When the SARS-CoV-2 began circulating around the globe last year, researchers familiar with the concept of challenge trials considered the idea of deliberately infecting humans with the virus in a controlled manner to learn more about this novel pathogen. Now, two such trials are underway in the UK, one aimed at determining the best dose for causing infection but not making volunteers too sick, and another at understanding their immune response to the virus.
The Scientist spoke with Helen McShane, a vaccine researcher at the University of Oxford who is heading up one of the trials, about the logistics, hurdles, and value of this research.
The Scientist: When and why did you first begin thinking about conducting challenge trials for COVID-19? And what were some of the obstacles you faced in getting them off the ground?
Helen McShane: I’ve spent twenty years working on tuberculosis and TB vaccine development. And for the last five years, I’ve been working on developing a controlled human infection model for TB. I work at Oxford, where we have a very well-established program using the malaria challenge model to test candidate malaria vaccines. So I’m very familiar with and experienced with challenge models per se. When a new pathogen emerges, it makes sense in that context to think about whether we can develop a challenge model, knowing the utility of these models to complement knowledge gained from other sources—from natural history studies, from animal studies, from in vitro studies.
What were the challenges? Well . . . a year ago, fifteen months ago, we didn’t know very much about SARS-CoV-2, and it’s difficult to think about establishing a human challenge model when you don’t really know much about what happens in a natural infection. I think we’ve come a long way and learned huge amounts, and by the time that we and Imperial College—who are also conducting a parallel study, these studies are very complementary—had started our challenge studies, we actually knew quite a lot about SARS-CoV-2 infection in humans. And we know that for the overwhelming majority of young, fit, healthy people who become infected, they have a very mild illness. So that’s when you start to think, ‘Okay, well, if we can identify a [subgroup of people for whom] this is a very mild or asymptomatic infection, then that’s the group that we could potentially use in a human challenge model.’
TS: How did you develop the actual challenge? Which virus variant did you start with? And what changes did you make, if any?
HM: The viral strain that we and Imperial College are using is the Wuhan strain. And I think that’s important as a strain to use for the first challenge studies because it is the strain that we know the most about clinically. . . . Most people in the world who’ve been infected with COVID [to date] were infected with the Wuhan strain. So that’s what we started with. There are discussions underway . . . about manufacturing a second strain, which might be one of the new variants, but we will first get some robust data using the Wuhan strain.
TS: Did you make any changes? How did you determine your dose?
HM: We didn’t make any changes at all; this is actually the unaltered virus that has been manufactured to . . . good manufacturing practice conditions to make it safe—that’s the highest standard for manufacture for things to be tested in humans. In terms of the dose, we and Imperial are both conducting a dose-escalation study. That’s the first part of the challenge model . . . to establish the best dose. Imperial College are infecting seronegative people, so people who haven’t previously had SARS-CoV-2 infection. We are, in contrast, deliberately infecting people who are seropositive, i.e., people who have a documented previous SARS-CoV-2 infection.
TS: What are the differing goals of those two studies—between infecting seronegative and seropositive individuals?
Because we know we are exposing everyone in a very controlled way, if people are not infected after that controlled exposure, we know they must have protective immunity.
HM: First of all, [we’re] aiming to identify the right dose. Our study, which is infecting people who are seropositive, is very specifically aiming at defining protective immunity. We’re interested in establishing a challenge dose that is safe, but then ideally [a dose that allows] us to reinfect some people who have previously been infected with SARS-CoV-2. We can then interrogate the baseline immune response in all of our subjects and determine, what is it about the immune response in those subjects who cannot be reinfected, compared with those who can? . . . We may see, for example, that people who have a level of neutralizing antibodies above a certain point cannot be reinfected. That then tells us [about] the immune response we want to induce with a new vaccine. And that then feeds very directly into vaccine design, but more importantly, vaccine testing. So that’s what our model is for. The model that Imperial are developing, where they’re looking at people who are naive, is developing a human challenge model where you can actually test vaccine efficacy using this challenge model. So slightly different, but very complementary aims.
TS: Tell me a little bit more about the logistics of these trials. So these individuals have to come in and be quarantined for some period of time during the study?
HM: They do. Subjects are admitted two days before the inoculation, where we do the last round of screening tests. And then after inoculation, they are in quarantine for at least fourteen days and until they are not infectious. We take twice daily swabs to test for PCR to see whether they’re infected, and it’s only when they are uninfectious that we can release them. And obviously during that time, they’re monitored extremely closely with scans of their lungs, scans of their heart, repeated blood [tests], and twenty-four/seven nursing and medical cover.
TS: It sounds like you’re planning on doing additional challenge studies. Can you elaborate on why you see this as a useful tool for studying COVID-19 going forward?
HM: I think the advantage of controlled human infection models is that they are controlled—so they control for both timing and dose of exposure. You can look at natural history studies, and you can see that the rates of reinfection are quite low. But the problem with those studies, or limitation of those studies, if you like, is we don’t know whether somebody has not been reinfected because they have protective immunity, or it may just be that they haven’t been exposed to SARS-CoV-2, and that’s why they’ve not been reinfected. Whereas in these studies, because we know we are exposing everyone in a very controlled way, if people are not infected after that controlled exposure, we know they must have protective immunity.
Editor’s note: This interview was edited for brevity.