COVID-19 Vaccines: Moderna, Pfizer/BioNTech, AstraZeneca/Oxford
More than 150 COVID-19 vaccines are in development around the world. The hope is that at least one of these could be brought to market quickly, something that could ease the global crisis in which we now find ourselves. The U.S. government has launched Operation Warp Speed, pledging $10 billion to developing, testing, and distributing an effective vaccine by early 2021. On July 20, Michael Weissman, GLG VP and Team Lead of GLG’s Healthcare Content team, spoke with Michael Farzan, PhD, Co-Chair of the Department of Immunology and Microbiology at Scripps Research, to discuss COVID-19 vaccines and the early-stage data coming from Moderna, Pfizer/BioNTech, and AstraZeneca/Oxford. This interview has been shortened and edited for clarity.
AstraZeneca and Oxford recently published their data from their phase one trial. Can you give us your initial thoughts on this data? To the extent that you’re able to compare apples to oranges, how does this compare with the recent published phase one trials from Moderna and Pfizer and BioNTech?
My first impression was a slightly positive surprise about the degree of neutralization that was observed in these studies, compared with some of the preclinical studies, especially in primates, where the neutralization levels of this class of vaccines was less impressive than, say, an mRNA or subunit vaccine. In that regard, it was slightly more impressive than I anticipated.
You’re right, these are somewhat contrasting comparisons. They are different groups doing different kinds of studies. That said, they’re all in roughly the same range of about a 1 to 100 or 1 to 200 neutralization titers, which is interesting because they have different antigens and they have different delivery systems.
Digging into that a little bit, you were mentioning different delivery systems, different antigen designs, and I’m somewhat surprised that it’s all similar.
Looking at these first few early-stage studies, is there anything noticeable in terms of adverse events (AE) or anything else that stands out?
Not yet. When you consider AEs, you want to think about several classes. One AE emerges when you have a reasonably good vaccine response. The vaccine is driving a lot of interferon, making the patient feel like they have a flu. That means that it probably is correlating with a better response to the antigen. Those individuals who get sicker have the better antibody response after that illness.
We don’t know yet about the other two classes of adverse events. One is the tendency of a vaccine to actually make an infection worse. That has been a concern in the coronavirus field, and other vaccines as well. Although that kind of concern seems to be receding, we should keep a careful eye out for it.
Other kinds of AEs are not necessarily connected to the immune response. There are no benefits to those AEs, just harm. I haven’t seen anything that’s flagging anything like that right now. So, most of the things seem to suggest the kind of response you get with sort of a well-adjuvanted vaccine — an immune response that is responding to what has been perceived as danger.
In the first three candidates that are farthest along, all require boosters. In the Oxford data, it was shown that over 90% of the people elicited some antibody response with just the first dose. And the second dose, with the booster, pushed it to 100%. If you look at the other two vaccines, they didn’t quite hit the same levels with just one injection. Do you think it’s possible that for the Oxford/AstraZeneca vaccine, that potentially only one shot would be necessary, or do you think that the booster will be required?
The main issue is how good the neutralizing response is. That tells you whether you’re going to get a little bit sick. In most cases, the companies working on this first crop of vaccines may not have made the best decisions in terms of their antigen design, so I think you’re going to need the booster. Which means they have room to improve.
I have no doubt they’re going to have second-generation vaccines that will learn from their own experience and everybody else’s experience to get primarily better antigens, allowing us to get back to a single vaccination rather than a prime-boost kind of arrangement. But so far it seems that all of them are going to be the prime boost. Because even then, that’s when they’re getting to the 200 or so neutralization titers. And really, if you were dreaming, you’d want 1,000 or 5,000 dilutions to neutralize 50% of the virus in a tissue-culture setting.
Can you share your thoughts on the impact of convalescent plasma comparator groups of the trials having mostly mild disease, and whether that impacts the hurdle for antibody response?
The main problem with convalescent plasma is the wide range of potencies. People who do not get that sick tend to be the donors, and they also do not tend to have very good neutralizing responses. We ourselves have been looking at that, at different convalescent sera, and it’s really surprising the number of people who have tested PCR positive for the virus and who actually have good anti-nuclear capsid antibodies but do not make good S-protein antibodies — the S-protein being the only real target for neutralization on the virus.
I know OneBlood is doing a better job now of screening through the sera to make sure that they’re using the more potent sera, but that was a problem with the early use of convalescent sera. I should also say that I suspect it’s going to be supplanted pretty quickly by a cocktail of two neutralizing antibodies that we know are very good and are well controlled and scalable, and you don’t have to draw from one person to give to another person, etc. Hopefully we’ll be fading out with convalescent sera and replacing them with some of the neutralizing antibody cocktails that should be out soon.
Do you think that the variability of titer assays is a confounding factor in the vaccine development, and as we’re trying to read the results that come out, there should be standardization across the board?
We as the public are going to be paying a lot for these vaccines in a variety of different ways. And because we’re paying a lot, we should have the right to ask for apples-to-apples comparisons of the vaccine responses in disinterested hands.
One of the sources of variability is different kinds of assays. Another source of variability is different groups of people with motivation to make sure that their vaccine is looking great. So, hopefully, we can get to a standardized set of protocols, with a disinterested group running these protocols. We’re paying for it.
In terms of individual vaccines, most people will naturally gravitate toward the more physiologic-sounding assays, like a plaque-forming assay or something of that nature. But there’s a buyer-beware there. First of all, you compare the various assays, and they all are not so bad. They’re all about within threefold of one another, and so you should mentally just put a threefold error bar on everything. So ideally everybody’s plotting things on a log scale, so those error bars are not that great.
But the different assays have different strengths and weaknesses. And one of the underappreciated weaknesses of assays relying on replicated virus, plaque-forming assays, etc., is that they’re not linear. Which is to say, that plaque formation is a complicated process that does not follow a straight line as you change the amount of antibody you apply to it. It’s a little bit easier to put your thumb on those assays and push yourself into a regime in which the curve is steeper. This means that you’re playing with the fact that it’s nonlinear in such a way as to get a better result.
I think AstraZeneca did a good job of combining multiple assays and showing their correlation and everything. I also think that the pseudovirus assays, which tend to be criticized because they’re not physiologic, are also good because there’s less play in those. There’s less sensitivity to the conditions in which you perform the assay.
The good news is all of these assays are within a threefold range of one another. And they’re all within the error or the resolution that we’re getting out of these different assays performed by different companies under different conditions.
About Dr. Michael Farzan
Dr. Michael Farzan has been Co-Chair and Professor in the Department of Immunology and Microbiology at Scripps Research since August 2012. His laboratory focuses on entry processes of enveloped viruses and the various ways that the innate and adaptive immune responses inhibit these processes. They also apply these insights toward developing therapeutic strategies that enhance or supplement these immune responses.
This article is adapted from the July 20, 2020, “COVID-19 Vaccines: Moderna, Pfizer/BioNTech, AstraZeneca/Oxford Data Review” teleconference. If you would like access to this teleconference or would like to speak with Dr. Farzan, or any of our more than 700,000 experts, contact us.
GLG is supporting nonprofits on the front line of COVID-19 relief, pro bono. If you represent or know of an organization that could use our help, let us know here. If you are a GLG network member whose expertise might be valuable to a relief organization, please get in touch here.