Phase 2 Study of CD19 CAR T Cells with a Fully Human Binder for Large B-Cell Lymphoma in Relapse or Progression after a Murine Binder-Bearing CD19 CAR T-Cell Therapy

Abstract

BACKGROUND T cells engineered with a CD19-targeted chimeric antigen receptor (CD19 CAR T cells) have revolutionized the treatment of high-risk, relapsed, or refractory large B-cell lymphoma (LBCL), but durable responses are only achieved in 30-40% of patients (pts). Outcomes are dismal in case of disease recurrence after CD19 CAR T-therapy; in this situation we have previously reported low efficacy and limited duration of response after a second infusion of the same CD19 CAR T-cell product (Gauthier, Blood, 2022). We observed that repeat CD19 CAR T-cell infusions were impaired by CD8+ T-cell responses against murine amino acid sequences located in the single chain variable fragment (scFv) of the CAR, which led to rapid immune rejection of CAR T cells. We hypothesized that CAR T cells engineered with a fully human scFv could circumvent anti-murine scFv immune responses in pts previously exposed to murine scFv-bearing CAR T cells. We report our phase 2 clinical trial results investigating treatment with fully human scFv-bearing CD19 CAR T cells (JCAR021) in R/R LBCL pts with relapsed or progressive disease after murine scFv-bearing CD19 CAR T-cell therapy. METHODS Pts with R/R LBCL were enrolled at the Fred Hutchinson Cancer Center on a phase II study (NCT03103971). Key eligibility criteria included: i) relapsed disease after complete response (CR) to prior CD19-targeted non-JCAR021 CAR T-cell therapy or persistent disease after partial response to prior CD19-targeted non-JCAR021 CAR T-cell therapy; ii) CD19-positive disease since completing the prior CD19-targeted CAR T-cell therapy. Pts received lymphodepletion (LD) with cyclophosphamide 300 mg/m2/d and fludarabine 30 mg/m2/d for 3 days. Cytokine release syndrome (CRS) and neurotoxicity (NT) were graded according to the Lee 2014 criteria and CTCAE v4.03 criteria, respectively. Disease response was evaluated at day +28 after JCAR021 infusion per PET-CT Lugano criteria. The planned sample size and CAR T-cell dose were n=27 and 7x10 6 CAR T cells/kg, respectively. RESULTS Pt characteristics, and outcomes are shown in the Table. All eight screened pts were enrolled, received LD, and JCAR021 infusion. Median age at enrollment was 63 (range: 41-77). Prior to lymphodepletion, 4 pts (50%) had bulky disease (largest lesion ≥5cm). All pts had extranodal disease. We observed elevated LDH prior to lymphodepletion in 6 pts (75%). Median prior lines of therapy was 5 (range: 3-7). Prior CD19 CAR T-cell product type was lisocabtagene maraleucel, n=4 (50%); axicabtagene ciloleucel, n= 3 (37%); or tisagenlecleucel, n=1. Response to the prior CD19 CAR T-cell therapy was CR in 5 pts (62%) and PR in 3 pts (38%). Median time from the first CAR T-cell infusion to relapse or progression was 165 days (range 49-434). Median time from the first CAR T-cell infusion to JCAR021 infusion was 304 days (range: 202-603). JCAR021 was administered at the dose of 7x10 6 (n=6, 75%) or 2x10 6 CAR T-cells/kg (n=2, 25%; dose reduction due to grade 5 NT in subject #6). We observed all-grade CRS in 4 pts (50%; grade 1, n=2; grade 2, n=2 not requiring vasopressors) and NT in 1 patient (dose limiting toxicity, grade 5 NT from intracerebral hemorrhage with concurrent acute kidney injury and toxic encephalopathy). At day+28 responses were seen in 3 of 8 pts (37%; CR, n=2; PR, n=1), all of whom received 7x10 6 CAR T-cells/kg. Two pts in CR after JCAR021 had bulky disease and elevated LDH prior to LD. In both pts, the prior CD19 CAR T-cell therapy had induced CR lasting ≥ 6 months. In the two evaluable responders (death in CR, n=1), we observed disease relapse or progression within 3 months. Median overall survival was 105 days (95%CI, 92-NA). We measured expansion of JCAR021 CAR T-cells in all pts. Exploratory analyses showed lower in vitro CAR T-cell expansion during manufacturing (fold expansion of CD8+ CAR T-cells, p<0.001) and lower in vivo CAR T-cell expansion (peak CD8+ CAR-T, p=0.002; AUC day0-28, p=0.01; JCAR021 dose, 7x10 6/kg) compared to pts receiving JCAR021 as their first CD19 CAR-T product (CAR-naïve cohort). CONCLUSION In LBCL pts with disease relapse or progression after a first CD19 CAR T-cell therapy, we observed low response rates and lack of durable responses after treatment with the fully human scFv-bearing product JCAR021, prompting early study termination. Lower CAR T-cell expansion during manufacturing in CAR-exposed pts suggest pre-existing T-cell dysfunction as a potential mechanism of failure.