I’m getting a lot of queries about antibody-dependent enhancement these days, and I can only assume that’s because there’s a lot of talk about this making the rounds of various social media platforms. Many of the people who are contacting me sound a lot more worried than I would have thought, so that prompts me to follow up on the post I did on the subject back in December.
What’s ADE, Again?
First, a quick recap. ADE is a problem that has shown up in several sorts of viral infection, although it also has to be said that there are other viruses in which it’s never really been seen at all. It happens when a previous infection or vaccination has generated antibodies that fit some specific criteria. First, these existing antibodies have to be non-neutralizing against the new viral infection: that is, they bind to the second virus, but not in a way that shuts down its activity. It’s important to realize, though, that *all* immune responses to a viral infection generate a mixture of neutralizing and non-neutralizing antibodies. That’s one of the things about the immune system – it revs up production of a wide variety of antibodies, selected from the untold billions of them circulating around in your bloodstream. Some of them bind to one part of the pathogen, and some to another. And they bind in different conformations, sticking to different parts of the surface of the invading virus from different directions.
Some of these are inevitably going to be more effective than others at stopping that virus’ activity – and remember as well that there are several ways that can happen, too. An antibody can bind to and cover some key part of a virus protein without which it can’t infect human cells (in the case of the current coronavirus, that could be the receptor-binding domain (RBD) out on the end of the Spike proteins that decorate its surface. Antibodies can also cause aggregation, sticking viral particles in clumps that can’t function as they would otherwise. And they can also signal various kinds of defensive cells to attack an antibody-bound viral particle directly and destroy it.
But if none of these work as hoped for, then you have a non-neutralizing antibody. The immune system is actually optimized for selecting and amplifying the neutralizing ones, though. So it’s usually not a problem having the non-neutralizing ones around at the same time, since the other more useful potent ones are out there taking care of business. But what if you don’t have any of those, just the non-neutralizing ones?
That’s what happens, for example, with Dengue fever. Dengue comes in four different varieties (which is probably the single hardest thing about trying to treat it or prevent it through vaccination). The antibodies you generate that can get you past one of the infections really don’t match up well enough with the others to be effectively neutralizing, and if you get one of those later on you can actually get a worse case of Dengue than you would have had otherwise. That’s the “enhancement” part of ADE. As mentioned in that December post, there are at least two different mechanisms that have been worked out for this. One of them (the most straightforward) seems to be that when some types of non-neutralizing antibody are stuck to the denguevirus particle, that it actually speeds up its entry into human monocyte cells. The monocyte membrane proteins treat the incoming antibody surface like an incompetent bar doorman letting people through with fake IDs: “Looks good to me, come on in”. Which is exactly what you don’t want.
That December post has links to times this has been seen with vaccinations, too: there has been one RSV vaccine candidate and one measles vaccine candidate that have certainly shown this problem (the antibodies they generated made the next exposure even worse). It’s not common, by any means, but it can most certainly happen. And you can believe that vaccine developers are aware of this. Which brings us to:
ADE and Coronaviruses
Now, when SARS appeared in the human population back in 2003, there was a lot of work done on it to try to make vaccines, should it erupt again. And (links in the earlier post) some of these candidate did show signs of ADE. When vaccinated animals were re-exposed to the same virus, some of them got even sicker than usual. (As an aside, this seems to have been through yet another different mechanism: an altered T-cell response, rather than a direct effect of binding antibodies on cell entry). Immunology being what it is, this certainly didn’t happen in every animal. Every mammal’s immune system is different, like a fingerprint, and it’s clear that with such a vaccine some people (through sheer bad luck, impossible to predict with current techniques) would be more vulnerable.
What you can do is see what the statistics are like – if you see any sign of ADE at all in an animal model experiment, that’s bad news, because the sample sizes for these are far, far smaller than the population that’s going to be getting vaccinated. And that would mean completely unacceptable risks in that human population. So animal studies (both rodents and primates) are specifically designed to look for such effects, and if ADE is seen, well, it’s back to the drawing board. You’ll also be watching your clinical trial data and (indeed) the eventual real-world rollout for any signs of this as well.
The SARS experience taught us a lot of extremely useful lessons, as it turned out. SARS-Cov-2 is rather closely related to the 2003 SARS coronavirus, and if you’re going to have a worldwide pandemic, you’re far better off with one that’s so much like something you’ve already poured R&D investments into! In this case, the two big take-homes were that coronavirus vaccines could indeed suffer from ADE, and that this seemed to depend on which protein you chose to base your vaccine around. Specifically, it was the vaccines that targeted the N (nucleoprotein) antigen of the coronavirus that had ADE problems,
while the ones that targeted the S (Spike) protein did not. Update: this isn’t accurate. There was trouble after immunization with a nucleoprotein-directed vaccine, but ADE could also be seen with some of the Spike-directed vaccine candidates as well – see reviews here, here, and here. That experience was thoroughly taken to heart in the vaccine developments of the last year: no one, to the best of my knowledge, even bothered to target the SARS-Cov-2 N protein at all, for just this reason. If you look at the antibodies generated in people who’ve been infected by the virus, they most certainly did make N-targeting ones, along with Spike-targeting ones and antibodies directed against the various ORF proteins. But for vaccine work, everyone has stuck with the Spike.
The Current Vaccine Data: Any Sign of ADE?
So now the Moderna and Pfizer/BioNTech vaccines have been rolled out in many parts of the world, along with the AstraZeneca/Oxford, Gamaleya, and CanSino adenovirus vector vaccines. Those look to be joined soon by J&J’s adenovirus vector and Novavax’s recombinant protein subunit vaccines, and likely more after that. So here’s the key question: did any of these show ADE hints during their development? And are any of them showing signs of it now?
The short answers: they did not. And they are not. Antibody-dependent enhancement was specifically tested for in the animal models as these candidates were being developed (re-exposure of vaccinated animals to coronavirus to see how protective the vaccine was). And no cases of more severe disease were seen – I’ve gone back through the reported preclinical studies, and I don’t think I’ve missed one, and what I’m seeing is not one single case of ADE for any of them. Indeed, as mentioned above, if something like that had shown up, it would have immediately released a bucket of clin-dev and regulatory sand into the gears of the whole project.
How about the human clinical trials? Again, no signs of ADE were seen. This is a bit less definitive, since we did not run deliberate “Here, have another blast of virus” challenges on the human participants the way we did in the preclinical studies. But at the same time, these trial participants were out there in the real world being monitored for signs of infection. The dramatic plots of the data after even one dose of the vaccines speak for themselves: the trials did hardly saw people getting infected at all after vaccination, and most certainly not with even more severe disease. To the contrary: one of the big features of the vaccines is that across the board they seem to almost totally wipe out the appearance of severe coronavirus symptoms. We’re still collecting data on transmissibility after vaccination and so on (things are looking good, though), but what seems to be beyond doubt is that the vaccinated subjects, over and over, show up with no severe coronavirus cases and no hospitalizations.
That is the opposite of what you would expect if ADE were happening. Remember, the bad thing about antibody-dependent enhancement is that it leads to more severe disease when you’re exposed again to the pathogen (or when you’re exposed after being vaccinated for it). And we’re just not seeing that. At all. We are, and I am very, very happy to be able to say this, seeing exactly the reverse. Watching the real-world data will alert us to any of the potential mechanisms (antibodies, T-cell effects, etc.) and nothing is showing up.
What About the Variants?
That’s a really good question. The earlier trials were run against what I’ve been calling “coronavirus classic”, and now we have several variant strains to contend with. The worse case is that one or more of these spread out to be as different as (say) the four types of Dengue fever, and that the antibodies raised by vaccination are inadequate to deal with them. That would mean several bad things: that people who are vaccinated would still be at significant risk for a regular infection, and even worse, that they cold be at risk for an even worse one than if they’d never gotten vaccinated at all. That seems to be the fuel for the current brushfire of ADE worries.
The news about vaccine efficacy against these variants is actually not as bad as you might have thought, based on some news reports. The B.1.1.7 variant (the one that was first characterized in the UK) seems actually to be handled quite well by the various vaccines, with only a very small dropoff in efficacy (and definitely not enough to start worrying about ADE). The B.1.351 variant that was first characterized in South Africa is a bit tougher. There’s no doubt that the antibodies generated by the various vaccines have a harder time with this one, but it looks like the degree of dropoff varies. On one end, the AstraZeneca/Oxford vaccine appears to lose potency to a degree that the South African government stopped vaccinating with it entirely. Now, that may or may not have been a hasty decision – the vaccine, even in South Africa, is a hell of a lot better than nothing – but that’s another topic. Meanwhile the J&J and Novavax vaccines show less efficacy against B.1.351, although apparently not to the degree that the AZ/Oxford one showed, and word has come within the past few days that the Moderna and Pfizer/BioNTech vaccines may be holding up even better than that: a drop in potency in lab experiments, but maybe not enough to even show up much in the real world population at all.
And so far, I have been able to find no reports of more severe disease after vaccination in South Africa. Please correct me if I’m wrong, those of you following this closely, but this would mean that there is (so far) no evidence of antibody-dependent enhancement against even this variant. There are of course other variants, and we most certainly need to keep an eye on them. But variants are what viruses do. This isn’t something weird and sinister – it’s expected and we know what to look out for. I’m going to have another post on these strains up on Monday or Tuesday, but I’ll just say that I’m actually relieved that we’re weathering these as well as we are.
The Bottom Line
So here’s the short version: no sign of ADE during the preclinical animal studies. No sign during the human clinical trials. No sign during the initial vaccine rollouts into the population. And (so far) no sign of ADE even with the variant strains in different parts of the world. We have things to worry about in this pandemic, but as far as I can tell today, antibody-dependent enhancement does not seem to be one of them. I understand why people would worry about it, and want to avoid it. But if you’re coming across reports that say that it’s a real problem right now and that you should avoid getting vaccinated because of it, well, I just don’t see it. Some of that is well-intentioned caution, and some of it is probably flat-out anti-vaccine scaremongering. Anyone with different data or different impressions, well, that’s why the comments are open around here!