ChemJam-CD47-SIRPalpha

The case for CD47-SIRPα small molecule inhibition

July 22nd, 2025 By John Widen

The other day I was perusing Arch Venture’s company portfolio and landed on Bitterroot Bio. Looking at their website, it appears that Bitterroot Bio has a single asset, BRB-002, which is a CD47 targeting antibody. According to a press release on Bitterroot’s website, the first patient has been dosed in a Phase 2a study (MATADOR) with BRB-002 to treat patients with atherosclerotic cardiovascular disease.

And so began my dive down the rabbit hole looking into CD47 and SIRPα (Signal Regulatory Protein alpha) as small molecule therapeutic targets.

I’ll begin with an introduction to CD47 (a.k.a Integrin-associated protein, IAP), which is a cell surface immune checkpoint protein that is ubiquitously expressed in normal tissues but also overexpressed in cancer cells. CD47 binding to SIRPα, which is expressed on macrophages, inhibits phagocytosis. CD47 is a ‘don’t eat me signal’ for macrophages. This is advantageous for cancer growth and progression to evade immune surveillance and create an M2-type (tumor promoting) microenvironment. CD47/SIRPα is complementary to the well-known PD-1/PD-L1 immune regulatory pathway.

The importance of immune regulatory checkpoint pathways as therapeutic targets for cancer became widely apparent with the huge successes of Pembrolizumab (Keytruda), Nivolumab (Opdivo), and many others as potent activators of the immune system toward inhibition of cancer growth and progression. FDA approved immune check point inhibitors only work for a subset of patients but works well for those individuals. A lot of effort has been put into understanding immune checkpoint mechanisms and approaches to improve upon these initial successes. I will stop there to focus on CD47 and SIRPα.

There are many companies pursuing CD47/SIRPα as therapeutic targets based on ongoing clinical trials, almost exclusively focusing on large molecules (e.g. antibodies) as the modality. Bitterroot was founded by a group that previously founded Forty Seven, which was acquired by Gilead for $4.9 BILLION. Gilead acquired Forty Seven for their lead asset magrolimab, an anti-CD47 antibody. At the time, Forty Seven had several ongoing clinical trials for magrolimab including myelodysplastic syndrome, acute myeloid leukemia (AML), and diffuse large B-cell lymphoma (DLBCL).

Unfortunately for Gilead (and patients), magrolimab had several issues during phase III clinical trials for all three indications that led to a partial clinical hold due to safety concerns in the middle of the study followed by discontinuation. The discontinuation was because magrolimab plus the standard of care did not outperform the standard of care alone. The efficacy of the therapy is confounded by the partial clinical hold and statistically significant increased risk of therapy related adverse events and death.

For SIRPα, the first in-human phase I clinical trial report for GS‐0189, a humanized IgG1 monoclonal anti‐SIRPα antibody, reported less severe adverse events but a lack of phagocytic response in this small cohort (n=9). The study was discontinued by Gilead in favor of further development of magrolimab (whoops!). That was likely a poor decision in hindsight but at the time magrolimab had already passed phase I trials.

The issues with magrolimab and GS-0189 did not deter others from developing their own anti-CD47 or anti-SIRPα large molecules. Many normal cells express CD47 on their cell surface. Magrolimab caused off-target toxicity likely due to the Fc domain of the antibody, which raised an immune response to normal cells. This hypothesis seems to be validated because there are >100 clinical trials ongoing involving CD47/SIRPα large molecule targeted therapy. The vast majority of them targeting CD47 in combination with standard of care chemotherapy or as bispecific antibodies that also target PD-1/PD-L1.

Coming full circle, Bitterroot was founded on the premise of a key finding in Irving Weismann’s lab at Stanford University (the father of hematopoiesis). His group published research demonstrating that anti-CD47 therapy can ameliorate atherosclerotic plaques in mice by activating macrophages (Kojima, Y. et al. Nature 2016). Bitterroot demonstrated safety of their new antibody (BRB-002) in a phase I clinical trial and ran a small human trial to demonstrate feasibility prior to their phase II trial (vide infra).

This is all to say that the safety concerns for targeting CD47 and SIRPα seem to be largely addressed. Thus far, the score card for developing small molecule therapeutics targeting CD47 and SIRPα are as follows:

Target indications:
Cancers with high expression of CD47 and have immune infiltration. Should be substantial!
There is a second solid indication in treating atherosclerosis. Without looking at any specifics I would hazard a guess that this population is sufficiently large.
Not an expert on the clinical side of things but there are other illnesses such as deep vein thrombosis. I believe this therapy would apply.

Safety Considerations:
Large and small molecule modalities targeting CD47 and SIRPα have passed phase I clinical trials. Low safety risk.

You may have just noticed that I mentioned that small molecule modalities have also passed phase I clinical trials. Despite the CD47 targeted therapies being DOMINATED by large molecules, there is one CD47 targeting small molecule program from Aurigene, AUR103, that completed a phase I trial and is recruiting for a phase 1b/2 (NCT07040059). This should reduce the safety risk even further because a small molecule has demonstrated tolerability in humans.

I looked through the literature for signs of other small molecule inhibitors that bind to CD47 and SIRPα. Surprisingly, I could not find much in the realm of small molecule inhibitors of the CD47-SIRPα interaction. I found the aforementioned Aurigene patent applications and a patent in Chinese from China Pharmaceutical University with three HTS hits (odd and not promising). I could not find any academic work regarding small molecule inhibition of CD47-SIRPα via CD47 binding. There are a few publications and patents regarding direct binding to SIRPα. I will discuss the chemical matter in Part 2 of this post (blog series!).

The lack of patent applications for small molecules towards these targets could be a really good thing or a bad thing. I’m sure there is a substantial amount of unpublished work for both CD47 and SIRPα small molecule inhibition. But the fact that there aren’t many published patent applications right now makes me think that this target is wide open for the taking! It’s a perfect time to make a few literature molecules to get a baseline on inhibition and establish assays.

There is a potential that many are scared away from the prospect that there are quite a few anti-CD47 antibodies being evaluated in the clinic as well as substantial preclinical development. I might be bias, but small molecules will always have several advantages over antibodies. I won’t go on a diatribe, but two obvious reasons are route of administration (I.V. vs P.O.) and manufacturing cost. Would you rather go to a hospital multiple times a month to get an injection or take a pill once/twice a day?

One potential pitfall to a small molecule approach to inhibition of the CD47-SIRPα interaction is that it is challenging to get potency. Antibodies can have exquisite potency and selectivity. Based on the small amount of published work targeting the CD47-SIRPα interaction with small molecules there is a >10x potency shift from binding (K_D) to cellular IC₅₀. This is a fairly common phenomenon in small molecule drug discovery. Therefore, that wouldn’t deter me from pursuing this target initially. But it's definitely something to keep in mind as the program progresses. There are plenty of challenging drug targets where potency was challenging. That is what medicinal chemistry is for! Since this is more focused on the chemical matter, I will dig into this a little more in the next post.

Searching the PDB database, there are no published structures of CD47 or SIRPα bound to a small molecule, but there are plenty of large molecule complexes. Additionally, there is a publication that models what is likely the Aurigene molecule, AUR103. So, there is a possibility to run a virtual screen in that defined pocket or find another likely pocket to do a screen. I will discuss the chemical matter in Part 2 of this blog series!

Did I make a convincing argument? Or am I missing a big piece of the puzzle that would explain why there is not much in the literature in terms of small molecule inhibition of the CD47-SIPRα interaction. Will someone pony up some cash to fund a screen and/or literature-based approach to discover and develop small molecule inhibitors of the CD47-SIPRα interaction? Unless someone has a good reason not to, I’ll take a closer look at the chemical matter for these targets, discuss potential screening approaches, and the assays required to start a CD47-SIPRα in a follow-up blog post (blog series!). Stay tuned!

The site does not have a comments section yet! Hopefully, very soon! Until then please drop me a line at jwiden@chemjam.com. If you provide comments on my articles I reserve the right to post them on this website as additional commentary. My goal is to have an open discussion!