Limits to simulating virus evolution in the lab
"On reading," by Simon Wain-Hobson, is a weekly discussion of scientific papers and news articles around gain of function research in virology.
Since January 2024, Dr. Wain-Hobson has written weekly essays for Biosafety Now discussing risky research in virology. You can read his entire series here.
On wondering about numerous lightweight papers on virus evolution and cross species adaption in the lab.
Virologists the world over are a pragmatic bunch. For years they have used what are called established cell lines to grow viruses, with an obvious bias to using those that allow plentiful production of virus. Vero cells, first established from African green monkey kidney tissue in 1962 spring to mind. More recently, they were widely used to isolate SARS-CoV-2 viruses.
There is a reason for this. Vero cells are unable to mount an efficient interferon response – a very powerful arm of what is called the innate immune response - which would clobber considerably virus growth. In the early days of virology growing up enough virus for experiments was hugely important, indeed #1, so it made sense to use Vero cells. However, when it comes to disease as well as the subtleties of virus evolution and adaptation to other hosts, they introduce a bias by allowing outgrowth of viruses best adapted to this unnatural cell line.
There is a large literature showing that when you isolate a bacterium, it is often accompanied by the loss of virulence factors. For rapidly mutating RNA viruses, you often isolate a minor form. On reading’s group showed this for HIV back in 1989. The reason is simple. The conditions used to culture the virus or bacterium in the lab are not the same as inside the body which is extraordinarily hostile to viruses. This results in the selection of a virus or bacterium better adapted to growth in the lab.
Many of the cell lines used to grow viruses were derived from cancers. Now cancer is often described as a disease of the genome which can carry tens of thousands of extra mutations compared to non-cancerous cells from the same patient. Sometimes there are more than a million mutations which are always accompanied by a bewildering array of DNA rearrangements and other chromosomal lesions. This is nowhere more evident than for the most renowned human cell line of them all, HeLa, a cervical carcinoma cell line transformed by human papillomavirus 18. Its genome has been sequenced. It contains thousands of mutations of which 71 were predicted to impact the function of 66 proteins, thousands of small insertions and deletions, as well as extensive shattering on four chromosomes. On top of which there are approximately three copies of each chromosome per cell. HeLa is anything but typical.
The human A549 cell line extensively used by flu researchers was established in 1972 from a patient with non-small cell lung cancer. Each A549 cell contains on average 62 chromosomes unlike noncancerous cells which harbor 46. Add to this, 25% of the A549 chromosomes have large scale structural abnormalities and a multitude of smaller ones.
And while A549 is a very useful cell line to think it can remotely help recapitulate the evolution of flu viruses is wide of the mark.
Another glaring shortcoming of growing microbes in the lab is that there is no adaptive immune response – antibodies and white blood cells - so it’s like isolating a virus from a patient without an immune system. And while those viral isolates can tell us a great deal, they can’t tell us everything. Certainly not about the adaptation to a new host.
Now let’s look at lab conditions for culturing a microbe. They represent an artificial mixture of nutrients that ‘work’ aka allows efficient microbial multiplication. It’s worth pointing out that most bacteria cannot be cultured in the lab - a long-standing issue – so you can understand the delight of a researcher who eventually finds conditions that allow them to grow the bacterium. Whether the conditions are physiological is another issue.
Next, there is an abundance of oxygen when growing human cells in the lab – around 20%. In arterial blood it ranges from 13-14%, yet only 0.5-4.5% in lymphoid organs, the sites of HIV replication. It is possible to use oxygen chambers with much less oxygen, however they ‘complicate’ lab work, cost more and are not widely used.
Yet for cells life under hypoxia is wildly different with some key proteins like HIF-1a, to mention but one, turning on or off hundreds of important genes. The importance of hypoxia was highlighted by the Nobel Prize in Medicine in 2019.
Next, temperature. Experiments are invariably performed at 37°C – the human core temperature even though that of the nasopharynx, where flu viruses grow, is a little lower. If you look at any number of papers on the culture of bird flu viruses 37°C is the norm. The problem here is that the body temperatures of birds are closer to 41°C. For humans a core temperature of 35°C is life threatening so 4°C of hypothermia for birds is not normal – the autumn understatement! The core blood temperature of cows is around 38.5°C… You get it; body temperatures vary yet few factor them into lab work.
The fact is that viruses can exploit such differences and a few degrees either way is enough to result in selection of a strain better adaptation to the new environment. Proof of this is the detection of mutations in the bird flu PB2 enzyme upon adaptation to growth in ferrets as per the Fouchier and Kawaoka experiments, although this was known before.
An increased awareness of cell culture constraints encouraged HIV researchers to work with organoid cultures, for example pieces of lymphoid tissue which brings isolation and growth closer to a more physiological setting.
Ron Fouchier has been doing this with a number of respiratory viruses, notably human metapneumoviruses (HMPV): the isolation efficiency of serotype A HMPV was comparable in both models, while isolation of serotype B viruses was profoundly more efficient in the ODB [organoid-derived bronchial] cultures than in Vero-118 cells, suggesting that primary cultures expressing ciliated cells should be considered as a superior isolation method for HMPV from clinical specimens.
Umpteen other papers from 2018 onwards have shown that organoid cultures are physiologically better models for influenza replication and the genesis of immune responses. Accordingly for studies of cross species transmission such models should be of more use compared to those using established cell lines like Vero or MDCK cells. Yet they still represent a minority of research papers.
Virologists’ ingenuity can be remarkable. It is common knowledge knows that HIV-1 jumped over from chimpanzees. Adaptation of several SIVchimp viruses to the humans has been studied in mice reconstituted with a human immune system. A series of papers using several strains resulted in the outgrowth of viruses that grew much better and killed the all-important CD4 T cells much faster. Twelve protein changing mutations emerged. You can’t get closer than that!
Finally, there is chance. Remember the paper where feline influenza H7N2 virus was adapted to human A549 cell line (Lessons not learned)? In three parallel adaptations, they came up with three genetically distinct viruses. Nobody knows why this occurred although it could be related to the amount of virus used in the experiments. The lower the amount the less genetic complexity in the inoculating virus which can give rise to so called founder effects, aka chance.
By the by, this feline flu virus was grown on the canine MDCK cell line at 37°C even though the cat core body temperature is between 37.7-39.8°C while the evolutionary split between dogs and cats goes back around 50 million years. But above all, the MDCK line is unable to express the interferon induced Mx proteins which are very important for controlling flu virus growth.
Despite these limitations, a lot of important work gets done and virology advances. It’s just that if you get into the weeds of virus adaptation and want to simulate cross species transmission of a bird flu virus such as H7N1 between chickens and humans why not try and make the experiments as physiologically relevant as possible?
Of course, ethically you cannot infect humans. However, you can make an effort to get as close as possible, always remembering that pandemic prediction is not possible.
The worst offenders are the computer jocks studying viral genomes writing papers on cross-species adaptation when some confuse serology and PCR studies (Cancel Virology). That or informing millions of people back in early 2020 that the COVID-19 virus jumped from an animal when there was no data for or against. They rarely do lab experiments to test their hypotheses expecting others to do so.
It would be good if far more lab virologists switched to studying virus replication in vivo, quantifying the tempo and magnitude of the titanic struggle between a virus and the immune system. Relatively few do because it is multidisciplinary requiring knowledge of physiology, virology, immunology and mathematical modeling. Its labor intensive, more costly and takes far more time to complete compared to quick tissue culture experiments.
All the above issues, concerns and limitations were obvious to the informed virologist before Drs. Fouchier and Kawaoka pulled the pin on the dangerous GOF research grenade. A deflagration didn’t occur because the NIH ‘bull’ published in the Washington Post (Chilled virology) closed down discussion. The importance of the NIH in biomedical funding resulted in a global omertà. The vast majority of virologists kept mum. Still do.
What we do know is that the Fouchier and Kawaoka papers were widely read and downloaded and that must include hostile nations free of charge, courtesy of USG. And while dangerous GOF research is now kaput in the US, the virologists and many of their allies are not all happy (An abundance of caution). That said something isn’t quite right. The Europeans haven’t seized the occasion to discuss this absurdity and to get to grips with Dual Use Research of Concern.
It’s a sad, sad situation and it’s getting more and more absurd.
Conclusions
Virologists, funders and publishers, the next time you come up against a manuscript or grant proposal attempting to tackle viral emergence, insist that the researchers make an effort to get as close to a physiological setting as possible with viruses that haven’t been extensively adapted to lab conditions. Don’t fall for tissue culture experiments.
Aside 1
Influenza A virus infection can be exhausting but short lived. In a patient without an immune system, it is a chronic infection with mutants being shed constantly. Similarly, in patients with hypogammaglobulinemia – unable to produce antibodies – and vaccinated with the live attenuated poliovirus vaccine constantly secrete the virus. Or just look at AIDS before modern therapy – horrible and multiple infections by all sorts of microbes running rogue. The impact of the adaptive immune system is colossal and another reason to be careful in extrapolating from what occurs in lab experiments. Fortunately, the development of antivirals is by and large unaffected by lab adapted viruses.
Aside 2
If you’re interested in the body temperature of bats take a look at this 2021 paper.
Aside 3
There is an interesting twist to the SARS-CoV-2 / Vero cell line story. The selective advantage for growth in Vero E6 cells is due to increased cleavage efficiency by cathepsins at the mutated S1/S2 site… The entry of Vero E6-adapted virus into human cells is defective because the mutated spike variants are poorly processed by furin or TMPRSS2. Minor subpopulation that lack the furin cleavage motif in the spike protein rapidly become dominant upon passaging through Vero E6 cells. Fascinating. VeroE6 is a derivative of the original Vero cell line. Cathepepsin and TMPRSS2 are enzymes capable of digesting specific sites in proteins.
As a concerned citizen, I find it highly suspicious that the European political elites don't speak about risky pathogen research, dual use or bioweapon research.