SARS-CoV-2 Spike S1 Subunit Triggers Pericyte and Microvascular Dysfunction in Human Pancreatic Islets

COVID-19 is caused by infection with SARS-CoV-2 (1). Its initial clinical manifestations are mostly pulmonary and cardiovascular, but SARS-CoV-2 can also compromise the function of several other organs, including the endocrine pancreas. Indeed, a bidirectional relationship between COVID-19 and diabetes has been proposed. On one hand, epidemiologic studies have revealed that type 2 diabetes and other metabolic diseases are risk factors for worse outcomes of COVID-19 (2,3). Conversely, several case reports have suggested that infection by SARS-CoV-2, or by the evolutionarily related coronavirus SARS-CoV, can lead to transient hyperglycemia, altered plasma levels of islet hormones, insulin resistance, ketoacidosis, and even diabetes (4–7). Infection by SARS-CoV-2 has been considered a potential risk factor for the development of autoimmune diseases, including type 1 diabetes (8). However, evidence of this association is mixed, and it is unclear whether dysglycemia persists or decreases in severity over time (9). Recent studies have shown that glycemic abnormalities are transient and disappear once SARS-CoV-2 infection resolves (10).

The potential diabetogenic effects of SARS-CoV-2, even if transient, are not well understood. Most of this uncertainty derives from the lack of consensus regarding the pancreatic distribution of angiotensin converting enzyme 2 (ACE2), the host cell receptor for SARS coronaviruses (11). ACE2 has been identified in multiple organs (12), including the human pancreas, with some contradictory findings; some reports show ACE2 in β-cells only (13–15), in both endocrine and vascular cells (16,17), or restricted to the microvasculature in islet and acinar regions of the human pancreas (12,18,19).

Abundant data support involvement of the vasculature in COVID-19 pathophysiology, and vascular complications (e.g., inflammation, coagulopathy, thrombosis with embolism, myocardial infarction, and stroke) are observed in different tissues (20–22). There is clinical evidence that SARS-CoV-2 viruses can directly attack the vascular system (23), and in vivo imaging approaches have revealed perfusion abnormalities (hypoperfusion) in the brain (24) and lungs (25) in patients with COVID-19. Interestingly, the mural cells of the microcirculation, known as pericytes, express ACE2 in different capillary beds (26–29) and, when integrated into cortical organoids, serve as SARS-CoV-2 viral replication hubs (30). Their targeting by this coronavirus would impair microvascular function and blood perfusion in different tissues, because pericytes are major regulators of tissue blood flow (31) and crucial for microvascular homeostasis (32).

In this article, we report the potential impact of SARS-CoV-2 on pericyte function in the pancreatic islet, an endocrine miniorgan whose microvasculature is essential for adequate nutrient sensing and efficient release of glucoregulatory hormones into the circulation (33). Pericytes are important regulators of islet capillary diameter and blood flow (34–36) and the source of trophic factors crucial for β-cell maturation and proper function (37,38). Importantly, a subset of pericytes expresses ACE2 in human islets (18), and advential cells and pericytes are among the major targets of SARS-CoV-2 viruses on in vitro infection of human pancreatic islets (39). Using confocal microscopy and living pancreas slices from different organ donors without diabetes, we show that acute incubation with SARS-CoV-2 recombinant spike proteins increases basal islet pericyte [Ca²⁺]i and constricts capillaries. These effects are mediated by a loss of ACE2 from the pericytes’ plasma membrane and mimicked by ACE2 pharmacologic inhibition. Our study thus supports that a dysfunctional islet microvasculature could contribute to COVID-19–associated loss of glucose homeostasis.

We obtained living pancreas slices (120–150 μm) from the Network for Pancreatic Organ Donors With Diabetes and used them for physiologic experiments within 4–36 h of arrival. We also produced slices from pancreas pieces obtained locally at the Diabetes Research Institute, University of Miami (40). For details on slice preparation, please see the Supplementary Material. Organ donors in this study were all autoantibody-negative male and female donors of different ethnicities without diabetes, with ages ranging from 14 to 55 years (Supplementary Table 1).

Living human pancreas slices were incubated for 1 h at room temperature with 80 nmol/L recombinant SARS-CoV-2 (2019-nCoV strain) spike S1 subunit (cat. no. 40591-V08H; SinoBiologicals) or HCoV-OC43 spike S1 protein (cat. no. 40607-V08H1; SinoBiologicals). Slices were also incubated with Fluo4-AM (6 μmol/L) in 3 mmol/L glucose solution (3G) containing aprotinin (25 KIU), with either NG2–Alexa 647 (1:50 for 2 h; cat. no. Fab2585R; R&D Systems) or DyLight 649 lectin (3.3 mg/mL for 1 h; cat. no. DL1178; VectorLabs), as published (35,41). Slices were placed in an imaging chamber with continuous perfusion and imaged with a 40× lens (numeric aperture 0.8) under an upright confocal microscope (Leica TCS SP8 upright).

We recorded changes in islet pericyte [Ca²⁺]i and capillary diameter induced by angiotensin II (angII; 100 nmol/L; cat. no. 1158; Tocris Bioscience), angiotensin (1-7) (ang1-7; 100 nmol/L; cat. no. 1562; Tocris Bioscience), ACE2 inhibitor MLN4760 (10 μmol/L; cat. no. 3345; Tocris Bioscience), norepinephrine (20 μmol/L; cat. no. 5169; Tocris Bioscience), and endothelin-1 (10 nmol/L). We used ImageJ software (https://imagej.nih.gov/ij/) to quantify changes in pericyte [Ca²⁺]i and capillary diameter, as previously described (35). For details on image analysis, please refer to the Supplementary Material.

Immunohistochemistry and colocalization analyses were performed as previously published (18). Details on antibodies used are included in the Supplementary Material.

Living human pancreas slices (n = 2 per donor) were incubated in 300 mL 3G in a 24-well plate for 20 min at room temperature, followed by 1-h incubation with either 80 nmol/L SARS spike or 80 nmol/L HCoV spike and then only 3G for 20 min. To measure insulin secretion, after incubating with spikes, slices were stimulated for 30 min with 11 mmol/L glucose. Insulin secretion experiments were conducted at 37°C. Supernatants were collected, and angII levels were measured using the Angiotensin II EIA Kit (cat. no. RAB0010), and insulin levels in supernatant and total content were measured using an insulin ELISA kit (cat. no. 10-1113-01; Mercodia).

For statistical comparisons, we used Prism 9 (GraphPad software). Unpaired t tests were performed when comparing data obtained for slices treated with either spike, and paired t tests were performed when comparing changes in the same cell or vessel.

All data generated or analyzed during this study are included in the published article (and its online supplementary files). No applicable resources were generated or analyzed during the current study. Requests for further information and for resources or reagents should be directed to the corresponding author.

Microvasculopathy is one of the hallmarks of COVID-19 infection (20–25), and pericytes, the mural cells of the microcirculation, are potential SARS-CoV-2 targets in different tissues, because they express ACE2 (16,18,26–28). We examined ACE2 expression in the human pancreas and found that it was present in both endocrine and exocrine compartments (Fig. 1A and Supplementary Fig. 1). In pancreatic islets in particular, ACE2 was not expressed by insulin-positive β-cells but instead by cells that localized in perivascular regions within the endocrine parenchyma (Fig. 1B and F and Supplementary Fig. 1). These ACE2⁺ cells expressed different pericyte markers, such as neuron-glial antigen 2 (NG2) (Fig. 1C) and platelet-derived growth factor receptor β (Fig. 1D), but not the endothelial cell marker CD31 (Fig. 1E and F).

To further evaluate pericytic expression of ACE2, we analyzed single-cell RNA sequencing data for human islets from a publicly available database (islet genomics database from Dr. Kyle Gaulton’s laboratory (42)). Stellate cells (quiescent and activated) expressed the highest levels of Ace2 transcripts in human islets when compared with other cells with a perivascular location or with islet endocrine cells (Fig. 1G). In particular, Ace2 was highly abundant in islet quiescent stellate cells, which included pericytes, as shown by their high expression of Cspg4, the gene encoding the chondroitin sulfate proteoglycan NG2 (Supplementary Fig. 2), and Ace2 transcript levels significantly increased with donor age within this cell population (Fig. 1H). These data are in line with the fact that the proportion of islet pericytes that expressed ACE2 protein significantly increased with donor age, as evidenced by a higher degree of colocalization between NG2 and ACE2 in older donors (age 44–54 years) (Fig. 1I). Our data thus indicate that in human pancreatic islets, ACE2 is expressed by vascular pericytes, and the fraction of pericytic ACE2 increases with donor age. This could indicate a potentially greater impact of SARS-CoV-2 on vascular function in older individuals and partly explain why age has been reported as an independent predictor of mortality in COVID-19 (43).

ACE2 has important physiologic functions in the body besides functioning as the host cell receptor for SARS coronaviruses. In particular, ACE2 is part of the renin-angiotensin system (RAS), where it functions as a carboxypeptidase that counterbalances the effects of ACE by degrading the vasoconstrictor angII into the vasodilator ang1-7 (44). To examine the potential impact of interfering with ACE2 activity on microvascular function in human islets, we adopted the pancreatic slice technique (35,41). Living human pancreas slices from organ donors without diabetes (Supplementary Table 1) were incubated with a membrane-permeant calcium indicator (Fluo4-AM) to record changes in cytosolic calcium levels ([Ca²⁺]i) in pericytes (labeled with NG2–Alexa 647 antibody) (Fig. 2A) and islet capillary diameter (visualized with lectins) (Fig. 2D).

We first examined the response of islet pericytes and capillaries to exogenous administration of RAS peptides angII (100 nmol/L) and ang1-7 (100 nmol/L). AngII induced a robust increase in [Ca²⁺]i in pericytes in islets (on average, a 45% increase in baseline fluorescence), whereas ang1-7 had a small inhibitory effect on pericyte [Ca²⁺]i (<10% decrease in baseline fluorescence) (Fig. 2B and C). The angII-induced increase in [Ca²⁺]i in islet pericytes was accompanied by a constriction of adjacent capillaries (average decrease of ∼8% in basal diameter) (Fig. 2D–F and Supplementary Movies 1 and 2). In contrast, ang1-7 application did not induce a significant change in islet capillary diameter (basal diameter 5.1 ± 0.2 and 5.2 ± 0.3 μm 4 min after ang1-7; n = 9 vessels per 2 donors). The vasoconstrictor effects of angII were mediated by angiotensin type 1 (AT1) receptors, because their specific blocker losartan (20 μmol/L), a commonly used drug in clinic, completely abolished angII-induced increases in pericyte [Ca²⁺]i (Fig. 2G and H). These data are in line with the fact that Agtr1 is the angII receptor gene mostly expressed by quiescent stellate cells in human islets when compared with the levels of Agtr2 or Mas1 receptor genes (Supplementary Fig. 3).

To determine whether angII could be produced endogenously in the human pancreas, we then examined the expression of ACE, the enzyme homologous to ACE2, which produces angII. ACE was present throughout the endocrine and exocrine pancreas compartments (Supplementary Fig. 3), most likely expressed by endothelial cells, as shown by RNA sequencing (Supplementary Fig. 3). We then administered AT1 receptor blocker losartan in 3 mmol/L glucose to determine whether signaling through this receptor occurred under basal glucose levels. Losartan significantly decreased pericyte [Ca²⁺]i levels in human islets (Fig. 2I), indicating that AT1 receptors were sensing endogenous angII being produced under basal conditions. In summary, our data indicate that a functional RAS exists in the human endocrine pancreas. Its different components, ACE, ACE2, and AT1 receptors, are expressed by islet vascular cells, which are responsive to vasoactive angiotensin peptides to different degrees.

Because the human islet microvasculature was responsive to angII (Fig. 2), and ACE2 is involved in the degradation of this potent vasoconstrictor peptide, we then hypothesized that occupation of ACE2 by SARS-CoV-2 viruses could interfere with islet microvascular function. To mimic SARS-CoV-2 interactions with ACE2, we used a recombinant SARS-CoV-2 (2019-nCoV strain) spike S1 subunit (SARS spike). As a control, we used a similar recombinant spike S1 subunit (HCoV spike) from the endemic coronavirus strain HCoV-OC43, another betacoronavirus that uses sialoglycan-based receptors to infect cells instead of ACE2 (45).

Living pancreas slices from donors without diabetes were treated either acutely or incubated for 1 h with SARS spike or control (HCoV) spike, as well as with the [Ca²⁺]i indicator Fluo4 and NG2–Alexa 647 to identify pericytes (Fig. 3A). Acute application of SARS spike (applied for 5 min in 3G) with continuous perfusion followed by 30 min of recording with no perfusion (stop flow) led to a progressive increase in Fluo4 fluorescence levels in islet pericytes (Fig. 3B). Similarly, incubating slices for 1 h with SARS spike, but not with HCoV spike, also led to an increase in fluorescence levels in islet pericytes (fluorescence levels normalized to those of a region of interest placed around the whole islet), indicative of higher [Ca²⁺]i levels in these cells after 1-h exposure to SARS spike (Fig. 3C). We then recorded pericyte [Ca²⁺]i responses to exogenous angII administration. As shown in Fig. 2, angII induced a robust increase in [Ca²⁺]i in islet pericytes in slices treated with control spike (Fig. 3C and Supplementary Movie 3). Islet pericytes in slices treated with SARS spike were still responsive to angII stimulation, but the magnitude of the calcium response was reduced when compared with pericyte responses in control conditions (Fig. 3C–E). Interestingly, in response to angII, islet pericytes from slices treated with SARS spike reached maximum fluorescence levels (peak) ∼1 min earlier than those in control slices (Fig. 3C, arrows).

We then examined whether SARS spike treatment interfered with pericyte calcium responses to exogenous administration of other vasoactive stimuli, such as norepinephrine, endothelin-1, and ang1-7. Although SARS spike slightly decreased pericyte responses to vasoconstrictors norepinephrine and endothelin-1 (Supplementary Fig. 4), it potentiated in turn the inhibitory effect of ang1-7 on islet pericyte calcium (Fig. 3F). These data indicate that a relatively short (1 h) exposure to the S1 subunit from SARS-CoV-2 spike is sufficient to activate islet pericytes, not affecting their viability but altering their capacity to respond to different vasoactive molecules.

Because pericytes control capillary diameter in mouse and human islets (34–36), we then examined whether changes in [Ca²⁺]i induced by incubation with SARS spike also resulted in changes in islet capillary diameter. To this end, we incubated slices with either SARS or control spikes and with a fluorescent lectin to label the microvasculature (Fig. 4A). Although lumens were visible in a majority of capillaries in islets from slices treated with control spike (average capillary diameter ∼5 μm), capillaries in islets on SARS spike treatment seemed to be fully or partially constricted at certain locations (Fig. 4B–D and Supplementary Fig. 4). Incubation with HCoV spike did not interfere with the vasoconstrictor effect that angII had on human islet vessels (Fig. 4C, E, and F and Supplementary Movie 3). However, in slices incubated with SARS spike, angII application did not elicit a further reduction in islet capillary diameter, most likely because they were already constricted under these conditions (Fig. 4E and F). In contrast, exogenous administration of ang1-7, which did not significantly change islet capillary diameter in control slices, elicited a greater vasodilatory effect (∼17% average increase in capillary diameter) in slices incubated with SARS spike (Fig. 4G). In summary, our data show that the S1 subunit from SARS-CoV-2 spike affects the function of vascular cells in human islets, constricting islet capillaries and interfering with their responses to exogenous administration of vasoactive RAS peptides.

Previous studies have shown that binding of SARS spike protein to ACE2 causes its endocytosis (46). We therefore hypothesized that the effects of SARS-CoV-2 spike on islet pericyte [Ca²⁺]i levels and capillary diameter were associated with a decrease in ACE2 carboxypeptidase activity at the cell surface. To determine whether ACE2 activity was altered, we compared the effects of SARS spike with those of a pharmacologic inhibitor of ACE2, MLN4760. Acute inhibition of ACE2 with MLN4760 triggered an increase in [Ca²⁺]i in islet pericytes in slices incubated with HCoV spike (Fig. 5A), further supporting that angII was processed in the pancreas under basal conditions. In contrast, incubation with SARS spike abolished islet pericyte [Ca²⁺]i responses to MLN4760 (Fig. 5A and B), suggesting that ACE2 activity was already reduced under these conditions. Acute inhibition of ACE2 with MLN4760 did not elicit a significant change in islet capillary diameter in slices treated with either recombinant spike, but again, capillaries were constricted on treatment with SARS spike (Supplementary Fig. 4).

We then hypothesized that decreased ACE2 activity could be due to ACE2 internalization on SARS spike interaction. To test this hypothesis, we immunostained pancreas slices treated with either HCoV or SARS spike with antibodies against NG2 and ACE2 (Fig. 5C). In slices treated with HCoV spike, ACE2 was expressed at the pericytes’ plasma membrane and colocalized with NG2, but it seemed to be internalized in slices treated with SARS spike (Fig. 5D). Line profiles drawn across pericytes in confocal images (in regions excluding the cell nucleus) also reflected this enrichment in ACE2 fluorescence intracellularly, indicative of ACE2 internalization on treatment with SARS spike (Fig. 5E).

Because the primary function of ACE2 in tissues was to degrade the vasoconstrictor angII into the vasodilator ang1-7, we then examined whether abnormal ACE2 localization/function on SARS spike treatment affected endogenous pancreatic levels of angII by ELISA (Supplementary Fig. 5). Incubating pancreas slices with SARS spike for 1 h, but not with HCoV spike, led to an increase in the levels of angII peptide in the extracellular solution (supernatant collected before and after spike administration) (Fig. 5F). Our findings thus indicate that SARS spike S1 subunit compromises the location/activity of ACE2 and thus pancreatic levels of angII, affecting in this way the function of vascular cells in human islets.

Because pericytes are the source of different trophic factors (e.g., the neurotrophin nerve growth factor [38] or bone morphogenetic protein 4 [37]) that are required for proper β-cell function and insulin secretion, we then examined whether pericyte dysfunction caused by SARS spike and elevated pancreatic levels of angII would affect β-cell responses to high glucose. To this end, we incubated living pancreas slices with either control or SARS spike in 3 mmol/L glucose for 1 h and then stimulated for 30 min with 11 mmol/L glucose. SARS spike incubation did not affect insulin secretion in 3 mmol/L glucose (Supplementary Fig. 6A) or the amount of insulin that was released on high-glucose stimulation (Supplementary Fig. 6B and C). In summary, our study supports that potential metabolic effects of SARS-CoV-2 on the human endocrine pancreas would most likely involve a functional vascular network.

In this report, we describe the effects of SARS-CoV-2 spike S1 subunit on pericyte activity and on their capacity to regulate vasomotion in human pancreatic islets. We found that incubating living pancreas slices from organ donors without diabetes with SARS spike significantly altered pericyte [Ca²⁺]i levels and contractile properties, leading to islet capillary constriction and compromising further vasomotive response to exogenous RAS peptides. Because the microvasculature is necessary for proper endocrine cell responses to glucose (36–38), our study supports that islet vascular dysfunction could contribute to the loss of glucose homeostasis observed in some patients with COVID-19. Although in this study we used a recombinant spike protein that does not represent the whole SARS-CoV-2 viral particle, literature supports that a substantial amount of spike S1 subunit is actually found free as a subproduct of virus processing and shedding (47). This is the best strategy we have found to examine potential changes in human islet function without the serious health concerns that live SARS-CoV-2 viruses could pose.

Our study supports that a local RAS exists in the human endocrine pancreas, as previously shown for other species (48). This pancreatic RAS seems to be largely associated with the vasculature in rodents and in humans (Fig. 1 and Supplementary Fig. 3). The precursor angiotensinogen was previously detected in ductal epithelia and endothelial cells in the rat pancreas (49), and prorenin mRNA was detected in reticular fibers and connective tissue surrounding human islet vessels (50). High levels of the AT1 receptor gene, Agtr1, were expressed by quiescent and active stellate cells in human islets that included pericytes (Supplementary Figs. 2 and 3). AngII activated islet pericytes and constricted capillaries by acting on AT1 receptors, because the AT1 receptor blocker losartan abolished pericyte [Ca²⁺]i responses to this vasoconstrictor peptide (Fig. 2). This pancreatic RAS has been implicated as an important regulator of islet blood flow and insulin release (51,52). AngII administration in vivo decreases blood flow in the pancreas and the amount of insulin that is released on stimulation with glucose (51,52). These data are in line with our previous observations that reducing islet blood flow by directly and acutely activating pericytes has metabolic consequences (36). In particular, activating pericytes in intraocular islet grafts and increasing their [Ca²⁺]i levels, either using optogenetics or mimicking sympathetic input, constricts capillaries and leads to a reduction in islet blood flow, which in turn affects plasma levels of islet hormones and compromises responses to a glucose challenge (36). Activation of pericytes and islet capillary constriction were changes observed on short incubation with SARS-CoV-2 spike protein (Figs. 3 and 4), suggesting that SARS-CoV-2 viruses could partly have diabetogenic effects in vivo by targeting the microvasculature and indirectly compromising islet responses (Fig. 6).

We found that 1-h treatment with SARS-CoV-2 spike protein had no effect on endocrine cell function or hormone secretion ex vivo (Supplementary Fig. 6). This is surprising, because islet pericytes are the source of trophic factors that are necessary for β-cell maturation and insulin secretion, such as NGF and BMP4 (37,38). However, our treatment with SARS spike was a relatively short one (1 h) and most likely not enough to compromise pericyte secretory capacity and consequently affect β-cell responses to glucose. Longer exposure to SARS-CoV-2 spike, for instance on in vitro infection of isolated islets with live viruses for 3 days, results in impaired glucose-stimulated insulin secretory responses (15). SARS-CoV-2 infection could affect endocrine cell function through additional mechanisms, because different islet cells express angiotensin receptors. For instance, angII induces a dose-dependent decrease in glucose-stimulated insulin secretion from mouse islets in vitro (53) and in vivo independently of its vasoactive effects (54). In summary, SARS-CoV-2 spike presence and disturbed levels of RAS peptides may affect the function of the endocrine pancreas by targeting both the vasculature and endocrine cells. Future studies investigating how the pancreatic RAS operates in vivo would be required to fully characterize how islet responses and glycemic control are affected by RAS peptides under physiologic conditions and then on infection by SARS-CoV-2. Some of these studies may have to be conducted in hamsters, for instance, because they, unlike mice, are suitable to study the pathophysiology of COVID-19 (55).

In our study, we find that incubating living pancreas slices with SARS spike compromised ACE2 plasma membrane location and activity, raising endogenous pancreatic levels of angII (Fig. 5). Similar results were observed in a recent study that determined that SARS-CoV-2 spike RBD constricted the microvasculature in living brain slices from hamsters (26). In that study, the authors found that SARS-CoV-2 spike–induced capillary constriction reflected a decrease in the conversion of angII to ang1-7, which was mediated by removal of ACE2 from the cell membrane and mimicked by blocking ACE2 (26). Interestingly, ACE2 was also restricted to the microvasculature (endothelial cells and pericytes) in other endocrine glands, such as the thyroid, parathyroid, and adrenal glands (12). This suggests that SARS-CoV-2 may compromise microvascular function in the brain and in different endocrine tissues, which may underlie the multisystemic nature of COVID-19.

How SARS-CoV-2 affects the function of the endocrine pancreas has been a matter of debate. Although a transient elevation of glycemia can have multiple causes, and vascular dysfunction can contribute to it in multiple ways (namely, by promoting an inflammatory milieu (56–58)), it has also been proposed that SARS-CoV-2 viruses directly attack the pancreas. Viral antigens have been found in pancreatic autopsy samples from patients with COVID-19, detected only in islet β-cells (14) or spread throughout the endocrine and exocrine pancreas (13,15,17), specifically in endothelial, ductal, acinar, and mesenchymal cells (13,19). Other histopathologic alterations are also evident in COVID-19 pancreata, such as thrombotic lesions, fibrosis, necroptotic cell death, and immune cell infiltration (17,19). All these alterations occur at vascular niches and could be due to an unstable microvasculature caused by pericyte dysfunction, because these mural cells are crucial for microvascular homeostasis (32). In summary, our study supports an involvement of the microvasculature (of pericytes, in particular) in the pathophysiology of the endocrine pancreas in COVID-19.

Acknowledgments. The authors thank Drs. Stephen Roper and Madina Makhmutova (University of Miami) for careful revisions of the manuscript. The authors also thank the Network for Pancreatic Organ Donors with Diabetes (nPOD), in particular, the organ donors, their families, and the nPOD slicing team in Gainesville for producing living pancreas slices, as well as the Human Islet Cell Processing Facility at the Diabetes Research Institute (University of Miami) for pancreatic tissue samples. This manuscript used RNAseq data extracted from Dr. Kyle Gaulton’s islet genomics database (42) (https://www.gaultonlab.org/pages/Islet_expression_HPAP/).

Funding. This work was funded by National Institutes of Health grants RO1 DK133483, RO1 DK138471, and U01 DK135017 (J.A.), by The Leona M. and Harry B. Helmsley Charitable Trust Pilot Award for nPOD Team Science AWD-007061 (J.A.), and by JDRF (Breakthrough T1D) postdoctoral fellowship 3-PDF-2024-1503-A-N (L.M.G.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. C.A.B. and L.M.G. performed immunohistochemistry and functional imaging with living human pancreas slices and analyzed data. E.P. performed immunohistochemistry. R.D.C. performed insulin ELISA. R.L. performed single-cell RNA sequencing analysis. M.B. analyzed vasomotive responses. J.A. designed the study, analyzed data, and wrote the manuscript. All authors discussed the data and revised the manuscript. J.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Portions of this work were presented in poster form at the 84th Scientific Sessions of the American Diabetes Association, Orlando, FL, 21–24 June 2024.