There is evidence from murine models that the host immune system has a dynamic relationship with a developing tumor and can recognize, control, and even eliminate cancer.75,76,77 Immunogenic proteins in human tumors have now been identified by screening of tumor complementary DNA (cDNA) libraries with tumor-specific T cells isolated from the blood or tumor environment,78 or by screening of patient sera for antibody responses to tumor-associated proteins.79 Distinct categories of tumor antigens have been uncovered, and several are being investigated as targets for T-cell therapy or vaccination (Table 26–1). However, the clinical translation of adoptive T-cell therapy and other immunotherapeutic modalities for human cancers has proven to be more challenging than for opportunistic viral infections. This reflects many issues, including the difficulty isolating highly avid tumor-specific T cells from cancer patients, and evasion mechanisms that tumors employ to avoid immune elimination including the local recruitment of regulatory T cells (TREG) or myeloid-derived suppressor cells, loss of antigen or HLA expression, and expression or secretion of inhibitory molecules or cytokines.80,81,82,83 Additionally, a problem distinct from the results of T-cell therapy for viruses is that transferred tumor-reactive T cells persisted only transiently in most early clinical trials, even if high-dose interleukin (IL)-2 was given to support their survival.84,85,86,87,88
Table 26–1.Categories of Tumor Antigens ||Download (.pdf) Table 26–1. Categories of Tumor Antigens
A. Classes of antigens for MHC-restricted T cells
Antigens arising from mutations or gene rearrangements (e.g., CDK-4, BCR/ABL)
Tissue-specific differentiation antigens (e.g., Tyrosinase, gp100)
Cancer-testes antigens (e.g., MAGE-1, NY-ESO-1)
Nonmutated overexpressed self-proteins (e.g., Her2/neu, WT-1)
Oncofetal antigens (e.g., CEA)
Viral proteins in virus associated malignancies (e.g., HPV E6 and E7, EBV LMP-1)
B. Classes of tumor cell surface molecules for chimeric antigen receptor-modified T cells
B-cell differentiation molecules (e.g., CD19, CD20, CD22)
Myeloid differentiation molecules (e.g., CD123)
Adhesion molecules (e.g., CD44v6, GD2, L1 CAM, mesothelin)
Oncofetal antigens (e.g., ROR1)
Signaling molecules (e.g., Her-2)
Hypoxia induced (e.g., CAIX)
The development of immune cell therapy for malignancy has focused on melanoma because target antigens have been identified and this tumor has responded to nonspecific immune therapy with IL-2,89 and on amplifying the graft-versus-leukemia (GVL) effect after allogeneic HSCT because of the evidence that donor T cells mediate tumor eradication in this setting.90,91,92 The ability to engineer T cells to have tumor specificity by introducing TCR genes that recognize tumor-associated antigens,93 or chimeric antigen receptor (CAR) genes that encodes a single chain mAb domain linked to the CD3ζ chain of the TCR, and confers recognition of a tumor-associated, cell-surface molecule, is facilitating the broader application of T-cell therapy for both human hematologic malignancies and common epithelial cancers.94,95,96,97
CELLULAR THERAPY OF MELANOMA
Early studies demonstrated that the adoptive transfer of autologous polyclonal tumor-infiltrating lymphocytes (TILs), isolated and expanded from resected melanoma specimens, combined with the administration of high-dose IL-2 resulted in a 31 percent response rate in patients with advanced melanoma.98 Most of the responses were transient, but these results validated the potential to eradicate a human solid tumor with immunotherapy. These results also encouraged efforts to define the antigens recognized by TILs in responding patients, and to refine the approaches to augmenting T-cell responses to tumor antigens.
Target Antigens for Melanoma-Specific T Cells
Melanoma has served as a model for the discovery of human tumor antigens because T cells specific for melanoma cells can often be detected in the blood or the tumor microenvironment. A landmark in cancer immunotherapy was the identification by cDNA expression cloning of MAGE-1, which is a member of the cancer-testes antigen class of tumor associated antigens.78 Several additional shared tumor/self-melanocyte differentiation antigens recognized by CD8+ and/or CD4+ T cells have been discovered, including differentiation proteins that function in normal melanocyte physiology such as tyrosinase, gp100, and MART-1; and other cancer testis antigens such as NY-ESO-1.78,99,100 Melanosome antigens are also expressed in normal tissues (skin, retina), and toxicity because of autoimmunity is a concern. Mutated proteins that arise as a consequence of the genetic instability of tumors are being identified as critical targets of immune recognition, and these offer the greatest promise for selectively targeting tumor cells without recognition of normal cells.101
Techniques for Isolation and Adoptive Transfer of Melanoma-Specific T Cells
The adoptive transfer of tumor-specific T-cell clones or oligoclonal populations of T cells expanded ex vivo can, in principle, allow control over the magnitude and function of the tumor-reactive T-cell response in the patient. If the tumor is easily accessible, tumor-reactive T cells can sometimes be isolated directly from the tumor biopsies by culture in high-dose IL-2.102 Melanoma-reactive T cells can also be isolated using autologous dendritic cells pulsed with synthetic peptide antigen, but this approach can enrich low-avidity T cells that have a limited capacity to persist and function in vivo.103
Initial clinical trials of T-cell therapy for melanoma employed CD8+ T-cell clones specific for MART-1 or gp100; or clonal or polyclonal melanoma-reactive T cells derived and expanded from TILs.84,104,105 The transferred T-cell clones given with low-dose IL-2 mediated transient antitumor activity in some patients with advanced disease, but did not persist for long term.104 The response rate was higher in patients treated with polyclonal TILs and high-dose IL-2.105 T-cell persistence was highly variable and only rarely sustained despite the infusion of large T-cell numbers (up to 1011).84,104 The inability of T cells to persist in vivo could reflect an inadequate antigen-specific CD4+ Th response, terminal differentiation of T cells during expansion, activation-induced T-cell death at the tumor site, or cell death as a consequence of IL-2 withdrawal.84,106
A major advance in the field was the demonstration that the transferred human T-cell persistence and therapeutic efficacy could be improved by pretreatment of the patient with a lymphodepleting regimen containing cyclophosphamide and fludarabine.107,108 In these studies, high-dose IL-2 was administered daily after TIL transfer until toxicity required it be discontinued. A subset of the patients achieved prolonged high-level engraftment of one or a few tumor-reactive CD8+ T-cell clonotypes present in the infused polyclonal T-cell product.107,108,109,110 In a followup analysis of a large cohort of patients, the overall and complete response rates were approximately 50 percent and 20 percent, respectively.108,111 The antitumor activity correlated with the persistence of high levels of transferred tumor-reactive CD8+ T cells. Several mechanisms make the lymphopenic environment favorable for T-cell transfer, including less competition for homeostatic cytokines such as IL-15 and IL-7 that promote lymphocyte survival,112,113 and the elimination of CD4+CD25+ TREG cells.114 Studies in murine models have confirmed that severe lymphodepletion can be exploited to improve the antitumor efficacy of the transferred T cells.115 The addition of 2 Gy or 12 Gy of total-body irradiation to the lymphodepleting treatment with cyclophosphamide and fludarabine before TIL transfer increased the response rate to 52 percent and 72 percent, respectively.116 The results of TIL therapy in a metastatic tumor that is unresponsive to conventional therapy were achieved with moderate toxicity, demonstrating the encouraging potential of this therapy.
Studies of the mechanisms of tumor eradication in melanoma are providing insights for treatment of other cancers with T-cell therapy. Exome sequencing of melanoma has shown that the frequency of mutational events is high,101,117 and detailed analysis of the specificity of TILs showed that in addition to T cells specific for shared tumor/self-melanocyte differentiation antigens,118 such mutated gene products encoded neoepitopes that were often targets of immune recognition.101,108,119 The identification of neoepitope T cells in nonmelanoma cancers is of considerable interest because mutations in other tumors may be similarly targeted. In one example, TIL therapy was used to successfully treat a patient with metastatic cholangiocarcinoma.120 Whole-exome sequencing of the tumor identified a mutation within the erbb2 interacting protein (ERBB2IP) that was recognized by a subset of CD4+ T cells in the TIL product. Treatment with a greater than 95 percent pure population of mutation-reactive T cells resulted in dramatic and durable tumor regression in this patient.120 These results illustrate the potential of patient specific T-cell therapy targeting immunogenic mutations present in their tumor and highlight the need to apply advanced genomic technologies to the discovery of targets for immune therapy in cancer.108,120
Collectively, significant progress has been made in cellular therapy for melanoma, but additional studies are necessary to define the optimal and safest regimens for adoptive therapy with tumor-reactive T cells. Advances in our understanding of the role of individual cytokines in T-cell survival in vitro and in vivo, or the regulation of T-cell activation and homeostasis will provide new opportunities for improving the persistence of in vitro expanded T cells after transfer, perhaps obviating use of toxic chemoradiotherapy to deplete lymphocytes before T-cell infusions. Most of the initial efforts have focused on the CD8+ T-cell response to tumor antigens, but newer data highlights the potential of tumor-specific CD4+ T cells. Combining T-cell therapy with targeted depletion of TREG, checkpoint inhibitors, and vaccines is also under investigation, and may help overcome mechanisms by which tumors evade elimination by limiting the quantity and quality of the host response.
CELLULAR THERAPY OF LEUKEMIA
Allogeneic donor T cells contained in or derived from the stem cell graft can mount a GVL effect that can contribute to the eradication of hematologic cancers, including leukemia.90 This is underscored by studies on the antitumor effects of infusions of unselected donor lymphocytes (DLI) given to patients who relapse after allogeneic HSCT. DLI can have potent antitumor effects in patients with relapsed chronic myeloid leukemia, but has been less effective in acute leukemias, and often complicated by the development of acute and chronic GVHD.91 The identification of leukemia-associated antigens that can be targeted to selectively promote a GVL effect without causing GVHD remains an important goal.92,121
Target Antigens for Leukemia-Specific T Cells
GVHD and GVL effects usually coexist, but a GVL effect can be observed after HSCT in the absence of GVHD.90 Thus, it is presumed there are antigens that are expressed by leukemia cells that can be targeted by allogeneic T cells. Several categories of such antigens have been identified. These include (1) minor histocompatibility antigens (mHAgs) that are selectively expressed in hematopoietic cells including leukemic cells, (2) tumor-specific proteins resulting from chromosome translocations or mutations, and (3) normal proteins that are overexpressed in leukemic cells. Proteins in the latter two classes could be targets both in the transplantation and nontransplantation setting, whereas mHAgs are only relevant after allogeneic HSCT.
Minor Histocompatibility Antigens
The increased potency of the GVL effect after allogeneic HSCT compared with syngeneic HSCT emphasizes the importance of disparity in major HLA and mHAgs for immune-mediated eradication of malignancy.90,122 Class I and class II molecules on recipient T cells display mHAgs, which are peptides derived from proteins that differ between the donor and recipient as a result of genetic polymorphism.92,123 In murine models, the adoptive transfer of T cells specific for a single mHAg eradicated leukemia without causing GVHD.124 In humans, donor T cells reactive with recipient mHAgs can be isolated after transplantation from most allogeneic HSCT recipients.125 Analysis of the specificity of such T-cell clones shows that many mHAgs are expressed preferentially in hematopoietic cells, including leukemic blasts, and might permit the separation of GVL from GVHD.125 mHAg-specific CD8+ CTLs prevent engraftment of human leukemia in nonobese diabetic/severe combined immunodeficiency mice, providing evidence that the leukemic stem cell can be recognized by allogeneic T cells.126
Most mHAgs result from nonsynonymous single nucleotide polymorphisms (SNPs) in the coding sequence of donor and recipient genes that alter the HLA binding or TCR contact of HLA-bound peptides. There are several million SNPs with an allele frequency of greater than 5 percent in the human genome, including approximately 50,000 SNPs that lead to amino acid changes in proteins.127 Identification of the polymorphic genes that encode mHAgs is facilitated by the data on genetic variation from the human HapMap project, which has enabled the use of whole-genome association analysis for mHAg discovery in addition to conventional techniques for antigen discovery.123 Identifying the subset of mHAgs that will be the most useful to target to augment the GVL effect requires consideration of several factors, including the allele frequency of the mHAg encoding gene, the HLA restricting allele that presents the mHAg, and the expression and presentation of the mHAg in leukemia cells and nonhematopoietic tissues.128 Most mHAg discovery efforts have focused on CD8+ T cells, but CD4+ T cells are likely to play a key role either as direct effector cells in the GVL response, or to support the function and persistence of CD8+ T cells, and efforts to define class II MHC-restricted mHAgs remain an important area of investigation.
Autosome-encoded mHAgs that could be targets for therapy of leukemia after allogeneic HSCT include HA-1 and HA-2, which are encoded by KIAA0023 and MY01G, respectively, and presented by HLA-A2; peptides encoded by BCL2A1, which encodes two mHAgs presented by HLA-A24 and HLA-B44, respectively; LRH-1, encoded by the P2X5 gene and presented by HLA-B7; SP110, which is derived by a novel peptide-splicing mechanism and presented by HLA-A3; and PANE-1, which is presented by HLA-A3 and selectively expressed on B-lymphoid malignancies.123,128,129 Additionally, the HLA-A2–restricted and hematopoietic-specific mHAg UTA2–1 has been described.130 Direct evidence for a role of these mHAgs in the GVL effect is provided by studies using HLA-A/peptide tetramers to detect expansion of mHAg-reactive T cells in patients who responded to treatment with DLI for treatment of relapse following transplantation.131,132
There is also evidence for a role of Y-chromosome–encoded mHAgs in the GVL effect. Male recipients of allogeneic HSCT from female donors have a higher risk of GVHD but exhibit a lower risk of leukemia relapse than do other donor/recipient gender combinations, even after controlling for GVHD.133 Several H-Y antigens are ubiquitously expressed in tissues, providing an explanation for the increased GVHD. Identifying mHAgs encoded by the Y chromosome that are selectively expressed on leukemia cells has been more challenging. A UTY epitope presented by HLA-B8 is preferentially presented in hematopoietic cells including acute myeloid leukemia, but the gene is expressed in other tissues and it is not clear if targeting this antigen could avoid GVHD.134
Leukemia-associated proteins that could be targets for cellular therapy have been identified.121 These include mutated proteins, such as p21/Ras, or the products of chromosome translocations, such as BCR/ABL, and the promyelocytic leukemia–retinoic acid receptor α protein (PML-RARα), which can provide unique peptides that represent potential tumor-specific targets.135,136 Nonpolymorphic proteins, such as proteinase 3, Wilms tumor antigen-1 (WT-1), or cyclin A1, which are overexpressed in some leukemias or the leukemic stem cells, also represent potential targets for T-cell therapy.137,138,139,140,141,142 T cells specific for WT-1, which is expressed at high levels in myeloid leukemias but at low levels in normal hematopoietic cells, have been isolated from normal donors by in vitro stimulation of PBMCs with synthetic peptides.143 WT-1–specific CTL selectively lysed leukemic blasts and prevented engraftment of leukemia in immunodeficient mice, suggesting that these T cells may mediate an antileukemic effect without affecting normal hematopoiesis in vivo.140,144 Recent work has identified WT-1–specific T cells after HLA-identical sibling HSCT and correlated these cells with a GVL effect.145 In a recent pilot trial, 11 high-risk leukemia patients received infusions of WT-1–specific T cells.146 In four of these patients, the infused T cells were generated in the presence of IL-21, which modulates the differentiation of T cells in culture. The transferred T cells exhibited evidence of leukemic activity, and CTL that were generated with IL-21 had superior antileukemic activity and survived long-term as memory T cells.146 Additional study in a larger number of patients is needed, but these results offer a potential safe and effective treatment for patients with limited treatment options.
GENETIC RETARGETING OF T CELLS
Extending cellular therapy to patients from whom tumor-reactive T cells cannot be isolated and to other malignancies can be accomplished by using gene transfer approaches to retarget patient T cells to recognize tumor antigens (Fig. 26–3). Two approaches that have already been translated to the clinic are discussed below. The first is the use of retroviral or lentiviral vectors to transfer of TCR α- and β-chain genes isolated from tumor-reactive T cells into T cells obtained from the patient, and the second is the expression of non–MHC-restricted synthetic CARs that target a tumor cell-surface molecule.
Engineering tumor-reactive T cells by insertion of genes that encode tumor-specific T-cell receptors (TCRs) or chimeric antigen receptors (CARs). After isolation of T cells from the desired T-cell subset, the tumor targeting receptor is introduced in the T cells by gene transfer, and the engineered redirected T cells are expanded for reinfusion to the patient.
Genetic Retargeting of T Cells with T-Cell Receptor Genes
Antigen specificity is conferred by the TCR, therefore the transfer of TCR α- and β-chain genes from a T cell of defined specificity into any T cell, will transfer antigen recognition. This strategy can impart specificity to viral antigens, tumor-associated antigens, or mHAgs,93,147,148 and can confer potent tumor recognition. However, TCRs are MHC-restricted and a given TCR construct can only be used to treat patients with tumors expressing the target molecule and also express the MHC restricting allele. Moreover, it can be difficult to achieve the same surface level of TCR expression in transduced T cells as observed in the parental T-cell clone from which the TCR genes were isolated. This problem was apparent in the first clinical trial in which MART-1 TCR-engineered T cells were used to treat melanoma and the low TCR expression likely contributed in part to the limited antitumor activity of the TCR-modified T cells.149 Another problem is that the TCR transfer endows T cells with additional rearranged TCR chains, leading to T cells that could potentially express four different TCR molecules on the cell surface: the natural endogenous TCR, the exogenously introduced TCR, and two mixed heterodimers consisting of endogenous and exogenous TCR chains. Such mismatched TCRs could result in potentially deleterious self-reactive specificities.150 This problem can be mitigated by the introduction of cysteine residues into the extracellular constant region of the introduced α and β TCR chains to provide for disulfide bond formation151,152 or by using murine constant domains in place of the human constant regions,153 both of which promote preferential pairing of the introduced chains. More recently, strategies to knock out or silence the endogenous TCR have been developed.154,155 These modifications provide for more stable pairing during assembly and export, and better competition for limiting components of the TCR complex.156
Strategies to enhance the potency of the TCR, such as by enhancing TCR affinity, may increase the risk of toxicity, particularly if self-antigens are targeted.157,158,159,160,161 In a clinical trial, autologous T cells were modified to express optimized high-affinity MART-1 or gp100 TCR transgenes and transferred to 36 melanoma patients. Nine of the patients exhibited clinical antitumor responses, but “on-target” toxicities to normal melanocytes in the skin, eye, and ear that required local glucocorticoid treatment were observed in a significant fraction of patients that received the high-avidity TCR.157 In a separate trial, autologous T cells modified to express a high-affinity TCR specific for the carcinoembryonic antigen (CEA) were transferred to three patients with colorectal cancer. The serum CEA levels decreased in all three patients and there was some regression of metastatic disease. However, the clinical trial was closed early when all three patients developed severe colitis, putatively from recognition of normal epithelial cells that express CEA.161 Toxicity was also observed in a study in which nine cancer patients were treated with autologous T cells modified to express an anti–MAGE-A3 TCR. Five patients experienced cancer regression, but three patients experienced serious and/or fatal neurologic toxicity. This occurred because of previously unrecognized expression of MAGE-A12, which encodes the identical epitope, in the human brain that resulted in neuronal cell destruction.160 Additional safety concerns of targeting MAGE-A3 using an affinity-enhanced TCR construct were identified after an unexpected fatal cardiac toxicity. Detailed analysis showed cross-reactivity of the engineered MAGE-A3 TCR-modified T cells with a titin-derived peptide.158,159
Such serious “on-target and off-target” toxicities of TCR gene transfer suggested that targeting antigens with restricted expression on tumor cells such as mutant epitopes or cancer testes antigens is preferable. NY-ESO-1 is expressed in many human cancers, but not in normal tissues, except testis. A TCR specific for an HLA-A2-restricted NY-ESO-1 epitope was used for the treatment of melanoma and synovial sarcoma.162 Toxicity was not observed, and encouraging results showed that the TCR-engineered T cells mediated objective responses in four of six patients with synovial cell carcinoma and five of 11 melanoma patients.
Genetic Retargeting of T Cells with Chimeric Antigen Receptors
A notable advance in T-cell therapy has been the development of CARs that link recognition domains of antibodies to molecules involved in signaling T-cell effector function.94,95,96,163,164 CARs typically consist of a single-chain variable fragment (scFV) derived from the VH and VL sequences of a mAb specific for a tumor cell-surface molecule, fused to a trans-membrane domain, as well as the CD3ζ-signaling domain alone, or in combination with one or more costimulatory signaling modules, such as CD28,4–1BB, OX40, or CD27 (Fig. 26–4).94,95,96,165,166 CARs can be introduced into T cells by gene transfer to target surface molecules expressed on tumors. Unlike conventional TCRs, CARs are not MHC-restricted and have the advantage that a single construct can be used to treat all patients expressing the tumor antigen.
Structure of a chimeric antigen receptor (CAR). The CAR is typically composed of a recognition module that is fused in tandem to nonsignaling extracellular and transmembrane domains and intracellular signaling elements. The recognition module, spacer domain, and signaling modules of the CAR can be modified to optimize tumor cell recognition and T-cell function.
A large number of CARs targeting a variety of molecules expressed on hematologic malignancies have been developed.96,167,168 Examples for candidate targets for the treatment of hematologic malignancies include CD19,96 CD20,169 and the orphan tyrosine kinase receptor ROR1170,171,172 expressed on B-cell lymphomas and leukemias; Lewis Y,173 CD44v6,174 CD33 and/or CD123 expressed on acute myeloid leukemia168,175,176,177; and NKG2D ligands178 or the B-cell maturation antigen (BCMA)179 expressed on myeloma.
Efficient methods for T-cell activation, gene transfer, and expansion of T cells for therapy have been developed and the results of pilot clinical trials of T cells modified to express CARs specific for CD19, CD20, ERBB2, the disialoganglioside GD2, and mesothelin have been reported and provided evidence of in vivo antitumor activity.180,181,182,183,184,185,186,187 However, serious toxicity has also been observed including a fatal toxicity shortly after infusion of a single high dose of ERBB2-specific CAR-T cells that contained both CD28 and 4–1BB costimulatory domains, and was attributed to cytokine release and recognition of normal lung epithelium that expresses low amounts of ERBB2.188 Toxicity to normal tissues was also observed in patients receiving CAIX CAR-engineered T cells for the treatment of metastatic renal cell carcinoma.189,190
The most encouraging results with CAR-T cells have been obtained targeting the B-cell lineage-restricted CD19 molecule that is expressed on B-cell leukemias and lymphomas. Dramatic and durable remissions in patients with chronic lymphatic leukemia (CLL) and acute lymphatic leukemia (ALL) have been reported after infusion of autologous T cells transduced with CD19-specific CARs that contained either a CD28 or a 4–1BB costimulatory domain.182,183,191,192 In these studies, infused T cells were shown to expand in vivo, induce tumor lysis and a deficiency of normal CD19+ B cells, and persist long-term in some patients. The complete response rate appears to be higher in patients with ALL than those with CLL or lymphoma for reasons that have not been elucidated. Tumor regression is often associated with a cytokine release syndrome (CRS) initiated by activation of CAR-T cells in vivo, associated with elevated levels of IFN-γ, IL-6, and tumor necrosis factor, and resulting in high fever, hypotension, and neurologic abnormalities.183,193 CRS is more severe in patients with high tumor burden and can require intensive supportive care, and treatment with glucocorticoids and/or anti IL-6 receptor antibodies. CAR-T cells often do not persist long term, although prolonged B-cell aplasia has been observed in a subset of patients.
The finding that durable responses can be achieved in some patients with advanced B-cell malignancies illustrates the potency of CAR-T cells, and suggests that future work to define optimal design of the CD19 CAR constructs and to identify the optimal subset(s) of T cells to modify may further improve outcomes, particularly in CLL and lymphoma where response rates and durability are lower. The demonstration of the superior engraftment properties of effector T (TE) cells derived from TCM would suggest that selection or enrichment of TCM prior to gene insertion may provide a superior T-cell product for adoptive therapy.48,194 This hypothesis is being examined prospectively in clinical trials of CAR–T-cell therapy. Moreover, integrating T-cell therapy earlier after diagnosis or after autologous HSCT, in which marked tumor cytoreduction can be achieved by intensive conditioning, may further improve outcome and reduce the toxicity resulting from CAR-mediated tumor lysis.
CAR-modified T cells also may have applications in the treatment of solid tumors. Candidate surface molecules on solid tumors that are being actively pursued as targets include GD2,186 mesothelin,187 L1-cell adhesion molecule (L1CAM),85 ROR1,170,171 prostate stem cell antigen,195 folate receptor,196 and the fibroblast activation protein.197 Significant antitumor activity without toxicity has been reported in patients with neuroblastoma treated with T cells modified with a first-generation GD2 CAR.186 The persistence of the transferred cells was relatively short in that study, however, perhaps owing to the lack of costimulation in the CAR. GD2 is expressed on normal peripheral nerves, and on-target toxicity from a sustained T-cell response will need to be monitored if more potent CARs are being examined. Suitable animal models for preclinical toxicity studies are urgently needed and are being developed.197,198,199 It will also be important to address the multiple evasion mechanisms that tumors employ to avoid immune elimination.96 For example, the programmed death-1 (PD-1) receptor is a negative regulator of TE mechanisms that limits immune responses against cancer, and the combination of checkpoint blockade and T-cell therapy is of future interest.200,201,202
Suicide Genes for Conditional Ablation
The results of clinical trials of T-cell therapy have revealed toxicity to normal tissues in some patients. The introduction of a suicide gene into the T cells that could be activated if toxicity occurred has long been the subject of research. The HSV-thymidine kinase (HSV-TK) gene has been used in gene therapy trials in the clinic and was effective in reversing GVHD after DLI.203 However, the viral thymidine kinase is immunogenic and can result in premature elimination of transferred T cells that do not cause toxicity.204 Novel suicide genes based on inducing conditional cell death through activation of CD95 (Fas) or caspases using a chemical dimerizer such as AP1903 to activate an engineered chimeric human CD95 or caspase transgene product may circumvent the problem of immunogenicity.205,206,207 Recently, 10 patients undergoing haploidentical HSCT for relapsed acute leukemia were treated with donor T cells modified to express an inducible caspase9 gene (iCasp9). The iCasp9 T cells engrafted and conferred protection against infectious diseases. A subset of the patients developed GVHD. A single dose of the dimerizing drug, given to those patients who developed GVHD ablated more than 90 percent of the iCasp9-modified T cells rapidly and eliminated the GVHD without recurrence or adverse events.208,209 A truncated epidermal growth factor receptor (EGFR) marker/suicide gene that could be targeted in vivo using a clinical grade anti-EGFR antibody (Erbitux) has been developed, but its efficacy for ablating transferred T cells in patients has not been established.