Let’s talk about a protein called CD47. It’s expressed all over the human body, and it’s a big player in immune signaling and several other processes. One of its immunological roles is serving as what’s called a “Don’t Eat Me” signal for immune macrophages – it’s part of the friend-or-foe signaling that tells these cells when to attack and when to let things pass. As you’d expect, that’s of great interest in research programs that are trying to get the immune system to attack things that it doesn’t, but usefully could. Many tumors overexpress CD47 on their cell surfaces, for example, and the scar tissue that builds up in various fibrotic diseases is another good candidate. It could be extremely useful to suddenly turn the power of the immune response loose on these things, but as usual with these ideas, it’s going to be a bit of a tightrope. Your immune system is capable of doing an enormous amount of damage if it gets mis-deployed, so getting it to just destroy the things you want destroyed is key.
In 2020, Gilead bought a company that was focused on this sort of work, focused enough that they were named “Forty-Seven”. That deal cost about five billion dollars, and the biggest proximal asset was magrolimab, an anti-CD47 antibody therapy. The idea was to go in and start blocking CD47 on the surface of cells, in the hopes that as you turned this signaling down that blood cancers like myelodysplastic syndromes, acute myeloid leukemia, and others could have their aberrant cells opened up to immune clearance. Forty-Seven got their first big clinical readout on that idea in 2019, combining magrolimab with another chemotherapy agent (azacitidine) used in these disease settings, and results were very promising. They had a 92% overall response rate in severe MDS patients, and a 64% ORR in AML patients who were ineligible for other types of therapy. No particular adverse events were noted. Clinically, this is a tough crowd, and these numbers got a lot of attention. You could have purchased Forty-Seven for a lot less than five billion dollars before those data, but the data are why you’d think about purchasing them at all, of course. AbbVie bought into lemzoparlimab, another anti-CD47 antibody from a company called I-Mab, and since then, Pfizer has paid $2.3 billion for another small company, Trillium, with two anti-CD47 antibody of their own. So there’s a lot of firepower in this area as of late.
Gilead went full speed ahead in the clinic, with the other anti-CD47 people chasing after them, but there are other companies with their own ideas about targeting MDS and related diseases. Takeda, for example, was going with pevonedistat, which works by a completely different mechanism. It’s an inhibitor of an enzyme called NEDD8, covalently linking to a key cysteine residue in that protein and thus shutting down its activity. NEDD8 itself is involved in at least two DNA repair pathways, and it’s also an irreplaceable step in the activation of another set of enzymes called Cullin-ring ligases, which are themselves key steps in the process of protein degradation. Cancer cells can be quite sensitive to both of those pathways – they have to clear a lot of proteins in their fast-moving metabolisms, and they are often just barely keeping ahead of their own genomic instabilities when it comes to DNA repair. So pevonedistat was in trials in AML and MDS patients, and initially showed promise of its own. But then it went on to miss its primary endpoint in the Phase 3 trials last fall.
That was food for thought for Gilead, since they had all those similarly promising early results in just those diseases with magrolimab, and if that sort of thing can blow up on you in a bigger trial, well. . .And there was also some need to talk to the FDA about what a New Drug Application would look like in these areas, in light of Takeda’s failure. The plan had been to ask for accelerated approval even before the Phase 3 trial got completely enrolled, but Gilead moved up the clinical readout on the magrolimab trials to 1Q 2022. Which is where we find ourselves right now!
But things haven’t even made it that far. The FDA just came in with a partial clinical hold on the magrolimab trial, saying that there was an imbalance between adverse events in the various study arms. Gilead is gathering more data to address the issue, and although this puts a hold on Phase 3 recruitment, the optimistic case is that the magrolimab data will still be good enough that they can file for accelerated approval as planned. The pessimistic case is that there’s something wrong with magrolimab, and that approval of any kind might be well delayed or not even happen at all. The really pessimistic case is that there’s something wrong with the whole idea of anti-CD47 antibodies as immuno-oncology therapies. It’s too early to say any of that, of course, but it’s not too soon to be worried, immunology being what it is. The only way to prove your good ideas in this field (or to find out that they’re not actually such good ideas) is to do definitive clinical trials in humans, which takes a definitive pile of money and represents a definitive risk. More exciting chapters in this story will follow; that at least we can be sure about.
