Alexander Robertson

If you’re interested in antiviral design, viral evolution, eco-evolutionary feedbacks, ploidy, etc. etc. etc., drop me a line! Also, check out this nice press release that UW news wrote about the paper: newsroom.uw.edu/news-release...

A cyclic plot depicting how viral density determines the degree of intracellular interactions, which in turn shapes realized phenotypes, influencing absolute fitness, which then feeds back into viral density in the next generation.

Big picture: the feedback loop between viral density leading to social interaction, which leads to realized phenotype, which alters fitness, leading to new viral densities, should be considered when designing optimal treatments.

Perhaps the most interesting thing to me is that this parallels recent work which frames viral infection in the terms of ploidy. What are the dis/advantages of having a “non-fixed ploidy”? How can broadening our understanding of population genetics lead to better pathogen treatments?

These ideas, which seem counterintuitive, have actually been described for therapies that treat cancers and bacteria. Similar ideas have also been proposed for the use of therapeutic interfering particles (TIPs).

This isn’t a clinical recommendation (!!!) but it highlights a very important consideration for antivirals that depend on virus-virus interactions: they must balance killing viruses 𝘯𝘰𝘸 with preserving the interactions that keep resistance low 𝘭𝘢𝘵𝘦𝘳.

We simulated clinical trials of individuals treated with drugs 1×, 10×, 100× and 1,000× weaker than pocapavir. Reducing drug potency delayed mean clearance time measured by days post infection but had a non-monotonic effect on the sum viral load over the course of the infection (n = 100 for each group). In both plots, dots represent individual simulations and are colored based on the frequency of resistance in the population over the course of the infection (scale in right panel)

When we simulated clinical trials using weaker drugs, they cleared later on average (mostly due to fewer early clearers). But the overall viral load was lower in the x100 group, and resistance frequency was likewise reduced.

Counterintuitively, our model suggests that a weaker treatment may maintain enough coinfection to keep resistant viruses wrapped in susceptible capsids, while still maintaining low viral load.

At low MOI, viruses singly infect cells, and rare resistant genomes are encapsidated by phenotypically resistant capsids, enabling selection for resistance.

The reason for this was due to the effectiveness of the treatment. As the treatment works, viral density collapses, coinfection plummets, and genotypes become linked to their phenotype. Resistant genomes are wrapped in their own proteins, and the drug now selects for them.

Clearance dates from the observed clinical trial and one simulated clinical trial of n = 93 viral populations treated with pocapavir. The dashed line indicates the date of the earliest placebo clearance. As in the clinical trial, we observed a population of early clearers (who cleared before the earliest placebo clearance date) and late clearers, who cleared on par with placebo. Late and early clearers both experienced drops in resistant and susceptible viral population size with diverging outcomes following the population bottleneck. Throughout the figure, resistant variants are illustrated in red and susceptible variants are illustrated in blue.

Adding a simple immune clearance step to our model and running out over multiple generations lets the model reproduce the pattern observed in the clinical trial, where late clearers predominantly had resistant infections.

At high density, viruses have a much higher rate of coinfection. Rare resistant genomes get packaged into capsids with lots of susceptible proteins. This means that the drug works well and resistance is suppressed.

Resistant viral yield under different intensities of susceptible virus coinfection shows a density-dependent effect in vitro (again, compared to Tanner et al.'s 2014 paper mentioned above) and in silico under our model.

Our model captures the dynamics observed in cell culture, where coinfection with larger susceptible virus populations reduces resistant viral yield. We realized, however, that the MOI was quite large, and this was only a single round of passaging.

A cartoon schematic of the eco-evolutionary model that we developed. We simulate intracellular poliovirus dynamics in four stages: (1) viral entry into host cells, (2) genome replication with mutation, (3) production of capsid subunits and (4) assembly and packaging of progeny virions. We then apply pocapavir.

So what gives? To investigate this discrepancy, we built an eco-evolutionary model for poliovirus treated by pocapavir and found that 𝐚 𝐬𝐢𝐧𝐠𝐥𝐞 𝐦𝐨𝐝𝐞𝐥 𝐜𝐨𝐮𝐥𝐝 𝐞𝐱𝐩𝐥𝐚𝐢𝐧 𝐛𝐨𝐭𝐡 𝐨𝐮𝐭𝐜𝐨𝐦𝐞𝐬.

(Link to Collett et al.'s 2017 clinical trial here: doi.org/10.1093/infd...)

In the clinical trial reported by Collett et al., pocapavir failed to significantly reduce time to infection clearance compared with a placebo in three of four matched groups of participants administered the live attenuated poliovirus vaccine, and resistance was enriched in the pocapavir group. Points represent the clearance dates of individual trial participants and are colored by resistance status (resistance in red, susceptible in blue), and grey boxes indicate dates that were not sampled during the trial (DPI, days post infection).

However, clinical trials with pocapavir had mixed results. In 3/4 placebo-matched groups, pocapavir did not significantly reduce clearance time in people infected with OPV. Additionally, resistance evolved in nearly half of the experimental group, highly enriched compared to placebo.

(Link to Tanner et al.'s 2014 fascinating study here: elifesciences.org/articles/03830)

In vitro experiments by Tanner et al. demonstrated that coinfection of drug-resistant (res.) and susceptible (sus.) poliovirus strains suppresses the yield of resistant virus under pocapavir treatment. Link to their study: https://elifesciences.org/articles/03830

A previous study by Tanner et al. showed the power of this phenomenon. They studied pocapavir, a capsid inhibitor that targets poliovirus. In coinfected cells, increasing amounts of susceptible viruses decreased resistant virus output, suggesting that resistance should be hard to evolve in PV.

This leads to an interesting phenomenon that can be exploited by capsid inhibitors: resistant genomes can be wrapped with susceptible capsid subunits when resistant/susceptible genomes coinfect the same cell. This can dramatically mute selection for resistance.

A two-part comparison of how resistance evolves in bacteria versus viruses. The top panels show bacteria: rare resistant cells (red) become common after treatment because each cell’s genes match its traits. The bottom panels show viruses: resistant and susceptible genomes (red and blue) can mix inside the same infected cell, creating viral progeny with "chimeric" capsids. Mixed particles carrying resistant genomes may still get blocked by the drug (yellow triangles), even if their capsid has some number of resistant subunits.

First, some background. Most organisms have a direct link between their genotype and their phenotype. However, since viruses replicate inside of cells, different genotypes can co-occupy a cell and “share” their proteins with each other.

#socialviruses #evosky #virosky 🧪

Intracellular interactions shape antiviral resistance outcomes in poliovirus via eco-evolutionary feedback - Nature Ecology & Evolution A model of intrahost poliovirus replication shows that, after several rounds of replication, pocapavir, a poliovirus capsid inhibitor, collapses viral density, preventing intracellular interactions th...

My first lead author paper is out with Ben Kerr and @alisonfeder.bsky.social! We found that making an antiviral too strong can sometimes make resistance easier to evolve. This has implications for how we design drugs, choose doses, and think about viral evolution in the face of treatment. (1/n)