IL18-armored CAR T cells in patients with CLL/SLL and Richter's transformation after prior BTK inhibitor and venetoclax failure Free

Background: The rate of complete remission (CR) in the pivotal trial that led to FDA approval of lisocabtagene maraleucel (liso-cel) for chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) was only 18% in patients (pts) after BTK inhibitor (BTKi) and venetoclax (ven) failure (Siddiqi et al, Lancet 2023). One strategy to improve the outcome of cellular therapies is armoring them to secrete cytokines that enhance anti-tumor activity. We previously described huCART19-IL18, an autologous CD19-directed CAR T-cell product secreting interleukin 18 and our experience in non-Hodgkin lymphomas (Svoboda et al, NEJM 2025).

Methods: Our trial using huCART19-IL18 for CD19+ malignancies includes a separate cohort of adult pts with CLL/SLL and RT. For CLL/SLL, pts must have disease relapsed/refractory after at least 2 lines of therapy and must have previously received or be intolerant to BTKi and ven. RT pts must have received at least 1 prior line of RT therapy. The primary objective is to determine the safety and maximum tolerated dose of huCART19-IL18 in CLL/SLL and RT. The secondary objectives are to evaluate manufacturing feasibility, determine preliminary efficacy, and perform correlative studies. Bridging therapy is optional. The starting dose level (DL) was 7x10⁶ of huCART19-IL18 with escalation and de-escalation strategies. The product is administered as a single intravenous infusion 2 to 5 days after lymphodepleting (LD) chemotherapy with bendamustine (90 mg/m²) x 2 days. Dose-limiting toxicity (DLT) observation period is 28 days after infusion. Responses are assessed using iwCLL criteria for CLL/SLL pts and Lugano criteria for RT pts at months (M) 3, 6, 9, and 12 following huCART19-IL18 infusion.

Results: As of July 1, 2025, 10 pts were enrolled. One pt is scheduled for infusion and 2 pts became clinically unstable while awaiting treatment and were not infused. Seven pts are evaluable for safety and efficacy. Six pts had CLL/SLL and 1 had RT at enrollment. The median age was 62 years (range 54-76), and majority were male (71%). Four (57%) had 17p deletion by FISH, 1 (14%) had p53 mutation by sequencing. The median number of prior therapies was 5 (range 4-13). All 7 pts (100%) had relapsed disease after prior covalent BTKi, ven, and anti-CD20 monoclonal antibody; 2 (29%) had prior CAR T-cells (liso-cel). All pts had successful manufacturing at the protocol-assigned target DL. Two pts received bridging therapy (obinutuzumab in 1, radiation in 1); an additional 3 were on a novel agent that started prior apheresis and continued until LD chemo (BTKi in 2, ven in 1). The median vein-to-vein time was long at 82 days (range 27-117) due to the protocol mandated safety staggers. Low grade cytokine release syndrome (CRS) occurred in all 7 (100%) pts, grade (G) 1 in 3 and G2 in 4. The median CRS duration was 6 days (range 2-10), and tocilizumab was used in 5 (71%) cases. Immune effector cell-associated neurotoxicity syndrome (ICANS) occurred in 2 of 4 (50%) pts treated at the initial DL (7x10⁶ cells), both G3. While transient (4 and 9 days), these fulfilled DLT criteria and resulted in de-escalation to 3x10⁶ cells for subsequent pts. There were no additional episodes of ICANS or DLTs in the 3 pts treated at this de-escalated DL. Other G3 or higher adverse events (AE) at least possibly related to the study product included hypoxia in 2 (29%), transaminitis in 1 (14%), and neutropenia in 1 (14%). There were no study-related deaths. The best ORR was 43% (90% CI: 13-77%), all complete responses (CR), including 2 pts with CLL/SLL and 1 with RT. Two CRs occurred at the de-escalated DL (3x10⁶ cells) including in one patient with prior liso-cel failure. One patient had SD and 3 had PD as their best responses. With the median follow-up 17.6 months (range 3.9-24.8), all responders continue to be in remission, including 1 patient with 2 years of follow-up. One-year PFS probability is 43% (90% CI: 14-70) and 1-year OS probability is 86% (90% CI: 45-97). Median peak expansion of huCART19-IL18 was 73,249 copies/µg DNA (IQR 37,323-89,125) in responders and 852 copies/µg DNA (IQR 538-6,576) in non-responders.

Conclusions: In the CLL/SLL and RT cohort, huCART19-IL18 produced durable remissions in heavily pre-treated pts after prior failure of a BTKi and ven. Following DL de-escalation, the treatment had a manageable toxicity profile. Our data supports further expansion of this cohort.