Current approaches to the prevention of type 1 diabetes

Type 1 diabetes (T1D, previously called insulin-dependent or juvenile diabetes) is usually the result of T cell - mediated autoimmune destruction of the insulin-producing beta cells of the pancreas.[1] While T1D affects males and females equally, there are very significant worldwide variations in prevalence, with people of European descent being most affected. Recent estimates of the incidence of this disease in Canada suggest an annual rate of 20 to 26 new cases per 100 000 children aged 0 to 14 years, among the highest reported in the world.[2] The prevalence of T1D (using 1990 American data) is estimated at 0.19%, or about 1 in 500 children aged 0 to 19 years.[3] This corresponds with our centre’s estimate of 0.15%, or 1477 children aged 0 to 18 years with T1D in British Columbia, based on insulin prescription data from PharmaNet (D.M., unpublished data, 2002). There is increasing evidence that the incidence of T1D is rising worldwide at a rate of about 3% per year, with the greatest increase being observed in children under 5; the reason for this remains unclear.[4]

Risk factors

Both genetic[5] and environmental factor[6] appear to contribute to an individual’s risk of developing T1D. As a result, first-degree relatives (parents, siblings, and offspring) of those with T1D have an increased lifetime risk of developing the disease, variably estimated at 5% to 7%, or approximately 30 times the risk for the general population. The risk rises to about 15% for an HLA-identical sibling and to 50% for an identical twin.[5]

There are now more than 20 genes that have been linked with susceptibility—as well as some that are linked with resistance—to developing T1D.[5] The locus most strongly associated with disease risk lies in the HLA complex on chromosome band 6p21.3. The presence of certain HLA-DR and particularly HLA-DQ alleles can explain about half of the increased risk for developing T1D seen in relatives of those with the disease.

Environmental factors are also felt to play a role in the initiation of the autoimmune process leading to T1D, although these remain largely uncharacterized.[6] Current interest is focused on the role of breastfeeding and early exposure to cow’s milk proteins. Several large epidemiological studies, primarily from Finland, have suggested that exclusively breastfed babies are statistically less likely to develop T1D than those exposed early in life to infant formulas containing cow’s milk protein.[7] Other possible environmental triggers include viral infections (e.g., rubella and coxsackie B) contracted in early childhood or even in utero; nitroso compounds (i.e., those containing the nitrosyl group) in food and drinking water; borderline vitamin D deficiency; and drinking water with an acidic pH. Recent evidence suggests that the autoimmune process begins early in life in many—if not most—individuals who develop T1D.

Progression of autoimmune response

It is now known that there is often a long, presymptomatic phase of T cell lymphocytic infiltration of the pancreatic islets (called “insulitis”) before hyperglycemia and the clinical symptoms of diabetes become manifest.[1] At the time of diagnosis, most patients with T1D have some residual beta cell function, which is often reversibly suppressed by high blood glucose levels (so-called glucose toxicity). Once euglycemia is restored, these surviving beta cells begin to function and secrete insulin during what is known as the “honeymoon period.” However, in most patients, the autoimmune process continues its inexorable destruction of these cells, eventually leading to total dependence on full replacement doses of insulin.

A number of autoantibodies have been identified in the blood of individuals during the presymptomatic phase.[8] Initially, islet cell autoantibodies (ICA) were seen in more than 90% of individuals who eventually developed T1D. Subsequently, more specific antibodies to three beta cell proteins (insulin, glutamic acid decarboxylase [GAD], and tyrosine phosphatase [IA2, or ICA512]) were identified in serum of patients with T1D. In research settings, the quantitation of these blood markers of autoimmunity in first-degree relatives—sometimes in conjunction with determination of HLA-DQ haplotypes and a clinical estimate of residual beta cell function—can be used to predict with great accuracy the future risk for development of T1D.

This information sets the stage for the implementation of strategies to prevent the development of disease in high-risk individuals. These strategies can be classified as either primary or secondary prevention. Primary prevention refers to both the eradication of environmental factors leading to the initiation of beta cell production and to the arrest of the already-present autoimmune process. Secondary prevention refers to the preservation of residual beta cell function and endogenous insulin production in those with new-onset T1D, in order to prolong the honeymoon period. (The most important primary and secondary prevention studies that have been performed to date or are currently in progress can be seen in the Table.)

Primary prevention

Most of the primary prevention trials that have been undertaken in humans are based on data from animal models, especially the non-obese diabetic mouse and the BioBreeding diabetes-prone rat. These two inbred strains develop autoimmune diabetes at very high rates. A number of dietary and environmental interventions have been demonstrated to influence—both positively and negatively—the onset of disease in these animals. Most of the large human prevention trials were preceded by smaller pilot studies in humans.

DPT-1

Based on some encouraging animal data and several small human pilot studies, the Diabetes Prevention Trial—Type 1 Diabetes DPT-1, (www.niddk.nih.gov/patient/dpt_1/dpt_1.htm) tested the hypothesis that exogenous insulin administration could delay or prevent the development of T1D.[8] This large study was sponsored chiefly by the US National Institutes of Health (NIH), and it involved the participation of a number of centres from across North America, including one at the University of British Columbia. More than 97 000 first-degree relatives were screened for the presence of ICA. Seropositive subjects (3.6% of those screened) were then offered further testing for the presence of insulin autoantibodies and for residual beta cell function using an intravenous glucose tolerance test. This allowed subjects to be stratified into three risk categories: ineligible (<25% chance of developing T1D within 5 years), intermediate risk (25% to 50% chance), and high risk (>50% chance).

The DPT-1 study actually consisted of two primary prevention arms, the first for intermediate-risk individuals and the second for high-risk individuals. A total of 339 high-risk subjects were randomized to receive either close monitoring only (i.e., no intervention) or parenteral insulin (4 days of intravenous insulin annually and twice-daily subcutaneous insulin). All subjects were tested at 6-month intervals for the presence of diabetes, for a median follow-up time of 3.7 years. The results of this arm of the DPT-1 trial were published recently.[9] Unfortunately, the parenteral insulin intervention had no effect on the rate of development of T1D in these subjects (15% per year in both groups). However, the study did confirm the accuracy of the prediction models employed to estimate risk, and this information will be quite useful for developing novel intervention trials in the future. Most importantly, the majority of subjects who did progress to T1D during the trial were diagnosed while still asymptomatic, thus reducing their risk of developing ketoacidosis and its complications.

Another 372 subjects in the DPT-1 intermediate-risk group were randomized in a double-blinded fashion to receive either oral insulin or a placebo. They were likewise tested at 6-month intervals for the presence of diabetes. Results of this arm of the study were announced at the Annual Meeting of the American Diabetes Association in June 2003. As with the parenteral insulin arm, intervention with oral insulin did not affect the rate of development of T1D in these subjects (9% per year in both the experimental and the control group). As before, the prediction model used to assign patients to the intermediate-risk group proved to be quite accurate.

Nondiabetic subjects from both the oral and the parenteral intervention arms will continue to be followed at 6-month intervals. This is to ensure that those who do develop diabetes are diagnosed early in the course of their disease, and to allow researchers to continue collecting data on the natural course of the development of disease in ICA-positive subjects.

ENDIT

Beginning about the same time as the DPT-1 study, the European Nicotinamide Diabetes Intervention Trial (ENDIT, www.bris.ac.uk/depts/DivMed/endit.html) investigated whether nicotinamide, a group B vitamin, could prevent or delay the development of T1D in ICA-positive first-degree relatives.[10] Again, animal and pilot human data suggested that this might be an efficacious therapy. Over 40 000 first-degree relatives were screened for ICA from 20 European and North American countries. Funded in part by the Medical Research Council of Canada, a number of Canadian centres also participated in this trial (CANENDIT), including the University of British Columbia. Worldwide, a total of 552 seropositive subjects were randomized in a double-blinded fashion to receive either oral nicotinamide or a placebo. Those eligible for the intervention trial were estimated to have a 40% risk of developing T1D over the 5-year observation period. All subjects were monitored at 6-month intervals for the development of T1D. The results of this trial were made public at the Annual Meeting of the European Association for the Study of Diabetes in September 2002. Unfortunately, nicotinamide had no effect on the rate of development of T1D in the intervention arm[10] (www.bris.ac.uk/Depts/DivMed/enditresults.html).

TRIGR

An exciting new primary prevention effort, the Trial to Reduce IDDM in the Genetically at Risk (TRIGR, www.trigrnorthamerica.org), is now underway. The TRIGR study hopes to determine if avoiding cow’s milk protein in infancy can reduce the rate of autoantibody seroconversion and, ultimately, the risk of developing T1D in first-degree relatives. Based on epidemiological evidence of a link between early weaning from breast milk to cow’s milk formulas and the development of T1D,[7] two pilot trials have already been carried out in Finland. After weaning from exclusive breastfeeding, first-degree infants were randomized to receive either a casein hydrolysate formula or a standard cow’s milk formula until the age of 6 to 8 months. Compared with those weaned to standard formula, the infants receiving the predigested formula had a relative risk of 0.34 for developing one or more islet autoantibodies.[11] It is still too early to ascertain whether this will protect those subjects from developing T1D.

In the international TRIGR trial, investigators from 14 countries will attempt to confirm and expand on the findings of these pilot trials. Study organizers hope to recruit more than 6200 expectant mothers worldwide during the next 2 years. Nationally, the Canadian Institutes of Health Research has provided $10 million to the 12 Canadian centres, including one in Vancouver, participating in this project. Mothers who deliver an at-risk infant (i.e., a baby with both an affected first-degree relative and a risk-associated HLA genotype) will be counseled and supported to breastfeed for as long as they desire. If they discontinue breastfeeding before 8 months, their babies will be randomized in a double-blinded fashion to receive one of two weaning formulas. The study formula is a proprietary casein hydrolysate formula (Nutramigen), and the control formula a standard cow’s milk formula (Enfalac, to which some Nutramigen has been added to mask its flavor). Babies will be in close contact with a study nurse and dietitian, and blood will be obtained regularly for the presence of the beta cell autoantibodies predictive of TID. Over a 10-year observation period, subjects will also be followed for the development of T1D, which is expected to occur in about 15% of controls. Eligible subjects from British Columbia are welcome to participate in this study (see box for contact information).

Other studies

Other, smaller primary prevention trials are also underway. The Diabetes Prediction and Prevention Project (www.utu.fi/research/dipp/engdexx.htm) at three centres in Finland and the IntraNasal Insulin Trial at the University of Melbourne in Australia are both investigating intranasal insulin in individuals at genetic and immunological risk for T1D.

Still in early research is a vaccine against T1D. At least one company is said to be working on this. Phase 2 studies of Diamyd, a Swedish GAD protein-based oral vaccine, are currently being conducted in individuals with latent autoimmune diabetes of adulthood.

Gene therapy is also being considered and may some day offer a means of providing protection against the development of T1D.[12]

Secondary prevention

A number of therapies have already been tested in secondary prevention trials in subjects recently diagnosed with T1D. These include plasmapheresis, gluten-fee diet, vaccination with Bacillus Calmette-Guerin (BCG) vaccine and Q fever vaccine, and the use of nicotinamide, cyclosporine A, azathioprine, and corticosteroids. While some of these treatments (e.g., azathioprine and cyclosporine) did result in short-term partial or complete (insulin-free) remission in some patients, none of these measures had a consistent sustained effect (i.e., beyond 24 months) in preserving endogenous insulin production.

However, two recent secondary prevention studies [13,14] offer hope to those with new-onset T1D. Investigators from Israel have published their results using a modified heat-shock protein called DiaPep277 in this population.[13] At 10 months, subjects who had received three subcutaneous injections of the peptide had higher basal C-peptide levels (indicating greater residual beta cell function) and lower insulin requirements than placebo-treated controls. Larger, longer-term studies are underway to confirm these results.

Another report appeared in 2002 from investigators at Columbia University.[14] They used a modified anti-CD3 monoclonal antibody called hOKT3γ1(Ala-Ala) in a similar group of patients newly diagnosed with T1D. Subjects were randomized to receive either no intervention or a 14-day course of the antibody intravenously. Stimulated C-peptide levels were maintained or improved to a greater degree in treated subjects than in controls, and this effect persisted for at least 12 months after diagnosis. Again, confirmatory studies are underway.

Type 1 Diabetes Trial Net

The Type 1 Diabetes TrialNet (www.diabetestrialnet.org) is a new North American diabetes trial network of cooperative clinical research groups. Funded by the US NIH, its main objective is the development of protocols for both primary and secondary prevention of T1D. A number of novel investigational therapies will be tested in both autoantibody-positive first-degree relatives and in individuals newly diagnosed with T1D. The initial therapies to be tested generally fall into the category of “immunomodulatory” compounds, that is, conventional immunosuppressive drugs (e.g., rapamycin, mycophenolate mofetil, interferon-α), monoclonal antibodies (e.g., daclizumab or anti-CD25, anti-CTLA4, anti-CD3), and modified beta cell peptides (e.g., insulin, DiaPep277, GAD). Eligible subjects from British Columbia are invited to participate in these studies (see Box for contact information).

Immune Tolerance Network

The Immune Tolerance Network (www.immunetolerance.org) is another collaborative research effort that aims to accelerate the clinical development of new tolerance therapeutics in a variety of human immune diseases, including T1D. It is funded by both the US NIH and by the Juvenile Diabetes Research Foundation. Secondary prevention protocols are also being developed and implemented through this network.

Conclusions

While no treatment has yet been identified that can interrupt the progression of the autoimmune process leading to type 1 diabetes, the future holds promise that one or more therapies will be developed to help lessen the burden of this chronic disease. Until that time, efforts at increasing our understanding of the genetic and environmental factors associated with the initiation of beta cell destruction are crucial.

Competing interests
None declared.

Table. Past and ongoing prevention trials for type 1 diabetes.

Daniel L. Metzger, MD, FAAP, FRCPC

Dr Metzger is a clinical associate professor in the Department of Pediatrics at UBC, and a pediatric endocrinologist in the Endocrinology & Diabetes Unit at BC’s Children’s Hospital.