Phase 1 Trial of CD30-Directed, CCR4 Co-Expressing Chimeric Antigen Receptor T-Cells in Patients with CD30+ Cutaneous T-Cell Lymphomas Refractory to Brentuximab Vedotin

Background CD30 CAR-T (CD30.CART) has shown activity in CD30+ lymphomas, including T-cell lymphomas (TCLs) and Hodgkin lymphoma (HL). We conducted a phase 1 trial in patients (pts) with cutaneous TCL (CTCL) and HL to study CD30.CART modified to co-express the chemokine receptor CCR4 (CCR4.CD30.CART) to improve T-cell trafficking along the CCL17 chemokine gradient in these malignancies (NCT03602157). The primary objective of maximum tolerated dose (MTD) was assessed in the complete trial population. Here we report the secondary objective of objective response rate (ORR, rate of CR or PR) at week (wk) 6 after CCR4.CD30.CART for the CTCL cohort. Methods Eligibility included: age ≥18; CD30+ mycosis fungoides (MF), Sézary syndrome (SS), primary cutaneous ALCL (pcALCL), or lymphomatoid papulosis with active disease prior to cell procurement and lymphodepletion (LD); failure of ≥2 lines of systemic therapy (tx) including brentuximab vedotin (BV); Karnofsky performance status (KPS) >60%; and adequate organ function. Prior to LD, pts could not receive CD30 antibodies for 4 wk or other systemic tx or radiation (RT) for 3 wk. Pts underwent LD with fludarabine (flu) 30 mg/m2/d and bendamustine (benda) 70 mg/m2/d -5, -4, and -3 d (+/- 1 d) before CAR-T infusion (d0). Pts intolerant to benda received cyclophosphamide (cy) 500 mg/m2/d for 3 d. We used 3+3 dose-escalation. Pts on dose levels (DLs) 1, 3 and 5 received CCR4.CD30.CART at 2x107, 5x107, and 1x108 CAR+ cells/m2 respectively; those on DLs 2, 4, and 6 received CCR4.CD30.CART at the preceding DL plus 1x108 CD30.CART/m2 to compare kinetics and biodistribution within the same pt. Results Six pts with CTCL enrolled 4/7/2020-9/6/2023. Baseline characteristics by frequency or median (range) were: age 63.9 years (25.2-70.9); 4 MF, 1 SS, and 1 pcALCL; 1 female pt; 1 black and 5 white pts; prior lines of systemic tx 3.5 (2-16); and 1 prior allogeneic stem cell transplant. Bridging tx were: 2 gemcitabine, 1 BV, 1 lenalidomide, 1 pembrolizumab, and 1 no bridging tx. Prior to LD, all pts had active disease. Median mSWAT was 32.5 (range 5.8-84.7). For LD, 5 pts had flu/benda; 1 had flu/cy. Three pts received CCR4.CD30.CART at DLs 1, 2, and 5, respectively, and 3 received both CCR4.CD30.CART and CD30.CART at DL6. No grade (G)5 adverse events (AEs) or dose-limiting toxicities occurred. No pts had CRS or ICANS. After infusion, G3-4 AEs occurring in >1 pt were all hematologic: 3 pts with neutropenia, 6 with lymphopenia, 2 with anemia, and 2 with thrombocytopenia (TCP). For G≥3 neutropenia, anemia, and TCP, median days to resolution to G≤2 (range) were: 80 (27-84), 131 (28-234), and 137.5 (49-226) respectively. Two pts had severe AEs. One who had recurrent diverticulitis in the 6 mo before CAR-T had G3 diverticulitis, G2 fall, and G2 confusion on d+34. The other had G3 neutropenia as well as G1 fever attributed to an infected CTCL lesion on d+1. Wk+6 response rates were: ORR 50% (80% confidence interval [CI] 20-80), CR rate 0% (80% CI 0-32), PR rate 50% (80% CI 20-80), SD rate 50% (80% CI 20-80), and PD rate 0% (80% CI 0-32). One of 2 MF/SS pts with large cell transformation had PR. Median change in mSWAT at wk+6 was -42.2% (80% CI -87.8- -4.3). All pts required subsequent tx. From CAR-T infusion, median follow-up was 26.0 mo; Kaplan-Meier median time to next systemic tx or death was 7.4 mo (80% CI 1.9-23.7); median time to next RT, systemic tx, or death was 6.9 mo (80% CI 1.9-7.1); and median OS was 23.9 mo (80% CI 17.3-not reached). The 3 deaths were associated with CTCL progression. CCR4.CD30.CART expanded in the blood in all pts. Tumor biopsy (bx) from a pt who received both products showed CCR4.CD30.CART alone, whereas CCR4.CD30.CART and CD30.CART were present in blood. Conclusions CCR4.CD30.CART has promising safety with low rates of G≥3 non-hematologic AEs. Despite the heavily pretreated population and within the limitations of our small sample size, CCR4.CD30.CART wk+6 ORR is similar to best ORR of BV in ALCANZA (65.6%), mogamulizumab in MAVORIC (34.9%), and pembrolizumab in CITN-10 (37.5%) (Prince et al., Lancet 2017; Kim et al., Lancet Oncol 2018; Khodadoust et al., JCO 2019). Tumor bx with only CCR4.CD30.CART, despite presence of both products in blood, supports the importance of CCR4-mediated trafficking for clinical activity. As with prior studies, responses were brief. Enrollment continues at the MTD in an expansion phase to better understand CCR4.CD30.CART efficacy in this difficult-to-treat population.