BCMA-Targeted CAR-T Cell Therapy for Generalized Myasthenia Gravis
(2026)Objective
To evaluate efficacy and safety of DSG3-CART (BCMA-targeted RNA CAR-T) in AChR-antibody-positive generalized MG
Study Summary
• A single infusion of RNA-based BCMA-targeted CAR-T cells produced clinically meaningful improvement in 67% of gMG patients vs 27% placebo at 3 months, with 83% sustaining response at 12 months and no CRS or neurotoxicity.
Intervention
DSG3-CART single IV infusion vs placebo infusion
Inclusion Criteria
AChR-antibody-positive gMG, MGFA class II–IV, inadequate response to ≥2 standard therapies
Study Design
Arms: Array
Patients per Arm: 15 DSG3-CART, 11 placebo
Outcome
Primary: MGC ≥5-point improvement at Month 3: 66.7% vs 27.3% (P=0.0472). Mean MGC change -7.1 vs -2.5. QMG ≥5-point improvement 53.3% vs 18.2%. 83% sustained response at 12 months.
Bottom Line
A single infusion of RNA-based BCMA-targeted CAR-T cells produced clinically meaningful improvement in 67% of gMG patients vs 27% placebo at 3 months, with 83% sustaining response at 12 months and no CRS or neurotoxicity.
Major Points
- First randomized controlled trial of CAR-T therapy in any autoimmune neurological disease. Uses RNA-based CAR-T cells that naturally decay — no lymphodepletion required.
- Primary endpoint met: MGC ≥5-point improvement at Month 3 in 66.7% (10/15) vs 27.3% (3/11) placebo (P=0.0472).
- Mean MGC change from baseline: -7.1 (DSG3-CART) vs -2.5 (placebo). QMG ≥5-point improvement 53.3% vs 18.2%.
- 83% of responders sustained response at Month 12, suggesting durable benefit despite transient nature of RNA CAR-T cells.
- AChR antibody levels declined significantly in treatment group, consistent with BCMA-mediated plasma cell depletion.
- Zero cytokine release syndrome, zero neurotoxicity (ICANS), zero serious treatment-related adverse events — dramatically better safety than DNA-based CAR-T.
Study Design
- Study Type
- Randomized, double-blind, placebo-controlled
- Randomization
- Yes
- Blinding
- Double-blind
- Sample Size
- 26
- Follow-up
- 12 months
Primary Outcome
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 27.3% | 66.7% | - | 0.0472 |
Limitations & Criticisms
- Very small sample size (n=26) limits generalizability and statistical power
- Only AChR-Ab+ patients — unclear if this approach works in seronegative or MuSK-Ab+ MG
- Dose cohorts pooled for analysis — optimal dose not established
- 12-month follow-up insufficient to determine true durability; re-dosing may be needed
- Placebo response rate of 27% is notable — may reflect natural disease fluctuation
Citation
Nat Med. 2026.