GLP1R agonists such as Exendin-4 (E) together with DYRK1A inhibitors such as Harmine (H) significantly increase human β-cell replication in vitro and in vivo. Importantly, 3-month treatment with H+E combination markedly increases human beta cell mass (∼7-fold) in islets transplanted in immunosuppressed mice, beyond the increase induced by proliferation. Therefore, we addressed whether H+E also enhances β-cell survival and that contributes to the increased human β-cell mass expansion in vivo. Treatment with H+E significantly decreased β-cell apoptosis in dispersed primary human islet cells treated in vitro with either thapsigargin (ER stress) , cytokines (inflammation) and H2O2 (ROS) , while the drugs alone or vehicle (V) did not induce such effect. Importantly, H+E treatment for 7 days significantly reduced β-cell apoptosis in human islet grafts transplanted in immunosuppressed mice assessed by insulin and TUNEL staining. Human β-cell mass analysis of these grafts by iDISCO+ revealed a significant ∼50% increase in H+E-treated mice compared with mice treated with drugs alone or V. Human α-cell mass was unchanged. To address the signaling pathways modulated by H+E after 7-day treatment, we performed RNA-seq analysis of these human islet grafts. We identified 29 differentially expressed human genes (> 2-fold, p<0.05) in H+E-treated human islet grafts. GSEA analysis revealed cell adhesion, survival, vascularization and secretion as the main pathways induced by H+E. Among the upregulated genes, VGF (VGF Nerve Growth Factor Inducible) is known to regulate insulin secretion and β-cell survival. We found that H+E treatment of human islets in vitro increases 2-to-3-fold VGF mRNA and secretion. We are now determining VGF involvement in H+E-induced human β-cell survival. In conclusion, we have uncovered a novel prosurvival function of H+E in human β-cells with the potential implication of VGF in these effects. These studies can lead to the discovery of future β-cell protection and regeneration therapies for diabetes. Disclosure C.Rosselot: None. A.F.Stewart: None. A.Garcia-ocana: Consultant; Sun Pharmaceutical Industries Ltd. Y.Li: None. D.Guevara: None. K.A.Beliard: None. R.Kang: None. P.Wang: None. K.Thakkar: None. G.Lu: None. R.J.Devita: None. Funding DYRK Inhibitors for Human Beta Cell Expansion RDK105015

Type 1 diabetes (T1D) results from loss of both immune tolerance and functional β-cells. Administration of harmine (H) plus exendin-4 (E) markedly induces human β-cell expansion. Anti-CD3 antibody treatment reduces C-peptide loss in T1D patients. Here, we tested whether combination therapy with anti-CD3 antibody and H+E enhances T1D remission in non-obese diabetic (NOD) mice. First, we tested whether H+E protects human β-cells against inflammation and ER stress. We found that H+E, but not the drugs alone, significantly reduced both thapsigargin- and cytokine-induced human β-cell apoptosis. Single-cell RNAseq of human islets treated with cytokines and H+E showed reduced IL1β and IFNγ signaling in β-cells. Apoptosis genes such as CYLD and RIPK1 were downregulated and prosurvival genes such as HIF1A and VEGFA were upregulated in β-cells of H+E-treated islets. Next, treatment of early-onset diabetic NOD mice with H+E (daily ip) for 8 weeks improved glucose homeostasis but failed to induce long-lasting immune tolerance and T1D remission. Therefore, we next treated early-onset diabetic NOD mice with low-dose anti-CD3 antibody daily for 3 days (40µg/day) followed by 8 weeks treatment with H+E or vehicle (V) (Alzet pumps) . Treatment with anti-CD3 and H+E eliminated hyperglycemia in 100% of the mice vs. only 20% remission in anti-CD3 and V-treated mice. CD3+ T cell numbers were significantly and similarly reduced (∼50%) in mice treated with H+E or V. Preliminary results indicate that activated T cells in pancreatic lymph nodes (PLNs) and islet insulitis were significantly reduced, and T regulatory cells in spleen and PLNs were increased in H+E-treated mice. TUNEL+ β-cells were decreased and Ki67+ β-cells were increased in the pancreas of H+E-treated mice. Collectively, these results indicate that combination therapy with low-dose anti-CD3 antibody and H+E enhances T1D remission in diabetic NOD mice by increasing β-cell viability and favoring immune tolerance. Disclosure G.Lu: None. R.Kang: None. Y.Li: None. P.Wang: None. C.Rosselot: None. R.J.Devita: None. A.F.Stewart: None. A.Garcia-ocana: Consultant; Sun Pharmaceutical Industries Ltd. Funding NIH DK105015

Single-cell RNA sequencing (scRNA-seq) of human islets requires cell dissociation and does not provide information on the transcriptional status of islet cells. On the other hand, single-nucleus RNA sequencing (snRNA-seq) does not require cell dissociation and provides abundant information on intronic sequences that can be used to identify actively transcribed genes. Based on this, we seek to compare scRNA-seq and snRNA-seq approaches in human islets to determine: 1) whether similar cell clusters could be detected; 2) whether new gene markers could be discovered in human islet endocrine cells using intronic reads; 3) whether human beta cell subpopulations could be identified based on INS expression dynamics; and 4) whether snRNA-seq could be used in human islet grafts transplanted in immunosuppressed mice. We performed scRNA-seq and snRNA-seq on three pairs of human islets obtained from three healthy adult human donors using exon only or exon plus intron reads, respectively. Analysis of integrated data revealed similar human islet cell clusters. In the snRNA-seq data, however, top differentially expressed genes were identified as new markers of human endocrine cells such as ZNF385D (β-cells) , PTPRT (α-cells) , LRFN5 (δ-cells) and CACNA2D3 (PP cells) . These markers also accurately define endocrine cell populations in human islet grafts. Additionally, we distinguished several beta cell sub-clusters- INS rich cluster, HNRNPA2B1 (vesicle) rich cluster and active INS transcribing cluster. In conclusion, snRNA-seq analysis of human islet cells is a previously unrecognized tool for the identification of human islet cell types in samples where nuclear RNA processing is required. By comparing INS expression between scRNA-seq and snRNA-seq data sets, we can detect different beta cell sub-populations with distinct gene expression patterns representing different biological dynamic states. Disclosure R.Kang: None. Y.Li: None. C.Rosselot: None. D.Scott: None. A.Garcia-ocana: Consultant; Sun Pharmaceutical Industries Ltd. G.Lu: None. Funding NIH DK105015

Diabetes results from diminished functional β-cell mass. Small molecule inhibitors of Dual Specificity Tyrosine Phosphorylation Regulated Kinase 1A (DYRK1A) markedly increase human β-cell replication in vitro and in vivo. Furthermore, combination of DYRK1A inhibitors such as Harmine (H) with widely clinically used glucagon-like peptide 1 receptor agonists such as Exendin-4 (E) further enhance human β-cell proliferation. However, whether H+E combination increases human β-cell mass and improves glucose homeostasis in diabetic immunosuppressed mice transplanted with a marginal number of human islets is unknown. Human islets from 5-6 different adult healthy donors were transplanted under the kidney capsule (200 islets per kidney) of streptozotocin-induced diabetic immunodeficient Rag1-/- mice. At the time of transplantation, osmotic Alzet minipumps were implanted for continuous delivery of vehicle V, H, E and H+E for a period of 12 weeks. At the end of the study, human islet-bearing kidneys were harvested, labeled with insulin antisera and cleared using a modification of the iDISCO+ method. Imaging and analysis of the β-cell volume was done using the Light-sheet Ultramicroscope II and Imaris software. Transplanted diabetic mice who received the combination of H+E showed a significant increase of human β-cell volume up to 600% compared with mice treated with V or the drugs alone. This increase correlated with a significant increase in plasma human insulin levels, sustained normalization of blood glucose levels and significantly improved glucose tolerance. Our studies provide unequivocal documentation that treatment with the harmine and exendin-4 combination can lead to actual increases in human β-cell mass in diabetic conditions suggesting that β-cell replenishment might be possible in humans with diabetes. Disclosure K.A.Beliard: None. S.Stanley: None. A.Garcia-ocana: Consultant; Sun Pharmaceutical Industries Ltd. C.Rosselot: None. Y.Li: None. A.Alvarsson: None. P.Wang: None. K.Thakkar: None. D.Guevara: None. R.J.Devita: None. A.F.Stewart: None. Funding NIH/NIDDK 1R01DK105015-05.