In Vivo CAR-T Could Reset the Immune System in Autoimmune Disease Without Weeks of Cell Manufacturing

Autoimmune diseases such as lupus, rheumatoid arthritis, and multiple sclerosis involve a breakdown of immune tolerance. Glucocorticoids, conventional immunosuppressants, and biologics such as rituximab and anti-BAFF antibodies often fail to produce lasting remission. They suppress immunity broadly, which raises the risk of infection, malignancy, and cumulative organ damage, and they often miss long-lived plasma cells and tissue-resident B cells that keep producing autoantibodies. This gap is why immune system 'resetting' is attracting so much interest.

This narrative review looks at CAR-T cell therapy for autoimmunity and at the newer in vivo approach. Ex vivo CD19 CAR-T cells have shown rapid B-cell depletion, normalized serum markers, and drug-free remission in many patients with refractory lupus, inflammatory myopathy, and systemic sclerosis. The authors argue that this approach is hard to scale to chronic, non-fatal disease. Autologous manufacturing takes several weeks from vein to vein, can fail because of poor T-cell yield or expansion problems, and can leave exhausted cells. Commercial products cost roughly USD 373,000–475,000, and total treatment costs can exceed USD 500,000 to 1,000,000. Cytokine release syndrome and neurotoxicity remain unpredictable. Broad CD19 or BCMA targeting also removes protective B cells, risking prolonged hypogammaglobulinemia and infections. Antigen escape and plasma cells that resist CD19-directed therapy add biological limits.

In vivo CAR-T delivers CAR-encoding genetic material directly to a patient's T cells, so no cells are collected, expanded, or cryopreserved. The authors describe two platform classes. Viral vectors (lentiviral and adeno-associated) provide durable expression. Lentiviral vectors carry about 8–10 kb and integrate into the genome, so insertional mutagenesis, off-target transduction, and prolonged CAR persistence are concerns. AAV carries about 4.7 kb, rarely integrates, but has common pre-existing neutralizing antibodies, liver tropism, and limited T-cell targeting. Ionizable lipid nanoparticles carrying modified mRNA give transient, controllable CAR expression without genomic integration and are more scalable to manufacture. Mitigation strategies in the comparison table include self-inactivating vector designs, integration-site mapping, T-cell-selective targeting ligands, dose control, and suicide switches.

The authors say preclinical and emerging clinical data show these platforms can quickly reprogram circulating T cells, clear B cells potently, and improve refractory disease. They expect that cutting production time from weeks to hours or days, along with lower cost and adjustable CAR expression, could make treatment accessible to far more patients. They also note complementary innovations that may combine with in vivo delivery: CAAR-T cells that spare protective B cells, CAR-Tregs, allogeneic products, and logic-gated CAR circuits. They call for work on delivery specificity, safety engineering, trial design, and cost-effective production.

Caveats: this is a narrative review rather than a systematic analysis or new trial, so it offers no pooled efficacy or safety estimates. Human in vivo CAR-T data are described as emerging and limited, particularly for AAV, and conclusions rest largely on animal and early-stage evidence. The long-term safety, durability, and real-world cost advantages remain unproven.