New paper shows that smoking cigs increases expression of the #SARSCoV2 receptor ACE2. The idea here is more ACE2, more infection, worse #COVID19.

This is a faulty assumption.
https://www.atsjournals.org/doi/abs/10.1164/rccm.202003-0693LE#.XqW1P4Cg3FM.twitter

More receptor does not automatically equal more virus. Yes, smoking predisposes a person to worse #COVID19, but that’s not necessarily because more ACE2 means more #SARSCoV2. Pathogenesis (the process by which viruses cause disease) is not determined solely by receptor expression

Virus titer isn’t even solely determined by receptor expression. More ACE2 on the surface of cells doesn’t necessarily mean those cells will produce more virus, and more virus doesn’t guarantee more disease.

This study analyzed transcriptomic datasets in public databases in a meta-analysis, or an analysis that mines datasets collected for another purpose to look at ACE2.

The trouble with meta-analyses is that it’s tricky to compare them. The studies they included looked at different tissues, recruited subjects with different criteria, and used different transcriptomic platforms (RNAseq, microarrays) to measure ACE2, furin, and TMPRSS2 expression.

Also the subjects in each study self-reported smoking habits, so it’s likely tobacco usage varies dramatically. This matters a lot, since being a former smoker could result in different gene expression from a current occasional smoker from a two-pack-a-day smoker.

And what does “gene expression” mean here anyway? Well, transcriptomics measures mRNA, the instructions that are made to build ACE2 protein when a cell makes more ACE2. More mRNA does not always correspond to similarly increased amounts of protein.

For #SARSCoV2, the virus has to bind ACE2 to enter a cell and then undergo priming by furin or TMPRSS2. These all are carried out by ACE2/furin/TMPRSS2 proteins, not mRNA. So with these data, we can’t conclude that increases in mRNA necessarily mean increases in protein.

They also did single cell RNAseq (scRNAseq) on 12 patients, 6 smokers, 6 not. scRNAseq lets you look at RNA in different types of cells. They concluded that smokers vs non have different types of airway cells with more or less ACE2.

I mean, maybe. There were few study subjects and based on figure 2, there seems to be different numbers of goblet vs club cells, but it’s hard to say if this matters since it was only in 12 patients. Could have been sampling or normal inter-individual variation.

And those of you more familiar with my work may be saying “but you use transcriptomics like ALL THE TIME”, yup, I do love me some RNAseq data. But for host response analysis I look at how gene expression reveals what functions are being turned on or off. Total protein amount...

...does not really matter. I don’t care how much interferon protein is there so long as I can see evidence of the downstream signaling it triggers and the time scale on which this occurs. The dynamics of pathway activation are different from just more mRNA = more protein.

But this study isn’t looking at ACE2 signaling or the effect of smoking on the renin-angiotensin axis. It’s just using transcriptomic data to make conclusions about protein expression. I’m not buying it. Immunohistochemistry of airway tissues or it didn’t happen!

Also this study fails to account for other variables and doesn’t look at ACE2 in the context of actual #COVID19! It’s all well & good to study smoking and ACE2, but if you’re trying to make a mechanistic connection between smoking, ACE2, & COVID, you gotta look at the COVID part.

It’s very easy, when testing a simple hypothesis using a lot of public data, to assign statistical significance to observed differences. It’s much harder to actually show significant relationships with disease across large, diverse populations of people.

Lung disease associated with smoking (asthma, COPD) and #COVID19 are both complex diseases that likely can’t be explained by the presence or absence of one single molecule. Time, space, genetics, and environment all likely play major roles.

Smoking is definitely a #COVID19 risk factor, but this paper doesn’t provide any useful info about why. Just looking at ACE2 expression in some random microarray and RNAseq datasets in GEO doesn’t really provide any new insight into #SARSCoV2 infectivity or pathogenesis.