neobiologia ATLAS OF THE NEW BIOLOGY
ESENZH
EXTRA PLATE / IN VIVO VECTORS TROPISM · DURABILITY · PHASE 1
VECTOR
ESCAPING THE LIVER TO REPROGRAM CIRCULATING T CELLS
ATGCCTAGCGTACG TACGATCAGTCAGT CGATCGTAACGTAC
VECTOR ENGINEERING · VERIFIED 2026-11-18

In vivo cell therapy vectors: extrahepatic tropism and Phase 1 trials

Engineering therapeutic cells directly inside the human body now hinges less on receptor design than on delivery vehicle physics. Extrahepatic antibody-targeted lipid nanoparticles, scarless circular mRNA, and dual EDV/AAV platforms bypass hepatic sequestration to reprogram lymphocytes in situ and enter human Phase 1 clinical trials.

LIVER OFF-TARGET
<5% in hepatocytes (Mol Ther 2026)
cmRNA DURABILITY
7 days in human T / 14d in mice
TRAC KNOCK-IN
18/20 complete responses (Nature 2026)
HUMAN PHASE 1
4/5 myeloma responses (ESO-T01)
<5%HEPATIC OFF-TARGET EXPRESSION
7 DAYSCAR WINDOW VIA CIRCULAR RNA
18 / 20COMPLETE RESPONSES EDV/AAV
4 / 5OBJECTIVE RESPONSES ESO-T01
01
EXTRAHEPATIC TROPISM

The hepatic sink and how engineered carriers bypass it

Four engraved scientific panels: antibody-targeted lipid nanoparticles bypassing liver cells in circulation, closed-loop circular RNA translating CAR proteins inside a T cell, dual EDV and AAV vectors integrating a transgene into the TRAC locus, and engineered CAR-T cells lysing target cells
1. Nanobody-decorated extrahepatic tLNPs bypass hepatic uptake in the bloodstream. 2. Scarless circular mRNA extends CAR translation inside T cells. 3. Dual EDV + AAV vectors insert a CAR cassette at the TRAC locus. 4. In situ CAR-T cells engage and lyse target cells. Conceptual diagram.

As outlined in our foundational plate on in vivo CAR-T cell generation, reprogramming lymphocytes directly in circulation promises to eliminate leukapheresis, multi-week cleanroom expansion, and toxic lymphodepleting conditioning. Yet between biological promise and clinical consistency sits an unforgiving anatomical barrier: the human liver acts as a high-capacity physiological sink for almost every systemically injected nanoparticle or viral capsid.

In conventional lipid nanoparticle formulations developed for mRNA vaccines, the neutral lipid shell rapidly adsorbs serum apolipoprotein E within seconds of intravenous infusion. That spontaneous protein corona funnels particles into hepatic sinusoids, where hepatocytes endocytose them via low-density lipoprotein receptors and Kupffer macrophages scavenge the remainder. When the therapeutic target is circulating or splenic T lymphocytes, hepatic trapping wastes most of the administered payload while exposing liver parenchyma to unnecessary inflammatory stress.

A 2026 study in Molecular Therapy by Lemgart and colleagues demonstrates how disciplined biophysical engineering reverses this default organotropism. By screening ionizable cationic lipids to select lipid 15, holding particle diameter near 100 nanometers (polydispersity index below 0.15), and tuning apparent pKa between 6.5 and 7.5, the team created nanoparticles that stay neutral at physiological blood pH 7.4 yet switch strongly cationic (~30 mV) inside acidic endosomes at pH 5.5 to trigger cytoplasmic escape. Conjugating anti-human CD8 nanobody (VHH) binders to that shielded shell redirected the particles away from liver clearance.

Following a clinically relevant intravenous dose of 0.3 mg/kg in PBMC-humanized NSG mice, more than 20% of human CD8-positive T cells across blood, spleen, and bone marrow expressed functional anti-CD22 CAR within 24 hours. Just as importantly, flow cytometry of hepatic cell populations showed that off-target GFP reporter expression across hepatocytes, sinusoidal endothelial cells, and Kupffer macrophages remained below 5% relative to vehicle controls.

A parallel non-viral strategy removes surface antibody conjugation altogether and tunes organotropism through Selective Organ Targeting (SORT) lipids. Adding a fifth anionic lipid component such as 18:1 PA or cardiolipin-mimic phosphoamide lipids shifts the adsorbed plasma corona in circulation, passively rerouting nanoparticles toward splenic T cells while minimizing manufacturing complexity associated with covalent ligand conjugation.

PRIMARY LITERATURE: Lemgart et al., Molecular Therapy, 2026, Molecular Therapy (2026) · Platform synthesis: Xu et al., Journal of Hematology & Oncology, 2025, J Hematol Oncol (2025).
02
DURABILITY & INTEGRATION

Linear mRNA, scarless circular RNA, and dual EDV/AAV editing

Solving tissue tropism answers only where a payload arrives; the second engineering axis governs how long the genetic instruction persists once inside the target T cell. Across 2025 and 2026, three distinct intracellular persistence profiles have emerged, each matched to specific disease biology.

EXTRAHEPATIC VECTOR TROPISM & IN VIVO CAR DURABILITY (2025-2026) A. TISSUE SELECTIVITY & KNOCK-IN (%) 50% 40% 30% 20% 10% 0% <5% Liver off-target >20% CD8+ tLNP 39.2% EDV+AAV6 vitro 40% EDV+AAV-hT7 B. CAR SURFACE DURABILITY (DAYS 0 TO 14) 100% 75% 50% 25% 0% 0d 1d 4d 7d 14d Linear mRNA (t1/2 ~2.5d, clears day 4) Scarless cmRNA (t1/2 ~4.5d, day 7+) EDV/AAV TRAC knock-in (stable >14d) Sources: Lemgart et al. (Mol Ther 2026), Hu et al. (STTT 2025), Nyberg et al. (Nature 2026).

The first profile relies on nucleoside-modified linear mRNA. Its central appeal is strict reversibility without chromosomal modification, ideal for resetting autoreactive B cells in autoimmune disease. However, open ends make linear transcripts vulnerable to rapid exonuclease turnover. As Hu and colleagues reported in Signal Transduction and Targeted Therapy (2025), linear mRNA encoding anti-CD19 or anti-GPRC5D CARs in primary human T cells peaked rapidly within 6 hours of delivery and faded to undetectable levels by day 4, corresponding to a protein expression half-life of roughly 2.5 days.

To stretch transient CAR activity without genomic integration, the same team built an engineered permuted intron-exon synthesis system (Hi-Scarless-PIE) that produces scarless circular mRNA (cmRNA), stripping out extraneous exon-junction residues that trigger innate RNA sensors. In primary human T cells, scarless cmRNA pushed peak CAR expression to day 1, doubled effective protein half-life to ~4.5 days, kept CAR detectable through day 7, and yielded higher surface receptor density alongside reduced T-cell exhaustion signatures. Packaged into ~75 nm lipid nanoparticles (~90% encapsulation), cmRNA sustained reporter expression in mice for 14 days.

When aggressive hematologic cancers or solid tumors require durable clonal expansion rather than a multi-day hit-and-run pulse, permanent site-specific genome insertion enters the picture via CRISPR-Cas9. In March 2026, Nyberg, Jennifer Doudna, and Justin Eyquem reported in Nature a coordinated dual-vector architecture designed to replace random lentiviral insertion. An anti-CD3 enveloped delivery vehicle (EDV) carrying Cas9 ribonucleoproteins tagged with four nuclear localization signals activates T cells and cuts the endogenous TRAC locus, while an evolved neutralizing-antibody-resistant adeno-associated virus capsid (AAV-hT7) delivers a promoterless CAR homology-directed repair template.

In activated primary human T cells in vitro, EDV plus AAV6 generated 39.2% CAR-positive, TCR-negative cells (rising toward ~80% with DNA-PK inhibition using M3814). In vivo across four PBMC donors in aggressive NALM6 B-ALL humanized mice, a single intravenous injection of anti-CD3 EDV plus AAV-hT7 produced complete responses in 18 of 20 mice, with up to 40% of human CD45-positive cells expressing CAR in bone marrow and spleen. Against head-to-head lentiviral benchmarks, EDV/AAV achieved 6/6 complete tumor control versus 1/6 complete remission with high-dose lentivirus, alongside 21- to 50-fold higher week-2 expansion, and delivered 8/8 complete responses in the OPM2 multiple myeloma model.

Vector Architecture Capacity & Diameter Genomic Integration? Typical CAR Durability Repeat Dosing Viable? Tropism & Key Constraint
Targeted LNP + linear mRNA (anti-CD8 VHH) Up to ~6 kb · 90 to 120 nm No (transient cytoplasmic) Peak 6 to 24 h; lasts 4 to 5 days Yes (4 doses verified) <5% liver off-target; requires repeat dosing in oncology
LNP + scarless cmRNA (Hi-Scarless-PIE) Up to ~6 kb · ~75 nm No (closed covalent RNA ring) Peak day 1; lasts 7 days in T (14d LNP) Yes (low innate sensing) Higher CAR surface density and lower exhaustion
Targeted Lentivirus (mutated VSV-G / Cocal) Up to ~10 kb · 80 to 120 nm Yes (semi-random insertion) Permanent (clonal memory) No (neutralizing antibodies) High potency in myeloma; acute CRS and insertion risk
Dual EDV + AAV-hT7 (TRAC knock-in) AAV ~4.7 kb + Cas9 RNP Yes (site-specific at TRAC) Permanent under native promoter Limited by anti-AAV immunity 18/20 CR in B-ALL; preclinical mouse validation only
Engineered AAV (DARPin / Ark313) ~4.7 kb · 20 to 25 nm Episomal unless paired with nuclease Weeks to months in resting cells Limited (anti-capsid Abs) ~40% T-cell transduction; residual hepatic retention
PRIMARY LITERATURE: Hu et al., Signal Transduction and Targeted Therapy, 2025, Signal Transduct Target Ther (2025) · Nyberg et al., Nature, 2026, Nature (2026).
03
REDOSING & SAFETY

Vector immunogenicity, repeat dosing, and acute human toxicity

The clearest operational divide between integrating viral platforms and synthetic lipid nanoparticles lies in whether a therapy behaves as a single-shot genetic intervention or a titratable, redosable medicine. Intravenous lentiviruses and adeno-associated viruses display dense foreign viral proteins on their surface, triggering rapid neutralizing antibody and complement responses that usually block effective repeat administration.

By contrast, Lemgart and colleagues tested repeat dosing of CD8-targeted VHH mRNA-LNPs at 0.3 mg/kg twice weekly for two weeks. Because a single transfection produces measurable CAR expression for roughly 5 days after peaking at 24 hours, blood sampling after the fourth dose proved that newly generated CAR-positive CD8+ T cells matched the peak percentages seen after the very first dose. No significant differences in body weight or spleen weight were observed, establishing that low-dsRNA purified tLNPs can be redosed pharmacologically without rapid vector neutralization.

On the integrating viral side, the first detailed peer-reviewed Phase 1 clinical trial of an in vivo lentiviral CAR-T product in oncology underscored both dramatic potency and sharp acute toxicity boundaries. Published in Nature Medicine in 2026 by An and colleagues, the single-arm ESO-T01 study (ClinicalTrials.gov NCT06791681) infused a nanobody-directed, immune-shielded anti-TCR lentiviral vector encoding a humanized anti-BCMA CAR at 0.2 x 10^9 transduction units into five heavily pretreated male patients with relapsed or refractory multiple myeloma, with zero leukapheresis and zero conditioning chemotherapy.

Efficacy signals were striking: four of five patients achieved objective responses, including three stringent complete remissions, and all four evaluable responders reached minimal residual disease negativity at 10^-5 sensitivity by day 60. Yet every treated patient experienced grade 3 or higher adverse events, including cytokine release syndrome in four of five patients (three grade 3, one grade 2) requiring tocilizumab or corticosteroids, transient cytopenias, reversible liver enzyme elevations, and one death following grade 1 neurotoxicity and extramedullary spinal cord compression. Enrollment stopped early in 2025 so investigators could recalibrate vector dosing and inflammatory prophylaxis.

Meanwhile, extrahepatic tLNPs continue pushing beyond oncology and autoimmunity into chronic organ fibrosis. In 2026, Yashaswini and colleagues reported in Science Translational Medicine that anti-CD5 targeted LNPs delivering anti-FAP CAR mRNA transiently programmed antifibrotic T cells in vivo, depleting profibrogenic hepatic stellate cells and restoring liver homeostasis in metabolic dysfunction-associated steatohepatitis (MASH).

PRIMARY LITERATURE: An et al., Nature Medicine, 2026, Nature Medicine (2026) · Yashaswini et al., Science Translational Medicine, 2026, Sci Transl Med (2026).
04
2026 TRIAL LEDGER

Phase 1 human clinical trial ledger and translational boundaries

Just three years ago, in vivo cell therapy existed solely in mouse models and primate pilot cohorts. By November 2026, both integrating lentiviral platforms and non-viral CD8-targeted mRNA-LNPs are actively enrolling patients in registered Phase 1 trials. The ledger below summarizes verified ClinicalTrials.gov records alongside the Phase 1/2 CASTLE basket trial (NCT06347718) in Nature Medicine (2026), which sets the clinical efficacy bar using conventional ex vivo autoimmune CAR-T.

Program / Registry Vector & Targeting Ligand CAR Target & Payload Clinical Indication Status & Cohort (Nov 2026)
ESO-T01 (NCT06791681) Immune-shielded LV · anti-TCR VHH Anti-BCMA CAR (integrating) R/R Multiple Myeloma Early Phase 1 · 4/5 ORR in Nat Med 2026
INT2104 (NCT06539338) Mutated VSV-G LV · anti-CD7 Anti-CD20 CAR (T & NK cells) R/R B-cell malignancies Phase 1 active, not recruiting (n=10 actual)
UB-VV111 (NCT06528301) Cocal LV · anti-CD3 + CD58/CD80 Anti-CD19 CAR (+ rapamycin switch) R/R DLBCL and CLL Phase 1 recruiting (est. n=106)
KLN-1010 (NCT07075185) Mutated VSV-G LV · anti-CD3 Human anti-BCMA CAR R/R Multiple Myeloma Phase 1 recruiting (est. n=70)
CPTX2309 (NCT06917742) Extrahepatic tLNP · anti-CD8 Transient anti-CD19 CAR mRNA Healthy, Rheumatoid Arthritis, SLE Phase 1 recruiting (est. n=64)
HN2301 (NCT06801119) ILB-3132 tLNP · anti-CD8 Transient anti-CD19 CAR mRNA SLE, Scleroderma, RA Phase 1 recruiting (est. n=30)
Arnovie101 (NCT07629596) Targeted mRNA-LNP In vivo CAR mRNA SLE and Autoimmune Hemolytic Anemia Early Phase 1 recruiting (est. n=5)
CASTLE ex vivo ref. (NCT06347718) Ex vivo autologous Zorpo-cel Anti-CD19 CAR + Flu/Cy chemo SLE, SSc, Myositis (n=24) Nat Med 2026: 22/24 efficacy at week 24

Comparing these nascent programs against CASTLE (where 22 of 24 autoimmune patients met 24-week remission or stabilization endpoints free of glucocorticoids, but only after leukapheresis and cyclophosphamide/fludarabine lymphodepletion) crystallizes the two diverging clinical paths. If redosable linear or circular mRNA tLNPs reproduce durable B-cell immune resets in outpatient clinics without chemotherapy, they are poised to dominate autoimmune and fibrotic indications. Conversely, targeted lentiviruses and dual EDV/AAV systems remain focused on refractory blood cancers where permanent CAR persistence is non-negotiable, provided cytokine release and off-target integration are rigorously controlled.

Honest boundary in November 2026: zero in vivo cell therapy vectors hold commercial FDA or EMA approval. Scarless cmRNA CAR kinetics and EDV/AAV TRAC knock-in efficiency come primarily from immunodeficient humanized mice, which cannot fully predict human hepatic clearance or anti-vector immunity; moreover, the first published myeloma Phase 1 trial of an integrating in vivo lentivirus (ESO-T01) halted early enrollment after grade 3+ adverse events occurred in all five patients.
CLINICAL SOURCES: An et al., Nature Medicine, 2026, Nature Medicine (2026) · Müller et al., Nature Medicine, 2026 (CASTLE), Nature Medicine CASTLE (2026) · Trials: NCT06791681 · NCT06917742 · NCT06801119.
<5%HEPATIC OFF-TARGET IN CD8-tLNP
~4.5 DAYSSCARLESS cmRNA HALF-LIFE
18 / 20COMPLETE RESPONSES EDV/AAV
7+ACTIVE PHASE 1 IN VIVO TRIALS
Why do standard lipid nanoparticles accumulate in the liver instead of T cells?
Upon entering the bloodstream, conventional lipid nanoparticles rapidly adsorb circulating apolipoprotein E and are cleared by low-density lipoprotein receptors on hepatocytes as well as hepatic Kupffer macrophages. Reaching circulating or splenic T lymphocytes requires tuning particle size near 100 nm, adjusting ionizable lipid pKa to 6.5-7.5, and adding anti-CD8 binders.
How does scarless circular mRNA improve on linear mRNA for CAR-T engineering?
Linear mRNA carries open ends vulnerable to cytoplasmic exonucleases, peaking near 6 hours and vanishing by day 4 with a ~2.5-day protein half-life. Scarless circular mRNA forms a closed covalent loop without extraneous splicing scars, extending protein half-life to ~4.5 days and maintaining measurable CAR expression on human T cells through day 7.
Can non-viral in vivo cell therapy vectors be redosed without losing potency?
Yes, unlike lentiviral or AAV vectors that elicit neutralizing antibodies after a single infusion. Preclinical studies with CD8-targeted VHH lipid nanoparticles at 0.3 mg/kg showed that four repeat intravenous doses over two weeks generated comparable circulating CAR-positive T-cell peaks after the fourth injection with under 5% off-target liver expression.
Extrahepatic tropismVector engineering that evades hepatic clearance to reach circulating or splenic lymphocytes
Circular mRNA (cmRNA)Closed-loop RNA lacking free ends, resisting exonucleases to extend CAR protein translation
EDVAntibody-targeted enveloped delivery vehicle delivering Cas9 ribonucleoprotein transiently
AAV-hT7Evolved adeno-associated virus capsid resistant to human neutralizing antibodies for DNA templates

Continue reading