Three thoughts on the novel coronavirus cell line study

Sometimes a paper is published and the real-world applicability of the study isn't easily concluded or communicated from the results. Yet despite that, these inferences spread among the media and can result in feelings of confusion, panic and dread when the public are faced with the prospect of a virus more pathogenic than the SARS coronavirus was.

This happened recently following the publication of a paper in the Journal of Infectious Diseases (Differential cell line susceptibility to the emerging novel human betacoronavirus 2c EMC/2012: implications on disease pathogenesis and clinical manifestation) It's OA too so go have a look. There's also a very good accompanying editorial outlining the issues with drawing clinical conclusions from these data. 

A number of news storys and tweets were communicated concluding that this virus is 'more deadly' than hCoV-SARS, which could only replicate in a few cell lines or does the study even provide evidence that the virus can replicate in many different tissue types? There was however a more muted story in CIDRAP. Can they really say that from their data?

Basically the Hong Kong group used the isolated novel coronavirus (hCoV-EMC) from Ron Fouchier's lab and infected a wide range of cell lines with one infectious virus particle per cell and measured production of viral RNA (I think the genomic positive sense strand) on day 0, 1 and 3 following infection as well as nucleoprotein protein expression as markers of replication and concluding from this viral tropism in a human person. From this they showed that the virus could replicate in nearly every cell line tested and could replicate their genome up to five logs (quite a lot).

MY THREE THOUGHTS:

1) The main issue with this paper is this: these cell lines, although originally human, are all immortalized cancer cell lines characterised by markedly different biological properties when compared to normal human cells of the same tissue/cell type. They can't be readily used a surrogates for normal human tissue/cell types. None were primary cells nor were any even from recently acquired tissue samples from biopsies etc. People have infected primary human airway epithelial cultures with hCoV-EMC - so this can be done successfully - , although it would be more difficult for other tissue types as these cultures haven't been developed. Some of these cell lines used may by chance lack key viral repressors of infection present in normal primary cells, which could skew results from cell culture infection experiments. Plus, a human tissue is not just a single cell type - they are composed of diverse kinds of cells that could together behave much, much differently than cell lines in culture. 

2) The pathogenesis and spread of virus relies on the complex interaction with the human immune system in a tissue specific manner. For example, the hCoV-EMC virus may never escape the human respiratory tract because tissue-resident immune cells and the innate immune system cripple virus replication before it can spread systemically in blood or lymph.

3) Virus spread and tropism also relies on physical cell-cell interactions. For example, measles and other paramyxoviruses gain access to diverse tissues in the human body including the brain and kidneys via infection of immune cells residing in near-by draining lymph nodes or those present in sites of primary replication like the lungs. If hCoV-EMC can't do this nor survive and persist in the blood stream/lymph then how is it go systemic?

All these processes can and should be modelled in some way in the lab but certainly not only through these basic cell culture infection experiments. And I should add that this study doesn't prevent others from doing so and encourage other groups across the world to look into this. The complex interactions of emerging viruses with all cell/tissues/biological processes should be investigated! However, that will require further work in more refined models or animal studies. 

One investigation that would prove extremely useful and answer these questions would be the pathological assessment of banked autopsy material from the fatal cases in the UK (this had been done in SARS). Assessment of the distribution of viral antigen could be used to infer virus tissue/cell tropism and point us in the direction of where and what the block or inhibitory factors act to limit virus transmission/severe pathogenesis like that seen with SARS. 

N.B - the idea for this post came from below.

Over twitter I took part in this brief discussion begun by Laurie Garret's tweeting of the link to the study:

I was probably over critical saying it was 'horribly flawed' - the study and science was OK (though see Matt Frieman's - who is a coronavirus group leader in the U.S - comments below) but it is the clinical conclusions that can be drawn that would be flawed if we took this as evidence that hCoV-EMC is more pathogenic than the SARS virus (it clearly isn't).

Then it was pointed out that there was an informative editorial accompanying the article: