ETIOLOGIC CLASSIFICATION

Type 1 Diabetes

Type 1 diabetes (T1DM) comprises several diseases of the pancreatic ß cells which lead to an absolute insulin deficiency. This is usually the result of an autoimmune destruction of the pancreatic ß cells (type 1A). Some patients with T1DM with no evidence of ß cell autoimmunity have underlying defects in insulin secretion or ß cell glucose sensing while others have no known cause for ß cell dysfunction or impaired insulin secretion (type 1B diabetes mellitus).

Type 2 Diabetes

Type 2 diabetes (T2DM) is by far the most common type of diabetes and is characterized by insulin resistance resulting from defects in the action of insulin on its target tissues, but complicated by varying degrees of insulin secretion insufficiency as well. Most patients with T2DM in the US and Europe are obese, however in India and China, most T2DM patients are lean.

Genetic Defects of ß-cell Function

Approximately 2 to 4 % of patients who present with mild diabetes at a young age [maturity-onset diabetes of the young (MODY)] have autosomal dominant transmission of a gene encoding a primary defect in insulin secretion (2). The clinical characteristics of these patients are sufficiency heterogeneous, that they are not reliable in predicting the underlying pathogenesis (3). Several genetic abnormalities have been found that account for the disorder in such patients. Some members of an affected family may have the genetic defect but not develop diabetes phenotype. Whether this is due to interacting genes or environmental factors is unclear. MODY differs from the classical type-1A form of T1DM in several ways. With MODY, a family history of diabetes (if adequately known) is always present, hyperglycemia is mostly mild with a modest tendency to ketosis, the insulin secretion in response to oral (OGTT) or intravenous (IVGTT) glucose administration is modestly decreased, and evidence of islet cell autoimmunity is of course absent.

The underlying genetic defects underlying many of the MODY subtypes have been identified as indicated below. It is anticipated that other genetic subtypes of T1DM and T2DM will also become more clearly defined in the future. To date, six genetic forms of MODY are generally recognized; related to defects in the glucokinase gene and five pancreatic ß cell-specific transcription factor genes, hepatocyte nuclear factor-1-alpha (HNF-1a), HNF-1b, HNF-4a, insulin promoter factor-1 (IPF-1) and NeuroD1/BETA2.

In 1987 a MODY variant occurring in young African-Americans was described that was characterized by an age at onset of less than 40 years, an initial acute onset of disease indistinguishable from type 1 diabetes with ketosis or even ketoacidosis possible at disease onset, the need for insulin replacement therapy to correct fasting hyperglycemia for months to years following diagnosis followed by the development of a non-insulin-dependent course months to years later, and an autosomal dominant mode of inheritance (9, 10). The term "atypical diabetes mellitus of African-Americans" or ADM was therefore coined.

Point mutations in mitochondrial DNA cause another form of diabetes with an insulin secretory defect that is commonly associated with neuro-sensory hearing impairment and a strict maternal inheritance (11). In addition, genetic abnormalities that result in the inability to convert pro-insulin to insulin (12) or the production of mutant insulin molecules (13) are other examples of specific genetic defects in ß cell function which are rare causes of diabetes.

Genetic Defects in Insulin Action

There are a series of rare genetic abnormalities in the insulin receptor or in the signal transduction events, which follow insulin docking to its receptor that results in diabetes.

Endocrinopathies

Several hormones, such as epinephrine, glucagon, cortisol, and growth hormone, antagonize the action of insulin. Whereas release of these hormones constitutes the protective counter regulatory response to hypoglycemia, primary over secretion of these hormones can result in glucose intolerance or overt diabetes.

• Cushing's syndrome, due to pituitary and ACTH secreting adenomas or adrenal hyperplastic disease or to exogenous glucocorticoid administration, can lead to diabetes, especially in elderly or obese patients (14). In our experience steroid induced diabetes is most common when there is pre-existing insulin resistance syndrome.
• Acromegaly is associated with overt diabetes in 10 to 15 percent of cases, and impaired glucose tolerance in a further 50 percent (15, 16). There is great insulin resistance and marked hyperinsulinemic responses.
• Pheochromocytomas are associated with both inhibition of insulin secretion and an increase in hepatic glucose output (17). These changes lead to impaired glucose tolerance, the severity of which is directly related to the magnitude of catecholamine production (18).
• Glucagon-secreting tumors (glucagonoma) are associated with an unusual constellation of clinical features, including skin rash, weight loss, anemia, and thromboembolic problems. In addition, approximately 80 percent of these patients have either impaired glucose tolerance or diabetes (19).
• Somatostatin-secreting tumors (somatostatinomas) are typically associated with the triad of diabetes mellitus, cholelithiasis, and diarrhea with steatorrhea.
• Although thyroxine is not a counter regulatory hormone, hyperthyroidism can interfere with glucose metabolism. It is associated with both increased sensitivity of pancreatic ß cells to glucose, resulting in increased insulin secretion, and antagonism to the peripheral action of insulin (20). The latter effect usually predominates, leading to impaired glucose tolerance in untreated patients (21).

Drug- or Chemical-induced Diabetes

A large number of drugs can impair glucose tolerance; they may act by decreasing insulin secretion, increasing hepatic glucose production, or causing resistance to the action of insulin (22). Included in this list are several classes of antihypertensive drugs, such as beta blockers (23), protease inhibitors used for the treatment of HIV infection (24-26), and tacrolimus and cyclosporine used primarily to prevent transplant rejection (27, 28).

There is a suggested association between obesity, insulin resistance, hypertension, and dyslipidemia, which has been called syndrome X or the insulin resistance syndrome (29-31). The administration of a thiazide diuretic or a ß-blocker to such patients can exacerbate the insulin resistance and may cause hyperglycemia (23). In comparison, angiotensin-converting enzyme (ACE) inhibitors and alpha-adrenergic antagonists (such as doxazosin) may improve insulin sensitivity. Because the former also protect against renal disease, they are the drug of choice for diabetic patients with hypertension.

Viral Infections

Certain viruses can cause diabetes, either through direct ß cell destruction or by inducing autoimmune damage e.g. Coxsackie B. Chronic hepatitis C virus infection has also been associated with an increased incidence of diabetes, but it is uncertain as yet if there is a cause-and-effect relationship.

Uncommon forms of Immune-mediated Diabetes

Several uncommon forms of immune-mediated diabetes have been identified.

• The stiff-man syndrome is an autoimmune disorder of the central nervous system, which is characterized by progressive muscle stiffness, rigidity, and spasm involving the axial muscles, with severe impairment of ambulation (32). Patients usually have high titers of glutamic acid decarboxylase (GAD₆₅) autoantibodies; diabetes occurs in approximately one-third of cases.
• Anti-insulin receptor antibodies can bind to insulin receptors and either act as an agonist, leading to hypoglycemia, or block the binding of insulin and cause diabetes (33).
• In the following, we will emphasize the common immune mediated form of type-1 diabetes, the growing predicament of type-2 diabetes affecting younger patients and the underlying insulin resistance syndrome and obesity which predisposes to it.

NATURAL HISTORY OF HUMAN TYPE-1 DIABETES

In Western countries, the incidence rates of type-1 diabetes appear to be increasing, with more than 50 persons/100,000 in Finland now developing the disease annually (O Semill. personal communication-DIPP study). According to the JDFI website, 35,000 Americans develop clinical type-1 diabetes each year of which 13,000 are children, while according to the ADA, the cost of health care of diabetes in the US has been estimated to exceed $100 billion per annum, with type-1 diabetes having a disproportionate share. However studies including our own, indicate that many more diabetic adults with a type-2 diabetic phenotype have type-1 diabetes than previously recognized (34).

a). Structure and Functions of Pancreas

Pancreatic ß cells that secrete insulin are found in the islets of Langerhans, are specialized groups of a few hundred to a few thousand endocrine cells that are anatomically and functionally separate from pancreatic exocrine tissue that primarily secretes pancreatic enzymes into ducts that drain into the duodenum. Normal subjects have about one million islets, which in total weigh 1-2 grams and constitute less than 1% of the mass of the pancreas. Further, islets are composed of various types of cells. At least 70% are ß cells that are localized in the core of the islets, surrounded by A cells that secrete glucagons, D cells that secrete somatostatin and PP cells that secrete pancreatic polypeptide. All the cells communicate with each other through their extracellular spaces and through gap junctions.

Insulin, a peptide hormone composed of 51 amino acids is synthesized, packaged and secreted in pancreatic ß cells. Insulin is synthesized as preproinsulin in the ribosomes of rough endoplasmic reticulum. The preproinsulin is then cleaved to proinsulin that is transported to the Golgi apparatus where it is packaged into secretory granules. Most of the proinsulin is cleaved into equimolar amounts of insulin and C-peptide in the secretory granules.

Glucose is the major regulator of insulin secretion. When extracellular fluid glucose concentrations are high, glucose is taken up by the ß cells via glucose transporters, GLUT2 and GLUT1. It is then phosphorylated into glucose-6-phosphate by an islet specific glucokinase. Glucose metabolism thus increases cellular ATP concentrations and closes potassium-dependent ATP (KATP) channels in the beta cell membrane, causing membrane depolarization and influx of calcium. The rise in intracellular free calcium in ß cells promotes margination of the secretory granules, their fusion with the cell membrane, and release of cell contents insulin into the extracellular space.

b). Metabolic Derangements of Type-1 Diabetes

As the pancreatic ß cell mass declines in an islet cell antibody (ICA) positive person, the first metabolic abnormality discernable is a decline in the first phase of insulin release (FPIR) to an IVGTT (35). The insulin level after a 3-4 minute infusion of glucose at 0.5Gms/kg rises abruptly in normal children at about 8 years of age, perhaps coincident with the onset of adrenal puberty or adrenarche (36). In the relatives and children from the general population with positive ICA, a decline in the FPIR is a strong predictive marker of impending diabetes (36-38). Subsequently, there is a decline in the ability to keep the two-hour, post OGTT glucose level below 200mgs/dl. Often later, there is a rise in the fasting glucose levels as well. Transient insulin resistance is also seen which may be due to raised levels of free fatty acids (FFAs) from uncontrolled lipolysis (39) and decreased levels of hepatic glucokinase and GLUT 4 glucose transporters in adipocytes (40-42) which conspire to force the onset of symptomatic diabetes itself. Once insulin replacement therapy begins, there is usually some recovery in the patient's ability to secrete insulin (the "honeymoon" period), but within months to years, this partial recovery in endogenous insulin secretion ultimately fails. Initially, the glucagon secreting cells within the pancreatic islets are relatively preserved with excessive secretion of glucagon relative to insulin after protein meals (43). These elevated glucagon levels exacerbate the effects of the insulin deficiency, and promote lipolysis and ketoacidemia, effects that can be partially reversed by an infusion of somatostatin (44). This occurs at a time when neo-islet formation is seen, with regenerating islets having increased numbers of cells secreting glucagon or pancreatic polypeptide. However with time, there is continued loss of islets with glucagon deficiency in established, long standing IMD, making the patients susceptible to insulin induced hypoglycemia (45, 46). Insulin is the hormone of "feasting", promoting deposition of ingested nutrients into body stores and having multiple anabolic effects in tissues. Insulin deficiency thus induces a starvation like state, associated with excessive hepatic and renal gluconeogenesis, decreased peripheral utilization of glucose with glycosuria and loss of water and sodium salts, proteolysis in muscle liberating amino acids such as alanine and glutamine for gluconeogenesis (47-49), uncontrolled lypolysis with rapid mobilization of triglycerides with increased formation of very low density lipoprotein (VLDL) and fatty acids with formation of ketone bodies. Specifically, hepatic glucokinase levels fall with insulinopenia, synthesis of hepatic triglyceride and glycogen levels decline, malonyl CoA falls and thereby carnitine palmitoyltransferase-I levels rise promoting the transport of fatty acyl-CoA into mitochondria with the formation of acetyl-CoA. (50-52). In the liver, acetyl-CoA is converted into ß-hydroxybutyrate and acetoacetate in a proportion that depends upon the prevailing redox state, which provide additional fuel substrates for muscle and brain (49, 53, 54). Lipoprotein lipases are also inactivated, leading to reduced hydrolysis of triglycerides that may turn the serum milky with increased VLDL characteristic of the type 4 lipemic phenotype (55-57). A milky serum is characteristic of diabetes onset in nonobese diabetic (NOD) mice, animal model for human type 1 diabetes.

c). Genetic Susceptibility

Individuals with type-1A T1DM have inherited a number of quantitative trait loci (QTL) that encode protective and predisposing alleles which have exceeded the net genetic threshold required to predispose them to the disease. However this genetic threshold (penetrance) is dependent in turn on chance interactions with greater predisposing than protective environmental forces. The multiple genetic influences in T1DM comprise a major effect from DR/DQ genotypes (some 50% of the genetic effect ), coupled to several other QTLs with minor influences (Table 1). All of the latter QTLs are not obligatory genetic elements themselves since they are of minor-influence, but they collectively conspire to create additive influences on the genetic threshold. Siblings of a diabetic patient develop IMD at about 15 fold greater frequency than persons in the general population (1:250-300), or at an ls value of 15. The HLA predisposition to IMD is encoded by cis- and trans complementation DQA1*/DQB1* heterodimers which have an arginine at residue 52 of the A chain and a neutral amino acid (DQB1*0302, *0201) rather than a charged aspartic acid at residue 57 of the B chain (DQB1*0602/3 and DQB1*0301) (58), as modified by DRB1*04 subtypes (*0401 and *0405 are susceptible and *0403 and 6 are resistant types) (59) in the HLA genotype. Further, HLA-DP alleles have also been implicated, even though they are at a considerable recombination frequency away from the closely linked DR/DQ loci (60). Other genes involved include the variable number of tandem repeat (VNTR) alleles 5' to the insulin (INS) gene on chromosome 11p15, where the protective class III alleles (>200 repeats) are associated with increased expression of insulin in the thymus and thus theoretically leading to a more efficient eradication of insulin autoreactive T cells as opposed to class I alleles (26-63 repeats) that confer susceptibility to develop diabetes (61, 62). There are also CTLA-4 gene polymorphisms on chromosome 2q that are associated with IMD. CTLA-4 is an induced accessory molecule that is expressed on activated T cells. Recently, Todd et al have found that SNP polymorphism analyses across the region indicates that the association is indeed with the CTLA-4 gene, albeit no specific mutants have been discovered (ADA Meeting, 2002). CTLA-4 interacts with B7.2 expressed by antigen presenting cells (APC), thereby signaling apoptosis of T cells that become activated as part of an immune response, thereby confining the response. The NOD mouse has an enlarged lymphoid mass because of resistance of their T cells to undergo apoptosis (63), as has the CTLA-4 knockout mice, which readily develop lymphocytic organ infiltrates like NOD mice. These genes thus collectively affect the general ability to be tolerant to "self" antigens. Another susceptibility locus, (the IDDM 4) in the genomic interval on chr. 11q13 harbors the high affinity IgE Fc receptor gene that has been linked to atophy and asthma, which are inherent Th2 responses that may help to protect individuals against the development of anti-islet Th1 responses and thereby protect against IMD. It is in fact the authors' clinical experiences that persons with an allergic diathesis, do appear to have lower frequencies of IMD. There are other genomic intervals associated with or linked to IMD that have been putatively mapped, but these mostly lack plausible candidate genes in the DNA region, and pathogenic mechanisms for them cannot yet be offered, even if they become unambiguously confirmed. The NOD mouse however has been subjected to extensive genetic mapping studies, in the hopes that genomic intervals harboring susceptibility or protective genes which are syntenic to the humans will be discovered, thus hastening the identification of equivalent defective genes (64).

d). Autoantigens and Autoantibodies

The Doniach group in London, first reported islet cell autoantibodies in patients with autoimmune polyglandular syndromes (APSs) (65), especially in those with APS type-1 (APS-1) (66), even though such patients did not often develop diabetes. Lendrum and colleagues, having failed to find serological evidence for an autoimmune basis for chronic pancreatitis, did succeed in finding Islet Cell Antibodies (ICA) detectable by indirect immunofluorescence in patients with IMD, while islet cell surface reactive autoantibodies and autoreactive peripheral blood T cells were also reported (67, 68). Over the years that followed, the presence of ICA in US patients was confirmed but with lower frequencies of ICA among African American diabetic patients (69). Insulin autoantibodies (IAA) were discovered in patients with IMD before their first dose of insulin had been received (70) while their presence together with ICA denoted a group of non diabetic relatives of probands with IMD, that were at high risk for IMD themselves (71), albeit insulin itself is not an ICA antigen. Subsequently, much of the antigenic nature of this ICA reactivity has become clearer. It was recognized that many patients with "stiff" man syndrome who were prone to develop diabetes, also had ICA and autoantibodies to GAD₆₅. These GAD autoantibodies penetrated the blood brain barrier. High concentrations of GAD in the cerebellum inducing reduce brain levels of the inhibitory neurotransmitter gamma amino benzoic acid (GABA), thereby causing the appearance of epilepsy, depressed cognition, muscle spasms and motor dysfunctions. That GAD₆₅ was the antigen that accounted for the 64KDa islet cell protein previously discovered by Baekkeskov to react with autoantibodies in IMD, was later confirmed by the same investigator (73). Antibodies to recombinant GAD₆₅ and GAD₆₇ in IMD patients were soon reported (74), while the autoantibodies reacted to the antigens by conformational rather linear epitopes, and thus with native rather than denatured antigens. Thus they were best detected by liquid phase assays such as radioimmunoassay, rather than by an ELISA technique. In stiff-man syndrome, the predominant GAD autoantibodies reacted with linear epitopes. It became known that besides islet cell 64 KDa sized proteins, autoantibodies in the sera of IMD patients also precipitated islet cell proteins of 50, 40 and 37 KDa as well (75). The next islet cell antigen discovered was one of the two-dozen tyrosine phosphatases expressed in islet cells, insulinoma antigen-2 (IA-2) (76). This antigen shared structural homologies with the ICA-512 antigen already reported (77). A second tyrosine phosphatase named IA-2ß was discovered next (78). These additional tyrosine phosphatase antigens allowed for the matching of the islet cell proteins previously identifiable only by their molecular weights. Thus GAD₆₅ and its tryptic fragment explained the 64 and 50 KDa proteins, while tryptic fragments of IA-2 and IA-2ß were identical with the 40 KDa and the 37 KDa islet precipitable proteins respectively (79). The tyrosine phosphatases are a family of transmembrane enzymes of which only these two are expressed by the pancreatic islets and react with IMD autoantibodies. The reactivity is almost exclusively with the internal domains of these molecules, suggesting that they often arise as a consequence of islet cell damage from autoimmunity. Antibodies to IA-2 cross-react with those of IA-2ß in about 50% of the patient sera. Some unusual patient sera however react exclusively with IA-2ß. The question of why only these two members of the tyrosine phosphatase family are targets of islet cell autoimmunity has been answered by the finding that they are relatively resistant to proteolytic enzymatic digestion, and once released from islet cells after their lysis, are insoluble and thus become better antigens for auto-immunization, than those that remain soluble and are more rapidly digested (80). Recently, another antigen of 38KDa size (GLIMA) was added to the islet cell group, albeit only a minority of patient's sera reacts to it (81). Still more islet cell autoantigens are likely to be discovered. The detection of islet cell autoantibodies is useful for diagnosis of IMD from diabetes of other causes, and can be used to predict onset of diabetes months to years before onset of the clinical disease (37, 38, 82, 83) in non-diabetic relatives of probands with IMD, and even in the general population (84). Importantly, the clinical onset of the disease is often long preceded by the appearance of autoantibodies reactive to islet cells (ICA) (83) and to insulin (71), as independent age-related variables in predicting a diabetic outcome (85). The German BABY-DIAB study and the Finnish TRIGR study showed that islet autoantibodies can be transferred through the placenta from islet antibody-positive mothers to their offspring (234, 235). In the German BABYDIAB study, it was found that progression to multiple islet autoantibodies was fastest in children who were antibody positive by age 2 years and that progression to diabetes was inversely related to the age of first positivity for multiple autoantibodies (P = 0.02) (244). The timing for the appearance of the autoantibodies seems to be important. Studies in the NOD mouse showed that elimination of maternally transmitted immunoglobulin prevented spontaneous diabetes in progeny, suggesting that humoral factors present during gestation, including immunoglobulin, and autoantibodies potentially play an important role in the pathogenesis of ß cell destruction (236). Koczwara et al., demonstrated that 729 offspring of mothers with T1DM had significantly lower risk of developing multiple islet autoantibodies (5 year risk 1.3%) and diabetes (8-year risk 1.1%) when they were GAD or IA-2 positive, than offspring who were islet autoantibody negative at birth (237). Islet cell autoantibodies (ICA) also show a strong tendency to disappear after diabetes onset when all ß cells are destroyed (86, 87). Similar antibodies (IAA and GAD₆₅A) have been reported in NOD mice but with much less consistency, unless found as part of a recall response following exogenous immunization by that antigen at low titers. The presence of multiple autoantibodies strikingly increases the risk of diabetes, whereas one in the absence of all of the others if tested for, denotes only a modestly increased risk (37, 38). This suggests that antigenic epitope spreading is involved in a sustained or accelerated autoimmune attack (88). Besides autoimmunity to islet cell autoantigens, patients with IMD are subject to other autoimmunities. Thus IMD is a component part of the autoimmune polyglandular syndromes commonly in APS-2 and APS-3 and only unusually in APS-1 (66). Accordingly, patients with IMD have high rates of thyroid autoimmunity, especially if they are females (89), and are at risk for Addison's disease (89), atrophic gastritis (90), pernicious anemia (91), celiac disease (92), and vitiligo (93).

e). Antigen Specific Cellular Immunity

Autoreactive T cells that develop in impending IMD, home to the pancreatic islets where they become a component part of the evolving insulitis lesions. Thus circulating autoreactive T cells are relatively sparse in impending IMD. Never the less, antigen specific T cells as identifiable through prolonged in-vitro cultures in the presence of purified or recombinant islet cell autoantigens such as GAD (94) and IA-2 (95) have been reported. In fact, autoreactivity to a large number of autoantigens have been reported in both human and murine diabetes (96). T cell proliferative responses to insulin and GAD₆₅, and more generally to islet extracts, have been repeatedly reported in both patients with IMD (97, 98) and NOD mice (99-101). However, both in humans and NOD mice, reports of spontaneous proliferative responses have been difficult to reproduce and validate, probably because of the relative paucity of autoreactive T cells in peripheral blood samples, and the ready contamination of recombinant "test" antigens by lymphotoxin or lipopolysaccharide (LPS), that by itself, can produce proliferative responses even when present in trace amounts. Furthermore, significant T cell responses reported can be to insulin, proinsulin or GAD₆₅ antigen, in normal controls and type-1 diabetic patients (102-104). Numbers of laboratories have reported T cell reactivities in diabetic patients against GAD₆₅ and IA-2 and their peptides with variable results (95, 97, 105-110). However in established diabetes, the loss of the majority of ß cell mass resulting in associated loss of GAD₆₅ and other ß cell antigens, results in turn in the inactivation of T cells due to the loss of the peptide antigens that were driving the response. Thus antigenic/epitopic spreading is an undesirable phenomenon associated with progression in autoimmune diseases like IMD to a clinically significant outcome.

PATHOGENESIS OF IMD

The availability of Biobreeding (BB) rats and nonobese diabetic (NOD) mice, the rodent models of IMD, has greatly enhanced our understanding of the possible pathogenic mechanisms involved in human IMD (Fig. 1). Furthermore, epidemiological studies have allowed for a developing picture of the natural history to emerge. The process of destruction of beta cells is chronic in nature, often beginning during infancy and continuing over the many months or years that follow. At the time of clinical diagnosis of IMD, about +80% of the beta cells have been destroyed, whereas the islets are infiltrated with chronic inflammatory mononuclear cells (insulitis), including CD8+ cytotoxic T cells. Once islet cell autoimmunity has begun, progression to islet cell destruction is quite variable, with some patients rapidly progressing to clinical diabetes, while others remain in a nonprogressive state.

Figure 1. The pathogenesis of islet cell destruction. Islet cell proteins are presented by antigen presenting cells (APCs) to naïve Th0 type CD4+ T cells in association with MHC class II molecules. Interleukin (IL)-12 is thus secreted by APCs that promotes the differentiation of Th0 cells to Th1 type cells. Th1 cells secrete IL-2 and IFN-g that further stimulate CD8+ cytotoxic T cells or macrophages to release free radicals (superoxides) or perforin/granzymes, leading to ß cell apoptosis or death. CD8+ cytotoxic T cells further mediate ß cell death by Fas mediated mechanisms. Interleukin (IL)-4, on the other hand, secreted mainly by natural killer T (NKT) cells drives Th0 cell to Th2 pathway leading to benign insulitis.

Diabetes risk and time to diabetes in relatives of patients directly correlates with the number of different autoantibodies present as already discussed. The pathogenesis of IMD has been extensively studied, but the exact mechanism involved in the initiation and progression of beta cell destruction is still unclear. The presentation of beta cell-specific autoantigens by antigen-presenting cells (APC) [macrophages or dendritic cells (DC)] to CD4+ helper T cells in association with MHC class II molecules is considered to be the first step in the initiation of the disease process. Macrophages secrete interleukin (IL)-12 stimulating CD4 + T cells, to secrete interferon (IFN)-gamma and IL-2. IFN-g stimulates other resting macrophages to release, in turn, other cytokines such as IL-1beta, tumor necrosis factor (TNF)-a, and free radicals, which are toxic to pancreatic beta cells. During this process, cytokines induce the migration of beta-cell autoantigen specific CD8+ cytotoxic T cells. On recognizing specific autoantigen on ß cells in association with class I molecules, these CD8+ cytotoxic T cells cause ß cell damage by releasing perforin and granzyme and by Fas-mediated apoptosis of the beta cells. Continued destruction of beta cells eventually results in the onset of diabetes.

THE INDUCTIVE EVENT

Various mechanism have been proposed as discussed:

a). Molecular Mimicry

Antigenic molecular mimicry is defined by cross-reactive immune responses because of statistically significant structural homologies shared by molecules encoded by dissimilar genes. Either linear amino acid sequences of the molecules or their conformational epitopes may be shared, even though their origins are separate.

The incidence of type 1 diabetes has increased over the last three decades in Europe, while the clinical disease exhibits preferential seasonal onset (111). These observations emphasize the role of environmental factors in the disease process. It has long been suggested that IMD in humans is caused by viral infections. However, despite a vast increase in the information regarding the various genetic factors controlling the disease, little is known about the role of the putative environmental factors that might provide a more direct approach to therapy.

Several studies have reported a viral etiology associated with IMD (112-114). The disease pathogenesis may involve multiple factors including the genetics of the host, strain of the virus, activation status of the autoreactive T cells, upregulation of pancreatic MHC class I antigens, molecular mimicry between viral and ß cell epitopes and direct islet cell destruction by viral cytolysis. Viruses, as one of the environmental factors affecting the induction of IMD, may act as triggering agents of autoimmunity or as primary injurious agents, which directly damage pancreatic ß cells. Immune responses against a determinant shared by host cells and a virus could cause a tissue-specific immune response by generation of cytotoxic cross-reactive effector lymphocytes or antibodies that recognize self-proteins located on the target cells. Monoclonal antibodies against viruses have been observed to be capable of cross-reacting with host determinants (115).

Moreover, direct evidence for virus-induced diabetes largely comes from experiments in animals (116). Several studies in humans also point to viruses as triggers of the disease (117). Coxsackie B virus and rubella virus have been linked with IMD and in few instances, Coxsackie B virus has even been directly isolated from pancreatic tissues of individuals with acute IMD. Inoculation of this virus into mice then produced IMD, fulfilling Koch's postulates (118). The possibility that viruses might cause some cases of IMD by infecting and destroying pancreatic ß-cells has received considerable attention. However, it is difficult to demonstrate in vivo that viruses replicate in human ß-cells and/or produce diabetes in man. An in vitro system was therefore developed to determine whether viruses are capable of destroying human beta cells in culture (119, 120). By this method, it was clearly shown that several common human viruses, including mumps virus (121), Coxsackie B3 virus (122), Coxsackie B4 virus (123), reovirus type 3 (124), could infect human ß-cells. In addition, by radioimmunoassay, it was shown that the infection markedly decreased the insulin content of the ß-cells.

A strong correlation was found between the CMV genome in the immunocytes and the islet cell autoantibodies in the sera from diabetic patients (125). About 15% of newly diagnosed autoimmune IMD patients have been reported to have persistent CMV infections. Furthermore, it has been proposed that a molecular mimicry between protein 2C (p2C) of Coxsackie virus B4 and the autoantigen GAD₆₅ may play a role in pathogenesis of IMD. Kaufman et al (126) and Vreugdenhil et al (114), showed that the amino acid sequence of p2C shares a striking homology with a sequence in GAD₆₅ (PEVKEK) and is highly conserved in Coxsackie virus B4 isolates as well as in different viruses of the subgroup of Coxsackie B-like viruses. These are the most prevalent enteroviruses and therefore the exposure to the mimicry motif should be a frequent event throughout the life. Furthermore, they suggested that molecular mimicry might be limited to the HLA-DR3 subpopulation of the IMD patients.

Although numerous sequence similarities between viral proteins and ß-cell autoantigens are plausible, the relationship between Coxsackie virus infection and GAD₆₅ autoimmunity has recently received the most attention.

Glutamate Decarboxylase (GAD)

The finding by Kauffman et al (126), of a striking sequence homology of 18 amino acid peptide between human GAD₆₅ and the Coxsackie virus p2-C protein, enhanced the evidence of a specific molecular mimicry model involving GAD. In addition, this specific region of GAD₆₅ contains a T cell epitope involved in the GAD cellular autoimmunity in humans with IMD (94) and this region is an early target of the cellular immunity in NOD mice (100, 101). GAD catalyzes the formation of the inhibitory neurotransmitter g-amino butyric acid (GABA) from glutamine (127). Two forms of GAD exist (GAD₆₅ and GAD₆₇). GAD₆₅ is the predominant form within the human pancreatic islet cells, while GAD₆₇ predominates in mouse islets. Within the islets, GAD is predominantly observed within the ß-cells, while its roles in the inhibition of somatostatin and glucagon secretion and in the regulation of proinsulin synthesis and insulin secretion, have also been suggested (128).

Another study further supports a link between Coxsackie virus and IMD, associating IgM antibodies to Coxsackie B virus as a marker of recent exposure to the virus in newly diagnosed IMD patients and age/sex-matched controls (129). In that report, humoral immunity to Coxsackie virus and GAD appeared to cluster, even in people without diabetes. A set of overlapping synthetic GAD₆₅ peptides were used to study the most reactive T cell determinants in individuals at increased risk for IMD i.e. autoantibody positive, first degree relatives of IMD patients. Elevated in vitro T cell responses were observed to GAD₆₅ peptides (amino acids 247-266 and 260-279) in newly diagnosed IMD patients and autoantibody positive at-risk individuals (130). The sequence of this region of GAD₆₅ (amino acids 250-273) is significantly similar to the p2-C protein of Coxsackie B virus (112). However, not all published reports have demonstrated a linkage between immunity to GAD and Coxsackie virus. For example, one study identified a non-Coxsackie-homologous region of GAD₆₅ as a predominant cellular immune epitope while studying the polyclonal human T cell responses (107).

Animal studies also provide a strong support to this GAD-Coxsackie mimicry hypothesis. In the NOD mice, T-cell responses to GAD appear to be a key event in the induction and propagation of immunity to ß-cells. Immunization of NOD mice with either Coxsackie virus p2-C protein or the Coxsackie virus peptide containing the region of sequence similarity with GAD₆₅ could induce T cell responses that cross-reacted with GAD or GAD peptides corresponding to the region of sequence similarity (128).

IA-2

As discussed above, tyrosine phosphatase IA-2 is another molecular target of pancreatic islet autoimmunity in IMD. In one recent study, the epitope spanning 805-820 amino acid elicited maximum T-cell responses in all at-risk relatives, out of a total of 68 overlapping, synthetic peptides encompassing the intracytoplasmic domain of IA-2 (131). This epitope was found to have 56% identity and 100% similarity over 9 amino acids with a sequence in VP7, a major immunogenic protein of human rotavirus. This dominant epitope also has 75-45% identity and 88-64% similarity over 8-14 amino acids to sequences in Dengue, cytomegalovirus, measles, hepatitis C and canine distemper viruses and the bacterium Haemophilus influenzae. Furthermore, three other IA-2 epitope peptides have 71-100% similarity over 7-12 amino acid stretch to herpes, rhino-, hanta- and flaviviruses. Two others have 80-82% similarity with dietary proteins of milk, wheat and bean proteins. These molecular mimicries could lead to triggering or exacerbation of ß-cell autoimmunity.

b). Superantigens

Besides molecular mimicry, retroviral expression of superantigens (Sags) may be able to activate clonal expansion of autoreactive T cell clones. Superantigens have been implicated in the pathogenesis of the various autoimmune diseases (132, 133). Originally described as minor-lymphocyte stimulating antigens, retroviral Sags expressed by B cells interact with the development of T helper cells of both Th1 and Th2 subtypes in mice. This is of interest, since ß-cell destruction appears to be Th1 dependent. Further, a glycemia related increased expression of p73 core protein of intra-cisternal particles has been noted in genetically diabetic (db/db) and NOD/Lt mice. These mice spontaneously develop insulin autoantibodies (IAA) that cross-react with the p73 antigen, suggesting that a molecular mimicry mechanism could be involved (134), despite the fact that they have little linear structural homology. Another study in patients with IMD demonstrated that two thirds of IAA positive sera also reacted with p73 (135). Conrad et al (136) isolated a novel mouse mammary tumor virus-related human endogenous retrovirus (HERV), in patients suffering from acute onset type 1 diabetes. He termed them the HERV IDDMK1,2 22 subtype. They further showed that the N-terminal moiety of the envelope (env) gene encoded a MHC class II-dependent superantigen. They proposed that expression of this Sag, induced extrapancreatically and by professional antigen-presenting cells, could lead to ß-cell destruction via the systemic activation of autoreactive T cells. He further reported the selective expansion of Vß7+ T cells in the islet cell infiltrates from two patients with recent onset IMD was associated with extensive junctional diversity of Vß7+ T cell clones. These investigators demonstrated that islet cell membrane preparations preferentially expanded Vß7+ T cells from non-diabetic peripheral blood mononuclear cells (137). Others and we were however unable to confirm IMD specificity of the IDDMK1,2 22, since it was equally recoverable as viraemia from controls as well as patients (138). Furthermore, both patients and controls made antibodies to env proteins. The Sag effect reported also needs confirmation.

In order to establish molecular mimicry as a mechanism responsible for the autoimmune diseases it is important to identify the precise epitope that initiates the putative cross-reactive immune response. Additional complexity that has come into limelight due to various animal studies is that of epitope spreading (139). An increasing array of autoantigens or autoantigenic peptides reactive with autoantibodies develop over time. Both intramolecular and intermolecular epitope spreading has been described in NOD mice (134, 140). These studies demonstrated that T-cell responses in NOD mice expand in vivo against a defined group of islet cell antigens in an orderly sequential manner. These responses in the young NOD mice first show a strong reactivity to GAD enzyme and not to other islet cell antigens. Furthermore, the initial response to GAD is first limited to one region of the protein only. Gradually, this response spreads intramolecularly to involve other regions of the protein. Eventually, after the destructive islet cell inflammation (insulitis) as a result of autoimmunity to ß-cells, the T-cell responses spread intermolecularly to involve other islet cell proteins (e.g. heat shock protein 60, carboxypeptidase H and insulin) as well (141). This epitope spreading makes it difficult to predict which putative cross-reactions, if any, are important in terms of disease induction, and which do not give rise to autoimmune pathology, particularly in humans who are exposed to many infections.

DEFICIENCIES IN IMMUNOREGULATION IN TYPE 1 DIABETES

There is both evidence for and speculation about defective central and peripheral mechanisms of immunoregulation in the IMD form of type-1 diabetes.

Deletion of autoreactive T cells in the thymus, is one mechanism for the induction of tolerance to self antigens (central deletion). This may involve diminished expression of insulin in the thymus of susceptible individuals due to the presence of class I VNTR alleles 5' to the insulin gene as already discussed. Others have suggested that it is the ineffective antigenic binding of the IMD-prone HLA-DQ or -DR that promotes islet cell autoimmunity, since this permits autoreactive T cells to escape thymic ablation and pass into circulation.

In addition to clonal T cell deletion and anergy in thymus, peripheral regulatory T (Treg) cells are essential for the down regulation of T cell responses to both foreign and self antigens, and for the prevention of autoimmunity. Recently progress has been made in characterizing different subsets of Treg cells and their modes of action. These regulatory T cells can be broadly classified into two main categories i) NKT cells and ii) "resting" CD4+CD25+ suppressor T cells. This is not an exclusive list as other cell types such as CD8+ Treg cells (142, 143), gd T (144) cells have also been shown to have roles in maintaining immunological homeostasis.

Various studies including ours have identified defects in the peripheral Treg cells in IMD patients (145, 146) as well as in NOD mice affecting both NKT cells (147, 148) as well as CD4+CD25+ suppressor T cells (149). Since these Treg cells are not absent in either species, ways to stimulate them should be actively sought to provide novel therapies for the future. The possibility of future therapeutic use of Treg cells in human autoimmune diseases lies heavily on basic studies that are designed to elucidate the mechanisms of induction and function of these cells. Therapy with immunomodulatory compounds that specifically target endogenous pool of Treg cells can be envisioned. This approach requires a more detailed investigation into the intracellular and extracellular events that regulate the differentiation and expansion of these cells in vivo.

ENVIRONMENTAL FACTORS

Besides the familial predispositions, much evidence points to a major role of environmental factors in the disease pathogenesis. More than 60% of identical twins affected by type-1 diabetes are discordant for the disease and most of these lack islet cell autoantibodies. The disease frequency is on a steep rise in Western countries over the past 3 decades that cannot be explained by the accumulation of the susceptible genes. Africans, which dominate the tropics, and Chinese, both have low frequencies of the susceptible genes and low incidence rates of IMD (69), except where there has been a high rate of Caucasian genetic admixture. In the genetically susceptible Arabian Gulf States, the incidence of type-1 diabetes has greatly risen with the life styles borne from increasing affluence and a decrease in infectious and parasitic diseases (NM personal communication). More persuasively, migrants from countries with low hygiene and low incidence rates of IMD to countries with high hygiene and high incidence become as susceptible as the natives within a generation (150). Lastly a number of non-antigen specific interventions can greatly affect the diabetes rate in the rodent models of the disease (151). Animals reared in sterile environments have early onsets and increased frequencies of diabetes while those infected with a variety of micro-organisms and parasites become protected (152-156). Taken together, all of the foregoing point to a strong relationship between prevailing level of community hygiene especially with respect to drinking water and the dramatic increase in the incidence of autoimmune diseases such as type-1 diabetes in the modern world.

ROLE OF DIET

Despite persuasive epidemiological evidence for environmental factors that precipitate type 1 diabetes in genetically susceptible individuals, their identity remains speculative. This may be due to long period between exposure and the onset of hyperglycemia, the complex genetics of the disease, and the likely multiple insults of perhaps different derivation involved in the initiation of the insulitis and subsequent ß cell destruction. Dietary habits such as consumption of dairy products and early weaning of infants, and dietary toxins such as nitrates and nitrites have been associated with this autoimmune disease (157, 158).

Close correlations between per capita consumption of unfermented milk proteins and the incidence of diabetes between countries (159-161) and within a country have been reported (162). The claimed negative association between diabetes incidence and a high frequency and long duration of breast-feeding is more controversial (157) and has not been confirmed by reports from Germany (163) and the United States. Several studies have found associations between the consumption of foods rich in nitrates and nitrites, which is reduced to nitrite in the gut, and the occurrence of type 1 diabetes (164-166). The active species is believed to be N-Nitroso compounds that can be formed from the reaction of nitrite with amines (167). These findings are of particular interest as N-Nitroso compounds are structurally related to streptozotocin, a ß cell-specific toxin that can induce diabetes in rodents (168).

The incidence of type 1 diabetes varies worldwide according to dietary patterns. In-depth exploration of dietary risk factors during pregnancy and early neonatal life is warranted to confirm whether and to what extent diet cooperates with genetic susceptibility in the early onset of type 1 diabetes.

SCREENING METHODS AND PREVENTION OF T1DM

T1DM is by far the most common metabolic disease of childhood and adolescence and its prevalence and incidence has been increasing worldwide (244). This increase of incidence is the highest among the children under 5 years of age over the past decade (242). Prevention of T1DM would constitute a major advance in the lives of pre-diabetic individuals and significantly relieve a major current and predicted burden on both the individual and the health care system. Identifying individuals at risk developing the disease and the prevention of the disease progression are two important steps before the onset of disease. The presence of islet autoantibodies, the genetic predisposition with specific HLA haplotypes are known risk factors associated with the development of diabetes. Most studies have been carried out on first-degree relatives of T1DM patients who have 15-fold increased risk of the developing diabetes in comparison to general population. However, approximately 85% of all patients developed T1DM does not have an affected family member. Therefore it is crucial to establish a standardized screening methods which will efficiently identify individuals at high risk in a general population. We screened school children between 5-18 years of age for 6-12 years to evaluate the predictive value of autoantibodies (245). Our study indicated that the risk of developing T1DM when ICA is detected in the absence of other autoantibodies is low, whereas with more than one autoantibody (either GAD65A, IAA, IA-2A or IA-2ßA) the risk of developing T1DM in a general population is high. Similar findings were also reported in other studies (246,247,248). These results support the value of multiple autoantibodies as good predictive markers for T1DM not only in first degree relatives but also in the general population.

Figure 2. Autoantibodies in ICS positive subjects. Shaded areas show the subjects who progressed type 1 diabetes in the presence of the autoantibodies (245).

Type-1A T1DM Prevention Trial

The elucidation of the natural history of pre-diabetes has allowed us to characterize, through the use of genetic, immunologic and metabolic markers, those individuals at greatest risk for developing T1DM, especially the type-1A form. This predictive ability has become possible in both high-risk relatives and the general population. The autoimmune destruction of ß-cells in the pancreas may last from a few months to several years. This pre-diabetic period have allowed investigators to test prevention strategies which mainly have focused in modulation of autoimmune process. A number of studies were initiated with general immunosuppressive agents, such as cyclosporin-A, (171-173) azathioprine and prednisone (174) in patients with new clinical onset T1DM, with positive results in that insulin free remission rates were increased and endogenous insulin (C-peptide) reserves were improved. However, despite continued immunotherapy with the attendant risks of renal damage and lymphomas at higher doses, relapses proved to be the rule and such treatments were abandoned. Cyclosporin given at a prediabetic phase of the disease delayed but did not prevent diabetes (175, 176).

With the observation that nicotinamide prevents pancreatic ß cell destruction from streptozotocin by raising otherwise depleted levels of islet cell NAD as a result of superoxide induced DNA breaks and repair, the vitamin was subjected to a large European and Canadian trial called The European Nicotinamide Diabetes Intervention Trial (ENDIT). However, nicotinamide failed to prevent progression to diabetes (238). One German study (DENIS) has been completed without any effect of nicotinamide on prevention of IMD (177). Other antigen based human trials have been built upon studies carried out in NOD mice. Daily doses of long acting insulin in NOD mice prevent their diabetes and strongly inhibit ß cell autoimmunity (178). The mechanism in part was due to ß cell rest with reduced islet cell expression of the autoantigens, however immunological effects are also likely. Use of close control after diagnosis by intravenous insulin in patients also gave relative remissions (179) and the combination of daily subcutaneous long acting insulin plus intermittent intravenous insulin appeared promising in at risk patients (180).

Thus the high risk arm (> 50%) based upon ICA plus FPIR below the first percentile, of the US multi-centered type-1 diabetes prevention trial (DPT-1) was born using such a combination. Daily doses of insulin at 0.25units/Kg/day were used in the trial, which could not show protection from diabetes nor a delay in diabetes onset with this dose (181). Oral feedings of insulin to NOD mice also delay their onsets of diabetes (182) as do oral feedings of GAD, whereas feedings of the combination had the best effects (183). Oral feedings of 7.5mgs recombinant, regular insulin per day is being given to relatives with a moderate (20-50%) risk of IMD based upon ICA plus IAA in the face of a normal FPIR.

This trial again was without benefit albeit those with high levels of IAA may have been benefited. One trial of oral insulin at 10mgs a day in newly diagnosed patients appeared to be ineffective in those who were diagnosed at < 20 years of age, but may have had improvement in endogenous insulin reserves after the initial 6 months (184). Smaller trials with oral insulin at different doses (2.5 mg/7.5 mg and 5 mg daily in newly diagnosed diabetic patients) performed in Europe also did not show any benefit of oral insulin therapy in the prevention of diabetes (239, 240). In our study though, we were able to show delay in progression of ß-cell failure by the administration of oral insulin (1 mg) in newly diagnosed patients over 20 years age (241).

It is clear that many interventions in NOD mice are capable of reducing the frequency and severity of their diabetes. This provides considerable encouragement that the situation may be no different in humans. The relatively low concordance rates in identical twins affected by T1DM give support for this view. Immunizations of insulin, insulin B chain and the 8-20 amino acid peptide in incomplete Freund's adjuvant (IFA) all dramatically reduced the diabetes rate in NOD mice (183, 185, 186), as did the B chain peptide when given with alum or the childhood DPT vaccine. The protection was transferable to irradiated NOD recipients by CD4+ T cells. A human trial using this approach is about to begin in newly diagnosed patients (187). Interestingly, the 8-23 amino acid insulin B chain peptide was the predominant epitope of T cell clones raised from the insulitis lesions of NOD mice and nasal inhalations of the peptide were found to also reduce the NOD diabetes rate (188). Two human trials of intranasal insulins are ongoing in Australia and Finland [Harrison and Simell: ADA abstracts/personal communication]. Further two peptides within this peptide were found to be targeted by CD8+ T cells, which were capable of transferring diabetes to NOD-scid mice (189). Immunizations by human GAD65 peptides or IA-2 were minimally or not effective in NOD mice (190). Another preliminary study using non-activating humanized anti-CD3 monoclonal antibody in new onset type-1 diabetes mellitus has recently been completed (191). The treatment has shown to mitigate the deterioration in insulin production and improved metabolic control during the first year of treatment. While the study requires confirmation, long-term benefits from the treatment are yet to be shown. Currently, the authors are aware of several other approaches that are in various stages of development for human trials. These include immunizations by GAD65 peptides and B25-Asp insulin, use of altered peptide ligands (APL) of insulin and GA₆₅, GAD₆₅ peptides in soluble class II MHC, oral insulin linked to cholera B chain carriers, immunizations by peptide 277 of the heat shock protein (hsp60), selective IL-4 agonists, humanized anti-CD3 monoclonal antibodies, anti-CD40L or CD80, and COX-2 inhibitors. Peptide 277 of the heat shock protein (hsp60) has been shown to prevent and reverse diabetes in NOD and streptozocin-induced diabetes mice (244). This rapidly advancing field will bear watching in the near future.

TREATMENT OF TYPE 1 DIABETES

Despite advances in technology, the management of diabetes is still cumbersome and insulin replacement not physiological. In regard to the latter, continuous glucose sensing is being developed but is not yet practically available. Insulin responses to changes in blood glucose are thus based upon intermittent blood glucose testing and corrections. Further insulin is given into the systemic circulation whereas endogenous insulin is secreted into the portal vein. However, it is crucial to normalize glucose levels in order to prevent long-term consequences of diabetes especially from micro-vasculopathies, leading to neuropathy, renal failure, and blindness. In 1993, The Diabetes Control and Complications Trial (DCCT) reported results demonstrating that the intensive therapy of T1DM reduces the risk of development and progression of micro-vascular complications. Furthermore, these benefits outweighed the increased risk of hypoglycemia that accompanied intensive diabetes therapy (169). Thereafter, The Epidemiology of Diabetes Interventions and Complications (EDIC) study assessed whether these benefits persisted after the end of DCCT. The findings of this study provide further support for the DCCT recommendation that most adolescents with T1DM receive intensive therapy aimed at achieving glycemic control as close to normal as possible to reduce the risk of microvascular complications (170). This is achieved increasingly well by combinations of newer long and short acting insulin (multidose insulin or MDI) and at best by insulin pumps.

Clinical Management

The clinical management for patients with type 1 diabetes concerns adequate insulin replacement matched to food intake as modified by exercise. Insulin is required throughout the whole day to prevent development of a starvation state. This is the basal insulin requirement. Glargine (Lantus) insulin provides day-long basal insulin without significant peaks of action. In some children however, Glargine may not last 24 hours and for this reason is usually given at night. This synthetic insulin cannot be safety mixed with other insulins in the same syringe due to pH incompatibilities. Throughout the day, short acting insulin such as Humalog (Lispro) is given to normalize blood glucose levels and cover carbohydrates consumed during meals. As 3 meals are eaten by most on a daily basis, short acting insulins should be given at least 3 times daily to prevent excessive hyperglycemic excursions. The dose depends upon the level of glycemia before the meal. The difference between the blood glucose found and the target of 120 is used to calculate a correction bolus dose. This may range from 1 unit for 10-100mgs/dl blood glucose depending upon age and body size. Next a meal bolus is calculated from an estimation of the carbohydrate content in gms and an individual factor relating insulin dosage to the amount of carbohydrate to be consumed. The range is from 1 unit per 10-50 gms. For infants, the intermittent, short acting insulin may be given after the meal when food consumption has been found to be unpredictable. In addition to the 3 meals, additional amounts of short acting insulin may be taken to cover snacks, and to reduce blood glucoses as necessary at bedtime. The burden of many injections of insulin each day can be reduced by use of an insulin "pen", which is a convenient way of carrying multiple doses in a single dispenser.

An alternative and better method is through the use of continuous subcutaneous insulin infusion (CSII) through use of an insulin pump. In this case, only short acting insulin e.g. Humalog or Novolog is taken as a continuous basal infusion and multiple boluses of insulin are given as above. The infusion site is best changed every 2 days to avoid skin infections. The advantages of CSII are that insulin is taken only when needed, and not in an anticipatory fashion as with long acting insulins. Insulin boluses are only taken, as needed, no special diets as required, hypoglycemic episodes are minimal and the system is convenient and portable. The one down side of CSII is that since only short acting insulin (Humalog effects are gone within 3 hours) are taken, any blockage or pump failure can lead to rapid onset of hyperglycemia and an uncontrolled diabetic state.

Besides insulin replacement therapy for type-1 diabetes, co-existing hypertension, dyslipidemia should be aggressively treated. Microalbuminuria and/or hypertension should be a call for use of angiotensin converting enzyme (ACE) inhibitors to minimize progression to chronic glomerulosclerotic damage. ACE inhibitors may induce angio-edema and produce a troublesome dry cough. Poorly controlled diabetes induces rise in VLDL and triglyceride levels. When severe or chronic, pancreatitis may be induced. Diet reduced in animal fat and administration of fibrates (e.g. gemfibrozil) should be given to combat hypertriglyceridemia. Co-existing Hashimoto's thyroiditis should be sought through thyroid autoantibody analyses, and hypothyroidism when identified treated by thyroid hormone replacement. Addison's disease, celiac disease and atrophic gastritis/pernicious anemia sought always be considered in patients with type-1 diabetes and if found, treated accordingly.

IMD Prevention Trials

A number of studies were initiated with general immunosuppressive agents, such as cyclosporin-A, (171-173) azathioprine and prednisone (174) in patients with new clinical onset IMD, with positive results in that insulin free remission rates were increased and endogenous insulin (C-peptide) reserves were improved. However, despite continued immunotherapy with the attendant risks of renal damage and lymphomas at higher doses, relapses proved to be the rule and such treatments were abandoned. Cyclosporin given at a prediabetic phase of the disease delayed but did not prevent diabetes (175, 176). With the observation that nicotinamide prevents pancreatic ß cell destruction from streptozotocin by raising otherwise depleted levels of islet cell NAD as a result of superoxide induced DNA breaks and repair, the vitamin was subjected to a large European and Canadian trial (ENDIT). The results are pending at the time of writing. One German study (DENIS) has been completed without any effect of nicotinamide on prevention of IMD (177). Other antigen based human trials have been built upon studies carried out in NOD mice. Daily doses of long acting insulin in NOD mice prevent their diabetes and strongly inhibit ß cell autoimmunity (178). The mechanism in part was due to ß cell rest with reduced islet cell expression of the autoantigens, however immunological effects are also likely. Use of close control after diagnosis by intravenous insulin in patients also gave relative remissions (179) and the combination of daily subcutaneous long acting insulin plus intermittent intravenous insulin appeared promising in at risk patients (180).

Thus the high risk arm (> 50%) based upon ICA plus FPIR below the first percentile, of the US multi-centered type-1 diabetes trial (DPT-1) was born using such a combination. Daily doses of insulin at 0.25units/Kg/day were used in the trial, which could not show protection from diabetes nor a delay in diabetes onset with this dose (181). Oral feedings of insulin to NOD mice also delay their onsets of diabetes (182) as do oral feedings of GAD, whereas feedings of the combination had the best effects (183). Oral feedings of 7.5mgs recombinant, regular insulin per day is being given to relatives with a moderate (20-50%) risk of IMD based upon ICA plus IAA in the face of a normal FPIR. One trial of oral insulin at 10mgs a day in newly diagnosed patients appeared to be ineffective in those who were diagnosed at < 20 years of age, but may have had improvement in endogenous insulin reserves after the initial 6 months (184). It is clear that many interventions in NOD mice are capable of improving the diabetes rates. This provides considerable encouragement that the situation may be no different in humans. The relatively low concordance rates in identical twins affected by IMD give support for this view. Immunizations of insulin, insulin B chain and the 8-20 amino acid peptide in incomplete Freund's adjuvant (IFA) all dramatically reduced the diabetes rate in NOD mice (183, 185, 186), as did the B chain peptide when given with alum or the childhood DPT vaccine. The protection was transferable to irradiated NOD recipients by CD4+ T cells. A human trial using this approach is about to begin in newly diagnosed patients (187). Interestingly, the 8-23 amino acid insulin B chain peptide was the predominant epitope of T cell clones raised from the insulitis lesions of NOD mice and nasal inhalations of the peptide were found to also reduce the NOD diabetes rate (188). Two human trials of intranasal insulins are ongoing in Australia and Finland [Harrison and Semil: ADA abstracts/personal communication]. Further two peptides within this peptide were found to be targeted by CD8+ T cells, which were capable of transferring diabetes to NOD-scid mice (189). Immunizations by human GAD₆₅ peptides or IA-2 were minimally or not effective in NOD mice (190). Another preliminary study using non-activating humanized anti-CD3 monoclonal antibody in new onset type-1 diabetes mellitus has recently been completed (191). The treatment has shown to mitigate the deterioration in insulin production and improved metabolic control during the first year of treatment. While the study requires confirmation, long-term benefits from the treatment are yet to be shown. Currently, the authors are aware of several other approaches that are in various stages of development for human trials. These include immunizations by GAD₆₅ peptides and B25-Asp insulin, use of altered peptide ligands (APL) of insulin and GAD₆₅, GAD₆₅ peptides in soluble class II MHC, oral insulin linked to cholera B chain carriers, immunizations by peptide 277 of the heat shock protein (hsp60), selective IL-4 agonists, anti-CD40L or CD80, and COX-2 inhibitors. This rapidly advancing field will bear watching in the near future.

Transplantation of the Pancreas

Clinical pancreas transplantations have been attempted for years without success until recently. For example, such transplants begun at the University of Minnesota in 1966 were initially associated with a high failure rate, but the outcomes have improved in parallel with other organ transplants as surgical techniques and immunosuppressive protocols have evolved. This procedure has been performed in over 15,000 patients worldwide with a success rate (insulin independence) of 85% at 1 year (192). An alternative to the invasiveness of whole pancreas transplantation is the possibility of reversing the diabetes state by islet cell transplantation/infusion. Research into islet cell replacement has been in progress for more than three decades and a total of 493 adult transplants have been performed worldwide. Data retrieved from International Islet Transplant Registry document insulin independence for more than 1 month in only 12.5% and for more than 1 year in only 8% of recipients reported to date (193). However a major commitment from the US Government to support allogeneic islet cell transplantation is based on the success reported by the group from the University of Alberta in Edmonton, Canada (194, 195). The major advances that have contributed to the success of the so-called "Edmonton Protocol" include improved immunosuppressive regimens that avoid steroid usage (namely, a combination of sirolimus, low dose tacrolimus and anti-IL-2 receptor antibody, daclizumab), and advanced organ procurement, the use of islets from more than one pancreas at a time and the islet isolation techniques used, particularly the availability of defined collagenase blends (e.g. liberase) (194, 196).

The lack of sufficient human donor pancreases for islet cell transplantation has led, inevitably to the search for alternative methods. These include xenotransplantation (197), ß-cell induction from pancreatic duct cells (198, 199), fetal islet precursor cells (200) and fetal pancreatic stem cells (201) and ß-cell engineering (202, 203). However, it goes without saying that several major hurdles remain before stem cell therapy can be applied for human use.

In an attempt to devise gene therapy for IMD, Lee et al. (204) used a recombinant adeno-associated virus that expresses a single-chain insulin analog (SIA), which possesses biologically active insulin activity without enzymatic conversion, under the control of hepatocyte-specific L-type pyruvate kinase (LPK) promoter, which regulates the SIA expression in response to blood glucose levels. They demonstrated that the SIA produced from the gene construct recombinant AAV-LPK-SIA caused remission of diabetes in streptozotocin-induced diabetic rats and autoimmune diabetic mice for up to 8 months without any apparent side effects. The rats or mice were normoglycemic within 7 days of treatment with recombinant adeno-associated virus. Lee et al. (204) concluded that this new single-chain insulin analog gene therapy might have potential therapeutic value for the cure of autoimmune diabetes in humans.

Cheung et al. (205) found that gut K cells could be induced to produce human insulin by providing the cells with the human insulin gene linked to the 5' regulatory region of the gene encoding glucose-dependent insulinotropic polypeptide. Mice expressing this transgene produced human insulin specifically in gut K cells. This insulin protected the mice from developing diabetes and maintained glucose tolerance after destruction of the native insulin-producing ß-cells.

INSULIN RESISTANCE SYNDROME (IRS)

This syndrome complex is centered upon genetic predispositions to insulin resistance and the hyperinsulinism that results from it. This medical state is also named syndrome X and the metabolic syndrome, however the descriptive term insulin resistance syndrome (IRS) is the one increasingly used in the literature (30, 206). In IRS, there are poorly understood genetic lesions that lead to insulin resistance from early life if not during embryogenesis. In many affected families, the disease occurrences suggest a dominant mode of transmission. In rare families, mutations affecting insulin receptors, or peroxisome proliferators-gamma (PPAR-g) expression may be the cause of it (207). IRS is the association of insulin and leptin resistance with obesity (typically with increased visceral fat), functional adrenal hyperandrogenism, functional ovarian hyperandrogenism, hypersecretion of pituitary LH, dyslipidemia, hypertension, and features of hyperinsulinism such as late reactive hypoglycemia and acanthosis nigricans. When the compensation by increased insulin secretion fails, glucose intolerance and type-2 diabetes result.

a). Natural history of IRS (Tables 2 and 3)

Several studies indicate that children and adults with type-2 diabetes were born small for gestational age. This suggests that the insulin resistant state existed in-utero since it is insulin rather than pituitary growth hormone that is the principal growth-promoting hormone of the unborn child, and decreased insulin action might be anticipated to impair embryonic growth. After birth, premature pubarche resulting from excessive adrenal androgens such as dihydroepiandrosterone (DHEA) may be seen, even before obesity has appeared. Excessive DHEA may be seen best after ACTH injection leading to a clinical suspicion that the 3ß hydroxysteroid dehydrogenase is underactive. Obesity can begin from infancy but often dates from about 8 years of age when physiological pubarche occurs. Acanthosis nigricans thought to result from hyperinsulin stimulation of insulin-like growth factor 1 (IGF-1) receptors often dates from this time. Menarche may be delayed in age at onset or menses may be missed after menarche, or else there can be dysfunctional bleeding resulting from anovulatory cycles. Hirsutism often becomes bothersome during adolescence, as may male pattern hair thinning, persistent acne and development of polycystic ovaries. An increase in very low density lipoprotein (VLDL) secretions by the liver are seen with increasing age, associated with diminished, atherogenesis protective, high density lipoprotein bearing cholesterol (HDL-chol), a dyslipidemic profile that promotes early and progressive onset of atherosclerosis, predisposing to coronary heart disease (CHD), stroke and peripheral vascular diseases in later life. The latter problems are compounded by the appearance of hypertension and type-2 diabetes. The glucose intolerance that precedes type-2 diabetes often first involves post-prandial glucose levels or the two-hour time point of the OGTT as discussed above, but later induces a rise in fasting glucose (impaired fasting glucose) levels as well. The mechanism is thought to be ß cell exhaustion or more likely a glucosamine and lipid mediated islet cell toxicity. Once this stage is reached, damage to the islets can become irreversible, resulting in the dual problems of insulin resistance and insulinopenia, both of which need to be addressed in therapeutic strategies.

b). Underlying mechanisms of IRS

Obesity: Affected patients commonly show polyphagia, and have voracious appetites that are characteristically resistant to dietary advice. When leptin deficiency was discovered in Ob/Ob mice and leptin receptor deficiency discovered in Db/Db mice, the fat cell became to be appreciated as an endocrine cell rather than one that was an inert repository of triglycerides. However the promise of a breakthrough in the understanding of human obesity was quickly dissipated when such lesions proved to be rare in humans. Obese patients with their greater degrees of adiposity also have the highest levels of leptin as expected, however these high levels do not reduce the appetites of IRS patients. Thus such patients are leptin resistant. Early trials of leptin therapy have not affected weight loss. However patients with lipodystrophy who have leptin deficiency develop insulin resistance, hyper-insulinism, dyslipidemia and type-2 diabetes, all of which respond dramatically to leptin given as therapy (208). Deficiencies in other appetite suppressing hormones such as resistin have more recently been implicated but not yet shown to have clinical relevance. Hyperinsulinism itself is a compounding variable, in that excessive carbohydrate containing diets stimulate the highest levels of insulin and the greatest degrees of adiposity. In our experience, therapies such as metformin that improve insulin sensitivity, can dramatically lower weight in children with IRS. Whatever the explanation of obesity in IRS, the more a subject gains in weight, the more insulin resistance becomes expressed.

Hyperandrogenism: It is uncertain whether the pituitary lesion (increased LH secretion) leads to the androgenic excess or visa versa. Probably, both are responses to the insulin resistance and hyperinsulinism of IRS by mechanisms that have yet to be clearly understood. Androgens of ovarian origins usually predominate over those of the adrenal gland. Sex hormone binding globulins in the circulation are often low, resulting in increased free androgens with their increased bio-availability (209). Such is often seen with testosterone, which can be raised or normal in hirsute girls when increased free testosterone levels are common. The mechanism whereby insulin resistance and hyperinsulinism can lower levels of hormonal binding globulins also needs clarification. Interestingly, we hold that there is a clinical overlap between Cushing's syndrome and IRS (210). Both tend to have visceral (central) obesity and striae suggesting gluco-corticoid excess. However, whereas the patient with Cushing's syndrome has high levels of serum cortisol, the patient with IRS has low normal levels, albeit both have increased levels of urinary free cortisol. Again, the explanation may lie in the low levels of corticosteroid binding globulins found in IRS where circulating cortisol is disproportionately free. Some investigators have suggested that there is an impaired conversion of cortisol to the metabolically inactive cortisone in IRS. Further, the child with Cushing's syndrome is invariably growth retarded in contrast to the child with IRS where growth is excessive. In IRS, the GH levels are suppressed making diagnosis of GH deficiency in IRS by stimulation testing difficult. Again, IGFBP-1 is depressed, giving an excessive free IGF-1 level, albeit the total level of IGF-1 is usually normal. A severely affected child with IRS may develop pseudo-acromegaly through this mechanism. Thus increased affective IGF-1 may explain the excessive growth, even when there are strongly suggestive features of corticosteroid excess.

Acanthosis nigricans: Stimulation of the IGF-1 receptors of skin kerotinocytes by high levels of circulating insulin is thought to explain their hyperplasia and excessive laying down of keratin in the skin of the neck, axillae, elbows and knees, skin creases and indeed most areas of skin (211). In addition, excessive free IGF-1 may have the same effect, albeit the greater the degree of insulin resistance, the higher the insulin levels, the more striking the acanthosis nigricans. We have shown among our patients that increased bioavailabilty of IGF-1 (high IGF-1 and low IGFBP-1) are directly correlated with the severity of acanthosis nigricans

Glucose intolerance and type-2 diabetes: Children and young adults affected by IRS are often hyperinsulinemic. In such persons, stimulation of insulin secretion by carbohydrates alone or with protein can induce a delayed but excessive rise in insulin secretion, causing early excessive rise in glucose followed by an excessive fall in glucose level 3-4 hours afterwards, of sufficient severity to provoke symptoms of hypoglycemia. As the ability to secrete insulin declines, glucose intolerance appears followed by 2-hour criteria on OGTT for diabetes with impaired fasting hypoglycemia followed by fasting criteria for diabetes.

Non-alcoholic steatohepatitis (NASH): It is also known as fatty liver or hepatic steatosis. The incidence of fatty liver among obese children is 2.6% (251), and hyperinsulinemia was found to be the major contributor for its’ development (252). A number of factors may play a role in the development of fatty liver including, induction of cytochrome P4502E1 during obesity, which is capable of generating free radicals, high dietary intake of polyunsaturated fatty acids or low intake of nutritional antioxidants contributes to the oxidative stress. Fatty liver alone appears to be a relatively benign disease can be reversible. However, it may progress over years to hepatic cirrhosis, liver failure or hepatocellular carcinoma. The onset of disease is usually insidious. Laboratory evaluation indicates mild to moderate elevation of serum aminotransferases in most children and serum alanine aminotransferase (ALT) levels had been shown a useful screening for fatty liver in obese children (249). The ratio of aspartate aminotransferase (AST) to ALT is usually less than 1, but this ratio increases as fibrosis advances. Serum aminotransferases, alkaline phosphatase and gamma glutamyltransferase (GGT) levels are proposed surrogate markers of fatty liver (250).

Renal involvement: A form of focal glomerulosclerosis appears to be associated with IRS, whereas hypertension is commonly so, especially with age. The mechanisms responsible have not been elucidated.

Inflammation: IRS and type-2 diabetes have increased markers of inflammation. This takes the form of increased levels of C-reactive protein, raised erythrocyte sedimentation rates (ESR) and increased cytokine (TNF-a) levels. Further, there is an increased incidence of autoimmune thyroiditis also. This is clinically complex, since thyroid binding globulin levels in IRS are often depressed, leading to confusion about the presence of hypothyroidism. Obese patients are thus often unnecessarily treated for hypothyroidism they do not have. They may however develop true hypothyroidism on the basis of associated Hashimoto's disease.

c). Treatment

The child with IRS should be aggressively treated to prevent the burgeoning complications of the condition. The approaches should include an exercise program such as walking or swimming for 30-40 minutes most days of the week, since at the level of the muscle, exercise provokes glucose entry into muscle without the involvement of insulin. Carbohydrate restriction is the key to reduce weight. However, where there is also an increased level of triglycerides, restriction of animal fats should be imposed. Metformin is approved for the treatment of type-2 diabetes in children, but is also the drug of choice for IRS. Some have suggested that it is the gastro-intestinal side effects of the drug that accounts for much of its' action. However the drug is effective in type-2 diabetes without weight loss, being found to reduce hepatic glucose output. The PPAR-g agonists are effective at insulin sensitization but are less useful in supporting weight loss. Oral contraceptive agents are often given in IRS where there is evidence of hyperandrogenisation, where they may have an effect because of their counteractions to androgens. However they also raise the level of hormonal binding globulins . Combined estrogen/progesterone therapy in a prepubertal patient raises the risk of premature closure of the epiphyses with loss of adult height. They also promote thrombosis and mitigate against weight loss. The goal of therapy in IRS should be to achieve ideal body mass index (kg/m2) for age.

TYPE-2 DIABETES

As the US passes into the 21st century, the epidemic of obesity and type-2 diabetes continues unabated. Perhaps under pressure of commercial interests, we have learned to eat too fast and much as a nation. For those with the energy conserving "thrifty" genes of IRS, this excess of food and especially of the insulin provoking carbohydrates, the high food consumption leads to obesity, the IRS phenotype and type-2 diabetes. At present in some US state where there are large numbers of ethnic groups prone to IRS and type-2 diabetes (Hispanics, American Indians, Asian Indians, African Americans), the numbers of children with type-2 diabetes is beginning to rival if not surpass the number with type-1 diabetes. The natural history of progression to type-2 diabetes is that a person with IRS begins to decompensate, with a fall in the disposition index. Subsequently, fed levels of blood glucose rise, followed by elevations in fasting blood glucose levels rising later. At this early stage, diet, exercise and insulin sensitisers are indicated. Latter still, the ability to secrete insulin becomes disabled, requiring the addition of insulin secretagogs to the therapeutic plan. When these additional agents also fail to maintain near normoglycemia, then insulin should be given instead of the secretagogs. The typical dyslipidemia associated with IRS and type-2 diabetes should be treated by reduced intake of animal fat and a fibrate such as gemfibrozil. However some patients have prominent elevations in LDL-cholesterol and should be treated by a statin. The mixed use of a statin and a fibrate should not be undertaken since the risks of muscle necrosis with renal failure is a greater risk than is the risk inherent in the hyperlipemic state. Hypertension, when present, should be aggressively treated.

SUMMARY

In the above chapter, we have reviewed type-1 diabetes, with particular emphasis on the most common immune mediated (IMD) form. Whereas type-2 diabetes appears to be an increasing price paid for societal affluence, there is also evidence worldwide of a rising tide of type-1 diabetes. The increase in understanding of the pathogenesis of type 1 diabetes mellitus has made it possible to consider interventions to slow the autoimmune disease process in an attempt to delay or even prevent the onset of hyperglycemia. Although the prevention of type 1 diabetes is still at the stage of research trials, the trials are often mentioned in the lay press. The results of antigenic immunostimulatory studies in NOD mice hold great promise for similar beneficial effects in humans who have just begun to develop the clinical disease or are at high risk for IMD. Current investigations will determine if antigen based therapies can in fact abrogate ongoing autoimmunity via immunostimulation and ultimately prevent diabetes in humans without the risks of general immunosuppression.