Does In Vivo CAR T Still Need Antibody Targeting? How ERTLNP Rethinks mRNA Delivery

If you look only at the therapeutic concept behind CAR T, it is actually quite intuitive: give a T cell a new recognition system so that it can find a specific target cell and attack it. What makes CAR T expensive, complex, and difficult to scale is often not the CAR itself. It is the fact that we first have to remove a patient’s T cells, genetically engineer them ex vivo, expand them, perform quality testing, and then infuse them back into the patient.

That is why I have been following in vivo CAR T closely over the past several years. The question is gradually shifting from “How do we make better CAR T cells?” to something more fundamental: can we avoid taking T cells out of the body altogether, and instead deliver a temporary program directly in vivo so that T cells become CAR T cells where they already are?

In Vivo CAR T: The Real Challenge Is Not the CAR, but Delivery

The core idea of in vivo CAR T is to deliver CAR-encoding mRNA or another genetic payload directly into the patient, reach T cells, and reprogram those cells inside the body. The appeal is obvious: if this can be done successfully, we may be able to eliminate some of the most complicated steps in today’s ex vivo CAR T manufacturing process.

But this immediately runs into a familiar problem in RNA therapeutics: delivery. Conventional intravenously administered LNPs tend to accumulate in the liver, but liver accumulation does not mean efficient entry into T cells. A common strategy has been to decorate nanoparticles with anti-CD3, anti-CD5, or other antibodies or ligands to pull the carrier toward T cells. Biologically, that makes sense. From a development perspective, however, it also adds complexity to manufacturing, analytics, and CMC.

What Makes ERTLNP Interesting: Reaching T Cells Without an Antibody Targeting Ligand

What caught my attention in this Nature Materials study is not simply that the authors made another LNP. ERTLNP challenges one of our assumptions about targeting: a carrier may not always need a surface antibody that precisely recognizes T cells in order to achieve useful T-cell delivery.

The researchers built ERTLNP using p-toluenesulfonyl arginine-modified oligoethylenimine (OEI-RT), together with DSPC, cholesterol, DMG-PEG, and mRNA. After systemic administration, the nanoparticles themselves were detected mainly in the liver and spleen, with stronger material-associated signals in the liver. Interestingly, functional mRNA translation showed a very different pattern: the 20% ERTLNP formulation produced strongly spleen-enriched expression, with 94.4% of major-organ bioluminescence detected in the spleen. This distinction is important because nanoparticle biodistribution does not necessarily predict where functional mRNA translation occurs. Instead of beginning with the question, “Which T-cell antibody should we put on the particle surface?”, this platform starts by engineering the material itself to influence biodistribution, cellular interactions, and functional expression.

More Importantly, ERTLNP Does More Than Deliver Cargo: It Changes T-Cell State

For me, the most interesting result is not the transfection percentage by itself. ERTLNP also induces T-cell activation. The study reported increases in CD69, CD25, and proliferation, together with changes in gene-expression programs associated with memory and persistence.

That distinction matters. We often design delivery platforms under the assumption that the carrier should be as biologically quiet as possible and simply deliver the payload. But immune cells are not passive containers. When the payload itself is intended to reprogram an immune cell, the biological response triggered by the carrier may be more than a side effect. It may become part of the therapeutic mechanism.

Why mTOR May Be Central to This Platform

The authors connected this behavior to mTOR signaling through two convergent activation mechanisms. One route begins with cell-surface IGF-1R engagement and proceeds through PI3K–AKT–mTOR signaling. A second route depends on ERTLNP uptake, with macropinocytosis contributing to internalization, followed by lysosomal sensing involving TM4SF5 and ERTLNP-derived arginine motifs, again converging on mTOR/S6 signaling. The delivery vehicle therefore begins to look less like packaging and more like a system that simultaneously controls cargo entry and the biological readiness of the target cell.

Placed back into the context of in vivo CAR T, the value becomes easier to see. It is not enough to deliver CAR mRNA into a T cell. Ideally, that T cell should also be in a state that supports CAR expression, proliferation, and therapeutic activity. Delivery efficiency and cell state may be much harder to separate than we often assume.

Generating CAR T Cells Directly in Mice: What Does the Study Actually Demonstrate?

The researchers loaded FAP CAR mRNA into ERTLNP, generated CAR-positive T cells in mice, and then evaluated therapeutic effects in disease models. In the spleen, CAR-positive CD3+ T cells were reported at about 7.7% on Day 1 and about 3.9% on Day 4. Because the payload is mRNA, CAR expression is inherently transient. Some may view that as a limitation, but I think transient expression may also be one of the potential advantages of in vivo programming.

Traditional viral vectors are often designed for durable expression. But if immune-cell programming eventually becomes repeat-dose therapy, with adjustable dosing and the ability to reprogram cells according to a patient’s changing disease state, transient expression provides a different kind of control. The study includes repeat-administration experiments, but these should not be interpreted as a systematic long-term safety assessment or chronic repeat-dose toxicology study. This connects directly with the areas I have been following across RNA therapeutics, in vivo CAR, and personalized medicine. The end product may not necessarily be a fixed cell therapy. It may instead become a therapeutic program that can be executed dynamically inside the patient.

From a CMC Perspective: Removing an Antibody Means More Than Removing One Material

If platforms like this eventually move toward the clinic, one issue I would watch very closely is manufacturability. Antibody-targeted LNPs can introduce additional conjugation steps, surface-density control, free-antibody removal, particle characterization, stability requirements, and comparability questions.

If material design can achieve sufficient immune-cell delivery without adding an antibody targeting ligand, the value may extend well beyond biology. It could directly affect CMC strategy, scale-up, and ultimately manufacturing cost. Of course, there is still a long distance between mouse studies and human therapy. Biodistribution, immune activation, therapeutic window, and repeated dosing will all need to be re-evaluated.

My Take: The Next Generation of Delivery May Be About Designing Cell State

When we talk about LNPs, it is easy to reduce the problem to three questions: can we encapsulate the payload, can we deliver it to the right place, and can we achieve enough expression? This study suggests that once the therapeutic target becomes a T cell, macrophage, or another immune cell, the next generation of delivery systems may need to ask one more question: after the nanoparticle enters the cell, what biological state does it push that cell toward?

That is why I think ERTLNP is more interesting than the simple statement that it can transfect T cells without an antibody. The platform brings targeting, delivery, cell activation, and therapeutic programming into the same material-design problem. If in vivo CAR T eventually reaches routine clinical use, the competition may not be about who has the best CAR construct. It may be about who can build a delivery platform that is controllable, manufacturable, repeat-dose compatible, and capable of precisely tuning immune-cell state.

The In Vivo CAR-T Landscape: Where Are We in 2026?

If we zoom out from this single paper to the broader industry, in vivo CAR-T is no longer only a preclinical concept in 2026. Two major technical directions are becoming increasingly clear. One uses targeted lentiviral vectors to pursue more persistent CAR expression. The other uses targeted mRNA-LNP systems to create transient, non-integrating, and potentially repeatable CAR expression.

Representative Clinical-Stage In Vivo CAR-T Programs in 2026

Below are selected clinical-stage or early-clinical in vivo CAR-T programs as of August 2026. I am presenting them program by program rather than in a wide table so the section remains readable in the Wix editor and on mobile devices.

KLN-1010 | Kelonia Therapeutics → Eli Lilly

Delivery: targeted lentiviral vector. CAR target: BCMA. Major indication: relapsed/refractory multiple myeloma. Clinical stage: Phase 1. KLN-1010 was originally developed by Kelonia Therapeutics. Lilly announced its acquisition of Kelonia in 2026, so I list the current ownership as Kelonia Therapeutics → Eli Lilly.

INT2104 | Interius BioTherapeutics → Kite / Gilead

Delivery: CD7-targeted lentiviral vector. CAR target: CD20. Major indication: relapsed/refractory B-cell malignancies. Clinical stage: Phase 1. One important distinction: CD7 is the delivery-targeting receptor, not the CAR target. INT2104 ultimately generates CAR-T and CAR-NK cells directed against CD20.

UB-VV111 | Umoja Biopharma / AbbVie Collaboration

Delivery: Umoja’s VivoVec engineered lentiviral vector. CAR target: CD19. Major indication: relapsed/refractory B-cell malignancies. Clinical stage: Phase 1. Umoja is the original developer and AbbVie is a collaboration partner, so I describe this as a partnership rather than an AbbVie-owned program.

CPTX2309 | Capstan Therapeutics → AbbVie

Delivery: CD8-targeted LNP carrying anti-CD19 CAR mRNA. CAR target: CD19. Major indications: rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). Clinical stage: Phase 1. This is an important benchmark for the mRNA-LNP route. The program was originally developed by Capstan Therapeutics, which was subsequently acquired by AbbVie.

GT801 | Grit Biotechnology

Delivery: targeted mRNA-LNP / T-LNP. CAR target: CD19. Major indications include B-cell malignancies and selected autoimmune diseases. Clinical stage: early Phase 1. Public information identifies Grit Biotechnology as the primary sponsor, with collaborators including Vivacta Biotechnology.

STR-P004 | Starna Therapeutics

Delivery: targeted RNA/LNP approach. CAR target: CD19. Major direction: autoimmune disease. Clinical positioning: exploratory / early clinical. Because publicly available studies include exploratory and investigator-initiated work, I think it is more accurate not to label STR-P004 simply as a conventional FDA-style Phase 1 program.

ESO-T01 | EsoBiotec → AstraZeneca

Delivery: targeted lentiviral vector. CAR target: BCMA. Major indication: relapsed/refractory multiple myeloma. Clinical stage: Phase 1. ESO-T01 was originally developed by EsoBiotec. AstraZeneca acquired EsoBiotec in 2025, so AstraZeneca should now be included in the program’s ownership.

This is not intended to be an exhaustive list. It highlights representative clinical-stage programs. Other programs worth following include UB-VV400, LB2501, HN2301, JY231, JCXH-213, and OriV508.

Integrating / Persistent Route

KLN-1010, INT2104, UB-VV111, and ESO-T01 represent the targeted viral-vector strategy. The major attraction of this route is that a single in vivo gene-transfer event may produce more persistent CAR expression. The tradeoff is that developers must also address integration, vector biodistribution, cell specificity, long-term follow-up, and manufacturing control.

Transient / Repeatable Route

Programs such as CPTX2309, GT801, and STR-P004 represent the RNA/LNP direction. mRNA does not integrate into the genome, so CAR expression is time-limited. That may require repeated dosing, but it also creates the possibility of controlling dose, exposure duration, and immune-cell depletion in a way that looks more like conventional pharmacology.

How I See the Current In Vivo CAR-T Landscape

I think the real competition is no longer simply about who can design the best CAR sequence. It is about who can control delivery pharmacology: which immune cells are reprogrammed, what fraction of those cells are modified, how long CAR expression lasts, what activation state the T cells enter, whether repeated dosing is possible, and whether the entire platform can ultimately support a scalable CMC and safety framework.

This is also why ERTLNP belongs in the broader landscape discussion. It is still a preclinical platform and should not be compared with clinical programs at the same maturity level. But it introduces another interesting possibility: if T-cell tropism and T-cell activation can be partially encoded into material chemistry rather than relying entirely on antibodies or other surface ligands, future platforms may be able to reduce the complexity associated with ligand conjugation, density control, particle heterogeneity, stability, and potency assays.

The landscape is also expanding quickly. Programs such as UB-VV400, LB2501, HN2301, JY231, JCXH-213, and OriV508 are worth watching. This section should therefore be read as a selected snapshot of the field in August 2026, not as a permanent or exhaustive list.

References / Further Reading

1. Cao Q, et al. An inherent T cell-activating mRNA delivery carrier for in vivo CAR T generation. Nature Materials (2026).

2. Rurik JG, et al. CAR T cells produced in vivo to treat cardiac injury. Science 375, 91–96 (2022).

3. Bot A, et al. In vivo chimeric antigen receptor (CAR)-T cell therapy. Nature Reviews Drug Discovery 25, 116–137 (2026).

4. Rafei H, Upadhyay R, Sharma P. A guide to CAR T cell therapies: development, current status and future prospects. Nature Reviews Immunology (2026).

5. Slaney CY, et al. Rewiring immunity with in vivo CAR T cell engineering. Nature Cancer (2026).

6. Ho PJ, et al. Updated results from inMMyCAR, the ongoing first-in-human phase 1 study of KLN-1010 in RRMM. Journal of Clinical Oncology 44, 7509 (2026).

7. Zhang X, et al. First-in-human in vivo CAR T-cell generation using a CD8-targeted lipid nanoparticle platform in R/R B-cell lymphoma. Journal of Clinical Oncology 44, 7014 (2026).

8. Liu Y, et al. First-in-human study of in vivo CAR-T therapy GT801 in adults with R/R CD19-positive B-cell hematologic malignancies. Journal of Clinical Oncology 44, 7079 (2026).

9. Pan J, et al. In vivo CD19 CAR T-cell therapy (STR-P004) to induce remission in autoimmune diseases. Journal of Clinical Oncology 44, e14519 (2026).

10. ClinicalTrials.gov. UB-VV111 Phase 1 study, NCT06528301.

11. ClinicalTrials.gov. CPTX2309 Phase 1 study in RA/SLE, NCT06917742.

12. ClinicalTrials.gov. STR-P004 exploratory clinical study, NCT07143617.

13. Interius BioTherapeutics. INT2104 / in vivo CAR pipeline.