In vivo CAR-T engineering: insights from the 2025 ASH annual meeting

Abstract

In vivo chimeric antigen receptor (CAR)-T engineering is a prominent field of research in cancer immunology, in which vectors are used to reprogram endogenous T-cells into CAR-T cells. Viral and nonviral platforms presented at the 67th Annual Meeting of the American Society of Hematology (ASH) offer precise T-cell targeting and efficient CAR expression. Preclinical models demonstrated robust generation of CAR-T cells, potent tumor clearance, and strong clinical translational potential. However, certain limitations remained to be addressed in future studies. This correspondence summarizes the key findings from the meeting, discusses current translational challenges, and highlights future directions.

To the editor,

In vivo CAR-T cells offer a promising alternative to the limitations of ex vivo CAR-Ts, including complex manufacturing, high costs, safety concerns related to lymphodepleting chemotherapy, and the risk of cytokine release syndrome (CRS). The 67th Annual Meeting of the ASH showcased groundbreaking advancements in preclinical studies toward the real-world application of in vivo CAR-Ts; therefore, this correspondence summarizes the most significant findings presented at the meeting.

Emergence of in vivo CAR-T engineering

In vivo CAR-Ts have advantages over ex vivo applications, but certain challenges persist. To mitigate these obstacles, viral and non-viral platforms have recently emerged (Table 1).

Table 1.

In vivo CAR-T engineering platforms

Platform	Delivery	Target cells	CAR targets	Features	Solutions to in vivo CAR-T challenges	Ref
Viral platforms
Qihan	Binder–fusogen combinatorial LV	Resting T cells (e.g., CD3, CD7, CD8)	Tumor antigens	Silent entry, minimal cytokine secretion	Decouples T cell activation from viral entry to prevent cytokine release and exhaustion	[1]
IASO	Engineered third-generation lentiviral vector	CD7+ T cells	CD19/CD20/BCMA	Precise targeting, CD47 expression and the knocking out of MHC I gene-enhanced persistence	Enhances vector persistence and reduces innate immune clearance in vivo	[4]
IMV101	Mutated MxV-pseudotyped lentiviral vector + TCM3 module	T cells	CD19	High T cell specificity, minimal epithelial/tumor cell infection, strong in vivo expansion	Improves T cell specificity while minimizing non-immune and tumor cell transduction	[3]
LV-169	CD3-targeted lentiviral vector (detargeted VSV-G + CD3-scFv)	Resting niche-resident T cells	BCMA	Low tonic signaling, no CAR incorporation into viral particles, efficient delivery to lymphoid tissues	Avoids off-target CAR delivery and tonic signaling, enables transduction of resting T cells in lymphoid niches	[2]
Non-viral platforms
Velvet	STAR nanoparticles + mcDNA	Immune cells	CD19	Clathrin uptake, repeat dosing, minimal off-target accumulation in the liver	Enables repeat dosing with reduced hepatotoxicity	[5]
Grit (GT801)	T-LNPs with VHH antibodies using CLAMP conjugation	T cells	CD19	Site-specific antibody conjugation with controlled ligand density, low off-targeting	Improves targeting precision while minimizing off-target uptake	[6]
Tessera	tLNP + RNA Gene Writer/CAR mRNA	HSCs, T cells	CD19/CD20	24–55% CAR+ T cells in vivo, Up to 75% HSC editing; extrahepatic delivery	Enables non-viral, permanent or programmable genome editing in vivo	[8, 9]
Aera	tLNP delivering CAR mRNA	CD8+ T cells	CD19	Transient CAR, immune reset	Prevents prolonged CAR persistence and chronic B cell aplasia	[7]
Orna (panCARoRNA)	oRNA + immunotropic mLNP	T cells, immune effector cells	CD19/ BCMA	Tunable CAR, Selective plasma cell depletion	Allows transient, re-dosable CAR expression without lymphodepletion	[10, 11]

BCMA B cell maturation antigen, CAR Chimeric Antigen Receptor, CLAMP Controllable Ligand Attachment Modification and Purification, HSC Hematopoietic Stem Cells, LV Lentiviral vector, mcDNA minicircle DNA, mLNP mRNA-lipid nanoparticles, T-LNP T cell-targeted lipid nanoparticles, tLNP targeted lipid nanoparticles

Viral platforms

Viral vectors offer high efficiency for T-cell transduction; nevertheless, they can cause undesired T-cell activation, off-target gene transfer, and long-term genomic integration. Wang et al. reported that the threshold of T-cell activation can be adjusted for efficient transduction in a quiescent state. Upon antigen exposure, these cells exhibit robust CAR-dependent proliferation [1]. Combining CD3-targeted lentiviral delivery with VSV-G pseudotyping enables CAR gene delivery to resting niche-resident T-cells, reducing off-target transduction [2]. IMV101 employs a mutated MxV-pseudotyped lentivirus with a T-cell targeting module, TCM3, achieving high T-cell specificity and robust in vivo expansion [3]. IASO’s third-generation lentiviral vectors integrate engineered binder-fusogen pairs, CD47 expression, and MHC-I knockout to extend vector persistence, reduce immune clearance, and enable precise transduction of CD7⁺ T-cells with CARs against BCMA (IASO 206), CD20 (IASO 208), and CD19 (IASO 207) [4]. These strategies hold promise, but they encounter challenges related to repeat dosing and genomic integration errors.

Non- viral platforms

Non-viral platforms can be administered repeatedly and are highly safe. The STAR-CAR mcDNA platform uses multi-arm polyaspartamide nanoparticles to deliver CAR-encoding mcDNAs to immune cells, leveraging clathrin-mediated uptake to minimize off-target accumulation in the liver and enabling repeat dosing [5]. The Grit platform uses T cell-targeted lipid nanoparticles (T-LNPs) that contain modified mRNA encoding an anti-CD19 CAR. The surface of these T-LNPs is engineered with CLAMP-linked VHH antibodies, which enable selective delivery, stable CAR expression, and robust in vivo expansion, while minimizing off-target transfer and CRS [6]. Aera’s platform developed an in vivo transient CAR-T that uses targeted LNPs to deliver mRNA encoding an anti-CD19 CAR to CD8⁺ T-cells, depleting differentiated B-cells while preserving early progenitors [7]. Tessera’s platform combines optimized extrahepatic LNPs with RNA-based gene writers to enable efficient in vivo editing of hematopoietic stem cells (HSCs) and T-cells with minimal off-target delivery [8, 9]. Orna’s platform also enables transient and tunable CAR expression in T-cells, achieving rapid and selective plasma cell depletion [10, 11].

Preclinical outcomes and translational challenges

Experimental models support the feasibility of immune reprogramming through in vivo CAR-T (Table 2). Results show complete tumor eradication [1, 2]. In B-cell malignancies, more than 95% B-cell depletion is achieved with re-dosable regimens and minimal hepatotoxicity [3, 4, 6]. In SLE and RA, B-cell depletion occurs while preserving naïve B-cells and maintaining normal organ function [5, 7]. For HSC editing, up to 75% efficiency has been reported, enabling combined CAR-T and stem-cell therapies [8, 9]. The Orna platform showed deep B-cell depletion in peripheral blood and lymphoid tissues. B-cell depletion was also observed in a lupus model, along with a decrease in anti-dsDNA titers [10, 11]. Other platforms, including LNPs conjugated to VHH antibodies using CLAMP, and engineered viral vectors, have also completed the preclinical phase [1–4, 6].

Table 2.

Preclinical studies of in vivo CAR-T platforms

Disease	Tumor/cell type	Models	Platforms	Route	Dosing	Outcomes	Translational consequences	Ref
Oncological diseases	Primary human T cells (CD3/CD7/CD8 targeted) & tumor xenografts	Humanized mice	Qihan	IV	N/A	Complete tumor clearance, low cytokine secretion	Supports one-time administration for aggressive malignancies while minimizing tonic signaling	[1]
B-cell malignancies	Human T cells targeted by VHH-T-LNP / B cells (CD19⁺)	Human PBMC - NOG mice	Grit (GT801)	IV	≥ 0.01 mg/kg	> 95% B cell depletion, low cytokines, re-dosable	Enables chronic disease management with controllable exposure	[6]
B-cell malignancies	Nalm6-Luc leukemia cells (CD19⁺)	NSG mice	IMV101	IV	1˜25 × 10⁶ TU/mouse	Strong CAR-T expansion, complete tumor regression, spleen-restricted biodistribution	Favorable biodistribution reduces off-target risk	[3]
Disseminated multiple myeloma	OPM-2-Luc myeloma cells (BCMA⁺)	MHC I/II-deficient NSG mice	LV-169	IV	Single injection of at different doses 3	Complete tumor clearance within 28 days, long-term remission, low cytokine levels	Long-term CAR persistence supports profound myeloma responses	[2]
RRMM, NHL, severe autoimmune diseases	Human T cells (CD7-targeted) /B cells (CD19⁺, CD20⁺) / BCMA⁺MM cells	HSC-humanized mice, human PBMC-NPG, myeloma xenograft	IASO	IV	Single dose	T cell transduction, B cell clearance, MM regression	Platform flexibility across oncology and autoimmunity	[4]
Oncological and autoimmune disease, SCD	HSCs (LT-HSC: CD34⁺CD38⁻CD90⁺CD45RA⁻), T cells, B cells	Humanized xenograft mice, CD34⁺ humanized mice, B6 mice, NHP	Tessera	IV	Single dose	Up to 55% CAR-T in vivo, tumor clearance; 75% HSC editing, potent CAR-T generation	Enables in vivo genome engineering beyond CAR-T	[8, 9]
Lupus, Multiple Myeloma	B cells (CD19⁺), plasma cells (BCMA⁺), autoreactive B cells (dsDNA⁺)	CD34⁺ humanized mice, lupus humanized mice, Tumor-engrafted humanized mice, NHP	Orna	IV	(0.1, 0.5, and 1.0 mg/kg) or three weekly doses of 0.1 mg/kg	Anti-dsDNA reduction, profound B cell loss, > 95% plasma-cell ablation	Transient CAR expression enables immune reset	[10, 11]
Autoimmune disease	Cytotoxic CD8⁺ T cells / Mature B cells (CD19⁺)	CD34⁺ humanized mice, NHP	Aera	IV	Two-dose cycle (≥ 0.3 mg/kg)	Durable B cell depletion, naïve B cell repopulation	Balances efficacy with immune reconstitution	[7]
Autoimmune disease (SLE/RA/MS)	Peripheral & splenic B cells (CD19⁺)	C57BL/6 mice	Velvet	IV	1.6 mg/kg, twice weekly	B cell depletion, normal serum chemistry in liver (ALT, AST, GGT) or kidney (BUN) function	Suitable for chronic autoimmune indications	[5]

HSC Hematopoietic Stem cell, IV Intravenous, LNP Lipid nanoparticle, LT-HSC long-term hematopoietic stem cell, MM Multiple Myeloma, MS Multiple sclerosis, NHL Non-Hodgkin lymphoma, NHP Nonhuman primate, Q3Dx2 twice three days apart, RA Rheumatoid Arthritis, RRMM Relapsed/Refractory Multiple Myeloma, SCD sickle cell disease, SLE Systemic Lupus Erythematosus, NSG NOD SCID gamma, NHP Non-human primates

Despite these promising results for both viral and non-viral platforms, clinical translation faces several challenges. Viral-based platforms provide durable but non-reversible expression, posing concerns in non-cancerous settings. Non-viral platforms can offer transient CAR expression; however, repeated administration may be required to maintain efficacy. Furthermore, as most of these platforms have been evaluated in immunodeficient models, questions about immunogenicity, clearance, and immune-related adverse events in humans remain unknown.

Conclusion

The rapid advancement of multiple in vivo CAR-T platforms is evident. Nevertheless, these platforms present challenges for regulatory agencies, requiring meticulous framework evaluation and risk mitigation strategies to foster innovation. If successful, applying in vivo CAR strategies to other immune cells, particularly NK cells and macrophages, could open new avenues for tumor therapy. Future clinical translation will depend on achieving a balance between durability, safety, scalability, monitoring immune responses, and conducting large-scale clinical trials.

Abbreviations

ASH

American Society of Hematology

BCMA

B-cell maturation antigen

CAR

Chimeric Antigen Receptor

CLAMP

Controllable Ligand Attachment Modification and Purification

CRS

Cytokine Release Syndrome

HSC

Hematopoietic stem cell

mcDNA

Minicircle DNA

NK

Natural Killer

RA

Rheumatoid arthritis

SLE

Systemic lupus erythematosus

tLNP

Targeted lipid nanoparticle

T-LNP

T cell-targeted lipid nanoparticle

Author contributions

GA drafted and prepared the tables. MR participated in the drafting, editing, and revising of the manuscript. All authors read and approved the final manuscript.

Funding

Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.