Abstract
Introduction
In the last 10 years, over 100 intervention studies have been conducted or are ongoing in new-onset type 1 diabetes (T1D), or in those at risk (multiple autoantibodies against islet antigens). The hitherto lack of causality of the initiating events leading to autoimmunity against β-cells of the pancreas, and heterogeneity of the disease process itself, have not made the path to prevention or cure easy. Recent studies provide more than a modicum of hope—both anti-CD3 and low-dose anti-thymocyte globulin (ATG) have shown some degree of preservation of β-cell function, and an alteration in the course of the disease. Although the mechanisms remain to be fully elucidated, the role of the immune system (as best determined from peripheral blood studies) in sustenance, breakdown of tolerance, and β-cell death is being closely studied.
Different therapies may be more effective at different stages of the disease. For example, some agents such as antiCD3 show efficacy in stage 2 T1D (two or more islet autoantibodies, normoglycemia), which may suggest a need for the presence of active immunity in order for the drug to be efficacious. Whether anti-CD3 and/or ATG will be effective earlier in the disease remains to be determined. In addition, it is likely that as we determine responders and nonresponders to therapy (identified from gene expression and immune signatures), approaches will need to be individualized.
Key Articles Reviewed for the Article
Herold KC, Bundy BN, Long SA, Bluestone JA, DiMeglio LA, Dufort MJ, Gitelman SE, Gottlieb PA, Krischer JP, Linsley PS, Marks JB, Moore W, Moran A, Rodriguez H, Russell WE, Schatz D, Skyler JS, Tsalikian E, Wherrett DK, Ziegler AG, Greenbaum CJ; for the Type 1 Diabetes TrialNet Study Group
Haller MJ, Schatz DA, Skyler JS, Krischer JP, Bundy BN, Miller JL, Atkinson MA, Becker DJ, Baidal D, DiMeglio LA, Gitelman SE, Goland R, Gottlieb PA, Herold KC, Marks JB, Moran A, Rodriguez H, Russell W, Wilson DM, Greenbaum CJ; on behalf of the Type 1 Diabetes TrialNet ATG‐GCSF Study Group
Pesenacker AM, Chen V, Gillies J, Speake C, Marwaha AK, Sun A, Chow S, Tan R, Elliott T, Dutz JP, Tebbutt SJ, Levings MK
Zhou X, Zhang S, Yu F, Zhao G, Geng S, Yu W, Wang XY, Wang B
Vatanen T, Franzosa EA, Schwager R, Tripathi S, Arthur TD, Vehik K, Lernmark Å, Hagopian WA, Rewers MJ, She JX, Toppari J, Ziegler AG, Akolkar B, Krischer JP, Stewart CJ, Ajami NJ, Petrosino JF, Gevers D, Lähdesmäki H, Vlamakis H, Huttenhower C, Xavier RJ
Jacobsen LM, Larsson HE, Tamura RN, Vehik K, Clasen J, Sosenko J, Hagopian WA, She JX, Steck AK, Rewers M, Simell O, Toppari J, Veijola R, Ziegler AG, Krischer JP, Akolkar B, Haller J; on behalf of the TEDDY Study Group
An anti-CD3 antibody, teplizumab, in relatives at risk for type 1 diabetes
Herold KC1, Bundy BN2, Long SA5, Bluestone JA6, DiMeglio LA7, Dufort MJ5, Gitelman SE6, Gottlieb PA8, Krischer JP2, Linsley PS5, Marks JB3, Moore W9, Moran A10, Rodriguez H2, Russell WE11, Schatz D4, Skyler JS3, Tsalikian E12, Wherrett DK13, Ziegler AG14, Greenbaum CJ5; for the Type 1 Diabetes TrialNet Study Group
1Departments of Immunobiology and Internal Medicine, Yale University, New Haven, CT; 2Departments of Epidemiology and Pediatrics, University of South Florida, Tampa, FL; 3Department of Medicine, University of Miami, Miami, FL; 4Department of Pediatrics, University of Florida, Gainesville, FL; 5Benaroya Research Institute, Seattle, WA; 6Diabetes Center, University of California at San Francisco, San Francisco, CA; 7Department of Pediatrics, Indiana University, Indianapolis, IN; 8Barbara Davis Diabetes Center, University of Colorado, Anschultz, CO; 9Children's Mercy Hospital, Kansas City, MO; 10Department of Pediatrics, University of Minnesota, Minneapolis, MN; 11Department of Pediatrics and Cell and Developmental Biology, Vanderbilt University, Nashville, TN; 12Department of Pediatrics, University of Iowa, Iowa City, IA; 13Hospital for Sick Children, University of Toronto, Toronto, Canada; 14Forschergruppe Diabetes, Technical University Munich, at Klinikum rechts der Isar, Munich, Germany
Background
Teplizumab is an Fc-receptor nonbinding anti-CD3 monoclonal antibody, and has shown promise in reducing β-cell loss for as long as 7 years, in recent onset T1D patients. The favorable immune modulating action of teplizumab in recent onset T1D is thought to be due to increased levels of exhausted CD8 T-cells, the major effector T-cells that destroy the β-cells. The current study, a phase 2, randomized, placebo-controlled, double-blind study, was undertaken to evaluate whether teplizumab will decrease disease progression of clinical T1D in stage 2 islet autoantibody positive nondiabetic subjects.
Methods
Participants were enrolled from TrialNet with the following criteria: (1) nondiabetic relatives of T1D ≥8 years of age, (2) presence of two or more autoantibodies within 6 months prior to randomization, and (3) evidence of dysglycemia defined as a fasting glucose 110125 mg/dL (6.1–6.9 mmol/L), or a 2-hour oral glucose tolerance test (OGTT) stimulated peak blood glucose of 140 mg/dL (7.8 mmol/L) to <200 mg/dL (11.1 mmol/L), or an intervening postprandial glucose level at 30, 60, or 90 minutes of >200 mg/dL on two occasions, within 52 days before enrollment. Randomization to teplizumab (n=44) or placebo (n=32) was done based on TrialNet site, age (<8 years or ≥18 years), and second OGTT result before treatment. Teplizumab was administered at a dose of 51, 103, 207, 413, and 826 μg/m2 BSA on days 0, 1, 2, 3, and 4–13, respectively. The primary endpoint was clinical diagnosis of diabetes based on random screening, and OGTT blood glucose values according to the study protocol.
Results
The median follow-up was 745 days with more than 75% of participants followed for more than 3 years. Nineteen of 44 (43%) of the teplizumab group and 23/32 (72%) of the placebo group were diagnosed with T1D. The annualized rates of T1D diagnosis were 14.9% per year in the teplizumab group and 35.9% per year in the placebo group. The median times to diagnosis were 48.4 months in the teplizumab group and 24.4 months in the placebo group (hazard ratio 0.41 [95% confidence interval 0.22–0.78]; two-sided t-test P=0.006). The largest effect of teplizumab treatment was found in the first year: diabetes was diagnosed in only 3 of 44 participants (7%) in the teplizumab group, in contrast to 14 of 32 participants (44%) in the placebo group (unadjusted hazard ratio 0.13 [95% CI 0.05–0.34]). The response to teplizumab was also greater among participants whose C-peptide responses to the OGTT at baseline were below the median (1.75 nmol/L) than among those whose responses were above the median (hazard ratio 0.19 [95% CI 0.08–0.40]). TIGIT+KLRG1+CD8+ T-cells were higher in the teplizumab group at months 3 and 6 than in the placebo group. The therapy was well tolerated with only minor side effects.
Conclusion
A 2-week course of teplizumab delayed progression to clinical T1D in high-risk participants for a median of two years.
This study showed that administration of teplizumab in dysglycemic, autoantibody positive but nonsymptomatic subjects delays the clinical diagnosis of T1D for as long as 48 months. The authors also stated that patients in the teplizumab group who had lower baseline C-peptide (possibly reflecting a more active or longer autoimmune response) had a more favorable response compared with those with baseline C-peptide above median value. With the study reporting a more favorable response to teplizumab in those negative for ZnT8 autoantibodies and HLA-DR3 but positive for HLA-DR4, the authors also commented on the potential use of these parameters to identify patients who might benefit the most from teplizumab treatment in stage 2 of the disease course.
The requirement for the presence of active autoimmune processes underlying the function of anti-CD3 monoclonal antibodies has been shown in preclinical studies. With this in mind, the authors have hypothesized that using teplizumab in stage 1 patients (normoglycemic islet autoantibody positive individuals) may not be as efficacious as using it in stage 2 patients. It is unknown if this could be related to lower circulating numbers of effector T-cells, the target of teplizumab, in stage 1 versus stage 2 disease. Thus, it will be informative to determine if there are differences in effector T-cell frequencies between stage 1 and stage 2 T1D, and if a lower number of circulating effector T-cells in stage 1 disease may lead to a poorer response compared with stage 2. However, given the side effects of lymphopenia and Epstein-Barr virus reactivation, it is difficult to comment if such a study will be undertaken in subjects with stage 1 disease, who understandably will be much younger than the age cut-off of 8 years in this study.
Low-dose anti-thymocyte globulin (ATG) preserves β cell function and improves HbA1c in new-onset type 1 diabetes
Haller MJ1, Schatz DA1, Skyler JS2, Krischer JP3, Bundy BN3, Miller JL3, Atkinson MA1, Becker DJ4, Baidal D2, DiMeglio LA5, Gitelman SE6, Goland R7, Gottlieb PA8, Herold KC9, Marks JB2, Moran A10, Rodriguez H3, Russell W11, Wilson DM12, Greenbaum CJ13; on behalf of the Type 1 Diabetes TrialNet ATG‐GCSF Study Group
1University of Florida, Gainesville, FL; 2Diabetes Research Institute, University of Miami Miller School of Medicine, Miami, FL; 3University of South Florida, Tampa, FL; 4University of Pittsburgh, Pittsburgh, PA; 5Indiana University, Indianapolis, IN; 6University of California, San Francisco, San Francisco, CA; 7Columbia University, New York, NY; 8University of Colorado Barbara Davis Center for Childhood Diabetes, Aurora, CO; 9Yale University, New Haven, CT; 10University of Minnesota, Minneapolis, MN; 11Vanderbilt University, Nashville, TN; 12Stanford University, Palo Alto, CA; 13Benaroya Research Institute, Seattle, WA
Background
Recent studies have highlighted that combination therapy may be more beneficial than monotherapy in slowing or halting autoimmunity in T1D. The current study is an extension of a pilot trial where a combination of low-dose ATG (2.5 mg/kg) and pegylated granulocyte colony-stimulating factor (GCSF; 6 mg subcutaneously every 2 weeks for 6 doses) administered to patients with established T1D (duration of 4 months to 2 years) preserved C-peptide. The patients treated with low-dose ATG/GCSF showed a relative increase of regulatory T-cells (Tregs) in circulation. However, the pilot study did not include an arm with subjects receiving low-dose ATG monotherapy. The current study was conducted to explore the benefits of either the combined therapy or low-dose ATG monotherapy in new-onset T1D (defined as diagnosed T1D <100 days).
Methods
T1D patients (n=89) of <100 days duration between the ages of 12–45 years from 14 TrialNet sites were enrolled. In order to qualify for enrollment, patients were also required to have at least one T1D-related autoantibody (microinsulin autoantibodies, tested only if duration of insulin therapy was <7 days; glutamic acid decarboxylase-65 autoantibodies, islet-cell antigen-512 autoantibodies [ICA-512Ab], zinc transporter 8 autoantibodies, or islet-cell autoantibodies [ICA]), as well as stimulated C-peptide levels ≥0.2 nmol/L after a mixed-meal tolerance test conducted at least 21 days after diagnosis of T1D and within 37 days of randomization. Patients were randomized in a 1:1:1 ratio to ATG/GCSF (n=29), ATG/GCSF placebo (n=29), ATG placebo/GCSF placebo (n=31). The study was double-masked.
ATG or placebo (2.5 mg/kg) was administered as two divided intravenous infusions of 0.5 mg/kg and 2 mg/kg on consecutive days. Premedication for ATG/placebo infusions included oral diphenhydramine 1.25 mg/kg up to 50 mg, oral acetaminophen 15 mg/kg up to 650 mg, and intravenous methylprednisolone or placebo at 0.25 mg/kg. Patients who developed serum sickness received prednisone as per protocol. GCSF or placebo was administered every 2 weeks for a total of 6 doses at a dose of 6 mg subcutaneously or, if the patient weighed <44.5 kg, a dose of 100 μg/kg. The primary outcome was comparison of the area under the curve (AUC) of stimulated C-peptide response over the first 2 hours of a 4-hour mixed-meal tolerance test at the 12-month visit.
Results
The 1-year mean AUC C-peptide was significantly higher in subjects treated with ATG (0.646 nmol/L) versus placebo (0.406 nmol/L) (P=0.0003) but not in those treated with ATG/GCSF (0.528 nmol/L) versus placebo (P=0.031). Glycated hemoglobin (HbA1c) level was significantly lower in both the ATG/GCSF (P=0.011) and ATG groups (0.002) versus placebo at 1 year. However, there were no differences in insulin use between the treated groups and placebo. Patients in the treatment arm had reduced total lymphocytes and CD4+ T-cells, but preservation of CD8+ T-cells that resulted in a reduction in the CD4/CD8 T-cell ratio compared with participants who received placebo. There were no serious side effects observed in the treatment groups.
Conclusion
A single low-dose administration of ATG in new-onset T1D preserved C-peptide and reduced HbA1c when compared with subjects treated with placebo 1 year after therapy.
One of the interesting outcomes from this study is the fact that low-dose ATG as a monotherapy is equally effective as ATG/GCSF combination therapy in preserving C-peptide. This is beneficial in terms of reducing the overall side effect profile and costs associated with a combined therapy. Although there were mild side effects associated with ATG treatment, including cytokine release and serum sickness, these all resolved spontaneously or were easily managed.
Low-dose ATG treated subjects had better C-peptide preservation compared with high dose ATG (6.5 mg/kg), which was evaluated in the START (Study of Thymoglobulin to Arrest Type 1 Diabetes) trial. It has been hypothesized that with a higher dose, there may be a greater release of inflammatory cytokines requiring subsequent use of high dose glucocorticoids, which in combination may worsen β-cell function. Furthermore, this low-dose ATG study showed relative preservation of CD8+ T-cells with reduction in CD4+ T-cells yielding a low CD4/CD8 ratio. This is in contrast to the START study, where there was overall depletion of T-cells including Tregs. In comparison to other immunomodulatory agents evaluated in other clinical trials, low-dose ATG provided a higher (57%) increase in AUC C-peptide over placebo in new-onset T1D patients (16% for rituximab and 23% for abatacept at 1 year). However, the most important question that remains is whether low-dose ATG will delay the onset of clinical T1D when administered to high risk, multiple autoantibody–positive subjects with stage 1 or 2 disease. We await these studies with zeal.
Treg gene signatures predict and measure type 1 diabetes trajectory
Pesenacker AM1, Chen V2, Gillies J1, Speake C3, Marwaha AK4, Sun A1, Chow S5, Tan R6, Elliott T7, Dutz JP5, Tebbutt SJ2, Levings MK1
1Department of Surgery, University of British Columbia (UBC), and BC Children's Hospital Research Institute, Vancouver, British Columbia, Canada; 2Department of Medicine and Centre for Heart Lung Innovation, UBC, and Prevention of Organ Failure Centre of Excellence, St. Paul's Hospital, Vancouver, British Columbia, Canada; 3Diabetes Clinical Research Program, Benaroya Research Institute, Seattle, WA; 4Department of Molecular Genetics, University of Toronto, Toronto, Ontario, Canada; 5Department of Dermatology, UBC, and BC Children's Hospital Research Institute, Vancouver, British Columbia, Canada; 6Department of Pathology, Sidra Medicine, Weill Cornell Medicine, Doha, Qatar; 7Department of Medicine, UBC, and BC Diabetes, Vancouver, British Columbia, Canada
Background
Studies performed in mice and humans over the past 10 years have suggested a critical role of FOXP3+ Tregs in ameliorating islet autoimmunity and consequently preventing or delaying T1D. In new-onset T1D studies, some immunoregulatory agents have shown partial preservation of β-cell function, most noticeably in the first few months posttherapy. It is clear that there are definite responders and nonresponders; however, it is unclear whether these differences may be related to variations in the autoimmune processes and, in particular, due to inherent differences in Treg function. In a previous study, the authors developed a composite biomarker assay that measured expression of 37 genes, which discriminated between Tregs and conventional T-cells (Tconvs). They demonstrated that Tregs from pediatric new-onset T1D patients have a significantly altered gene signature. In this study, the authors sought to determine whether Treg gene signatures (TGSs) could further identify Treg alterations in a bigger cohort of adult patients with new-onset and established T1D, and investigated the predictive power of TGS-based biomarkers to determine the rate of C-peptide decline in treated and placebo patients.
Methods
Peripheral blood was collected from participants in three separate trials: UST1D (Ustekinumab in Patients with New-onset T1D), START (Study of Thymoglobulin to Arrest Type 1 Diabetes), and TIDAL (Type 1 Diabetes and Alefacept), and stored as peripheral blood mononuclear cells. Participants (n=20 for USTID, and n=41 for START and TIDAL) were matched for age and sex (n=10 for UST1D and n=37 for START and TIDAL). CD4+CD25hiCD127lo Tregs were isolated from cryopreserved PBMC samples. Sorted Tregs and whole PBMC lysates were analyzed for TGS using nanoString technique to compare individuals with T1D (from TIDAL and START), healthy controls (from TIDAL and START), or new-onset T1D receiving ustekinumab (αIL-12/23p40, from UST1D).
Results
TGS distinguished T1D from healthy control subjects. When genotypes at various T1D risk associated single nucleotide polymorphisms (SNPs) were included, the rate of C-peptide fall was more predictable. Treg cell expression of four genes incorporated into an algorithm either increased (C8ORF70, PMSL11, STAM) or decreased (ICA1), with progressively worsening C-peptide, and discriminated individuals with rapid versus slow decline. This algorithm was able to stratify participants with rapid versus slow C-peptide decline in the ustekinumab trial. In contrast, analysis of PBMC lysates was ineffective at discriminating between these, suggesting that intrinsic alterations in Tregs, rather than an overall change in immunoregulatory balance, may be better able to predict future disease course.
Conclusion
TGS may be a novel approach to the long-standing challenge of predicting and measuring T1D disease trajectory and outcomes of immunotherapeutic interventions.
Studies over the last few years have revealed that T1D is a heterogeneous disease, which may explain the variability in responses after immunotherapeutic interventions either before or after clinical disease. Treg alterations together with other yet unknown different gene signatures may be contributory. The authors of this complex mechanistic study have commented that measuring or monitoring TGS in patients at risk for T1D may be helpful to prognosticate disease course and identify responders to immunotherapeutic interventions. It remains to be seen if the gene signatures in the algorithms used in the study can be further evaluated with a view to manipulate disease course.
Tolerogenic vaccine composited with islet derived multi-peptides and cyclosporin A induces pTreg and prevents type 1 diabetes in murine model
Zhou X1, Zhang S2, Yu F2, Zhao G2, Geng S2, Yu W2, Wang XY1, Wang B2,3
1Institutes of Biomedical Sciences, Shanghai, China; 2Key Laboratory of Medical Molecular Virology of the Ministry of Health and the Ministry of Education, Shanghai Basic Medical College, Shanghai, China; 3Children Hospital of Fudan University, Shanghai, China
Background
Various studies, including antigen-based therapies, have attempted to either upregulate Treg function or increase Treg concentrations directly in patients with T1D with the goal of increasing tolerogenicity and thus suppressing autoimmunity. The authors of this study have sought to evaluate if inducing an immunosuppressive environment (using cyclosporine [CsA]) during the presentation of autoantigens by dendritic cells (DCs) to T-cells, would enhance tolerogenicity of autoreactive T-cells.
Methods
Four islet-derived peptides (GAD65206-220, GAD65536-550, insulin B9-23, and insulin C17-A1), were combined with CsA (GAD-IN+CsA). Based on BALB/c mice dosing studies, 20 μg of GAD-IN +10μg CsA regimen achieved the highest increase in CD4+CD25+Foxp3+ peripherally induced Tregs (pTregs). NOD male mice (6–8 weeks old) were administrated 50 mg/mL streptozotocin daily for 5 days, and then administered 20 μg GAD-IN+CsA vaccine subcutaneously on days 0, 6, and 12. Control mice were administered mixed peptides (GAD-IN) alone, CsA alone, or vehicle. A subset of the NOD mice were sacrificed 24 hours after vaccination, while the rest underwent an OGTT. CD4+CD25+Foxp3+ pTreg or CD4+CD25+Foxp3− effector T-cells were isolated from spleen and pancreatic lymph nodes of the sacrificed animals and analyzed by flow cytometry.
Results
A higher frequency of CD4+CD25+Foxp3+ pTregs was observed in pancreatic lymph nodes of the 20 μg GAD-IN+CsA immunized NOD mice at 24 hours compared with that of control mice (multipeptide mix without CsA, CsA alone, or vehicle). There was increased expression of the immunoregulatory cytokines interleukin (IL)-10 and tumor growth factor-β in these induced Foxp3+ pTregs. In contrast, there was decreased expression of proinflammatory interferon-γ, tumor necrosis factor-α, and IL-2 by antigen specific CD4+ effector T-cells in the GAD-IN+CsA treated animals compared to control mice. Over 70% of the immunized NOD mice remained diabetes free to the end of the study, while only 25% of the controls remained diabetes free. Evaluation of islet tissue sections revealed that no insulitis to mild peri-insulitis was present in the GAD-IN+CsA vaccinated mice, whereas most of the control mice exhibited more severe insulitis. In vitro analysis showed that the GAD-IN+CsA vaccine modulates DCs by upregulating their expression of IL-10 and downregulating CD40. Upon in vitro co-culture, these tolerogenic DC, converted CD4+CD25− T-cells to CD4+Foxp3+ Tregs. Ki67 expression (indicating proliferation) by these in vitro induced Tregs was significantly higher upon further stimulation with cognate antigen.
Conclusion
Islet-derived multipeptides with CsA induce tolerogenicity in DCs, increase antigen-specific pTregs in NOD mice, and prevent T1D.
Over the past few years, there have been numerous antigen-based studies aimed at preventing T1D. Most of these studies have been performed in either recent onset T1D patients or high-risk autoantibody positive (stage 1–2) T1D subjects. None of these studies have been able to prevent progression of diabetes either in the preclinical or clinical phase. This mouse study explored the immunomodulating effect of CsA in inducing tolerogenicity during the interaction of antigen presenting DCs with islet antigen-specific T-cells. The results are encouraging with potential for translation into human studies. However, as with most NOD mice studies, beneficial effects have not always been successfully extrapolated into human studies, whether due to the quality of the model, differences in pathobiology, optimal drug dosing, etc. Nonetheless, the findings from this study may provide new avenues for antigen-based studies in human T1D patients.
The human gut microbiome of early onset type 1 diabetes in the TEDDY study
Vatanen T1, Franzosa EA1,2, Schwager R2, Tripathi S1, Arthur TD1, Vehik K3, Lernmark Å4, Hagopian WA5, Rewers MJ6, She JX7, Toppari J8,9, Ziegler AG10,11,12, Akolkar B13, Krischer JP3, Stewart CJ14,15, Ajami NJ14, Petrosino JF14, Gevers D1,19, Lähdesmäki H16, Vlamakis H1, Huttenhower C1,2, Xavier RJ1,17,18
1Broad Institute of MIT and Harvard, Cambridge, MA; 2Department of Biostatistics, Harvard T. H. Chan School of Public Health, Boston, MA; 3Health Informatics Institute, Morsani College of Medicine, University of South Florida, Tampa, FL; 4Department of Clinical Sciences, Lund University/CRC, Skåne University Hospital SUS, Malmo, Sweden; 5Pacific Northwest Research Institute, Seattle, WA; 6Barbara Davis Center for Childhood Diabetes, University of Colorado, Aurora, CO; 7Center for Biotechnology and Genomic Medicine, Medical College of Georgia, Augusta University, Augusta, GA; 8Department of Pediatrics, Turku University Hospital, Turku, Finland; 9Institute of Biomedicine, Research Centre for Integrative Physiology and Pharmacology, University of Turku, Turku, Finland; 10Institute of Diabetes Research, Helmholtz Zentrum München, Munich, Germany; 11Forschergruppe Diabetes, Technische Universität München, Klinikum Rechts der Isar, Munich, Germany; 12Forschergruppe Diabetes e.V. at Helmholtz Zentrum München, Munich, Germany; 13National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD; 14Alkek Center for Metagenomics and Microbiome Research, Department of Molecular Virology and Microbiology, Baylor College of Medicine, Houston, TX; 15Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK; 16Department of Computer Science, Aalto University, Espoo, Finland; 17Gastrointestinal Unit, Center for the Study of Inflammatory Bowel Disease, and Center for Computational and Integrative Biology, Massachusetts General Hospital and Harvard Medical School, Boston, MA; 18Center for Microbiome Informatics and Therapeutics, Massachusetts Institute of Technology, Cambridge, MA; 19Present address: Janssen Human Microbiome Institute, Janssen Research and Development, Cambridge, MA
Background
The gut microbiome of infants differs from adults, undergoes change in the first few years of life, and is influenced by multiple environmental factors. Previous studies have reported an association between alterations in the gut microbiome and T1D. Associations linking decreased diversity of gut flora and islet autoimmunity (IA) have also been suggested. NOD mice fed specialized diets leading to higher bacterial release of the short chain fatty acids (SCFA) butyrate and acetate have shown to be protected from T1D. The Environmental Determinants of Diabetes in the Young (TEDDY) group study sought to characterize potential changes in the gut microbiome in the largest newborn genetically at-risk cohort associated with the subsequent development of IA and T1D.
Methods
This nested case-control study evaluated the gut microbiome of patients from the TEDDY study cohort from the age of 3 months who progressed to either IA (n=418) or T1D (n=114). Case-control pairs were matched 1:1 by clinical center, sex, and family history of T1D, with controls being those participants who had not developed persistent and confirmed IA (IA on 2 or more occasions) or T1D by the time the case to which they were matched had developed IA or T1D, within ±45 days of the event time. Stool samples (n=10,903 samples, n=783 subjects) were collected monthly from participants 3–48 months of age, then every 3 months until 10 years of age, and then biannually until development of IA or T1D. Metagenomic sequencing was performed for all samples along with 16S rRNA sequencing. This was followed by taxonomic and functional profiling of the metagenome samples. Common and rare life events from infancy through 5 years of age, as clinical covariates relevant to the gut microbiome, were recorded.
Results
A high degree of heterogeneity (both taxonomically and functionally) in the gut microbiome was seen in the earliest stool samples of all subjects. However, the dominant species were Bifidobacterium bifidum, B. breve, and B. longum or of the phylum Proteobacteria. Strain specific carriage of genes involved in human milk oligosaccharide utilization within a subset of B. longum was specifically seen in breastfed infants. Antibiotic use depleted Bifidobacterium members with the exception of B. longum and B. breve. Healthy controls harbored higher levels of Lactobacillus rhamnosus (q=0.055) and B. dentium (q=0.054) compared with IA cases, whereas IA cases had higher levels of Streptococcus group mitis/oralis/pneumoniae species (q=0.11). T1D-matched controls had more Streptococcus thermophilus (q=0.078) and Lactococcus lactis (q=0.094) species, whereas T1D cases harbored higher levels of Bifidobacterium pseudocatenulatum (q=0.078), Roseburia hominis (q=0.11), and Alistipes shahii (q=0.14). Finnish IA cases had more Streptococcus group mitis/oralis/pneumoniae species (q=0.0008), while IA-matched controls from Colorado had more Streptococcus thermophilus (q=0.0059), and Swedish IA cases harbored more Bacteroides vulgatus (q=0.090).
Conclusion
The microbiomes of control children harbored more bacteria expressing genes related to fermentation and SCFA biosynthesis, but these were not consistently associated with particular taxa across geographically diverse clinical centers, suggesting that microbial factors associated with T1D are taxonomically diffuse but functionally coherent. SCFA may have a potential protective effect on early onset human T1D.
An earlier observational study by the TEDDY group reported a decreased risk for IA with probiotic supplementation in the first month of life, further suggesting a role of gut microbiota on T1D development. This study group sought to characterize the gut microbiota longitudinally during the early years in those genetically at risk for T1D, and to determine whether a difference existed in those who develop IA and/or progress to T1D.
Metagenomic sequencing of stool samples from enrolled participants revealed a high degree of diversity in the gut bacterial population, some of which may be due to geographical location. However, despite the microbial diversity, the presence of various bacterial strains carrying genes responsible for fermentation and biosynthesis of SCFA was seen more often in those without IA. Whether this would imply that SCFA supplementation in those with high risk for T1D is protective is too premature a conclusion. Despite studies about the use of antibiotics with subsequent perturbations in gut microbiota, especially a depletion in the Bifidobacterium species during the breastfeeding period, it is still not clear if there are durable effects of antibiotics on gut microbiome, subsequent development of T1D-related autoimmunity, and disease progression. An earlier study by the same group did not find any association between antibiotic use and development of T1D autoimmunity in those with high genetic risk for T1D. Nonetheless, this study has yet again highlighted that gut immunity and questions surrounding how the gut microbiota may shape immune responses in early childhood, especially in the development of autoimmune disorders such as T1D, cannot be ignored. The potential association of differences in microbial populations with differences in disease presentation, such as reactivity of the first appearing autoantibody, subsequent appearance of other autoantibodies, and disease progression, should also be studied. Variables such as geography, ethnicity, and disease-associated alleles, as well as other microbes including viruses and fungi, may also need to be assessed. Longitudinal follow up of this population may reveal novel potentially causal, predictive, or protective microbial components in T1D.
Predicting progression to type 1 diabetes from ages 3 to 6 in islet autoantibody positive TEDDY children
Jacobsen LM1, Larsson HE2, Tamura RN3, Vehik K3, Clasen J3, Sosenko J4, Hagopian WA5, She JX6, Steck AK7, Rewers M7, Simell O8, Toppari J8,9, Veijola R10, Ziegler AG11, Krischer JP3, Akolkar B12, Haller J1, on behalf of the TEDDY Study Group
1Department of Pediatrics, University of Florida, Gainesville, FL; 2Department of Clinical Sciences Malmö, Lund University, Skåne University Hospital SUS, Malmö, Sweden; 3Health Informatics Institute, Morsani College of Medicine, University of South Florida, Tampa, FL; 4Division of Endocrinology, University of Miami, Miami, FL; 5Pacific Northwest Diabetes Research Institute, Seattle, WA; 6Center for Biotechnology and Genomic Medicine, Medical College of Georgia, Augusta University, Augusta, GA; 7Barbara Davis Center for Childhood Diabetes, University of Colorado, Denver, CO; 8Department of Pediatrics, Turku University Hospital, Turku, Finland; 9Department of Physiology, Institute of Biomedicine, University of Turku, Turku, Finland; 10Department of Pediatrics, Medical Research Center, PEDEGO Research Unit, Oulu University Hospital and University of Oulu, Oulu, Finland; 11Institute of Diabetes Research, Helmholtz Zentrum München and Forschergruppe Diabetes e.V. Neuherberg, Neuherberg, Germany; 12Division of Diabetes, Endocrinology, and Metabolism, National Institute of Diabetes, Digestive, and Kidney Diseases, National Institutes of Health, Bethesda, MD
Background
Recent data supports the concept of heterogeneity in the preclinical and clinical course of T1D. The Diabetes Prevention Trial-type 1 (DPT-1) study group had previously developed and validated a DPT-1 risk score using body mass index (BMI), age, log-fasting C-peptide, and 2-hour OGTT data to predict T1D risk in nondiabetic relatives aged 3–46 years. The TEDDY cohort represents a unique group of children with an increased genetic risk of progression to T1D followed since birth. The authors of this study have attempted to develop a similar risk scoring system in younger participants in the TEDDY study.
Methods
Infants with high-risk human leukocyte antigen (HLA) genotypes in the TEDDY study group are followed for the development of T1D autoantibodies. Participants in this study had ≥1 autoantibody on two consecutive tests by the age of 3 years 5 months but were not diagnosed with T1D (n=363). The development of T1D at 6 years of age was chosen as the primary outcome. 38 risk candidate predictors determined from clinical, immunologic, metabolic, and genetic data were examined by logistic regression modeling and a four-fold cross validation method for development of T1D. Important risk predictor variables included BMI z-scores at 3 years of age, autoantibody number (single versus multiple) and type, age at confirmation of autoantibody positivity, HbA1c and OGTT at two time points (fasting and 2 hours), fasting C-peptide, insulin, and homeostatic model assessment. SNPs associated with autoantibody positivity and T1D in previous TEDDY analyses were also included.
Results
Of the 363 TEDDY participants who were autoantibody positive at the age of 3 years, 76 developed T1D by the age of 6 years. The presence of IA-2A, HbA1c level (>5.2%), BMI z-score, SNP rs12708716_G, and low fasting insulin level (<2.0 mcU/mL) with a background of multiple autoantibodies were identified as significant predictors of clinical T1D. Positivity for IA-2A had the most significant effect at 3 years of age with an odds ratio of 8.7. The presence of SNP rs12708716_G (tagged to CLEC16A, a susceptibility locus for islet autoantibody development) demonstrated an odds ratio of 2.4.
Conclusion
Logistic regression modeling can predict progression to T1D from age 3 to 6 years in TEDDY children with high sensitivity but low specificity.
Previous birth and nonbirth cohort studies such as Diabetes Autoimmunity Study in the Young, BABYDIAB and BABYDIET, Type 1 Diabetes Prediction and Prevention, and DPT-1, respectively, confirmed a very high risk for clinical T1D in those with multiple autoantibodies. The current study has validated the previous findings in younger children, including a higher risk of progression to clinical disease with positivity for IA-2A and risk-associated genotypes at non-HLA SNPs. Given the number of variables that affect the disease course in T1D, it is difficult to comment if all of the risk predictors in this study reported to be associated with a faster progression to clinical T1D can be utilized in a real world setting for individual patient stratification. The novel findings associating IA-2A and the non-HLA SNP rs12708716_G with progression to T1D certainly warrants further evaluation.
Footnotes
Author Disclosure Statement
B.N., N.B., and D.S. have no conflicts of interest and no financial interests to disclose.