SARS-CoV-2 infects cells of the human exocrine and endocrine pancreas and interferes with beta-cell function

Preexisting diabetes increases the risk of a severe course of the pandemic coronavirus disease 2019 (COVID-19). Vice versa, exacerbations of a preexisting diabetes as well as new-onset diabetes have been reported upon SARS-CoV-2 infection. Thus, there is an imperative need to clarify whether human pancreatic endocrine cells organized within an islet of Langerhans are permissive for and affected by SARS-CoV-2 infection, and to elucidate the mechanisms underlying the development of diabetes upon COVID-19. Here, we (i) dened ACE2 and TMPRSS2 expression patterns in human pancreatic endocrine and exocrine cell types, (ii) employed human pancreatic islet cultures to demonstrate susceptibility to SARS-CoV-2 infection and to viral replication in β-cells, (iii) showed that SARS-CoV-2 attenuates glucose-stimulated insulin secretion, and (iv) tested remdesivir as eventually effective to prevent β-cell failure. In addition, we (v) visualized viral particles replicating in endocrine pancreatic cells and dene their subcellular localization patterns via transmission electron microscopy, and nally (vi) present examples of cell type specic pancreatic infection patterns of COVID-19 deceased patients. Overall, our data demonstrate that SARS-CoV-2 can infect both the exocrine and endocrine compartments of the pancreas and can perturb β-cell integrity, which might lead to an increased risk for diabetes.


Introduction
Initially, the pandemic coronavirus disease 2019 , caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was considered an exclusive lung disease eventually leading to serious respiratory symptoms 1 . In the meantime, accumulating experimental and clinical knowledge shows that SARS-CoV-2 also causes lesions in kidneys, heart, brain, and gastrointestinal organs [2][3][4][5][6][7] . SARS-CoV-2 tropism towards distinct tissues is governed by cellular factors expressed on target cells such as the viral entry receptor angiotensin-converting enzyme 2 (ACE2) 8 or the transmembrane serine protease 2 (TMPRSS2) 8 . ACE2 expression within islets of Langerhans has been reported but not yet shown to allow SARS-CoV-2 entry [9][10][11] . Diabetes mellitus presents Janus-like in COVID-19: First, preexisting diabetes per se increases the risk of a severe disease, requiring more intense interventions and increasing mortality 12, 13 .
Second, severe exacerbations of a preexisting diabetes as well as new-onset diabetes have been reported [13][14][15][16] and cases of non-immune mediated diabetes in previously healthy individuals become increasingly evident 17 . So far, it is unclear whether SARS-CoV-2 triggers immune-mediated β-cell ablation as in type 1 diabetes (T1D) or directly perturbs β-cell function leading to non-autoimmune-mediated diabetes. Recent evidence suggests that SARS-CoV-2 can infect human endocrine cells in vitro 18 . However, stem-cell derived immature human β-cells were employed and neither viral replication, β-cell function, viral tropism, nor putative rescue strategies were determined 18 . Moreover, single cell RNAsequencing performed by three individual studies did not reveal conclusive information on ACE2 and TMPRSS2 expression patterns across distinct pancreatic cell types 9,10,19 . Immunostaining showed even partially opposing expression patterns for ACE2 and TMPRSS2 in the exocrine and endocrine compartment of the pancreas 11, [20][21][22] . Thus, there is an imperative need to clarify whether human pancreatic endocrine cells organized within an islet of Langerhans are permissive for and affected by SARS-CoV-2 infection, and to elucidate the mechanisms underlying the development of diabetes upon COVID-19 [12][13][14][15][16] . Finally, promising antiviral drugs to treat SARS-CoV-2 infected patients such as remdesivir, which shorten the time to recovery in adult COVID-19 patients 23 , warrant testing of their capacity to resolve non-pulmonary facets of COVID-19.
Here, we (i) de ned ACE2 and TMPRSS2 expression patterns in human pancreatic endocrine and exocrine cell types, (ii) employed human pancreatic islet cultures to demonstrate susceptibility to SARS-CoV-2 infection and to viral replication in β-cells, (iii) showed that SARS-CoV-2 attenuates glucose-stimulated insulin secretion, and (iv) tested remdesivir as eventually effective to prevent β-cell failure. In addition, we

Methods
Speci c information about material and methods is provided in the supplemental information.

Results
ACE2 and TMPRSS2 expression in endocrine cells and a ductal subpopulation.
To determine susceptibility to ex vivo infection, human pancreatic islets isolated from three human donors were exposed to SARS-CoV-2, and expression of viral spike (S) and nucleocapsid (N) protein as well as endocrine cell markers were analyzed. S and N proteins were not detected at day 1 but became readily detectable at day 3 and 5 post infection (Figure 2A 35,36 and found that many of the infected hormone-negative cells were still positive for PDX1, suggesting that endocrine cells lose hormones upon infection ( Figure 2C). Increasing intra-and extracellular viral RNA levels in islets of donors 2 and 3 in the absence of remdesivir treatment indicated progressive viral replication ( Figure 2D,E). Productive viral replication in islets of all donors was con rmed by increasing infectious viral titers in the respective supernatants ( Figure 2F). Upon remdesivir treatment, almost no infectious virus was detected in supernatants of islets ( Figure 2F), indicating e cient inhibition. This is in line with low viral RNA levels ( Figure 2D,E) and absence of N or S protein in confocal microscopy analyses (Figure 2A  coronavirus morphology 37-39 indicating productively infected endocrine cells. The virus containing vesicles are formed in the perinuclear region and processed to the cell surface. Furthermore, infection resulted in a marked reduction of endocrine secretory vesicles which seemed to be enlarged and maintained in the perinuclear region. In contrast, we were unable to detect intracellular virus particles or morphological alterations in remdesivir-treated islet cells ( Figure 3A; Supplementary Figure 4A). Thus, SARS-CoV-2 caused a damage pattern in human endocrine cells similar to previously reported TEM phenotypes of lung and gut derived cells 37-39 . To analyze whether SARS-CoV-2 infection and associated cell damage of the islets affects function, we assessed the islet response towards a high glucose pulse. We found that glucose-stimulated insulin secretion (GSIS) was induced in all conditions but the magnitude of induction was reduced in infected islets ( Figure 3B,C). However, GSIS in SARS-CoV-2infected human islets was only marginally restored by remdesivir treatment ( Figure 3B;C). The responsiveness to glucose, however, was lower in two islet preparations, most likely due to the limitations of prolonged ex vivo culture (Supplementary Figure 4C,D). These data suggest that glucose sensitive insulin secretion of β-cells could be impaired by SARS-CoV-2 infection.
Transcriptional changes in human islets after SARS-CoV-2 infection.
Of note, ISGs were not only upregulated after SARS-CoV-2-infection compared to uninfected, but also to remdesivir treated islets (Supplementary Figure 5C). Gene ontology (GO)-term analysis con rmed an initiation of a transcriptional cellular defense reaction in response to SARS-CoV-2-infection. Terms like defense response to virus and regulation of viral genome replication were strongly upregulated after SARS-CoV-2-infection (Supplementary Figure 5D). Albeit not all of these terms were depleted after remdesivir treatment, several interferon related terms like interferon alpha/beta signaling and type I interferon signaling pathway were reverted (Supplementary Figure 5F). Partial e cacy of remdesivir treatment was further supported by the speci c downregulation of COVID-19 related disease terms after remdesivir treatment (Supplementary Figure 5E,G). In contrast, no terms were signi cantly enriched in the uninfected or remdesivir treated infected islets compared to SARS-CoV-2 infected islets. Gene set enrichment analysis (GSEA) further con rmed the enrichment of interferon signaling in SARS-CoV-2 infected islets against uninfected and remdesivir treated infected islets (Supplementary Figure 5H-K). In addition, a trend indicating loss of β-cell identity as revealed by several gene sets 48 as well as defects in protein secretion in virally infected islets could be detected (Supplementary Figure 5H-K). Vice versa, these defects were attenuated upon remdesivir treatment pointing toward a partial transcriptional rescue in accordance with our functional ex vivo experiments (Supplementary Figure 5H-K). Thus, on transcriptional level infected islets show innate defense reactions and loss of β-cell identity that is reduced upon remdesivir treatment.
Finally, we aimed to validate the relevance of these ex vivo ndings for the effects of SARS-CoV-2 infection in the pancreas of infected individuals. Therefore, we obtained pancreatic specimens after autopsy of four COVID-19 deceased patients (two with type 2 diabetes, one with otherwise not speci ed endocrine insu ciency, one without available past medical history), which were reviewed by a pathologist based on hematoxylin and eosin stained sections (Supplementary Table 2). Staining for SARS-CoV-2 N protein detected varying numbers of positive cells in all four patients indicating robust pancreatic infection during severe COVID-19 ( Figure 4A). Speci cally, viral N protein was detected in some small ducts, single or grouped acinar cells as well as in endothelial cells, in agreement with ACE2 expression patterns in intra-islet vessels ( Figure 4A, close ups; compare to Supplementary Figure 2-C and 21 ). Notably, some neighboring cells of infected cells frequently showed faint red staining potentially pointing toward N protein levels crossing the detection threshold ( Figure 4A). N protein positive cells (stained in red) were not randomly scattered across the human pancreas but instead occurred in clusters of infected cells indicating localized viral spread ( Figure 4A,B). To probe infection of human β-cells, we performed immunohistochemical double staining for the viral N protein and insulin, but only observed a few double positive cells ( Figure 4A, close ups marked with #). Nevertheless, N protein positive cell clusters were located in close vicinity to the islets of Langerhans, indicating a certain degree of association between SARS-CoV-2 infection and the endocrine compartment. This was quanti ed by a vicinity score based on the distance between N protein to insulin positive cell clusters and classi ed in cells with a distance < 100 µm or ≥ 100 µm against a randomly calculated reference distance. On average, 51% of SARS-CoV-2 infected cell clusters were located close to human islets, with a signi cant maximum of 83% and 60% in patients no. 1 and no. 4 as well as similar trends with 40% and 31% in patients no. 2 and 3, respectively ( Figure 4B,C). Again, some insulin positive cells revealed a faint red N protein signal pointing towards spreading infection ( Figure 4D, arrow heads). In line, morphology of some of the clearly infected cells neither resembled ductal, acinar or endocrine morphology indicating a certain degree of plasticity occurring after infection. We hypothesized that relevant viral infection in β-cells could lead to perturbed hormone secretion and loss of endocrine granules as suggested by TEM and immunostaining of infected islet explants (Figures 2, 3). Indeed, high N protein signals appeared to go along with low insulin staining intensity ( Figure 4D,E, close up) supporting our hypothesis that SARS-CoV-2 infection can trigger hormone loss. To further address this, we co-stained for the N protein and NKX6.1, which is exclusively expressed by β-cells within the adult pancreas 49 . Indeed, we detected N-/NKX6.1-double positive cells in four out of four patients in close proximity to islets of Langerhans and SARS-CoV-2-infected cell clusters ( Figure 4F; Supplementary Figure 6). This indicates that infection in β-cells might lead to loss of hormone positivity, an observation supporting the results of our ex vivo assays (Figure 2; Supplementary Figure 3). Taken together, pancreatic SARS-CoV-2 infection appears likely to occur frequently in severe cases of infection, and in a complex probably multistep process including the exocrine and endocrine compartment.

Discussion
The disease course in COVID-19 patients can be perturbed by diabetes mellitus in two ways. On the one hand, diabetes is a risk factor for severe disease 16,17 , and contrariwise SARS-CoV-2 infection may directly compromise glucose metabolism 16,17 . Speci cally, ketosis and ketoacidosis were observed during and after COVID-19, both being strong clinical indicators of an absolute lack of insulin due to β-cell loss or malfunction 14  We demonstrate that endocrine cells of human pancreatic islets express ACE2 and TMPRSS2. Consistently, there was a strong ACE2 positivity of intra-islet endothelial cells as similarly reported in other studies. In contrast, these studies did not report ACE2 expression in endocrine cells 20,21 . However, we observed that ACE2 expression in β-cells is variable among various human subjects suggesting that these differences are not technical problems but rather donor variations 11, 18,20,21 . Notably, the strong inter-donor variation in the frequencies of ACE2-expressing endocrine cell types could also provide a potential explanation for varying diabetic phenotypes in COVID-19 patients. According to the current state of knowledge, ACE2 expression is the major determinant of SARS-CoV-2 entry and thus organ tropism, whereas several proteases can prime the coronavirus S protein 6,8,22,56,57 , albeit pancreatic TMPRSS2 expression also matched our infection pattern of deceased COVID-19 patients. Our ex vivo data employing human islets preparations revealed productive SARS-CoV-2 infection. Thus, our results agree with those of a previous study reporting SARS-CoV-2 susceptibility of human pluripotent stem cell-derived endocrine cells 18 . We found that ACE2 and TMPRSS2 are hardly co-expressed in δ-and α-cells. However, at low frequency these cell types could be a target of infection too. Previous data support this hypothesis 18 , however, the latter might be due to a fetal maturation grade of their pluripotent stem cell-derived origin. Future studies to comprehensively explore the precise SARS-CoV-2 infection pattern in human islets and endocrine subpopulations are still warranted.
In contrast to other model systems e.g. gut organoids 39,58 , our data suggest low to moderate replication of SARS-CoV-2 in pancreatic islets. We visualized SARS-CoV-2 particles inside vacuoles in the perinuclear region 37 of endocrine cells by transmission electron microscopy (TEM). The most striking observation was an enlarged and vacuolized ER-Golgi intermediate compartment, similar to observations in SARS-CoV-2 infected intestinal, kidney, and airway epithelial cells 37-39 . The hallmarks of endocrine differentiation, namely secretory granules, are displaced and reduced. However, a more comprehensive TEM-based analysis across a complete viral replication cycle in human islets as well as more samples from infected patients is required to draw more de nite conclusions. Nevertheless, the TEM observations are in line with the impaired insulin secretion observed in our study albeit we faced experimental variations across the three investigated islet preparations. Of note, e.g. β-cells infected by enterovirus display decreased glucose-stimulated insulin secretion and loss of Golgi structure 59 . However, also dedifferentiation of β-cells mimicking reversal to a progenitor state accompanied by decreased β-cellspeci c gene transcription may occur after viral 54 but also chemical 60 injury. Our RNA-sequencing, confocal microscopy analysis and TEM data would be in line with both hypotheses namely ER stress followed by β-cell degranulation and dedifferentiation. However, pancreatic virus-induced injury is also a self-potentiating damage driver due to cytokine release. In fact, SARS-CoV-2 infection provoked a broad signature of cytokines and interferon (IFN)-stimulated genes (ISGs) attributed to type I and III IFN responses in human islets. We recently showed that Interferon-induced transmembrane (IFITM) proteins promote SARS-CoV-2 infection of human lung cells 61 . Of note, IFITM1-3 ranged top amongst upregulated transcripts in SARS-CoV-2-infected human islets. Similar gene ontology (GO)-terms have been reported in gut-derived organoids after SARS-CoV-2 infection 39 identifying such intrinsically triggered immune response as a general feature across distinct organs during COVID-19.
Our results show that viral replication in ex vivo infected islets was e ciently inhibited by remdesivir. However, inhibition of viral replication was neither associated with an entire rescue in β-cell function nor full restoration of transcriptomes. This is most likely due to a delay in full β-cell recovery, which cannot be reached in the present experimental setting due to the deterioration of islets upon prolonged ex vivo culture. However, diabetic exacerbation or manifestation in COVID-19 patients might follow several mechanistical routes being eventually transient, as dedifferentiated or hormone-negative β-cells may fully recover after immune clearance of the virus 60 .
Our investigation of deceased COVID-19 patients and their pancreatic infection pattern sheds light on an obviously complex and multilayered infection process. The most striking observation in all four investigated samples was the scattered distribution of the infected cell clusters across the pancreas, most visible in the exocrine compartment but with a high vicinity to the islets of Langerhans in four patients. Such pattern faithfully recapitulates spread to neighboring pancreatic cells originating from a few infected cells eventually reached by viral particles directly via the blood stream during temporary viremia typically occurring in severe COVID-19 62,63 . Of note, systemic dissemination of genomic material of SARS-CoV-2 is associated with a sepsis-like biological response and critical illness in patients with COVID-19 63 . While insulin-positive β-cells in vivo showed only subtle signs of SARS-CoV-2 infection, the βcell lineage label NKX6.1 clearly con rmed infection occurring in β-cells in all four investigated COVID-19 patients. Notably, pancreatic NKX6.1 expression is unique as no other transcription factor is restricted exclusively to β-cells within the adult pancreas 49 . As we also observed those hormone-negative cells in our human islet preparations, SARS-CoV-2 infection likely perturbs hormone-positivity by cytokine and/or ER stress followed by β-cell degranulation and dedifferentiation. Still, further analysis is necessary to fully understand the underlying pathomechanism. Robust infection of pancreatic ducts and acinar cells further explains elevated lipase levels and acute edematous pancreatitis in COVID-19 patients with SARS-CoV-2 associated pneumonia 5,15,64 .
Overall, our data demonstrate that SARS-CoV-2 can infect both the exocrine and endocrine compartments of the pancreas and can perturb β-cell integrity, which might lead to an increased risk for diabetes.
Treatment with remdesivir during ex vivo infection of islets with SARS-CoV-2 inhibited viral replication and partially restored β-cell integrity.

Declarations
Con ict of interest statement: The authors have no con icts of interest to declare.