Here’s a cautionary tale from late last year that you might have missed – there’s a coronavirus angle, for sure, but there’s also a vivid warning not to go around making confident predictions about human immunology. As became clear during the Omicron wave – and I am very, very happy to use the past tense in this sentence – several of the existing monoclonal antibodies against SARS-Cov-2 turned out to be useless against the new variant. At the same time, the efficacy of vaccination went down, too, but still stayed at very useful levels after a booster shot. The difference is that a monoclonal is just what it says on the label: a single antibody, given in quantitiy. It binds with terrific specificity to one particular spot on its target antigen – and if that antigen has thrown a few mutations, it may not be able to bind any more. But after vaccination (or infection), your body produces a strongly polyclonal response: hundreds, thousands of different antibodies answer the call, binding all over the target, and the most useful ones get amplified and refined by B cells.
But unfortunately, attempts earlier in the pandemic to dose people with polyclonal antibody fractions from recovered patients really didn’t work out so well – the efficacy was quite disappointing. Perhaps that had to do with the (rather variable) pooled samples used for dosing, the amounts that were given to patients, the timing of the blood collections used to assemble them, or perhaps it speaks to the relative importance of T cells as opposed to antibodies alone. A large dose of the approved monocolonal antibodies, on the other hand, did seem to be effective. Until it wasn’t, of course. And since pretty much everyone was getting infected with Omicron, and the Omicron proteins themselves were where the problem lay, there was nothing to be done about it.
That’s where a small company called Adagio comes in. You can still see the pitch on their web site: their ADG20 antibody was meant to be a broadly neutralizing one that would hit a whole range of coronavirus types (and coronavirus mutants). That’s what you’d want for a therapeutic antibody in general, but it’s not easy to get there. You need to find some viral protein that has a very conserved sequences across a whole range of variants and types, but at the same time this has to be a target for which antibody binding really shuts down the infection. Not all antibodies are neutralizing, remember, and at times antibody binding can even make things worse. That’s the too-famous-by-now Antibody Dependent Enhancement, which we are still seeing no signs of whatsoever with the current coronavirus by the way (not after infection, not after vaccination) and that’s an excellent thing. But it just means that your broad-spectrum antibody had better do the job, and there’s no guarantee of that.
Adagio, it’s safe to say, raised expectations. Here’s a press release from late November, where they claimed that none of the Omicron mutations were associated with escape from the relentless ADG20. They expected it to work, to work well, to deal with Omicron and all of the other variants of concern so far, to be well-equipped to handle whatever else was coming down the chute, and for a single injection to be effective for up to a year. The folks who had bought into their August IPO (or their earlier private financing of over $300 million) were probably happy to hear about all this; the stock went from around $18 to over $46 on the news. But the party didn’t last for very long. Two weeks later, the clinical results showed that ADG20 was not, in fact, effective against Omicron. The stock went to $7, and if you feel a desire for some today you can get it for less than that. This result was “not suggested by prior data”, the company said, which I suppose is fair, although “not suggested in our prior in vitro data” is more accurate, or perhaps “not suggested in our prior press releases”.
Everything has to be proven in the clinic, out in the real world. One of the reasons that I was so loud about not seeing ADE with the vaccines early on was because no sign of it had been seen in the animal models nor in the human clinical trials. Then as the various vaccines started rolling out into the human population we continued not seeing it, and we continue not seeing it today, even after several variants have had their shot. These data are more valuable than all theoretical considerations about what might cause trouble, because we don’t understand the human immune response well enough to make finely targeted predictions. Even as the companies making the earlier monoclonal antibodies were running in vitro experiments against Omicron (and not liking what they saw), reports were coming in from hospitals that they indeed seemed to be losing efficacy. Clinical data is and always will be the coin of the realm.
