Randomized, Phase II Dose Optimization Study of Chimeric Antigen Receptor Modified T Cells Directed Against CD19 (CTL019) in Patients with Relapsed, Refractory CLL

BACKGROUND:

We have shown that infusion of autologous T cells genetically modified to express a chimeric antigen receptor (CAR) consisting of an external anti-CD19 domain, with the CD3z and 4-1BB signaling domains (CTL019 cells), can mediate potent anti-tumor effects in patients (pts) with advanced, R/R CLL and other CD19+ malignancies. In our initial pilot study, doses of 0.14-5.9 x 10⁸ CTL019 cells were given to 14 pts with R/R CLL and the overall response rate (PR plus CR) was 57%. The majority of responses were sustained, and associated with marked expansion and long-term persistence of transduced cells. Given that CARTs can be self-replicating, it is unknown if there is an optimal biologic dose or merely a necessary threshold dose that must be achieved. To better define an optimal dose, we are performing a randomized phase II study of 2 doses of CTL019 cells in pts with R/R CLL.

METHODS:

Eligible pts have relapsed or persistent CLL after at least 2 previous treatments and progressed within 2 years of their last therapy. Pts are randomized to receive either 5x10⁸ or 5x10⁷ CTL019 cells. All pts receive lymphodepleting chemotherapy ending 3-5 days before T cell infusion. In stage 1, 12 evaluable pts per arm will be treated and in stage 2, an additional 8 pts will be treated at the selected dose. Data is examined through 7/1/14.

RESULTS

To date, 30 pts have been randomized; T cells did not adequately expand in 1, 1 pt was not eligible after screening and 26 pts were treated. 13 pts have been randomized to the higher dose (5 x 10⁸ CTL019 cells) and 13 pts have been randomized to the lower dose (5 x 10⁷ CTL019 cells). Pts were 67% men with a median age of 62 yrs (range 54-76). 10 pts had a known mutation of p53 or 17p del, and 2 pts had failed prior ibrutinib. Pts had received a median of 3 prior therapies (range 2-10) and all had active disease at the time of CTL019 infusion. 3 pts who did not receive the targeted dose are evaluable for toxicity but not response, leaving 23 are evaluable for response.

Median follow-up for all 26 pts was 7.3 mo (range 1-16 mo). Of 23 evaluable pts, 5 (22%) achieved a CR and 4 (17%) achieved PR, for an overall response rate of 35% (95% exact CI 20-61%). 3 responding pts progressed; 1 pt in CR and 1 in PR progressed at 1 and 6 mo after infusion with aggressive CD19-negative transformed lymphoma, and one pt who achieved PR progressed 9 mo after infusion with CD19+ CLL having lost persisting CTL019 cells.

14 of 26 pts experienced a delayed CRS. We devised a novel CRS grading scale that will be presented. Detailed cytokine profiles show marked elevations in IL6 and other cytokines during CRS CRS required intervention in 3 pts with tocilizumab (with or without steroids and additional anti-cytokine therapy) for hemodynamic or respiratory instability and was effective in reversing symptoms of CRS in all pts. 2 of these 3 responded to CTL019. There was no treatment-related mortality.

This preliminary analysis does not suggest a dose:response or dose:toxicity relationship. 6 of 11 evaluable pts on the high dose responded, and 3 of 12 recipients at the lower dose responded (p=0.21). 6 of 11 evaluable on the high dose and 7/12 recipients of the low dose experienced CRS (p=1.0). The 13 evaluable pts who experienced CRS included 6 of 9 responding pts and 7 of 14 non- responders (p=0.67), and one pt not evaluable for response. All statistical tests were based on Fisher’s exact test.

CTL019 expansion in-vivo was noted in all responding pts. By flow cytometry, at peak expansion CTL019 cells represented 0.1-84.6% of CD3+ cells in the 9 responding pts (median 14.7%) and 0-10.6% (median 0.35%) in the 14 non-responders.

Neither response nor expansion correlated with typical patient or disease characteristics including age, number of prior therapies or mutated P53. Data regarding T cell phenotype before and after CAR modification will be presented.

CONCLUSIONS:

In this ongoing dose optimization study of CTL019 cells, 9 of the first 23 evaluable pts have responded. As yet, there is no suggestion of a dose:response or dose:toxicity relationship at the doses being studied. CTL019 cells from R/R CLL pts can undergo robust in-vivo expansion and are associated with a CRS, manageable with anti-cytokine directed therapy. A novel grading scale was developed to better capture the clinical severity of CRS and help guide timing of intervention. This ongoing trial confirms that CTL019 cells can induce potent and durable responses for pts with advanced, R/R CLL with manageable toxicity.