This review addresses the basics of CAR T-cell design and reviews data from published clinical studies in leukemia.
Pallawi Torka, MD
Department of Medicine,
Roswell Park Cancer Institute
Buffalo, New York
Elizabeth A. Griffiths, MD
Department of Medicine
Roswell Park Cancer Institute
Buffalo, New York
Chimeric antigen receptor T cells (CAR T cells), engineered from individual patients to specifically target tumor-associated antigens, have shown extraordinary promise in the management of B-cell malignancies. Moreover, CAR T cells are being developed for a myriad of other hematologic malignancies. First-generation CAR T cells demonstrated only modest efficacy because of inadequate activation, persistence, and proliferation, but the addition of one (second-generation) or more (third-generation) costimulatory molecules has produced a marked improvement in efficacy. Clinical trials are recruiting patients with leukemia, and CARs are being tested for patients with a large variety of both solid and liquid tumors. It seems likely that CAR T-cell therapy will become an established option in the near future for patients with B-cell malignancies. Significant questions remain about optimal timing for administration, product preparation, and CAR design to optimize response across malignancy types. This review addresses the basics of CAR T-cell design and reviews data from published clinical studies in leukemia. It also contains commentary on the prospects of this promising new therapeutic strategy.
The most familiar immunotherapeutic approach for treatment of leukemia is allogeneic hematopoietic stem cell transplant (allo-HSCT). Manipulation of graft-vs-leukemia (GVL) effects can be further accomplished through the use of donor lymphocyte infusions (DLIs) and donor natural killer (NK) cell infusions. These conventional approaches are unfortunately characterized by unpredictable donor T-cell-mediated graft-versus-host disease (GVHD), and in some cases, by a failure to achieve antileukemic activity. More specific immune-based approaches have long been the topic of significant research, but until recently had been unsuccessful.
Hematopoietic malignancies express a variety of antigens that could be potential targets for immunotherapy, but overlap between leukemic and normal cell antigen expression has been a significant barrier. Leukemias have been shown to evade host immunosurveillance through a variety of different mechanisms, including downregulation of human leukocyte antigens and expression of programmed cell death ligand-1 (PD-L1 [among others]). Furthermore, leukemia-associated self-antigens are not targeted endogenously because most T cells directed against these self-antigens are deleted during thymic education, and the residual population is held in check by regulatory T cells. Ex vivo gene modification of T cells through the introduction of recombinant receptors directed against these self-antigens can overcome host immune tolerance. Chimeric antigen receptor T-cell (CAR T) therapy based on this concept is a form of “living” immunotherapy that has shown unprecedented responses in leukemia, specifically B-cell acute lymphoblastic leukemia (BALL). Currently, CAR T therapy is under simultaneous development at many centers, and although the most well-developed CAR T therapies are directed at CD-19, a pan B-cell marker present on both normal and malignant B cells, numerous other targets are also promising. Despite the intense enthusiasm for CAR T therapy, many questions and uncertainties remain about its optimal design, long-term safety, and optimal timing.
Adapted from Brentjens et al, 2012, and Kochenderfer et al, 2013.
In its simplest form, a CAR consists of an antigen recognition domain derived from an antibody that is fused to a T-cell receptor (TCR) signaling domain. Key steps in the manufacturing process include the selection and activation of isolated T cells, transduction of T cells to express CARs, ex vivo expansion of the modified T cells, and cryopreservation in infusible media.1Once the patient has received lymphodepleting chemotherapy, the CAR T cells are reinfused. In this way, CARs graft the specificity of a monoclonal antibody onto the dynamic and persisting characteristics of an effector T cell.
The advantage of an engineered CAR T cell is that, unlike native TCRs, CAR T cells can recognize a variety of cell surface antigens without the constraint of major histocompatibility complex (MHC) human leucocyte antigen (HLA) class restriction. This HLA independence overcomes the tumor’s ability to escape immunodetection by downregulating HLA molecules on its surface. CARs can be engineered against any tumor antigen that is expressed on the cell surface for which a monoclonal antibody can be generated, including proteins, carbohydrates, and glycolipids, thus opening a world of opportunities for cancer treatment. Although the generation of a large quantity of CAR T cells is labor intensive and requires significant infrastructure and expertise, the turnaround time is just a few days, making this strategy feasible for use in the clinical setting. Because the patient’s own cells are reprogrammed to express tumor-specific antigens, the risk of GVHD is limited, a distinct advantage over all HSCT-based immunotherapy. Furthermore, because of long-term persistence of the CAR T cells, a single infusion may be sufficient to produce a sustained response, thus offering the potential for “living” immunotherapy.
Gross et al are credited with introducing a T-cell hybridoma in 1989.2A CAR consists of a polypeptide that is the single-chain variable fragment (scFv) from an antibody, linked to the signaling machinery of a T-cell receptor, typically the ζ chain.3The scFv binds to a defined target antigen and triggers effector cell activation upon target engagement.
The first generation of CARs consisted of:
In clinical trials involving patients with various cancers, including lymphoma and ovarian cancer, the first generation of CAR-T cells triggered T-cell killing but showed only modest efficacy because of inadequate activation, persistence, and proliferation of the modified T cells, all of which were attributed to a lack of costimulatory molecules.6-9CD28 was the first costimulatory molecule used in the CAR construct and dramatically enhanced antigen-specific cytokine production (eg, interleukin 2 [IL-2]), cytotoxicity, and proliferation.10-12These second-generation CARs were also relatively resistant to suppression by regulatory T cells.13Since then, many other costimulatory molecules have been introduced, including CD27,14CD134 (OX40, TNFRSF4), CD137 (4-1BB, TNFRSF9),15,16CD244,17and inducible T-cell costimulator (ICOS).18
CAR constructs with a single costimulatory molecule are known as second generation, and those with two or more costimulatory molecules are referred to as third generation. A variety of studies have proved the superiority of second- over first-generation CARs in terms of persistence and cytotoxic activity,8but the superiority of third-generation CARs over second-generation CARs remains unproven. Preclinical data comparing CD19 CARs containing either CD28 or 4-1BB costimulatory domains suggested that while cytotoxicity is similar with both modifications, the CAR containing 4-1BB may be associated with improved proliferation,15although this observation may be restricted to CD19 CARs. An ongoing clinical trial is designed to identify the best costimulatory domain for CD19 CARs (NCT00466531).
Currently, various CAR constructs that differ in scFv, hinge, transmembrane, stimulatory, and costimulatory domains are being studied in clinical trials at different centers. The optimal combination is yet to be identified and may differ according to the clinical indication.Peripheral blood mononuclear cells (PBMCs) are collected from the patient by apheresis, typically before the administration of lymphodepleting chemotherapy.1An absolute T-lymphocyte count below 200 to 300/μL may result in poor T-cell collection and, therefore, chemotherapy immediately before harvest is not ideal.1If the patient is stable post allo-HSCT, then donor PBMCs can be engineered. To date, no significant GVHD has been observed after infusion of allogeneic donor-type CAR T cells in patients with B-cell ALL who relapsed after allo-HSCT. Leukapheresis is typically performed without granulocyte-colony stimulating factor (G-CSF) support, since some reports suggest that G-CSF mobilization may skew T cells toward a Th2 phenotype and impair signaling through CD28.19The potency of T cells may vary, depending on the mobilization regimen, patient treatment history, method of T-cell isolation, and stimulation. The apheresed PBMC product is then transduced and expanded before reinfusion into the patient.
CAR indicates chimeric antigen receptor; PBMC, peripheral blood mononuclear cell.
Stable gene transfer is a prerequisite for sustained CAR-T cell expansion and is required for the persis- tence of effector cells in patients following reinfusion. The genetic material encoding the CAR can be introduced into T cells by several methods, each of which has its own pros and cons in terms of complexity, stability of transgene expression, safety, regulations, and cost. The commonly used techniques are:
Viral vectors can integrate into the host genome, leading to permanent transgene expression. These vectors are highly efficient in transducing T cells after in vitro activation with various cytokines. The advantage of gamma retroviral and lentiviral vectors compared with adenoviral and adeno-associated viral vectors is low immunogenicity. In addition, lentiviruses can integrate in nondividing quiescent cells, are less susceptible to silencing by host restriction factors, and can deliver larger DNA sequences.4The disadvantages of viral vectors are the high cost, need for highly specialized staff and facilities required for production, strict regulatory requirements, and risk of mutagenesis.20
Nonviral approaches are cheaper and potentially safer; however, the CAR T-cell constructs are usually not durable. Transposon/transposase systems such as Sleeping Beauty introduce plasmid DNA by elec- troporation along with DNA or RNA that codes for a transposase.21They can deliver a large payload and have been used by MD Anderson Cancer Center in Houston, Texas. Specific terminal repeats flanking the CAR expression cassette are recognized by the transposase and the cassette is inserted into the T-cell genome. RNA electroporation is the process of introducing synthetic mRNA into T cells without integration into the host genome.22It leads to generation of short-lived CAR-T cells, with limited antitumor activity, that last 3 to 5 days. Their short lifespan translates into limited toxicity, and as a result they have been used in early experiments to study the safety and efficacy of new CAR T-cell constructs.T-cell expansion ex vivo requires activating signals, appropriate culture media, and expertise. The optimal duration of ex vivo expansion is unknown, but longer expansion periods may result in more terminally differentiated T cells that may have impaired antitumor efficacy.23Most reported clinical studies have involved transfer of bulk-transduced T cells with wide variation in the CD4:CD8 ratio, proportion of regulatory T cells (Tregs), and memory T-cell subsets. Graft engineering consists of selection and expansion of a particular cohort of T cells from the apheresed product and is being developed at some centers (eg, Fred Hutchinson Cancer Research Center in Seattle). Preclinical studies suggest that naiÌˆve T cells, central memory T cells, and Th17 or memory stem cellsrather than terminally differentiated effector cells—possess the optimal characteristics for adoptive cellular therapy because they have a high proliferative potential, persist long term, and have the ability to repopulate both memory compartments.24,25Immunomagnetic separation of distinct T-cell subsets before CAR engineering may make the infused product more homogeneous but adds significantly to the complexity of their generation.
Currently, cell culture is done on an individual basis at a few large academic centers. The challenge for the future is streamlining this resource-intensive process into a more automated, assembly linelike culture system. Before infusion into the patient, the CAR T-cell product must be tested to ensure safety, sterility, purity, potency, and identity.1
allo-HSCT indicates allogeneic hematopoietic stem cell transplant; CR, complete response; MRD, minimal residual disease; NR, not reported.
Preclinical and clinical studies have shown that T- cell homeostatic proliferation and persistence are increased by lymphodepletion, probably by eliminating Tregs26and by reducing competition for avail- able cytokines (IL-2 and IL-7).27Functional immune reconstitution is superior when T-cell infusion occurs on day 2 after high-dose conditioning compared with later timepoints. The optimal conditioning regimen is not defined, but regimens that include cyclophosphamide and fludarabine have been used in almost all current CAR studies.28-31
The level of engraftment and expansion does not correlate strongly with the infused T-cell dose; hence, the optimal dosing strategy for T-cell infusion is unknown. After infusion, an initial peak in CAR T-cell number is noted in the peripheral blood, followed by a reduction within the first few days from redistribution into tissues, and subsequent resurgence in the peripheral blood from proliferation.6Strategies to augment T-cell persistence after transfer include exogenous cytokine administration, engineered autocrine production of cytokines (IL-12), and overexpression of prosurvival signals or reversal of antisurvival signals. In the first human studies, IL-2 was administered pre-infusion, but it was eliminated later because of lack of evidence of benefit.32
CD19 is a transmembrane glycoprotein expressed on a multitude of B-cell malignancies, including >95% of all B-cell ALLs, B-cell non-Hodgkin lymphomas, and chronic lymphocytic leukemia (CLL). It is also expressed on normal B cells from the early pro-B to mature B-cell stages but not on any other cell types. The restricted expression of CD19 to B-lineage cells means that anti-CD19 CAR T cells have on-target/off- tumor effects resulting from B-cell depletion. Loss of B-cells produces hypogammaglobulinemia, which can be easily managed with periodic intravenous immunoglobulin (IVIG) infusions, but is associated with limited other toxicity. As a result, most of the clinical trials to date have been done with CD19 CAR T cells as described in the subsequent paragraphs. A sum- mary of these studies is presented in TABLE 1. Other target antigens under investigation for hema- tologic malignancies include CD20,33ROR1,34immunoglobulin kappa light chain,35CD33,36,37IL-3 receptor α chain/CD123, and Lewis Y antigen.38
The first studies in ALL were published in the spring of 2013 by investigators from the University of Pennsylvania29in Philadelphia and the Memorial Sloan Kettering Cancer Center (MSKCC)39in New York City. Updated reports with expanded patient cohorts have been published since then. Maude et al reported on 30 patients with relapsed/refractory CD19+ ALL who received lentivirally transduced second-generation CD19 CARs using the FMC63 scFv and 4-1BB costimulatory domain.31This study enrolled 25 children and five adults (26-60 years of age); three patients had primary refractory ALL; the remainder had relapsed disease. One patient had CD19-expressing T-cell ALL; 18/30 patients had relapsed following allo-HSCT; three were refractory to blinatumumab, the CD3/CD19 bispecific antibody; and two had asymptomatic central nervous system (CNS) involvement. At first assessment after 1 month, 27/30 patients had morphologic complete remission (CR) and 22/30 patients were negative for minimal residual disease (MRD) by multiparametric flow cytometry, implying a CR rate of >90%. After a median follow-up of 7 months, 19 patients remained in CR, and 15 did not require any further therapy. Event-free survival (EFS) was 67% (95% CI, 51-88) and overall survival (OS) was 78% (95% CI, 65-95) at 6 months. No differences in OS and EFS were seen between patients who had received allo-HSCT and those who had not. Two patients with CNS disease showed clearance of CNS blasts. There was no discernable effect of CD19 density or cell dose on efficacy or toxicity; however, durable responses appeared to correlate with higher peak levels of circulating CAR T-cells and duration of B-cell aplasia. Of the seven patients who relapsed after CAR T-cell therapy, three occurred early from loss of CAR T cells, one had refractory disease, and three relapsed because of loss of CD19 expression on the leukemia cells in the face of persistent CAR T-cell circulation. CAR T cells were detectable up to 11 months, and the probability of persistence at 6 months was 68% (50%-92%).
Cytokine release syndrome (CRS) occurred in all patients, with 8/30 (27%) requiring care in the intensive care unit (ICU). Severe CRS (sCRS) was noted to present early (after 1 day compared with 4 days in patients with mild-moderate CRS and was associated with higher baseline disease burden and higher levels of IL-6, C-reactive protein (CRP), ferritin, interferon-γ, and soluble IL-2 receptor. Because of high IL-6 levels, tocilizumab was incorporated into the management of early, sCRS and resulted in significant abrogation of symptoms. Neurologic toxicity was noted in 13 patients and was self-limiting.
Davila et al treated 16 adult patients with re-lapsed/refractory B-ALL with autologous T-cells expressing a CD19-specific CD28/CD3ζ second- generation CAR.28Fourteen patients had detectable disease at the time of CAR T-cell infusion, despite salvage chemotherapy and cyclophosphamide conditioning. The overall CR rate was 88%, with 86% of patients from the CR group further classified (12/14) as MRD negative. Seven of 11 eligible patients were able to undergo allo-HSCT; no relapses were noted post-transplant. Of the remaining patients, three had suboptimal response to CAR T-cell therapy, two declined transplant despite achieving molecular CR, and three had medical comorbidities that ruled out a transplant. This precluded assessment of durability of responses in these patients. In four of eight non-transplanted patients, morphologic remission was noted at up to 24 months. The persistence of 19-28z CAR T cells was noted to be about 3 months, in contrast to the University of Pennsylvania study, where patients exhibited CAR T-cell persistence for several months. One possible explanation for this difference is that the 19-28z CAR T-cell expansion and subsequent contraction are CD19 antigen dependent, resulting in T-cell clearance on elimination of normal and malignant B cells, as seen in a normal T-cell immune response to antigen.39The persistence of CD19-targeted CARs incorporating a 4-1BB moiety rather than the CD28 costimulatory molecule may therefore result from antigen-independent signaling through the 4-1BB CAR, as previously demonstrated in preclinical studies.15
Toxicities were similar across the two studies. Stringent criteria were applied for diagnosis of sCRS, which included a triad of persistent fever for ≥3 days, selected cytokine elevation (75-fold in two cytokines or 250-fold in one cytokine), and evidence of clinical toxicity (hypotension requiring at least one pressor, or hypoxia, or neurologic manifesta- tions). Patients with sCRS required a prolonged hospital stay (56.7 days vs 15.1 days).
The National Institutes of Health (NIH) recently published results of a phase I intention-to-treat dose-escalation study assessing feasibility, toxicity, maximum tolerated dose (MTD), response rate, and biological correlates of response in children and young adults with relapsed or refractory ALL treated with 19-28z CAR T cells.30In all, 21 patients were enrolled and all received fludarabine and cyclophosphamide preconditioning. The MTD was 1 × 106 CD19-CAR T-cells per kg; all but two patients received the anticipated T-cell dose. Efficacy was similar to that reported in earlier studies, with 67% of patients achieving a CR and 60% achieving MRD negativity. Ten patients who were MRD negative and underwent allo-HSCT remained disease-free at last follow-up. Two patients who were MRD negative and ineligible for transplant relapsed with CD19-negative ALL at 3 and 5 months. With a median follow-up of 10 months, OS was 51.6%. Eleven of 18 evaluable patients demonstrated the presence of CAR T cells in the cerebrospinal fluid (CSF), and CNS disease was cleared in two patients with asymptomatic CNS involvement. Grade 4 CRS was noted in 3/21 (14%) patients, and all toxicities were fully reversible. Neurotoxicity was noted in 6/21 patients; the concentration of CAR T-cells in the CSF was higher in patients who developed neurotoxicity than in those who did not. In evaluable patients, CAR T-cell persistence was lost after 68 days.
Overall, CAR T-cells have shown unprecedented response rates of up to 90% in patients with relapsed/ refractory ALL (TABLE 1). Response rates appear to be higher in ALL than in CLL, likely from defects in T-cell function in CLL40and from preferential homing of CD19 CAR T cells to the bone marrow over the lymph nodes. Efficacy appears to depend on the degree of expansion of CAR T-cells rather than on the dose infused. CAR T-cell therapy demonstrates considerable promise as a bridge to transplantation, even in chemorefractory disease, but the potential to replace transplant with engineered T-cells remains a tantalizing possibility. Important prerequisites for response appear to include host lymphodepletion, incorporation of a costimulatory domain, short-term culture, and persistence of the infused product. Longer follow-up is required to assess the optimal duration of persistence for response durability.The initial report of activity of CARs in CLL came from the University of Pennsylvania, where three patients with refractory CLL were treated with CD19 CARs with a 4-1BB costimulatory domain.41Among the treated patients were two prolonged CRs and one partial response (PR). Investigators at MSKCC treated eight patients with relapsed, heavily pretreated, purine-analogue refractory CLL with CD28 costimulatory domain-based CD19 CARs.42They found that T cells from patients with CLL expanded poorly when compared with T cells derived from patients with ALL, but overall gene transfer was more effective. The median age of the eight treated patients was 68 years; all patients had bulky adenopathy. Of four evaluable patients from this cohort, one patient had a PR and two had stable disease (SD) without concomitant development of B-cell aplasia. An inverse correlation was noted between preexisting peripheral blood tumor burden and persistence of CAR T cells, suggesting that T cells are more rapidly cleared from the circulation in the presence of greater target cell numbers. The investigators concluded that clinical benefit was more likely to occur in the context of MRD. A third study was reported by investigators from the National Cancer Institute (NCI), who reported results from four patients with advanced, heavily-treated CLL responding to CD19 CAR T-cell therapy. Three of four patients attained CR, which was durable for up to 23 months.43Investigators at Baylor College of Medicine in Texas had one patient with CLL in their treatment cohort. This patient demonstrated SD for 10 months after T-cell infusion without prior lymphodepleting chemotherapy.8All four trial investigators used second-generation CARs with a CD28 costimulatory domain and gammaretroviral vector transduction, with the exception of the University of Pennsylvania investigators, who used 4-1BB as their costimulatory molecule and a lentiviral transduction approach. The best response in these studies, a CR with long-term B-cell aplasia, was observed in a patient with low estimated disease burden, suggesting that maximal cytoreduction before CAR T-cell infusion may be necessary for success in patients with CLL.Unlike lymphoid malignancies, where abrogation of CD19 results in B-cell aplasia but no other on-target/ off-tumor effects, the challenge in AML is identification of an appropriate target that is not present on normal hematopoietic cells. The lack of AML-specific surface markers has been a significant barrier to the development of CAR-based approaches. The IL-3 receptor α chain/CD123 is expressed on most AML cells, and human CD123-redirected T cells have resulted in AML remissions in mice.44However, because CD123 is also present on marrow precursors, anti-CD123 CAR T cells have the potential to eradicate myelopoiesis, which would require allo-HSCT as a rescue strategy. Similarly, CD33 is expressed on immature myeloid cells, including AML; anti-CD33 CAR T cells have been shown to have equal anti-AML efficacy but more severe myeloablation compared with anti-CD123 CAR T cells.45Despite the potentially chilling side effect profile of such strategies, clinical trials targeting CD123 and CD33 have commenced at the City of Hope (NCT02159495) and in China (NCT01864902).
Recently, researchers at the University of Pennsylvania have published preclinical data on 4-1BB- costimulated anti-CD33 CAR T cells in a humanized mouse model bearing primary AML xeongrafts.46These authors confirmed the antileukemic activity of CD33-targeted CARs at the expense of marked toxicity toward the myeloid compartment. They further developed a “biodegradable” RNA-based anti-CD33 CAR T-cell construct that persisted for a limited period.46This construct demonstrated less robust antileukemic activity, but limited myeloid toxicity, potentially opening the door to a CAR T-cell-based bridging approach to enable allogeneic bone marrow transplantation in patients with chemotherapy-refractory AML.
CD44v6 is another potential target present on AML cells. This molecule is an adhesion protein expressed on some AML blasts and myeloma cells, but its presence on keratinocytes has the potential for lethal epithelial toxicity, a phenomenon seen previously in one patient treated with the anti-CD44v6 monoclonal antibody, bivatuzumab.47Lewis Y antigen has also been targeted by a CD28-costimulated retrovirally tranduced CAR T-cell construct at the University of Melbourne. So far, four patients have been treated with minimal toxicity.38,48Another study evaluating the safety of CAR T cells directed against NKG2D-ligands is recruiting patients with AML and myelodysplastic syndrome (MDS) with refractory anemia with excess blasts (RAEB) (NCT02203825).CRS is the most common adverse effect (AE) of CAR T-cell therapy and can be potentially life-threatening. The symptoms range from mild ‘flu-like’ illness to multiorgan/respiratory failure requiring intubation and blood pressure support. Some degree of CRS is seen in most patients receiving anti-CD19 CAR T cells, with sCRS reported in about 30% of patients. CRS symptoms appeared earlier with CD28-containing CD19-CAR T cells, compared with 4-1BB-containing CD19-CAR T cells, likely because of a more rapid T-cell expansion in the former group. Severe CRS is associated with elevations in serum IL-6, IL-10, and interferon-γ concentrations. In addition, retrospective analysis of serum samples in the MSKCC study28has shown a significant difference between the C-reactive protein (CRP) levels of patients with sCRS compared with patients demonstrating mild or no CRS; CRP levels ≥20 mg/dL were associated with sCRS. CRP might therefore have use as a predictive biomarker of sCRS. Tocilizumab, the anti-IL-6 antibody, is the preferred first-line therapy for sCRS, with the use of corticosteroids restricted only to the most severe cases because administration of high-dose steroids has the potential to ablate the CAR T cells.28,29,39Suppression of CAR T-cell expansion would be presumed to have a negative impact on leukemic control, especially if required in the early period after CAR T-cell infusion.
Reversible neurologic complications, including delirium and seizure-like activity, have been noted in all CD19 CAR T-cell studies in ALL; similar symp- toms have been noted in patients receiving blina- tumomab, the bispecific CD19-directed CD3 T-cell engager. Patients may develop a gradual progression of confusion, word-finding difficulty, and aphasia, and ultimately become obtunded. Imaging, electroencephalograms (EEGs), and lumbar punctures are typically unrevealing, but a higher concentration of CAR T cells has been noted in the CSF in these patients.30The possible mechanisms responsible for these symptoms include either passage of systemic cytokines across the blood-brain barrier or cross reactivity of CAR T cells with targets in the CNS. The symptoms are typically transient and management is supportive.
This is an on-target/off-tumor effect that serves as a surrogate for persistence and activity of the CAR T cells. The duration of aplasia depends on the CAR construct and subjects patients to increased risk of infection. Patients are typically given prophylactic monthly IVIG infusions, in a manner analogous to that used for the management of patients with X- linked agammaglobulinemia.NK cells are of great clinical interest because they contribute to the graft-versus-leukemia effect, but unlike T cells, play no role in GVHD. Feasibility and safety of NK-cell therapies have been shown for leukemia in both the pretransplant and posttransplant settings.49In a fashion analogous to that previously described for CAR T cells, CARs have been introduced into NK cells by either viral-based transduction or by electroporation with varying degrees of success(<10%-93%). In preclinical studies, CAR-NK cells directed against CD19 and CD20 have shown promise against B-cell malignancies.50-53Two clinical studies using CAR-NK cells are ongoing and both are assessing the safety of anti-CD19-BB-ζ CAR-NK cells in patients with B-lineage ALL in the relapsed/ refractory setting (NCT00995137, NCT01974479).CAR T-cell therapy is a valuable addition to the battle against leukemia, but it is still in the early stages of development and much work needs to be done to maximize its efficacy and limit its toxicity. To enhance safety, the next advance in construction of CARs will be the introduction of a suicide gene so that CAR T cells can be depleted in cases of severe, unacceptable toxicity. Various approaches are being studied, including cotransduction of the herpes simplex virus-derived thymidine kinase gene (killed by ganciclovir),54expression of an inducible caspase or Fas that can induce apoptosis when exposed to a synthetic dimerization agent,55and incorporation of extracellular portions of the EGF receptor, CD34 or CD20,56,57which are susceptible to existing therapeutic agents (eg, rituximab). The introduction of two separate CARs into the same T cell could also be used to prevent off-tumor/on-target side effects and enhance specificity. In such a construct, one CAR would provide a low-affinity activating CD3ζ receptor specific for a first tumor-associated antigen, while another CAR would contain a receptor harboring a costimulatory domain fused to an antigen-binding domain of a second antigen. In this scenario, only interactions that trigger both CAR components would be sufficient to activate T-cell cytolysis.
Multispecific CAR T cells could combat relapse from antigen loss escape by targeting multiple disease-specific B-cell antigens essential for survival/proliferation. This can be achieved by several methods: designing multiple recognition sites on the same CAR, expressing a number of CARs on the same cell using multicistronic vectors, or transducing independent T-cell populations with CARs of different specificities.58Developing chimeric receptors that provide activating signals on engagement of inhibitory ligands in the tumor microenvironment, such as PD-L1, may help overcome the immunosuppressive molecules secreted by the cancer cells.59Furthermore, CAR T cells that express proliferative T-cell costimulatory ligands (4-1BBL) or proinflammatory cytokines (IL-12), called armored fourth-generation CAR T cells, are already under construction.60,61
Currently, CAR T-cell therapy is personalized, with the raw T cells being derived from the patient. Clearly, there are technical and financial challenges in manufacturing single-patient product lots. To make CAR T-cell therapy more commercially feasible, universal off-the-shelf products need to be developed.1Preclinical work is ongoing in this regard.CAR T-cell therapy holds great promise to revolutionize treatment for hematologic malignancy and has already shown impressive results in ALL, although in small clinical studies. However, significant variables in the CAR constructs, generation, manufacturing process, and clinical trial design make direct comparisons between these studies difficult. At least 30 clinical trials targeting CD19 are recruiting patients in North America, Europe, and Asia.4Ideally, CAR therapy should be experimentally optimized for each target antigen and application, and this can only materialize through collaborative efforts of the key players. A clinical trial is planned at the University of Pennsylvania Children’s Hospital and at MSKCC funded by an NIH Special Translational Research Acceleration Project award to address the issue of optimal costimulatory domain choice. In this study, patients treated with the MSKCC 19-28z CAR will be compared head to head with those treated with the University of Pennsylvania’s 19-BBz CAR for cytotoxicity, expansion, and persistence. It is a forgone conclusion that engineered T cells will transform the way we treat B-cell ALL, and it is likely they will significantly affect the management of other B-cell disorders. What remains to be seen is whether or not the potential of engineered immunity will fulfill its promise in the management of myeloid malignancy, and even more provokingly, provide a significant shift for the management of all cancer types.The authors would like to thank Michael J. Nemeth, PhD, for critical review of the manuscript prior to submission. PT and EAG wrote the review and reviewed the literature. PT has no disclosures. EAG would like to disclose honoraria from Celgene, Inc and Alexion Pharmaceuticals and grant funding from Astex Pharmaceuticals.