In Vivo CAR-T Cell Therapies: Platform Landscape, Clinical Progress, and Strategic Signals as of June 2026

Overview and Conceptual Foundation

In vivo CAR-T（体内生成CAR-T细胞疗法）represents a fundamental departure from the ex vivo paradigm that has defined approved chimeric antigen receptor T-cell therapies since 2017. Rather than extracting a patient's T cells, engineering them in a centralized manufacturing facility over two to four weeks, and reinfusing the product after lymphodepleting chemotherapy, in vivo approaches deliver genetic instructions—via viral vectors, lipid nanoparticles/LNPs（脂质纳米颗粒）, or other targeted carriers—directly into the patient's body, where endogenous T cells are reprogrammed in situ to express a CAR construct and mount an antitumor or autoimmune-regulatory response. This shift promises to compress time-to-treatment from weeks to days, eliminate patient-specific manufacturing bottlenecks, reduce cost, and expand access to patients with poor-quality T cells or limited proximity to manufacturing centers.123

Compared with ex vivo allogeneic "off-the-shelf" CAR-T, in vivo platforms avoid the graft-versus-host disease risk inherent in donor-derived products but introduce distinct genetic medicine considerations: biodistribution control, vector shedding, integration-site safety for integrating systems, and immunogenicity of delivery vehicles. As of mid-2026, the field has decisively entered clinical validation, with multiple programs in Phase 1 trials across hematologic malignancies and autoimmune diseases, and with deal-making activity at an unprecedented scale.

Platform Landscape

Five principal delivery modalities are currently in active development, each with a distinct risk-benefit profile. The fifth modality, targeted polymeric nanoparticles, offers a non-viral, non-lipid synthetic delivery platform engineered for active tissue targeting. These nanoparticles typically incorporate ligands (e.g., antibodies, peptides, or aptamers) on their surface to enable cell-specific uptake following systemic administration. Unlike LNP-mRNA systems, polymeric nanoparticles can be designed to achieve sustained release and enhanced tumor penetration, though their clinical translation has been constrained by manufacturing complexity and limited in vivo transfection efficiency relative to viral vectors. Early-phase data across oncology and rare disease programs are emerging, with risk-benefit profiles largely defined by the specific polymer composition and targeting moiety employed.

Lentiviral vectors offer stable genomic integration, durable CAR expression, and the potency benchmark of conventional ex vivo CAR-T. Kelonia Therapeutics' iGPS (in vivo gene placement system) uses envelope-modified lentiviral particles co-expressing modified fusogen and anti-CD3 moieties to achieve T-cell-selective transduction without lymphodepletion. The clinical advantage is a single-dose, outpatient-compatible infusion; the principal concerns are insertional mutagenesis risk, the requirement for 15-year long-term follow-up mandated by regulatory gene-therapy frameworks, vector shedding assessment, and manufacturing scale-up complexity.13

LNP–mRNA platforms deliver CAR-encoding messenger RNA that is transiently expressed without genomic integration. Ionizable lipid formulations with antibody-based T-cell targeting (anti-CD3, anti-CD4, anti-CD8) enable cell-selective transfection and support repeat dosing. The non-integrating safety profile is attractive, particularly in autoimmune indications where the acceptable toxicity threshold is lower. However, transfection efficiency in circulating T cells remains technically challenging, hepatic and myeloid off-target uptake requires optimization, and the need for repeated administration to sustain therapeutic effect has not been fully resolved in human studies.123

Engineered circular RNA (circRNA) paired with optimized LNPs extends protein expression duration beyond linear mRNA while retaining a non-integrating profile. Strand Therapeutics' EverScript platform incorporates SignalLock microRNA-responsive regulatory sequences that modulate CAR expression in off-target cell types, building programmable safety controls directly into the construct.23

CRISPR-based site-specific knock-in approaches combining enveloped delivery vehicles (EDVs) carrying Cas9-RNP complexes with AAV HDR templates enable targeted integration at the TRAC locus, preserving physiologic T-cell promoter control and reducing random insertional mutagenesis risk. Translational barriers include two-component delivery, pre-existing anti-AAV antibodies, and Cas9 immunogenicity requiring patient stratification.1

Table 1. In Vivo CAR-T Platform Comparison

Clinical Progress

As of June 2026, disclosed human efficacy data remain concentrated in early Phase 1 cohorts, predominantly in lentiviral programs targeting hematologic malignancies.123

Kelonia KLN-1010 (R/R multiple myeloma, anti-BCMA): The inMMyCAR Phase 1 trial is currently the most mature in vivo CAR-T program with published human data. Early cohort results presented at ASH 2025 and updated at ASCO 2026 reported 100% overall response rate (ORR) in 18 evaluable patients, 100% MRD-negative bone marrow responses at one month, and stringent complete responses (sCR) or very good partial responses (VGPR) in patients with ≥4 months follow-up, all MRD-negative. Critically, no lymphodepletion was required. CAR-T expansion peaked around Day 15 with memory-like phenotype. Safety included Grade 1–2 CRS in the majority of patients (one Grade 3 ICANS, limited to three days), with no infusion reactions after dexamethasone premedication. Median time from consent to infusion was 13 days, and outpatient dosing was approved. Kelonia received FDA Fast Track Designation in May 2026.13

EsoBiotec ESO-T01 (R/R multiple myeloma, anti-BCMA): An investigator-initiated trial conducted in Wuhan (first patient dosed December 2024) demonstrated rapid responses. Within 28 days of first dosing, one patient achieved complete MRD negativity. Updated data from five patients showed four objective responses including three stringent complete responses, confirming rapid kinetics and depth of response comparable to ex vivo CAR-T benchmarks.3

Legend Biotech LB2501 (R/R B-cell non-Hodgkin lymphoma, CD19/CD20 dual-targeting): Late-breaking Phase 1 data presented at EHA 2026 (June 2026) reported 100% ORR (6/6) and 5/6 CR at dose level 2, with no lymphodepletion, no dose-limiting toxicities, no ICANS, and no SAEs or deaths. CRS and infusion-related reactions were Grade 1–2 only. CAR-T cells were detectable in peripheral blood for up to 116 days, and vector integrations were polyclonal and diverse.3

CREATE Medicines MT-303 (frontline HCC, GPC3-targeted myeloid CAR): First patient dosed in December 2025 in a combination trial with atezolizumab and bevacizumab. Early signals described as "extremely compelling" by the company across >50 patients dosed across programs, though detailed efficacy data were not publicly disclosed as of the search cutoff.23

AbbVie/Capstan CPTX2309 (anti-CD19, B-cell autoimmune diseases): Phase 1 program using CD8-targeted tLNP to deliver anti-CD19 CAR mRNA, enabling transient B-cell depletion without lymphodepletion. NHP studies demonstrated dose-dependent systemic B-cell depletion at doses as low as 0.3 mg/kg, with subsequent naïve-predominant B-cell reconstitution, consistent with immune "reset."23

Table 2. Selected In Vivo CAR-T / In Vivo Immune-Cell Engineering Programs (2025–2026)

Translational and Regulatory Considerations

In vivo CAR-T fundamentally compresses the clinical workflow: a single intravenous infusion replaces leukapheresis, ex vivo expansion, and in most disclosed programs, lymphodepleting chemotherapy. This enables outpatient administration and dramatically reduces the 4–8 week manufacturing delay that often results in disease progression before infusion.12

Regulatory expectations for in vivo CAR-T mirror gene therapy frameworks. Lentiviral programs require comprehensive integration-site analysis, clonal tracking, biodistribution characterization, vector shedding monitoring, and 15 years of long-term follow-up. Non-integrating LNP-mRNA platforms require demonstration of extrahepatic T-cell targeting reproducibility, innate immune activation profiling, and repeat-dose immunogenicity surveillance. Patient selection parameters likely to influence outcomes include baseline lymphocyte counts, immune competence, concomitant immunosuppression, and serostatus for pre-existing antibodies (especially anti-AAV for vector-based platforms).13

Safety switch integration (inducible caspase-9, truncated EGFR) remains largely preclinical for in vivo constructs. The programmable miRNA-responsive SignalLock system (Strand Therapeutics) and EverScript platform represent early efforts to embed controllability directly into the RNA construct—a regulatory priority as in vivo platforms scale.23

China's regulatory environment has emerged as a significant translational accelerator. The National Health Commission's dual-track IIT framework permitted EsoBiotec to generate first-in-human data from its Wuhan trial—translating to an estimated 18-month development advantage that directly contributed to AstraZeneca's acquisition rationale. State Council Order No. 818 (effective May 2026) has elevated IIT data integrity standards toward international norms, likely increasing the global translatability of China-originated datasets.3

Competitive and Deal-Making Signals

The in vivo CAR-T field has experienced a wave of high-value M&A consolidation in 2025–2026 that reflects pharma confidence in platform validation and commercial scalability.

Table 3. Deal-Making and Strategic Signals (2024 – June 2026)

The consolidation pattern reveals several strategic themes: (1) lentiviral platforms dominate large-cap M&A due to their durable integration and validated clinical responses; (2) targeted LNP platforms are preferred for autoimmune indications where transient expression and repeat dosing offer a safer toxicity profile; (3) both Lilly and AbbVie frame acquired platforms as broadly applicable beyond initial indications, suggesting multi-indication pipeline ambitions; and (4) the rapid absorption of leading platforms by top-tier pharma signals a consolidation phase that may accelerate clinical development while narrowing near-term competitive diversity.123

Outlook

The most clinically advanced and commercially validated modality at mid-2026 is lentiviral in vivo CAR-T in hematologic malignancies, anchored by KLN-1010 and ESO-T01 data demonstrating MRD-negative responses without lymphodepletion. Near-term catalysts include multi-center expansion of the inMMyCAR trial under Lilly, Phase 1 dose escalation and RP2D determination for LB2501, and first autoimmune efficacy readouts from CPTX2309 and CRT-402 in 2026–2027.123

For LNP-based platforms, the central question remains whether transient CAR expression can achieve clinically durable responses without excessive repeat dosing—a threshold not yet established in human subjects. Solid tumor efficacy data are limited to early biological signals in myeloid-targeted programs (MT-303/HCC) and require substantially larger, controlled datasets before conclusions can be drawn.23

Critical unresolved barriers include: durability of non-integrating platforms without frequent re-administration; consistency of T-cell targeting across diverse patient phenotypes; long-term clonal safety surveillance for integrating vectors; manufacturing cost-of-goods and GMP scalability for targeted nanoparticle formulations; and competitive positioning relative to improving ex vivo CAR-T products and bispecific T-cell engagers. The convergence of early clinical validation, operational simplification, and sustained pharma investment positions in vivo CAR-T as a potentially transformative modality—but the field remains in early Phase 1, and caution regarding preliminary efficacy and safety claims is warranted until larger, longer-duration cohort data mature.123