key: cord-0787086-53db9whw
authors: Aguiar, Jennifer A.; Tremblay, Benjamin J-M.; Mansfield, Michael J.; Woody, Owen; Lobb, Briallen; Banerjee, Arinjay; Chandiramohan, Abiram; Tiessen, Nicholas; Cao, Quynh; Dvorkin-Gheva, Anna; Revill, Spencer; Miller, Matthew S.; Carlsten, Christopher; Organ, Louise; Joseph, Chitra; John, Alison; Hanson, Paul; Austin, Richard; McManus, Bruce M.; Jenkins, Gisli; Mossman, Karen; Ask, Kjetil; Doxey, Andrew C.; Hirota, Jeremy A.
title: Gene expression and in situ protein profiling of candidate SARS-CoV-2 receptors in human airway epithelial cells and lung tissue
date: 2020-07-16
journal: Eur Respir J
DOI: 10.1183/13993003.01123-2020
sha: f332bad3c29297fe115a932a096a810918691c41
doc_id: 787086
cord_uid: 53db9whw

In December 2019, SARS-CoV-2 emerged causing the COVID-19 pandemic. SARS-CoV, the agent responsible for the 2003 SARS outbreak, utilises ACE2 and TMPRSS2 host molecules for viral entry. ACE2 and TMPRSS2 have recently been implicated in SARS-CoV-2 viral infection. Additional host molecules including ADAM17, cathepsin L, CD147, and GRP78 may also function as receptors for SARS-CoV-2. To determine the expression and in situ localisation of candidate SARS-CoV-2 receptors in the respiratory mucosa, we analysed gene expression datasets from airway epithelial cells of 515 healthy subjects, gene promoter activity analysis using the FANTOM5 dataset containing 120 distinct sample types, single cell RNA sequencing (scRNAseq) of 10 healthy subjects, proteomic datasets, immunoblots on multiple airway epithelial cell types, and immunohistochemistry on 98 human lung samples. We demonstrate absent to low ACE2 promoter activity in a variety of lung epithelial cell samples and low ACE2 gene expression in both microarray and scRNAseq datasets of epithelial cell populations. Consistent with gene expression, rare ACE2 protein expression was observed in the airway epithelium and alveoli of human lung, confirmed with proteomics. We present confirmatory evidence for the presence of TMPRSS2, CD147, and GRP78 protein in vitro in airway epithelial cells and confirm broad in situ protein expression of CD147 and GRP78 in the respiratory mucosa. Collectively, our data suggest the presence of a mechanism dynamically regulating ACE2 expression in human lung, perhaps in periods of SARS-CoV-2 infection, and also suggest that alternate receptors for SARS-CoV-2 exist to facilitate initial host cell infection.

multiple mechanisms. Recent in vitro reports have demonstrated that similar host proteins are involved in facilitating cell entry by SARS-CoV-2, such as ACE2 and TMPRSS2 [5, 24] Biophysical and structural evidence strongly support an interaction of ACE2 with SARS-CoV-2 spike protein, similar to SARS-CoV spike protein [12, 13] . Molecular docking studies have also suggested that SARS-CoV-2 spike protein can interact with cell-surface GRP78 [25] . Indirect evidence for a role of CD147 in SARS-CoV-2 binding has been demonstrated in vitro with the use of an anti-CD147 intervention that prevented virus replication [26] . Furthermore, a clinical study with an anti-CD147 intervention reduced symptoms and duration of hospital admission for COVID-19 patients [27] . In summary, although there is evidence that SARS-CoV-2 and SARS-CoV both utilize ACE2 as a receptor to facilitate virus entry, it is possible that differences in host-entry mechanisms play a role in the large epidemiological differences between the two viruses, which may include additional unidentified receptors.

ACE2 and TMPRSS2 were identified as cellular entry determinants for SARS-CoV using mechanistic studies. The original report of in situ human lung ACE2 expression described positive immunohistochemical staining for alveoli and airway epithelial cells, and immunocytochemical staining in A549 type II alveolar epithelial cells [28] . ACE2 protein expression is also present in the human lung adenocarcinoma cell line, Calu-3 [29] . Similar to ACE2, the original report describing the expression of TMPRSS2 in human respiratory mucosa described expression in airway epithelium and type II alveolar epithelial cells [30] . The specificity of the ACE2 and TMPRSS2 antibodies used for analysis of expression patterns in human lung tissues remains to be addressed.

To address the uncertainties related to SARS-CoV-2 receptors in human lung, we performed gene expression and in situ protein profiling of candidate receptors in human airway epithelial cells and lung tissue. Our computational analysis used publicly available microarray gene expression datasets from airway epithelial cells of 515 unique subjects, single cell sequencing data from 10 subjects, and the FANTOM5 dataset for promoter activities of 74 lung-related cell and tissue types. Our proteomic analysis used data from the Human Proteome Map[31] and a dataset from primary human airway epithelial cells grown under air-liquid interface culture conditions [32] . For our in situ protein profiling, we performed immunohistochemical analysis of 98 human lung tissue samples. To determine antibody specificity, we performed immunoblots on protein isolated from Calu-3 cells, primary human airway epithelial cells, primary type II alveolar epithelial cells, the human bronchial epithelium cell (HBEC)-6KT cell line, the A549 type II alveolar epithelial cell line, and HEK cells.

Collectively our data contrast previous reports, demonstrating rare ACE2 protein expression in the airway epithelium and alveoli of human lung. Our protein expression data are consistent with low ACE2 promoter activity in a variety of lung epithelial cell samples and low ACE2 gene expression in both microarray and single cell RNA sequencing (scRNAseq) datasets. We present confirmatory evidence for the presence of TMPRSS2, CD147, and GRP78 protein in vitro in airway epithelial cells and confirm broad in situ protein expression of CD147 and GRP78 in the respiratory mucosa. Our data suggest that the presence of a mechanism dynamically regulating ACE2 expression in human lung, perhaps in periods of SARS-CoV-2 infection and/or that alternate receptors for SARS-CoV-2 exist to facilitate initial host cell infection in lung tissue.

with chronic lung diseases (Figure 6) . A single healthy human sample contained one positive airway epithelial cell with additional positive staining in the peripheral lung in cells with type II alveolar epithelial cell morphology ( Figure 6Asecond row) . A representative image of a sample from a smoker with chronic obstructive pulmonary disease ( Figure 6Bsecond row) shows no ACE2 protein staining in the airway epithelium and a rare positive cell in subbasement membrane tissue. Quantification of positive pixel count for ACE2 staining normalized to total tissue pixel count revealed no differences between healthy non-smokers and tobacco smokers (Supplement Figure 4) ACE2 protein expression is rare in human lung tissue and found in select cells in both healthy individuals and those with chronic lung diseases. TMPRSS2 and CD147 protein expression are potentiated in individuals with a history of tobacco smoking and a diagnosis of COPD.

The global COVID-19 pandemic that emerged in late 2019 is caused by SARS-CoV-2. The possible host receptor(s) for SARS-CoV-2 have not been exhaustively surveyed in human lung tissue at the gene and protein level. Understanding the expression levels and localization of candidate SARS-CoV-2 receptors in host lung tissue may provide insights into therapeutic interventions that might reduce disease spread, viral replication, or disease pathology. To address this knowledge gap, we performed gene expression, proteomic profiling at the tissue and cell level, and in situ protein profiling of candidate receptors in human airway epithelial cells and lung tissue (summarized in Figure 7) . Collectively our data demonstrate rare ACE2 protein expression in human airway epithelial cells in vitro and in situ. Our protein expression data are consistent with low ACE2 promoter activity in a panel of lung epithelial cell samples and low ACE2 gene expression in bronchial epithelial cells (microarray) and lung cells (scRNAseq). We present confirmatory evidence for the presence of TMPRSS2, CD147, and GRP78 protein in vitro in airway epithelial cells and confirm broad in situ protein expression of CD147 and GRP78 in the respiratory mucosa. Our data suggest that for ACE2 to be an integral receptor for SARS-CoV-2, mechanisms are likely to exist that dynamically regulate expression in human lung, perhaps in periods of SARS-CoV-2 infection [54] . It is also possible that alternate receptors for SARS-CoV-2 are important in initial host cell infection.

Using a curated microarray gene expression dataset generated from bronchial brushings of 504 healthy subjects that considers the limitations of merging multiple datasets from distinct sources, we observed that sex did not correlate with gene expression of any candidate host molecule involved in SARS-CoV-2 infection and that ACE2 and TMPRSS2 were the lowest expressed genes of interest examined. In one dataset, ACE2 gene expression modestly decreased with age, although protein level confirmation was not possible. The low level of ACE2 and TMPRSS2 gene expression in bulk bronchial epithelial cell gene expression samples suggests low levels of cells expressing both of these genes within this lung tissue. We confirm that tobacco smoking is associated with elevated ACE2 gene expression levels in bronchial epithelial cell samples [55] , although we were unable to confirm this with immunohistochemistry analysis of protein on human lung samples.

Advances in transcriptomics have enabled scRNAseq that has identified unique and rare cell types in human lung that may have importance in health and disease [56, 57] . scRNAseq provides an opportunity to look at transcriptional profiles in subsets of cell populations, which may isolate a cell signal from a bulk sample. We therefore utilized scRNAseq data from healthy Consortium based publicly available datasets represent another parallel approach to confirm our data. We have used the FANTOM5 dataset containing CAGE promoter activation data for 1,866 primary cells, cell lines, and tissue samples from humans [39] to examine the level of promoter activity for each candidate SARS-CoV-2-receptor genes. The FANTOM5 CAGE data provides an additional and complementary approach to quantifying gene expression since a given gene's shared promoter can yield multiple transcripts at different expression levels, as well as being partially independent of any given transcript's half-life in the cell. In general, the promoter activity of ACE2 in airway-related tissues is low or absent; only a single sample originating from an adult lung yields a normalized CAGE promoter expression level above one transcript per million, while expression was observed in gut cells, consistent with known patterns of ACE2 expression [60] . Consistent with the microarray data, CD147 promoter activity is elevated relative to ACE2 across airway-related cells and tissues, although the relatively low CTSL (cathepsin L1) promoter activity is incongruent with modest levels of gene expression.

The expression of genes does not always correlate with protein expression [49] . With this in mind we performed combination proteomic analyses with immunoblot analyses. For our immunoblots, we used the human Calu-3 adenocarcinoma cell line as this cell is susceptible and permissive to SARS-CoV-2 infection and expresses ACE2, an observation we confirm [24, 29] . We also used primary human airway epithelial cells and the bronchial epithelial cell line (HBEC-6KT). We performed immunoblots for ACE2 and TMPRSS2 as these have been highlighted as interacting with SARS-CoV-2, while we probed CD147 as recent pre-clinical and clinical studies have provided proof of concept as a candidate SARS-CoV-2 receptor [26, 27] .

Lastly, GRP78 was dominantly expressed throughout transcriptomic studies and was selected as a positive control as previous expression has been confirmed in human airway epithelial cells [61] . Cathepsin L was excluded from the present analysis due to low promoter activity (Figure 3) , while ADAM17 was excluded as the proposed function in coronavirus infections is via ACE2 [5, 24] , which was included in analysis. Immunoblot analysis with all antibodies revealing dominant bands of predicted molecular weight, with the anti-TMPRSS2 polyclonal antibody revealing additional minor bands in all cell samples examined. The identity of these other bands remains unclear and suggest downstream immunohistochemical analysis may be confounded by the specificity of this antibody. In contrast, antibodies for ACE2, CD147, and GRP78 were specific and could be used for immunohistochemistry without concerns of specificity. Interestingly, ACE2 protein could only be detected with a supersensitive ECL solution and only in Calu-3 cells, suggesting absent protein in primary human airway epithelial cell and the HBEC-6KT cell lines. Our data are consistent with previous immunoblots of primary human airway epithelial cells grown under submerged monolayer conditions using the same primary antibody, where ACE2 protein was absent, and only expressed under air-liquid interface culture conditions [62] . The observation that CD147 and GRP78 are also expressed in Calu-3 cells encourages further interrogation into these host proteins, as they may contribute to function of ACE2 and TMPRSS2 in SARS-CoV-2 binding and fusion in this cell type.

Collectively, the profiling of antibodies by immunoblot of airway epithelial cells revealed distinct band patterns demonstrative of antibody specificity for ACE2, CD147, and GRP78, and to a lesser extent for TMPRSS2. Immunohistochemical analysis has been performed for localization of ACE2 and TMPRSS2 in human lung [28, 30] . The observation of positive staining in human lung tissue for these proteins was not accompanied by companion immunoblot or complementary approaches to define the specificity of the antibody used [51] . In the absence of determination of antibody specificity, the historical data presented should be interpreted with caution. To address the issue of antibody specificity for immunohistochemical staining, we used the same antibodies we validated by immunoblot and confirmed findings using proteomics as an orthogonal, antibody independent, approach. We again focused on ACE2 and TMPRSS2 as these are candidate proteins important for SARS-CoV-2 infection of host cells. Our immunohistochemical staining patterns of ACE2 were consistent with transcriptional profiling and immunoblots with only 1 of 98 human samples demonstrating rare staining in the airway and alveolar epithelium. Positive ACE2 staining in heart tissues and areas of lung microvasculature suggest our staining protocol was successful. These results directly contrast those reported with antibodies that lacked validation for specificity [28, 30] . TMPRSS2 was expressed more frequently across all samples examined with variability in the airway epithelium, associated with history of smoking and/or chronic obstructive pulmonary disease status. In contrast, CD147 expression was observed in airway epithelium of all samples. Similar to TMPRSS2, elevated CD147 expression was associated with history of smoking and/or chronic obstructive pulmonary disease status, consistent with previous reports [50] . Our original GRP78 antibody selected for immunoblotting was not validated for immunohistochemistry. We therefore performed confirmatory GRP78 immunoblotting and immunohistochemistry with an additional antibody (HPA038845) and provide demonstration of expression in human airway epithelial cells in vitro and in situ. Our study has several limitations that have not already been addressed above. Our observation of differences in gene expression between upper and lower airways and along the airway tree were not corroborated at the protein level. It remains possible that entirely different protein expression profiles for the candidate molecules examined exist in the upper airway, presenting a different environment for SARS-CoV-2 interaction with the respiratory mucosa. data. We would like to thank all of the personal and professional support from the author's respective families (CLB, JMH, and HLH) and research institutes, and most importantly, the frontline healthcare workers during the COVID-19 pandemic. shown for samples related to lung, gut, heart, and prostate tissues (n=120). Dot sizes are proportional to promoter activity, depicted as log 10 -transformed normalized transcripts per million (TPM). Notably, ACE2 is either not expressed or expressed at low levels (less than 1 TPM in all but one sample) in the airway, including measurements from healthy (white rows) and cancerous cells (grey rows). 

10 20 30 40 50 60 

Dynamically modeling SARS and other newly emerging respiratory illnesses: past, present, and future

Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

Functional assessment of cell entry and receptor usage for SARS-CoV-2 and other lineage B betacoronaviruses. Nat. Microbiol

Structural basis for the recognition of the SARS-CoV-2 by full-length human ACE2

American Association for the Advancement of Science

Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus