In Vivo CAR-T Engineering: The Emerging Role of mRNA–LNP Technologies

Introduction

Cancer immunotherapy has long been limited by a fundamental biological constraint: T cells can only recognize disease when antigens are presented via the Major Histocompatibility Complex (MHC). However, cancer cells frequently evade immune detection by downregulating MHC expression or altering antigen presentation pathways, rendering conventional T cell responses ineffective.

The emergence of Chimeric Antigen Receptor T-cell (CAR-T) therapy marked a revolutionary shift in oncology. By engineering CAR T cells with synthetic receptors capable of recognizing tumor surface antigens independently of MHC, CAR T cell therapy has achieved unprecedented remission rates in refractory hematological malignancies [1]. Yet, despite these breakthroughs, the full potential of CAR T therapy remains underutilized due to the complex, time-consuming, and costly ex vivo viral manufacturing processes required.

Recent advances in mRNA and lipid nanoparticles (LNPs) are now paving the way for a transformative approach : in vivo programming of CAR T cells through a fully non-viral platform. This shift could democratize access to CAR T cell therapy, reduce production timelines, and expand therapeutic possibilities beyond current limitations.

This review explores how mRNA-LNP technologies are redefining CAR T therapy, from overcoming MHC restrictions to enabling scalable, in vivo immune cell engineering.

1. From MHC Biological Constraints to CAR Engineering Solution and Limits

A. CAR as the Solution to MHC-Restriction

Under physiological conditions, T-cell activation requires peptide presentation via MHC molecules. Tumor cells often exploit this dependency by downregulating MHC class I expression or modifying antigen processing, leading to immune escape.

CAR technology bypasses this restriction. By combining an antibody-derived antigen-recognition domain with intracellular signaling domains, CAR constructs enable direct surface antigen recognition independent of MHC. Since one of the first CAR designs by Eshhar et al. [2], successive generations have incorporated co-stimulatory domains to enhance persistence, cytokine production, and cytotoxicity. Figure 1 gives an overview of the CAR design evolution.

The biological breakthrough was clear. The challenge became engineering: how to efficiently introduce CAR constructs into T cell membranes?

B. The Viral Ex Vivo CAR-T Approach: Clinically Powerful, Structurally Constrained

The dominant clinical approach to CAR-T engineering has so far relied on lentiviral transduction within a five-step ex vivo workflow:

1. Leukapheresis
2. Viral transduction (commonly lentiviral vectors)
3. Ex vivo expansion
4. GMP quality control
5. Reinfusion

While clinically powerful, this model is constrained.

• Manufacturing Timeline

Production typically requires three to four weeks.

• Infrastructure Dependence

Advanced GMP facilities and cold-chain logistics are mandatory.

• Cost

Treatment costs often exceed $400,000 USD per patient and may surpass $1 million [5].

These limitations have sparked strong interest in finding ways to bypass the external manufacturing steps by generating CAR T cells directly in vivo.

C. The Interest of In Vivo Solution Through the Viral Approach

Viral vectors have been explored for direct in vivo CAR engineering. Adeno-associated viruses (AAV) and lentiviral systems can deliver CAR constructs directly into immune cells within the patient, enabling long-term genomic integration and sustained expression.

However, viral in vivo strategies remain constrained by several factors:

• Risk of insertional mutagenesis (mainly for lentiviral vectors)
• Limited cargo capacity (particularly for AAV)
• Pre-existing anti-viral immunity
• Complex vector manufacturing

These limitations illustrate that in vivo delivery does not inherently eliminate the challenges associated with viral systems.

This is where mRNA and Lipid Nanoparticle (LNP) technologies emerge as transformative in vivo & non-viral platforms.

2. How RNA-LNP Technology Opens the Possibility of an In-Vivo CAR-T Approach

A. mRNA as a Transient Genetic Instruction

Messenger RNA (mRNA) is a transient genetic intermediary that directs protein synthesis in the cytoplasm without altering the host genome. Once delivered into cells, it is translated by ribosomes and naturally degraded within hours to days. The therapeutic relevance of this mechanism was clinically validated by the rapid development and global deployment of mRNA vaccines against SARS-CoV-2, which demonstrated robust and controllable in vivo protein expression [6].

In the CAR-T context, mRNA can encode the CAR construct itself. Importantly, CAR activity depends on receptor expression—not necessarily permanent genomic integration. Transient expression may provide sufficient anti-tumor activity while limiting long-term toxicity. Several RNA modalities may be considered for this application, including conventional mRNA and self-amplifying RNA (saRNA). For a detailed comparison between conventional mRNA and self-amplifying RNA (saRNA), please refer to our dedicated reviews.

However, mRNA presents two challenges. First, it must reach the cytoplasm to be translated. Second, it is highly fragile and is rapidly degraded by extracellular nucleases when exposed outside of the cell [7]. This is where LNP technology plays a crucial role, acting as a nanovector solution for RNA.

B. Lipid Nanoparticles as Enablers of In Vivo RNA Delivery

Since the COVID-19 pandemic, LNP technologies have gained widespread attention. But what defines their structure and function?

Lipid nanoparticles are composed of 4 different lipids:

• Ionizable lipids
• Helper phospholipids
• Cholesterol
• PEG-lipids

If you want a deep dive into the role of each lipid, please have a look to our dedicated reviews, “A complete guide to understanding Lipid nanoparticles (LNP)”. The role of LNP is to encapsulate and protect mRNA, facilitate cellular uptake, and enable endosomal escape. Compared with viral vectors, they do not induce genomic integration, which significantly reduces their potential toxicity. We also detailed here the use of LNP for RNA delivery.

Their specific application for targeted in vivo T-cell engineering was demonstrated in 2022 by Rurik et al [8]. In this paper, scientists administered in vivo RNA-LNP to a mouse model of cardiac fibrosis, and they showed that:

• Modified mRNA encoding a CAR was selectively delivered to T cells
• Functional CAR-T cells were generated transiently in vivo
• Cardiac fibrosis was reduced
• Heart function was restored

This study provided an important proof of concept that CAR-T cells can be generated in vivo using a fully non-viral approach. This led us to the following question: what could be the benefit of this approach? And what are its actual limits?

3. Strengths and Limitations of mRNA-LNP Platforms for In Vivo CAR-T Engineering

A. RNA-LNP vs Viral Vector Technology

RNA-LNP technology offers several potential advantages (see Table 1 below).

Table 1. A comparative framework: Viral vs mRNA-LNP CAR engineering.

Importantly, as Aranud Deladeriere explains in his recent article “In Vivo’s Biggest Threat — Comparison To Old Models” the distinction between viral and non-viral platforms should not be reduced to a question of technological superiority.

In oncology, where long-lasting CAR expression is needed, lentiviral vectors with genomic integration remain advantageous despite the uncertainties linked to permanent insertion. In contrast, mRNA‑LNP platforms providing transient, non-integrative expression make them better suited for indications requiring controlled and reversible immune modulation, such as autoimmune or inflammatory diseases. Today, the real question is not which platform is superior, but which expression duration best fits the biology of the target condition.

Although the field is advancing rapidly, RNA–LNP systems still present significant limitations and offer considerable room for further development. The following points highlight the most relevant constraints observed today.

Looking to get start with mRNA-LNP formulation?
Discover our LNP Formulation system

Discover

B. The Current Limitations of RNA–LNP Technologies for In Vivo CAR‑T Therapy

Despite their promise, in vivo mRNA-LNP CAR-T strategies still face several technical challenges:

• Achieving precise T-cell targeting in oncology
  Infiltrating T cells remains difficult due to biodistribution constraints and significant off-target uptake, particularly in the liver and spleen.
• Overcoming inefficient endosomal escape
  Although LNPs promote cellular uptake, only a small fraction of internalized mRNA reaches the cytoplasm (a few %), limiting functional CAR expression.
• Balancing transient expression with therapeutic durability
  mRNA-encoded CARs are inherently short-lived. This improves safety but may reduce long-term anti-tumor activity in aggressive malignancies.

Additional constraints—such as PEG-related immune responses or intellectual property restrictions surrounding ionizable lipid chemistries—also shape the current landscape. These limitations, being primarily technical rather than conceptual, have stimulated substantial innovation across LNP engineering.

C. Beyond Oncology: Why Fibrotic and Autoimmune Diseases May Be Strong Use Cases for mRNA–LNP In Vivo CAR Engineering

While most discussions of CAR-T innovation remain anchored in oncology, autoimmune disease may represent one of the most compelling future applications for in vivo mRNA–LNP CAR engineering. In several refractory autoimmune diseases, CAR-T strategies have been used to target pathogenic B-cell compartments, particularly via CD19-directed depletion, with the aim of eliminating autoreactive populations and enabling reconstitution of a more tolerant immune repertoire.
This “immune reset” concept differs from the classic oncology objective of persistent tumor surveillance. As a result, the transient and controllable nature of mRNA expression may be particularly attractive in autoimmune settings, where excessive persistence could increase toxicity while reversible activity may be sufficient to induce durable remission

D. Advances in LNP Targeting: Overcoming Biodistribution Challenges

Among these innovation areas, targeting has emerged as a particularly dynamic field of development. Table 2 below outlines key targeting strategies currently employed to refine biodistribution and enhance delivery precision.

Table 2. LNP targeting strategies.

You can find more information on LNP targeting strategies in our dedicated review.

More globally, the substantial room for improvement has fueled growing scientific interest, driving rapid innovation and accelerating efforts to develop more efficient RNA–LNP solutions.

E. In Vivo CAR-T RNA-LNP Approaches Drive the Scientific Interest

As described in Deborah Day Barbara’s article, recent analyses highlight a sharp acceleration in the adoption of non-viral RNA technologies, with nanoparticles now representing nearly half of all non-viral delivery approaches entering the CAR pipeline in 2024. This momentum is reinforced by the diversification of targeted immune cells—extending beyond T cells to macrophages and NK cells—and by high-profile collaborations such as Moderna and Carisma’s in vivo CAR‑macrophage programs, which have already shown promising preclinical activity in solid tumor models. Major pharmaceutical companies, including Sanofi, are now advancing multiple in vivo CAR‑T candidates and reporting encouraging preclinical data using LNP systems with high transfection efficiency and minimal systemic toxicity. Early clinical signals, such as those from MT‑303 in advanced liver cancer, further support the feasibility and safety of mRNA‑LNP in vivo CAR therapies. Together, these developments illustrate a rapidly expanding ecosystem, strong industrial investment, and growing clinical validation—all of which position mRNA‑LNP platforms as a transformative and increasingly credible alternative within the future landscape of in vivo cell engineering.

These considerations naturally raise a broader question: where do mRNA‑LNP in vivo strategies sit within the expanding ecosystem of CAR engineering?

Sum Up: The Landscape of CAR Engineering Strategies

The Rossi–Breman framework is valuable not only because it maps existing CAR‑T strategies, but because it reveals the strategic trade-offs that differentiate them: durability versus control, manufacturing complexity versus scalability, and integration versus transient expression.

Positioned at the intersection of in vivo delivery and non-viral engineering, mRNA‑LNP platforms own a strategic place in this landscape. Explaining why it triggers as much enthusiasm and expectation in the field. Based on the Beacon Intelligence analysis, LNP-based in vivo CAR therapies accounted for nearly one-third of newly disclosed assets in 2024, a proportion that is expected to increase as RNA delivery platforms continue to mature [12].

Conclusion & Future Directions

The evolution of CAR-T therapy illustrates that no single engineering model is universally superior. While ex vivo viral platforms have demonstrated strong clinical efficacy, in vivo approaches—both viral and non-viral—are rapidly advancing. There is currently no perfect model: neither in vivo nor ex vivo strategies can be considered inherently superior, and viral and non-viral systems each present distinct strengths and limitations. The optimal approach will likely depend on the therapeutic context, disease biology, and required duration of expression.

At the same time, the field is clearly shifting toward increased exploration of in vivo cell engineering. Growing numbers of preclinical and early clinical programs are investigating mRNA-based platforms, not only for T cells but also for other immune populations such as NK cells and macrophages. The landscape is expanding beyond traditional T-cell–centric paradigms, reflecting a broader vision of programmable immune modulation.

mRNA-LNP technologies are central to this transformation. By enabling transient, modular, and potentially repeatable immune programming, they offer a flexible alternative to permanent genomic modification. Beyond CAR engineering, the same technological foundation is also helping expand adjacent modalities such as cancer vaccines and other programable immunotherapies. Rather than replacing viral systems, RNA-based approaches may complement them, broadening the therapeutic toolbox.

The future of CAR engineering is therefore unlikely to be defined by a single dominant platform. Instead, it will be shaped by context-dependent strategies integrating viral and non-viral tools, ex vivo and in vivo models, and increasingly diverse immune cell targets. In this evolving landscape, mRNA-LNP systems stand as a powerful enabler of next-generation programmable cell therapies.

References

[1]       J. Lu and G. Jiang, “The journey of CAR-T therapy in hematological malignancies,” Dec. 01, 2022, BioMed Central Ltd. doi: 10.1186/s12943-022-01663-0.

[2]       Z. Eshhar, T. Waks, G. Grosst, and D. G. Schindler, “Specific activation and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the y or C subunits of the immunoglobulin and T-cell receptors,” 1993. [Online]. Available: https://www.pnas.org

[3]       A. C. Uscanga-Palomeque et al., “CAR-T Cell Therapy: From the Shop to Cancer Therapy,” Nov. 01, 2023, Multidisciplinary Digital Publishing Institute (MDPI). doi: 10.3390/ijms242115688.

[4]       I. Zugasti et al., “CAR-T cell therapy for cancer: current challenges and future directions,” Dec. 01, 2025, Springer Nature. doi: 10.1038/s41392-025-02269-w.

[5]       E. Pinto et al., “From ex vivo to in vivo chimeric antigen T cells manufacturing: new horizons for CAR T-cell based therapy,” Dec. 01, 2025, BioMed Central Ltd. doi: 10.1186/s12967-024-06052-3.

[6]       P. F. P et al., “Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine,” New England Journal of Medicine, vol. 383, no. 27, pp. 2603–2615, Dec. 2020, doi: 10.1056/NEJMoa2034577.

[7]       N. Pardi, M. J. Hogan, F. W. Porter, and D. Weissman, “mRNA vaccines-a new era in vaccinology,” Mar. 28, 2018, Nature Publishing Group. doi: 10.1038/nrd.2017.243.

[8]       J. G. Rurik et al., “CAR T cells produced in vivo to treat cardiac injury.” [Online]. Available: https://www.science.org

[9]       L. Yue, X. Gao, W. Qi, L. Zhang, and Y. Wang, “Optimizing the targeting of lipid nanoparticles for gene therapy,” Nanoscale Horiz., vol. 11, no. 2, pp. 334–356, 2026, doi: 10.1039/D5NH00351B.

[10]     X. Wang et al., “Hybrid Extracellular Vesicles for Efficient Loading and Functional Delivery of mRNA,” J. Extracell. Vesicles, vol. 14, no. 12, Dec. 2025, doi: 10.1002/jev2.70201.

[11]     M. Rossi and E. Breman, “Engineering strategies to safely drive CAR T-cells into the future,” 2024, Frontiers Media SA. doi: 10.3389/fimmu.2024.1411393.

[12]     “In Vivo CAR Therapies Report.” Accessed: Mar. 05, 2026. [Online]. Available: https://beacon-intelligence.com/market-reports/in-vivo-car-therapies/