Angiotensin I-Converting Enzyme type 2 expression is increased in pancreatic islets of type 2 diabetic donors.

Aims . Angiotensin I-converting enzyme type 2 (ACE2), a pivotal SARS-CoV-2 receptor, has been shown to be expressed in multiple cells including human pancreatic beta-cells. A putative bidirectional relationship between SARS-CoV-2 infection and diabetes has been suggested, confirming the hypothesis that viral infection in beta-cells may lead to new-onset diabetes or to a worse glycometabolic control in diabetic patients. However, whether ACE2 expression levels are altered in beta-cells of diabetic patients has not yet been investigated. Here, we aimed at elucidating the in-situ expression pattern of ACE2 in T2D respect to non-diabetic donors which may account for a higher susceptibility to SARS-CoV-2 infection in beta-cells. Material

The presence of SARS-CoV-2 receptor Angiotensin I-Converting Enzyme type 2 (ACE2) in pancreatic beta-cells was initially highly debated, with observed discrepancies among studies (21)(22)(23).At present, ACE2 expression in beta-cells is supported by multiple reports which clearly showed the expression of ACE2 and other co-factors (i.e.TMPRSS2, NRP1) in beta-cells (23)(24)(25)(26)(27)(28).As a consequence, pancreatic islets are susceptible to SARS-CoV-2 infection mainly due to the expression of ACE2 and its co-factors, as demonstrated in several studies showing that beta-cells can be infected in vitro and in vivo in patients who died as a consequence of COVID-19 (24,25,27,28).
ACE2 expression has been shown to be modulated by pro-inflammatory stress with multiple studies showing increased expression of ACE2 upon exposure to several cytokines and/or pro-inflammatory molecules (23,29,30,31,32).Notably, chronic inflammation is a welldescribed feature of obesity and T2D (33)(34)(35).Inflammatory processes are also activated in pancreatic islets, as demonstrated by multiple evidence from animal models and humans with obesity and/or T2D (reviewed in 36).In this context, pancreatic islets low-grade inflammation has been shown to be associated with progressive beta-cell failure (37)(38)(39).
In T2D, ACE2 expression has been shown to be increased in several organs exposed to a diabetic milieu (40)(41)(42).However, a systematic assessment of ACE2 expression in human pancreatic islets of T2D patients is still missing.
We here took advantage of the availability of a cohort of T2D and of non-diabetic multiorgan donors to investigate pancreatic islet ACE2 expression levels employing confocal immunofluorescence analysis and multiple antibodies directed against different ACE2 epitopes.

Ethics Statement and multiorgan donors pancreata
Studies involving human participants were reviewed and approved by local ethics committee at the University of Pisa (Pisa, Italy).Pancreata not suitable for organ transplantation were obtained with informed written consent by organ donors' next-of-kin and processed following standardized procedures.Human pancreatic tissue sections were obtained from pancreata of brain-dead adult non-diabetic and T2D multiorgan donors, COVID-
To confirm the validity of the staining 1 μg of polyclonal rabbit anti-human ACE2 (cat.ab15348 -Abcam, Cambridge, UK) was combined with or without 10 μg of the immunizing human ACE2 peptide (cat.15325 -Abcam, Cambridge, UK) and the staining was performed as describe above.

CD68-positive cells immunohistochemical staining
Pancreatic CD68-positive cells were detected using enzymatic-colorimetric immunohistochemical staining.After deparaffinization and rehydration (see above), pancreatic sections were subjected to blocking of peroxidase with 3% hydrogen peroxide in PBS 1X for 20 minutes.Then, the sections were subjected to heat-induced antigen retrieval using Tris-EDTA buffer (10 mmol/l Tris, 1 mmol/l EDTA, 0.05% Tween-20, pH 9.0) for 20min at 100°C.After cooling and incubation in 3% BSA in PBS 1X for 30min at RT to reduce nonspecific reactions, sections were stained in 3% BSA in PBS 1X for 1h at RT with mouse monoclonal anti-human CD68 (cat.M0876 -Agilent Technologies, Santa Clara, CA, USA) (final concentration: 0,4 mg/ml).After three washes in PBS 1X, sections were incubated with secondary antibody polyclonal goat anti-mouse HRP-conjugate (cat.115-036-003-Jackson ImmunoResearch, Philadelphia, PA, USA) diluted 1:500 in PBS 1X for 1h at RT. Subsequently, the sections were incubated with one drop of 3,3-Diaminobenzidine (DAB) chromogen solution (cat.RE7270-K, Novolink MAX DAB, Leica Microsystems, Wetzlar, Germany) for 5 minutes, to trigger the colorimetric reaction.After 10 minutes of incubation in water, the sections were incubated for 1 hour at RT with ready to use polyclonal guinea pig anti-human insulin (cat.IR002 -Agilent Technologies, Santa Clara, CA, USA) further diluted 1:5 in PBS 1X supplemented with 3% BSA.After three washes in PBS 1X, sections were incubated with secondary polyclonal antibody goat anti-Guinea Pig conjugated with Alkaline Phosphatase (AP) (cat.A18772-ThermoFisher Scientific, Waltham, MA, USA) (final concentration: 0,3 µg/ml).Subsequently, the sections were incubated with one drop of Liquid Fast Red (cat.K0640 -Agilent Technologies, Santa Clara, CA, USA) (a drop of chromogen in 3 ml of substrate Levamisole (cat.X3021 -Agilent Technologies, Santa Clara, CA, USA) one drop per ml of LFR of Levamisole) for 5 minutes.
Stained sections were then counterstained with hematoxylin (cat.MHS31 -Sigma Aldrich, St. Louis, MO, USA) for 4 minutes.After 1 hour of air dry the sections were covered with a drop of Faramount, Aqueous Mounting Medium, Ready-to-Use (cat.S302580-2 -Agilent Technologies, Santa Clara, CA, USA).

Image analysis
Images were acquired, as a single stack focal plane, employing a Leica TCS SP5 confocal laser scanning microscope system (Leica Microsystems, Wetzlar, Germany).
For confocal laser scanning microscope system sections were scanned and images acquired at 40× magnification.The same confocal microscope setting parameters were applied to all stained sections before image acquisition, in order to uniformly collect detected signal related to each channel.
Colocalization analysis between ACE2 and insulin was performed using LasAF software (Leica Microsystems, Wetzlar, Germany).The region of interest (ROI) was drawn to calculate the colocalization rate (which indicates the extent of colocalization between two different channels and reported as a percentage) as a ratio between the colocalization area and the image foreground.Evaluation of ACE2 expression intensity in human pancreatic islets was performed using LasAf software (www.leica-microsystem.com).This software calculates the ratio between intensity sum ROI (which indicates the sum ROI of the greyscale value of pixels within a region of interest) of ACE2 channel and Area ROI (μm 2 ) of human pancreatic islets.Both in colocalization and intensity measurement analysis, a specific threshold was assigned based on the fluorescence background.The same threshold was maintained for all the images in all the cases analyzed.
Insulin positive area was measured using Volocity 6.3 software (Perkin Elmer, Waltham, MA, USA).Relative insulin signal positive area was calculated as a ratio between Area Intensity sum ROI (μm 2 ) of insulin channel and Area ROI (μm 2 ) of each human pancreatic islets.
For CD68-positive cells detection and quantification, images of the entire section were acquired using NanoZoomer S60 Digital slide scanner (cat.C13210-01 -Hamamatsu Photonics, Hamamatsu City, Japan) and were displayed using the proprietary NDP.view2 software.Manual count of CD68 + cells on the entire section area was performed.

Statistical analysis
Results were expressed as mean ± Standard Deviation (S.D.).Comparisons between two groups were carried out using Mann-Whitney U test for non-parametric data (normality checked using Kolgomorov-Smirnov test).Differences were considered significant with p values less than 0.05.Clinical variable associations with ACE2 expression were checked using multiple least square regression analysis.Statistical analyses were performed using Graph Pad Prism 8 software.

Results
To detect pancreatic ACE2 protein expression and distribution, and to evaluate differences between non diabetic (ND) and type 2 diabetic (T2D) donors, we performed a quadruple immunofluorescence analysis on FFPE pancreatic sections obtained from n = 20 ND and n = 20 T2D multiorgan donors (Table 1 and ESM Table 1).To cross-validate the ACE2 staining results, we used two different anti-ACE2 antibodies: (i) a monoclonal mouse IgG2a anti-human ACE2 (R&D, MAB933), whose specificity was confirmed through an isotype primary antibody staining (ESM Figure 1A) and (ii) a rabbit polyclonal anti-ACE2 (Abcam, Ab15348), whose specificity was tested through a peptide competition assay and subsequent staining in ND donors pancreatic sections (ESM Figure 1A).
In ND and T2D donor pancreata, both antibodies showed signals indicating that ACE2 is expressed in pancreatic islets where it is mostly colocalized with insulin signal (ESM Figure 2A).A triple staining on ND FFPE pancreatic sections using ACE2-MAB933, insulin and glucagon antibodies confirmed the prevalent colocalization of ACE2 with insulin in comparison to ACE2 with glucagon (ESM Figure 2A), in line with previous data (23).
The present results confirm that ACE2, in pancreatic islets, is prevalent in beta-cells as expected.
Outside pancreatic islets, ACE2-positive cells showed a vasculature-like morphology and distribution; such results were confirmed by using two different ACE2 antibodies (ESM Figure 3A); a subsequent co-staining with ACE2 and vascular-endothelial marker CD31 showed the juxtaposition of the two signals (ESM Figure 3A), thus confirming our previous observations (23), in line with other reports (22) , which demonstrated the localization of ACE2 also in pancreatic vascular cells.
Next, we focused on ACE2 expression in pancreatic islets.To evaluate putative ACE2 expression differences between ND and T2D, we performed an analysis of the intensity of ACE2 signals including a total of n=1082 islets.Both antibodies revealed a higher intensity of ACE2 in T2D compared to ND pancreatic islets (Figure 1A).Analysis of ACE2 intensity confirmed the significantly increased expression of ACE2 in T2D pancreatic islets compared to ND donors as measured by R&D and Abcam antibodies (greyscale values of ACE2-MAB933 in T2D=52.5±34.6 and in ND =37.1±28.1,p < 0.001; greyscale values of ACE2-ab15348 in T2D=53.2±63.5 and in ND =27.3±22.3,p < 0.001) (Figure 1C, 1D).
Since ACE2 expression in pancreatic islets is mostly prevalent in beta-cells, we performed a colocalization rate analysis (reported as the percentage of the overlap of INS and ACE2 signals) of ACE2 MAB933/INS and ACE2 Ab15348/INS in T2D and ND pancreatic islets.
In the multiple linear regression analysis (ESM Table 2), ACE2 expression (reported as staining intensity) was not associated with age, BMI, gender, ICU stay or duration of cold ischaemia time, thus excluding the influence of these putative confounding variables on ACE2 levels.Of note, ACE2 expression was not associated with gender or blood glucose

levels (ESM Table 2).
To investigate a possible link between ACE2 expression and inflammation triggered by innate immune cells in the pancreas of T2D donors, we analysed CD68 + macrophages in pancreatic tissue.In the whole pancreatic section (Figure 2A), CD68 + macrophages showed an abundance of about 6.2 cells per mm 2 in T2D and 5.1 cells per mm 2 in ND.Interestingly, CD68 + macrophages in the peri-islets showed an increased abundance trend in T2D compared to ND donors (0.22 vs. 0.15 CD68+ cells/islet) (Figure 2B); although not statistically significant, this result is consistent with the increased expression of ACE2 observed in pancreatic islets of previous serial sections.
Overall, these data show an increased expression of ACE2 in pancreatic beta cells in T2D compared to ND donors.Although such increase is independent of available clinical variables related to glycometabolic outcomes, we observed a tendency to increase in periislets CD68 + -macrophages thus putatively associating ACE2 expression increase to inflammatory insults.
In this study, we analysed an extended cohort of T2D multiorgan donors in comparison to age-and sex-matched non diabetic ones, to evaluate ACE2 expression and distribution.We demonstrated that ACE2 is increased in pancreatic islets of T2D donors and showed a higher colocalization rate in beta cells of T2D versus ND donors.Notably, we considered a total of n=1082 pancreatic islets across all ND and T2D donors, and the results were obtained using two different anti-ACE2 antibodies (monoclonal ACE2-MAB933 from R&D and polyclonal ACE2 Ab1538 from Abcam) adopted in the immunofluorescence analysis in an experimental cross-validation approach.The higher colocalization rate between ACE2 and insulin in T2D suggests that, in pancreatic islets, ACE2 hyperexpression is mainly occurring in beta-cells; however, at this stage, we cannot decipher whether (i) ACE2 expression is increased in beta-cells already expressing the receptor, (ii) it is increased due to de-novo expression occurring in ACE2-negative beta-cells, or (iii) a combination of both mechanisms.Additional analyses using imaging machine learning approaches and single cells segmentation are required to further decipher the intra-islet expression pattern of ACE2 in T2D.Overall, our data support an increased expression of the SARS-CoV2 receptor in beta-cells of T2D donors, and are in-line with a recent report demonstrating the upregulation of ACE2 in pancreatic islets of T2D donors subjected to microarray and RNA sequencing (43); Indeed, Taneera and colleagues showed that ACE2 is elevated in diabetic islets but no correlation between its expression and HbA1c, age or BMI was detected, similarly to what we have observed in the present study.In contrast to our results, other previous reports did not observe the upregulation of ACE2 in T2D islets (21,22); this can be due to the high heterogeneity of ACE2 expression or differences among T2D cohorts and/or reagents adopted.It is worth noting that in the present study we two different antibodies after a detailed analysis of their specificity and efficiency testing.
Previous studies showed that other organs exposed to a diabetic milieu such as lung, kidney and heart showed the upregulation of ACE2, thus corroborating our findings in a different context (40)(41)(42).Indeed, ACE2 expression was found increased in bronchial epithelium and alveolar tissue of T2D donors and a linear relationship was detected between blood glucose levels and ACE2 expression in alveolar tissue (40).In the heart tissue, ACE2 expression was significantly increased in cardiomyocytes of T2D patients with poor glycaemic control respect to ND patients and T2D patients with good glycaemic control (41).In kidney organoids ACE2 was expressed in tubular-like cells and an oscillatory glucose regimen induced the expression of ACE2 (42).Collectively, we can hypothesise that ACE2 expression is increased upon exposure to inflammation and/or high glucose or other stressors and that such chronic stress stimuli also exert their deleterious effect on beta cells favouring the upregulation of ACE2.However, unlike other reports, we cannot find a significant correlation between ACE2 expression and blood glucose levels.This can be explained by the high level of glycaemia already observed in ND patients during the ICU stay (Table 1); alternatively, we can hypothesize that, at least in beta-cells, high glucose is not the main factor leading to the hyperexpression of ACE2 and that pro-inflammatory molecules may play a major role in ACE2 modulation.The latter hypothesis is also supported by Van der Heide and colleagues who did not find any association between ACE2 expression and high glucose exposure in beta-cells (26).In addition, our previous study showed that the in vitro exposure of the beta-cell line EndoC-βH1 or primary pancreatic islets to pro-inflammatory molecules (i.e.IFNγ+IL-1β+TNFα or IFN α), but not metabolic stressors such as palmitate, can significantly increase ACE2 expression (23), thus supporting the hypothesis of a major inflammatory-mediated mechanism governing ACE2hyperexpression in beta-cells.
In line with this hypothesis, we explored the potential contribution of inflammation in ACE2 upregulation in T2D pancreatic islets, by analysing pancreatic-tissue resident CD68 +macrophages.We observed an increase of peri-islets CD68 + macrophages, even though not significant.However, such increase is supported by other previous observations which showed a peri-islets increase of CD68 + -macrophages in T2D pancreata in comparison to ND donors (44,45).Thus, additional analyses should be considered to further explore the inflammatory mechanisms leading to ACE2 upregulation in beta cells in T2D.
It has been suggested that increased expression of ACE2 may explain the increased infectivity or severity of COVID-19 in patients with diabetes (46).In this context, we can argue that increased expression of ACE2 in beta cells may lead to increased susceptibility of beta-cells to SARS-CoV-2 infection, making beta cells more prone to virus tropism in patients already infected with SARS-CoV-2.However, further analyses are needed to explore the expression of ACE2 cofactors (i.e.NRP1, TMPRSS2) in T2D islets and to decipher their contribution in the putative enhancement of the susceptibility of beta-cells to SARS-CoV-2 infection during COVID-19.
In conclusion, we observed the upregulation of ACE2 in pancreatic islet beta-cells of T2D donors, putatively driven by inflammatory-mediated mechanisms.Higher ACE2 expression in T2D islets might increase their susceptibility to SARS-CoV-2 infection during COVID-19 disease in T2D patients, thus exacerbating glycometabolic outcomes and worsening the severity of the disease.

Table 1 .
Main clinical characteristics of ND and T2D subjects included in the study.