Identification and structural characterization of a mutant KRAS-G12V specific TCR restricted by HLA-A3.

Mutations in KRAS are some of the most common across multiple cancer types and are thus attractive targets for therapy. Recent studies demonstrated that mutant KRAS generates immunogenic neoantigens that are targetable by adoptive T-cell therapy in metastatic diseases. To expand mutant KRAS-specific immunotherapies, it is critical to identify additional HLA-I allotypes that can present KRAS neoantigens and their cognate T-cell receptors (TCR). Here, we identified a murine TCR specific to a KRAS-G12V neoantigen (7VVVGAVGVGK16) using a vaccination approach with transgenic mice expressing HLA-A*03:01 (HLA-A3). This TCR demonstrated exquisite specificity for mutant G12V and not WT KRAS peptides. To investigate the molecular basis for neoantigen recognition by this TCR, we determined its structure in complex with HLA-A3(G12V). G12V-TCR CDR3ß and CDR1ß formed a hydrophobic pocket to interact with p6 Val of the G12V but not the WT KRAS peptide. To improve the tumor sensitivity of this TCR, we designed rational substitutions to improve TCR:HLA-A3 contacts. Two substitutions exhibited modest improvements in TCR binding avidity to HLA-A3 (G12V) but did not sufficiently improve T-cell sensitivity for further clinical development. Our study provides mechanistic insight into how TCRs detect neoantigens and reveals the challenges in targeting KRAS-G12V mutations.

Using patient-derived tumor organoids from common epithelial cancers to analyze personalized T-cell responses to neoantigens.

Adoptive cell transfer of tumor-infiltrating lymphocytes (TIL) can mediate durable complete responses in some patients with common epithelial cancers but does so infrequently. A better understanding of T-cell responses to neoantigens and tumor-related immune evasion mechanisms requires having the autologous tumor as a reagent. We investigated the ability of patient-derived tumor organoids (PDTO) to fulfill this need and evaluated their utility as a tool for selecting T-cells for adoptive cell therapy. PDTO established from metastases from patients with colorectal, breast, pancreatic, bile duct, esophageal, lung, and kidney cancers underwent whole exomic sequencing (WES), to define mutations. Organoids were then evaluated for recognition by autologous TIL or T-cells transduced with cloned T-cell receptors recognizing defined neoantigens. PDTO were also used to identify and clone TCRs from TIL targeting private neoantigens and define those tumor-specific targets. PDTO were successfully established in 38/47 attempts. 75% were available within 2 months, a timeframe compatible with screening TIL for clinical administration. These lines exhibited good genetic fidelity with their parental tumors, especially for mutations with higher clonality. Immunologic recognition assays demonstrated instances of HLA allelic loss not found by pan-HLA immunohistochemistry and in some cases WES of fresh tumor. PDTO could also be used to show differences between TCRs recognizing the same antigen and to find and clone TCRs recognizing private neoantigens. PDTO can detect tumor-specific defects blocking T-cell recognition and may have a role as a selection tool for TCRs and TIL used in adoptive cell therapy.

Proteogenomic Analysis Unveils the HLA Class I-Presented Immunopeptidome in Melanoma and EGFR-Mutant Lung Adenocarcinoma.

Immune checkpoint inhibitors and adoptive lymphocyte transfer-based therapies have shown great therapeutic potential in cancers with high tumor mutational burden (TMB), such as melanoma, but not in cancers with low TMB, such as mutant epidermal growth factor receptor (EGFR)-driven lung adenocarcinoma. Precision immunotherapy is an unmet need for most cancers, particularly for cancers that respond inadequately to immune checkpoint inhibitors. Here, we employed large-scale MS-based proteogenomic profiling to identify potential immunogenic human leukocyte antigen (HLA) class I-presented peptides in melanoma and EGFR-mutant lung adenocarcinoma. Similar numbers of peptides were identified from both tumor types. Cell line and patient-specific databases (DBs) were constructed using variants identified from whole-exome sequencing. A de novo search algorithm was used to interrogate the HLA class I immunopeptidome MS data. We identified 12 variant peptides and several classes of tumor-associated antigen-derived peptides. We constructed a cancer germ line (CG) antigen DB with 285 antigens. This allowed us to identify 40 class I-presented CG antigen-derived peptides. The class I immunopeptidome comprised more than 1000 post-translationally modified (PTM) peptides representing 58 different PTMs, underscoring the critical role PTMs may play in HLA binding. Finally, leveraging de novo search algorithm and an annotated long noncoding RNA (lncRNA) DB, we developed a novel lncRNA-encoded peptide discovery pipeline to identify 44 lncRNA-derived peptides that are presented by class I. We validated tandem MS spectra of select variant, CG antigen, and lncRNA-derived peptides using synthetic peptides and performed HLA class I-binding assays to demonstrate binding to class I proteins. In summary, we provide direct evidence of HLA class I presentation of a large number of variant and tumor-associated peptides in both low and high TMB cancer. These results can potentially be useful for precision immunotherapies, such as vaccine or adoptive cell therapies in melanoma and EGFR-mutant lung cancers.

αß T cell receptors that do not undergo major histocompatibility complex-specific thymic selection possess antibody-like recognition specificities.

Major histocompatibility complex (MHC) restriction is the cardinal feature of T cell antigen recognition and is thought to be intrinsic to αß T cell receptor (TCR) structure because of germline-encoded residues that impose MHC specificity. Here, we analyzed αßTCRs from T cells that had not undergone MHC-specific thymic selection. Instead of recognizing peptide-MHC complexes, the two αßTCRs studied here resembled antibodies in recognizing glycosylation-dependent conformational epitopes on a native self-protein, CD155, and they did so with high affinity independently of MHC molecules. Ligand recognition was via the αßTCR combining site and involved the identical germline-encoded residues that have been thought to uniquely impose MHC specificity, demonstrating that these residues do not only promote MHC binding. This study demonstrates that, without MHC-specific thymic selection, αßTCRs can resemble antibodies in recognizing conformational epitopes on MHC-independent ligands.

Expression profiling of TCR-engineered T cells demonstrates overexpression of multiple inhibitory receptors in persisting lymphocytes.

Despite significant progress in the development of adoptive cell-transfer therapies (ACTs) using gene-engineered T cells, little is known about the fate of cells following infusion. To address that, we performed a comparative analysis of gene expression between T-cell receptor-engineered lymphocytes persisting in the circulation 1 month after administration and the product that was infused. We observed that 156 genes related to immune function were differentially expressed, including underexpression of stimulators of lymphocyte function and overexpression of inhibitory genes in postinfusion cells. Of genes overexpressed postinfusion, the product of programmed cell death 1 (PDCD1), coinhibitory receptor PD-1, was expressed at a higher percentage in postinfusion lymphocytes than in the infusion product. This was associated with a higher sensitivity to inhibition of cytokine production by interaction with its ligand PD-L1. Coinhibitory receptor CD160 was also overexpressed in persisting cells, and its expression was associated with decreased reactivity, which surprisingly was found to be ligand-independent. These results contribute to a deeper understanding of the properties of transgenic lymphocytes used to treat human malignancies and may provide a rationale for the development of combination therapies as a method to improve ACT.

An antigenic peptide produced by reverse splicing and double asparagine deamidation.

A variety of unconventional translational and posttranslational mechanisms contribute to the production of antigenic peptides, thereby increasing the diversity of the peptide repertoire presented by MHC class I molecules. Here, we describe a class I-restricted peptide that combines several posttranslational modifications. It is derived from tyrosinase and recognized by tumor-infiltrating lymphocytes isolated from a melanoma patient. This unusual antigenic peptide is made of two noncontiguous tyrosinase fragments that are spliced together in the reverse order. In addition, it contains two aspartate residues that replace the asparagines encoded in the tyrosinase sequence. We confirmed that this peptide is naturally presented at the surface of melanoma cells, and we showed that its processing sequentially requires translation of tyrosinase into the endoplasmic reticulum and its retrotranslocation into the cytosol, where deglycosylation of the two asparagines by peptide-N-glycanase turns them into aspartates by deamidation. This process is followed by cleavage and splicing of the appropriate fragments by the standard proteasome and additional transport of the resulting peptide into the endoplasmic reticulum through the transporter associated with antigen processing (TAP).

Molecular identification of an MHC-independent ligand recognized by a human {alpha}/{beta} T-cell receptor.

During an analysis of T-cell responses against human renal cell carcinoma (RCC), we identified a CD4(+) T-cell line that showed TCR-mediated recognition and lysis of nearly all RCC lines regardless of MHC type. We have now elucidated the nature of the ligand for this α/ß TCR, and it contains no MHC-related moiety and does not involve classic peptide processing. First, matrix metalloproteinase 14 (MMP14) expressed on RCC cells releases membrane-bound TRAIL expressed by the T cell; then, soluble TRAIL binds to its receptor DR4 (TRAIL-R1), which is expressed on tumor cells, and this TRAIL-DR4 complex is recognized by the TCR through a complementarity-determining region 3α (CDR3α)-mediated interaction. Direct and specific antigen-TCR interaction was demonstrated when the immobilized recombinant TRAIL/DR4 complex stimulated the TCR. In addition, amino acid substitutions in the CDR3α of the TCR either obliterated or enhanced target-specific recognition. This description of the molecular nature of a non-MHC target structure recognized by a naturally occurring α/ß TCR not only broadens our concept of what the TCR can recognize, but also raises the question of whether such a T cell could be of clinical utility against RCC.

Identification of neoepitope reactive T-cell receptors guided by HLA-A*03:01 and HLA-A*11:01 immunopeptidomics.

BACKGROUND: Tumor-specific mutated proteins can create immunogenic non-self, mutation-containing 'neoepitopes' that are attractive targets for adoptive T-cell therapies. To avoid the complexity of defining patient-specific, private neoepitopes, there has been major interest in targeting common shared mutations in driver genes using off-the-shelf T-cell receptors (TCRs) engineered into autologous lymphocytes. However, identifying the precise naturally processed neoepitopes to pursue is a complex and challenging process. One method to definitively demonstrate whether an epitope is presented at the cell surface is to elute peptides bound to a specific major histocompatibility complex (MHC) allele and analyze them by mass spectrometry (MS). These MS data can then be prospectively applied to isolate TCRs specific to the neoepitope. METHODS: We created mono-allelic cell lines expressing one class I HLA allele and one common mutated oncogene in order to eliminate HLA deconvolution requirements and increase the signal of recovered peptides. MHC-bound peptides on the surface of these cell lines were immunoprecipitated, purified, and analyzed using liquid chromatography-tandem mass spectrometry, producing a list of mutation-containing minimal epitopes. To validate the immunogenicity of these neoepitopes, HLA-transgenic mice were vaccinated using the minimal peptides identified by MS in order to generate neoepitope-reactive TCRs. Specificity of these candidate TCRs was confirmed by peptide titration and recognition of transduced targets. RESULTS: We identified precise neoepitopes derived from mutated isoforms of KRAS, EGFR, BRAF, and PIK3CA presented by HLA-A*03:01 and/or HLA-A*11:01 across multiple biological replicates. From our MS data, we were able to successfully isolate murine TCRs that specifically recognize four HLA-A*11:01 restricted neoepitopes (KRAS G13D, PIK3CA E545K, EGFR L858R and BRAF V600E) and three HLA-A*03:01 restricted neoepitopes (KRAS G12V, EGFR L858R and BRAF V600E). CONCLUSIONS: Our data show that an MS approach can be used to demonstrate which shared oncogene-derived neoepitopes are processed and presented by common HLA alleles, and those MS data can rapidly be used to develop TCRs against these common tumor-specific antigens. Although further characterization of these neoepitope-specific murine TCRs is required, ultimately, they have the potential to be used clinically for adoptive cell therapy.

Splicing of distant peptide fragments occurs in the proteasome by transpeptidation and produces the spliced antigenic peptide derived from fibroblast growth factor-5.

Peptide splicing is a newly described mode of production of antigenic peptides presented by MHC class I molecules, whereby two noncontiguous fragments of the parental protein are joined together after excision of the intervening segment. Three spliced peptides have been described. In two cases, splicing involved the excision of a short intervening segment of 4 or 6 aa and was shown to occur in the proteasome by transpeptidation resulting from the nucleophilic attack of an acyl-enzyme intermediate by the N terminus of the other peptide fragment. For the third peptide, which is derived from fibroblast growth factor-5 (FGF-5), the splicing mechanism remains unknown. In this case, the intervening segment is 40 aa long. This much greater length made the transpeptidation model more difficult to envision. Therefore, we evaluated the role of the proteasome in the splicing of this peptide. We observed that the spliced FGF-5 peptide was produced in vitro after incubation of proteasomes with a 49-aa-long precursor peptide. We evaluated the catalytic mechanism by incubating proteasomes with various precursor peptides. The results confirmed the transpeptidation model of splicing. By transfecting a series of mutant FGF-5 constructs, we observed that reducing the length of the intervening segment increased the production of the spliced peptide, as predicted by the transpeptidation model. Finally, we observed that trans-splicing (i.e., splicing of fragments from two distinct proteins) can occur in the cell, but with a much lower efficacy than splicing of fragments from the same protein.

A phenotypic signature that identifies neoantigen-reactive T cells in fresh human lung cancers.

A common theme across multiple successful immunotherapies for cancer is the recognition of tumor-specific mutations (neoantigens) by T cells. The rapid discovery of such antigen responses could lead to improved therapies through the adoptive transfer of T cells engineered to express neoantigen-reactive T cell receptors (TCRs). Here, through CITE-seq (cellular indexing of transcriptomes and epitopes by sequencing) and TCR-seq of non-small cell lung cancer (NSCLC) tumor-infiltrating lymphocytes (TILs), we develop a neoantigen-reactive T cell signature based on clonotype frequency and CD39 protein and CXCL13 mRNA expression. Screening of TCRs selected by the signature allows us to identify neoantigen-reactive TCRs with a success rate of 45% for CD8+ and 66% for CD4+ T cells. Because of the small number of samples analyzed (4 patients), generalizability remains to be tested. However, this approach can enable the quick identification of neoantigen-reactive TCRs and expedite the engineering of personalized neoantigen-reactive T cells for therapy.

Molecular signatures of antitumor neoantigen-reactive T cells from metastatic human cancers.

The accurate identification of antitumor T cell receptors (TCRs) represents a major challenge for the engineering of cell-based cancer immunotherapies. By mapping 55 neoantigen-specific TCR clonotypes (NeoTCRs) from 10 metastatic human tumors to their single-cell transcriptomes, we identified signatures of CD8+ and CD4+ neoantigen-reactive tumor-infiltrating lymphocytes (TILs). Neoantigen-specific TILs exhibited tumor-specific expansion with dysfunctional phenotypes, distinct from blood-emigrant bystanders and regulatory TILs. Prospective prediction and testing of 73 NeoTCR signature-derived clonotypes demonstrated that half of the tested TCRs recognized tumor antigens or autologous tumors. NeoTCR signatures identified TCRs that target driver neoantigens and nonmutated viral or tumor-associated antigens, suggesting a common metastatic TIL exhaustion program. NeoTCR signatures delineate the landscape of TILs across metastatic tumors, enabling successful TCR prediction based purely on TIL transcriptomic states for use in cancer immunotherapy.

Regression of human kidney cancer following allogeneic stem cell transplantation is associated with recognition of an HERV-E antigen by T cells.

Transplanted donor lymphocytes infused during hematopoietic stem cell transplantation (HSCT) have been shown to cure patients with hematological malignancies. However, less is known about the effects of HSCT on metastatic solid tumors. Thus, a better understanding of the immune cells and their target antigens that mediate tumor regression is urgently needed to develop more effective HSCT approaches for solid tumors. Here we report regression of metastatic renal cell carcinoma (RCC) in patients following nonmyeloablative HSCT consistent with a graft-versus-tumor effect. We detected RCC-reactive donor-derived CD8(+) T cells in the blood of patients following nonmyeloablative HSCT. Using cDNA expression cloning, we identified a 10-mer peptide (CT-RCC-1) as a target antigen of RCC-specific CD8(+) T cells. The genes encoding this antigen were found to be derived from human endogenous retrovirus (HERV) type E and were expressed in RCC cell lines and fresh RCC tissue but not in normal kidney or other tissues. We believe this to be the first solid tumor antigen identified using allogeneic T cells from a patient undergoing HSCT. These data suggest that HERV-E is activated in RCC and that it encodes an overexpressed immunogenic antigen, therefore providing a potential target for cellular immunity.

A machine learning model for ranking candidate HLA class I neoantigens based on known neoepitopes from multiple human tumor types.

Tumor neoepitopes presented by major histocompatibility complex (MHC) class I are recognized by tumor-infiltrating lymphocytes (TIL) and are targeted by adoptive T-cell therapies. Identifying which mutant neoepitopes from tumor cells are capable of recognition by T cells can assist in the development of tumor-specific, cell-based therapies and can shed light on antitumor responses. Here, we generate a ranking algorithm for class I candidate neoepitopes by using next-generation sequencing data and a dataset of 185 neoepitopes that are recognized by HLA class I-restricted TIL from individuals with metastatic cancer. Random forest model analysis showed that the inclusion of multiple factors impacting epitope presentation and recognition increased output sensitivity and specificity compared to the use of predicted HLA binding alone. The ranking score output provides a set of class I candidate neoantigens that may serve as therapeutic targets and provides a tool to facilitate in vitro and in vivo studies aimed at the development of more effective immunotherapies.

Characterization of a novel nonclassical T cell clone with broad reactivity against human renal cell carcinomas.

A CD4(+) T cell clone (HC/2G-1) was established by stimulating peripheral blood T cells from a patient with renal cell carcinoma (RCC) with dendritic cells preincubated with the autologous apoptotic renal tumor line in the presence of IFN-alpha. It recognizes the autologous RCC and most allogeneic RCC lines by IFN-gamma release (10 of 11 lines) and lysis (9 of 10 lines), but does not recognize multiple EBV B cells or fibroblasts. It shows little or no recognition of a panel of melanomas, breast cancers and non-small-cell lung cancers. Phenotypically, HC/2G-1 is CD3(+)CD4(+) TCR alphabeta(+), but CD161(-)CD16(-)NKG2D(-). Tumor recognition by clone HC/2G-1 was not blocked by Abs to HLA class I or class II, but was significantly reduced by anti-TCR alphabeta Ab. Furthermore, tumor recognition was beta(2)-microglobulin-independent. HC/2G-1 does not use a Valpha or Vbeta described for classical NKT cells, but rather Valpha14 and Vbeta2.1. Allogeneic T cells cotransfected with mRNAs encoding the alpha and beta chains of the HC/2G-1 TCR recognized renal tumor lines, demonstrating that tumor recognition is TCR-mediated. Interestingly, TRAIL appears to play a role in tumor recognition by HC/2G-1 in that reactivity was blocked by anti-TRAIL Ab, and soluble TRAIL could enhance IFN-gamma secretion by HC/2G-1 in response to renal tumors. Our findings suggest that clone HC/2G-1 represents a novel type of CD4(+) cell that has broad TCR-mediated recognition of a determinant widely expressed by RCC.

Immune recognition of a human renal cancer antigen through post-translational protein splicing.

Cytotoxic T lymphocytes (CTLs) detect and destroy cells displaying class I molecules of the major histocompatibility complex (MHC) that present oligopeptides derived from aberrant self or foreign proteins. Most class I peptide ligands are created from proteins that are degraded by proteasomes and transported, by the transporter associated with antigen processing, from the cytosol into the endoplasmic reticulum, where peptides bind MHC class I molecules and are conveyed to the cell surface. C2 CTLs, cloned from human CTLs infiltrating a renal cell carcinoma, kill cancer cells overexpressing fibroblast growth factor-5 (FGF-5). Here we show that C2 cells recognize human leukocyte antigen-A3 MHC class I molecules presenting a nine-residue FGF-5 peptide generated by protein splicing. This process, previously described strictly in plants and unicellular organisms, entails post-translational excision of a polypeptide segment followed by ligation of the newly liberated carboxy-terminal and amino-terminal residues. The occurrence of protein splicing in vertebrates has important implications for the complexity of the vertebrate proteome and for the immune recognition of self and foreign peptides.

Novel MHC-Independent αßTCRs Specific for CD48, CD102, and CD155 Self-Proteins and Their Selection in the Thymus.

MHC-independent αßTCRs (TCRs) recognize conformational epitopes on native self-proteins and arise in mice lacking both MHC and CD4/CD8 coreceptor proteins. Although naturally generated in the thymus, these TCRs resemble re-engineered therapeutic chimeric antigen receptor (CAR) T cells in their specificity for MHC-independent ligands. Here we identify naturally arising MHC-independent TCRs reactive to three native self-proteins (CD48, CD102, and CD155) involved in cell adhesion. We report that naturally arising MHC-independent TCRs require high affinity TCR-ligand engagements in the thymus to signal positive selection and that high affinity positive selection generates a peripheral TCR repertoire with limited diversity and increased self-reactivity. We conclude that the affinity of TCR-ligand engagements required to signal positive selection in the thymus inversely determines the diversity and self-tolerance of the mature TCR repertoire that is selected.

An effective mouse model for adoptive cancer immunotherapy targeting neoantigens.

The adoptive cell transfer (ACT) of T cells targeting mutated neoantigens can cause objective responses in varieties of metastatic cancers, but the development of new T cell-based treatments relies on accurate animal models. To investigate the therapeutic effect of targeting a neoantigen with ACT, we used T cells from pmel-1 T cell receptor-transgenic mice, known to recognize a WT peptide, gp100, and a mutated version of the peptide that has higher avidity. We gene-engineered B16 cells to express the WT or mutated gp100 epitopes and found that pmel-1-specific T cells targeting a neoantigen tumor target augmented recognition as measured by IFN-γ production. Neoantigen expression by B16 also enhanced the capacity of pmel-1 T cells to trigger the complete and durable regression of large, established, vascularized tumor and required less lymphodepleting conditioning. Targeting neoantigen uncovered the possibility of using enforced expression of the IL-2Rα chain (CD25) in mutation-reactive CD8+ T cells to improve their antitumor functionality. These data reveal that targeting of "mutated-self" neoantigens may lead to improved efficacy and reduced toxicities of T cell-based cellular immunotherapies for patients with cancer.

Cytotoxic T-lymphocyte-associated protein 4 expressed by melanoma cells does not affect melanoma-specific cytotoxic T lymphocytes in the effector phase.

Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is one of the important molecules that regulate the anti-melanoma T-cell response. Currently, there are some reports showing that CTLA-4 is expressed not only by T cells but also by various kinds of tumor cells, including melanoma cells. However, there is no report that shows the role of CTLA-4 expressed by melanoma cells in melanoma-specific cytotoxic T-lymphocyte (CTL) response. In this report, we confirmed substantial CTLA-4 expression and the localization of CTLA-4 in melanoma cell lines and tissues. Also, we examined its impact on melanoma-specific CTL in vitro, and found that CTLA-4 expressed by melanoma cells does not affect melanoma-specific CTL in the effector phase. Our findings suggest the importance of elucidating the role of CTLA-4 expressed by melanoma cells, particularly in anti-CTLA-4 antibody therapy.

Multiple protein-protein interactions by RNA polymerase I-associated factor PAF49 and role of PAF49 in rRNA transcription.

We previously demonstrated the critical role of RNA polymerase I (Pol I)-associated factor PAF53 in mammalian rRNA transcription. Here, we report the isolation and characterization of another Pol I-associated factor, PAF49. Mouse PAF49 shows striking homology to the human nucleolar protein ASE-1, so that they are considered orthologues. PAF49 and PAF53 were copurified with a subpopulation of Pol I during purification from cell extracts. Physical association of PAF49 with Pol I was confirmed by a coimmunoprecipitation assay. PAF49 was shown to interact with PAF53 through its N-terminal segment. This region of PAF49 also served as the target for TAF(I)48, the 48-kDa subunit of selectivity factor SL1. Concomitant with this interaction, the other components of SL1 also coimmunoprecipitated with PAF49. Specific transcription from the mouse rRNA promoter in vitro was severely impaired by anti-PAF49 antibody, which was overcome by addition of recombinant PAF49 protein. Moreover, overexpression of a deletion mutant of PAF49 significantly reduced pre-rRNA synthesis in vivo. Immunolocalization analysis revealed that PAF49 accumulated in the nucleolus of growing cells but dispersed to nucleoplasm in growth-arrested cells. These results strongly suggest that PAF49/ASE-1 plays an important role in rRNA transcription.

Preclinical Evaluation of Chimeric Antigen Receptors Targeting CD70-Expressing Cancers.

Purpose: CD70 expression in normal tissues is restricted to activated lymphoid tissues. Targeting CD70 on CD70-expressing tumors could mediate "on-target, off-tumor" toxicity. This study was to evaluate the feasibility and safety of using anti-human CD70 CARs to treat cancer patients whose tumors express CD70.Experimental Design: Seven anti-human CD70 CARs with binding moieties from human CD27 combined with CD3-zeta and different costimulatory domains from CD28 and/or 41BB were constructed. In vitro functionality of these receptors was compared and in vivo treatment efficacy was evaluated in a xenograft mouse model. A homologous, all murine anti-CD70 CAR model was also used to assess treatment-related toxicities.Results: The CAR consisting of the extracellular binding portion of CD27 fused with 41BB and CD3-zeta (trCD27-41BB-zeta) conferred the highest IFNγ production against CD70-expressing tumors in vitro, and NSG mice bearing established CD70-expressing human tumors could be cured by human lymphocytes transduced with this CAR. In the murine CD27-CD3-zeta CAR model, significant reduction of established tumors and prolonged survival were achieved using CAR-transduced splenocytes in a dose-dependent manner. Host preirradiation enhanced treatment efficacy but increased treatment-related toxicities such as transient weight loss and hematopoetic suppression. The treatment did not appear to block adaptive host immune responses.Conclusions: Preclinical testing supports the safety and efficacy of a CD27-containing CAR targeting CD70-expressing tumors. Clin Cancer Res; 23(9); 2267-76. ©2016 AACR.