Presenter
Chaitra Rao, PhD (Indiana School of Medicine)

Authors
Chaitra Rao, Fei Huang, Matthew B. Johnson, Zhengjie Zhou, Jennifer Nelson, Charanya Muralidharan, Xiaoyan Yi, Soumyadeep Sarkar, Bobbie-Jo Webb-Robertson, Ernesto Nakayasu, Decio L. Eizirik, Carmella Evans-Molina, Sarah May, Sarah Tersey, Yun Fang, Richard Oram, Raghavendra G. Mirmira, Emily K. Sims

Purpose
The nature of intercellular crosstalk between β cells and immune cells is a pivotal early determinant in the progression or not of type 1 diabetes (T1D). For example, surviving β cells in T1D respond to inflammation by upregulating programmed death-ligand 1 (PD-L1) that engages immune cell programmed death-1 (PD-1) to limit β cell destruction by immune cells. Extracellular vesicles (EVs) and their cargo contribute to islet intercellular communication but β cell EV PD-L1 has never been described. We hypothesized that the inflammatory milieu of T1D increases PD-L1 in β cell EV cargo and that EV PD-L1 may protect β cells against immune-mediated cell death. In this study we sought to identify the molecular determinants of PD-L1 production and extracellular vesicle shuttling in human β cells.

Methods
EV PD-L1 was analyzed in EVs emanating from EndoC-βH1 human β cells and human islets by ultracentrifugation followed by immunoblotting, flow cytometry and by Exoview®. To mimic the microenvironment of early T1D in vitro, cells were treated with IFN-α (2000 U/ml) for 24 h. The capacity of EV PD-L1 to bind PD-1 was tested in decreasing concentrations of EndoC-βH1 EVs using a PD-1/PD-L1 competitive binding assay. mRNA encoding either wild-type PD-L1 or a homozygous PD-L1 variant resulting in an in-frame deletion associated with neonatal T1D (c.682+1G>A, p.Gly177_Pro227del) was transduced in EndoC-βH1 cells using proprietary lipid nanoparticles, with Exoquick® isolation of EVs. Plasma EV PD-L1 was assayed in children with recent-onset T1D vs. age, sex, and BMI-matched nondiabetic controls.

Summary of Results
PD-L1 protein colocalized with EV-associated proteins intracellularly and was detected in β-cell EVs. IFN-α treatment of EndoC-βH1 cells and human islets yielded a 2-fold upregulation of EV PD-L1, without changing total EV concentration. Flow cytometry indicated that PD-L1 is present on the surface of EVs, and EV PD-L1 dose-dependently bound to PD-1. Plasma EV PD-L1 levels were similar in children with recent-onset T1D compared to controls. However, plasma EV PD-L1 positively correlated with circulating C-peptide in children with T1D but not controls. Finally, we studied the cell biology of a homozygous PD-L1 variant (PD-L1v) causing neonatal T1D and identified that PD-L1v transduced EndoC-βH1 exhibited: (a) decreased overall expression compared to PD-L1; (b) an intracellular pattern of distribution with reduced cell membrane localization; (c) significantly impaired interaction with PD-L1; and (d) reduced deposition into EVs.

Conclusions
Our data indicate that PD-L1, a pro-tolerogenic protein, is expressed in β cell EVs in response to IFN exposure. PD-L1 EV cargo has the capacity to bind to PD-1, suggesting potential for a functional interaction with PD-1 on immune cells. EV PD-L1 levels correlate with residual beta cell function in children with recent-onset type 1 diabetes, whereas a human deletion variant (PD-L1v) exhibits reduced production, altered intracellular trafficking and reduced sorting to EVs, potentially promoting the early development of T1D in the individuals harboring this variant. Collectively, the present data provide new insight into the molecular mechanisms by which β cells generate a protective response against autoimmunity in T1D and suggest that EV PD-L1 could be exploited to inhibit immune-mediated beta cell death.

Presenter
Christiana Lekka (University of Exeter)

Authors
Christiana Lekka, Jessica Hopkinson, Jessica Hill, Fatoumata Samassa, Javier Perez-Hernandez, Christine Flaxman, Roberto Mallone, Noel Morgan, Sarah Richardson

Purpose
The upregulation of both classical and non-classical HLA class I (HLA-I) molecules in the insulin-containing islets of those diagnosed with Type 1 diabetes (T1D) has been previously reported. However, it is not clear if specific classical HLA-Is (HLA-A/B/C) are differentially regulated under these conditions. Moreover, the role of non-classical HLA-I molecules has not been extensively studied in islet cells in T1D. Here, we explore the expression of classical and non-classical HLA-I in the pancreata of people with and without T1D using multiplex immunofluorescence staining.

Methods
nPOD and Exeter Archival Diabetes Biobank (EADB) pancreas sections were stained with antibodies raised selectively against HLA-A, HLA-B, HLA-E, HLA-F, HLA-G or insulin using the OPAL platform and/or conventional multiplex IF. Whole slide scans were spectrally unmixed and imported into Indica HALO software for image analysis using the HighPlex module.

Summary of Results
Following careful validation of HLA-A and HLA-B antibodies, immunostaining of pancreas sections confirmed that both HLA-A and HLA-B are upregulated on the insulin-containing islets of individuals with T1D. Interestingly, the elevation of HLA-B expression was significantly higher than that of HLA-A, suggesting a preferential upregulation of HLA-B in subjects with T1D. In accord with previous findings, the expression of HLA-E, HLA-F and HLA-G was minimal in the islets of individuals without diabetes whereas it was upregulated in insulin-containing islets but diminished once more in insulin-deficient islets of those diagnosed with type 1 diabetes. HLA-E was predominantly localised to alpha cells, whereas HLA-F and -G were present both in beta cells and in other endocrine cells. HLA-A and HLA-F expression was localised predominantly to the plasma membrane but this differed by comparison with the distribution of HLA-B, -E and -G-E which appeared predominately cytoplasmic. The expression of HLA-I isoforms was distinctly lobular within insulin-containing islets, such that select islets expressed more HLA-A whereas others expressed more HLA-F.

Conclusions
This study provides important insights into the relative distribution of classical and non-classical HLA-I molecules in the pancreata of individuals with and without diabetes. The differential expression of classical and non-classical HLA-I molecules among the islets of individuals with T1D could indicate differences in the immediate islet milieu and we suggest that the balance of HLA-I isoform expression may influence the extent and pace of beta cell demise during autoimmune attack.

Presenter
Noel Morgan, PhD (University of Exeter)

Authors
Kaiyven Afi Leslie, Christiana Lekka, Sara Richardson, Mark Russell, Noel Morgan

Purpose
Recent evidence has implicated the activation of STAT1 as a critical component of the mechanism by which proinflammatory cytokines, including interferons (IFN)-alpha and gamma, drive beta-cell demise during the development of type 1 diabetes. Activated STAT1 can exert detrimental effects on the beta-cells by promoting the expression of pro-apoptotic genes and it can also increase the visibility of beta-cells to influent autoreactive CD8+ T-cells by virtue of upregulation of MHC class I on the cell surface. Targeting of the mechanisms that promote STAT1 activation may, therefore, be a fruitful means to attenuate these actions and to improve beta-cell viability in the context of type 1 diabetes. Accordingly, trials are underway with JAK Kinase inhibitors (such as Baricitinib) since JAK kinases are central to IFN signal transduction via STAT1. In the present work, we have considered an alternative approach to intervene in this pathway by studying the activity of a cytosolic lysine deacetylase, HDAC6, which controls the acetylation of STAT1 and thereby alters its capacity for activation.

Methods
Experiments were conducted using EndoC-βH1 cells, isolated human islets and sections of fixed human pancreas. Protein expression was monitored by immunocytochemistry, Western blotting, immunoprecipitation of specific target molecules and mass spectrometric analysis of peptide composition. STAT1 reporter activity was assessed by transfection of cells with a luciferase construct under the control of an IFN-responsive promoter. Target gene transcription was also monitored by RT-PCR.

Summary of Results

Immunoprecipitation of STAT1 from human islet cells followed by probing of the immunoprecipitate with an antiserum directed against acetylated lysine residues revealed that the transcription factor is acetylated under control conditions. The extent of STAT1 acetylation was reduced rapidly upon addition of IFNγ to cultured EndoC-βH1 cells (control – 14.46±2.15AU; IFNγ – 1.94±0.13AU; p<0.01 by densitometry) and this correlated with enhanced phosphorylation of STAT1 at Tyr701. Inclusion of an inhibitor (BRD9757) selective for the cytosolic lysine-deacetlyase, HDAC6, resulted in enhanced STAT1 acetylation in cells treated with IFNγ (IFNγ + BRD9757: 9.5±2.5AU; p>0.05) and this was associated with impaired phosphorylation and attenuated activation of STAT1 (luciferase reporter assay: IFNγ: 15.08±0.84 fold activation; IFNγ + BRD9757 – 7.2±0.4 fold; p<0.001). Targeted knockdown of HDAC6 using siRNA approaches also resulted in impaired STAT1 activation by IFNγ when measured either by reporter assay or via the induction of a downstream target gene, Mx1. To verify the importance of HDAC6 in human islets in situ, its expression was monitored in human pancreas samples recovered from control individuals or those with recent-onset type 1 diabetes. HDAC6 was present in the cytosol of islet cells of subjects without diabetes and was retained in the residual pancreatic beta-cells of people with recent-onset type 1 diabetes; conditions under which STAT1 levels are enhanced.

Conclusions
The present results imply that a previously unrecognised level of regulation occurs in the signalling pathways employed by IFNs to exert their influence in beta-cells. This is achieved by control of the extent of STAT1 acetylation (and, thereby, its state of activation) mediated, at least in part, by alterations in the functional activity of HDAC6. Repurposing of HDAC6 inhibitors might, therefore, represent an alternative (or additional) therapeutic approach to the use of JAK kinase inhibitors as a means to attenuate the rate of beta-cells loss during progression to type 1 diabetes.

Presenter
Esra Karakose, PhD (Icahn School of Medicine at Mount Sinai)

Authors
Esra Karakose, Xuedi Wang, Peng Wang, Saul Carcamo, Deniz Demircioglu, Luca Lambertini, Olivia Wood, Randy Kang, Geming Lu, Don Scott, Adolfo Garcia-Ocaña, Carmen Argmann, Robert Serba, Dan Hasson, Andrew Stewart

Purpose
Type 1 diabetes results from inadequate numbers of insulin-producing beta cells. Although the current attempts to replenish the remaining beta cell pool in people with diabetes (whole pancreas transplantation, pancreatic islet transplantation, transplant of human stem cell-derived beta cells) are encouraging, scalability and cost issues limit the access to these therapies for the millions of people with diabetes. We previously showed that the DYRK1A inhibitors, either alone or in combination with GLP1 receptor agonists (GLP1) or TGF beta superfamily inhibitors (LY), induce beta cell replication and increase beta cell mass both in vitro and in vivo. However, the precise mechanisms of action of these regenerative drugs remain elusive.

Methods
To more deeply explore the beneficial mechanisms of action of DYRK1A inhibitors, we performed single cell RNA sequencing on four different human pancreatic islets treated with a DYRK1A inhibitor, either alone, or with GLP1 or LY. This is the first in-depth single cell transcriptomic analysis on human cadaveric islets treated with regenerative drugs.

Summary of Results
This is the first in-depth single cell transcriptomic analysis on human cadaveric islets treated with regenerative drugs. First, we confirm the presence of all previously reported cell types in human islets. More importantly, we identify a cluster of Cycling Alpha Cells as the endocrine cell type most responsive to regenerative drug treatment. Interestingly, we observed that Cycling Alpha Cell numbers diminish drastically in T1D patients, underlining the importance of this understudied cell cluster. Our in-depth analyses showed that in addition to the significant increase in the abundance of Cycling Alpha Cells, a sub-cluster of Cycling Alpha Cells gained beta cell identity with regenerative drug treatment. Our velocity and pseudotime lineage trajectory analyses confirmed this finding and also suggested that Cycling Alpha Cells serve as the primary target cell type for the regenerative drugs, and may serve as precursor cells that transdifferentiate into functional human beta cells in response to the regenerative drug treatment.

Conclusions
Collectively, these findings suggest a novel mechanism of action through which DYRK1A inhibitors are able to expand human beta cell numbers. This mechanism involves Cycling Alpha Cells which serve as a beta cell regenerative reservoir, and readily transdifferentiate into functional human beta cells upon treatment with beta cell regenerative drugs of the DYRK1A inhibitor class. Importantly, these studies raise the question as to whether “beta cell drug targeting” strategies are necessary or appropriate for human beta cell regeneration, since alpha cells may be the principal target for “beta cell regenerative drug therapy”. These findings and potential implications will be further discussed.

Presenter
Jessie Barra, PhD (University of Florida)

Authors
Jessie Barra, Rob Robino, Roberto Castro-Gutierrez, Leonardo Ferreira, Holger Russ

Purpose
The loss of functional beta-cell mass is a hallmark of Type 1 diabetes (T1D). Exogenous insulin therapies reduce glucose fluctuations, yet effective glycemic control is only partially restored. Beta-cell replacement strategies such as islet transplantation represent a promising alternative approach, but eventual immune-mediated graft destruction remains a major clinical hurdle. Human stem cell-derived beta-like cells (sBC) can solve the problem of low availability of high-quality human islets. However, like primary islets, sBC are still susceptible to both allogeneic and recurrent autoimmune attack. Unlike primary islets, sBCs can easily be genetically modified to provide localized immune suppression. Chimeric antigen receptor (CAR)-regulatory T cell (CAR-Treg) therapies represent a unique way of suppressing immune responses toward a specific antigen. By harnessing the suppressive power of CAR-Tregs at the site of beta cell engraftment, we predict that sBCs will be protected from, both allo- and autoimmune rejection.

Methods
Taking advantage of a previously published CAR against epidermal growth factor receptor (EGFR), we engineered sBCs to constitutively overexpress a truncated form of this receptor (EGFRt). By transducing the EGFR CAR into purified polyclonal human CD4+CD25hiCD127low Tregs, we predict that we should induce localized suppression of effector T cell responses at the graft site through engagement with EGFRt overexpressing sBCs.

Summary of Results

We demonstrate uniform overexpression of EGFRt within human pluripotent stem cells (hPSC) while maintaining normal expression of pluripotent cell markers compared to wild type control cells (WT). Directed differentiation of WT and EGFRt hPSCs into sBCs displayed comparable efficiency in the generation of insulin positive cells (40-50%). We can successfully transduce Tregs with the EGFR CAR with greater than 80% efficiency. CAR-Tregs are stable and continue to express key Treg markers including FOXP3, HELIOS, and CTLA-4. EGFR CAR-Tregs are activated specifically when co-cultured with EGFRt-expressing hPSC or sBCs. Stimulated EGFR CAR-Tregs suppress effector T cell proliferation and dendritic cell responses during in vitro co-culture assays. Employing a novel transplantation model, we show that EGFR CAR-Tregs prevent T cell mediated immune destruction of EGFRt sBC grafts in vivo.

Conclusions
We have successfully integrated a truncated form of EGFR into a human stem cell line to serve as a unique target for CAR-Treg activity at the site of beta cell engraftment. We observe efficient transduction of EGFR CAR into human Tregs, and CAR-Tregs are robustly activated when co-cultured with EGFRt overexpressing cells. Based on these promising results, we predict that the use of CAR-Tregs combined with the genetic engineering of a unique ligand for the CAR on sBCs represents a novel way to suppress immune responses in a highly localized manner, ultimately providing long-term sBC graft survival without systemic immunosuppression.

Presenter
Melanie Shapiro, PhD (University of Florida)

Authors
Melanie Shapiro, Puneet Rawat, Michael Widrich, Keshav Motwani, Daniel Perry, Milena Pavlović, Leeana Peters, Amanda Posgai, Michael Haller, Desmond Schatz, Mark Atkinson, Clive Wasserfall, Geir Sandve, Victor Greiff, Todd Brusko

Purpose
Despite significant progress in defining genetic loci that contribute to type 1 diabetes (T1D) risk, via genome-wide association studies (GWAS), our understanding of the impact for such variants on immune function remains limited. Human leukocyte antigen (HLA) haplotypes are known to alter thymocyte development and peripheral immune tolerance, suggesting that T1D risk HLA may alter the T cell receptor (TCR) repertoire allowing for autoreactive TCR persistence in genetically-predisposed individuals.

Methods
We performed deep TCRβ sequencing on peripheral blood (PB) samples and microarray-based precision medicine genotyping of 793 unrelated living participants from the University of Florida Diabetes Institute (UFDI) biobank (no-diabetes [ND], n=502; islet autoantibody positive [AAb+], n=56; T1D, n=235). A machine learning algorithm, Deep Repertoire Classification (DeepRC), was used to identify a TCR motif significantly enriched in the PB of T1D subjects. Genetic ancestry was inferred using Admixture software for projection analysis on the 1000Genomes reference. A continuous T1D genetic risk score (GRS2) was calculated to determine associations with the frequency of an identified TCR motif. Quantitative trait loci (QTL) analysis was performed to detect associations between TCR motif frequency versus 240 T1D risk variants. Observations in PB were tested in TCRβ-sequenced sorted conventional CD4+ T cells (Tconv), regulatory CD4+ T cells (Treg), and CD8+ T cells from pancreatic lymph nodes (pLN) as well as spleen from 75 nPOD donors (ND, n=29; islet AAb+, n=4; T1D, n=37; T2D, n=3; other diabetes, n=2) in order to understand matters of cell type and tissue specificity.

Summary of Results

GRS2 was significantly associated with the TCR prediction motif frequency in PB of European ancestry ND individuals (r=0.329, p<0.0001) and, to a lower magnitude, in persons with T1D (r=0.133, p=0.048). HLA-DRB1*0301-DQA1*05:01-DQB1*02:01 (p=1.62e-11), HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02 (p=1.43e-9), HLA-DQA1*05:05-DQB1*03:01 (p=5.03e-6), and HLA-DQA1*03:0X-DQB1*03:01 (p=2.31e-5) haplotypes were significantly associated with a TCR motif frequency independent of disease status such that risk haplotype correlated with an increased motif score. A T1D-associated variant tagging the XL9 super enhancer (p=9.41e-11), known to regulate HLA-DRB1 and HLA-DQA1 expression, and an intergenic HLA-DRA1-DRB1 variant (p=4.82e-5) were likewise associated with a PB motif score. Of these findings, the protective allele of the XL9 variant (p=5.88e-3) and HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02 (p=4.46e-3) were significantly associated with decreased TCR motif frequency in pLN CD8+ T cells.

Conclusions
T1D HLA risk genetics may confer selection pressure for a TCRβ motif enriched in PB and CD8+ T cells in pLN of individuals with T1D. Efforts to understand the antigenic specificity of TCR containing an enriched motif are in progress. We propose that the TCR motif score may complement existing biomarkers to track disease progression in individuals at-risk for T1D.

Presenter
Tim Tree (King’s College London)

Authors
Tim Tree

Purpose
The argument for a role for T cells in type 1 diabetes is compelling. T cells dominate the insulitis; selected HLA gene polymorphisms confer the major genetic risk; and anti-T cell therapies have shown success in halting disease extension. HLA class II genes confer the highest disease risk of any gene polymorphism; this strongly implicates CD4 T cells as disease initiators/drivers, because of their requirement for HLA class II molecules for antigen presentation. However, whereas islet-specific T cells can be detected in the peripheral blood of most individuals regardless of disease status, detailed phenotypic studies have suggested subtle differences between the phenotype of these cells in those with and without T1D but detailed unbiased profiling of these cells remains challenging. Development of assays that can enumerate and characterise both islet-specific effector (Teff) and Treg cells would provide powerful analysis tools which could be employed within INNODIA to increase our understanding at each stage disease: from initial loss of tolerance, progression to clinical disease and the rate of beta-cell loss following diagnosis, as well as providing valuable insights into the efficacy and specificity of immunotherapies.

Methods
We have developed an activation-induced marker (AIM) assay for the detection, isolation and phenotypic characterisation of antigen-specific CD4 T cells incorporating unbiased single-cell transcriptional profiling, detailed cell surface phenotyping by CITE-Seq and TCR use via full-length paired TCRA/B sequencing using the 10x Genomics platform. In theory, this assay makes no assumption about the phenotype of cells isolated and can identify cells with a variety of effector and regulatory functions. Using this technology, we characterised islet-specific (cells responsive to proinsulin or a pool of proinsulin and IA-2 peptides) and compared this to vaccine-specific CD4 T cells from individuals with 24 individuals with new-onset T1D (<100d from diagnosis) and compared this to cells isolated from 12 age and sex-matched individuals with a family history of T1D but with no evidence of islet autoimmunity (islet AAB-ve).

Summary of Results

Single-cell transcriptional analysis revealed islet-specific T cells displayed a distinct transcriptional profile compared to vaccine-responsive T cells in both T1D and AAB-ve groups including an increase in populations of FOXP3 Tregs and cells with a cytolytic profile consistent with a regulatory phenotype. Unbiased clustering analysis of islet-specific T cells revealed cells with a variety of effector and regulatory functional profiles in both T1D and AAB- cohorts. Comparison of islet-specific responses between cohorts revealed individuals with T1D displayed a significant expansion of cells with a highly proinflammatory phenotype (in particular clusters characterised by expression of IIL-17 family cytokines) whereas individuals in the AAB- cohort displayed an increased number of cells characterised by markers of immune regulation including subpopulations of FoxP3 Tregs. Analysis of islet-specific T cells from individuals who developed T1D at different ages (range 2-18y) revealed distinct transcriptional profiles were associated with younger and older ages of diagnosis including a higher expression of transcripts associated with a highly proinflammatory phenotype (IL21, TNF, IFNG) in those diagnosed young and a more regulatory phenotype in those diagnosed later (FOXP3, IL32). Finally, we observed distinct clonotypic expansions in individuals with and without T1D and were often restricted by the phenotype of responding cell. Re-expression of these paired TCR A and B genes is ongoing and will determine if structural differences in TCR contribute to the phenotype of responding cells.

Conclusions
In conclusion, islet-specific T cells display a distinct phenotypic profile which differs significantly in those with and without T1D and are related to the age of disease onset. Ongoing analysis seeks to link different profiles of islet-specific immunity with disease development.

Presenter
Miguel Medina-Serpas (University of Florida)

Authors
Miguel Medina-Serpas, Gregory Golden, Maigan Brusko, Trevor Rogers, Richard Musca, Michael Betts, Klaus Kaestner, Ali Naji, Mark Atkinson, Todd Brusko

Purpose
Spatial transcriptomic technologies enable in situ whole transcriptome detection thereby preserving biologically relevant signals typically lost in disaggregated single cell approaches – rare cells, stromal cell populations, and cell:cell interactions. This is advantageous in studying the histopathology of type 1 diabetes (T1D), where both immune and pancreatic cells have been implicated in initiation and progression of insulitis, and where striking islet and lobular heterogeneity has been described within the pancreas of T1D organ donors. To investigate key features of disease pathogenesis, we acquired formalin-fixed paraffin-embedded (FFPE) sections from paired pancreas and pancreatic lymph node (pLN) of 16 total donors (Mean Age: 20.25 y.o.[ (SD: +/- 6.12)(Range: 10 – 31 y.o.)], 37.5% Female) and employed the Visium Spatial Gene Expression Array (10x Genomics). Resultant spatial transcriptome data was linked with annotated single cell RNAseq data from matched tissues and individuals to bolster cell type annotation.

Methods
We analyzed a cross-sectional cohort comprised of non-diabetic controls (n = 6), at-risk subjects presenting with a single (n = 3) or multiple (n = 2) autoantibodies (AABs), and T1D (n = 6) subjects with variable disease duration (mean+SD: 4.67 yr +3.20 yr). The Visium Spatial Gene Expression Array contains defined capture areas (6.5mm x 6.5mm or 11.5mm x 11.5mm capture areas), composed of a grid of 55µm diameter capture spots containing a probe-based amplification system . Spatial gene expression libraries were pooled at equimolar ratios and sequenced at minimum to 25,000 reads/ capture spot according to manufacturer protocol. Low quality capture spots were identified and removed based on total number of reads, total number of detected genes, and <10% mitochondrial gene content. Gene counts were normalized and scaled for interdonor differential expression analysis. Visium capture spots are. To overcome this technical limitation, in silico cell deconvolution was performed using a single-cell resolution reference, and tissue functional regions (e.g. islets, follicles) were annotated.

Summary of Results
In silico cell deconvolution of pancreas spatial data revealed an immune cell signature within at-risk and T1D donors compared to non-diabetic controls, with greater enrichment detected among donors that present with multiple AABs and/or closer to disease onset. Furthermore, we identified increased inflammatory chemokine expression (CXCL12, CCL2) (LogFC ≥ 0.25), with significant enrichment of CCL2 (p = 0.0245, Wilcoxon rank sum test) in the pancreatic islets of the aforementioned subject groups. Differential expression analysis of inflammation-associated genes revealed significant increases in TNFRSF1A, TNFAIP3, and IFNGR2 expression (p ≤ 0.05, Wilcoxon rank sum test) in the pancreatic islets, with parallel increases observed in downstream signaling gene networks. Finally, preliminary analysis of paired pLN data via cell deconvolution has revealed enrichment of cell-type specific gene expression programs at specific stages of T1D. Notably, we identified relative enrichment of Tcm/Treg GEX programs in non-diabetic controls, as well as effector CD8+ T-cell programs in at-risk and T1D subjects. Differential expression analysis of pLN revealed upregulation of various chemotactic (CXCL13, CCL19, CCR7) (LogFC ≥ 0.25) and immune modulatory genes (ZAP70, TNFRSF4), and significant upregulation of TXNIP (p = 0.036, Wilcoxon rank sum test) within the T-cell zone of the pLN.

Conclusions
Collectively, we demonstrate the potential of emerging technology amd bioinformatic tools to unvocer unique insights in the immune histopathological features of T1D. Application of in silico cell deconvolution to pancreas spatial data identified a distinct immune cell signature is enriched in at-risk and T1D donors presenting with multiple islet-specific autoantibodies relative to non-diabetic controls. We further identified a global and compartment-level inflammatory chemokine signature in donor pancreata which follows a similar pattern of enrichment to the identified bulk immune signature which may represent a non-specific mechanism contributing to pancreatic immune infiltration in T1D. Lastly, we identified cell-type specific GEX programs and disease stage patterns of enrichment within matched donor pLN which we intend to leverage in resolving specific transcriptional phenotypes from the bulk pancreas signature. Ongoing work includes validating ST signature genes and pathways in serially prepared tissue sections using high parameter single cell resolution probe and protein staining technologies which we expect will provide novel leads for new therapeutic targets for therapeutic interventions in T1D.

Presenter
Geetha Mylvaganam, PhD (Abata Therapeutics)

Authors
Timothy Nelson, Juliana Barrios, Jeremy C. Burns, Ellen Cahir-McFarland, Katie Callow, Yuan Feng, Michelle Fleury, CJ Ives, Matthias John, Josh Lengieza, Devan Moodley, Elissa Murphy, Niranjana Nagarajan, Lawrence Schweitzer, Andrea Van Elsas, Grace Voorhees, Yizhou Wang, Yanbo Zhang, Richard Zhou, Jiang Zhu, Geetha Mylvaganam

Purpose
Abata Therapeutics is dedicated to developing novel targeted Treg cell therapies for the treatment of tissue-specific autoimmune disorders. Type 1 diabetes (T1D) is an autoimmune disease characterized by the destruction of insulin producing beta cells. Regulatory T (Treg) cell therapy offers a promising therapeutic avenue to restore immune tolerance in T1D patients by limiting ongoing immune activation and beta cell destruction. Here, we developed a proprietary innovative TCR discovery pipeline that facilitated the identification of novel MHC class II restricted islet reactive TCRs against two critical autoantigens in T1D, proinsulin and GAD65. These novel islet reactive TCRs have the potential to redirect Tregs to the sites of tissue inflammation, provide targeted immune suppression, and promote tissue repair resulting in targeted immune tolerance in T1D.

Methods
Antigen-specific CD4+ T cells were enriched in vitro, in silico reconstructed, and screened in a human T cell line and in primary human CD4+ T cells enabling ranking of candidate TCRs based on high functional avidity, Treg functional characteristics and on-target reactivity against lead antigens. Primary Tregs engineered with candidate TCRs against our lead target antigen, proinsulin, were subjected to in vitro suppression assays to further confirm their functional rank.

Summary of Results
Our robust end-to-end TCR discovery pipeline allows for the isolation of antigen-specific T cells, in silico TCR analysis, functionalization and screening that identified novel proinsulin and GAD65 specific TCRs. Ranking of TCRs based on TCR functional avidity facilitates further functional testing in a primary human Treg assay, providing an additional filter for candidate TCRs based on avidity and function within Tregs, as well an assessment of on-target specificity and screening against potential alloreactivity. This process enables efficient ranking of high functioning TCRs leading to the nomination of a candidate TCR for use in IND-enabling studies. Furthermore, downstream antigen-specific Treg functional suppression assays confirm TCR avidity ranking observed in the Treg activation assays.

Conclusions
In summary, we isolated, functionalized, and validated a novel set of candidate TCRs specific for two immunodominant islet autoantigens. Our robust end-to-end TCR discovery platform enabled rapid identification and tiered selection of lead candidate TCRs that exhibited high function, on target specificity and low cross reactivity. A candidate TCR targeting proinsulin was selected for use in the development of a TCR-engineered Treg cell therapy, for the treatment of T1D.

Presenter
Victoria Kuznetsova (University of Miami)

Authors
Victoria Kuznetsova, Daria Ivanova, Boutheina Marnissi, Luciana Mateus, Joana Almaca, Silvio Bicciato, Paolo Serafini

Purpose
T cell exhaustion and the interaction between programmed cell death protein 1 (PD1) and PD1 ligand 1 (PDL1) play a crucial role in the progression of Type 1 Diabetes (T1D). Residual functional islets from patients with a long history of T1D express PDL1; patients with autoreactive antibodies undergoing PD1 inhibition as cancer treatment develop fulminant autoimmune diabetes; the length of the honeymoon phase correlates with circulating exhausted T cells. In mice, transgenic or viral expression of PDL1 on allogeneic islet grafts prevents their rejection. This aligns with the observations in oncology and viral infection settings that PD1/PDL1 interaction gradually modulates T cell differentiation, increases coinhibitory receptors, reduces their effector function, and plays a role in peripheral tolerance. While in these settings, the use of checkpoint inhibitors restored the anti-tumor and anti-viral immunity and revolutionized the treatment of these diseases, the opposite (induction of PDL1 to inhibit autoimmunity) has not been therapeutically explored. This study aims to develop and test a bifunctional RNA therapeutic to modulate PDL1 expression on β cells in vivo. We hypothesize that our smart drug will upregulate PDL1 specifically on β cells, modulate the immune landscape in the pancreatic environment, and induce exhaustion on autoreactive T cells, potentially halting or reversing autoimmunity in T1D.

Methods
We generated a bifunctional RNA therapeutic (hereafter called PS03) by conjugating mouse and human β cell-targeting RNA aptamers with small activating RNA (saRNA) for upregulating PDL1. We first identified and empirically tested saRNA specific for PDL1 using previously described algorithms and immortalized cell lines. Then, we conjugate the one with the largest effect size to optimized RNA aptamer 1-717 and m12-3773 by RISC cleavable loop. We evaluated the resulting construct by adding it to MIN6 cultures and assessing PDL1 expression by flow cytometry and qRT-PCR. To test PS03 specificity in vivo, we intravenously injected NOD female mice and evaluated PDL1 expression by quantitative fluorescence immunocytometry on the pancreas and other tissues. We then evaluate the effect of PS03-driven PDL1 upregulation on the immune system. First, we employed living slices from female NOD pancreas to evaluate its effect on β cells and islet infiltrating leukocytes. Then, we evaluate the phenotype of pancreatic T cells from NOD mice treated TIW for 4 weeks with PS03 or vehicle by multicolor flow cytometry. Finally, we assessed the effectiveness and safety of PS03 by treating 15-week-old NOD mice TIW for six weeks.

Summary of Results
We evaluated PS03 on MIN6 and observed a significant upregulation of the PDL1 starting at 24 hours and peaking three days after adding the culture media. Similarly, we observed upregulation of PDL1 on the living slice islets that were treated with PS03 for 48 hours. This PDL1 upregulation correlated with the inhibition of Ca2+ influx of islets infiltrating leukocytes, suggesting inhibition of TCR signaling. In vivo, we observed specific PDL1 upregulation on endogenous β cells of NOD mice, with a peak at 5 days after treatment and persistence observed at day 10 after administration. In sharp contrast, we did not observe PDL1 upregulation in acinar and alpha cells and all other tissues evaluated. We then evaluated the phenotype of autoreactive T cells after prolonged PS03 treatment on β cells and observed a three-fold increase in the exhaustion-associated phenotype population compared to untreated controls. Finally, a long-term treatment of female NOD mice with a still unoptimized PS03 formulation significantly delayed T1D onset and prevented it in 40% of mice.

Conclusions
In summary, PS03 is a new RNA therapeutic that, given systemically, provides effective and β cell-specific PDL1 upregulation and protects NOD mice from T1D, likely by blocking autoreactive T cells’ function and inducing their exhaustion.

Presenter
Guido Sebastiani, PhD (University of Siena)

Authors
Mattia Toniolli, Giuseppina Emanuela Grieco, Marco Bruttini, Stefano Auddino, Elena Aiello, Alessia Mori, Daniela Fignani, Giada Licata, Andrea Berteramo, Erika Pedace, Laura Nigi, Caterina Formichi, Chantal Mathieu, Noel Morgan, Guido Sebastiani, Francesco Dotta

Purpose
The age at which type 1 diabetes mellitus (T1DM) is diagnosed differs according to genetic predisposition, presence of islet autoantibodies, clinical presentation of the disease and the decline in beta-cell function. Younger individuals typically exhibit a more severe clinical presentation together with a more rapid disease progression. This heterogeneity has led to the hypothesis that multiple endotypes of T1DM exist. Histopathological evidence has defined one such endotype (T1DE1) as being characterized by an early age of onset (below 7 years of age), a substantial infiltration of CD20+ B cells in the pancreas, and a pronounced and rapid loss of beta-cell function. However, it remains uncertain how the recognition of this endotype can be leveraged for specific interventional therapies and how it can be readily identified in clinical practice. Circulating microRNAs have demonstrated consistent associations with T1DM onset and some have also been linked to disease progression, particularly the decline in beta-cell function. As such, microRNAs hold promise as potential biomarkers for tracking T1DM progression and/or staging and/or personalized intervention. Consequently, the objective of this study is to establish whether specific circulating microRNAs are associated with the T1DE1 endotype.

Methods
Small RNA sequencing was carried out on plasma samples collected during the initial visit (v1) from two independent cohorts of individuals recently diagnosed with T1DM (within 4 weeks of onset). These cohorts were recruited as part of the European INNODIA consortium and consisted of n=115(sex: 58F/57M; mean age: 12.4±7.7 years; age range: 2-38y) and n=147 subjects (sex: 55M/92F; mean age: 11.9 ± 7.9 years; age range: 1-42y). T1DM individuals underwent programmed follow-up visits at 3- (v2), 6- (v3), and 12-months (v4). During these visits, they were tested for autoantibodies, fasting and mixed-meal tolerance test (MMTT) glycaemia, C-peptide levels and HbA1c. T1DM individuals were stratified based on their age at onset (<7 or ≥7 years), and miRNA differential expression analysis was conducted using the DeSeq2 Wald test (P-value adjusted using FDR<0.05). Differences in clinical parameters were evaluated using monovariate regression analysis. Detected miRNAs were validated using miRCury LNA reverse transcription and droplet digital PCR (ddPCR). Additionally, publicly available small RNA sequencing datasets from control (n=28), celiac disease (n=46), and asthmatic cohort individuals (n=462) with an age range similar to INNODIA T1DM cohorts, were re-analyzed to verify the disease specificity of the detected microRNAs or their association with age.

Summary of Results
Stratification based on an age cutoff of 7 years revealed that in the first T1DM cohort, there were n=18 individuals aged <7 and n=97 individuals aged ≥7, while in the second cohort, there were n=30 individuals aged <7 and n=117 individuals aged ≥7. One particular circulating miRNA, miR-150-5p, was found to be selectively and significantly elevated in plasma of T1DM individuals aged <7 in comparison to those aged ≥7 in both cohorts (log2 fold change: 0.65; adjusted p-value (Padj)=0.0008 and log2 fold change: 1.18; Padj: 9.6×10^-8). Additionally, miR-150-5p displayed a slight negative correlation with the age at onset in both cohorts (first cohort: R = 0.41, p-value = 5.6×10^-6; second cohort: R=-0.36, p-value: 5.7×10^-6). ddPCR confirmed the upregulation of miR-150-5p in T1D subjects <7 with respect to those ≥7 years, in both cohorts (first cohort p-value = 0.00048; second cohort p-value = 0.0066, non-parametric Wilcoxon test). As expected, T1DM individuals aged <7 exhibited reduced fasting C-peptide levels, a lower fasting C-peptide/glucose ratio, and a diminished MMTT AUC C-peptide over the 12 months of follow-up. They also showed an increase in MMTT AUC glucose. Furthermore, at baseline, T1DM individuals aged <7 displayed higher IAA titers (first cohort p-value = 0.024; second cohort p-value = 0.019) and lower GADA (first cohort p-value = 0.015; second cohort p-value = 0.00033) and ZnT8A (first cohort p-value = 0.0054; second cohort p-value = 9.6×10^-5) titers in both cohorts compared to those aged ≥7 at the time of onset. Notably, circulating miR-150-5p did not exhibit any association with age (<7 , ≥7y) in non-diabetic control individuals or in individuals with celiac disease or asthma. Additionally, in these cohorts miR-150-5p did not correlate with age in a linear regression model, thus excluding any association with age.

Conclusions
Plasma circulating levels of miR-150-5p, a microRNA associated with early B-cell development and activation, are increased in T1DM subjects aged <7 years (endotype T1DE1). Interestingly, this phenomenon was not observed in control subjects, or in individuals with celiac disease or asthma. This microRNA could be potentially used to categorize individuals with the T1DE1 endotype and, in doing so, shed light on additional mechanisms underlying the heterogeneity of T1D.

Presenter
Carla Di Dedda, PhD (San Raffaele Diabetes Research Institute, IRCCS Ospedale San Raffaele)

Authors
Carla Di Dedda, Isaac Snowhite, Alexandra Amaya, Carlos Blaschke, Janine Sanchez, George Burke, Paolo Monti, Alberto Pugliese

Purpose
Beta cell replacement, through pancreas or islet transplantation, represents a therapeutic option for patients with severe T1D. However, while allo-rejection can be efficiently controlled by immunosuppressive regimens, the long-lived autoreactive memory T cells can reactivate after the encounter with the graft antigens and initiate the recurrence of T1D. The glucose transporter GLUT1 is upregulated in T cells upon activation as part of the metabolic switch from oxidative phosphorylation to glycolysis. This project aims to test pharmacological GLUT1 blockade as a potential strategy to control T cell activation.

Methods
Ex- vivo and in vitro studies were performed using PBMCs from healthy donors (HD), patients with T1D (T1D), and patients with T1D who received a simultaneous pancreas-kidney (SPK) transplant. These were sub-divided into three groups: normal glucose-tolerant recipients without autoantibodies on follow-up (SPK-neg) or who experienced autoantibody conversions (SPK-Conv), and recipients who developed recurrence of autoimmunity and T1D (T1DR). We characterized T cell subsets involved in T1DR and assessed whether GLUT1 expression is associated with autoimmune reactivation. Pools of class I and class II dextramers were used to characterize autoreactive T cells among SPK patients. Moreover, using the small molecule GLUT1 inhibitor WZB117, we studied phenotypic, differentiation, and metabolic changes of T cells activated under GLUT1 blockade in vitro.

Summary of Results
The results showed that all SPK groups had higher frequencies of CD4+ Tem cells compared to HD and T1D patients, while CD8+ Tem cells were increased among T1DR patients compared to HD, T1D, and, importantly, compared to SPK-Neg patients. There was also a trend toward increased Tscm cells, whose expression of GLUT1 was consistently higher compared to other T cell subsets. Although the frequency of autoreactive T cells did not differ in T1DR patients, their phenotype was more activated and less exhausted compared to SPK-Neg and SPK-Conv, and this was associated with an upregulation of GLUT1 expression. This upregulation was not observed within the total CD4+ and CD8+ T cell populations. In vitro studies showed that GLUT1 blockade reduced T cell proliferation in HD, T1D patients, and SPK-Neg recipients, causing greater effects on cells from T1D and SPK patients, compared to HD. Other key changes that consistently occurred in T1D and SPK patients after GLUT1 blockade were reductions in the size of naïve, Tcm, and Tscm subsets, accompanied by a downregulation of several activation markers. Only T1D patients showed decreased GLUT1 expression following WZB117 treatment, especially in Tscm and Tcm subsets.

Conclusions
GLUT1 was highly expressed in autoreactive T cells from patients experiencing autoimmunity recurrence post-transplant, but not in T cells with irrelevant reactivity. GLUT1 expression levels were also consistently higher in Tscm compared to other T cell subsets, a population that was reported to be a major driver of islet autoimmunity. This, together with the demonstrated ability of GLUT1 blockade to inhibit T cell activation and proliferation in vitro, provides a rationale for further investigation of GLUT1 blockade approach for selectively targeting autoreactive T cells, preventing or inhibiting their activation, minimizing off-target effects, and enhancing the specificity of therapeutic interventions.

Presenter
Giovanna Bossi, PhD (Immunocore Ltd)

Authors
Adam P. Curnock, Matthew W Becker, Mollie K Huber, David X Overton, Veronica Gonzalez, Fiona McCann, Susan Aungier, Nitha Mulakkal, Andrece Powell, Todd M Brusko, Edward A. Phelps, Tara M Mahon, Giovanna Bossi

Purpose
Localized immune-mediated destruction of beta cells is a hallmark of type 1 diabetes. To-date, the majority of therapeutic approaches has involved systemic immunosuppression which carries risks particularly in the young population. We sought to develop a tissue targeted localised immunomodulatory molecule, called ImmTAAI, to protect beta cells from T cell attack but preserving the systemic immune response.

Methods
To create a b-cell specific ImmTAAI, an affinity-enhanced T cell receptor (TCR) specific for the peptide-HLA-A*02 complex of pre-pro-insulin PPI15-24 was fused to an agonistic PD-1 antibody. The impact of the b-cell bound PD-1 ImmTAAI was evaluated using a variety of assays that quantify TCR signalling, antigen-specific T cell effector and functional responses in co-culture experiments. To test the targeting to b-cells in tissue, live human pancreas slices obtained through nPOD were incubated with CF647 labeled PD-1 agonist ImmTAAI together with AF568 labeled anti-ENTPD3 to label beta cells. Furthermore, slices from a nPOD donor with T1 diabetes at onset were treated with PD-1 ImmTAAI and the mobility of T cells was tracked by time-lapse microscopy.

Summary of Results
In co-culture experiments, the ImmTAAI was able to suppress TCR signaling at picomolar concentrations and demonstrate modulation of inflammatory cytokine secretion in addition to protection of b-cell killing by autoreactive CD8+ T cell clones. Importantly, this PD-1 agonist ImmTAAI did not inhibit T cells when free in solution. The molecule is active only in the presence of b-cells since no binding or activity was detected with antigen negative cell lines. Confocal microscopy of live pancreas tissue slices from non-diabetic and diabetic nPOD donors showed that PPI PD-1 ImmTAAI co-localized strongly with ENTPD3-labeled beta cells demonstrating specific binding to b-cells within the islets in a physiologically relevant setting. Finally, treatment of pancreatic slices from nPOD donor 6578, a T1 diabetic donor at onset, resulted in an increase of T cell mobility within the islets suggesting PD-1 ImmTAAI-mediated reduction of b-cell-T cell interaction.

Conclusions
We have generated a TCR bispecific inhibitory ImmTAAI molecule that binds specifically to b-cells to suppress autoreactive T cells and potentially to prevent pancreatic islets damage. Importantly, the molecule is inactive when free in solution and therefore has the potential to deliver localized immune suppression while avoiding systemic immunosuppression. These features make PD-1 agonist ImmTAAI molecules an attractive and novel approach to potentially treat T1D.

Presenter
Yuval Dor, PhD (The Hebrew University – Hadassah Medical School)

Authors
Liza Zamashanski, Shani Peleg, Maya Israeli, Roy Novoselsky, Roni Cohen-Fultheim, Chunhua Dai, Klaus Kaestner, Shalev Itzkovitz, Erez Levanon, Alvin Powers, Agnes Klochendler, Yuval Dor

Purpose
One little understood hallmark of type 1 diabetes is the heterogenous pattern of islet inflammation, whereby intact islets are observed adjacent to islets that are heavily inflamed. We have generated a mouse model that mimics early T1D-associated insulitis. We are interrogating this system to understand the molecular processes that may lead to heterogenous islet destruction.

Methods
We have established a mouse model for deficient RNA editing, using knockout of the RNA editing gene Adar in pancreatic beta-cells to model the interferon-mediated inflammatory response associated with early Type I diabetes (T1D). We used immunofluorescent staining of pancreatic sections to examine inflammation heterogeneity. To assess the interferon response at a single cell level, we performed single-cell RNA-sequencing and achieved spatial resolution using single molecule RNA in situ-hybridization.

Summary of Results
We found that deletion of the RNA editing enzyme Adar in beta cells causes an accumulation of double-stranded RNA (dsRNA), which triggers a massive interferon response followed by insulitis and beta-cell destruction, eventually leading to diabetes. Strikingly, despite efficient disruption of RNA editing in beta cells, islet inflammation is asynchronous. Some islets are infiltrated and destroyed within few days, while others remain intact for a long time. Using single cell RNA sequencing and RNA in situ hybridization, we found that only a small subset of beta cells mounts an interferon response. These cells are concentrated in individual islets, which are targeted for rapid infiltration and destruction. With time, more islets turn on an interferon response and are destroyed. Interestingly, preliminary experiments indicate that hyperglycemia induced by an insulin receptor antagonist accelerates the onset of insulitis and amplifies its severity. This suggests that beta cell metabolic activity is a determinant of the heterogenous islet interferon response and consequent insulitis.

Conclusions
The asynchronous and heterogenous insulitis seen in mice with Adar-deficient beta cells is reminiscent of observations in the pancreas of recently diagnosed T1D patients. We propose that the underlying mechanism involves a positive feedback loop, wherein potential metabolic cues within a few neighboring beta cells modulate their transcription landscape, promoting the accumulation of dsRNA structures. This results in the production of type 1 interferon, rapidly leading to massive, islet-wide interferon response followed by infiltration and islet destruction. While our findings indicate that beta-cell metabolic activity regulates the onset of interferon response and islet inflammation, it remains unclear why and when a given islet is drawn into this cycle, both in Adar-deficient mice and in human T1D.