Effects of Gluten Free Diet on Type 1 Diabetes Mellitus in Children

Research Strategy A. SPECIFIC AIMS

The continuing rise in incidence of type 1 diabetes mellitus (T1DM) globally has not been matched by successful efforts for prevention or amelioration of the disease. Additionally, a huge gap exists in our understanding of the adjacent gastrointestinal milieu in which the pancreatic islet destructive process occurs and leads to insulin-dependence in T1DM. The largest increase in incidence of type 1 diabetes over the past decade has occurred in children under 5 years of age, a group particularly susceptible to diabetes related morbidity and mortality. Invasive methodologies via finger pricking for glucose monitoring and subcutaneous insulin injection remain the standard of care required for patient survival. Many intervention studies have been conducted based on inadequately substantiated knowledge of the mechanisms (presumed to be exclusively autoimmune) causing permanent beta-cell destruction in type 1 diabetes. Almost all intervention trials in new-onset type 1 diabetes show only transient improvement of insulin production (up to 6 months) with subsequent decline to levels similar to the control arms. Moreover, there are scientific arguments and data which suggest that non autoimmune pathogenic mechanisms may be involved, perhaps in a more toxic and generalized affliction of the pancreas. Recent evidence that beta-cells continue to regenerate in the pancreas of patients with long-standing type 1 diabetes offers hope that the disease process leading to insulin-dependence may be reversed if islet destruction is arrested at a critical window of time.

The link between type 1 diabetes and the gut microbiome is actively being explored. Emerging evidence suggests a role of the gut microbiome in the destruction and protection of islets in type 1 diabetes-prone experimental models. In addition, type 1 diabetes was prevented in genetically established murine models of the disease when these animals were restricted to a gluten-free diet. The specific hypothesis addressed in this application is that a gluten-free diet introduced shortly after diagnosis can reverse or halt the progression of type 1 diabetes in children and adolescents. Our proposed mechanism is that a gluten-free diet preserves endogenous insulin production via altering the gut microbiota towards strains that are inherently less toxic to islets. The study is designed for a comprehensive assessment of the role of a gluten-free diet in children with T1DM. A specific objective is to reduce or modify patients' gut microbiota early in the disease process as a possible means of arresting islet destruction. The specific measurable aims of the current application outline an innovative approach to begin the identification of beta-cell protective effects of specific changes in gut flora that can be therapeutically developed to ameliorate type 1 diabetes.

Specific Aim 1: To determine the effect of a gluten-free diet on endogenous insulin secretion and pancreatic morphology in children and adolescents with new-onset type 1 diabetes.

Specific Aim 2: To determine the effect of a gluten-free diet on the gut flora of patients with new-onset type 1 diabetes.

Resultant data will allow determination of the relation between gut flora and beta- cell function in patients with new-onset diabetes.

Collectively, the proposed studies are innovative for new-onset T1DM in children. A large population of these patients currently receive care at the Pediatric Diabetes Clinics of Loma Linda University (LLU) which currently provide care to approximately 2000 patients with T1DM. On average, three to five patients with new-onset T1DM are diagnosed at LLU Children's Hospital every week. This proposal is consistent with the research theme of the university which is focused on diet and health, and is very much in line with the 'Wholeness' mission of the institution. In collaboration with the new Genomic Center at LLU, we have a unique opportunity to evaluate a novel mechanism of T1DM. In the future, this can lead to a sustained research program on diet as therapy for type 1 diabetes, allow our site to design and pioneer multi-center trials in this field, and optimize our chances to attract generous extramural funding.

The preliminary data from this study will form the basis for an RO1 application.

1. Significance Type 1 diabetes mellitus is the most common metabolic disease of childhood. Every year, approximately twenty thousand children are diagnosed with type 1 diabetes mellitus (T1DM) in the United States alone according to the Centers for Disease Control and Prevention. The incidence of T1DM continues to increase worldwide. The largest increase in global incidence of T1DM over the past two decades has occurred in children under 5 years of age, a group particularly susceptible to diabetes-related morbidity and mortality. As one of the most common chronic illnesses in children, T1DM has been extensively studied and concluded to be a consequence of autoimmune destruction of the insulin-producing pancreatic islet beta-cells in genetically susceptible individuals. Despite the increasing prevalence of type 2 diabetes, T1DM remains the most common form of diabetes in childhood, accounting for approximately two-thirds of new diagnoses of diabetes in patients under 19 years of age in the United States. The acute complications of T1DM include hypoglycemia, ketoacidosis, and severe psychopathology, all of which are potentially fatal. The origins of profoundly disabling long-term complications such as blindness, kidney failure and amputations can be significantly obliterated if glycemic control is maintained early at the onset of the disease.

A large body of data is emerging on the role of the gut microbiome in health and disease. Recent studies have provided some evidence for a role for structural modulation of gut microbiota in the pathogenesis of type 2 diabetes, the type of diabetes that is more common in adults. Similar evidence has been observed in other disorders such as obesity and cardiovascular disease. These studies have suggested that alteration in the composition of intestinal microbiota may be linked to the development of several diseases, but this is not yet clearly implicated in patients with type 1 diabetes.

More recently, special focus was directed to the role of gut bacteria in the destruction versus protection of pancreatic islets. This has first been studied in T1DM-prone experimental models. Since a type-1 diabetes-related protein was identified in wheat , T1DM murine models showed a reduction in diabetes-associated cell subsets when put on a wheat-free diet. Moreover, type 1 diabetes was prevented in genetically established murine models of the disease when they were restricted to a gluten-free diet.

In young children who are genetically susceptible to T1DM, gut bacteria were notably less diverse and less stable in the group which progressed to develop type 1 diabetes autoimmunity. Of interest are reports of increased incidence of type 1 diabetes in children who, on retrospective review, had received proton pump inhibitor therapy (NIH, personal communication). Additionally, there is a published report of remission of T1DM in a 6 year old child maintained by strictly adhering to a gluten-free diet without any insulin therapy. This triggered a multicenter pilot study currently taking place in Northern Europe with preliminary data suggesting that a gluten-free diet can preserve insulin secretion in adults with type 1 diabetes. To our knowledge, none of the patients maintained on a gluten free diet have undergone studies of the gut microbiome to date.

At LLU, three years following diagnosis of type 1 diabetes, one 17-year old patient at Loma Linda University pediatric diabetes clinics has maintained excellent glycemic control despite unusually low insulin requirements (a so called prolonged honeymoon phase) by adhering to a gluten free diet. Two other patients at LLU, ages 3 and 4 years, experienced a significant decline in insulin requirements (to honeymoon doses, i.e. total daily insulin less than 0.5 units per kg), necessitating discontinuation of an insulin pump in one case, following six months of a gluten-free diet.

Taken together, the above data and cases suggested to us that altering the gut microbiota at onset of the disease via a gluten-free diet may significantly preserve endogenous insulin production, thereby minimizing the need for insulin injections and optimizing long-term health in patients with type 1 diabetes.

Access to a large T1DM patient population at Loma Linda University Children's Hospital and clinics offers a unique opportunity for the exploration of this novel dietary approach that may contribute to tipping the balance within the pancreas towards islet function and regeneration. When optimized, this will pave the way for patients with T1DM to achieve insulin independence.

1. Innovation This proposal challenges the traditional and widely accepted autoimmune mechanism for the development of T1DM, and adds to our previous data on the role of ambient direct toxic factors in the development of T1DM in children. In a large case-control study, we previously described a novel significant association between type 1 diabetes and cumulative ambient ozone exposure birth to the time of diagnosis. Additionally, we have already shown that patients under 5 years of age, the group with the largest increase in incidence globally, frequently lack evidence of known type 1 diabetes autoimmune markers. Moreover, the rapid rate of increase in incidence and decline in age of T1DM suggests that aggressive toxic elements rather than a slow autoimmune process play a pivotal role in the pathogenesis of the disease in children.

The project introduces a new potential etiology for type 1 diabetes in humans: the Bacterial Hypothesis. Here, we investigate a novel, outstandingly safe, and potentially curative approach to T1DM, consisting of merely dietary modification to change the gut microbiota towards a more permissive milieu for insulin secretion. Most intervention trials in new-onset T1DM have had an open design, and in many trials the investigational agent's side-effects allowed the participating subjects to know whether or not they received active drug. Thus, interpretation of results of most intervention studies to date cannot be deemed accurate given a probable placebo effect. The current study design incorporates the placebo effect in both study and control subjects. To our knowledge, this is the first randomized placebo controlled trial testing the effect of dietary modification on the gut flora and the glycemic control of children with newly diagnosed type 1 diabetes.

1. Approach

This is a randomized controlled clinical trial testing the effect of altering gut bacteria via a gluten-free diet on the pace and severity of islet destruction in T1DM within six months of diagnosis. The project entails modifying meal content, frequency and rate of consumption, and assessing bacterial gut flora and endogenous insulin production up to one year following enrollment. The current proposal is based on our current level of knowledge of the T1DM disease process, caveats in current intervention studies, and the desirability of interventions with minimum side effects in children.

The objective of the current proposal is to change the gut flora to reduce the bacterial escape in the stomach and duodenum in a way which may have direct detoxification effects on the adjacent pancreas and its insulin producing islet cells. The proposal tests the effect of eliminating gluten from the diet on reducing or changing the gut flora in a manner conducive to insulin secretion by beta cells. As the amount of bacteria entering the duodenum is extremely dependent on the type of food, as well as on the volume of each meal, the established benefit of eating small frequent meals may prove to be particularly useful in patients type 1 DM. The proposed dietary intervention is likely to substantially influence the bacterial flora in the duodenum without inflicting any significant harm to children's health, growth or development.

D1. Specific Aim 1: To determine the effect of a gluten-free diet on endogenous insulin secretion in children and adolescents with new-onset type 1 diabetes.

Subject Recruitment and Sample Size. Subjects diagnosed with type 1 diabetes will be recruited from LLU Children's Hospitals and clinics within the first six months of diagnosis. Participating subjects will be randomized to one of two study groups, each consisting of 25 to 30 subjects to provide 80% power to detect a significant difference in endogenous insulin production (Area under the C-peptide curve in a mixed meal tolerance test).

Inclusion criteria: Ages: 1-17 years; diagnosis of type 1 diabetes within 6 months of enrollment; parent or guardian willing to sign informed consent to comply with basal bolus insulin treatment as well as study procedures and diet guidelines.

Exclusion criteria: Diabetes diagnosed under one year of age; patients with known malabsorption or malnutrition; glycosylated hemoglobin HbA1C value over 9% at three months following diagnosis (a general screen for noncompliance with treatment guidelines).

Study assessment time points are: enrollment, 3 months, 6 months, 9 months, and 12 months.

The only procedures to be conducted beyond the scope of routine care for type 1 diabetes are: stool sample collections at enrollment and at 3, 6, 9 and 12 months; and mixed meal tolerance tests (MMTT) at enrollment, and at 6 and 12 months. MMTT is a standard research tool to assess islet functional reserve by measuring basal and stimulated blood glucose and C-peptide at 30 minute intervals over two hours before (time zero) and after a standard high caloric formula.

At diagnosis, all subjects will undergo routine institutional diabetes protocol laboratory testing which includes diabetes autoimmune antibodies: Glutamic acid decarboxylase, Islet Cell Antibodies, Insulin Autoantibodies, Zinc transporter 8 antibodies, as well as thyroid and celiac antibodies. A fasting lipid panel and thyroid stimulating hormone (TSH) will also be measured at initial evaluation. The level of 25 hydroxyvitamin D will be measured at baseline to screen for malabsorption and nutritional deficiency particularly in view of a suspected role for vitamin D in T1DM.

The standard of care for type 1 diabetes includes quarterly encounters with a diabetes team including physicians, nurses, dietitians, together with quarterly assessment of glucose profile, glycosylated hemoglobin (HbA1C), and appropriateness of insulin treatment regimens with adjustments as needed.

Study subjects will be divided into two groups as follows:

Study group A subjects will receive standard T1DM basal bolus insulin regimens guided by physicians, nurses, and a nutritional consult to ensure adequate knowledge of diabetes diet recommendations and carbohydrate- based insulin dosing. In addition, subjects will receive instructions to consume a "gluten-free" diet (e.g. replace white bread and potatoes with rice, gluten-free bread and other foods). Group A will also be encouraged to increase the time of consumption of each meal and to reduce the volume of meals including the volume of liquids (i.e. a subject can eat the same volume of food desired but in smaller more frequent meals).

Study group B will receive standard T1DM basal bolus insulin regimen guidelines by the same physicians, nurses, and nutritionist but will be given the normal meal recommendations for subjects with T1DM without any gluten intake or volume restrictions. Caloric and carbohydrate daily recommendations will be tailored to age and will not be different between the two study groups.

Subjects who show strong serologic evidence of gluten hypersensitivity will be excluded from the two main study groups and started on a gluten-free diet. Their data will be tracked retrospectively for any significant outcome differences which may be attributable to pre-existing gluten hypersensitivity alone.

Gut bacteria will be studied at every assessment time-point via stool sampling in both study groups.

Glycosylated hemoglobin A1C (HbA1C) will be measured at diagnosis and every three months together with a dietitian's assessment to ensure compliance with designated diet instructions.

Mixed meal tolerance testing for basal and stimulated C-peptide measurements will be done at enrollment, and at 6 months and one year post-enrollment. Peak levels and areas under the curve for stimulated C-peptide will be compared between the intervention and control group.

Every assessment time-point will include stool sampling for characterization of the bacterial flora of the gut over time. Stool testing at each time-point will include regular bacterial strain identification as well as genomic assessment.