In Vivo Site-Specific Large DNA Insertion for Efficient CAR-T Cell Generation

Engineered T cells, modified to express Chimeric Antigen Receptors (CAR) or T Cell Receptors (TCR), have revolutionized cancer treatment and are currently being explored for the treatment of autoimmune and infectious diseases. Gene editing to enhance T cell function—whether through the disruption of endogenous genes or the precise insertion of DNA payloads—has demonstrated significant potential.

However, the current ex vivo manufacturing process is lengthy and expensive, limiting the accessibility of these therapies. Generating CAR-T cells in vivo could overcome these obstacles, but existing methods rely either on transient expression with limited persistence or on random integration of DNA payloads that lack specificity.

Recently, a collaborative study by the teams of Justin Eyquem and Jennifer A. Doudna at the University of California, San Francisco, was published online in Nature, titled "In vivo site-specific engineering to reprogram T cells". This research demonstrates that stable and cell-specific transgene expression can be achieved through site-specific integration of large DNA payloads in vivo. The researchers developed a dual-vector system using an enveloped delivery vehicle (EDV) and an adeno-associated virus (AAV) to deliver the CRISPR-Cas9 nuclease complex and the DNA template, respectively. Both vectors were optimized to achieve specific delivery to T cells and high gene-targeting efficiency.

By transducing a CAR gene into a T-cell-specific locus, the study successfully generated therapeutic levels of CAR-T cells in vivo within humanized mouse models of B-cell deficiency, as well as in hematologic and solid tumor models. These findings pave the way for more efficient, precise, and accessible T-cell therapies.

CAR-T cells represent a promising approach for treating hematologic malignancies; to date, the U.S. Food and Drug Administration (FDA) has approved seven CAR-T cell therapies. Standard CAR-T therapy requires personalized production for each patient, which is limited by inconsistent product quality, long production cycles, and high costs.

Typically, CARs are delivered via retroviral vectors, leading to variability in expression results due to random integration. By using CRISPR-Cas9 and adeno-associated virus (AAV)-mediated homology-directed repair (HDR), the CAR can be targeted for integration into the native human TCRα gene locus (TRAC).

TRAC-CAR T cells exhibit dynamic CAR expression, which can delay cell exhaustion and improve tumor control in xenograft and immunocompetent models. This work is crucial for the development of allogeneic CAR-T cell therapies, as it disrupts the TCR upon transgene insertion—a necessary step to limit Graft-versus-Host Disease (GvHD).

Clinical trials using allogeneic TRAC-CAR T cells derived from healthy donors or induced pluripotent stem cells (iPSCs), combined with intensive lymphodepletion conditioning, have achieved complete remission in patients with hematologic malignancies. Allogeneic therapies can address manufacturing limitations by creating "off-the-shelf" drugs from healthy donors. However, allogeneic CAR-T cells are eventually rejected by the body, and frequent relapses have been observed.

Generating CAR-T cells directly in vivo may bypass the various hurdles encountered during leukapheresis and the manufacturing process. It also has the potential to promote the formation of a less-differentiated CAR-T cell population, a trait associated with enhanced anti-tumor activity.

To date, efforts to generate CAR-T cells in vivo have utilized randomly integrating viral vectors for sustained CAR expression or lipid nanoparticles (LNPs) for transient expression. Both methods were recently validated in non-human primates and evaluated in a Phase I clinical trial. These approaches face several challenges, including how to efficiently deliver genes to therapeutic doses while avoiding the risk of off-target transduction.

Figure 1. Co-delivery of Cas9-EDV and HDRT-AAV generates TRAC-CAR T cells in vitro and in vivo. (NYBERG, William A., et al., 2026)

Both delivery and CAR expression should be T-cell specific, as off-target modification of hematopoietic stem cells (HSCs) could lead to mutagenic transformation. Furthermore, CAR expression in tumor cells could prevent the surface expression of target proteins, leading to antigen-negative relapse. LNP delivery of CAR mRNA results in transient expression, which prevents insertional mutagenesis or stable expression in tumor cells, but the required dosage remains unclear.

While the envelopes of lentiviral vectors can be engineered to improve T-cell specificity, any non-T cell that is transduced would also express the CAR—posing a potential risk of insertional mutation unless a cell-lineage-specific promoter is used. The researchers hypothesized that integrating a promoterless CAR transgene into the TRAC locus in vivo would enable T-cell-specific and physiological CAR expression while avoiding the ex vivo manufacturing process. Until now, the precise in vivo integration of large DNA payloads in human T cells had not been achieved.

This study developed a method combining AAV with an enveloped delivery vehicle (EDV) for site-specific transgene integration in primary human T cells in vivo. By optimizing the AAV and EDV tools to improve cell-specific delivery efficiency and enhance resistance to human neutralizing antibodies, the researchers were able to generate therapeutic levels of TRAC-CAR T cells in vivo and control tumor growth in multiple humanized mouse models.