AstraZeneca's $1B acquisition of EsoBiotec's in vivo CAR T platform showed impressive efficacy (100% response rate, all MRD negative) but significant toxicity—all five trial patients experienced grade 3+ adverse events including cytokine release syndrome and one death attributed…
Every patient in the trial had a serious adverse event. That is the most important sentence in the Nature Medicine paper on AstraZeneca's experimental CAR-T therapy for multiple myeloma — and it is also the sentence most likely to get buried in the deal-story framing that follows a $1 billion acquisition.
AstraZeneca bought Belgium-based EsoBiotec for up to $1 billion ($425 million upfront, up to $575 million in milestones), announced in March 2025, betting that its in-vivo CAR-T platform could solve the manufacturing problem that has kept cell therapies out of reach for most patients. The idea is elegant: instead of extracting a patient's T cells, modifying them in a lab, and reinfusing them — a process that takes weeks and costs hundreds of thousands of dollars — ESO-T01 would be injected directly into the patient, reprogramming T cells inside the body. If it works, you eliminate the factory.
The first clinical data, published in Nature Medicine and reported by Inside Precision Medicine, came from five heavily pretreated male patients at a single site in Wuhan, China. All five experienced grade 3 or higher serious adverse events. Four developed cytokine release syndrome, three of them severe. One patient died: rapid neurological deterioration followed by spinal cord compression from an expanding tumor mass. The investigators attributed the death to disease progression, not the therapy.
What complicates the picture — and what makes this worth writing about rather than killing — is what showed up alongside the toxicity. Of four evaluable patients at the lowest dose, two had all signs of their multiple myeloma disappear completely and two showed partial responses; all four were MRD-negative in bone marrow by day 28. That is a response rate that would look extraordinary for a conventional drug in this population. In CAR-T, it is less surprising but still significant.
The hyperacute reaction in one patient, which occurred within hours of infusion, was severe enough to force a mid-trial protocol amendment. Subsequent patients received preemptive steroids. The trial stopped early in 2025 with no further enrollment. AstraZeneca has not disclosed whether a redesign is planned or whether the program is on hold.
The biology behind ESO-T01 is worth understanding because it is the reason AstraZeneca paid a premium. The therapy uses a lentiviral vector engineered with CD47 overexpression and MHC class I knocked out — an attempt to shield the viral payload from the patient's immune system long enough to reach T cells. This immune-shielding strategy is novel and the toxicity pattern suggests the shield may be incomplete, or that the immune activation from the vector itself is substantial even when the payload reaches its target. A 100 percent SAE rate at the lowest dose is a signal, not a verdict — five patients is not a trial — but it is a serious one.
The in-vivo CAR-T field is not monolithic. Umoja Biopharma, Interius BioTherapeutics (acquired by Kite/Gilead for $350 million in 2025), and Kelonia Therapeutics (collaborating with Johnson & Johnson on an anti-BCMA in-vivo program) are all pursuing variations of the idea. The field is watching AstraZeneca's data closely, precisely because the manufacturing advantage, if it holds at scale, would be transformative for patient access. Existing CAR-T therapies require custom manufacturing for each patient, with list prices in the hundreds of thousands of dollars per infusion. An in-vivo approach could theoretically cut both cost and time-to-treatment dramatically.
But the lesson from ESO-T01 is that the biology does not always cooperate with the business plan. The immune system is not a clean delivery corridor. Dodging it long enough to reprogram T cells and then surviving the immune activation those same T cells generate — that is a narrow path, and this first passage through it was rough.
For AstraZeneca, the question is not whether the data is discouraging. It is whether the efficacy signal justifies another pass. The company declined to comment on trial plans beyond what is in the published paper. Given the investment already made, walking away is expensive. But so is running a Phase II trial on a therapy where every patient so far has been hospitalized.
The death, at least as the investigators characterized it, was not the therapy's fault. The serious adverse events almost certainly were. That distinction will matter for regulators, for the trial design that follows, and for the patients with relapsed/refractory multiple myeloma who have run out of other options and are watching from the other side of a very high wall.