The Sliding Scale of In Vivo CAR T: When Chemistry Meets Reality

by James ColeScience and Modalities Editor, HMND

The past week offered a blunt reminder that in vivo CAR T is not a clean escape from ex vivo manufacturing, but a different manufacturing argument with harsher physics. The field is moving for real, with an IND for a relapsed and refractory multiple myeloma program and phase 1 data from inMMyCAR showing MRD negative responses in all four patients at one month, including a complete response in the first patient treated and deepening responses over time. That is meaningful progress, but it is still early human proof, not a solved platform.

What readers are tired of hearing

Senior engineering and R&D teams are not tired of the science. They are tired of the oversimplification. They know that "just deliver the payload in the body" is not a strategy, and they have lived long enough with translational systems to know where these stories break.

They are tired of hearing that in vivo CAR T will compress weeks into days without anyone saying what has to happen to the dose, the biodistribution, the durability, and the safety margin for that to be true. They are tired of the idea that moving the work from a clean facility into a human body somehow makes the work easier. It does the opposite. The manufacturing problem does not disappear. It just moves into delivery, targeting, and proof.

Why the adoption problem is real

The hard part is not the slogan. It is the biology and the transport layer. In vivo CAR T depends on getting genetic instructions into the right cells inside the body, at a dose that is high enough to matter but not so high that it becomes toxic. That sounds elegant in a slide deck. In practice, it means solving endosomal escape, cell specificity, payload efficiency, and immune compatibility at the same time.

That is where teams stall. Delivery systems can look promising in preclinical models and still fall apart in people because they do not reach enough target cells, or they reach the wrong ones, or the dose needed to overcome poor efficiency creates a toxicity problem. Pre existing antibodies can also interfere with vector delivery, which is a very ordinary way for a supposedly transformative platform to get stuck in the real world. When the approach is wrong, failure does not look dramatic at first. It looks like weak transduction, shallow responses, short durability, and a lot of expensive explanation.

What the recent data actually says

The most useful signal from the recent readout is not hype. It is that the early study reported a tolerability profile that did not collapse immediately, with two grade 3 cytokine release syndrome events, one grade 4 neutropenia event, and no ICANS reported. That matters because delivery platforms are often judged on exactly this tradeoff: can they hit hard enough to matter without turning systemic dosing into a safety problem.

The broader literature keeps circling the same constraints. Reviews of in vivo CAR T point to the same cluster of obstacles: targeting accuracy, biodistribution, immune response to the vector, and the need for a reliable proof package that regulators can trust. The promise is attractive because it could reduce the ex vivo burden, but the burden does not vanish. It shifts into a harder part of the system.

Why scale is still the bluntest test

This is where many therapies look good on slides and then break in the clinic. A platform can show a clean mechanistic story and even a small human response, but still fail when the engineering has to scale. If the vector dose has to keep climbing to get effect, the economics and toxicity both worsen. If specificity is not tight enough, the wrong cells get modified or destroyed. If the manufacturing or formulation path cannot be reproduced across batches, the platform becomes fragile before it becomes routine.

That is the part readers in pharma and biotech already know too well. The first patient is not the product. The first response is not a platform. The failure mode is not just "it does not work." It is that it works inconsistently, only in selected settings, at a cost and complexity profile that never really leaves the lab.

The real industry movement

The industry is trying to escape ex vivo bottlenecks through next generation delivery systems, including lipid nanoparticle based approaches, viral vectors, and other in vivo payload platforms. The point is not novelty for its own sake. It is to reduce the friction of patient specific cell handling, shorten the path to treatment, and make cell therapy feel less like a bespoke manufacturing operation and more like a deployable modality.

That shift is attractive precisely because ex vivo CAR T has made everyone fluent in what breaks at scale. But the new stack still has to earn trust in the body, not just in process diagrams. Regulators will not bless a concept because it is elegant. They need reproducible human evidence, a credible safety profile, and a delivery story that survives contact with actual physiology.

For now, in vivo CAR T is still best understood as a manufacturing argument wearing a therapeutic face. The science is moving, but the field is still paying for every shortcut it imagines.

If you are tracking this space closely, the interesting question is not whether the headline survives. It is which part of the stack fails first, and whether the teams building it are honest enough to name that early. That is usually where the useful conversation starts.