Embattled Researchers See Path Forward for Anti-CD47 Cancer Drugs

Don't eat me signal (pacman)_istock_artisteer

Pictured: Pacman, representing the “don’t eat me” signal broadcasted by CD47/iStock, artisteer

Companies researching CD47 inhibitors can’t seem to catch a break. Gilead’s multibillion-dollar bet on the CD47-targeting antibody therapy magrolimab appears in jeopardy, and AbbVie recently ended its collaboration with I-Mab, the maker of another anti-CD47 drug, after terminating two early-phase studies of the candidate last August. What once seemed like a competitive field has thinned out considerably, leading some insiders to speculate that CD47 antibodies may be fatally flawed as a therapeutic class.

But there is a glimmer of hope. Earlier this month, Bay Area–based ALX Oncology announced positive interim data from a Phase II trial of evorpacept in HER2-positive gastric cancer, bucking the recent downturns across the field. This positive readout, coupled with ALX’s concurrent announcement that it will move into a Phase III trial in the second half of 2024, underscores that the story of anti-CD47 therapies is still being written, said ALX CEO Jason Lettmann.

Jason Lettman_ALX Oncology [square]
Jason Lettmann

“Like most areas of biotech, there's all sorts of ups and downs, but this one has more than its fair share of incredible highs and moments of a lot of doubt,” Lettmann told BioSpace.

Still, a fundamental question remains: Why haven’t efforts by some of biotech’s biggest names translated into effective therapeutics? Experts told BioSpace that the apparent setbacks can be attributed to known safety concerns for anti-CD47 antibodies combined with an under-diversified field with too many drugs going after the same target. Even so, they remain hopeful that a CD47-targeting drug can make it to market.

Toxic Contradictions

CD47 is a transmembrane protein expressed in many human cells, including red blood cells, platelets and immune cells, as well as in a wide variety of cancer cells. The reason cancer cells express this protein is clear: when it binds to its receptor, CD47 inhibits the innate immune system by broadcasting a “don’t eat me” signal. In the past decade, researchers have engineered antibodies to decrease this signal in cancer cells and, in some cases, provide an unambiguous “eat me” signal to macrophages. 

But these antibody drug candidates often have off-target effects, said Laurence Blumberg, a biotech executive who was formerly the CEO of Arch Oncology, which shuttered earlier this year despite receiving the FDA’s Orphan Drug Designation for its anti-CD47 antibody.

Laurence Blumberg_self courtesy [square]
Laurence Blumberg

“The problem is that CD47 is broadly expressed in normal cells and particularly broadly expressed in components of blood—platelets, white blood cells, red blood cells—and the big issue has been the safety and toxicity of all the different molecules,” Blumberg told BioSpace. “We haven’t gotten to the point where the efficacy has been compelling enough to justify arguably life-threatening toxicities.”

Blumberg said these toxicities vary depending on the type of anti-CD47 drug. Monoclonal antibodies, for example, are often associated with conditions such as anemia and thrombocytopenia. This is because the antibodies bind to CD47 on healthy blood cells as well as cancer cells and induce macrophages to engulf and digest both the healthy and diseased cells.

Indeed, safety concerns have plagued Gilead’s magrolimab. In January 2022, the FDA placed a hold on trials of the candidate, which was being tested in multiple studies in combination with azacitidine, due to “an apparent imbalance” in suspected unexpected serious adverse reactions. The hold was lifted several months later, but the FDA issued a second partial clinical hold on magrolimab in acute myeloid leukemia (AML) in August. Additionally, a Phase Ib study found that 28.7% of AML patients treated with magrolimab plus azacitidine developed anemia related to the anti-CD47 drug. Gilead has also paused or discontinued multiple clinical trials of magrolimab in recent months.

“The efficacy data has been tantalizing,” Blumberg said, “but not really robust enough in light of the fact that you’re dealing with a drug class that is fraught with toxicity risks, and all the molecules have a nuanced toxicity. I don’t think it’s over for the class, but I do think that there’s headwind.”

A Field in Flux

Prior to the spate of disappointing results, the funding landscape for anti-CD47 immunotherapies looked rosy, buoyed by two multibillion-dollar acquisitions. In March 2020, Gilead acquired magrolimab’s manufacturer, Forty Seven, for $4.9 billion. Soon after, AbbVie entered into its agreement with I-Mab, and Pfizer purchased Trillium Therapeutics, the maker of two anti-CD47 proteins, for $2.22 billion.

The acquisitions that occurred alongside promising preclinical and clinical results resulted in the field being flooded with “a lot of money to chase maybe too many drugs going after the same target,” Blumberg said. “But there was nothing wrong with the pursuit.”

Some, like Lettmann, believe that many drugmakers unknowingly kneecapped their anti-CD47 antibody therapies by coupling the inhibition of the “don’t eat me” signal with the promotion of the “eat me” signal. “If you try to accomplish both as part of the same molecule, it will be very difficult to effectively target cancer over healthy [cells],” Lettmann said.

Magrolimab’s makers tried to skirt the issue by switching from one subclass of immunoglobulin G to another, he continued. Pfizer has similarly stuck to this subclass. In a statement, a Pfizer spokesperson told BioSpace that its anti-CD47 candidate maplirpacept has demonstrated minimal red blood cell binding and few reported cases of anemia. Maplirpacept acts on both cancer cells and macrophages, the spokesperson said, and has demonstrated “encouraging activity” when used as a monotherapy in studies with more than 75 patients. Currently, the company is conducting Phase I and II trials in ovarian cancer, multiple myeloma and diffuse large B-cell lymphoma.

ALX has approached drug development with safety in mind, Lettmann said. Evorpacept decouples anti-CD47 monoclonal antibodies’ two mechanisms of action, blocking the cancer cells’ “don’t eat me” signal without turning macrophages into indiscriminate killers—an approach that, when combined with other cancer therapies, has shown preclinical and clinical promise.

This month, ALX reported that 52% of patients with gastric/gastroesophageal junction (GEJ) cancer enrolled in a Phase II trial responded to a combination of evorpacept and trastuzumab, ramucirumab and paclitaxel, compared to 22% of patients who received only the three latter drugs. ALX plans to report a final analysis from the trial next year, along with a half-dozen readouts from trials of evorpacept in combination with chemotherapy, a checkpoint inhibitor and an antibody-drug conjugate.

Even so, ALX’s journey has not been without its bumps as the company recently axed CD47 programs where evorpacept attempted to treat myelodysplastic syndrome and acute myeloid leukemia—though Lettmann clarified that the mechanism tested in the two blood cancers differs from evorpacept’s mechanism of action in solid tumors.

For Lettmann, winning FDA approval for evorpacept would be the culmination of a nine-year journey for ALX and its leadership. Contingent on positive final results from the Phase II trial and the commencement next year of a Phase III trial for gastric/GEJ cancer, Lettmann said he believes the company has “a very logical path to an approval” in this indication.

Maddie Bender is a freelance science journalist based in Honolulu, HI. You can contact her through her website, maddiebender.com

Back to news