in the vast majority (60% to 95%) of IDDM subjects at the time of diagnosis. The prevalence of these islet cell antibodies decreases with increasing duration of IDDM. Only about 20% to 25% of patients have detectable antibody three to five years after diagnosis, and those with persistent antibody often have evidence of other autoimmune endocrine disease. These patients may have polyendocrine autoimmune syndrome and so-called type IB diabetes.
Definite proof that these islet cell antibodies destroy islet beta cells in the subject with IDDM is not available, ie, it is not known whether islet antibodies are primary or secondary to beta cell damage. The latter is likely with ICA, since these antibodies are directed against cytoplasmic targets not ordinarily accessible to serve as immunogens. Although the demonstration of these various antibodies does not necessarily imply a causal role for them in the etiology of IDDM, nevertheless, islet cell antibodies may serve as useful markers of immune activity in IDDM.
Current interest has focused on potential cell-mediated immunologic mechanisms for beta cell destruction, although these mechanisms require further definition.21'23 Recently diagnosed IDDM patients show increases in la-antigen bearing T-lymphocytes (thought to represent activated T cells); inhibited lymphocyte migration caused by components of the endocrine pancreas; adherence of lymphocytes to and dissolution of cultured human insulinoma cells (lymphocytes inhibit insulin release from or lyse rat islet cells) ; and defects in functional suppressor cells. These findings are similar to those observed in several autoimmune diseases.
Amongst the strongest evidence in favor of immune mechanisms mediating beta cell destruction is the accelerated "re-enactment" of the pathogenetic sequence following pancreatic transplantation in identical twins (from a nondiabetic twin to a diabetic cotwin) in the absence of immunosuppression.24·25 The genetic identity precluded rejection (which was also excluded because the transplanted kidney from the same donor continued to function, and the histology was not consistent with rejection), yet insulitis and pancreatic graft failure occurred in a rapid course.
In the best available animal models for IDDM - the BB (biobreeding) rat, the NOD (nonobese diabetic) mouse, and the low dose streptozotocin- induced diabetic mouse - immune mechanisms seem to be crucial to the appearance of the diabetic state. 15,26,27 Each of the animal models develop insulitis, followed by loss of beta cells, and the appearance of diabetes. The diabetes can be prevented or reversed by a variety of immune interventions.
There is a subtype of IDDM (sometimes called type IB diabetes) in which a familial autoimmune abnormality involves not only pancreatic islets, but other endocrine organs as well (polyendocrine autoimmune syndrome).28'30 This form of IDDM is associated with other disorders of presumed autoimmune etiologySome individuals have combined autoimmune disorders (ie, Schmidt's syndrome and polyendocrine failure). Others have a variety of organ-specific autoantibodies. Often there is also a family history of autoimmune diathesis. IDDM patients with polyendocrine autoimmune syndrome often have a later age of onset and are more likely to be female.
Recent evidence suggests that the islet immunopathology may commence several years prior to the clinical recognition of IDDM. Such evidence comes from at least three sources.
In a large prospective study in England, the first degree relatives (parents and siblings) of 198 probands with IDDM were monitored.31 Of 719 unaffected first degree relatives, 24 were demonstrated to have positive complement fixing ICA (CF-ICA) at some time during follow-up. Of these, 13 (54%) developed IDDM during follow-up. In all cases, complement fixing ICA (CF-ICA) was demonstrable from 1 to 10 years prior to the clinical onset of IDDM. In contrast, only 2 of 665 (0.3%) relatives who lacked both ICA and CF-ICA developed IDDM during the first 10 years of follow-up.
In another study conducted in Boston, involving unaffected monozygotic twins (or triplets) and other first degree relatives of IDDM probands, ICA could be demonstrated retrospectively in the frozen sera of twins who eventually developed IDDM, up to 8 years prior to the clinical onset of the disease. 8,32,33 Moreover, the appearance of ICA was associated with a progressive decline in islet beta cell function, as measured by a decrease in early insulin release after an intravenous glucose load.
Finally, in addition to the classical insulitis lesion frequently seen at the time of IDDM diagnosis, the majority of islets are devoid of beta cells and demonstrate only other types of islet cells (alpha cells or delta cells), and are termed "Pseudoatrophie" islets.17,18,34,35 Presumably, the inflammatory immune process has abated, having occurred much earlier in those islets, with IDDM appearing only when sufficient beta cells have been destroyed and carbohydrate tolerance can no longer be maintained. This evidence is consistent with the theory that there is a slow, continuing immune process antedating the clinical diagnosis of IDDM.
If IDDM is an immunologically mediated disease, then immune intervention should alter the natural history of the disease, and potentially abort the syndrome. This would be true whether the disease was due to spontaneous autoimmunity, or was simply immune mediated. Indeed, if immune intervention does prevent beta cell destruction, this would provide the strongest evidence in support of immune mechanisms being involved in the pathogenesis of IDDM. In fact, immune intervention does prevent or reverse the disease in several animal models of IDDM, including the spontaneous disease both in the BB rat and the NOD mouse, and the immune-mediated, environmentally triggered disease in the low dose streptozotocin mouse model.
The evidence that an immune mechanism may be important in the etiopathogenesis of human IDDM, coupled with the success of immune intervention in animal models, has led to a number of clinical trials of various immune intervention therapies in IDDM. In human beings, immune intervention trials have been initiated to confirm that the immune system is involved in the pathogenesis of IDDM, to clarify and define the immune mechanisms involved in the pathogenetic sequence, and to lead to clinically applicable intervention. The latter may depend on the development of specific forms of immune intervention targeted at those pathways involved in beta cell destruction.
Most studies have involved treating patients shortly after the clinical onset of IDDM, usually within the first several weeks after diagnosis. Immune intervention strategies tested have included the use of glucocorticoids; plasmapheresis; interferon; methisoprinol; a combination of glucocorticoids, azathioprine, and antilymphocyte globulin, with or without plasmapheresis; antithymocyte globulin; azathioprine; cyclosporine; ciamexone; levamisole; gamma globulin; theophylline; leukocyte transfusions; monoclonal antibodies CBLl, OKT3, orT12; nicotinamide; and transfer factor.36 Several of these have been tested in controlled clinical trials. Two large trials with cyclosporine (25% in remission) have demonstrated an increased frequency of insulin-free remissions at one year in comparison with a placebo group (2% to 10% in remission).37,38 In spite of this, the existing data are too meager to permit a conclusion to be drawn about the clinical utility of any immune intervention strategy. Longer follow-up is needed to define the duration of remissions and to ascertain benefits in relationship to risks.
Yet, immune intervention trials in recent onset insulin dependent diabetes mellitus clearly have demonstrated that immune intervention can indeed alter the natural history of the disease. This provides convincing support for the hypothesis that immune mechanisms are important in the etiopathogenesis of IDDM. Accompanying studies of the immune system may provide insight into the nature of the immune abnormalities in IDDM.
Much effort is being expended to identify the nature of islet beta cell antigens that might be targets of the immune system. The further identification and characterization of such antigens will allow us to take intervention approaches much further. For example, antigen identification should permit development of methods for early recognition of the pathogenetic process, presumably with greater accuracy than the use of ICA or CF-ICA does today. Moreover, it will permit the design and use of immune intervention strategies specific for the processes leading to pancreatic islet beta cell destruction. Such intervention might involve the development of monoclonal antibodies that can serve as blocking antibodies or anti-idiotypic antibodies, or monoclonal antibodies specific to macrophage or cytotoxic T-lymphocytes that might be involved in islet cell destruction.
As the pathogenetic sequence is clarified, it is possible to project the following sequence. At birth the population would be routinely screened for diabetogenic genes or gene products. Once the individuals with the potential of developing IDDM were identified, they would be followed longitudinally in search of evidence of immune damage to islet beta cells. In such individuals, evidence of altered or diminishing beta cell function would be sought. If identified, these individuals would become candidates for intervention therapy designed to abort the pathogenetic sequence.
1. Nerup J; HLA studies in diabetes mellitus: A review. Advances in Mewbotk Disorders 1978; 9:263-277.
2. Christy M. Green A, Christan B, et al: Studies of the HLA system and insulindependent diabetes mellitus. Diabetes Care 1979; 2:209-214.
3. Cudworth AG. Wolf E: The genetic susceptibility to type I (insulin-dependent) diabetes mellitus. CIm Endocrinol Mewb 1982; 11:389-408.
4. Nerup J1 Mandrup-PoulsenT, MolvigJ: The HLA-IDDM association: Implications for etiology and pathogenesis of IDDM. Diabetes Metabolism Reviews 1987; 3:779-802.
5. Nerup 1, Christy M1 PIaH P, et al: Aspects of the genetics of insulin-dependent diabetes mellitus, in Andreani D, Di Mario U, Federlin KF, et al (eds): immunology m Diabetes. London, Kimpton Medical Publications, 1984, pp 63-70.
6. Owerbach D, Rich C, Carnegie S, et al: Molecular biology of Ae HLA system in insulin dependent diabetes mellitus. Diabetes Metabolism Reviews 1987; 3:819-834.
7. Barnen AH, Eff C. Leslie RDC, et al: Diabetes in identical twins. A study of 200 pairs. Diabetoiogia 1981; 20:87-93.
8. Elsenbarth GS: Genes, generator of diversity, glycoconjugates. and autoimmune betacell insufficiency in type I diabetes. Diabetes 1987; 36:355-364.
9. Craighead JE: Viral diabetes in man and experimental animals. Am J Mfd 1981; 70:127-134.
10. Rayfield EJ, Ishimura K: Environmental factors and insulin dependent diabetes mellitus. Diabetes Metabolism Reviews 1987; 3:(in press).
11. Rubinstein P. Walker ME, Fedun B, et al: The HLA system in congenital rubella patients with and without diabetes. Diabetes 1982; 31:1088-1091.
12. Fleegler FM, Rogers KD, Drash A, et al: Age, sex and season of onset of juvenile diabetes in different geographic areas, ftdiafrics 1979; 63:374-379.
13. Jenson AB, Rosemberg HS, Notkins AL: Pancreatic islet cell damage in children with fetal viral infections. Lancet 1980; 2:354-358.
14. Yoon JW1 Austin M, Onodera T, et at: Vims-induced diabetes mellitus. Isolation of virus from the pancreas of a child with diabetic ketoacidosis. N Engl J Med 1979; 300:1173-1179.
15. Mordes JP, Rossini AA: Animal models of diabetes. AmJ Med 1981; 70:353-360.
16. Toniolo A, Onodera T, Yoon JW, et al: Induction of diabetes by cumulative environmental insults from viruses and chemicals. Nature 1980; 288:383-385.
1 7. Gepts W: fóthologécal anatomy of the pancreas in juvenile diabetes mellitus. Diabetes 1965; 14:619-633.
18. Gepts W, VeU PAV: islet morphologic changes. Diabetes Metabolism Reviews 1987, 3:(in press).
19. Bottazzo GF, Dean DM, McNaIIy JM, et al: In situ characterization oí autoimmune phenomena and expression of HLA molecules in the pancreas in diabetic insulitis. N Engl J Med 1985; 313:353-360.
20. NerupJ, Lern mark A: Autoimmunity in diabetes mellitus. Ami Med 1981; 70:135-141.
21. Lernmark A, Li S, Baeklteskov S, Christie M, et al: Islet specific immune mechanisms. Diabetes Metabolism Reviews 1987; 3:(in press).
22. Palmer JP: Insulin autoantibodies: Their role in the pathogenesis of IDDM. Díateles Metabolism Reviews 1987; 3:(in press).
23. Barbosa J, Bach FH: Cell-mediated autoimmunity in type I diabetes. Diabetes Metabolism Reviews 1987; 3:<in press).
24. Sutherland DER, Sibley R, Xu X-Z, et al: Twin-to-twin pancreas transplantation: Reversal and reenactment of the pathogenesis of type I diabetes. Tram Assoc Am Physicians 1984; 97:80-87.
25. Sibley RK, Sutherland DER, Goetz F, et ah Recurrent diabetes mellitus in the pancreas iso- and allograft. Lab Invest 1985; 53:132-144.
26. Mordes JP, Desemone J, Rossini AA: The BB rat. Diabetes Metabolism Reviews 1987; 3:725-750.
27. KoIb H: Mouse models of insulin dependent diabetes: Low-dose streptozotocininduced diabetes and nonobese diabetic (NOD) mice. Diabetes Metabolism Reviews 1987; 3:751-778.
28. MacCuish AC, Irvine WJ: Autoimmunological aspects of diabetes mellitus. CIm Endocrinol Metab 1975; 4-435-471.
29. Christy M, Dcckcrt T, Nerup J: Immunity and autoimmunity in diabetes mellitus. Clin Endocrinol Metab 1977; 6:305-332.
30. Eisenbarth GS, Wilson PW, Ward F, et al: The polyglandular failure syndrome: Disease inheritance, HLA type, and immune .function. Ann Intern Med 1979; 91:528-533.
31. Gorsuch AN, Spencer KM, Lister J, et al: Evidence for a long prediabetic period in type 1 (insulin-dependent) diabetes mellitus. Lancet 1981; 2:1363-1365.
32. Srikanta S, Ganda OP, Eisenbarth GS, et al: Islet-cell antibodies and beta-cell function in monozygotic triplets and twins initially discordant for type I diabetes mellitus. N Engl J Med 1983; 308:322-325.
33. Srikanta S, Ganda OP1 Jackson RA, et al: Type I diabetes mellitus in monozygotic twins: Chronic progressive beta cell dysfunction. Ann Intern Med 1983; 99:320-326.
34. Rahler J, Goebbels RM, Henquin JC: Cellular composition of the human diabetic pancreas. Diabetokgia 1983; 24:366-371.
35. Faults AK, Stewart JA: The pancreas in recent onset type I (insulin-dependent) diabetes mellitus: Insulin content of the islets, insulitis, and associated changes in the exocrine acinar tissue. Diabetologia 1984; 26:456-461.
36. Skyler JS: Immune intervention studies in insulin dependent diabetes mellitus. Diabetes Metabolism Reviews 1987; 3:(in press).
37. Feutren G, Lapoz L, Assan R, et al, for the Cyclosporin/Diabetes French Study Group: Cyclosporin increases the rate and length of remissions in insulin dependent diabetes of recent onset: Results of a multicentre double-blind trial. Lancet 1986; 11:119-124.
38. KoIb H, Nerup J, for the Canadian/European Diabetes Study Group: The Canadian/ European cyclosporine study. Prcxnced sc the Juvenile Diabetes foundation Immunosuppressive Therapy Workshop, Houston, Texas, February 9-10, 1987.