Type 1 (Insulin-Dependent) Diabetes Mellitus
Etiology, Pathogenesis, Prediction, and Prevention
Type 1 (insulin-dependent) diabetes mellitus (T1DM) is associated with several immune abnormalities. Classically, when age at onset is young, classification of the disease is straightforward. Classification may be changing in the current obesity epidemic among the young, making differentiation between T1DM and type 2 diabetes mellitus (T2DM) more difficult.1,2 Although it has been known for decades that diabetes mellitus can occur in various degrees of severity, it was not until approximately 40 years ago that evidence was presented that indicated different modes of inheritance for what were then classified as “maturity-onset” and “juvenile-onset” types of diabetes.3 It is now evident that T1DM may occur at any age. The fact that T1DM in adults may fulfill clinical criteria of T2DM demonstrates the limitation of a disease classification based on clinical symptoms rather than on the etiology and pathogenesis. Despite the fact that autoantibodies are the main markers that predict and differentiate T1DM from other forms of diabetes, these markers are not yet included among the diagnostic criteria of T1DM.4
The diagnostic criteria used to differentiate between T1DM (insulin-dependent) and T2DM (non-insulin-dependent) are important for understanding the cause and pathogenesis of these two disease entities. The diagnostic criteria are primarily recommendations for the classification, the aim of which is better understanding of the cause of diabetes and optimization of care of diabetic patients. By using molecular genetics, it has been possible to clarify fully several monogenetic forms of diabetes. The etiology of these diabetes phenotypes has been explained by mutations in the genes for insulin, insulin receptor, glucokinase (maturity-onset diabetes in the young; MODY-2), HNF-4α (MODY-1), or HNF-1α (MODY-3).5 The clarification of the etiology of these diabetes phenotypes is a major advancement that has increased the diagnostic precision of diabetes syndromes.4,5 The understanding of the more complex and multifactorial T1DM syndrome also has undergone significant advances. This chapter is an attempt to present an evidence-based medicine review of the etiology, pathogenesis, and natural history of T1DM.
History
Early histologic studies of pancreatic tissue of diabetic patients who died shortly after clinical onset revealed that the pancreatic islets were altered by fibrosis, hyalinosis, atrophy, and infiltration of inflammatory cells.6 The inflammatory lesion of the islets of Langerhans was later described as insulitis,7 and quantitative studies of the pancreatic islets showed a specific loss of insulin-producing cells in association with clinical onset of T1DM.8,9 The rediscovery of insulitis9 was of major significance, especially because it was later observed that autoimmune thyroid disease often developed in diabetic patients treated with insulin or, conversely, that patients with diseases of autoimmune character (e.g., Grave’s disease, Hashimoto’s thyroiditis, pernicious anemia, and Addison’s disease) had an increased prevalence of T1DM.10 It was therefore suggested that the pathogenesis of T1DM involved autoimmune reactions directed toward the endocrine pancreas. This notion was supported by leukocyte migration inhibition to pancreatic islet antigens.11 Numerous studies have confirmed the presence of insulitis12,13; however, in spite of the assumption that T1DM is a T-cell mediated disease, reproducible and standardized tests of blood T-cell reactivity against islet autoantigens are yet to be established.14 Long-sought-for islet cell antibodies (ICA) were described in 1974,15,16 islet surface antibodies (ICSA) in 1978,17 and complement-dependent antibody-mediated islet cell cytotoxicity in 1980.18 The first antigen recognized by islet antibodies, an islet protein of 64,000 relative molecular mass (Mr), or 64K, was described in 1982.19,20 Later, the 64K protein was found to have glutamic acid decarboxylase (GAD) activity,20 but molecular cloning showed that the human islet GAD was a novel isoform, GAD65.21 Autoantibodies to insulin were demonstrated in 1983,22 to insulinoma-associated antigen 2 (IA-2), a receptor-type protein tyrosine phosphatase, in 1994,23 and to the ZnT8 transporter in 2007.24 In contrast to T-cell analyses, GAD65 and IA-2 autoantibodies may be determined in standardized assays.25 Taken together, ample evidence exists that islet cell autoimmunity is of major importance in the etiopathogenesis of T1DM.
Definition
T1DM results from selective destruction of the β cells in the pancreatic islets, leading to absolute lack of insulin and hyperglycemia.4 The β cells are principally destroyed by an aggressive autoimmune process, which is mediated but not limited to infiltration of CD4+ and CD8+ T cells, as well as macrophages, resulting in insulitis. An improved understanding of these processes is obtained in animals with spontaneous T1DM, most prominently the NOD mouse and the BB rat. These animals allow detailed genetic and experimental observations prior to onset of diabetes and facilitate preclinical trials to develop approaches of prevention and intervention, as recently reviewed.26
Symptoms and Signs
T1DM is often thought to be a disorder of acute onset, and the clinical onset may be dramatic. Nevertheless, T1DM may be discovered accidentally27 or associated with serious, life-threatening diabetic ketoacidosis (DKA). Over the years, however, numerous reports have noted that signs of subclinical diabetes precede the clinical onset. In addition, in adult diabetes patients classified with T2DM, a change sometimes occurs from an insulin-independent to insulin-dependent state.
T1DM is said to have four major clinical phases: preclinical diabetes, overt diabetes, partial remission phase (honeymoon), and chronic phase of lifelong dependency on injected insulin.28 It is now accepted that autoantibodies against GAD65, insulin, IA-2, or ZnT8, alone or in combination, may be present up to several years before the clinical onset of the disease.24,29,30 The possibility that islet autoimmunity might be present long before symptoms of hyperglycemia occur makes it difficult to define a causative factor. The clinical onset is not likely to occur until a major loss (80% to 90%) of the islet β cells has occurred. Currently it is not possible to determine the human β-cell mass or volume. β-Cell function tests are affected by both β-cell mass and insulin sensitivity.31 It is therefore not possible to relate clinical onset to β-cell mass, especially since some reports suggest that around 40% to 50% of β cells may be viable, owing to β-cell regeneration at the time of overt hyperglycemia.27
The occurrence of DKA is more commonly seen with new-onset disease, especially in children younger than 4 years of age (Fig. 14-1), but it is not specific to T1DM and may also occur in T2DM.32 DKA occurs invariably in different populations, ranging from 15% to 67% of patients in Europe and North America, whereas in the developing world, around 80% of children with T1DM present initially with DKA.33 Typically the child is acidotic with acetone fruity odor, air hunger and Kussmaul respiration, abdominal pain, nausea and vomiting, and polyuria and polydipsia (see Fig. 14-1). The child has hyperglycemia, glucosuria, ketonemia, and ketonuria. Without timely management, severe fluid and electrolyte depletion soon develop, with signs of hypoperfusion and deterioration in level of consciousness due to cerebral edema that may lead to coma and death (see Fig. 14-1).
FIGURE 14-1 Pathophysiologic mechanisms and clinical presentations in type 1 diabetes mellitus. The clinical presentation in T1DM can be classical or emergency diabetic ketoacidosis (DKA), depending on the extent of insulin deficiency, effects of counteracting hormones, and other conditions. DKA can be viewed as an exaggerated compensatory physiologic mechanism.
Because T1DM is detected in the majority of patients after a relatively short period of symptoms, such as increased thirst, polyuria, and unexplained weight loss, the natural history of the disease has been poorly defined until recently. With the current ability to define individuals at high risk for developing T1DM based on high-risk HLA and autoantibody positivity, our understanding of the prediabetic period is improving. Current prospective studies of children with increased risk for T1DM (TEDDY, DIPP; DAISY, DiPiS, BABYDiab, PANDA)34,35 have revealed that autoantibodies to GAD65, IA-2, or insulin and decreased ability to release insulin in response to glucose36 may develop several years before the clinical diagnosis. The sequence of events preceding the diagnosis of overt T1DM would include the following: (1) genetic predisposition; (2) overt immunologic abnormalities with normal glucose levels; (3) development of β-cell dysfunction; (4) development of overt hyperglycemia with detectable C peptide; and (5) the final stage of insulin dependency, with disappearance of C peptide (Fig. 14-2). The fact that T1DM develops in persons of all ages must be taken into account when studying the natural history in children and adults. It is of significance that T1DM is often associated with other autoimmune diseases such as autoimmune thyroiditis (15% to 30%), Addison’s disease (0.5%), and celiac disease (4% to 9%). Additionally, the risk of autoimmune diseases is increased among relatives to T1DM patients.37
Diagnostic Criteria
The basis for distinction between T1DM and T2DM is islet autoimmunity and the patient’s dependence on insulin. Most patients have a history of polyuria, polydipsia, and unexplained weight loss (see Fig. 14-1). The American Diabetes Association (ADA)4 and the International Society for Pediatric and Adolescent Diabetes (ISPAD)28 have recommended diagnostic criteria for T1DM (Table 14-1).
Table 14-1
Criteria for the Diagnosis of Diabetes
| 1. | Fasting plasma glucose (FPG) ≥126 mg/dL (7.0 mmol/L). Fasting is defined as no caloric intake for at least 8 hr.* |
| OR | |
| 2. | Symptoms of hyperglycemia and a casual plasma glucose ≥ 200 mg/dL (11.1 mmol/L). Casual is defined as any time of day without regard to time since last meal. The classic symptoms of hyperglycemia include polyuria, polydipsia, and unexplained weight loss. |
| OR | |
| 3. | 2-hr plasma glucose ≥200 mg/dL (11.1 mmol/L) during an oral glucose tolerance test (OGTT). The test should be performed as described by the World Health Organization, using a glucose load containing the equivalent of 75 g anhydrous glucose dissolved in water.* |
*In the absence of unequivocal hyperglycemia, these criteria should be confirmed by repeat testing on a different day.
From American Diabetes Association: Standards of medical care in diabetes–2008. Diabetes Care 31(Suppl. 1):S12–S54, 2008.
Evidence has been presented that adult-onset T2DM may progress to insulin dependence at a rate of 1% to 2% per year.3 This rate is increased in GAD65 autoantibody–positive patients classified with T2DM,38 and this form of T1DM has variably been termed type 1.5 diabetes, latent autoimmune diabetes in adults (LADA), or slowly progressive T1DM.39
Glucose tolerance tests, both oral and intravenous, are being used to evaluate diabetic states. The oral glucose tolerance test (OGTT) is a diagnostic criterion (see Table 14-1). Although glucose tolerance tests are more important in disorders of impaired glucose tolerance and T2DM, they may also find an increased use in the prediction and diagnosis of T1DM.40
Classification
T1DM can be clinically classified into two main types: immune-mediated, or type 1A, and the uncommon idiopathic, or type 1B (~10% of patients). Unlike immune-mediated T1ADM, T1BDM is not associated with islet-cell antibodies or HLA, although both forms can involve ketosis acidosis.4,5
Epidemiology
The prevalence of T1DM is low compared with that of T2DM. Among individuals aged 30 years or younger, the prevalence of T1DM does not usually exceed about 0.3%, compared with prevalence rates for T2DM of 4.2% worldwide and nearly 25% in certain high-risk populations.33 While current attention is directed toward the trends of the “epidemic” of T2DM, recent data suggest that there is a “parallel rise” in the incidence rates of both T1DM and T2DM.41 Both geographic and ethnic variations are seen in the prevalence rates. The International Diabetes Federation (IDF) estimates that in 2006 there were 440,000 cases of T1DM among children younger than 15 years of age, and that Southeast Asia contributes to a quarter of the prevalent cases, while Europe contributes with one fifth.33 It has been pointed out that the results of prevalence studies should be viewed with caution. Different age groups were studied,4,33 and although geographical variation may be related to variations in genetic predisposition, environmental risk factors, or both, it may also be due to some research methodological problems.33 In addition, it is unclear whether recently observed increases in incidence rates worldwide reflect a change in the age at onset of diabetes or a true increase in prevalence of T1DM.42
Incidence
Globally, the current rise in incidence rates by 3% (2% to 5%) is expected to increase 40% higher in 2010 compared to figures from 199843,44 (Fig. 14-3). Currently it is estimated that more than 70,000 new patients are diagnosed each year worldwide.33 There are more suggestions that the increasing incidence rates and changing trends are seen in populations with high incidence and populations with previously lower incidence rates.27 Over the last 3 decades, T1DM has shown temporal trends in most of the regions of the world except Central America and the West Indies, while European data showed that the main recent rise in incidence rates is in Central-Eastern countries.43 Other studies such as EURODIAB44 and DiaMond45 have attempted to better define the incidence rates in various populations by using prospective, population-based, geographically defined registries. Overall, the average annual incidence rate varies widely, from 0.1 per 100,000 children in parts of Asia and South America43,46 to 64.2 per 100,000 in Finland.47 Worldwide epidemiologic investigations of T1DM as a noncommunicable disease indicate a 4% to 6% annual increase in incidence rate in Scandinavia,33,44 and similar data are being collected in other countries.48 Pooled data from all sites in the EURODIAB study demonstrated an annual rate of increase in incidence of 3.4%, with more rapid rates of increase occurring in certain regions.43,44 About 1% of all children born in Scandinavian countries will manifest T1DM during their lifetimes. The causes of this rapid rise are not understood, and further studies are needed to define to what extent the rise may be due to environmental factors, genetic admixture, or both.
FIGURE 14-3 Incidence rates of type 1 diabetes mellitus (T1DM), 0 to 14 years: global trends in childhood T1DM. The incidence of T1DM varies widely, depending on genetic background and geographic location. (From Diabetes Atlas, 3rd ed. Brussels: International Diabetes Federation, 2006, with permission.)
Geographical Distribution
T1DM can occur in any region of the world, and recently it was observed that the incidence is increasing in countries with both high and low incidence rates; in the latter group, the rise is steeper.43,48 However, it is suggested that the incidence of T1DM increases with increasing distance away from the equator (see Fig. 14-3). The annual incidence of T1DM is higher in northern Europe than in the Mediterranean area, with the exception of Sardinia. Furthermore, the incidence in Iceland is lower than that in Sweden or Finland.43,44 Surprisingly, the T1DM incidence rate in Estonia is about 25% of the rate in Finland, in spite of their close geographic proximity.49 The incidence rates within countries also show stable differences. The eastern and southern parts of Finland,50 as well as the central and southern parts of Sweden,51 have higher incidence rates than the northern parts of those countries.
The cause of geographic variation remains unknown, but it has been speculated to be related to genetic factors primarily associated with different HLA-DR or DQ genotypes of the major histocompatibility complex (MHC) on chromosome 6. HLA class II molecules from the DQ and DR loci that are necessary (but not sufficient) for disease vary greatly between countries and thereby affect disease incidence.49,52,53 In addition, environmental factors are important in understanding the pathogenesis of T1DM and may also help to explain differences in geographic distribution. This is supported by the observation that monozygotic twins show less than 20% to 30% concordance rates for T1DM.54
In the European region, incidence rates show correlation with the frequency of HLA susceptibility genes in the general population.52,55 In this region, there is also wide variation in incidence rates, ranging from 6 to 62 cases/100,000/year among children 0 to 14 years of age, with higher rates observed in Scandinavian countries, especially Finland and Sweden.44,47 In countries with populations of European origins, there is less variation: Canada reports 21/100,000/year, the United States 16/100,000/year, and New Zealand and Australia have incidence rates ranging from 16 to 20 cases/100,000/year.43 In the Middle East and North Africa, few reliable data are available. The incidence rates range from 1.0/100,000/year in Pakistan to 8 to 12 cases/100,000/year in Egypt, Sudan, and Libya; however, reports from Kuwait showed a rise in incidence rates from 15.4 in 1992 to 20.9 in 1997. There are few published data from the Southeast Asian region, and data from the African region are mostly lacking. In South and Central America, the incidence rates are low, except for Argentina, 6.8, and Uruguay, 8.3 cases/100,000/year. The Western Pacific countries have low incidence rates, especially China, which has one of the lowest incidence rates worldwide, ranging from 0.1 to 0.6 cases/100,000/year (see Fig. 14-3).43,48
It has been documented that migration of people from relatively low incidence countries into countries with a “diabetogenic” environment (i.e., with high incidence rates) increases the risk among the migrant groups. Migrants, within a short period of time, assume comparative risks of the native population, and several studies confirmed the importance of country of origin in the risk of T1DM.56,57 Improved epidemiology and better diagnostic criteria to distinguish different forms of diabetes are critical in obtaining reliable incidence rates for various countries and states.
Variation with Age
Until recently, T1DM was thought to occur almost exclusively in children and adolescents (Table 14-2). Epidemiologic studies with rigorous diagnostic criteria4 suggest, however, that the clinical onset of T1DM may occur at any age. The incidence rate varies with both age and gender (Fig. 14-4). The peak for both girls and boys, age 11 to 14 years, has been discernible in most studies and seems to be present irrespective of the country or area studied.44,46,48 This peak is associated with puberty and the maximal velocity of pubertal growth, and it may be associated with reduced insulin sensitivity due to hormonal factors related to growth. Previous studies have suggested that the annual rate of increase in incidence is higher in younger age groups, with rates of 6.4% in 0- to 4-year-olds, 3.1% in 5- to 9-year-olds, and 2.4% in 10- to 14-year-olds.44 In children, minor incidence peaks occur at 4 to 6 years58 and 7 to 8 years,59,60 which have been associated with entrance into preschool or school programs.61 Recent data suggest that around 50% to 60% of cases of T1DM in Western countries occur before the 15th birthday, and that more than 90% of childhood and adolescent diabetes is T1DM.55 Epidemiologic investigation has indicated that the incidence rate of T1DM in patients older than 20 years is lower than that seen in children, except for a possible peak at around 50 to 65 years.62 In addition to adults having what is considered a more classic T1DM clinical picture, some adult patients initially classified and treated as having T2DM may require insulin after 1 to 5 years of therapy with diet, exercise, or oral hypoglycemic agents.38,39 As noted earlier, this type of diabetes is referred to as type 1.5 diabetes, latent autoimmune diabetes in the adult (LADA), or slowly progressive insulin-dependent diabetes mellitus (SPIDDM).39,63 These patients are positive for islet autoantibodies and have lower body mass index and higher frequencies of high-risk HLA types compared with other “type 2” patients.63 It is speculated that these patients in fact have T1DM in addition to genetic factors that may inhibit rapid progression of β-cell killing.62
Table 14-2
Severity of Symptoms at Onset, A1C Goals, and Treatment for T1DM by Age Group
A1C, Hemoglobin A1c; PBG, plasma blood glucose; T1DM, type 1 diabetes mellitus.
Modified from American Diabetes Association: Standards of medical care in diabetes–2008. Diabetes Care 31(Suppl. 1):S12–S54, 2008.
FIGURE 14-4 Age-adjusted incidence of insulin-dependent diabetes mellitus in relation to age at onset in male and female subjects. (Data from Nystrom L, Dahlquist G, Ostman J et al: Risk of developing insulin-dependent diabetes mellitus [IDDM] before 35 years of age: indications of climatological determinants for age at onset. Int J Epidemiol 21:352–358, 1992, with permission.)
Although rare, some forms of diabetes mellitus can occur in the newborn period, at birth, or during the first months of life. Neonatal diabetes (NDM) describes one kind of diabetes that occurs in the first month (up to 6 months) of life at a rate of 1 in 500,000 live births. It is characterized by insulin-sensitive hyperglycemia. About half of cases of NDM are transient (TNDM), but the remaining progress into a permanent (PNDM) type, with lifelong dependence on exogenous insulin. Clinically the infant is thirsty, with excessive urination and hyperglycemia and/or glycosuria, but the condition may progress into ketoacidosis. PNDM usually proceeds after the second year, and differentiation between the two forms can only be done through genetic testing. TNDM occurs at birth up to 3 to 6 months and is correlated to defects in specific genes: ZAC/HYMAI (ZAC pleomorphic adenoma gene–like 1, or PLAG1, and hydatidiform mole–associated and imprinted transcript, HYMAI), or KCNJ11, the moderately activating mutation of the Kir6.2 gene, or hepatocyte nuclear factor 1β (HNF1β) and SUR1 (sulfonylurea receptor 1). The PNDM, on the other hand, is related to several gene defects, such as pancreatic aplasia, activating mutations of KCNJ11 of the Kir6.2 gene that encodes for ATP potassium channel (KATP), the ABCC8 gene of sulfonylurea receptor 1 (SUR1), complete deficiency of glucokinase IPF1 (PDX1) gene, which encodes insulin promoter factor1, PTF1A, which encodes the pancreas transcription factor 1A, and also mutations in forehead box p3 (FOXP3) gene (also known as T-cell regulatory gene).64
Variation with Gender
It has been reported that the peak incidence in girls occurs earlier than in boys59 (see Fig. 14-4). If the clinical onset of T1DM is linked to pubertal growth, this difference in incidence rate can be explained by the fact that pubertal growth occurs earlier in girls. In children followed from birth because of T1DM genetic risk, the male-to-female ratio was 1.4 : 1.0 before 6 years of age at T1DM diagnosis. This ratio was increased (1.7 : 1.0) among children younger than 18 months of age.65 Prepubertal boys were found to be taller at the clinical onset of T1DM.60 In addition, newly diagnosed children of both genders showed advanced skeletal maturity.66 Even if boys tend to show an increased height compared with controls, their growth seems to cease about 35 weeks before the clinical onset of T1DM.60 It therefore is possible that processes affecting the pancreatic β-cell mass and the ability to produce insulin may have profound effects on body growth and function at a young age. Because these processes differ slightly between boys and girls, growth characteristics may offer a simple explanation for the differences seen in incidence rates between the genders.
A study that examined sex differences in T1DM incidence before the age of 15 showed a slight male preponderance in many but not all European countries, whereas a female preponderance was found in most African and Asian countries.67 Interestingly, the male excess was seen in all countries with the highest incidence rates (>20/105/yr), whereas a female excess was seen in all countries with low incidence rates (<4.5/l05/yr). Studies of older age groups have consistently shown a male preponderance for new cases of T1DM, with male/female ratios ranging from 1.3 to 2.5 : 1.68 One registry of T1DM among 15- to 34-year-olds (see Fig. 14-4) demonstrated that T1DM was 1.5 times more common among men than women.61 Further studies are necessary to document gender-dependent incidence rates and to explain their mechanisms.
Seasonal Variations
T1DM shows cyclic, sinusoidal appearance of seasonal variation in date of diagnosis, with a peak in winter months.48 This variation is proposed to be related to the timing of occurrences of certain precipitating factors such as viral infection and cold climate.48 It is of interest that islet cell autoantibodies appearing during the prediabetic state follow similar seasonal variations, being more prevalent in colder months and rare in summer and spring. Additionally, the triggering of β-cell autoimmunity appears to have variable trends over subsequent years and does not equally affect genetically susceptible siblings.69 These observations suggest that triggering of immunologic markers may be related to factors (most likely viruses) that prevail mostly in colder months with variable frequency of occurrence.
Etiology
The absence of an unambiguous mode of inheritance, the presence of a period of subclinical islet autoimmunity preceding clinical onset of disease, human leukocyte antigen (HLA) genes that control the immune response, and age and seasonal variation must be taken into account in attempts to explain the cause of T1DM. A defined etiologic factor, endogenous or exogenous, capable of causing T1DM remains to be identified (see Table 14-5). Because evidence exists of genetic heterogeneity in T1DM, it is possible that different causative factors are responsible. In experimental animals (see Table 14-5), both viral and chemical agents have been used to induce diabetes reproducibly, and certain strains of animals are at higher risk for developing diabetes due to genetic factors. In addition, only indirect evidence suggests that environmental factors that are clearly diabetogenic in animals are involved in initiating T1DM in humans. The following is a brief summary of possible genetic or environmental factors that are associated with the appearance of T1DM.
Inheritance
The mode of inheritance is complex, and around 80% to 85% of cases of T1DM occur sporadically without familial aggregation.52 Among HLA-identical siblings of T1DM-affected patients, about 20% eventually manifest the disease. The overall lifetime risk for first-degree relatives has been estimated at about 8% for siblings—15 times higher than the general population—and 5% for children of parents with T1DM.70 The risk of T1DM among offspring ranges from 2% to 4% if the mother was affected, compared to 6% to 9% if the father was affected, and the risk tolls to 30% when both parents are affected.52,70
Genetic Factors
T1DM is both genetically associated with and linked to certain HLA genetic factors of the major histocompatibility complex (MHC).71,72 By using DNA sequence information in the genetic analysis, it is found that more than 95% of all patients in whom T1DM onset occurred before age 30 years are positive for the chromosome 6 HLA haplotypes DRB1*04-DQAI*0301-BI*0302, DRBI*03-DQA1*0501-BI*0201, or both. Although some 40% to 50% of the background population carry these HLA factors, they represent necessary but insufficient prerequisites for the development of T1DM. It has been estimated that HLA contributes about 60% of T1DM risk among first-degree relatives to T1DM patients. Hence it is not surprising that less than 10% of genetically susceptible subjects develop the disease.68 Other genetic factors have indeed been identified,73,74 but none has shown a level of importance comparable to that of HLA.
The HLA haplotypes DRB1*04-DQAI*0301-BI*0302 and DRBI*03-DQA1*0501-BI*0201 are the two major-risk haplotypes71,75 (Table 14-3). The most important alleles are DQB1*0302 and DQB1*0201, along with DRB1*03. DRBI*04 is a large family of related molecules, and DRB*0401 confers an independent risk, whereas DRB1*0403 is negatively associated with T1DM and may protect or decelerate an ongoing disease process.75 DRB1*03 seems to be more important than DQB1*0201, as only DRBI*03-DQA1*0501-BI*0201, not DRBI*07-DQA1*0501-BI*0201, confers T1DM risk. DQB1*0401 and DQB1*0404 are susceptibility alleles on the DRB1*04-DQAI*0301-BI*0302 haplotype. DQB1*0604 and DQB1*0501 are susceptibility alleles, together with either DRBI*03-DQA1*0501-BI*0201 or DRB1*04-DQAI*0301-BI*0302. The genetic linkage and association between T1DM and HLA is remarkable in that certain HLA haplotypes are protective. Most prominently, DQA1*0102-B1*0602 and DQA1*0102-B1*0603 are protective before age 15 years (see Table 14-3). The detailed mechanisms by which HLA confers either risk or resistance is not fully understood.76 The function of these molecules is to display peptide antigens to be recognized by T-cell receptors (TCR). The disease association may therefore be related either to an inability to induce immunologic tolerance to certain autoantigens or to antigen presentation of an endogenous autoantigen. In the first case, the subject may be exposed to infectious agents that mimic autoantigens. Reactivity to the infectious agent sets off an immune response that cross-reacts with self. In the second case, the immune reaction may result in a direct attack on the individual’s own cells.
Table 14-3
Class II HLA Haplotype Risk Genes of Type 1 Diabetes in Caucasians
The commonest high-risk haplotype is: HLA DR4-DQ8 (DQB1*0302-A1*0301).
The commonest moderate-risk haplotype is: HLA DR3-DQ2 (DQB1*0201-A1*0501).
The commonest high-risk genotype is: heterozygous DR3-DQ2/DR4-DQ8.
*The five most susceptible (S).
†The four most protective (P).
Modified from Erlich H, Valdes AM, Noble J, et al: HLA DR-DQ haplotypes and genotypes and type 1 diabetes risk. Analysis of the Type 1 Diabetes Genetics Consortium families. Diabetes 57:1084–1092, 2008.
The genetics of T1DM is studied extensively because it represents a paradigm for genetically complex diseases. Genome screens and studies on candidate genes have provided evidence for genetic linkage between polymorphic DNA markers and more than 10 putative T1DM susceptibility genes (refer to T1DBase at http://www.t1dbase.org/page/Welcome/display). Currently, it has been shown that PTPN22 (chr.1p13), CTLA-4 (chr.2q33), IFH1 (chr.2q24), IL2 (chr.4q27), ITPR3 (chr.6p21), IL2RA (chr.10p15), INS-VNTR (chr.11p15), TH (chr.11p15), ERBB3 (chr.12p13), C12orf30 (chr.12q24), CLEC16A/KIAA0350 (chr.16p13), PTPN2 (chr.18p11) genes and two recently identified novel loci, BACH2 (chr.6q15) and UBASH3A (chr.21q22)77 are all associated with increased T1DM risk (Table 14-4). References to these individual genetic factors are found at the T1DBase website.
Environmental Factors
T1DM is characterized by a multifactorial web of environmental factors (Table 14-5). The interaction between genetic and environmental factors and the aggressive islet autoimmune phenomena is complex. The fact that the disease has a relatively long latent period, in which autoimmunity is triggered long before the onset of the clinical syndrome (see Fig. 14-2), indicates a role of environmental factors as initiators or promoters in relation to genetic predisposition.68 Several environmental factors are incriminated in the etiology of T1DM, but linking is indirect, and a definite proof of confirmed association for each factor is yet to be achieved, since the majority of investigations have been carried out at the time of clinical diagnosis and not in relation to the time of islet autoimmunity.
Table 14-5
Viruses and Other Environmental Factors Implicated In or Able to Induce Type 1 (Insulin-Dependent) Diabetes
Viral Infection
There are numerous reports of individual cases in which T1DM onset followed an acute viral infection, borne out by subsequent animal studies in which viruses have been shown to have diabetogenic activity.78 The first case linking T1DM to an acute viral infection was reported in the late 19th century; the onset of T1DM appeared to be precipitated by a mumps infection in a child.79 Many similar reports have followed since,78,80 and taken together, these reports suggest a relation between the clinical onset of T1DM and several viruses, including rubella, mumps, coxsackievirus B, rotavirus, cytomegalovirus, and Epstein-Barr virus, have been implicated.78 The true relation between these viral diseases and the clinical onset of T1DM remains conjectural.
Viruses are thought to increase the risk of T1DM through different proposed mechanisms. Whereas some viruses induce rapid and direct effect, causing β-cell destruction and subsequent insulin deficiency, other viruses induce slower, longstanding effect through activation of autoreactive T cells. Viruses may also inhibit insulin production by inducing interferons and HLA antigen expression, or they may mimic autoantigens of β cells.78,80
T1DM develops in individuals who are positive for HLA-DR3-DQ2, DR4-DQ8, or both (discussed earlier). The DQ6 haplotype containing DQB1*0602 confers resistance among children. Because these specificities are present in about half the population, a virus inducing either islet autoimmunity, clinical onset of T1DM, or both, may not be spread effectively enough to cause disease. In addition, variations in the annual incidence rate are often taken as evidence of an involvement of virus. An annual variation was found in a population of 15- to 34-year-olds, with lower numbers of new patients being identified during the summer months.48,69 In the group of children younger than 6 or 7 years who have T1DM, however, this annual variation is not always present.51,81
Earlier studies suggested that the congenital rubella syndrome was strongly associated with T1DM; approximately 10% to 20% of children develop autoimmune T1DM, and around 50% of them develop autoimmune antibodies.80 However, most countries with high incidence, such as Finland and Sweden, had successfully implemented vaccination programs against rubella virus. Vaccination practices have prevented rubella epidemics, but they have not affected the incidence rate of T1DM.
Maternal enteroviral infection during pregnancy also appears to be a risk factor for childhood T1DM.82 It is therefore possible that gestational infections by many types of viruses affect the maturation of the immune system, causing certain children to be more predisposed to autoimmunity and thereby increasing the risk for T1DM. Prior exposure to measles, mumps, and rubella, but not vaccination, decreased prevalence of pancreatic and thyroid autoantibodies.83 Maternal viral infections or reduced exposure to natural infections may be associated with an increase in T1DM.
Enteroviruses, in particular the coxsackie B serotype have been extensively studied at the time of clinical onset84 but rarely in relation to the appearance of islet autoimmunity.80 Coxsackie B4 virus was isolated from the pancreas of a child who died at presentation with T1DM, propagated in the in-vitro cultures of endocrine pancreatic cells, and then shown to have diabetic activity in certain mouse strains.85 Coxsackievirus infection in the mouse seems to be associated with virus replication in the β cells, followed by the formation of GAD65 antibodies.86,87 Two important hypotheses follow from these experiments. One hypothesis is that the coxsackievirus induces β-cell neoantigens, which initiate an (auto)immune reaction. This hypothesis can be tested by analyzing the appearance of such neoantigens. The neoantigen may initiate a devastating reaction if its structure mimics a self-protein. In line with this hypothesis, a sequence in GAD65 is identical to that in a coxsackievirus antigen.88,89 Another hypothesis is that Coxsackie B virus replication in β cells results in β-cell necrosis and the formation of antibodies against β-cell constituents or “hidden antigens” not normally surveyed by the immune system.90 An autoimmune reaction is initiated that may escalate with time.
Other viruses incriminated in T1DM etiology include rotaviruses, which are common causes of childhood gastroenteritis. These viruses showed peptide sequences similar to T-cell epitopes in IA-2 and GAD antibodies, suggesting their role in T1DM.82 Cytomegalovirus is a DNA virus that is thought to induce autoimmunity through a molecular mimicry mechanism. Mumps virus was correlated to autoimmunity and even to overt T1DM. Human endogenous retroviruses (HERV) were also reported to correlate with T1DM.80
In experimental animals (see Table 14-4), several viruses are known to induce T1DM, either through direct effects on the β cells, causing rapid destruction (e.g., encephalomyocarditis virus EMC-D) or through the disruption of the normal immune regulatory mechanisms (e.g., Kilham rat virus [KRV], which causes autoimmune T1DM in diabetes-resistant BioBreeding [DR-BB] rats).78 Other viruses (e.g., reovirus) were found to affect primarily the immune system of the host to induce a polyclonal autoantibody response. The production of autoantibodies appeared to be closely related to the pathogenesis of disease in mice.86,87 Captured bank voles may develop T1DM with islet autoantibodies and signs of parechovirus infection, raising the question whether there are zoonotic virus–inducing islet autoimmunity, T1DM, or both.91
The virus-induced disease in mice depends on strain, because some mouse strains are resistant to the pathogenesis of a virus that induces T1DM in another strain. Similar strain dependency to the β-cytotoxic agent streptozotocin, followed by the inoculation of a diabetogenic virus, rendered otherwise virus-resistant mice diabetic.92 This observation may be significant to humans because it is possible that repeated injuries to the pancreatic β cells over several years of life may eventually induce T1DM. An additional role of HLA has been suggested, because it cannot be excluded that repeated injuries are particularly detrimental if the T1DM-associated HLA alleles are linked with a poor regenerative capacity of the pancreatic β cells.
Hygiene Hypothesis
The hygiene hypothesis proposes that as living environment is improved, children become less exposed to infectious agents, which leads to inadequate maturation of their immune systems. This hypothesis suggests that early exposure to pathogens may enhance the immune responses of those children, thereby suppressing autoimmune reactions involved in T1DM pathogenesis. The latter suggestion is based on increasing incidences of diseases like asthma or other atopic disorders, in addition to the fact that T1DM is more prevalent in developed societies. The hypothesis also suggests that younger children are more prone to infections, since they did not acquire antibodies against viruses (e.g., enterovirus antibodies from their mothers), resulting in an increased risk for T1DM or other autoimmune diseases.80 Indeed, other studies reported higher maternal enterovirus antibodies in countries with lower incidence of T1DM compared to countries with high incidence.82
Dietary Factors
The dietary hypothesis is based on earlier observations that breastfed children have a lesser risk of T1DM. It was also noted that early exposure of infants younger than 6 months to cow’s milk proteins may double the T1DM risk, particularly in HLA high-risk children. Other studies described an association between islet autoimmunity and shorter duration of breastfeeding, as well as early exposure of infants to bovine protein.93 The Diabetes Prediction and Prevention (DIPP) study of Finland reported a fivefold higher risk of multiple islet autoantibodies among children with HLA DQB1*0302 who received cow’s milk formula before the age of 4 months. Similar associations between islet autoimmunity or T1DM and intake of glutens, foods rich in protein, carbohydrates, and nitrosamine compounds were reported.94 Additionally, early exposure to bovine insulin in cow’s milk may lead to formation of autoantibodies against human insulin through cross-reaction.95
