In Vivo CAR-T Therapy Delivers First Human Proof-of-Concept for Autoimmune Reset, Bypassing Lab Manufacturing
MagicRNA's mRNA lipid nanoparticle therapy reprogrammed patients' own T cells inside the body, wiping out disease-driving B cells in refractory lupus patients without conventional cell manufacturing.
Overview
For more than a decade, chimeric antigen receptor T-cell therapy — CAR-T — has transformed the treatment of blood cancers. But it has remained inaccessible to the vast majority of patients who might benefit, hampered by a manufacturing process that requires removing a patient’s immune cells, engineering them in a specialized laboratory, and reinfusing them days or weeks later. A study published in September 2025 in the New England Journal of Medicine has now demonstrated, for the first time in living patients, that this bottleneck can be bypassed entirely.
The First-in-Human Milestone
Shenzhen-based MagicRNA Biotech enrolled five patients with treatment-refractory systemic lupus erythematosus (SLE), four of whom also had lupus nephritis, in a first-in-human trial of its investigational therapy HN2301. Rather than extracting T cells for ex vivo modification, HN2301 delivers synthetic mRNA encoding a CD19-targeting chimeric antigen receptor directly into the bloodstream via the company’s proprietary lipid nanoparticle platform, known as EnC-LNP. The nanoparticles are engineered to seek out and enter CD8-positive T cells within the patient’s own body.
According to the NEJM publication, CAR expression was detectable in up to 60 percent of CD8-positive T cells within six hours of infusion at higher dose levels. The newly programmed cells promptly attacked CD19-bearing B cells — the immune cells that drive lupus pathology by producing autoantibodies — achieving complete depletion of circulating B cells that persisted for seven to ten days before the mRNA degraded and the CAR expression faded.
The clinical effect was notable. All five patients saw their Systemic Lupus Erythematosus Disease Activity Index (SLEDAI-2000) scores decline by as much as 20 points within three months, a reduction that would be considered clinically significant by any standard measure. No patient experienced grade 3 or higher cytokine release syndrome — a life-threatening inflammatory complication that has historically limited conventional CAR-T use — and no neurotoxic effects were reported.
Why the Manufacturing Question Matters
Conventional CAR-T therapy, as approved for lymphomas and multiple myeloma, requires a highly specialized and expensive manufacturing chain. A patient’s white blood cells are harvested through a process called leukapheresis, shipped to a central facility, genetically engineered using viral vectors, expanded over two to four weeks, quality-tested, then shipped back and infused. The cost per patient regularly exceeds $400,000, and patients must be medically stable enough to tolerate the wait — a serious constraint in rapidly progressing diseases.
For autoimmune diseases, the calculus is even more complicated. The patients most likely to benefit are those who have exhausted conventional immunosuppressants, yet who may not have the same clinical infrastructure as a tertiary oncology center. Researchers at Erlangen University Hospital, led by rheumatologist Prof. Georg Schett, who pioneered the use of ex vivo CD19 CAR-T in SLE beginning in 2021, collaborated with MagicRNA on the trial and described the result as showing “functional CAR T cells generated directly in patients’ bodies, achieving rapid B-cell clearance and clinical improvement of autoimmune disease.”
Because the mRNA in HN2301 degrades within days, CAR expression is transient — the immune reset is temporary but potentially repeatable. This contrasts with viral-vector CAR-T, which integrates permanently into the genome and is a one-time treatment. The transient profile may prove advantageous in autoimmune applications, where indefinite B-cell depletion carries infection risks, but will require longer follow-up to characterize durability of remission.
A Growing Field
MagicRNA’s result is not isolated. Capstan Therapeutics published parallel preclinical data in Science in June 2025 demonstrating that targeted lipid nanoparticles could reprogram CD8-positive T cells in vivo, achieving tumor control in humanized mice and B-cell depletion in non-human primates. Aera Therapeutics announced in September 2025 that it had nominated its own candidate, AERA-109, a B cell-targeted in vivo CAR-T delivered via lipid nanoparticles, and has targeted a Phase I clinical trial start in mid-2026. Unlike MagicRNA’s approach, AERA-109 is designed to avoid the need for lymphodepleting chemotherapy before treatment, a step that currently adds toxicity and hospitalization to the conventional CAR-T protocol.
The convergence of multiple programs on the same technical approach — mRNA delivered via lipid nanoparticles to reprogram T cells in vivo — reflects the lessons drawn from the rapid maturation of mRNA vaccine technology during the COVID-19 pandemic. The same ionizable lipid chemistry that delivers SARS-CoV-2 spike protein mRNA into muscle cells has been adapted to deliver CAR-encoding mRNA to specific immune cell populations, a repurposing that would have seemed speculative before 2021.
Open Questions and Context
The MagicRNA study enrolled only five patients, and the follow-up period reported in the NEJM letter was three months — too short and too small to draw conclusions about long-term remission, optimal dosing, or comparative efficacy against the ex vivo approach that generated more extensive data in SLE over the preceding three years. The dose-escalation phase of the HN2301 trial is continuing.
SLE affects an estimated five million people worldwide, predominantly women, and currently has no cure. Standard treatment — hydroxychloroquine, corticosteroids, and biologic agents such as belimumab — can control symptoms for many patients but leaves a substantial minority with refractory disease and progressive organ damage, particularly to the kidneys. The ex vivo CAR-T approach pioneered at Erlangen showed durable drug-free remissions in a small series, but its complexity limits deployment to a handful of academic centers globally.
If the in vivo approach can be confirmed in larger trials, it could change that calculus significantly. A therapy that can be manufactured once, stored, and administered as a standardized infusion — without the logistical requirements of personalized cell manufacturing — would be amenable to deployment at the scale the disease demands. Whether the transient nature of mRNA-based CAR expression can achieve the same depth and durability of immune reset as permanent viral-vector engineering remains the central clinical question the field now faces.