SARS-CoV-2 receptor ACE2 and TMPRSS2 are predominantly expressed in a transient secretory cell type in subsegmental bronchial branches

Soeren Lukassen; Robert Lorenz Chua; Timo B. Trefzer; Nicolas Carlos Kahn; Marc Andre Schneider; Thomas Muley; Hauke Winter; Michael A. Meister; Carmen Veith; Agnes W. Boots; Bianca P. Hennig; Michael Kreuter; Christian Conrad; Roland Eils

doi:10.15252/embj.20105114

The EMBO JournalApr 14, 2020GenomicsPeer-reviewed journal article

SARS-CoV-2 receptor ACE2 and TMPRSS2 are predominantly expressed in a transient secretory cell type in subsegmental bronchial branches

Soeren Lukassen iD, Robert Lorenz Chua iD, Timo B. Trefzer iD, Nicolas Carlos Kahn iD, Marc Andre Schneider iD, Thomas Muley iD, Hauke Winter iD, Michael A. Meister iD, Carmen Veith, Agnes W. Boots iD, Bianca P. Hennig iD, Michael Kreuter iD, Christian Conrad iD, Roland Eils iD

PDF DOI Preprint Data

In-depth full-text digest available · Jump to digest ↓

Abstract

The SARS-CoV-2 pandemic affecting the human respiratory system severely challenges public health and urgently demands for increasing our understanding of COVID-19 pathogenesis, especially host factors facilitating virus infection and replication. SARS-CoV-2 was reported to enter cells via binding to ACE2, followed by its priming by TMPRSS2. Here, we investigate ACE2 and TMPRSS2 expression levels and their distribution across cell types in lung tissue (twelve donors, 39,778 cells) and in cells derived from subsegmental bronchial branches (four donors, 17,521 cells) by single nuclei and single cell RNA sequencing, respectively. While TMPRSS2 is expressed in both tissues, in the subsegmental bronchial branches ACE2 is predominantly expressed in a transient secretory cell type. Interestingly, these transiently differentiating cells show an enrichment for pathways related to RHO GTPase function and viral processes suggesting increased vulnerability for SARS-CoV-2 infection. Our data provide a rich resource for future investigations of COVID-19 infection and pathogenesis.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (30 Mar 2020) · CC BY 4.0

Single-cell profiling of lung and bronchial cells places ACE2 mainly in a transient secretory cell state of cultured bronchial epithelium, between goblet and ciliated cells, co-expressing TMPRSS2.

The airway cell types carrying the SARS-CoV-2 entry factors ACE2 and TMPRSS2 were poorly resolved. This work supplies an annotated single-cell reference of lung tissue and subsegmental bronchial epithelial cells to localize them.

Key results

ShownIn cultured bronchial epithelial cells (HBECs), ACE2 is strongest in a secretory subset (secretory3), which also shows high TMPRSS2. ACE2+/TMPRSS2+ double-positive cells are enriched across the HBEC dataset.
HBEC dataset: 17,451 cells from four donors. ACE2+/TMPRSS2+ double-positive enrichment 1.6 fold, p=0.002 (hypergeometric). About 40% of ACE2+ transient secretory cells co-express TMPRSS2. · ACE2 and its co-factor TMPRSS2 are expressed in…; DISCUSSION
SuggestedPseudotime places secretory3 cells between club/goblet and ciliated cells. Their marker genes are enriched for RHO GTPase pathways, seen in no other cell type, and not in any whole-lung cell type. Increased infection susceptibility is inferred, not tested.
Monocle3 pseudotime on HBECs; marker-gene overlap with goblet and ciliated markers; pathway enrichment via Metascape (Benjamini-Hochberg adjusted). No infection experiments were performed. · Transient secretory cells display a transient…; Pre-processing and data analysis; Statistical analyses
ShownIn lung tissue, ACE2 is very low in every cell type. AT2 cells show the highest ACE2 and a TMPRSS2 specificity, and TMPRSS2 is expressed much more strongly than ACE2.
snRNA-seq of 39,778 nuclei from twelve lung adenocarcinoma patients (non-tumor tissue). Median ACE2 below five counts per million, so reads were aggregated into pseudo-bulks per patient and cell type. · ACE2 and its co-factor TMPRSS2 are expressed in…
SuggestedFURIN is co-expressed with ACE2 and TMPRSS2 across cell types. FURIN priming could potentially make more cells permissive, but the triple-positive enrichment in HBECs was not significant.
Authors estimate 25% more cells vulnerable with FURIN than with TMPRSS2 alone. Transient secretory cells expressing ACE2 plus either or both proteases reach 50% (vs about 40% for TMPRSS2 alone). ACE2+/TMPRSS2+/FURIN+ enrichment in HBECs did not reach significance. · FURIN expressing cells overlap with…; DISCUSSION
SuggestedNo sex or age effect on ACE2 was seen within individual cell types. Pseudo-bulk lung data showed trends: age in females (p=0.065) and higher ACE2 in males aged 40 to 50 (p=0.048). Smoking could not be separated from age.
Female lung samples: ACE2 vs age R^2=0.40, p=0.065. Ages 40 to 50: higher ACE2 in three males vs five females, p=0.048. Lung cohort: three male and nine female donors; HBEC cohort: three male and one female. · Sex, age and history of smoking in correlation…

Methods, data and reproducibility

Single nuclei RNA-seq of frozen non-tumor lung tissue and single cell RNA-seq of air-liquid-interface cultures of bronchial epithelial cells, analyzed with Seurat integration, Monocle3 pseudotime, pseudo-bulk aggregation and hypergeometric overlap tests.

Snap-frozen non-tumor lung tissue from lung adenocarcinoma patients, snRNA-seq (10x Chromium, HiSeq4000) · Homo sapiens · n = 12 donors, 39,778 nuclei
Air-liquid-interface cultures of primary bronchial epithelial cells from subsegmental airways, scRNA-seq (10x Chromium) · Homo sapiens · n = 4 donors, 17,451 cells (abstract states 17,521)
Processed count and metadata tables for both datasets (raw reads under controlled access) · Homo sapiens · doi:10.6084/m9.figshare.11981034.v1

Methods & tools: CellRanger 2.1.1, Seurat 3.0.0 (FindIntegrationAnchors, IntegrateData), Monocle3, Metascape (KEGG, GO, Reactome, Corum), 10x reference genome hg19 1.2.0, Wilcoxon rank sum test, Hypergeometric test, Benjamini-Hochberg correction, scipy 0.14.1, statsmodels 0.9.0, 10x Genomics Chromium v2 chemistry
Data & accessions: https://doi.org/10.6084/m9.figshare.11981034.v1 https://data.mendeley.com/datasets/7r2cwbw44m/1
Reagents & environment: CellRanger 2.1.1; Seurat 3.0.0; Python 3.7.1 with scipy 0.14.1 and statsmodels 0.9.0; R; hg19 reference 1.2.0.
Reproducibility gaps: Raw sequencing data under controlled access (accession pending); samples only on request with a material transfer agreement; no analysis code repository stated.

Limitations (5)

Small cohorts (twelve lung, four HBEC donors) leave the data underpowered for sex, age and smoking effects on ACE2; significance is at best weak. · stated by the authors
Lung (snRNA-seq) and HBEC (scRNA-seq) data come from different techniques and sample types, so absolute expression should not be compared directly; only relative patterns are comparable. · stated by the authors
Donors had no SARS-CoV-2 infection history, and smoking prevalence differed from COVID-19 cohorts (25% lung cohort, 100% HBEC cohort). · stated by the authors
Susceptibility of transient secretory cells and a FURIN role are inferred from transcript co-expression and pathway enrichment only, with no protein or infection validation. · noted in this summary
The bronchial findings come from cultured, differentiated cells after cryopreservation, so the transient secretory state may differ from native airway epithelium; ACE2 detection is sparse. · noted in this summary