Optimizing Chimeric Antigen Receptor T-Cell Therapy for Adults With Acute Lymphoblastic Leukemia

Noelle V. Frey, MD; Pamela A. Shaw, PhD; Elizabeth O. Hexner, MD; Edward Pequignot, MS; Saar Gill, MD, PhD; Selina M. Luger, MD; James K. Mangan, MD, PhD; Alison W. Loren, MD; Alexander E. Perl, MD; Shannon L. Maude, MD, PhD; Stephan A. Grupp, MD, PhD; Nirav N. Shah, MD; Joan Gilmore, BS; Simon F. Lacey, PhD; Jos J. Melenhorst, PhD; Bruce L. Levine, PhD; Carl H. June, MD; David L. Porter, MD


J Clin Oncol. 2020;38(5):415-422. 

In This Article


Study Design

Thirty-five adult patients with r/r ALL were treated at the University of Pennsylvania in 2 single-arm, open-label studies. The studies were designed to determine the safety and efficacy of autologous T cells transduced through a lentiviral vector with a CAR composed of a murine anti-CD19 single-chain variable (scFv) region and T-cell receptor CD3ζ/4-1BB costimulatory domains (CTL019). Our pilot first-in-human study (ClinicalTrials.gov identifier: NCT01029366) was open to all patients with CD19+ malignancies, of whom 5 had r/r ALL and are reported here. The second study (ClinicalTrials.gov identifier: NCT02030847) was a follow-up specific to patients with r/r ALL. Eligible patients in both studies were age ≥ 18 years with r/r CD19+ B-cell ALL. Patients with relapse after blinatumomab or allogeneic hematopoietic stem-cell transplantation (HSCT) were eligible provided that they did not have active graft-versus-host disease. Patients with active disease in their CNS were excluded. The study protocols and their amendments were approved by the institutional review board of the Hospital of the University of Pennsylvania and conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonization of Good Clinical Practice guidelines, including written informed consent. Research that involved recombinant DNA was conducted under biosafety level 2 containment and was approved by the institutional biosafety committee at the University of Pennsylvania in accordance with the National Institutes of Health Guidelines for Research Involving Recombinant DNA Molecules.

Peripheral blood T cells for manufacturing were collected from patients by leukapheresis and stimulated with beads coated with antibodies directed against CD3 and CD28. Cells were transduced with a lentiviral vector that encoded anti-CD19 scFv linked to 4–1BB and CD3ζ signaling domains, as described previously.[8,9] Lymphodepleting chemotherapy before CTL019 infusion was recommended but not mandated for patients with WBC ≤ 1,000/μL. In both studies, the choice of lymphodepletion was left to the discretion of the treating physician.

The dosing and timing of CTL019 varied with protocol amendments in response to toxicity and efficacy. Depending on timing of enrollment, patients were treated in 1 of 3 dosing cohorts (Table 1). Patients either received a planned high dose of CTL019 cells (5 × 108) or a low dose of cells (5 × 10[7]). The cells were then administered as a single infusion, or the total planned dose was split into fractions, with 10% of the dose delivered on day 1 (D1), 30% on D2, and the remaining 60% on D3. In the cohorts treated with a fractionated dosing approach, D2 and/or D3 doses were held if the patient experienced early signs of CRS, including fever (Table 1).

Patients treated in the initial pilot trial were treated with the high-dose fractionated (HDF) approach according to the design of our first-in-human study. In the follow-up study, the first patients were treated with a high-dose single infusion (HDS) of 5 × 108 CAR T cells. Because of CRS-related deaths, the protocol wasmodified, and patients were treated with a lower dose of 5 × 107 cells with either single infusion or fractionated dosing (low-dose [LD]) cohort. Because of low efficacy in the LD cohort, the protocol was again modified, and patients were treated with the higher dose but with the fractionated dosing scheme (HDF), which was safe and effective in the pilot study (Figure 1).

Figure 1.

Flow diagram for adult patients with acute lymphoblastic leukemia treated with CTL019. CR, complete remission; CRS, cytokine release syndrome; NR, no response; HDF, high-dose fractionated; LD, low dose; HDS, high-dose single infusion.

Toxicity Assessment

CRS was graded according to the Penn grading scale[10,11] (Appendix Table A1, online only). Other adverse events were graded according to the National Cancer Institute CTCAE (version 4.03).

Response Assessment

Disease response assessments were done at protocol-defined time points. The primary efficacy end point was overall complete remission (CR) rate at D28. CR was defined as < 5% bone marrow blasts; CRi was defined as CR without full hematopoietic recovery. Patients were determined to be minimal residual disease (MRD) negative if bone marrow blasts were < 0.01% by multiparameter flow cytometry. Secondary efficacy end points were time-to-event analysis of overall survival (OS) and event-free survival (EFS).

Statistical Analysis

Means, medians, and ranges were calculated for groups. OS and EFS were estimated by Kaplan-Meier curves, and dosing cohort differences were tested using the log-rank test. OS and EFS by subsequent HSCT were compared using a landmark analysis, which compares only those patients in both groups that were free of an event at 5.2 months postinfusion, the maximum time of HSCT.[12] This approach addresses the immortality time bias in survival curves caused by HSCT-eligible patients having to remain event free until the time of HCST.[13] D28 response rates and frequency of severe (grade 4/5) CRS by cohort were tested using Fisher's exact test. The association between the number of doses received and baseline disease burden was assessed using Spearman's rank correlation coefficient. Statistics were calculated using SAS 9.4 (SAS Institute, Cary, NC) and R version 3.4 (R Foundation for Statistical Computing, Vienna, Austria) software.