Safety and Response of Incorporating CD19 Chimeric Antigen Receptor T Cell Therapy in Typical Salvage Regimens for Children and Young Adults with Acute Lymphoblastic Leukemia

CD19 chimeric antigen receptor (CAR) T cells have shown significant promise in multiple early phase trials including our own (Lancet 385:517-28). We manufacture CAR T cells containing CD28 and CD3z domains in 7 days using a retroviral platform. Several challenges remain to its widespread use: 1) reduction in the incidence of grade 4 cytokine release syndrome (CRS) and 2) incorporation with standard salvage regimens. Here, we update our experience with 39 patients.

In the first 21 patients we defined the maximally tolerated dose as 1x10⁶ CAR T cells/kg, grade 4 CRS occurred in 16%, and noted that severity of CRS correlated with disease burden. We stratified the current cohort (n=18) by disease burden. Subjects 1-21 and subsequent patients with low burden disease (Arm 1: isolated CNS disease or <25% marrow blasts) received a low dose preparative regimen of fludarabine (25 mg/m2/day Days-4 to -2) and cyclophosphamide (900 mg/m2 Day-2). Those with high burden disease (Arm 2: ³25% marrow blasts, circulating blasts or lymphomatous disease) received a high dose regimen to reduce tumor burden prior to cell infusion in an attempt to decrease severity of CRS. Arm 2 regimens were individualized based on prior therapies and risk from comorbidities. FLAG (n=6), ifosfamide/etoposide per AALL0031 (IE; n=2) and high dose fludarabine (30 mg/m2/day Days -6 to -3) with cyclophosphamide (1200 mg/m2/day Days -4 and -3) (HD flu/cy; n=3) were used.

All products in the second cohort met cell dose though contaminating monocytes tended to inhibit maximal growth and transduction (see companion abstract by Stroncek). All patients received 1x10⁶ CAR T cells/kg. Using grading criteria and an algorithm for early intervention to prevent grade 4 CRS (Blood 124:188-95) no grade 3 and only 1 grade 4 (5.6%) CRS occurred. Having significant comorbidities, Pt 34 was electively intubated for airway protection, did not require vasopressors, and rapidly recovered after tocilizumab and steroids. A brief seizure occurred, though he had a history of seizures. None others in the current cohort had neurotoxicity.

Using intent to treat analysis, the complete response (CR) rate was 59% overall and 61% in ALL. 13/16 (81%) low burden and 10/22 (46%) high burden ALL patients had a CR across both cohorts. Low burden patients treated on either cohort had similar CR rate of 8/10 (80%) and 5/6 (83%). Although not statistically significant and underpowered, 7/11 (64%) high burden patients treated with low dose flu/cy had a CR while 3/11 (27%) had a CR with high dose regimens. Specifically, 3/6 (50%) receiving FLAG achieved MRD-CR while none receiving IE or HD flu/cy responded. 8/8 with primary refractory ALL had MRD-CR regardless of disease burden or preparative regimen raising the prospect that T cell fitness in these patients was superior to others.

Of the 20 patients achieving an MRD-CR, the median leukemia free survival (LFS) is 17.7 months with 45.5% probability of LFS beginning at 18 months. Only 3 did not have a subsequent hematopoietic stem cell transplant as their referring oncologist determined the risk of such was unacceptable. Two relapsed with CD19-leukemia at 3 and 5 months, while 1 remains in CR with detectable CAR T cells at 5 months. Reliance on multiple infusions of cells is problematic as 0/5 CD19+ patients receiving a second dose responded.

Preclinical models have demonstrated that T cell exhaustion has a role in limiting the efficacy of CAR T cells. We evaluated CAR products and the T cells used to generate them for phenotypic markers of exhaustion and will present data evaluating the relationship between these and response. Our results demonstrate that CD19 CAR T cell therapy is safe and effective with aggressive supportive care and use of an early intervention algorithm to prevent severe CRS and provides a potential for cure in primary refractory ALL.