A review.Two independent groups have now corroborated and extended our earlier report that the administration of autologous cells transduced with a chimeric antigen receptor (CAR) targeting the CD19 mol. could eradicate CD19+ B-lineage cells in humans and was associated with the regression of advanced follicular lymphoma.These studies have opened opportunities to tailor the genetic modification of autologous lymphocytes with receptors suitable for antigens presented on a patient's unique cancer and thus personalize cancer immunotherapy.The clearest demonstration of the ability of adoptive cell transfer (ACT) therapy to mediate the regression of metastatic cancer comes from clin. trials in patients with refractory metastatic melanoma.The adoptive transfer of autologous tumor-infiltrating lymphocytes following a lymphodepleting regimen mediated durable complete regressions in up to 40% of patients.Melanoma, however, is the only cancer that naturally gives rise to autologous cells with antitumor activity.The success of ACT in melanoma thus stimulated attempts to genetically engineer normal circulating T cells to express antitumor receptors reactive with a wide variety of antigens.16 In neither of these two trials was any antitumor effect seen.The first successful cancer regressions induced by the adoptive transfer of autologous cells genetically engineered to express a conventional T-cell receptor (TCR) were reported by Morgan et al.,10 who showed that cells transduced with a TCR reactive with the MART-1 melanoma/melanocyte antigen could mediate regression of melanoma metastases and, using improved receptors in a later study, observed regressions in up to 30% of patients.In contrast to conventional TCRs that recognize antigens in a major histocompatibility complex (MHC)-restricted fashion, the recognition of cell surface antigens by CARs is not restricted by the MHC because these receptors comprise a single chain of the heavy- and light-chain variable regions of an antibody fused to T-cell intracellular signaling chains such as CD3ζ, CD28, and 4/1BB.Early attempts to use CARs to treat patients targeted an α-folate receptor overexpressed on ovarian cancers or carbonic anhydrase IX overexpressed on kidney cancer.16 In neither of these two trials was any antitumor effect seen.In neither of these two trials was any antitumor effect seen.The first successful CAR gene therapy targeted the glycoprotein GD2 in patients with relapsed neuroblastoma; three of 11 patients experienced complete regressions that were sustained in 2 of them, and evidence of tumor necrosis was seen in addnl. patients (personal communication from the reference 17 group).In applying ACT therapy to a wide variety of cancers, the choice of antigen is criticalThese target antigens fall into three main categories.The first includes antigens uniquely expressed on cancers.Targeting antigens encoded by shared mutations that are essential for maintenance of the malignant phenotype would be ideal, but this has not yet been successfully realized.Cancer/testis (CT) antigens can be ideal targets because they are expressed during fetal development but not expressed on any normal adult tissue (except the testes, which are immunol. protected) and are re-expressed on about 25% of common solid cancers.Because there are more than 100 CT antigens, testing tumors for specific antigen expression represents an attractive approach to personalized immunotherapy.ACT therapy with T cells specific for the CT antigen NY-ESO-1 mediated good antitumor responses in patients with melanoma and synovial cell sarcomas without toxicity to any normal tissues.Ongoing studies are targeting the MAGE-A3 CT antigen, which is highly expressed on a variety of solid cancers such as cancers of the lung, esophagus, and head and neck.A second category of targets includes proteins that are overexpressed on tumors relative to normal tissues.Because this has resulted in severe on-target toxicity to normal tissue in studies targeting carbonic anhydrase IX (biliary tract toxicity),carcinoembryonic antigen (colitis), melanoma/melanocyte antigens (uveitis), and HER-2 (lung toxicity), this class of antigens should be targeted with caution.A third class of targets includes antigens that are expressed on cancer as well as in nonessential normal tissues, so the cancer can be impacted without severe sequelae to the host.An example of this approach is the targeting of the CD19 protein, expressed on normal B cells and on the great majority of B-cell lymphomas and B-cell leukemias.Although critical for the induction of humoral immune responses, transient elimination of CD19-expressing cells is well tolerated.The first effective treatment using a CAR against CD19 was reported by Kochenderfer et al. in a patient with advanced follicular lymphoma treated with chemotherapy followed by the adoptive transfer of T cells transduced with a gammaretrovirus encoding an anti-CD19 CAR and administration of interleukin-2 (IL-2).The patient experienced a dramatic regression of his lymphoma, and, interestingly, his B-lineage cells were completely eradicated from the blood and bone marrow for 36 wk after infusion of CAR-transduced T cells.Consistent with an eradication of B-lineage cells, the patient also developed severe hypogammaglobulinemia.Seven months after his initial CAR-transduced T-cell infusion, the patient's lymphoma started to progress.He was treated a second time with the same regimen, and he remained in a partial remission as of 17 mo after his second cell infusion (unpublished data; Figure 1).We subsequently treated an addnl. seven patients using this regimen (three with B-cell lymphomas and four with chronic lymphocytic leukemia, CLL) (unpublished observations).One patient with lymphoma died of a documented H1N1 pneumonia shortly after treatment.Five of the remaining six patients experienced an objective remission, including one complete remission (Kochenderfer et al., unpublished data).Two recently published pilot trials using anti-CD19 CAR-expressing T cells confirm many of the findings we reported.Kalos et al. reported three patients with CLL who received chemotherapy followed by infusions of T cells transduced with a lentiviral vector encoding an anti-CD19 CAR.Two patients obtained complete remissions of CLL, and the third obtained a partial remission.Eradication of normal B cells and hypogammaglobulinemia were again observedBrentjens et al. recently reported results in eight patients with progressive CLL treated with an anti-CD19 CAR (as well as one patient with acute lymphocytic leukemia treated in remission).Three CLL patients were treated without cyclophosphamide conditioning, and no objective responses were seen.A fourth patient died soon after T-cell infusion and was not evaluable for clin. response.Four patients with CLL received cyclophosphamide preconditioning chemotherapy prior to CAR-transduced cell infusion; one of the four exhibited an objective reduction of peripheral lymphadenopathy lasting 6 mo.Thus, all the reported responding patients received preconditioning chemotherapy with cyclophosphamide.The impact of this chemotherapy must be considered in evaluating the effectiveness of these treatments, although the prolonged depletion of normal B cells seen by us and by Kalos et al. is highly suggestive evidence for an in vivo effect of the CAR-transduced cells.The CAR used in our clin. trial1 and that of Brentjens et al.4 contained a CD28-costimulatory domain plus the cytoplasmic portion of the CD3ζ mol., whereas the CAR used by Kalos et al. used a CAR containing a 4-1BB domain but not CD28 (reference 3).Human T cells transduced with CARs containing a CD28 moiety have recently been shown to exhibit modestly enhanced proliferation and persistence compared with T cells transduced with CARs that lack CD28 (reference 35).Kalos et al. reported more extensive proliferation of the transferred cells in two of three patients and long persistence of CAR-transduced T cells.These results suggest that 4-1BB moieties may provide enhanced proliferation and persistence of CAR-transduced T cells compared with the CD28 costimulatory domain, but a larger cohort of patients is needed before any firm conclusions can be drawn about the role of these intracellular costimulatory signals.The toxicities seen in the three trials were similar and resulted primarily from the sequelae of neutropenia, cytokine release, and tumor lysis syndrome.Fever, hypotension, capillary leak, and severe malaise were common.There was no clear evidence that the different signaling chains used in the CAR constructs impacted toxicity.It is not known whether IL-2 administration after cells contributes to the antitumor effects of the CAR administration, although in several animal models IL-2 significantly improved the effectiveness of ACT therapy.The introduction of genes encoding herpes simplex thymidine kinase (reference 38) or an inducible caspase 9 (reference 39) provide an opportunity to eliminate transduced cells in vivo by providing prodrugs to activate cell death pathways.This approach has been of value in averting graft-vs.-host disease following allogeneic stem cell transplants and may be of value in situations in which the transduced cells may be expected to have long-term deleterious effects.Despite the above successes, significant questions remain concerning the optimization of ACT therapy for cancer treatment-particularly for solid tumors, which may have an intensely immunosuppressive tumor environment.Experiments in animal models have shown that preconditioning lymphodepletion was essential for the effectiveness of cell transfer studies based on several factors, including the depletion of host regulatory T cells and myeloid suppressor cells (reference 37), whereas elimination of endogenous lymphocytes also increases circulating levels of homeostatic cytokines such as IL-7 and IL-15 (reference 6), which are thus able to supply the transferred cells with an endogenous source of growth-promoting cytokines without competing with endogenous lymphocytes.There is no apparent significant difference between the use of gammaretroviruses or lentiviruses, although the intracellular signaling chains used may have an impact on results.The use, individually or in combination, of CD3ζ, CD28, and 4-1BB is being explored and may impact the persistence and clin. effectiveness of the transduced cells.In addition, preclin. studies have shown that transducing genes encoding other mols., such as IL-2 (reference 40), IL-12 (reference 41), and CD80 (reference 42), in conjunction with antitumor receptors may increase the antitumor effects of ACT.Perhaps most importantly, these pilot trials utilizing CD19 CAR-transduced cells to treat patients with B-cell lymphomas and CLL provide further evidence of the potential usefulness of ACT therapy as a new form of immunotherapy for the treatment of cancer.The ability to genetically engineer T cells to express antitumor receptors has opened the door to an era of personalized immunotherapy in which tumors will be tested for the expression of antigens and the receptor for therapy will be selected that is appropriate for each individual patient.
