Integrated analyses of single-cell atlases reveal age, gender, and smoking status associations with cell type-specific expression of mediators of SARS-CoV-2 viral entry and highlights inflammatory programs in putative target cells
bioinformatics cell biology inflammation virology
Authors: Muus + 20 further first authors et al.
Journal/ Pre-Print: bioRxiv
Tags: Bioinformatics, Modelling, Statistics
1. Nominate potential target cells of the virus by looking at cells expressing a combination of ACE2 and TMPRSS2 or CTSL in a large cross-tissue single-cell study. They found ACE2 and TMPRSS2 expression to be particularly enriched in specific cells of the ileum, liver, lung, nasal, eye, olfactory epithelium, heart, bladder, testis, prostate and kidney.
2. Meta-analysis study examining the relationship between specific cells types and covariates related to disease severity namely age, sex, and smoking. They found lower ACE2 expression in young paediatric samples, and upregulated ACE2 expression in airway epithelial cells in past and present smokers.
3. Further identified gene expression programs in ACE2+TMPRSS2+ cells in epithelial subsets from different tissues (lung, nasal, gut) and cell types (e.g. AT2 cells). The genes identified include genes that may mediate viral entry, and be involved in immune function, where the authors highlight IL6 and IL1R.
The authors assembled 107 single-cell/nucleus RNA-seq datasets from multiple consortia to compute expression patterns of ACE2, TMPRSS2 and CTSL. They find ACE2+TMPRSS2+ cells in 10 tissues, and describe pathways and genes (including multiple PCSK-family proteases) enriched in these cells. Using a (slightly non-standard) meta-analysis methodology, they test association between these genes’ expression and sex, age and smoking status (N=164) in different cell-types. Results were heterogeneous across cell-types and genes, but ACE2 levels increased with age in airway epithelial and AT2 cells, and with smoking in airway epithelial cells. The authors replicate many of these results in mice of different age, sex and nicotine smoke exposure
Impact for SARS-CoV2/COVID19 research efforts
Understand the virology and/or cell biology of SARS-CoV2/COVID19
Clinical symptoms and pathogenesis of SARS-Cov2/COVID19
· In silico study / bioinformatics study
Strengths and limitations of the paper
· First large meta-analysis of single-cell datasets to provide a comprehensive overview of the expression patterns of ACE2 receptor and accessory proteases genes such as TMPRSS2 across tissues. They further analysed the effect of covariates (age, sex, smoking) on co-expression of these genes in different cell types.
Standing in the field:
· Supports similar results reported in other pre-prints and papers
· Yes, they used standard single-cell analysis tools and methods
· They assembled a very large dataset from human samples across multiple tissues which should provide sufficient power to assess co-variates such as sex, age and smoking in cell types at single-cell resolution.
· Association testing in their Poisson model suffers from inflation due to pseudo-replication (which they attempt to correct for by repeating analyses in pseudobulk).
Viral model used:
· Healthy human and mouse tissue samples from published and unpublished sources
· Helps assess virus tropism to further understand clinical manifestations
· Only validation of gene expression of ACE2, TMPRSS2 and CTSL was shown but no further functional studies included.
· They discuss the limitation that cells have been treated as independent observations in their Poisson model which can lead to inflated p-values when looking at the significance of association with covariates – age, sex, smoking.