The leptin receptor has no role in delta-cell control of beta-cell function in the mouse

Introduction Leptin inhibits insulin secretion from isolated islets from multiple species, but the cell type that mediates this process remains elusive. Several mouse models have been used to explore this question. Ablation of the leptin receptor (Lepr) throughout the pancreatic epithelium results in altered glucose homeostasis and ex vivo insulin secretion and Ca2+ dynamics. However, Lepr removal from neither alpha nor beta cells mimics this result. Moreover, scRNAseq data has revealed an enrichment of LEPR in human islet delta cells. Methods We confirmed LEPR upregulation in human delta cells by performing RNAseq on fixed, sorted beta and delta cells. We then used a mouse model to test whether delta cells mediate the diminished glucose-stimulated insulin secretion in response to leptin. Results Ablation of Lepr within mouse delta cells did not change glucose homeostasis or insulin secretion, whether mice were fed a chow or high-fat diet. We further show, using a publicly available scRNAseq dataset, that islet cells expressing Lepr lie within endothelial cell clusters. Conclusions In mice, leptin does not influence beta-cell function through delta cells.


Introduction
Leptin, a hormone secreted by adipose tissue, impairs hunger and promotes energy expenditure by activating its cognate receptor (1). It exerts these effects primarily through the hypothalamus (2,3). The human and mouse leptin receptor exists in six isoforms resulting from alternative LEPR gene splicing (4). These receptors are referred to as LepRa-f. The full-length receptor, LepRb, has a strong signal in response to leptin binding. However, the short isoforms, LepRa-c and LepRf, lack varying lengths of the intracellular domain and demonstrate a weak signaling response to leptin binding (4). The short isoforms also aid in transporting leptin across the blood-brain barrier and may promote leptin internalization and degradation in some cell types (5). Lepr db/db mice become obese and, on some background strains, exhibit diabetes. Lepr db/db mice have a mutation within the Lepr gene that results in exon usage only for short LepR isoforms and not LepRb (6). This mutation phenocopies that of mice with a mutation in leptin, indicating the relative importance of the signaling capacity of the short versus long isoforms (6,7). LepRe, known as soluble leptin receptor, is a secreted isoform lacking the transmembrane and intracellular domains. It can bind circulating leptin, thereby preventing its transport and decreasing bioavailability (8,9).
Leptin receptor signaling acts through a variety of pathways. Leptin binding of the short or long LepR isoforms can trigger the Jak signaling cascade (4). In contrast, Signal Transducer and Activation of Transcript (STAT) protein signaling can only be activated by the full-length LepRb.
In addition to its hypothalamic localization, the leptin receptor (LEPR) is also expressed in immune cells, pericytes, and endothelial cells, affecting various immune processes and vessel constriction (4,10). LepRb has also been detected by RT-PCR and Northern blot in islets, and multiple labs have reported that leptin directly inhibits insulin secretion from isolated mouse, human, and rat islets (11)(12)(13)(14)(15)(16)(17). Moreover, deletion of the leptin receptor (Lepr) throughout the pancreatic epithelium using Pdx1-Cre alters glucose homeostasis (18). Pdx1-Cre;Lepr fl/fl mice fed a normal chow diet display improved glucose tolerance, while those provided a high-fat diet display glucose intolerance (18). Since leptin inhibits insulin secretion, deleting Lepr would be expected to promote insulin secretion, as observed on a normal chow diet. Perhaps this increased insulin secretion in times of high metabolic demand, such as when on a high-fat diet, leads to beta-cell exhaustion and thus the glucose tolerance observed in high-fat-diet-fed Pdx1-Cre; Lepr fl/fl mice.
While the altered glucose homeostasis in Pdx1-Cre;Lepr fl/f suggests that the cell responsible for this phenotype has an epithelial origin, narrowing down the exact cell type has proven difficult. Two different models of Lepr deletion within beta cells have yielded conflicting results. Rat insulin promoter (RIP 25Mgn )-Cre;Lepr mice display hyperglycemia and reduced insulin secretion, but they also have high adiposity due to RIP-Cre Mgn activity in the brain (19,20); the disrupted glucose homeostasis is likely a secondary effect to obesity associated with loss of Lepr in the hypothalamus. A second model of beta-cell Lepr ablation using Insulin 1 (Ins1)-Cre results in only a slight change in blood glucose regulation, with females at eight weeks of age exhibiting slightly improved glucose tolerance and enhanced glucose-stimulated insulin secretion (21). These changes disappeared by sixteen weeks of age. In addition, alpha-cell ablation of Lepr using either Glucagon (Gcg)-Cre or Gcg-Cre ER did not lead to altered glycemia (21,22).
Another primary islet endocrine cell type is the somatostatinsecreting delta cell. Somatostatin is a paracrine hormone that inhibits adjacent secreting cells. Single-cell RNA sequencing (scRNAseq) of human islets has revealed that LEPR is enriched within delta cells (23). Leptin signaling is generally inhibitory for secretory cells since the Jak cascade through phosphoinositide 3 kinase and phosphodiesterase 3B ultimately converts cAMP into AMP (24). However, LepR signaling through Jak or STAT can activate PKC (25,26). In multiple cell types, including beta cells and heart cells, some PKC isoforms enhance Ca 2+ influx and insulin secretion (27,28). Delta cells are glucose-responsive (29). In addition to expressing all the machinery for insulin secretion, except insulin itself, they also integrate several paracrine signals, such as GLP-1 and Urocortin 3, which elevate cAMP levels, and electrical signals from beta-cell through gap junctions (16, 30-33) ( Figure 1A). Although cAMP likely has a more prominent role in regulating delta-cell than beta-cell function, delta-cell Ca 2+ concentration is correlated with somatostatin secretion (29,35). Considering these data, we proposed that leptin acts on delta cells by stimulating PKC, elevating Ca 2+ and somatostatin release, thereby inhibiting beta cells ( Figure 1A). We tested this premise in a mouse model of delta-cell-specific Lepr ablation.

Materials and methods
Mice Sst-rtTA (36), Tet-O-Cre (37), and Lepr fl/fl (38) mice and genotyping have been described previously. 2% doxycycline in water was administered at eight weeks of age for two weeks. Mice were fed chow or 60% HFD (Research Diets) ad libitum and maintained on a 12-hour light/dark cycle. Control mice were littermates of mutants, had one allele of Sst rtTA , and included Sst rtTA ;Tet-O-Cre;Lepr +/fl and Sst rtTA ;Lepr fl/fl mice. All procedures followed the Institutional Animal Care and Use Committee guidelines at the University of Pennsylvania or Johns Hopkins University.

Glucose tolerance tests, mouse islet isolations, and perifusions
For glucose tolerance tests (GTTs), mice were fasted overnight for 16 hours, and blood glucose was measured with an AlphaTRAK glucometer. Mice were injected intraperitoneally with 1g/kg or 2g/kg glucose in sterile 1X PBS; blood glucose was measured 15, 30, 60, 90, and 120 minutes after injection. Islets were isolated by collagenase digestion followed by Ficoll separation and hand-picking, as previously described (39). Perifusions were performed by the University of Pennsylvania Pancreatic Islet Cell Biology Core.

Immunofluorescence and immunohistochemistry
Pancreata were fixed in 4% PFA for 4 hours, rinsed in 1X PBS, and embedded in paraffin. 4-micron sections were attached to PermaFrostPlus slides. For beta-cell mass assessments, five sections spread throughout the pancreas were immunolabeled with guinea pig-anti-insulin (DAKO) and a secondary biotinylated anti-guinea pig antibody. HRP was bound to the antibody with an ABC kit (Vector Laboratories) and visualized with DAB (Vector Laboratories). Slides were counterstained with eosin. Imaging was performed with an Aperio microscope at 400X. DAB-and eosinpositive areas were quantified using QuPath (40). The DAB-toeosin ratio was multiplied by pancreas weight to obtain beta-cell mass. For immunofluorescence, the primary antibodies guinea piganti-insulin (1:500, DAKO), mouse anti-SST (1:250, Santa Cruz), and rabbit anti-glucagon (1:500, Santa Cruz) were used. Secondary antibodies (Jackson Immunolaboratories) were applied at 1:600. Nuclei were stained with DAPI. Images were acquired using a BZ-X Series All-in-One Fluorescence Microscope.

Recombination analysis
One to four paraffin-embedded pancreatic sections were scraped with a clean razor into a sterile microcentrifuge tube and digested overnight with 100 microliters of Proteinase K solution before standard PCR, using 1-10 microliters of heat-inactivated digested sections. Primers for recombination analysis were previously described (41). Testing whether leptin mediates inhibition of insulin secretion through the delta cell. (A) RNAseq of sorted mouse alpha, beta, and delta cells has revealed the expression of all the components involved in canonical and amplifying pathways of insulin secretion, except insulin itself, in the delta cell. These components include a glucose transporter, voltage-gated potassium channels, glucokinase, and adenylate cyclase. Delta cells are also regulated by paracrine interactions by other endocrine cells, including by Urocortin 3 and by GLP1 (30-34). Although cytosolic Ca 2+ may be secondary to cAMP in regulating delta-cell function, Ca 2+ concentration correlates with somatostatin secretion. In human islets, the leptin receptor is enriched within the delta cell population. We, therefore, hypothesized that leptin receptor signaling enhances SST release by activating a PKC isoform that enhances Ca Dv activity and increases intracellular Ca 2+ (pathway denoted by black arrows). Proposed leptin receptor signaling is shown with black arrows while established signaling pathways in the delta cell are depicted with grey arrows. (B) Mouse model: Sst rtTA ;Tet-O-Cre;Lepr fl/fl (Lepr Dd ) and littermate control mice were placed on doxycycline from eight weeks to ten weeks of age. Mice were assigned to a high-fat or normal chow diet at ten weeks of age. GTTs were performed at eight, 14, and 36 weeks of age, corresponding to before dox treatment, and four and 26 weeks after doxycycline cessation.
islets. These primers span exons 18-19 and demonstrated the same melting curve product in islet and hypothalamic cDNA. RNA was isolated and cDNA was generated from whole islets or hypothalamus using Trizol plus the Qiagen RNeasy kit, as previously described (42). cDNA was synthesized using 100 ng RNA, as previously described (42). qPCR was performed using 1/3 of the total cDNA made, spread over six wells (42).

Bulk RNAseq
Human islets were obtained from the Integrated Islet Distribution Program (IIDP). Islet preparation and RNA extraction after intracellular marker cell sorting have been described previously (43). Briefly, islets from five non-diabetic human donors (Table 1) were suspended into a single-cell solution using 0.05% Trypsin, then fixed in 1% PFA, 0.1% saponin, and 1:100 RNasin (Promega). Cells were then indirectly fluorescently immunolabeled against somatostatin and insulin before sorting. RNA was extracted with a Recoverall Total Nucleic Acid kit (Life Technologies). RNA library was prepared and sequenced as previously described (44). Data was analyzed by the University of Pennsylvania Next Generation Sequencing Core, as previously described (44). scRNAseq scRNAseq data was downloaded f rom PANC-DB (pmacs.hpap.upenn.edu) on March 31, 2023, using cell calling performed by the Human Pancreas Analysis Program (HPAP). Non-diabetic, autoantibody-negative male and female donors >17 years of age were included in our analysis. Differential expression analysis was performed as previously described (39).

Results
Differential expression between bulksorted human beta and delta cells Current sorting protocols for live human islet cells cannot reliably distinguish delta cells from beta cells (45,46). Therefore, gene expression differences have been reported between bulk-sorted human alpha and beta cells (47) but not between bulk-sorted human beta and delta cells. To separate these two closely related cell types, we fixed islets from five non-diabetic human donors, suspended them in a single-cell solution, permeabilized them, and indirectly immunolabeled them with antibodies against insulin and somatostatin. These cells then underwent fluorescence-activated cell sorting. RNA was extracted from the purified beta-and deltacell populations and subjected to RNAseq and analysis. As expected, known lineage-specific genes such as SST and HHEX were enriched in delta cells (Table 2), while genes such as INS and MEG3 were more highly expressed in beta cells (Table 3).
An additional gene enriched within delta cells was LEPR, in agreement with a previous scRNAseq study (23). LEPR appears twice, once as the full-length (NM_002303, >39-fold enriched) and once as a short (NM_001198689, >44-fold enriched) isotype. Since our bulk-sorted RNAseq was performed on a mixture of male and female donors, we also examined gene expression differences between male and female beta and delta cells using publicly available scRNAseq data from the Human Islet Research Network's Human Pancreas Analysis Program (Table 4; (48)). The genes enriched within male and female delta cells were similar, although the magnitude of differential expression varied. LEPR was elevated in delta cells compared to beta cells by 1.490-fold and 1.328-fold with an adjusted p-value of 3.80 x 10 -308 or 9.99 x10 -311 in females and males, respectively. Although LEPR was not among the top 50 enriched delta-cell genes in males, it was in females ( Table 4). The difference in enrichment of LEPR in human scRNAseq data and bulk-sorted data was observed in most genes and likely reflects the drop-outs that occur in scRNAseq, since only a fraction of expressed genes are detected in any given cell, thus lowering mean expression for the population (49).

Delta-cell-specific deletion of Lepr in mice
Ablation of the Lepr within the mouse pancreatic epithelium results in alterations in glucose homeostasis, whereas ablation within alpha or beta cells does not. Together with the enrichment of LEPR in human delta cells, these results suggested to us that delta cells mediate the inhibited insulin secretion observed in response to leptin treatment. We, therefore, derived a mouse model (SST rtTA/+ ; Tet-O-Cre;Lepr fl/fl , henceforth referred to as Lepr Dd mice) to test this idea ( Figure 1B). Mice were treated with doxycycline for two weeks. In addition to its presence in islet delta cells, somatostatin is also expressed in gut enteroendocrine cells. We, therefore, performed a four-week washout to allow the replenishment of intestinal SST +

Analysis of Lepr gene recombination and RNA expression
We next examined Lepr locus recombination using DNA extracted from pancreatic sections from 36-week-old Sst rtTA/+ ;Tet-O-Cre;Lepr +/fl , Sst rtTA/+ ;Tet-O-Cre;Lepr fl/fl , Sst rtTA/+ ;Lepr +/fl , and Sst rtTA/+ ;Lepr fl/fl mice treated with doxycycline from eight to ten weeks of age. The Lepr fl allele was designed with loxP elements flanking the first exon of Lepr (38). We used primer sets spanning either one loxP element or both loxP elements to directly examine DNA recombination (Figure 2A). Primers traversing only one loxP element amplify unrecombined DNA, which in our model would be expected from non-delta cells. Amplification with these primers resulted in one (Lepr fl/fl mice) or two bands (Lepr +/fl ) in all samples ( Figure 2B), with one band indicating homozygosity of the floxed allele and two bands indicating one wild-type and one floxed allele.

Gene Symbol Fold Enrichment FDR
In the absence of recombination, primers flanking both loxP elements span an approximately 5kb distance, a distance too long for amplification using standard PCR reagents (Figure 2A To test whether Lepr expression was reduced in Lepr Dd delta cells, whole islets were isolated from 14-week-old Lepr Dd and control mice for quantitative RT-PCR analysis. No difference was observed in Lepr expression between Lepr Dd and control mice ( Figure 2D). Others have reported decreased Lepr expression when this Lepr fl mouse was used for deletion in other tissues (38). No difference in mRNA, together with the recombination observed in some pancreatic cells, suggested that Lepr is expressed in cells other than delta cells within the islet. We reasoned that we were not detecting Lepr deletion because delta cells make up only a small percentage of mouse islet endocrine cells (~5%), and expression in other cell types would overwhelm differences generated by Lepr absence in delta cells.

Delta-cell-specific Lepr ablation in the mouse results in normal glucose homeostasis and insulin secretion
No difference in weight was observed between Lepr Dd and control mice on a high-fat or chow diet (Figures 3A-D). There were no significant differences in beta-cell mass (Figures 3E, F) or islet morphology (Figures 3G, H).
Intraperitoneal GTTs were performed with 2g/kg glucose at 8 weeks, 14 weeks, and 36 weeks of age, corresponding to the predoxycycline time point, and then four weeks or six months after doxycycline removal. No difference was observed in glucose tolerance in male (Figure 4) or female ( Figure 5) Lepr Dd mice at any time point, whether the mice were on a standard chow or highfat diet. To rule out the possibility that we were missing subtle differences, GTTs were also performed on males and females on a HFD diet using 1g/kg glucose ( Figure 6). Again, no alterations in glucose tolerance were observed. Insulin secretion was examined by islet perifusion at 36 weeks of age. Islets from male and female Lepr Dd mice displayed no significant differences in insulin secretion compared to controls (Figures 7A-D).

Lepr is expressed in endothelial cell clusters within the mouse islet
We next examined Lepr expression in mouse alpha, beta, and delta cells using a published dataset built from fluorescently sorted populations of live lineage-traced cells. Lepr was not detected in mouse alpha, beta, or delta cells ( Figure 8A, (31)). This result starkly contrasts those for the transcription factor Pdx1, which is observed at the RNA level in each population ( Figure 8B) despite the limitation of PDX1 protein expression to beta and delta cells.
Examination of a publicly available scRNAseq dataset uploaded to the PanglaoDB webserver with visualization capability (51) confirmed a lack of Lepr expression within the mouse islet endocrine population ( Figure 8C). Most Lepr-positive cells are located in endothelial cell clusters (labeled in Figure 8C). An additional cell cluster exhibits Lepr expression; cells in this cluster expressed markers of several cell types, including Tie-1, Sst, Ins, Pecam, and Pdx1, among others. These cells may be doublets, the droplets capturing these cells may have been severely contaminated  Asterisks denote genes with differing rank order. "No difference" denotes no differential expression compared to beta cells. z denotes opposite directions of differential gene expression.
with RNA from lysed cells, or the number of genes sequenced in these cells may be too low for good segregation, since the mean number of genes per cell was 1528 (52, 53). This cluster contains many Tie-1 + and Pecam + cells, both of which are endothelial cell markers. Because islet Lepr expression appears within the mouse endothelial cell population, we assessed whether it is also expressed in human islet endothelial cells, again using publicly available data uploaded to Panglao DB (51). Unlike in the mouse islet, only a small percentage of cells within the human islet endothelial cluster expressed LEPR and most are observed in the delta-cell cluster ( Figure 8D).

Conclusions
Multiple labs have demonstrated that leptin decreases insulin secretion from isolated islets from multiple species. In addition, beta cells from control mice hyperpolarize in response to leptin, but this response is lost in beta cells from global Lepr null mice (54). These data suggest that a cell type within the islet mediates altered insulin secretion in response to leptin. However, pinpointing this cell type has proven to be challenging.
Pdx1-Cre is expressed throughout the pancreatic epithelium; this Cre line induces recombination of Lepr in the islet endocrine and exocrine cells but not in the immune, endothelial, or smooth hypothalamic neurons, and RIP-Cre 25Mgn deletion of Lepr results in obesity, impaired glucose-stimulated insulin secretion, and a seemingly paradoxical reduced fasting blood glucose (19,56). Pdx1-Cre is also expressed in the brain, primarily in the hypothalamus, and the overall Pdx1-Cre;Lepr fl/fl phenotype may result from this limited neuronal Lepr ablation (22,56); however, whether neuronal deletion can explain altered insulin secretion Ca2 + influx in islets isolated from Pdx1-Cre;Lepr fl/fl mice is less clear. Neuronal effects on islets have been reported to be lost ex vivo (57). Glucagon and Gcg-Cre are expressed in the hindbrain, olfactory neurons, and gut enteroendocrine L cells, as well as in islet alpha cells (21). The transgenic Gcg-Cre exhibits silencing, with recombination in only 30-45% of alpha cells, and the inducible gene replacement Gcg-Cre ERT2 line suffers from glucagon haploinsufficiency (58). Somatostatin, Sst-Cre, and the Sst-rtTA used here are all expressed in gut enteroendocrine D cells, hypothalamic neurons, and the motor cortex (50).
Had the alpha or delta-cell Lepr ablation models resulted in altered glucose homeostasis, the cell type responsible would have been unclear. When using inducible Sst-and Gcg-driven lines, though, it is important to note a washout period can allow highturnover cell populations, like gut enteroendocrine cells, to regenerate while the more slow-growing islet endocrine cells retain their deletion.
The Pdx1-Cre;Lepr fl/fl phenotype data suggest that pancreatic epithelial cells are responsible for the leptin-induced depression of glucose-stimulated insulin secretion in isolated islets. However, loss of Lepr in alpha cells using an inducible Glucagon-Cre (21), in beta cells using Ins1-Cre (21), or in delta cells (demonstrated here) does not result in altered glucose tolerance or insulin secretion. Additionally, bulk RNAseq of sorted mouse alpha, beta, or delta cells cannot detect Lepr. In addition, when we examined Lepr expression in control and Lepr Dd islets, we saw no difference in Lepr expression. We initially attributed this lack of change despite seeing recombined DNA to the rarity of the delta cell. However, after examining Lepr in mouse islet scRNAseq data and in RNAseq data from sorted alpha, beta, and delta cells, we now deduce that detectable Lepr in the islet is likely expressed in endothelial cells (11,18).
We unsuccessfully attempted to detect LEPR protein expression in the mouse pancreas using immunofluorescence and immunohistochemistry using an antibody previously used to detect LEPR mouse endothelial cells in other tissues (59). These data may indicate that no LEPR protein is synthesized in the pancreas; however, multiple groups have had difficulty immunolabeling LEPR, even in the hypothalamus (60). Because of the difficulty of LEPR immunolabeling, Lepr-Cre lines have been utilized in conjunction with a reporter mouse to detect Lepr expression patterns. In a Lepr-Cre;TdTomato line, TdTomato did not colocalize with insulin or glucagon in tissue sections, and the images shown had no TdTomato expression within the islets (61). However, TdTomato was detected in cells located in circular structures outside of the islets resembling blood vessels.
Lepr is expressed in mouse T cells (62,63), endothelial cells (64), and smooth muscle cells (64). Here, we show the presence of Lepr in mouse islet endothelial cells using publicly available RNAseq data. These data suggest that, in the mouse, islet endothelial cells mediate the effect of leptin on insulin secretion. However, mice with Lepr deleted specifically in endothelial cells using a Tie2-Cre display unchanged ad libitum-fed blood glucose (41,59). A more systematic analysis of glucose homeostasis was not performed. Endothelial cells regulate beta-cell function and insulin release through factors such as Endothelin-1 and Thrombospondin-1 (65, 66). Thus, signaling changes between endothelial and beta cells in response to leptin treatment may mediate decreased insulin secretion in mouse islets. However, Pdx1-Cre should not have activity in endothelial cells; therefore, the endothelial cells are unlikely to underlie altered insulin secretion in the Pdx1-Cre;Lepr fl/fl model. In human islets, LEPR is seen in scant endothelial cells by scRNAseq; thus, in human islets, any effect of leptin on insulin release is unlikely through endothelial cells. Instead, LEPR enrichment within the human delta-cell population suggests that the delta cell mediates the reduced insulin secretion response to leptin in human islets.
In contrast to what likely occurs in humans, we provide evidence that the mouse delta cell does not control this response. Further work will be necessary to definitively determine how leptin affects insulin secretion in various species. These data should serve as a deterrent from extrapolating the effects and mechanisms of islet leptin treatment from one species to another. Lepr is not expressed in mouse delta cells. (A, B) RNAseq reads for Lepr (A) and Pdx1 (B) in sorted alpha, beta, and delta cells were obtained from publicly available data. An inset for exon 1 Pdx1 expression in alpha cells at a zoomed-in scale is included to demonstrate the appearance of lowerexpressed genes. (C) scRNAseq expression in mouse islets cells for Lepr, Sst, Pdx1, Ins, and two endothelial markers, Tie1 and Pecam, was obtained from the public database PanglaoDB (https://panglaodb.se SRA745567). (D) scRNAseq expression in human islet cells for LEPR, SSST, PDX1, INS, TIE1, and PECAM was obtained from PanglaoDB (SRA701877).

Data availability statement
Original datasets are available in a publicly accessible repository: https://www.ncbi.nlm.nih.gov/geo/, accession number: GSE243984. Publicly available datasets were analyzed in this study and can be found at hpap.pmacs.upenn.edu. Publicly available datasets were used as previously analyzed and can be found at https://huisinglab.com/islet_txomes_2016/index.html and at https://panglaodb.se/, accession numbers SRA745567 and SRA701877. All other data is available upon reasonable request.

Ethics statement
For studies involving human cadaveric material, a waiver was obtained by Johns Hopkins Institutional Review Board. The studies were conducted in accordance with local legislation and institutional requirements. The human samples used in this study were acquired from the Integrated Islet Distribution Program, funded by Special Statutory Funding Program for Type 1 Diabetes Research, administered by the NIDDK. The IIDP works with islet isolation centers throughout the United States to provide de-identified islets from cadaveric donors to researchers. Written informed consent for participation was obtained from relatives of donors by OPOs before the IIDP received and disseminated islets. The animal studies were approved by Johns Hopkins University Institutional Animal Care and Use Committee. The studies were conducted in accordance with the local legislation and institutional requirements.