Glioblastoma’s hidden driver: CD47’s new role beyond immunity

Glioblastoma’s reputation as the most aggressive brain cancer isn’t just about its speed. It’s about how relentlessly it outmaneuvers every treatment thrown at it. Now, researchers at Australia’s Centre for Cancer Biology (CCB) have uncovered a previously unknown way the disease sustains itself—and it hinges on a protein already familiar to immunologists: CD47.

The study, published in Proceedings of the National Academy of Sciences, reveals that CD47 doesn’t just help tumors evade the immune system, as prior research showed. It also directly fuels glioblastoma’s growth and spread through a separate mechanism. In lab models, blocking this pathway slowed tumor progression—a finding that, if replicated, could open a new therapeutic angle.

But here’s the critical context: This is preclinical work. The experiments used cell lines and mouse models, not human patients. The sample size and methodology (detailed in the PNAS paper) are rigorous for this stage, but they don’t yet answer whether the same effect holds in people—or whether blocking CD47 in humans would be safe enough to outweigh risks like anemia, a known side effect of earlier CD47-targeting drugs.

For patients and clinicians, the question isn’t whether this is a breakthrough—it’s whether it’s even relevant yet. The answer, for now, is no. No CD47 inhibitors are approved for glioblastoma, and the few in trials (like Magrolimab) target its immune-evasion role, not this newly identified pathway. The CCB team’s work suggests a potential dual mechanism, but translating that into a therapy would require years of testing to confirm efficacy and rule out off-target effects.

What’s more, glioblastoma’s heterogeneity means a single-protein target rarely suffices. Past failures—like the disappointment of bevacizumab in extending survival—underscore how easily promising lab results falter in real-world tumors. The CD47 finding is biologically intriguing, but its clinical value hinges on two unknowns: whether the protein’s tumor-driving role is consistent across glioblastoma subtypes, and whether blocking it doesn’t trigger compensatory growth pathways.

The study’s lead authors are clear-eyed about the gaps. As they note in their university statement, this is ‘a step toward understanding,’ not a treatment. The next step? Validating the mechanism in patient-derived models—then, perhaps, designing inhibitors tailored to this specific function of CD47.