Trispecific CAR T Cells Demonstrate Safety, Early Activity in B-Cell Lymphoma

A trispecific chimeric antigen receptor (CAR) T-cell product targeting CD19, CD20, and CD22 with an OX40 costimulatory domain was safely administered to patients with relapsed/refractory B-cell malignancies, with particularly encouraging activity in B-cell non-Hodgkin lymphoma (B-NHL), according to results from a first-in-human phase 1 trial (NCT05418088) published in Blood Cancer Discovery.

Successful manufacturing was achieved in 16 of 17 patients, with 1 failure occurring in a heavily pretreated patient with Richter transformation. The median manufacturing time was 7 days (range, 6-10), the median transduction efficiency was 38.8% (range, 16.9%-57.6%), and the median fold expansion was 5.4 (range, 1.4-12.9).

No dose-limiting toxicities (DLTs) were observed across the 3 dose levels tested. Cytokine release syndrome (CRS) occurred in 25% of patients and was limited to grade 1 or 2, with a median onset of 8.5 days after infusion. No patients developed immune effector cell–associated neurotoxicity syndrome (ICANS) or hemophagocytic lymphohistiocytosis (HLH). The non-relapse mortality was 0, and the day-100 mortality rate of 18.75% was attributed entirely to disease progression.

Across all 16 treated patients, the overall response rate (ORR) was 50%, and the complete response (CR) rate was 38%; partial responses, stable disease, and progressive disease occurred in 13%, 25%, and 25% of patients, respectively. The CR rate was 83% among patients with B-NHL (n = 6), with 4 of the 5 complete responders remaining in CR more than 1 year after infusion. CRs were seen at all dose levels and included patients with mantle cell lymphoma, follicular lymphoma, and diffuse large B-cell lymphoma. The CR rate was 25% in B-cell acute lymphoblastic leukemia (B-ALL; n = 4), all patients with chronic lymphocytic leukemia/prolymphocytic leukemia (CLL/PLL; n = 4) had stable disease, and no responses occurred among those with Richter transformation (n = 2).

In the full cohort, the 1-year relapse-free survival (RFS) rate was 27.5% (95% CI, 8.6%-50.7%), and the 1-year overall survival (OS) rate was 60.6% (95% CI, 32.3%-80.1%). Among patients with B-NHL, the 1-year progression-free survival (PFS) rate was 67% (95% CI, 19%-90%), and the 1-year OS rate was 100%.

“[T]his phase 1 study enrolled a heterogeneous population of patients with B-cell malignancies and demonstrated a favorable safety profile. Given efficacy signal in B-NHL with minimal toxicity, exploration of trispecific CAR T in this disease population seems justified,” lead author Sumithira Vasu, MD, professor of internal medicine in the Division of Hematology at The Ohio State University James Comprehensive Cancer Center, concluded in the publication with study coinvestigators.

The open-label, single-institution, dose-finding study followed a standard 3 + 3 design. Of 23 patients screened, 18 were enrolled and 16 were treated across 3 cohorts: cohort A (n = 6) included CLL, B-PLL, and B-NHL with lesions smaller than 5 cm; cohort B (n = 9) included B-ALL, B-NHL with lesions larger than 5 cm, and Richter transformation; and cohort C (n = 1) enrolled pediatric patients. Eligible patients had progressed after at least 2 prior lines of therapy. The median age was 64 years (range, 11-73), and patients had received a median of 3 prior lines of therapy (range, 2-7); 2 patients had received prior CAR T-cell therapy.

Patients received lymphodepleting cyclophosphamide and fludarabine, prophylactic tocilizumab (Actemra), and CAR T cells at 5 × 10^5/kg, 1 × 10^6/kg, or 2 × 10^6/kg. Cohort A received a single fresh infusion, whereas cohorts B and C received 40% of the dose fresh on day 0 and 60% as a cryopreserved product on day 7.

The coprimary end points of the study were feasibility and safety to determine the maximum tolerated dose (MTD). Secondary end points included ORR, minimal residual disease–negative remission, duration of remission, relapse rates, RFS, and 1-year OS.

CAR T-cell expansion did not correlate with dose or response. However, spectral flow cytometry in patients with lymphoma showed that pretransduction CD8-positive T cells from nonresponders were enriched for a cluster expressing exhaustion markers including PD-1, TIM-3, TOX, and TIGIT, whereas a less activated, TCF1-positive cluster was enriched in complete responders. By days 7 and 14 after infusion, exhausted CAR T-cell phenotypes were associated with progression, and memory-like phenotypes with response. Relapses were predominantly CD19-positive, suggesting that limited persistence rather than antigen escape drove treatment failure.

The investigators cautioned that the sample size was small and that the study introduced multiple novel variables simultaneously, making it unclear whether early exhaustion stemmed from antigen overstimulation or the OX40 domain. They also noted that activity was modest in B-ALL and CLL and that validation in a larger cohort is needed.

Reference

Vasu S, Denlinger N, Song NJ, et al. Safety and clinical outcomes of a first-in-human trial of point-of-care manufactured trispecific CAR T cells targeting CD19, CD20, and CD22. Blood Cancer Discov. 2026;7(5):711-725. doi:10.1158/2643-3230.BCD-26-0186