Could we go on the offensive against emerging diseases?

Viruses like Ebola and the original SARS have highlighted the risks that emerging diseases pose to our modern, highly connected society. While the standard approach of isolating the infected and limiting the spread of the disease worked in those cases, it works slowly enough to make many people nervous. But the global spread of Zika and SARS-CoV-2 shows that these approaches have their limits, leaving us at risk.

Is there anything else we could do? A perspective by Scott Nuismer and James Bull of the University of Idaho suggests we now have the tools to go on the offensive against viruses before they transfer to humans. The proposal: treat animal hosts of threatening viruses with virus-based vaccines that can spread through wild populations. While there are a lot of details to work out here, the article lays out how we might determine if this could be a viable approach.

Threats and their hosts

There are a huge number of hosts that share virus with our species. These range from familiar threats, like the mammals that carry the rabies virus, to our agricultural species that have spanned flu pandemics, as well as newly emerging dangers, such as hantaviruses and coronaviruses, carried by mice and a variety of species, respectively. While there’s no real pattern to the species that transfer viruses to humans, there have been successful efforts to identify the hosts from which viruses originated. Nuismer and Bull highlight the PREDICT program, run by the US Agency for International Development, which identified nearly 1,000 previously uncharacterized viruses before the Trump administration terminated it in March.

Figuring out which of those viruses might pose a threat is not a simple matter. But for the time being, there are a large collection of viruses that we know circulate in animals and are a threat to humans, so there’s no shortage of potential targets. If we actually get a preemptive approach to work with them, we can start worrying about potential threats.

So how do you stop a virus that’s not even infecting us yet? The basic idea is simple: develop a vaccine and give it to the animals that carry the virus. The obvious challenge to this approach is delivering the vaccine to a wild animal population. Not only are these populations often widely dispersed and difficult to access, but many of the animals (such as mice, in the case of hantavirus) have pretty short lifespans.

The solution that Nuismer and Bull consider is to use a virus as the vaccine—specifically, a virus that can spread beyond the population given the initial dose. In other words, the vaccine will make copies of itself and ensure that the unvaccinated population has a chance to receive a dose. This basic idea has been explored using epidemiological models, and it would likely work, but it’s only received a single, limited test in an animal population so far.

Options and risks

The epidemiological models indicate that this spread can be fine-tuned based on the infectivity of the virus being used as a vaccine. With a high-enough infectivity, the vaccine should spread throughout any populations of hosts that aren’t sufficiently isolated. A weaker virus with a lower infectivity might spread once or twice after the initial inoculation before fizzling out. Depending on how well we know and can manipulate the virus being used as the vaccine, it might be possible to tune its properties to match the size and distribution of a population, as well as the ease with which we can deliver additional doses.

There are two options for doing this. The first would involve starting with the virus that we’re trying to vaccinate against and generating a weakened form, often termed “attenuated.” This approach has been used for some human vaccines. Unfortunately, there have been a number of instances where a weakened virus has re-evolved virulence while circulating in a population. If this were to happen with a virus that poses a threat to humans, it would be possible for our vaccination efforts to inadvertently expand the pool of animals that could transfer it to us. For that reason, Nuismer and Bull don’t consider this a viable option.

Two for one

The alternative is to do one of the things that is being tried with SARS-CoV-2: engineer a gene that encodes a protein from the virus being targeted into an innocuous virus that can spread through the population. Ideally, as the harmless virus infects new animals, the immune response they generate will target both the virus’s proteins and the one engineered into it, thus providing immunity to two viruses.

Engineering a different virus’s gene into a virus is the least challenging aspect of this approach. It’s completely dependent upon our ability to find or generate harmless viruses that infect the target species. If we do use a naturally occurring virus, then we run the risk of the targeted population having a pre-existing immunity to it. We may need to spend time tuning its infectivity to match our needs as well. And finally, after all that’s done, there’s a chance that the protein, which is superfluous to the virus, will end up being lost.

Of course, if we plan on reintroducing the vaccine regularly, then the loss of the protein won’t be a major factor. But for something like Ebola, where new outbreaks seem to originate in remote areas, this may be more of a challenge.

In the end, the authors recommend a set of basic guidelines: use something that is based on a harmless virus, make sure it’s species-specific, and make sure that it’s engineered to limit its spread once it’s put in a wild population.

What to try first

So far, the authors indicate that this method has been tried a grand total of once. A rabbit virus that had naturally evolved into a harmless form was engineered to carry the gene for a protein that would confer immunity to a more dangerous virus that also targeted rabbits. A group of rabbits were inoculated with this virus before being set loose on a small island. After some time to allow the virus to spread, a bunch of other rabbits were tested, and about half of them were found to have been infected. This suggests that the infectivity was high enough that the rabbit population could eventually hit herd immunity from a single release.

Nuismer and Bull think that’s a good model for testing the approach using a virus that targets humans. They suggest something like rabies, which has been intensely studied and has a number of known hosts. Again, they think that an island population is a good choice, as it will allow a detailed tracing of the vaccine’s spread through the animals. If that works out, we can start considering the method’s use in more widely dispersed populations.

So does it make sense to take the offensive and start a pre-emptive vaccination program in animals against viruses that might be a threat to us? The approach recommended here, which involves identifying harmless, species-specific viruses and then engineering them to be vaccines for a dangerous one, involves a significant amount of work. Safety testing in a controlled environment—with involuntary participants like bats—will add considerably to the effort involved. At some point, it’s going to become similar to the effort of designing a human vaccine instead.

But…

For something that poses a regular health risk, like rabies, all this effort may be worthwhile. But to just take a currently relevant example, there are a large collection of coronaviruses in bats alone—along with a very large collection of bat species—and bats aren’t the only species that has been the source of a coronavirus that’s gone on to infect humans. Most of these are probably harmless, and extensive work will be needed to determine which might pose a threat to us. Can we really expect to protect ourselves from everything relevant there?

It’s an intriguing idea, and once we have a better grip on the threats posed by emerging viruses, the method may prove to be a useful way of neutralizing them. For now, while we’ve certainly got the technology to do it, the number of targets that it makes sense to go after is small enough that this doesn’t seem likely to be widely useful.