CD19 CAR-T Cells Are Highly Effective in Ibrutinib-Refractory Chronic Lymphocytic Leukemia

BACKGROUND:

Ibrutinib, a Bruton Tyrosine Kinase (BTK) inhibitor causes partial responses (PR) in a majority of patients (pts) with chronic lymphocytic leukemia (CLL). However, complete responses (CR) are rare and high-risk pts who progress on ibrutinib have short survival. Lymphodepletion chemotherapy followed by infusion of CD19-specific chimeric antigen receptor (CAR)-modified T cells has produced encouraging responses in CLL in phase 1 clinical trials, but the majority of pts in those studies had not previously received or failed ibrutinib.

METHODS:

We treated 18 adults with CLL who had previously received ibrutinib with anti-CD19 CAR-T cells that were manufactured from defined CD4⁺ and CD8⁺ T cell subsets obtained by immunomagnetic selection of leukapheresis products, formulated in a final 1:1 ratio of CD8⁺:CD4⁺ CAR⁺ T cells, and infused at 1 of 3 dose levels (2x10⁵, 2x10⁶ or 2x10⁷ CAR-T cells/kg) after lymphodepletion chemotherapy.

RESULTS:

Eighteen pts (median age 60; range 40-73) with a median of 5 previous therapies (range 3-9), including 3 pts that failed prior allogeneic stem cell transplant were enrolled and treated on the study. All pts were refractory to or had relapsed after receiving a regimen containing fludarabine and rituximab, and all pts had previously received ibrutinib; 11 were ibrutinib-refractory, 3 were ibrutinib-intolerant, and 4 were refractory to venetoclax. Twelve pts had a complex karyotype and 11 pts had 17p deletion. The median percentage of abnormal B cells in marrow was 77% (range 0.4 Ð 96). All pts had extramedullary disease and 2 had CNS disease. Lymphodepletion chemotherapy consisted of cyclophosphamide (Cy) 30-60 mg/kg x 1 and fludarabine (Flu) 25 mg/m²/day x 3 days (n=15); Flu 25 mg/m²/day x 3 days alone (n=2); and Cy 60 mg/kg alone (n=1). CAR-T cells were manufactured for all pts and 16/18 received CD4⁺ and CD8⁺ CAR-T cells in the defined 1:1 ratio. Four pts received 2x10⁵, 13 received 2x10⁶ and 1 received 2x10⁷ CAR-T cells/kg.

Seventeen pts have completed response and toxicity assessment. Analysis of all pts with B cell malignancies treated with Cy/Flu and CAR-T cells on our trial showed that the highest dose level (2x10⁷ CAR-T cells/kg) was too toxic for an initial CAR-T cell infusion, and identified a maximum tolerated first dose of 2x10⁶ CAR-T cells/kg. Cytokine release syndrome (CRS) was graded according to Lee et al (Blood, 2014). After a single cycle of lymphodepletion chemotherapy and CAR-T cell infusion, 8 pts developed grade (gr) 0-1, 5 had gr 2, 3 had gr 3, and 1 had gr 4 CRS. Four pts developed gr ³3 neurotoxicity (NT). No gr 5 events were observed, no pts were intubated, and only 1 pt required pressors. Three pts received tocilizumab and dexamethasone to treat CRS and/or NT. Four pts received a second cycle of lymphodepletion chemotherapy and CAR-T cells at a 10-fold higher dose than the 1^st infusion for persistent disease. CRS and NT (gr3) was observed in 3 of 4 pts after the second cycle of therapy.

Restaging was performed 4 weeks after the last CAR-T cell infusion. The ORR was 76% (8 PR and 5 CR). Two of the pts with PR by lymph node size criteria (IWCLL 2008) had negative PET scans after therapy. Among ibrutinib-refractory (n=10) or intolerant pts (n=3), the ORR was 77% (7 PR and 3 CR). In venetoclax refractory pts, 2 of 4 responded (PR). Only 1 of the 3 pts who did not receive Cy/Flu lymphodepletion responded. At day 28, 11 of 13 (85%) pts who received Cy/Flu lymphodepletion and a CAR-T cell infusion at ²2x10⁶ CAR-T cells/kg had complete elimination of marrow disease by flow cytometry; 10/13 (77%) with nodal disease had a PR or CR at restaging, 1 had a mixed response, and 2 had progressive disease (PD).

No malignant IGH sequences were detected in marrow of 4/4 pts in CR who underwent IGH deep sequencing. Pts with CR had a higher peak percentage of CD8⁺ (p=0.006), but not CD4⁺ CAR-T cells in blood. Robust CAR-T cell expansion was seen in some non-responders, which in conjunction with the lower response rate in nodal sites compared to blood, suggests that factors in the malignant lymph node environment may inhibit CAR-T cell activity. No pt in CR has relapsed or died with a median follow-up of 8.4 months. For the 13 pts that received Cy/Flu lymphodepletion and ≤ 2 x 10⁶ CAR-T cells/kg, OS is 100% and PFS is shown in Fig. 1.

CONCLUSION:

CD19 CAR-T cells of defined CD4:CD8 ratio are highly active in CLL and can induce high response rates and durable CRs in poor prognosis pts who have previously failed ibrutinib.

View largeDownload slide

View largeDownload slide

Close modal