Overexpression of the SARS-CoV-2 receptor angiotensin converting enzyme 2 in cardiomyocytes of failing hearts

Hospitalized patients who die from Covid-19 often have pre-existing heart disease. The SARS-CoV-2 virus is dependent on the ACE2 receptor to be able to infect cells. It is possible that the strong link between cardiovascular comorbidities and a poor outcome following a SARS-CoV-2 infection is sometimes due to viral myocarditis. The aim was to examine the expression of ACE2 in normal hearts and hearts from patients with terminal heart failure. The ACE2 expression was measured by global quantitative proteomics and RT-qPCR in left ventricular (LV) tissue from explanted hearts. Immunohistochemistry was used to examine ACE2 expression in cardiomyocytes, fibroblasts and endothelial cells. In total, tissue from 14 organ donors and 11 patients with terminal heart failure were included. ACE2 expression was 2.6 times higher in 4 hearts from patients with terminal heart failure compared with 6 healthy donor hearts. The results were confirmed by immunohistochemistry where more than half of cardiomyocytes or fibroblasts showed expression of ACE2 in hearts from patients with terminal heart failure. In healthy donor hearts ACE2 was not expressed or found in few fibroblasts. A small subpopulation of endothelial cells expressed ACE2 in both groups. Upregulated ACE2 expression in cardiomyocytes may increase the risk of SARS-CoV-2 myocarditis in patients with heart failure.

. Clinical background of included multi-organ donors. Table summarizes    Relative protein quantification. Tissue from six organ donors (age 19-63 years) and from four patients (age 28-65 years) in the heart failure group was used. The tissue was snap-frozen and stored at −80 °C until use. Proteins were extracted using a lysis buffer (50 mM triethylammonium bicarbonate (TEAB), 2% sodium dodecyl sulfate (SDS)). 30 μg from each sample and a reference sample were digested into peptides using filteraided sample preparation (FASP) 16 . The reference consisted of aliquots from each group to provide representative references. Peptides were labeled using TMT 11-plex isobaric mass tagging reagents (Thermo Fisher Scientific), according to the manufacturer's instructions. The TMT set was fractionated using basic reverse phase liquid chromatography (pH 10) into 20 fractions and analyzed in an Orbitrap Fusion Tribrid mass spectrometer interfaced with the Easy-nLC1200 liquid chromatography system (Thermo Fisher Scientific). The peptides were separated on an analytical C18 column using a gradient from 4 to 28% acetonitrile in 0.2% formic acid over 75 min. Relative quantification was performed using Proteome Discoverer version 2.4 (Thermo Fisher Scientific) and the Mascot search engine (v. 2.5.1 Matrix Science, London, UK) matching against the SwissProt H. sapiens database (July 2019). Peptide and fragment tolerances were set to 5 ppm and 0.6 Da, zero missed cleavages, variable methionine oxidation, fixed cysteine methylthiolation and TMT-6 modifications on lysine and peptide N-termini. Percolator was used for PSM validation at the 1% FDR threshold. TMT reporter ions were identified in the MS3 HCD spectra with a 3 mmu mass tolerance, and samples were normalized on the total peptide amount. The reference sample was the denominator and used to calculate abundance ratios. Statistical analysis was performed with Perseus software 17 (version 1.6.10.45), two-side students t-test on the Log2 protein abundance ratios for p value calculations. For the volcano plot the number of randomization was set to 250, FDR to 0.05 and S0 to 0.1.

RNA isolation and gene expression by qPCR.
Left ventricular tissue was obtained from 14 organ donors (age 19-74 years) and 11 patients with severe heart failure in conjunction with heart transplantation (age 27-67 years). The tissue was preserved in RNAlater. Total RNA was extracted using the reagents and equipment from Qiagen. Briefly, the tissue was homogenized with TissueLyser LT and Qiazol and then purified using a RNeasy Mini column with DNase1 treatment for removal of residual genomic DNA. The cDNA was prepared from total RNA using a High-Capacity cDNA reverse transcription kit with RNase Inhibitor #4374967 and TaqMan Gene Expression Master Mix #4369542 (Applied Biosystems).
The human TaqMan gene expression assay ACE2 Hs01085333_m1 was used for the gene of interest, and PPIA Hs99999904_m1 as the reference gene. The relative comparative method was used to analyze the RT-qPCR data (Sequence Detector User Bulletin 2, Applied Biosystems) and the relative quantification (RQ) values were calculated using PPIA as the reference gene and an in-house calibrator sample. Gene expression data are presented in relative units.

Immunohistochemistry.
Biopsies from the LV from seven organ donors and six heart failure patients were embedded in Tragacant mounting medium (Histolab Products AB, Gothenburg, Sweden), frozen in liquid nitrogen and stored at −80 °C. The frozen tissues were sectioned into 7 μm serial sections that were fixed in −20 °C acetone for 10 min and washed in Phosphate Buffer Saline (PBS). A 30-min blocking step followed, with 2% bovine serum albumin, 0.3% Triton-X100 and 5% goat serum (Invitrogen, Carlsbad, CA, USA) diluted in PBS.
Primary antibodies were diluted according to Table 3, added to the sections and incubated at 4 °C overnight in a humidified chamber. Results were visualized by staining with secondary antibodies: goat anti-rabbit Alexa Fluor 546 or goat anti-mouse Alexa Fluor 647 (Invitrogen) for 1-2 h at RT. To enable triple staining, cTnT antibody was conjugated with Alexa 488 using the Zenon kit (Invitrogen) and a 1:6 molar ratio. After incubation with secondary antibodies, the samples were washed and incubated with the Zenon conjugated cTnT antibody for another hour. Sections were fixed with Histofix (Histolab) for 15 min and mounted with Prolong Gold Antifade reagent with DAPI (Invitrogen). Corresponding isotype controls for the primary antibodies were used for determining the background and did not show any specific staining.
Image analysis and quantification. The results were visualized using an ECLIPSE Ti inverted microscope (Nikon Corporation, Tokyo, Japan). A Nikon DS-2Mv camera was used for brightfield histology images. For analyses of immunohistochemistry, fluorescence images were acquired with an Andor Zyla camera. Large images: 7 × 7 fields shot with the 20× objective were scanned at three Z levels to capture all parts of the large images in focus. Generally, four channels (DAPI, Alexa 488, Alexa 546 and Alexa 647) were acquired.
All images were exported to Image J software (v. 1.47 h, Fiji distribution) 18 for further analysis. For each channel, displayed pixel ranges were set so that most of the background was extinguished. Isotypic controls were treated in the same way. The composite photos were used for analysis of the expression of biomarkers. Two large images, composed of stitched photos of 49 fields by 20× objective were analyzed for each individual.
For quantification of ACE2 expression, an intensity threshold was first set for each staining using a customized plugin, to reduce background staining. Pixel values below the threshold were set to zero. The threshold was then subtracted from pixel values above the threshold in order to get a continuous distribution of pixel values. All images were treated equally. The threshold level was set based on background staining of isotype control images. Areas of quantification were created include most of the stained tissue, but excluding obvious artefacts and parts of the images not in focus (Suppl. Figure 1). Mean pixel intensity measurements were carried out on ACE2 expression for each image. Outliers were excluded based on the 1.5 interquartile range (IQR) method. In total, measurements of 3 images were regarded as outliers (two donors, one heart failure subject), and were excluded. In each case, the other replicate image was not regarded as outlier. After outlier exclusion, a mean pixel

Results
Histology. Tissue sections were stained with Hematoxylin Eosin and Picric Sirius red for histology examination (Fig. 1). LV tissue from the organ donor group displayed a normal myocardium histology. Tissue from the Heart Failure (HF) group showed histopathological changes, including excessive fibrosis, infiltration of adipocytes and hypertrophy of cardiomyocytes. Cardiomyocyte nuclei were enlarged and irregular, often with lipofuscin accumulation.
Relative protein quantification. We analyzed tissue from six organ donors and four patients from the HF group, using global quantitative proteomics, for relative quantification of proteins between samples. A total of 5989 proteins were quantified and among these, 669 proteins were statistically differentially expressed between the two groups (p value < 0.05). For an overall assessment of the proteomics of the two groups, we applied principal component analysis (PCA) to all expressed proteins (Fig. 2a). Principal component 1 (PC1) clearly separates the Donor and HF groups where the individuals within the groups cluster together. Donor six, 63 years old, with cardiovascular disease, diabetes and obesity (Table 1), reflects the biological variability among the organ donors and is an outlier compared with the other five healthy donors. ACE2 was one of the major contributors to the segregation of the two groups. Vulcano plot of al included proteins shows that ACE2 was one of the top 20 proteins that was overexpressed in the HF group (Fig. 2b) and was upregulated 2.6 times (Fig. 2c). Proteases TMPRSS2 and CTSL, relevant for SARS-CoV-2 virus entry into the cell were not detected (data not shown).

Gene expression.
Relative ACE2 mRNA levels were quantified using RT-qPCR and performed on individual tissue samples from 14 organ donors and 11 patients from the HF group. The results showed no difference in ACE2 mRNA levels between the groups (Fig. 2d). H He em ma at to ox xy yl li in n E Eo os si in n P Pi ic cr ri ic c S Si ir ri iu us s R Re ed d were analyzed for protein expression of ACE2 in different cell types. ACE2 antibody was combined with antibodies detecting the most common cardiac lineage markers; cTnT staining cardiomyocytes, CD31 for endothelial cells and TE-7 for fibroblasts. Tissue from three donors were negative for ACE2 expression while a few positive cells were detected in tissue from four donors (Fig. 3a-c). Donor six, the outlier with higher age and cardiovascular disease, displayed a few more ACE2 + cells (Suppl. Figure 2). In contrast, ACE2 + cells were detected in tissue from all six HF patients (Fig. 3d-f). When ACE2 expression in the images was quantified, the heart failure group had a significant higher expression compared to the Donor group (Fig. 3g). Cardiomyocytes in tissue from donor hearts did not express ACE2 since no co-expression of the cTnT and ACE2 was observed. The few ACE2 + cells were found between cTnT + cardiomyocytes (Fig. 4a-c).
For the most part, ACE2 and CD31 staining did not overlap in either of the groups (Suppl. Figure 5). However, a small subpopulation of CD31 + endothelial cells that did express ACE2 could be found (Fig. 4j arrows).

Discussion
We found that ACE2 levels were roughly three times higher and that ACE2 was one of the top 20 proteins that were relatively overexpressed in terminal heart failure patients that underwent heart transplantation. The relative overexpression was found in cardiomyocytes and fibroblasts. Upregulation of the SARS-CoV-2 virus receptor ACE2 could result in a greater likelihood of myocarditis and mortality in patients with heart failure, a possibility that has not been explored. Using immunohistochemistry, we confirmed a higher expression level of ACE2 in heart failure subjects compared to donors. Two different expression patterns were observed, where ACE2 was expressed by cTnT + cardiomyocytes in the myocardium from three of the heart failure patients or by TE7 + fibroblasts in the other three patients. Different underlying diseases behind the heart failure might explain the expression of ACE2 in fibroblasts or cardiomyocytes. Clinical background of the included patients did not provide any clear answers to evaluate this question further. Several previous studies have shown that ACE2 is expressed in the cardiovascular system. Using proteomics, Wicik et al. 19 found similar ACE2 expression levels in the respiratory and cardiovascular systems, supporting the theory that heart tissue is a potential target of SARS-CoV-2.
In the present study, the ACE2 mRNA levels showed no difference between healthy myocardium and heart failure in line with a previous study by Battle et al. 20 . In contrast increased mRNA levels of ACE2 in failing hearts were reported by Goulter et al. 21 . This discrepancy between our results remains unexplained. Notably, in the study by Goulter the control group consisted of donor hearts without chronic heart failure similarly to ours, whereas in the study by Battle limited data on the clinical background was provided. It could be hypothesized that differences in clinical background and medication could account for the discrepancies between studies.
Previous studies indicate that endothelial cells express ACE2 in many organs 22 as well as the heart 23,24 . The tendency for thromboembolic conditions and edema to be induced in severe COVID-19 has been ascribed to viral infections and the destruction of endothelial cells. Evidence of direct viral infection of the endothelial cells and diffuse endothelial inflammation was reported in COVID-19 patients. Varga et al. 25 therefore suggest that SARS-CoV-2 infection induces endothelialitis in several organs. Our results show that most of the high ACE2 expression in the terminal heart failure was located in cardiomyocytes or fibroblasts. Most of the CD31 + cells did not express ACE2 even if a small subpopulation of CD31 + /ACE2 + cells was found showing expression in endothelial cells and possibly pericytes. Previously, myocardium was examined and ACE2 expression found in macrophages, endothelium and myocytes 24 and in line with our results, qualitatively increased expression of ACE2 was found in heart failure. However, these studies 22,24 used light microscopy and histology for detection ACE2 but no co-staining with specific biomarkers was performed.
It should be acknowledged that global quantitative proteomics may not detect less abundantly expressed proteins. This might be an explanation for why we did not detect proteases TMPRSS2 nor CTSL, important for the SARS-CoV-2 virus entry into the cell. In line with our results Liu et al. 26 reported far higher ACE2 mRNA levels in human heart with low expression of TMPRSS2.
In conclusion, we report three times greater expression of the COVID-19 receptor in heart failure. Most of the ACE2 + cells co-expressed Troponin T, indicating that cardiomyocytes are possible target of the SARS-CoV-2 virus. The other cell type expressing ACE2 are the cardiac fibroblasts. This might explain the involvement of the myocardium in COVID-19 and the high mortality among patients with cardiovascular comorbidities. Ethics approval. The study was approved by the Research Ethics Board at the Sahlgrenska Academy, University of Gothenburg, Sweden, following the Helsinki Declaration. Consent to participate. Documentation of consent from the multi organ donors, stating that their organs could be used for other medical purposes than organ donation. Signed informed consent was obtained from heart failure patients before cardiac transplantation surgery.