This piece at Forbes by Bill Haseltine has set off a lot of comment – it’s a look at the Oxford group’s vaccine candidate as compared to the SinoVac candidate, and you may recall (background here) that these are the two teams that have separately reported that their vaccines appear to protect rhesus monkeys from infection after exposure to the coronavirus. Haseltine has some criticisms of the Oxford data, and as you will see from that link to his name, his opinions deserve to be taken seriously. So what’s going on? Update: here’s the take on this at BioCentury.
Looking at the preprint on the Oxford results, Haseltine has a problem with the claim that the monkeys were protected from infection by a dose of ChAdOx1 nCoV-19. The key data are in the preprint’s Figure 3. The Oxford team checked for viral RNA several different ways. One was using bronchoaveolar lavage (BAL fluid), a sampling technique that involves running a bronchoscope down into the lungs and washing out aveolar spaces – a pretty darn invasive assay, which is why you don’t hear about it all that much compared to the still-not-so-nonivasive nose swabs. BAL fluid of the virus-exposed unvaccinated animals showed coronavirus genomic RNA throughout the study, and viral subgenomic RNA (more indicative of active replication) at days 3 and 5 after exposure. Meanwhile, the vaccinated animals showed the genomic RNA in only two monkeys, and no subgenomic RNA at all.
So far, so good. But both vaccinated and unvaccinated monkeys showed the same amount of viral genomic RNA from nose swab samples (Figure 3c). That’s the test that’s used out in the human population, and that means that the vaccinated animals would still be declared as positive for the coronavirus after being exposed to it. And the other thing that Haseltine notes is that the amount (the “titer”, in the lingo) of neutralizing antibodies in the blood of the vaccinated animals does not appear to be that high. You’d like to be able to dilute the blood antibody samples down by hundreds of times or even a thousandfold and still see antiviral activity in an in vitro assay, but in the Oxford case the activity started disappearing at about fortyfold dilution (Figure 2b).
On the positive side, 2/3 of the unvaccinated animals showed clear evidence of viral pneumonia at autopsy, but none of the vaccinated ones did. The conclusion is that the vaccinated animals were indeed infected – the vaccine did not protect against that – but that the disease was definitely less severe. But these results mean that the virus might well still be transmissible from people who had been so vaccinated, even if the disease course itself was not as deadly. You’d want to do better than that, if you can. Haseltine’s take is “Time will tell if this is the best approach. I wouldn’t bet on it.”
Haseltine compares these results to the SinoVac inactivated virus vaccine, and finds that that one looks better – at its highest dose, no viral RNA was recovered from the tissues of the vaccinated animals, for example. This sort of “sterilizing immunity” is what you’d want to aim for – it gives the virus nowhere to go in the human population if you can vaccinate enough people. But it’s worth noting that the SinoVac results were from three doses of their vaccine (versus one of the Oxford candidate), and the viral exposure challenge was about half as strong (total viral particles) as what the Oxford paper used. The Oxford group also inoculated their monkeys in both the upper and lower respiratory tract, while the SinoVac team used a single inoculation in the trachea. So I agree with that tweet linked from AndyBiotech; I don’t think that a head-to-head comparison is fair. But Haseltine’s point stands, that the results as we have them from the ChAdOx1 nCoV-19 vaccine did not actually protect monkeys from infection.
If you dosed it more than once, perhaps it would. Update: that’s not going to work, of course. After a shot of the adenovirus vector, you’re not only going to have neutralizing antibodies to the desired antigen, but you’re going to have them to the adenovirus as well. It’s probably one shot and done. Update 2: maybe not! See this trial; they’re going to try it and see what happens. But we’re going to have to keep an eye on dosing schedules, of course. We’re looking at a massive effort to vaccinate people all over the world, in the end, and if we have to vaccinate them all twice (at, say, a two-month interval) then the problem becomes just that much harder. Three times, well, that’s something that I think we’re only going to seriously consider if we absolutely have to. If you want to picture an ideal vaccine, it’s one dose, sterilizing immunity, extremely low incidence of side effects. We have no idea at present if we’re going to get that, or even be able to get that, and the only way we’ll find out is human data. Lots and lots of human data. May the best vaccine win, and may the best vaccine be very good indeed. . .