Abstract/Flash Poster Session C

Background: Autoimmune diabetes (T1D) is a chronic disease that causes severe loss of insulin-producing β cells in the islet. CD4 T cells have a central role in the destruction of β cells, yet the limited knowledge on the phenotype of circulating islet-reactive CD4 T cells restricts our understanding of T1D and impedes the development of more effective diagnostic approaches.
Methods: A modified tetramer-based assay that improves the sensitivity of direct ex vivo detection was developed to study the rare autoreactive CD4 T cells in the periphery. An integrated immunological and bioinformatics approach was used to characterize islet-reactive CD4 T cells specific for HLA-DQ8-restricted immuno-dominant epitopes. 
Results: A highly heterogeneous population was observed within islet-reactive CD4 T cells. Despite the heterogeneity, increased percentages and frequencies of islet-reactive CD4 T cells expressing the Th17 lineage marker CCR6 were found in at-risk (AAb+) and T1D subjects, compared to healthy controls. Correlated with the skewed CCR6 expression, only islet-reactive CD4 T cells from subjects with T1D but not from healthy controls produced IL-17 and IL-21. Notably, cells from one T1D patient predominantly produced IL-4. Single-cell transcriptome analysis of longitudinal at-risk samples revealed extensive TCR clonotype sharing in islet-reactive CD4 T cells, with no public TCR specificity among subjects. Unique genetic signatures, namely the Th17 and Th2 pathways, were also identified from single-cell transcriptomics. 
Conclusion: Our observations suggest that distinct phenotypes are present during progression of T1D, and immune intervention targeting the Th17 pathway could be a promising treatment for some individuals at-risk of or with T1D. In addition, the HLA-DQ8-restricted islet-reactive CD4 T cells investigated in this study could be potential prognostic and predictive biomarkers for the disease.

Background: Type 1 diabetes (T1D) is almost invariably preceded by positivity for islet autoantibodies. Recent studies have shown that circulating T follicular helper (Tfh) cells are increased in patients with T1D. Together, these findings support an important role for T-B-cell interactions in T1D pathogenesis. A new subset of T cells, peripheral helper T cell (Tph) capable of promoting B cell responses and antibody production in inflamed tissues has recently been observed to be associated with rheumatoid arthritis. It is unclear, however, whether Tph cells play a role in the development of T1D. 
Methods: We used multicolor flow cytometry and cell sorting to characterize peripheral blood CXCR5-PD1hi memory CD4+ T cells. The frequency of CXCR5-PD1hi T cells was analyzed in a clinical cohort of 48 children with newly diagnosed T1D, 40 autoantibody-positive at-risk children and 86 age-matched autoantibody-negative control children. 
Results: We confirmed that CXCR5-PD1hi Tph cells share several features with CXCR5+PD1hi Tfh cells, such as the high expression of ICOS and IL-21 but also displayed higher expression of chemokine receptors associated with trafficking to inflamed tissues, such as CCR2, CX3CR1 and CCR5, than Tfh cells. Tph cells were also able to activate memory B cells into antibody-secreting plasma cells as efficiently as Tfh cells. We observed that the frequency of peripheral blood Tph cells was increased both in children with newly-diagnosed T1D and in autoantibody-positive children at-risk for the disease when compared to healthy children. Additional flow-cytometric analyses revealed that the expanded population of CXCR5-PD1hi Tph cells in individuals with T1D-associated autoimmunity also expresses the marker TIGIT.
Conclusion: Our results demonstrate that a novel population of circulating CXCR5-PD1hi Tph cells with B-cell activating potential is expanded in individuals with T1D-associated autoimmunity. Consequently, Tph cells could have potential both as a biomarker of T1D progression and as a target for immunotherapy.

Insufficient regulatory T cell (Treg) control of autoreactive T cell-mediated destruction of beta-cells contributes to type 1 diabetes (T1D). Multiple therapeutic strategies to restore a normal immunoregulatory balance and stop T1D progression are in development and testing. However, a major challenge has been defining biomarkers that can prospectively identify subjects most likely to benefit from immunotherapy and/or measure intervention effects on immunity. We previously found that Tregs from children with new-onset T1D have an altered gene signature compared to healthy controls, suggesting that this could be a useful biomarker to monitor immunoregulation. Here we investigated whether the Treg gene signature is also altered in T1D adults, finding significant differences in new-onset and cross-sectional cohorts compared to controls (AUC=0.830, AUC=0.953). Furthermore, we also found that significant differences in the Treg gene signature are detected in unfractionated peripheral blood mononuclear cells, with statistical power increased by including genotype for T1D-associated SNPs (e.g. signature + Cd25 rs2104286 AUC=0.996). Treg gene signatures in T1D were distinct from those with T2D, indicating disease-specific Treg alterations. We further tested the predictive potential of the Treg gene signature by longitudinal testing in a phase I/II clinical trial of immunotherapy with ustekinumab (aIL-12/23p40). When looking at the change in Treg gene signature between weeks 0 and 40, we found an algorithm that could accurately discriminate between patients who had slow versus rapid c-peptide decline (AUC=0.818). Moreover, when examining data at week 40 alone there was a near perfect separation of slow versus rapid c-peptide decline subjects, possibly indicating a restoration of immunoregulatory balance in subjects who responded to ustekinumab. Overall these data confirm that Treg gene signature measurement is a simple and useful biomarker to measure immunoregulatory status and could possibly be useful for predicting disease trajectory, as well as stratifying and monitoring immunotherapy patients.

Type 1 diabetes (T1D) is primarily characterized by immune-mediated destruction of pancreatic beta-(β)-cells resulting in loss of insulin production. Its incidence is significantly increasing in developed countries and there is currently no prevention or cure. Existing rodent and in vitro models lack many functions critical for understanding the onset and progression of this disease in humans. Although isolated primary islets are considered the in vitro gold standard of diabetes research, they present multiple limitations to their experimental use due to their inherent heterogeneity in size, cellular composition, function, and purity, as well as a rapid functional decline (glucose-responsiveness) and viability. The current study investigated the use of a novel co-culture platform comprising peripheral blood mononuclear cells (PBMC) and uniform 3D human islet microtissue. The microtissues, produced using an optimized dissociation protocol and controlled scaffold-free reaggregation of primary islet cells, were cultured one-islet per well in 96-well plates, and unstimulated and anti-CD3/CD28-stimulated HLA-matched PBMC were cultured together, with or without inflammatory cytokines. The microtissues displayed uniform, long-term (>28 days), and robust viability and function enabling high-throughput and longitudinal study of immune cell-endocrine cell interactions and β-cell function. Cultures of naïve PBMCs and islet microtissues had no impact on β-cell function; however, combination of activated PBMCs and islet microtissues resulted in a PBMC number-dependent decline in islet health demonstrated by decreased stimulated insulin secretion, insulin content, and PDX-1+ nuclei count within each microtissue. The observed destruction correlated with the amount of CD3+ cells infiltrating the islet microtissues and was significantly increased by cytokine preconditioning of the islets. The established platform serves as a valuable biomimetic islet model that may allow for better understanding of T1D progression and immune-islet interaction, as well as for screening compounds that could prevent immune-cell mediated destruction of islets.

Background: In the pre-clinical phase of disease, monitoring anti-islet antibodies is the only means to evaluate b cell destruction. As T1D is a T cell mediated disease, and can develop in agammaglobulinemia patients, detecting and profiling activated anti-islet CD4 T cells from peripheral blood would appear more appropriate and sensitive.
Methods: Single cell RNAseq of antigen-specific CD4 T cells in the islets of 4-8 week old NOD mice. FACS validation of identified upregulated cell surface markers on islet specific CD4 T cells. Isolation of peripheral blood T cells with similar FACS profile and sequencing. Validation of antigen specificity using pMHC tetramers in mice and humans.
Results: An exhaustive single cell analysis of infiltrating CD4 T cells in islets of pre-diabetic NOD mice has distinguish three main populations of cells based on surface phenotypes and signaling signatures. Using parabiosis, we demonstrate pathogenic CD4 T cells re-circulate and could be captured in peripheral blood using either surface markers and/or antigen-specific reagents. We have moved these results into translational studies to assess our ability to diagnose the islet-specific autoimmune reaction from peripheral blood. First, a cohort of 20 control and 20 established type 1 diabetic patients were examined for the possible presence of islet-activated CD4 T cells in blood. Unexpectedly, a population of activated cells could be isolated from T1D patients but not from control, most likely indicating that even in established disease, the autoimmune process is ongoing. Less surprisingly, in just-diagnosed patients, the same diagnostic population is increased to high percentages. The prognosis and treatment management value of following these CD4 cell population will be evaluated in a large follow-up cohort.
 

Background: Regulatory T cells (Tregs) play an important role in maintaining tolerance to self-antigens. Defects in frequency and function of polyclonal Tregs have been reported in type 1 diabetes (T1D). However, the characteristics of islet-specific Tregs have not yet been determined in juvenile-onset T1D (JOT1D) or adult-onset T1D (AOT1D) subjects, distinguished by their age at onset of T1D.
Methods: In the current study, we analyzed the frequency and phenotype of proinsulin (PI)-specific Tregs and PI-specific effector T (Teff) cells in peripheral blood of JOT1D and AOT1D subjects. For this, peripheral blood mononuclear cells were stimulated with PI derived peptides for 72 hours and PI-specific Tregs and PI-specific Teff cells were identified and characterized using the combination of HLA class II TMRs, FOXP3, and CD45RA by flow cytometry.
Results: At first, we observed no significant difference in the frequency of polyclonal Tregs between the T1D and HC subjects. However, a significant positive correlation between age and frequency of polyclonal Tregs was observed among all the subjects (r=+0.35, p=0.04). Upon segregation of T1D subjects in the two groups, we observed that JOT1D and AOT1D subjects harbor comparable frequency of PI-specific Teff cells as well as PI-specific Tregs in their peripheral blood, although the frequency of polyclonal Tregs was significantly higher in AOT1D subjects (p=0.02). Further, a higher proportion of PI-specific Tregs (p=0.002), as well as PI-specific Teff cells (p=0.04), exhibited memory phenotype in both the JOT1D and AOT1D subjects. Lastly, we observed no significant difference in the expression of FOXP3 and IL-2 receptor on PI-specific Tregs in the JOT1D and AOT1D groups.
Conclusions: Chronic antigen presentation results in a distinct memory-like phenotype of PI-specific Tregs in all T1D subjects irrespective of the age of disease onset.

Background
The course of beta cell function decline is very heterogenous after T1D onset. We aimed to describe the natural course of residual insulin secretion over the first year after T1D diagnosis and to determine the predictors of the partial remission within metabolic and immune parameters.
Methods
A total of 38 children with new onset T1D (20 girls), aged 9.4 years (SD±4) were prospectively followed-up for 1 year. HbA1c and total daily dose of insulin (TDD) were obtained at baseline, 6 and 12 months. The insulin-adjusted A1c (IDAA1c) and partial remission (IDAA1c≤9) were calculated. A mixed-meal tolerance test was performed at 12 months and area under the curve (AUC) of C-peptide was estimated. All subjects were screened for immune markers (Treg,Th17,Th1,MFI BAFFR CD4,MFI BAFFR CD8,CD19,mDC,pDC) at baseline, 6 and 12 months.
Results
A total of 66% of patients achieved partial remission at 6 months as compared to only 45% at month 12. We observed a linear increase of C-peptide AUC with the age of patients (p<0.0001). The TDD of insulin (p<0.009) and IDAA1c (p=0.02) increased significantly with the decrease of C-peptide AUC regardless of age. HbA1c at 12 months was not associated with C-peptide AUC. Dynamics of immune parameters was reflected in an increase of Tregs (p=0.02) and a significant increase of originally low pDCs (p<0.001). Interestingly, the expression of a receptor for BAFF was markedly increased on B lymphocytes (MFI BAFFR CD19; p<0.001) and to a lesser extent also on T lymphocytes over 12 months period regardless of their C-peptide AUC. Curiously, the number of CD4+ BAFFR (p<0.015) and also CD8+ BAFFR expressing cells (p<0.014) were the only ones positively correlated to C-peptide AUC. The correlation between HbA1c and immune parameters varied at all visits.
Conclusion
These data imply that low HbA1c is not associated with partial remission at 12 months. The exact role of changes immune markers for beta cell function decline remains hypothetical.

Background: Inflammation and oxidative stress in the pancreas amplifies various post-translational modifications (PTMs) on self-proteins. The loss of immune tolerance to PTMs within the stressed islets subsequently impacts the autoreactive T cell epitope repertoire and contributes to the destruction of insulin-producing beta cells in type 1 diabetes (T1D). In this study, we examined the effects of the novelPTM autoantigens on immunogenicity and beta cell functions. Moreover, we also examined autoantibodies arising in diabetes elicited by checkpoint inhibitor therapy (CPI) in cancer patients.

Methods: Mass spectrometry was performed to identify proteins and map specific PTM sites. Autoantibody and T cell responses to modified islet proteins were evaluated by ELISA and HLA tetramers, respectively. Moreover, proinsulin and insulin secretion upon glucose stimulation was examined in human islets in the context of beta cell PTMs. Finally, the autoantibodies profile was screen from CPIs-treated patients serum.

Results:We identified a number of PTM beta cell proteins. In particular, we found carbonyl residues within P4Hb (an insulin folding protein) and nineteen citrullination modifications within glucokinase. Carbonylated-P4Hb is amplified in stressed islets coincident with decreased glucose-stimulated insulin secretion and altered proinsulin to insulin ratios. Autoantibodies against both P4Hb and glucokinase arise in human T1D patients. Likewise,CD4+ T cells specific for citrullinated glucokinase are present in the circulation of T1D patients. Interestingly,anti-glucokinase antibodies correlated with anti-ZnT8 immunity. InCPI-treated cancer patients, we observed autoantibodies to both established T1D biomarkers (insulin, GAD, IA2, ICA and ZnT8) and also to novel T1D autoantibodies (P4Hb and glucokinase) as well as antinuclear antibodies (ANAs) typical of lupus autoimmunity. Taken together, our studies implicate these crucial enzymes as biomarkers, providing new insights into creating autoantigens and define the impact of PTMs on the biological function of beta cells in T1D and CPI-associated diabetes. 

The role of type 17 immune responses in the pathogenesis of type 1 diabetes (T1D) has been tested in multiple studies however the importance of these pathways remains enigmatic. The function of the prototypic IL-17 family cytokine, IL-17A, has been tested most extensively and while blockade of IL-17A in NOD mice protects from the development of T1D, IL-17A deficient NOD mice have normal disease incidence. To further elucidate the role of IL-17 pathways in T1D we have tested the functions of the cytokine IL-17F and IL-17RC, a receptor for IL-17A and IL-17F, in islets and NOD mice.
We show that NOD islet infiltrates contain a population of IL-17F/IL-17A coexpressing T cells and that mouse beta cells express the IL-17RC/IL-17RA receptor complex required for IL-17F responsiveness. Stimulation of mouse islets with IL-17F, in combination with TNFa and IFNg, potently increased the expression of inflammatory markers, including chemokines and Nos2, and induced apoptotic cell death. IL-17F also suppressed the expression of beta cell signature transcripts including Pdx1, Insulin1 and Glut2, consistent with a loss of beta cell function. To determine how the combined effects of IL-17F and IL-17A regulate the development of T1D we generated IL-17RC deficient NOD mice. Islets from IL-17RC deficient mice failed to respond to both IL-17F and IL-17A. Importantly IL-17RC deficient NOD mice exhibited perturbations in the islet autoimmune response with reduced insulin autoantibody production and were protected from development of T1D. Together these data demonstrate that IL-17F has pathogenic activities in mouse islets and that the combined actions of IL-17F and IL-17A regulate the development of T1D in NOD mice.