(Update, March 17: See also later entry here.
There's been a severe problem with the first-in-man dosing of a new antibody in England. TeGenero, a small company in Germany, was testing their most advanced candidate, TGN1412, which is a monoclonal antibody against CD28. That's a glycoprotein on the surface of many types of T cells, and it seems to be extremely important in several mechanisms that activate them. Accordingly, finding something to modulate its activity would seem to be a promising way to attack autoimmune and other inflammatory diseases.
But when TGN1412 went into six volunteers in Phase I this week, all of them ended up in the hospital, and two of them are gravely ill. The press reports make it seem like some sort of anaphylactic reaction, but it sounds like they're not treating it like simple anaphylaxsis, so there must be something more going on. That's a very surprising outcome, since if anything you'd expect the immune response to be downregulated in response to the drug. (Update: this is only partly true - see end of post.) But there's an awful lot about human immunology that we don't understand, to put it mildly. TeGenero says that:
"These events were completely unexpected and do not reflect the results we obtained from initial laboratory studies which enabled us to progress investigations into human volunteers. . .The drug was developed in accordance with all regulatory and clinical guidelines and standards. In pre-clinical studies, TGN1412 has been shown to be safe and the reactions which occurred in these volunteers were completely unexpected”
I have to believe that this is correct. Neither TeGenero, the company conducting the trial for them (Parexel), or the British regulatory authorities have anything to gain from taking a drug into humans that had shown severe effects in animals, despite rumors to the contrary that seem to be going around in England. There's a breathless report in The Independent (I know, that's redundant) to this effect:
"The men, who were offered £2,000 to take part, were recruited by the US company Parexel, for the trial in its 36-bed unit on the Northwick Park hospital campus. They reportedly signed a contract warning that side-effects in rats and mice included "increased urine volume, decreased faeces, redness of the skin". Dogs experienced "increased heart rate and decreased blood pressure"."
I have news for the Independent, though - if we killed off every drug that showed effects like this, we'd never take anything into humans at all. The dog cardiovascular effects would be worth checking out in detail before going forward, naturally, but that's what dogs are for, to show you cardiovascular side effects. And I note that none of these effects have anything to do with devastating immune reactions. If that's all that TeGenero saw in the animal studies, I can see why they were shocked at the human results.
But antibodies are powerful - no one should forget that. Every antibody therapy has a small but real risk of setting off something terrible, and if you're targeting the immune response itself, well, the risk just goes up. The immune system is a bit like demonlogy: don't call anything up that you don't know how to send back down. Unfortunately, we don't know many effective spells yet.
My sympathies go out to those affected and their relatives and friends, of course. They shouldn't, as far as I can see, torture themselves thinking that this could have been foreseen. I hope that everyone makes it through.
Update: I've inadvertently glossed over some of the mechanism of TGN1412. Its binding to CD28 actually sets off the receptor's signaling - it's an antibody agonist. This makes it a T-cell activator, but it seems to particularly activate the class known as regulatory T-cells. These are modulators of the activity of other classes of T-cells involved in autoimmune responses.
So there's a way that things could have gone wrong - if TGN1412 isn't quite as selective in the real-world human immune system as it is in the animal models. Alternatively, even if it is selective but the spectrum of human T-cell response to CD28 ligands is intrinsically somewhat different, things could have gone off the rails very quickly. There appears to have been little reason to suspect either of these possibilities going into the clinic. We're going to learn something important about human immunology from this incident, but this is certainly a hell of a way to do it.
More: BlackTriangle has a roundup of some reactions in Britain.