Type 1 diabetes (T1D) is a chronic autoimmune disorder resulting from immune-mediated destruction of insulin-producing beta cells within the pancreatic islets. Prediction of T1D is now possible, as having 2 or more islet autoantibodies confers a 100% risk of diabetes development. With the ability to predict disease development, clinical trials to prevent diabetes onset have been completed and are currently under way. This review focuses on the natural history, prediction, and prevention trials in T1D. We review the lessons learned from these attempts at preventing a chronic autoimmune disease and apply the paradigm from T1D prevention to other autoimmune disorders.
Type 1 diabetes is a chronic progressive autoimmune disorder with a preclinical phase of disease before clinical onset.
Type 1 diabetes can be predicted based on the measurement of antibodies directed against islet antigens.
Large prospective randomized double-blinded placebo-controlled secondary prevention trials for Type 1 diabetes have been completed.
As rheumatoid arthritis and other rheumatic diseases have a defined preclinical stage of disease, the lessons learned from diabetes prevention efforts can be applied to rheumatic diseases.
Type 1 diabetes mellitus (T1D), the immune-mediated form of diabetes requiring insulin treatment, is a prevalent chronic autoimmune disease affecting both children and adults. The incidence of T1D is increasing dramatically, doubling in the past 20 years. The vast majority of T1D cases result from autoimmune-mediated, nonreversible destruction of insulin-producing beta cells within the pancreatic islets. Progressive beta cell destruction and decreased endogenous insulin production occur during a silent preclinical phase in which blood glucose levels remain normal. During the preclinical phase of disease, autoantibodies directed toward beta cell–specific antigens can be measured in a patient’s blood, and measurement of islet autoantibodies has made T1D a predictable disease. Inflammation and T-cell–mediated destruction of islet beta cells result in the development of clinically apparent disease marked by abnormal glucose homeostasis. With the ability to assess diabetes risk and predict disease onset, many large clinical trials aimed at disease prevention have been completed over the past decade. These studies have not completely prevented disease onset but hold promise for identifying an intervention to slow disease progression. This review focuses on the natural history of T1D, with brief sections on clinical diagnosis and treatment, prevention efforts in preclinical T1D, and a final section applying the lessons learned from diabetes prevention to rheumatic diseases.
Great strides in understanding the natural history and pathogenesis of T1D have occurred in large part from longitudinal studies following children from birth for the development of islet autoantibodies and diabetes development (DAISY in the United States, EURODIAB in Germany, and the Type 1 Diabetes Prediction and Prevention Trial [DIPP] in Finland). T1D incidence has also been well defined through these studies. The incidence of T1D varies greatly by geographic location, with an average annual incidence of 2.3% per year. The incidence among Caucasians in the United States is 17.8/100,000 patient years for children younger than 14 years. Unlike most other autoimmune diseases in which female individuals are affected more than male individuals, male and female individuals are equally affected with T1D. The age of diabetes onset has 2 peaks, 1 in children 5 to 7 years of age and again in adolescents 10 to 14 years old. Adults also develop T1D, with approximately 25% of new T1D cases diagnosed in individuals older than 18 years of age. With few exceptions, the incidence rate for T1D is rising in all age groups between 2.4% and 3.3% per year, with the largest increase among children who are younger than 5 years.
T1D is still the predominant form of diabetes in youth, even though the incidence of type 2 diabetes (T2D) mellitus is increasing. More than 85% of people with T1D or T2D who are younger than 20 years old have T1D. Although most individuals diagnosed with T1D have no family history of T1D, the development is strongly influenced by genetic factors. In the general population, there is a 1 in 300 lifetime risk for developing T1D. Individuals with a first-degree relative with T1D have a 1 in 7 to 1 in 30 lifetime risk of developing the disease depending on the affected relative. Children of mothers with T1D carry an approximately 3% lifetime risk of developing T1D, whereas the risk increases to approximately 5% for a father with diabetes. A recent analysis of monozygotic twins who were initially discordant for T1D showed that by 60 years of age, persistent autoantibody positivity, T1D, or both had occurred in 78% of these individuals.
T1D is clearly a polygenic disorder, as evidenced by genome-wide association studies, which have identified more than 40 genetic polymorphisms that confer susceptibility to T1D development. The HLA antigen class II region on chromosome 6 confers greater than 50% of the genetic susceptibility to T1D. Specific major histocompatibility complex class II alleles can confer both risk and dominant protection. Individuals having a haplotype containing DR4 and DQ8, which are in close linkage disequilibrium, have the highest risk for disease development. Approximately 60% of all T1D patients have this haplotype, whereas 90% of patients have either or both the DR4/DQ8 and DR3/DQ2 haplotypes. Those individuals with a DQ6 allele (DQB1*06:02) are protected from diabetes development with a striking odds ratio of 0.03 for disease development.
Besides the HLA class II genes, non-HLA genes also confer genetic risk for T1D. Of the non-HLA genes, the insulin gene (INS), protein tyrosine phosphatase nonreceptor type 22 (PTPN22), CTLA4, and IL2RA have the greatest association with T1D development. Similar to the class II genes, the insulin gene polymorphisms confer both risk and protection for T1D. Variable number of tandem repeats 5′ of the insulin gene allow for more or less insulin message to be expressed in the thymus. More insulin expressed in the thymus results in central tolerance and increased protection of T1D. On the other hand, less insulin message in the thymus correlates with a lack of negative selection for autoreactive CD4 T cells and risk for T1D development. Many non-HLA genes appear to be markers of generalized autoimmunity, as a number are present in other autoimmune disorders. For example, PTPN22 is associated with rheumatoid arthritis (RA), Crohn disease, Graves disease, and systemic lupus erythematous. PTPN22 is expressed in lymphocytes and is involved in T-cell activation.
Despite significant research efforts aimed at identifying an environmental trigger that leads to a loss of tolerance and T1D development in genetically susceptible individuals, no clear risk factor has been causally linked to islet autoimmunity (detected by the measurement of serum autoantibodies) or T1D onset. T1D natural history studies indicate that islet autoantibodies can be detected between 9 months and 2 years of age in genetically high-risk newborns, suggesting that the initiating environmental trigger may occur in utero or early in postnatal life. Many viruses have been proposed to break tolerance, initiate the preclinical phase of T1D, and accelerate the onset of clinical T1D. These include coxsackie virus, cytomegalovirus, and enterovirus, with the most recent research focusing on enteroviruses. In the Diabetes Autoimmunity Study in the Young (DAISY) cohort, there was a strong correlation between children who developed autoantibodies and enterovirus in their serum.
Other hypothesized environmental triggers include the hygiene hypothesis, in which a lack of exposure to bacterial pathogens during childhood leads to altered protective immunity and increased risk of autoimmune diseases. The north-south gradient hypothesis postulates that a lack of vitamin A and D early in life triggers the onset of autoimmune disorders, such as T1D and multiple sclerosis. Vitamin D is implicated to have immune modulatory and anti-inflammatory effects, thereby protecting from disease development. Studies suggest that supplementation with vitamin D may have a protective effect on T1D development. Another well-studied environmental factor is that the introduction of cow’s milk or gluten early in life may trigger islet autoimmunity. Two recent studies found that exposing an infant to gluten before 3 months of age or after 7 months of age is associated with the development of islet autoantibodies. There is limited evidence suggesting that other environmental factors, including vaccinations, confer any risk for islet autoimmunity or progression to diabetes. Potential environmental triggers that are under further evaluation include delivery via cesarean, mothers with preeclampsia or advanced maternal age, exposure to nitrosamine compounds, increased body weight, and exposure to rubella virus.
Pathophysiology and islet autoantibody development
Three decades ago, it was hypothesized that T1D is a chronic autoimmune disorder that develops in stages. The model combined genetic, immunologic, and metabolic markers for disease development and still remains relevant today. The adapted model in Fig. 1 proposes that in genetically susceptible individuals there is a precipitating event that leads to the development of preclinical disease where there are overt immunologic abnormalities leading to progressive loss of beta cell mass, reduced insulin release, and intermittent dysglycemia. Serum autoantibodies in the preclinical stage of disease development are present years before metabolic decompensation. The 4 islet autoantibodies are directed against insulin (IAA), glutamic decarboxylase (GAD), tyrosine phosphataselike insulinoma antigen (IA-2), and zinc transporter 8 (ZnT8). These autoantibodies are markers of disease, whereas T-cell–mediated destruction results in beta cell loss. As the disease progresses and there is approximately 10% to 20% of beta cell mass remaining, a patient will become symptomatic and satisfy clinical criteria for T1D diagnosis. At this time, there will be minimal endogenous insulin release and abnormal blood glucose levels.
Multiple antibodies tend to develop simultaneously in an individual, indicating that epitope spreading occurs rapidly. If a single antibody is detected in children, it is most often to insulin. Insulin autoantibody levels directly correlate to T1D progression. In adults who develop T1D, GAD and IA-2 tend to be the most common detected antibodies. The risk of developing T1D is strongly correlated with the number of positive antibodies, and individuals with 2 or more go onto develop T1D with almost 100% certainty. Autoantibodies are most useful in predicting disease when they are present at a young age, in high levels, or in individuals with high-risk HLA genotypes. In general, islet autoantibodies persist until diagnosis and up to several decades thereafter. It is estimated that 5% to 10% of all T1D patients are islet autoantibody negative, indicating that other islet autoantigens may exist.
There are many hypotheses as to how beta-cell autoimmunity is initiated after a precipitating event, none of which are proven. One hypothesis is that the inciting event is an environmental determinant that shares amino acid sequence homology with beta-cell proteins and through molecular mimicry induces autoimmune destruction of beta cells. Another hypothesis proposes genetic susceptibility may lead to abnormal central tolerance (ie, specific HLA alleles do not allow for negative selection of islet antigens) followed by an inciting event that causes acute immune activation and targeting of beta cells when there are higher levels of antigen to be presented to T cells. One final proposal is that beta cells become more sensitive to free-radical or cytokine-activated inflammation that is present following a triggering event. Recent efforts to study pancreata of T1D organ donors, the network for pancreatic organ donors (nPOD), have shed light on the islet infiltrates of T1D patients. In those cases that have mononuclear cell infiltration within the islets (insulitis), there is a reduction of insulin-producing beta cells and an infiltrate consisting of CD8 T cells most frequently along with CD4 T cells, B-lymphocytes, and macrophages.
The classic clinical presentation of T1D consists of a triad of symptoms: polyuria, polydipsia, and weight loss. From the American Diabetes Association (ADA), an individual meets diagnostic criteria for diabetes if he or she has one of the following: (1) a fasting (≥8 hours) plasma glucose ≥126 mg/dL; (2) a random plasma glucose ≥200 mg/d with symptoms of hyperglycemia; (3) a 2-hour plasma glucose after a standard oral glucose tolerance test ≥200 mg/dL; (4) a hemoglobin A1C ≥6.5%. In distinguishing which type of diabetes an individual has, the ADA defines T1D as a disease caused by cellular-mediated autoimmune destruction of pancreatic beta cells; autoantibodies against beta-cell antigens (insulin, GAD, IA-2, and ZnT8) are markers of this destruction. Measurement of autoantibodies remains the best way to distinguish T1D from other forms of diabetes at this time.
Studies screening genetically at-risk individuals for islet autoantibodies have identified patients early in the disease course (ie, more remaining beta-cell mass). Hemoglobin A1c values rise 1.0 to 1.5 years before clinically apparent disease, indicating that hyperglycemia occurs before acute disease onset. Without an effective treatment to prevent beta-cell destruction, all individuals eventually develop hyperglycemia and T1D. After diagnosis, many patients enter a “honeymoon” period in which there is endogenous insulin production, and they require small amounts of exogenous insulin to maintain good glycemic control. Unfortunately, the destruction of beta cells continues and all individuals eventually become reliant on exogenous insulin to maintain euglycemia. Although intensive insulin therapy decreases the risk for complications in T1D, macrovascular and microvascular complications are still prevalent. Over time, sustained hyperglycemia damages small vessels and leads to nephropathy, retinopathy, and neuropathy. Large vessels also are damaged and cardiovascular events occur 10 times more often in people with T1D than in age-matched controls.
The mainstay of treatment for T1D is the subcutaneous administration of insulin. Currently, there are no Food and Drug Administration–approved therapies that effectively preserve residual beta-cell mass. In large diabetes centers, patients are almost exclusively started on multiple daily injection therapy with a long-acting insulin analogue to counteract basal glucose release from the liver and a short-acting insulin analogue to cover the carbohydrate content of meals and correct for hyperglycemia (ie, basal-bolus therapy). Many patients go on to use insulin pump therapy that provides continuous subcutaneous insulin infusion (CSII). Studies show that CSII has a clear effect on lowering the rate of hypoglycemic events by approximately 70%. To dose insulin and keep blood glucoses in target range, T1D patients perform self-monitoring of their blood glucose with a glucometer on average 4 to 6 times daily. Recent advances in technology now provide patients the option of a continuous glucose monitoring system (CGM). With CGM, a subcutaneous sensor is inserted and interstitial blood glucose is continuously measured. A significant amount of research is focused on creating a closed loop system, combining CGM data with algorithms to automatically deliver insulin from an insulin pump, which would function autonomously as an artificial pancreas.
With the incidence and disease burden of T1D increasing, international efforts at delaying or preventing diabetes onset have occurred over the past 15 years. Large clinical trial networks have helped define the natural history of T1D, identify those individuals with preclinical disease, and provide infrastructure for coordinating prevention trials. Both primary and secondary prevention efforts in T1D are reviewed below.
Primary prevention trials have been conducted in high-risk individuals who have no islet autoantibodies or metabolic abnormalities. These individuals are generally first-degree relatives of T1D patients and have the presence of high-risk HLA genes. These trials involve infants and young children, necessitating minimal risk, and therefore they have been mostly limited to early in life dietary interventions. A clinical trial completed in Finland enrolled 230 high-risk infants, based on HLA and a first-degree relative with T1D, to receive either a casein hydrolysate formula or a conventional cow’s milk formula at the time when breastfeeding was decreased or no longer available within the first 6 to 8 months of life. Casein hydrolysate contains more processed proteins cow’s milk or traditional infant formula. Islet autoantibodies were measured periodically over 10 years, and the study showed that individuals who received the casein hydrolysate formula had a decreased incidence of developing positive autoantibodies. Based on these data, the Trial to Reduce Incidence of Diabetes in Genetically at Risk (TRIGR) study was initiated. Enrollment at 77 centers in 15 countries was completed in 2006. In this trial, infants with HLA susceptibility and a first-degree relative with T1D were randomized to casein hydrolysate or conventional formula at birth and were allowed to breastfeed as often as desired. The main outcomes from this trial include islet autoantibody positivity and clinical development of T1D. Although this trial is still ongoing, an interim analysis suggests that hydrolyzed infant formula may lessen the risk of developing a single islet autoantibody and a trend toward developing 2 islet autoantibodies ( P = .07). Further follow-up is needed to determine if clinical diabetes onset is delayed with this dietary intervention. Another pilot study took place in a similar genetically at-risk population, the Finnish Dietary Intervention Trial for the Prevention of Type 1 Diabetes (FINDIA). The study examined the removal of bovine insulin from whey-based hydrolyzed formula and randomized children to 1 of 3 intervention groups: (1) cow’s milk formula, (2) whey-based hydrolyzed formula that contains bovine insulin, or (3) whey-based hydrolyzed formula without bovine insulin. Islet autoantibodies were followed for 3 years and the infants assigned to the whey-based formula without bovine insulin had a lower risk of developing islet autoantibodies.
A final primary prevention trial under way is the TrialNet Nutritional Intervention to Prevent (NIP) Type 1 Diabetes Pilot Trial, which aims to assess the effect of omega-3 fatty acid supplementation on prevention of T1D autoimmunity. Dietary omega-3 fatty acid intake has been shown to reduce the risk of islet autoimmunity in genetically at-risk infants when analyzed retrospectively and is believed to have anti-inflammatory properties. At-risk infants were recruited into the NIP study and received docosahexaenoic acid or placebo for 6 months. Enrollment is now complete and the study is currently in the follow-up phase assessing for islet autoantibody development.
T1D develops in stages, and the preclinical period is readily defined by the presence of 2 or more islet autoantibodies. Many secondary prevention trials are aimed at delaying or preventing the progression to T1D in individuals with islet autoantibodies have been completed ( Table 1 ). To conduct these large randomized double-blinded placebo-controlled trials, infrastructure in the form of financial support and a clinical trials network has been necessary. The first T1D clinical trials network emerged in the early 1990s as the National Institutes of Health–sponsored Diabetes Prevention Trial-Type 1 (DPT-1). DPT-1 became known as TrialNet, and conducts clinical intervention trials focused on delaying/preventing T1D onset and identifying therapies to preserve residual beta-cell mass in newly diagnosed T1D patients. TrialNet is an international consortium with 18 clinical centers and an additional 150 affiliated sites. TrialNet has already completed many studies and provides ongoing support for clinical trials with several prevention trials currently enrolling participants.
|Study||Agent||Number Randomized||Age, y||Comments|
|DIAPREV-IT||GAD/Alum||50||4–18||Trial in follow-up phase|
|a Abatacept||CTLA-4 Ig||Currently enrolling||6–45||TrialNet sponsored|
|a Teplizumab||Anti-CD3 mAb||Currently enrolling||8–45||TrialNet sponsored|
|a Intranasal Insulin Trial (INIT-II)||Intranasal Insulin||Currently enrolling||4–20||Less-frequent dosing than prior studies|
|a Oral Insulin||Oral Insulin||currently enrolling||3–45||TrialNet sponsored|
|Diabetes Prevention Trial – Type 1 Study Group (DPT-1)||Oral Insulin||372||3–45||Post hoc analysis shows delay to T1D diagnosis in subset of patients|
|European Nicotinamide Diabetes Intervention Trial (ENDIT)||Nicotinamide||552||3–40||No effect on prevention|
|Diabetes Prediction and Prevention Trial (DIPP)||Intranasal Insulin||224||1–5||No effect on prevention|
|Diabetes Prevention Trial–Type 1 Study Group (DPT-1)||Parenteral Insulin||339||4–45||No effect on prevention|