What you need to make in vivo CAR-T
Reprogramming T cells directly in the body sounds straightforward, but in practice, it’s far from a single technology problem.
First, there is the genetic payload, typically mRNA, which encodes the CAR itself. This allows T cells to be engineered directly in vivo, but the constraint is that expression is transient. In early proof-of-concept work, mRNA delivered to T cells in vivo produced functional CAR-T cells, but only for a limited time window.
Then comes the delivery vehicle, most often lipid nanoparticles (LNPs), which protect the mRNA and enable it to enter cells. These systems have become widely used, but without additional targeting, they tend to accumulate in organs like the liver, which limits their usefulness for immune cell engineering.
To address that, a targeting or guidance layer is needed. This is typically achieved by decorating nanoparticles with ligands that direct them toward specific T cell populations. For example, targeted LNPs carrying mRNA have been shown to generate CAR-T cells in vivo by selectively delivering payloads to T lymphocytes.
To build its own in vivo CAR-T platform, Sanofi went the acquisition route, putting together these technologies over the years.
Sanofi’s strategy: mix and match
Large pharmaceutical companies rarely build therapeutic platforms from scratch. Instead, capabilities tend to accumulate over time, through acquisitions that are not necessarily initially designed to fit into a single system. Sanofi’s in vivo CAR-T platform is a good example of how those pieces can later converge.
One of the earliest elements came with the acquisition of Ablynx in 2018. The deal, worth €3.9 billion, was not positioned around cell therapy, but it gave Sanofi access to a large library of nanobody-based binders, small antibody fragments. Their size and stability make them particularly well-suited for targeting hard-to-reach epitopes and for incorporation into more compact or multispecific constructs.
“Nanobodies can play an important part for Sanofi, as they provide optimal LNPs decoration, compact mRNA payloads, and the ability to target hard-to-reach epitopes. This is valuable beyond CAR-T, and Sanofi can use and combine them like Lego blocks,” explained Samara Rocha Sousa, researcher at the University of Helsinki, focused on sherpabodies, antibodies half the size of nanobodies.
A second layer came with the acquisition of Translate Bio in 2021 for $3.2 billion. Translate had developed an end-to-end mRNA platform, including manufacturing capabilities, initially applied to vaccines and rare diseases. Beyond the payload itself, the company also brought experience with lipid nanoparticle formulation.
Just months earlier, Sanofi had acquired Tidal Therapeutics, a biotech specifically focused on in vivo reprogramming of immune cells using targeted nanoparticles. Tidal’s approach was built around directing mRNA payloads toward selected immune cell populations, including T cells, using surface targeting strategies, an attempt to move beyond the non-specific biodistribution typically seen with standard LNP systems.
“In the paper Sanofi published in Molecular Therapy, you see how it brings all pieces together. Tidal provided the intellectual property and knowledge on targeted LNP formulations, and Translate Bio provided the mRNA capability. The nanobodies came with the acquisition of Ablynx in 2018. The Sanofi paper summarises how all three techs together were necessary and, interestingly, sufficient,” said Rocha.
This platform was not built as a single, coherent programme from the outset. It emerged from the integration of capabilities acquired over several years, and it is only when combined that these technologies begin to resemble a functional in vivo CAR-T platform. And the platform is starting to show.
How does Sanofi’s platform perform?
In February, Sanofi published preclinical data on its in vivo CAR-T approach, offering a first look at how these technologies perform once combined, and suggesting the platform does work, at least in a controlled setting.
Sanofi was able to deliver mRNA encoding a CAR selectively to CD8+ T cells and generate functional CAR-T cells directly in vivo. This translated into anti-tumour activity, with reprogrammed T cells inhibiting tumour growth.
A persistent limitation of LNP-based systems is their tendency to accumulate in the liver, which constrains dosing and reduces effective targeting. In this case, the formulation was engineered to minimise off-target expression and improve specificity, with low levels of liver uptake and more selective transfection of T cells.
Rocha pointed out that this is significant as reducing liver accumulation directly expands the therapeutic window, making repeated dosing more feasible and bringing the approach closer to a usable system.
However, because the platform relies on mRNA, CAR expression remains transient, lasting only a few days in preclinical models. That’s the trade-off with in vivo CAR-T: transient expression avoids the risks of permanent genetic modification but raises uncertainty about the durability and sustainability of the benefits.
Also, the results remain confined to hematological models, where targeting circulating T cells is more straightforward. Extending this approach to other settings may introduce additional challenges.
Beyond Sanofi: is this a model?
Building an in vivo CAR-T platform such as Sanofi’s, internally, from scratch, would mean investing for years across mRNA design, delivery chemistry, targeting biology, manufacturing, and cell therapy expertise. Acquisitions offer a route where technologies have already survived part of that filtering, then try to integrate them at scale.
Large pharma, Rocha argued, often waits for universities and startups to “develop the key ideas” before acquiring and accelerating them toward development. That model is not specific to Sanofi, and recent deals suggest that in vivo cell engineering has become one of the areas where pharma is willing to buy platform capability rather than slowly build it: Lilly agreed to acquire Kelonia Therapeutics for up to $7 billion in April, as part of a wider wave of recent in vivo CAR-T deals involving Gilead buying Interius BioTherapeutics for $350 million, AbbVie acquiring Capstan Therapeutics for $2.1 billion, Bristol Myers Squibb Orbital Therapeutics for $1.5 billion, and AstraZeneca spending $1 billion on EsoBiotec.
It’s not the individual building blocks that are the main constraint; mRNA platforms, delivery systems, and targeting technologies already exist. The difficulty is making them work together. Sanofi’s platform is one attempt at that integration, and early data suggest it might become a successful one.
Rocha also noted that, given the current position of the U.S. on the biotech scene, it’s increasingly important for Europe to be aware of this platform-building trend. “If you want to make sure your public investment in life sciences reverts fully to EU citizens, you need to look at the whole process: invest in basic research, stimulate university spin offs, provide money for growth, develop a venture capital ecosystem, have big companies headquartered on the continent, or else big U.S. pharma will ultimately own the patents, the drugs and the profits.”
Editor’s note: This article was originally published in European Biotechnology Magazine Summer 2026.