The Promise of Chimeric Antigen Receptor T-Cell Therapy
The Promise of Chimeric Antigen Receptor T-Cell Therapy
Chimeric antigen receptors (CARs) are engineered molecules that can be introduced into T cells to enable them to target specific tumor antigens. CAR T cells targeting CD19 have shown promise in patients with relapsed and refractory B-cell neoplasms, including those with acute lymphoblastic leukemia, chronic lymphocytic leukemia, and non-Hodgkin lymphomas. Notably, durable responses have been observed in patients who had not undergone consolidative stem cell transplant, a finding that correlates with reports of T-cell persistence and B-cell aplasia in studies of anti-CD19 treatment in vivo. Cytokine release syndrome, correlating with activation and expansion of T cells, and neurologic toxicity are the most significant treatment-related adverse effects. Efforts are underway to extend the benefits of immunotherapy with anti-CD19 CAR T cells to other targets and tumor types.
Cancer immunotherapy aims to harness the antitumor potential of the immune system and translate it into effective therapies for patients. One such approach, chimeric antigen receptor (CAR) T-cell therapy, has yielded promising initial clinical results. CAR T-cell therapy employs gene transfer techniques to reprogram endogenous T cells to target a specific tumor antigen. Most of the collective experience with this treatment has involved the use of CAR T cells to target CD19, a pan–B-cell antigen expressed in most B-cell malignancies. Genetically modified T cells expressing an anti-CD19 CAR have shown promise in several clinical trials for patients with acute lymphoblastic leukemia (ALL), chronic lymphocytic leukemia (CLL), and B-cell non-Hodgkin lymphoma (NHL).[1-7] However, CAR T-cell therapy is associated with toxicities including cytokine release syndrome (CRS), a novel inflammatory process correlating with the in vivo activation and expansion of CD19 CAR T cells; neurologic effects; and B-cell aplasia.
The promise of CAR-modified T-cell therapy derives from its combined immunologic benefits; these include the specificity of a targeted antibody, the ability to expand the T-cell population in vivo for an amplified antitumor response, and the potential for long-term persistence to facilitate ongoing tumor surveillance. Given the successes of early-phase trials with anti-CD19 CAR T cells, larger studies are underway to assess the feasibility of evaluating this treatment modality across multiple centers and in larger patient cohorts. In addition, modification of CARs to target new antigens and a variety of tumor types, including solid tumors, is the focus of ongoing investigations. In this review, we will describe the mechanism of action of CAR T cells, discuss outcomes of current clinical trials, and highlight emerging directions for this exciting approach to cancer treatment.
CAR T Cells
CARs are engineered to combine an extracellular single-chain antigen-recognition domain (usually the variable regions of a specific monoclonal antibody) with one or more intracellular T-cell signaling domains. With the use of gene transfer techniques, CARs can then be introduced into normal T cells, redirecting them to target new antigens (such as CD19) in a manner independent of major histocompatibility complexes.[8,9] Transduction of T cells with CARs encoded in lentiviral or retroviral vectors results in permanent modification of the genome and thus the potential for ongoing expression of the CAR protein for the life of the T cell. The components of the CAR signaling domain are critical for maximal activation, expansion, and persistence of CAR T cells, and therefore are a key target for manipulation. The so-called “first-generation” CARs, which included only the antigen-recognition domain with an intracellular CD3ζ signaling domain, had limited clinical activity[10-12] (Figure 1). They were unable to adequately activate T cells in vivo, especially when tumor cells lacked expression of T-cell costimulatory molecules.[8,9,11,13]
Several groups have worked to improve treatment efficacy by developing “second-generation” CARs, which include a costimulatory domain typically derived from CD28 or CD137 (also known as 4-1BB), in addition to the CD3ζ domain. The strong indication from preclinical trials that second-generation CARs would have greater antitumor potency[14,15] has now been proven by the clinically significant treatment outcomes seen in patients with relapsed and refractory (R/R) CD19-positive malignancies (see Figure 1).[1-4,6] Although the optimal costimulatory domain for second-generation CARs is not known, preclinical studies suggest that 4-1BB CARs may improve in vivo expansion and persistence. In clinical studies, 4-1BB CARs have exhibited prolonged persistence when used in the treatment of patients with CLL and ALL, which may correlate with sustained remissions.[4,6,15] Investigation of third-generation CARs (which have multiple costimulatory domains) is underway to assess whether the efficacy of CAR T-cell therapy can be further improved (see Figure 1).
The anti-CD19 CAR T cells used in various clinical trials may differ in regard to the design of the CAR molecule, the method of CAR transfer into T cells, and other aspects of the CAR T-cell culture environment. The basic steps involved in the manufacture of anti-CD19 CAR T cells at the University of Pennsylvania (Penn) and in most other clinical trial programs are summarized in Figure 2. Patients undergo an apheresis procedure to collect autologous T cells. Once collected, T cells are genetically modified with the CAR construct and expanded ex vivo for clinical use. While lentiviral and retroviral vectors are the methods of gene transfer in the represented studies, other methods such as electroporation and RNA-based methods can be employed.[12,17] The type of cell culture manufacturing system used may impact the final phenotypic composition of the CAR T-cell product, yielding different clinical outcomes.[18-20] As an example, the manufacturing process at our institution uses an anti-CD3/CD28 antibody-coated magnetic bead system that results in a final product capable of memory function and persistence. Prior to anti-CD19 CAR T-cell infusion, patients typically receive chemotherapy; the goal is to induce lymphodepletion and thereby enhance CAR T-cell expansion and persistence in vivo. Lymphodepletion may have the additional benefit of tumor cytoreduction, which can potentially improve CAR T-cell treatment efficacy and minimize toxicity. Notably, however, some patients have responded to CAR T-cell therapy in the absence of prior lymphodepletion. Besides variations in the investigational product, other differences across CAR T-cell treatment protocols include the type and intensity of lymphodepletion, the timing and dose of anti-CD19 CAR T-cell infusions, and of course the target patient population and malignancy.
Clinical Outcomes of Anti-CD19 CAR T-Cell Therapy
The greatest clinical experience with CAR T cells is with targeting CD19, which is a pan–B-cell antigen expressed in the majority of B-cell malignancies, including ALL, CLL, and NHL. This ubiquitous expression of CD19 on malignant B cells, plus the limitation of off-tumor expression on healthy B cells, make CD19 an attractive target associated with limited “on-target/off-tumor” cytotoxicities.
The prognosis for adult patients with R/R B-cell ALL is poor, and almost all patients will die from their disease.[22-29] The only possible curative approach in this setting is the use of a salvage strategy, followed by allogeneic stem cell transplantation (SCT) in patients who achieve complete remission. In a retrospective review of 547 adult patients with first relapse of Philadelphia chromosome–negative ALL, no patient who had not undergone SCT was alive after 1 year, compared with 38% of patients who received SCT after initial salvage therapy. Even poorer outcomes are seen in adult patients who relapse after SCT, most of whom have chemotherapy-resistant disease with low response rates to donor leukocyte infusion. While the majority of children diagnosed with ALL respond well to traditional treatment approaches, those with primary refractory disease have a poor prognosis, with an overall survival rate of about 30% at 10 years. Children with relapsed disease are traditionally more responsive to initial attempts at salvage therapy than their adult counterparts, but relapsed ALL remains a leading cause of cancer deaths in children.[32,33] The dismal prognosis and lack of conventional treatment options for patients with R/R ALL have driven the pursuit of therapies with alternative mechanisms of action. Immunotherapeutic approaches that target ALL antigens (such as CD19, CD20, and CD22) are proving to be remarkably effective but transient in the R/R ALL population.[1,3,4,29,34-36] Outcomes using CAR T cells directed at anti-CD19 highlight the potential of an immunologic approach.
Unprecedented remission rates of 67% to 90% have been observed in adult and pediatric patients with R/R ALL treated with second-generation anti-CD19 CAR T cells (see the Table).[1,3,4,7] In our program, we reported a 90% complete response (CR) rate in 25 pediatric and 5 adult patients treated with CTL019 cells, which are anti-CD19 CAR T cells incorporating the 4-1BB costimulatory domain. Flow cytometry to assess for minimal residual disease (MRD) was negative in 22 patients, positive in 3 patients (with 0.1%, 0.09%, and 0.22% levels of MRD), and not performed in 2 patients who achieved a CR. Of interest, responders included 2 patients who were refractory to blinatumomab, an agent that uses a bispecific antibody to redirect cytotoxic T cells to eliminate B cells with its anti-CD3 and anti-CD19 arms. Eighteen of the 30 patients treated had relapsed after a prior allogeneic SCT; T cells were successfully collected and manufactured from these recipients, and no post-infusion graft-vs-host disease (GVHD) was observed. Of the 27 responding patients, 7 relapsed (6 weeks to 8.5 months after infusion) and 1 developed myelodysplastic syndrome. Three of these relapses occurred in patients with CD19-negative leukemia. Importantly, sustained remissions lasting from 2 months to more than 24 months were noted in the remaining 19 patients, 15 of whom did not receive further treatment (3 were consolidated in remission with SCT). The durable remissions observed in patients not bridged to allogeneic SCT correlated with CAR T-cell persistence and the biological correlate of ongoing CAR T-cell activity, B-cell aplasia.
Similarly high response rates have also been observed in an ALL treatment study by researchers from Memorial Sloan Kettering Cancer Center (MSKCC), using anti-CD19–directed CAR T cells containing a CD28 costimulatory domain (as opposed to the 4-1BB domain used in Penn/Children’s Hospital of Philadelphia [CHOP] protocols). Of 16 adults with R/R ALL who were treated with the regimen, 14 (88%) achieved a CR. Four patients had relapsed after allogeneic SCT, and no patient developed GVHD. Of the 16 patients treated, 7 underwent allogeneic SCT, with 2 deaths related to SCT complications and the remaining 5 patients in ongoing remission (with follow-up ranging from 2 to 24 months). The duration of CAR T-cell persistence in this study was limited to 1 to 3 months after infusion.
In a National Cancer Institute (NCI) study (ClinicalTrials.gov identifier: NCT01593696), a CR rate of 67% was observed in a phase I intent-to-treat analysis of 20 children and young adults with B-cell ALL treated with anti-CD19–directed CAR T cells containing a CD28 domain. Ten of the 14 responding patients subsequently underwent SCT. Two patients who were not consolidated with SCT developed CD19-negative relapse, similar to the Penn/CHOP reports. Three patients who did not respond to initial treatment were reinfused with CAR T cells, but they did not respond to therapy. CAR T cells did not persist beyond 68 days. Similar to the MSKCC experience described previously, no patient treated with CAR T cells after allogeneic SCT (n = 8) developed GVHD.
Recently, investigators from the Fred Hutchinson Cancer Research Center reported similar outcomes in 29 adult patients with ALL in a trial with several unique features. First, CD4-positive and CD8-positive central memory T cells were enriched. CAR T cells were manufactured separately from the CD4-positive and CD8-positive T-cell subsets; this was achieved by using an anti-CD19 CAR incorporating the CD3ζ signaling and 4-1BB costimulatory domains transduced via a retrovirus that also contained a truncated epidermal growth factor receptor (EGFR). Inclusion of this truncated EGFR allowed identification and selection of virally transduced cells. Cells were then reconstituted in a 1:1 ratio of CD4:CD8 cells for infusion. The authors reported a CR rate of 93% (86% of whom were MRD-negative). Ten of 12 patients who received cyclophosphamide or etoposide alone achieved a CR, but 7 of them subsequently relapsed. The next 17 patients received fludarabine and cyclophosphamide. They had higher peak expansion and persistence through day 28. While 94% of these patients achieved a CR, the follow-up duration was short; however, disease-free survival seemed to be improved compared with that of patients treated with cyclophosphamide alone. This suggests that, indeed, the intensity of the lymphodepletion may impact overall outcomes. This interesting approach is the first attempt at using a defined T-cell population for CAR therapy; it is hoped that future studies will identify the ideal ratio of CAR-modified T-cell subtypes, and evaluate whether or not this approach can improve outcomes, compared with trials that do not preselect T-cell populations.
Anti-CD19 CAR T-cell therapy can be used successfully as a bridge to curative allogeneic SCT. In patients who do not undergo consolidation with SCT, long-term disease control may be dependent in large part on the persistence of anti-CD19 CAR T cells, although the optimal length of persistence (months to lifelong) is unknown. An important problem (even in patients with CAR T-cell persistence) is CD19-negative relapse. Anti-CD19 CAR T cells are effective in the post-SCT setting, where it is feasible to collect and manufacture donor-derived T cells from the recipient and safely infuse anti-CD19–directed CAR T cells into the patient without triggering GVHD.[1,3,4] Multicenter studies using anti-CD19–directed CAR T cells in adult and pediatric ALL are underway or in development.
CLL and NHL
Despite a wide array of available treatment options, CLL remains incurable without an allogeneic SCT, and the prognosis is quite poor for patients with refractory or multiply relapsed disease and short progression-free intervals between treatments. Anti-CD19 CAR T cells have been used by several centers to treat patients with R/R CLL.[2,6,38] We recently reported outcomes of long-term follow-up of the first 14 patients with CLL treated at Penn using CTL019 cells. The overall response rate (ORR) in this heavily pretreated cohort was 57%, with 4 CRs and 4 partial responses (PRs). Similar to our experience with CAR T-cell immunotherapy for ALL, durable remissions have been observed that correlate with anti-CD19 CAR T-cell expansion and persistence. Remarkably, 2 of these patients with CLL completed CAR T-cell infusion more than 5 years ago, and both are in molecular remission with detectable levels of functional anti-CD19 CAR T cells. In the Penn program, we have not yet been able to identify baseline disease- or patient-related characteristics that differentiate CLL responders from nonresponders.
In an NCI-led study, researchers used anti-CD19 CAR T cells with a CD28 costimulatory domain to treat 15 patients with advanced CD19-positive B-cell malignancies. All 4 patients with CLL in the study responded to treatment, with 3 having sustained CRs lasting longer than 1 year. However, the initial experiences of MSKCC investigators using anti-CD19 CAR T cells with a CD28 costimulatory domain in patients with chemotherapy-refractory or relapsed B-cell lymphomas were more disappointing, with 0 of 7 evaluable CLL patients responding. The reasons for this lack of response are not clear but may be related to differences in the protocol (a phase I design with 3 of 7 CLL patients not undergoing lymphodepletion) or in patient selection criteria, compared with the NCI and Penn studies.
Data on the use of anti-CD19 CAR T cells as treatment for NHL are less mature. Nine patients with diffuse large B-cell lymphoma (DLBCL) or primary mediastinal B-cell lymphoma received this therapy at the NCI, with 4 achieving a CR, 2 a PR, 1 stable disease, and 2 not evaluable for response. Preliminary results describing outcomes using anti-CD19 CAR T cells in CD19-positive NHL from Penn and the University of Washington, Seattle, were presented at the 2015 American Society of Hematology Annual Meeting.[39,40] Researchers from Penn reported an ORR of 68% following treatment with CTL019 in 22 patients with heavily pretreated NHL (13 with DLBCL, 7 with follicular lymphoma, and 2 with mantle cell lymphoma). At a median follow-up of 11.7 months, the rate of progression-free survival was 62%. The group from Seattle reported their experience using an anti-CD19 CAR containing the 4-1BB costimulatory domain in 28 patients with NHL (18 with DLBCL, 6 with follicular lymphoma, and 4 with mantle cell lymphoma). A differential response was noted in 12 patients treated with cyclophosphamide as lymphodepletion, compared with 12 (of 16 evaluable) patients who received fludarabine plus cyclophosphamide as lymphodepletion. Response rates were improved in patients pretreated with fludarabine (ORR, 67% vs 50%), and correlated with higher peak levels of CAR T cells in vivo and improved CAR T-cell persistence; similar to results observed in patients with ALL, these findings suggest that the process of lymphodepletion may have an important impact on overall treatment outcomes for patients with B-cell malignancies.