T-Cell Engagers in Autoimmunity: CRS From 12% to 67% Across Six Reports Save

Six ACR 2026 reports put BCMA, CD19, CD20 and gamma-delta engagers side by side in autoimmune disease. The efficacy signal is consistent. The safety signal depends almost entirely on how the molecule was built.

Bispecific T-cell engagers redirect a patient's own T cells against B-lineage targets without the manufacturing delay of a cell product. Six reports at ACR Convergence 2026 tested that idea across four targets in autoimmune disease: BCMA, CD19, CD20 and a gamma-delta approach. Response signals were broadly similar. Cytokine release syndrome ranged from 12% to 67%, a spread that tracks molecular design more closely than it tracks the patient population.

The mechanistic question came first. Gauthier and colleagues (Université Paris-Saclay and Hôpital Bicêtre) measured soluble BCMA in 120 plasma samples from patients with rheumatoid arthritis, Sjögren's disease and lupus (Abstract 1114). sBCMA was higher in autoimmune disease than in healthy controls and tracked disease activity in Sjögren's (r=0.42, p=0.004). It correlated tightly with BCMA-positive plasma cells counted in minor salivary gland biopsies (r=0.80, p=0.003) and barely at all with BCMA expression on circulating B cells (r=0.07, p=0.70). The same group then addressed a clinical puzzle: why a plasma-cell-directed drug empties the whole B-cell compartment. In 48-hour PBMC cultures, teclistamab killed switched memory B cells and plasmablasts, the subsets carrying BCMA above a minimal threshold. Surviving naive B cells responded to T-cell engagement by upregulating CD38 and BCMA and downregulating CD20, which made them targets in turn. CD3/CD28 beads reproduced the effect with no BCMA arm involved. One refractory Sjögren's patient treated in vivo followed the same sequence, reaching pan-B-cell depletion at one month.

Clinical data came from Charité. Albach and colleagues collected 18 patients treated off-label with teclistamab at five European centres: systemic sclerosis (n=10), inflammatory myopathy (n=4), lupus (n=2), undifferentiated connective tissue disease (n=1) and IgG4-related disease (n=1). Median five prior therapies, median follow-up 5.1 months. Major clinical responses occurred in 11 patients (61%) and minimal-to-moderate responses in 4 (22%), with background immunosuppression discontinued. The cost column is long. Grade 1/2 CRS in 12 patients (67%), hypogammaglobulinemia below 4 g/l in all 17 evaluable patients, severe IBD-like colitis in 2, and two deaths among patients with advanced cardiac scleroderma, at 20 and 24 days. Four patients had relapsed by data cut, the first at 31 weeks.

Attenuating the CD3 arm changes that picture. Masri and colleagues pooled 83 patients treated with cizutamig, a BCMA×CD3 engager built with reduced CD3 affinity, across myasthenia gravis, lupus, RA, scleroderma, thyroid eye disease and other indications (Abstract 1629). CRS occurred in 10 patients (12%), all grade 1 except one grade 2. No ICANS. Grade 3 or higher infections in 4 patients (5%), no deaths, mean follow-up 33 weeks. There is no head-to-head comparison with teclistamab and the populations differ, but a fivefold gap in CRS is hard to attribute to case mix alone.

The CD19 and CD20 reports point the same way. CC312, a CD19/CD3/CD28 tri-specific that supplies costimulation in cis, produced only grade 1 CRS in 11 lupus patients, depleted B cells in bone marrow as well as blood, and held SRI-4 beyond 48 weeks in 6 of 7 patients dosed at 10 µg or above (Abstract 1212). Plamotamab, given subcutaneously to 9 patients with longstanding RA, caused one grade 1 CRS and lowered DAS28-ESR by a mean of 3.3 points at the highest dose level by week 12 (Abstract 1417). In both, reconstituting B cells were predominantly naive and memory recovery lagged behind.

INB-619 changes the effector cell (Abstract 0965). It engages pan-gamma-delta T cells, and in SLE donor PBMCs it expanded both Vδ1 and Vδ2 subsets while clearing B cells by day 6, with no expansion of CD4 or CD8 αβ T cells. Blinatumomab and mosunetuzumab depleted B cells without that expansion. These data are preclinical.

Depth of depletion now looks like a solved problem. What stays open is durability, the hypogammaglobulinemia that followed every teclistamab course, and whether the low-CRS designs hold their efficacy with longer exposure. The next data to watch are the twelve-month cizutamig and CC312 cohorts.