We explored the possibility to target Ewing's sarcoma family of tumors (ESFT) by redirecting T cells. To this aim, we considered NKG2D-ligands (NKG2D-Ls) as possible target antigens. Detailed analysis of the expression of MICA, MICB, ULBP-1, -2, and -3 in fourteen ESFT cell lines revealed consistent expression of at least one NKG2D-L. Thus, for redirecting T cells, we fused a CD3ζ/CD28-derived signaling domain to the ectodomain of NKG2D, however, opposite transmembrane orientation of this signaling domain and NKG2D required inverse orientation fusion of either of them. We hypothesized that the particularly located C-terminus of the NKG2D ectodomain should allow reengineering of the membrane anchoring from a native N-terminal to an artificial C-terminal linkage. Indeed, the resulting chimeric NKG2D receptor (chNKG2D) was functional and efficiently mediated ESFT cell death triggered by activated T cells. Notably, ESFT cells with even low NKG2D-L expression were killed by CD8pos and also CD4pos cells. Both, mRNA transfection and lentiviral transduction resulted in high level surface expression of chNKG2D. However, upon target-cell recognition receptor surface levels were maintained by tranfected RNA only during the first couple of hours after transfection. Later, target-cell contact resulted in strong and irreversible receptor down-modulation, whereas lentivirally mediated expression of chNKG2D remained constant under these conditions. Together, our study defines NKG2D-Ls as targets for a CAR-mediated T cell based immunotherapy of ESFT. A comparison of two different methods of gene transfer reveals strong differences in the susceptibility to ligand-induced receptor down-modulation with possible implications for the applicability of RNA transfection.
Citation: Lehner M, Götz G, Proff J, Schaft N, Dörrie J, Full F, et al. (2012) Redirecting T Cells to Ewing's Sarcoma Family of Tumors by a Chimeric NKG2D Receptor Expressed by Lentiviral Transduction or mRNA Transfection. PLoS ONE 7(2): e31210. doi:10.1371/journal.pone.0031210
Editor: Rachel Louise Allen, University of London, St George's, United Kingdom
Received: September 5, 2011; Accepted: January 4, 2012; Published: February 15, 2012
Copyright: © 2012 Lehner et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was financially supported by the Wilhelm Sander-Foundation (project Nr. 2002.034.1 and 2009.002.1), Staedtler-Foundation (project 24/09), the Interdisciplinary Center for Clinical Research (IZKF: Genesis, Diagnostics and Therapy of Inflammation Processes), the German Research Foundation (DFG: project SCHA 1247/1-1), EN 423/2-1, and SFB796) and a grant from the Fondazione Berlucchi, Brescia, Italy. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
A subgroup of patients with Ewing's sarcoma family of tumors (ESFT) is still threatened by a poor long term prognosis. Despite modern multimodal therapy (chemotherapy, radiation, and surgery) ESFT relapse in about 30% of patients with localized disease. Long-term survival of those who relapsed and of patients with metastatic disease at diagnosis is currently below 30% –. Therefore, new treatment options are needed.
Tumors cells frequently up-regulate “stress” induced ligands recognized by the NK cell activating receptors DNAM-1 (CD226) and NKG2D (CD314), whose ligands have been found recently also on ESFT cells . Therefore, the infusion of NK cells has emerged as a promising new treatment strategy for malignant tumors in general and has been also suggested for the treatment of ESFT , . NK cells due to their innate specificity allow tumor targeting without extensive ex vivo modification, and have not been reported to cause auto- or allo-immune side-effects following transfusion even across MHC barriers , . CD8pos T cells, on the other hand, are characterized by the capacity to differentiate into effector cells or long term memory cells, and have been adopted in the past with a broad range of new antigenic specificities by receptor transfer (for review see ).
NKG2D recognizes several ligands (MICA, MICB, ULBP-1 to ULBP-6) with only restricted expression in normal tissues , . Utilizing this receptor for redirecting T cells, Sentman and co-workers recently reported the construction of an NKG2D-based chimeric T cell antigen receptor (CAR) and demonstrated its efficacy against a variety of malignant cells in vivo and in vitro –. We wanted to further develop this approach for targeting ESFT and decided to construct a similar NKG2D-based CAR, but - based on the known 3D-structure of NKG2D - focused on reengineering the extra-cellular domain of NKG2D to allow its fusion in type I membrane protein orientation to the signaling domain of a second generation CAR, which contains also CD28 for enhanced signal transduction , , . Here we report that this new chNKG2D successfully allowed us to establish a T cell based approach for efficient ESFT cell elimination.
For expression of this chimeric receptor we compared two different methods of gene transfer, namely mRNA transfection and lentiviral transduction, since both methods are currently considered for clinical application. Lentiviral mediated gene transfer mediates stable transgene expression and therefore is attractive, because the clinical effectiveness of adoptive tumor therapy has been correlated with the persistence of transferred cells . On the other hand, the transient nature of receptor expression following mRNA transfection might be an advantage under the aspects of safety and other regulatory issues. Although mRNA transfection thus has important benefits, we asked how such transient expression influences receptor stability following target cell recognition.
Materials and Methods
Written informed consent was obtained from all healthy voluntary donors involved in the study (commissioned by the local ethic committee of the Friedrich-Alexander Universität Erlangen-Nürnberg, Ethics proposal 2247, 10.07.2000). All data were analyzed anonymously.
Primary human mononuclear cells (MNCs) were obtained from blood of healthy voluntary donors (commissioned by the local ethic committees) by density gradient centrifugation using endotoxin-free Ficoll-Paque-PLUS® (Amersham Pharmacia Biotech AB, Uppsala, Sweden). CD8pos and CD4pos T cells were isolated from MNCs by negative magnetic selection using microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany). T cells were cultured in R10-IL2 medium consisting of RPMI 1640 supplemented with 10% FBS and 100 U/ml IL-2 (Millipore, Billerica, MA). 293T cells and the myeloid cell lines MONO-MAC-6, U-937, HL-60, NB-4, OCI-AML5, MV4-11, KG-1a, and THP-1 were obtained from the German Resource Center for Biologic Material (DSMZ, Braunschweig, Germany), the glioblastoma cell line U-373 was obtained from ATCC. The ESFT cell lines STA-ET-1, -2, -3, -6, -8.2, -10, -11, ER-ESFT-1 were generated in house , , CADO-ES1, TC-71, WE-68, VH-64 were provided by Frans van Valen . All ESFT cell lines were maintained in culture dishes, which were pre-coated for 2 hours at RT with a 0.1% gelatin solution in PBS. Murine cell lines stably, expressing either human ULBP2 or MICA, were generated by transduction of Ba/F3 pre-B cells with a pMX-pie based retroviral expression vector . All cell lines were maintained in RPMI 1640 supplemented with 10% FBS and IFN-γ when indicated.
Generation of chNKG2D
The extracellular portion of human NKG2D (aa82–216) was amplified by PCR from a plasmid containing the NKG2D cDNA sequence by using the primers 5′-cgaatagtccatggaagcttagccaccatgccgctgctgctactgctgcccctgctgtgggcaggggccctggctatgggtacctctagattattcaaccaagaagttcaaattccc-3′ containing a Kozak-sequence (bold), CD33 signal peptide sequence (underlined), restriction sites (italic) of NcoI and HindIII downstream; 5′-gatgtcacggatccccgccaccgcccgatccaccacctcctgaca-cagtcctttgcatgcagatg-3′ with a BamHI-site (bold) and the Gly4Ser-Gly4Asp-linker (underlined). The PCR product was cloned by NcoI×BamHI restriction into the retroviral vector pBullet 607 containing the IgG1-Fc/CD28/CD3ζ backbone . For in vitro transcription, the NKG2D portion was recloned from the pB607/NKG2D vector into the pGEM4Z-CEA vector, also containing the IgG1-Fc/CD28/CD3ζ backbone, by NcoI×BamHI restriction. The NKG2D portion was additionally recloned from the pB607/NKG2D vector into a derivate of the vector pST1 (provided by U. Sahin) for transcription of “RNA #2”. The scFv-portion of a Cytomegalovirus glycoprotein gB (CMV-gB)-specific IgG1-Fc/CD28/CD3ζ-CAR-construct  was replaced by HindIII×BamHI restriction. After recloning, the polyA tail was found to be shortened from 120 to approximately 100 A nucleotides as estimated by Eco53KI restriction. This vector was linearized for in vitro transcription with LguI to produce a polyA-tail devoid of non-A nucleotides. For lentiviral expression NKG2D/IgG1-Fc/CD28/CD3ζ was excised from pB607/NKG2D by HindIII×BspEI restriction and cloned into the PmeI site of the lentiviral plasmid pWPI (D. Trono, Geneva).
In vitro transcription and RNA-electroporation
In vitro transcription and electroporation was performed as previously described . Briefly, in vitro transcription was performed with linearized pGEM4Z-NKG2D or pST1-NKG2D using the mMESSAGE-mMACHINE-T7 Ultra kit (Applied Biosystems/Ambion) followed by polyadenylation. The vector pST1 was developed by Holtkamp et al.  to allow in vitro transcription of a more stable mRNA and the kit used for in vitro transcription was optimized for a more efficient translation initiation by using the modified anti-reverse cap analog (ARCA, 7-methyl(3′-O-methyl)GpppG)m7G(5′)ppp(5′)G). RNA from a cognate CMV-gH-specific IgG1-Fc/CD28/CD3ζ-CAR (Goetz G., unpublished) served as control. Electroporation was performed with 10 µg RNA/100 µl Opti-MEM containing ≤6×106 CD8pos or CD4pos T cells either immediately after isolation or 12 days after activation with an anti-CD3-antibody (clone OKT3). Anti-CD3-activation was performed by plating 0.2×106 T cells/ml on wells pre-coated with 10 µg/ml of anti-CD3-antibody in R10-IL2 medium. After two days the cells were transferred to fresh wells without anti-CD3-antibody. Half of the medium was replaced twice a week. Electroporated T cells were further cultured in R10-IL2 medium and used for functional analysis one day after RNA transfection.
Production of lentivirus and transduction of T cells
293T cells were seeded in 75 cm2 flasks and transfected using Lipofectamine 2000 (Invitrogen Corporation, Bethesda, MD) according to manufacturer's instructions. The pWPI vector plasmid containing the chNKG2D was cotransfected together with psPAX2 (packaging) and pMD2.G (VSV-G, envelope) in a ratio of 4∶3∶1 (total 15 µg plasmid) with 12.5 µl Lipofectamine 2000 per flask. The supernatants were collected 48 hours after transfection by centrifugation (10 min, 2000 rpm) to remove cell debris, and were stored at −80°C until further usage. Prior to transduction primary human cells were activated over night by 2.4 µg/ml PHA in R10-IL2 medium. Transduction was performed by spinoculation of 0.5–1×106 cells with 0.5 ml of virus containing supernatant (minimum 0.5 µg p24; quantified by ELISA), supplemented with 5 µg/ml Polybrene (Sigma), at 1500×g for 4 hours at 33°C. After an additional incubation over night at 37°C the cells were washed twice and further maintained in R10-IL2 medium. For enrichment of chNKG2Dpos cells irradiated ULBP2 expressing murine Ba/F3 cells were added 1∶1 in weekly intervals. Control T cells transduced with a pWPI vector containing a cognate CMV-gB-specific CAR  were enriched by co-culture with irradiated murine IIA1.6 cells stably expressing CD64 (provided by Jan van de Winkel; University Medical Center Utrecht, The Netherlands). The immunophenotype of transduced T cells was controlled by flow cytometry before functional analysis (data not shown). Only pure CD3pos TCRα/βpos cultures not containing NK cells were used in functional assays seven to ten days after the last stimulation with ligandpos target cells.
Human cells were analyzed by anti-CD3 (clone S4.1), anti-CD4 (clone SK3), anti-CD8 (clone SK1), anti-CD56 (clone NCSM16.2), anti-pan TCR-alpha/beta (clone BMA031), anti-MHC-class-I (clone Tü149), and appropriately labeled IgG1 and IgG2b isotype controls. Expression of chNKG2D was detected either by anti-CD314 (clone 1D11) or via its human IgG1-Fc portion using biotinylated anti-human-Fc (clone JDC-10) and streptavidin-PE. NKG2D-Ls were analyzed with anti-MICA (clone AMO1), anti-MICB (clone BMO2), anti-ULBP1 (clone AUMO2), anti-ULBP2 (clone BUMO1), anti-ULBP3 (clone CUMO3) (all from Axxora Life Sciences Inc., San Diego, CA) and an APC-labeled goat-anti-mouse-IgG used as secondary antibody.
50 000 activated T cells were co-cultured with 100 000 target cells in 0.5 ml of R10-IL2 medium supplemented with 0.3% Sandoglobulin® (CSL Behring, King of Prussia, PA). Co-culture with ESFT cells was performed in gelatin coated wells. Supernatants were harvested after 20 hours and TNF, FasL, and TRAIL were determined by using ELISA kits from Mabtech, Diaclone, and BD Biosciences, respectively, according to the manufacturer's instructions. The figures show total cytokine levels without background subtraction.
The cytoxic activity was analyzed either by determining target cell lysis using a non-radioactive Europium (Eu)-based assay or, if Eu-labeling of target cells was inefficient, by determining apoptosis induction in target cells by flow cytometry. Europium assay: ESFT cells were harvested by trypsin/EDTA treatment and labeled with Eu as previously described . Eu released from 1000 labeled target cells was determined in the supernatants after 4 hours of co-culture. Specific lysis was calculated by the formula: % lysis = (experimental counts minus spontaneous release)×100/(maximum release minus spontaneous release). Annexin V apoptosis assay: Target cells were harvested, washed in PBS and labeled by incubation with 2 µM SNARF-1 carboxylic acid, acetate succimidyl ester (Invitrogen) in PBS for 15 minutes at 37°C. Labeled cells (5000) were preincubated with 1% Sandoglobulin® for 20 minutes and then co-cultured with effector T cells at the indicated ratios in 200 µl R10-IL2 medium in a 96 well round bottom plate. After 5 hours, the cells were harvested by trypsin/EDTA and then supplemented with 1% FBS, 2 mM CaCl2, and Annexin V-FITC. The fraction of Annexin Vpos cells of SNARF-labeled target cells was determined by flow cytometry (measured in FL4). The percentage of spontaneous Annexin Vpos target cells (in absence of T cells) was subtracted as unspecific background and the fraction of “specific” Annexin Vpos targets actually induced by T cells was calculated by the formula: % Annexin Vpos = (Annxin Vpos of co-culture minus spontaneous Annexin Vpos)×100/(100 minus spontaneous Annexin Vpos). The fraction of spontaneous Annexin Vpos and maximum possible Annexin Vpos (100%) cells were corrected by the fraction of actually ligand expressing cells of the stably transfected cell line Ba/F3 (70–90%).
Expression of NKG2D-ligands in ESFT
The expression of different NKG2D-Ls has been reported previously for several ESFT cell lines . Here, we screened 14 ESFT cell lines for the expression of MICA, MICB, and ULBP1, ULBP2, and ULBP3 by flow cytometry. The results emphasize that NKG2D-Ls appear to be broadly expressed by ESFT cell lines, although some (STA-ET-1, -6, -8.2, -10, -11) displayed only fairly low ligand expression (Fig. 1). Most cell lines expressed MICA and frequently also ULBP1 and/or ULBP2. MICB was only weakly expressed in some cell lines and ULBP3 was consistently absent. From the data we concluded that NKG2D-Ls might allow targeting of a broad range of ESFT cells.
The histograms show the surface expression of the respective NKG2D-Ls recognized by specific monoclonal antibodies and analyzed by flow cytometry (filled histograms = specific antibodies; open histograms = isotype controls). The two stably transfected murine B cell lines (Ba/F3) expressing either human MICA or ULBP2 are displayed for comparison.
Construction of the chNKG2D
Existing signaling platforms of CARs are derived from type I transmembrane proteins with extra-cellular N- and intra-cellular C-terminus. NKG2D, contrary to most transmembrane proteins, however, is a type II receptor with inverse orientation. Fusion of NKG2D with the existing signaling platforms therefore could alter their configuration and function. To avoid this and to allow future improvements of signaling we tested the option of constructing a chNKG2D with type I orientation. The ectodomain of wildtype NKG2D forms non-covalent homo-dimers and is linked to the transmembrane domain via a 21 amino acid residues (aa) long highly flexible anchor (aa75–95). Analyzing the published crystal structure (Protein Data Bank; accession code 1HYR) revealed that the C-terminal end of NKG2D protrudes from the globular ectodomain in close proximity to the entry site of the membrane anchor (Fig. 2A) . Thus, we hypothesized that the NKG2D ectodomain could be alternatively anchored via a C-terminal linkage without disturbing its relative positioning. This should allow its integration into a CAR composed of a well established IgG1-Fc/CD28/CD3ζ signaling platform with type I orientation . We chose to truncate the NKG2D ectodomain within its native membrane anchor at aa81 thereby removing the transmembrane and intracellular portion of NKG2D and to fuse its C-terminal end (aa216) via an artificial flexible Gly/Ser-linker to the N-terminus of the IgG1-Fc/CD28/CD3ζ molecule. The flexible upper hinge residues (DKTHT) followed by the hinge core residues (CPPC) containing the inter-chain disulfide bonds of IgG1-Fc served as N-terminal acceptor site of the signaling platform . We calculated that a linker consisting of two Gly/Ser-repeats (GGGGSGGGGD; last serine replaced by hydrophilic aspartic acid) would by appropriate to allow sufficient flexibility for homo-dimeric pairing of the extra-cellular NGK2D portions but not to provoke sensitivity to proteolytic degradation. Wildtype NKG2D after translation into the endoplasmatic reticulum is directed to the cell surface via an uncleaved type II transmembrane signal anchor. This anchor was replaced in the chimeric receptor by the type I transmembrane signal anchor of CD28. Furthermore, as type I membrane proteins require also cleavable N-terminal signals for appropriate insertion into the cell membrane, we introduced the cleavable N-terminal signal peptide of CD33, which is removed in the mature protein. Figure 2A displays a scheme of the final mature fusion protein.
A, The figure schematically illustrates the domain architecture of wildtype NKG2D and chNKG2D (N = N-terminus, C = C-terminus, CM = cytoplasma membrane). The N-terminal truncation of the extracellular NKG2D domain at aa81 is indicated by a bold bar. The native C-terminal end of NKG2D was fused to the N-terminus of the IgG1-Fc/CD28/CD3ζ transmembrane signaling platform creating a transmembrane receptor of type I orientation in which the relative orientation and positioning of the NKG2D ectodomain has not been changed. B, efficient transfection and strong surface expression of chNKG2D by electroporation of mRNA into primary T cells. The histograms show the surface expression of chNKG2D 24 hours after electroporation of quiescent or anti-CD3-activated human T cells (thick line, open histogram). Analysis was performed by flow cytometry using an anti-CD314 antibody. The expression of endogenous wildtype NKG2D in T transfected with a CMV-gH-specific CAR is shown for comparison (thin line, open histograms; filled histograms = isotype controls). C, chNKG2D triggers cytotoxicity of activated T cells. Quiescent or anti-CD3-activated CD8pos or CD4pos T cells were transfected with RNA encoding either the chNKG2D or a CMV-gH-specific control receptor. One day after transfection, the T cells were co-cultured with dye-labeled murine B cells (Ba/F3-MICA) stably expressing the human NKG2D-L MICA. The diagram shows the percentage of Annexin Vpos murine target cells after 5 hours of co-culture at the indicated E∶T ratios (data pooled from three independent experiments using different cell donors; mean ± S.D.). Statistical significance was calculated for the effects of activated T cells expressing the chNKG2D versus T cells expressing the control receptor.
Transfer of chNKG2D into different T cell populations by electroporation of mRNA
T-cell activation might be required for triggering effector functions also in the CD45ROpos T cell subpopulation , however, induced proliferation might accelerate the loss of chNKG2D expression due to dilution of transfected mRNA. Thus, we compared resting and activated T cells. CD8pos and CD4pos T cells were purified from peripheral blood and electroporated either directly or 12 days after activation via CD3, when proliferation had declined again. Figure 2B shows the expression of chNKG2D one day after transfection. Detection was done by an anti-NKG2D(CD314)-antibody, which also detects endogenous NKG2D expressed in CD8pos T cells. T cells transfected with a CMV-gH-specific control CAR  were thus included in Figure 2B for comparison. ChNKG2D-mRNA-transfected T cells displayed strongly increased anti-NKG2D-staining compared to CMV-gH-CAR controls. Based on peak channel numbers from three independent experiments using cells from different donors, we calculated a 5–9-fold increased chNKG2D expression density in chNKG2D-RNA-transfected activated CD8pos T cells compared to endogenous NKG2D (Fig. 2B and data not shown). The expression of chNKG2D was homogeneous and augmented following cell activation both in CD4pos and CD8pos T cells compared to the resting cells.
NKG2D-Ls are monomeric and are recognized only by the homo-dimerized native NKG2D receptor by asymmetric binding of the α1- and the α2-epitopes of the ligands . Here, we evaluated the functional activity of the novel chNGK2D type. Figure 2C shows that chNKG2Dpos T cells can induce apoptosis in a murine cell line stably expressing the NKG2D-ligand MICA (Fig. 1). This experiment also illustrates that resting chNKG2Dpos T cells were clearly inferior to CD3-activated chNKG2Dpos T cells, which killed their target cells even at an effector∶target (E∶T) ratio of only 1∶1. Importantly, apoptosis in MICApos targets was induced both by chNKG2D-expressing CD8pos as well as CD4pos T cells. The data so far confirmed that the novel recombinant chNKG2D is functional and that RNA transfection allows high level expression on the surface of primary T cells.
ESFT cells are killed by T cells transfected with chNKG2D mRNA
In the following experiments we tested the capacity of RNA transfected activated T cells to recognize ESFT cells with varying and partially low ligand density (Fig. 1). Figure 3A illustrates the lysis of ESFT cells by T cells expressing either chNGK2D or an irrelevant CMV-specific control receptor. ESFT cells were specifically lysed by chNKG2Dpos T cells. Notably, chNKG2Dpos CD4pos T cells lysed over 20% of ESFT cells at an E∶T ratio of only 4∶1 and even targets with low-level ligand density such as STA-ET-8.2 and -11. Figures 3B and 3C demonstrate that chNKG2D also mediated specific cytokine release in mRNA transfected CD4pos T cells, as shown by detection of TNF and soluble FasL. Together, these results confirm that chNKG2D-RNA transfected T cells specifically recognize ESFT cells even with low ligand density. Furthermore, the data suggest that also the CD4pos subset of T cells could mediate an efficient anti-ESFT response.
A, Lysis of ESFT cells. The diagram shows the lytic activity of anti-CD3-activated CD8pos and CD4pos T cells transfected with mRNA encoding either chNKG2D or a similar chimeric CMV-gH-specific control receptor (αCMV-gH-CAR) in response to ESFT cells with different ligand expression (stated in brackets). Displayed is the specific lysis of the indicated ESFT cell lines at the indicated E∶T ratio (measured by Eu-release; mean ± S.D. of triplicates). Statistical significance was calculated for the effects of T cells expressing the chNKG2D versus T cells expressing the control receptor where indicated. B and C, Cytokine secretion of chNKG2D-mRNA transfected CD4pos T cells in response to ESFT cells. The diagrams show TNF and FasL secretion by CD4pos T cells transfected with RNA encoding chNKG2D or a CMV-gB-specific control receptor in response to the indicated cell lines or only medium as control (measured by ELISA; mean of duplicates).
Lentiviral expression of chNKG2D
For lentiviral expression we cloned chNKG2D into the bicistronic vector pWPI, which was previously optimized for high-level transgene expression in hematopoietic cells . Transduction of PHA-activated T cells yielded expression of chNKG2D in different cell types at similar proportions (16.1% of chNKG2Dpos CD8neg cells and 19.3% of chNKG2Dpos CD8pos cells in a representative experiment, data not shown). After removal of PHA but in the continued presence of IL-2 chNKG2Dpos cells enriched during prolonged culture (data not shown), suggesting an auto-stimulatory effect possibly through endogenous expression of NKG2D-Ls induced by PHA stimulation . Accelerated enrichment and expansion was achieved by restimulation with an ULBP2pos murine Ba/F3 cell line (Fig. 4A). Repeated restimulation in IL-2 supplemented medium resulted in highly enriched (>80%) chNKG2Dpos CD8pos and CD4pos T cells that could be expanded for at least three months. Expression of the receptor was stable during the whole period of restimulation, and waned only gradually over a time period of few weeks after cessation of periodical restimulation (data not shown). Figure 4B shows the specific TNF production of enriched chNKG2Dpos CD4pos T cells in response to different NKG2D-Lpos tumor cells. Finally, we directly compared the lytic activity of CD8pos T cells modified with chNKG2D either by lentiviral gene transfer or by mRNA electroporation. As shown in Figure 4C, both transfer methods conferred lytic capacity to a comparable extent against ULBP2 expressing target cells but not against ligand negative control cells. Not surprisingly, we observed also weak lysis of ULBP2pos target cells, but not of ULBP2neg control cells, when co-cultured with receptor unmodified CD8pos effector cells, which we ascribe to endogenous expression of (costimulatory) CD314 (NKG2D).
A, antigen-specific enrichment of lentivirally transduced chNKG2Dpos MNCs. On day 13 after transduction the cells were co-cultured either with Ba/F3, ULBP2pos Ba/F3 or without target cells. FACS analysis was performed 6 days later (open histogram, bold line: co-culture with the ULBP2pos Ba/F3, 86.7% chNKG2Dpos T cells; filled histogram, dotted line: co-culture with Ba/F3, 60.7% chNKG2Dpos T cells; thin line: cells cultured without target cells, 54.1% chNKG2Dpos T cells; open histogram, dotted line = secondary antibody only). B, cytokine production. The diagram shows TNF secretion of purified CD4pos T cells lentivirally transduced with either chNKG2D (enriched to 83–92% positive cells) or a similar chimeric CMV-specific control receptor (enriched to 69–81%) in response to the indicated cell lines or only medium as control (measured by ELISA; mean of duplicates). Ba/F3 or ULBP2pos Ba/F3 cells were used for negative and positive control. Statistical significance was calculated for the effects of T cells expressing the chNKG2D versus T cells expressing the control receptor. C, Comparison of the cytotoxic response triggered by chNKG2D either stably or transiently transferred. CD8pos T cells either unmodified (“without chNKG2D”) or expressing chNKG2D by lentiviral transduction (“LV”; enriched for chNKG2Dpos >90%) or mRNA transfection (“mRNA”; 20 hours after electroporation) were co-cultured with dye-labeled murine Ba/F3 cells expressing human ULBP2 (ULBP2-Ba/F3) or not (CTRL-Ba/F3). The diagram shows the percentage of Annexin Vpos Ba/F3 target cells after 4 hours of co-culture at the indicated E∶T ratios (mean ± S.D. of triplicates).
Together the data demonstrate that lentiviral transduction allowed ligand induced long-term proliferation and the generation of highly enriched chNKG2Dpos T cells, which were characterized by a strong NKG2D-L specific responsiveness comparable to that observed with mRNA modified T cells.
Stability of chNKG2D surface expression upon target cell recognition
Our data so far confirmed that both mRNA transfection and lentiviral transduction mediated high level expression of the chNKG2D in T cells. However, ligand-induced receptor internalization, which is known for T cell receptors (TCRs) and has been reported also for CARs, could result in a significant loss of receptor molecules from the cell surface upon target cell recognition . For this purpose, we compared the stability of chNKG2D surface expression obtained by either RNA electroporation or by lentiviral transduction, after co-culture with MICApos/ULBP2pos STA-ET-3 target cells (Fig. 5A). Co-culture of the RNA-modified cells one day after electroporation with a 2.7 fold excess of STA-ET-3 target cells for 24 hours led to a reduction of the surface expression of chNKG2D by more than 90% compared to the expression observed in T cells without co-culture, as estimated by peak channel numbers (without co-culture: 1762; with co-culture: 87). This was in contrast to lentivirally modified T cells, in which chNKG2D expression even slightly increased during co-culture. As the extent of CAR internalization is strongly determined by the target antigen density , , we investigated the effect of interaction with ESFT cell lines characterized by different levels of NKG2D-L expression. Figure 5B illustrates that chNKG2D down-modulation in mRNA transfected cells indeed correlated with ligand density on ESFT cells.
The histograms show the flow cytometric analysis of chNKG2D expression (detected in CD4pos T cells by an anti-CD314 antibody or in CD8pos T cells by an anti-human-IgG1-Fc antibody; log scale in all histograms; note: amplification voltage in flow cytometric analysis was lowered for lentivirally transduced T cells due to the strong expression of the receptor). A, comparison of lentivirally transduced and mRNA transfected T cells. 75,000 chNKG2Dpos CD4pos T cells were cultured 24 hours after electroporation for further 24 hours in presence or absence of 200,000 STA-ET-3 cells (filled grey histograms = presence of STA-ET-3, open histograms bold line = absence of STA-ET-3, open histograms dotted line = T cells not transfected). Shown is one representative experiment out of three independent repetitions. B, chNKG2D down-regulation in response to different ESFT cell lines. 100,000 chNKG2Dpos CD8pos T cells were co-cultured 50 hours after mRNA electroporation for further 24 hours with 100,000 cells of different ESFT cell lines as indicated (unmodified T cells = “without chNKG2D”; without co-culture = “without target cells”). C, kinetics of chNKG2D down-regulation upon target-cell encounter at different times after transfection. 20,000 chNKG2Dpos CD4pos T cells were co-cultured with 60,000 STA-ET-3 target cells for different times (1, 4, or 48 hours) before chNKG2D expression was determined (6, 30, or 74 hours after transfection) (one experiment). D, comparison of differentially stabilized mRNA species. ChNKG2Dpos CD4pos T cells were co-cultured with STA-ET-3 target cells at different effector∶target ratios and at different times after transfection. Target cells were added at the indicated ratios either two or 26 hours after transfection and chNKG2D expression was determined by flow cytometry 6, 30, 50, or 74 hours after transfection. Shown is one out of three experiments. E, effect of patient serum containing soluble NKG2D-Ls. 15,000 chNKG2Dpos CD4pos T cells were cultured 24 hours after electroporation for further 24 hours with or without serum of two different neuroblastoma patients (final serum concentration 33%).
In the following experiments we investigated the influence of the ESFT cell line STA-ET-3, most effective in CAR down-modulation with mRNA transfected T cells, in more detail. Figure 5C illustrates, that the capacity of mRNA-electroporated T cells to compensate for ligation induced chNKG2D down-regulation strongly decreased already early after transfection. Target-cell encounter three days after transfection almost completely down-modulated chNKG2D from the cell surface within only 4 hours of contact at an E∶T ratio of 1∶3. Finally, we compared two differentially engineered RNA species encoding chNKG2D for their capacity to compensate receptor down-modulation by lower number but prolonged presence of target cells (Fig. 5D). RNA #1 refers to the RNA used in all experiments described above and RNA #2 was modified according to Holtkamp et al. with RNA stabilizing elements . The figure illustrates the comparably slow decline of chNKG2D expression over a time period of 6, 30, 50, and 74 hours after mRNA transfection, when T cells were cultured without target cells (open histograms, bold line). A markedly accelerated decline occurred, when target cells were added 2 or 26 hours after transfection. The receptor was strongly down-regulated at time point 30 hours after transfection, when tumor cells were added 2 hours after transfection at an E∶T ratio of 1∶1 (filled histograms). Addition of tumor cells at later time points (26 hours after transfection) resulted in more pronounced chNKG2D down-regulation at time point 50 hours after transfection. Then, even lower amounts of tumor cells (E∶T 1∶0.33; open histograms, thin line) resulted in marked down-regulation, which after an additional day of co-culture (FACS analysis 74 hours after transfection) was very similar to cultures with an E∶T of 1∶1. Importantly, almost identical effects were observed with both types of mRNA, although RNA #2 slightly mitigated the ligation induced down-modulation (visible at E∶T 1∶0.33). Together, these experiments had shown that the extent of down-modulation correlates with target cell numbers and increases with time following transfection.
It has been reported that endogenously expressed NKG2D is down-regulated by cell-bound ligands but not by soluble ligands , . In view of the observed reduction of chNKG2D cell surface density in mRNA-transfected T cells, we wanted to investigate this issue with sera from tumor patients, which are known to frequently contain high levels of a soluble fraction of different NKG2D-Ls and thereby could inactivate injected T cells. Thus, we co-cultured RNA-modified CD4pos T cells with two sera of neuroblastoma patients containing high detectable levels of MICA (171 and 494 pg/ml; “NB1” and “NB2” respectively). In these experiments, however, we did not observe any reduction of chNKG2D surface density (Fig. 5E).
The goal of our study was to evaluate the potential of a T cell based approach for adoptive therapy of EFST using a chimeric NKG2D receptor and to compare the suitability of two different methods for receptor transfer. For this aim we investigated NKG2D-Ls, which have previously been reported to be expressed in ESFT , as potential ESFT target antigens. In confirmation and extension of Verhoeven et al.'s data, we found NKG2D-Ls consistently expressed in all of 14 tested ESFT cell lines (Fig. 1). Based on this and on the fact that our chNKG2D modified T cells efficiently recognized ESFT cells also with low NKG2D-L expression, we propose a chNKG2D based approach for immunotherapy of ESFT lesions.
Originally, the feasibility of redirecting T cells to NKG2D-Ls has been demonstrated by Sentman et al. using NKG2D fused as a full-length protein to the cytoplasmic domain of CD3ζ –. For our study, we intended to construct a NKG2D-based CAR with an integrated costimulatory CD28 domain for enhanced T cell activation , . Notably, CD28 signaling enhances the resistance to TGF-β1 and regulatory T cells (Tregs) and thus could be particularly relevant in metastasized ESFT, for which secretion of TGF-β1 and increased numbers of infiltrating Tregs have been described , , . Further, signal amplification by CD28 reduces the threshold of endocytosed receptor molecules required for T cell activation , and, hence may be favorable for targeting ESFT, which express NKG2D-Ls often at rather low levels (Fig. 1). Low target antigen density may impair the activation of CAR modified T cells, as CARs require about 100 fold higher target antigen densities than T cell receptors to activate a T cell, which is explained by the lack of serial triggering due to high affinity interaction , , , . Specifically, using a first generation CAR, James et al. have shown that it requires ∼30,000 target molecules/target cell (inducing endocytosis of ∼20,000 CAR molecules) to trigger maximum lytic activity , . Increased ligand densities may be required if receptor expression is low .
In line with these theoretical considerations we indeed found that CD8pos T cells redirected by our novel type of chNKG2D at high expression levels could efficiently lyse ESFT cells even when their ligand expression was low (e.g. STA-ET-11; Fig. 3A). This might be important and appears to be in line with the reported NK cell lysis of ESFT cells, for which also no relation between susceptibility to lysis and the expression levels of the triggering ligands has been reported .
Activated NK cells have been shown to kill ESFT cells with high efficiency , . By using a flow cytometry-based method, Cho et al. have observed up to 87% propidium iodide positive cells within 4 hours of co-culture at an E∶T ratio of 1∶1. In our study we relied on a cytotoxicity-assay based on the release of non-radioactive Europium (Fig. 3A) or flow-cytometry using Annexin V staining (Figs. 2C and 4C). In general, the Annexin V based method seems to give higher estimates of cytotoxicity (data not shown), probably due to the fact that Annexin staining detects early steps of apoptosis. Taken into account the different methods used by this group and by us, our results with CAR modified T cells appear to be within the same magnitude of killing efficiency compared to the NK cells tested by Cho et al.
The cytotoxicity results shown in Figures 2C and 3A underline that CD4pos T cells in addition to the CD8pos subset are attractive candidates for expression of the chNKG2D. The observed strong cytotoxicity of CD4pos T cells might be explained by several potentially cytotoxic proteins. We found that chNKG2D modified CD4pos T cells in response to ESFT cells did not express detectable amounts of TRAIL or perforin (data not shown). However, we found soluble FasL in the supernatants of these cells (Fig. 3C), which may contribute to killing since ESFT cells are sensitive to Fas cross-linking (data not shown and –).
Apart from chNKG2D-dependent cytotoxicity, we repeatedly observed some chNKG2D-independent background killing by particularly CD8pos T cells (unmodified or expressing an irrelevant control-CAR; Figs. 2C, 3A, and 4C) which we could not attribute to eventually contaminating NK cells (purity of the cell population after cell sorting and cell expansion: <1% NK cells; data not shown). Such CAR-independent cytotoxicity, however, may likely be due to endogenously expressed NKG2D and other activating receptors, particularly DNAM-1 (CD226), which has been described to mediate strong TCR-independent lysis of many tumor cell lines , . Its ligands, CD155 and CD112, are also frequently up-regulated in tumors including ESFT. Together with NKG2D-Ls the ligands of DNAM-1 efficiently trigger NK cell cytotoxicity against ESFT cells , ,  and likely also mediate some CAR-independent killing of ESFT by activated CD8pos and CD4pos T cells. With regard to the background FasL production observed also in cultures containing control-CAR transfected T cells (Fig. 3C), we cannot exclude a contribution from the ESFT cells themselves . However, when cultured without T cells, our ESFT cell lines produced less than 60 pg/ml sFasL (data not shown).
For redirecting T cells we used RNA transfer and lentiviral transduction which both allowed high level expression of chNKG2D and comparable elimination of target cells. Regarding the persistence of the expression, it is known that transfected mRNA maintains substantial translation only within the first day after electroporation , but that the translated protein persists at high density on the cell surface for several days. According to our data, the persistence on the cell surface actually might be much more limited due to rapid ligand induced receptor internalization upon target cell recognition. Ligand induced receptor internalization is important for proper signal transduction , , and has been quantified for the T cell receptor complex and also for CARs , . CARs are internalized by ligands roughly at a ratio of 1∶1 whereas 100 fold more TCR molecules per target antigen are internalized due to serial triggering.
Cross-linking of transmembrane receptors generally results in rapid redistribution from the cell surface into intracellular compartments, where a fraction of the receptors enters lysosomal degradation instead of being recycled back to the cell surface , . As a consequence, a high density of ligand molecules on target cells can rapidly reduce the cell surface density of retrovirally expressed CAR molecules below the critical threshold required for T cell activation resulting in the loss of serial killing capacity .
Here we show that the strong lentiviral expression of our chNKG2D could effectively compensate ligand induced internalization in response to STA-ET-3 cells, expressing several different NKG2D-Ls (Figs. 1 and 5A), whereas mRNA transfection was similarly effective only during the first hours after electroporation (Fig. 5D). Target cell contact at later time points, however, resulted in strong and irreversible down-modulation of chNKG2D (Figs. 5C and 5D). Notably, this was seen with mRNA, for which we had applied state of the art techniques to maximize stability and translation, i.e. extensive polyadenylation and capping with a specially modified cap analog (see materials and methods) , . Further attempts to improve mRNA stability according to Holtkamp et al. yielded only negligible effects (“RNA #2” in Fig. 5D) , .
Such characteristics of mRNA mediated CAR expression could be significant, since clinical studies have shown, that intravenously administered T cells in humans due to extensive organ trapping and long migration distances reach tumor sites only two or three days after infusion , . Our observations would argue for injecting mRNA transfected T cells in humans not only repetitively but also perhaps intra-tumorally or intra-arterially to ensure that the T cells meet their targets without any further delay.
Regardless of the applied receptor transfer strategy, our data illustrate for the first time the potential of an NKG2D based approach for targeting ESFT cells by receptor modified T cells. The new receptor efficiently triggered the elimination of ESFT cells in vitro by CD8pos and CD4pos T cells and thereby opens a new avenue for developing an adoptive T cell approach that could have the potential to complement conventional multimodal therapy of ESFT.
We thank Prof. Didier Trono (Geneva, Switzerland) for providing the plasmids of the lentiviral vector system, Prof. Ugur Sahin (Mainz, Germany) for providing the pST1 plasmid, Dr. Frans van Valen (Münster, Germany) for providing ESFT cell lines, and Dr. Ulrike Köhl (Frankfurt, Germany) for providing the sera of neuroblastoma patients.
Conceived and designed the experiments: ML YAM AE OP WH. Performed the experiments: ML GG NS JD JP. Analyzed the data: ML FF WH. Contributed reagents/materials/analysis tools: YAM HA PFA HK AC. Wrote the paper: ML WH.
- 1. Leavey PJ, Mascarenhas L, Marina N, Chen Z, Krailo M, et al. (2008) Prognostic factors for patients with Ewing sarcoma (EWS) at first recurrence following multi-modality therapy: A report from the Children's Oncology Group. Pediatr Blood Cancer 51: 334–338.
- 2. Miser JS, Goldsby RE, Chen Z, Krailo MD, Tarbell NJ, et al. (2007) Treatment of metastatic Ewing sarcoma/primitive neuroectodermal tumor of bone: evaluation of increasing the dose intensity of chemotherapy–a report from the Children's Oncology Group. Pediatr Blood Cancer 49: 894–900.
- 3. Rodriguez-Galindo C, Liu T, Krasin MJ, Wu J, Billups CA, et al. (2007) Analysis of prognostic factors in ewing sarcoma family of tumors: review of St. Jude Children's Research Hospital studies. Cancer 110: 375–384.
- 4. Verhoeven DH, de Hooge AS, Mooiman EC, Santos SJ, ten Dam MM, et al. (2008) NK cells recognize and lyse Ewing sarcoma cells through NKG2D and DNAM-1 receptor dependent pathways. Mol Immunol 45: 3917–3925.
- 5. Cho D, Shook DR, Shimasaki N, Chang YH, Fujisaki H, et al. (2010) Cytotoxicity of activated natural killer cells against pediatric solid tumors. Clin Cancer Res 16: 3901–3909.
- 6. Rubnitz JE, Inaba H, Ribeiro RC, Pounds S, Rooney B, et al. (2010) NKAML: a pilot study to determine the safety and feasibility of haploidentical natural killer cell transplantation in childhood acute myeloid leukemia. J Clin Oncol 28: 955–959.
- 7. Kloess S, Huenecke S, Piechulek D, Esser R, Koch J, et al. (2010) IL-2-activated haploidentical NK cells restore NKG2D-mediated NK-cell cytotoxicity in neuroblastoma patients by scavenging of plasma MICA. Eur J Immunol 40: 3255–3267.
- 8. Sadelain M, Brentjens R, Riviere I (2009) The promise and potential pitfalls of chimeric antigen receptors. Curr Opin Immunol 21: 215–223.
- 9. Nausch N, Cerwenka A (2008) NKG2D ligands in tumor immunity. Oncogene 27: 5944–5958.
- 10. Eagle RA, Traherne JA, Hair JR, Jafferji I, Trowsdale J (2009) ULBP6/RAET1L is an additional human NKG2D ligand. Eur J Immunol 39: 3207–3216.
- 11. Zhang T, Lemoi BA, Sentman CL (2005) Chimeric NK-receptor-bearing T cells mediate antitumor immunotherapy. Blood 106: 1544–1551.
- 12. Zhang T, Barber A, Sentman CL (2006) Generation of antitumor responses by genetic modification of primary human T cells with a chimeric NKG2D receptor. Cancer Res 66: 5927–5933.
- 13. Barber A, Zhang T, DeMars LR, Conejo-Garcia J, Roby KF, et al. (2007) Chimeric NKG2D receptor-bearing T cells as immunotherapy for ovarian cancer. Cancer Res 67: 5003–5008.
- 14. Barber A, Zhang T, Megli CJ, Wu J, Meehan KR, et al. (2008) Chimeric NKG2D receptor-expressing T cells as an immunotherapy for multiple myeloma. Exp Hematol 36: 1318–1328.
- 15. Savoldo B, Ramos CA, Liu E, Mims MP, Keating MJ, et al. (2011) CD28 costimulation improves expansion and persistence of chimeric antigen receptor-modified T cells in lymphoma patients. J Clin Invest 121: 1822–1826.
- 16. McFarland BJ, Kortemme T, Yu SF, Baker D, Strong RK (2003) Symmetry recognizing asymmetry: analysis of the interactions between the C-type lectin-like immunoreceptor NKG2D and MHC class I-like ligands. Structure 11: 411–422.
- 17. Rosenberg SA, Dudley ME (2009) Adoptive cell therapy for the treatment of patients with metastatic melanoma. Curr Opin Immunol 21: 233–240.
- 18. Kovar H, Auinger A, Jug G, Aryee D, Zoubek A, et al. (1993) Narrow spectrum of infrequent p53 mutations and absence of MDM2 amplification in Ewing tumours. Oncogene 8: 2683–2690.
- 19. Kovar H, Jug G, Aryee DN, Zoubek A, Ambros P, et al. (1997) Among genes involved in the RB dependent cell cycle regulatory cascade, the p16 tumor suppressor gene is frequently lost in the Ewing family of tumors. Oncogene 15: 2225–2232.
- 20. Van Valen F (1998) Ewing's sarcoma family of tumors. In: Masters JRW, Palsson BO, editors. Human Cell Culture 1. Great Britain: Kluwer. pp. 55–85.
- 21. Textor S, Durst M, Jansen L, Accardi R, Tommasino M, et al. (2008) Activating NK cell receptor ligands are differentially expressed during progression to cervical cancer. Int J Cancer 123: 2343–2353.
- 22. Hombach A, Wieczarkowiecz A, Marquardt T, Heuser C, Usai L, et al. (2001) Tumor-specific T cell activation by recombinant immunoreceptors: CD3 zeta signaling and CD28 costimulation are simultaneously required for efficient IL-2 secretion and can be integrated into one combined CD28/CD3 zeta signaling receptor molecule. J Immunol 167: 6123–6131.
- 23. Full F, Lehner M, Thonn V, Goetz G, Scholz B, et al. (2010) T cells engineered with a cytomegalovirus-specific chimeric immunoreceptor. J Virol 84: 4083–4088.
- 24. Birkholz K, Hombach A, Krug C, Reuter S, Kershaw M, et al. (2009) Transfer of mRNA encoding recombinant immunoreceptors reprograms CD4+ and CD8+ T cells for use in the adoptive immunotherapy of cancer. Gene Ther 16: 596–604.
- 25. Holtkamp S, Kreiter S, Selmi A, Simon P, Koslowski M, et al. (2006) Modification of antigen-encoding RNA increases stability, translational efficacy, and T-cell stimulatory capacity of dendritic cells. Blood 108: 4009–4017.
- 26. Lehner M, Morhart P, Stilper A, Petermann D, Weller P, et al. (2007) Efficient chemokine-dependent migration and primary and secondary IL-12 secretion by human dendritic cells stimulated through Toll-like receptors. J Immunother 30: 312–322.
- 27. Saphire EO, Stanfield RL, Crispin MD, Parren PW, Rudd PM, et al. (2002) Contrasting IgG structures reveal extreme asymmetry and flexibility. J Mol Biol 319: 9–18.
- 28. Barber DL, Wherry EJ, Ahmed R (2003) Cutting edge: rapid in vivo killing by memory CD8 T cells. J Immunol 171: 27–31.
- 29. Salmon P, Kindler V, Ducrey O, Chapuis B, Zubler RH, et al. (2000) High-level transgene expression in human hematopoietic progenitors and differentiated blood lineages after transduction with improved lentiviral vectors. Blood 96: 3392–3398.
- 30. Molinero LL, Fuertes MB, Rabinovich GA, Fainboim L, Zwirner NW (2002) Activation-induced expression of MICA on T lymphocytes involves engagement of CD3 and CD28. J Leukoc Biol 71: 791–797.
- 31. James SE, Greenberg PD, Jensen MC, Lin Y, Wang J, et al. (2010) Mathematical modeling of chimeric TCR triggering predicts the magnitude of target lysis and its impairment by TCR downmodulation. J Immunol 184: 4284–4294.
- 32. James SE, Greenberg PD, Jensen MC, Lin Y, Wang J, et al. (2008) Antigen sensitivity of CD22-specific chimeric TCR is modulated by target epitope distance from the cell membrane. J Immunol 180: 7028–7038.
- 33. Salih HR, Goehlsdorf D, Steinle A (2006) Release of MICB molecules by tumor cells: mechanism and soluble MICB in sera of cancer patients. Hum Immunol 67: 188–195.
- 34. Waldhauer I, Steinle A (2006) Proteolytic release of soluble UL16-binding protein 2 from tumor cells. Cancer Res 66: 2520–2526.
- 35. Loskog A, Giandomenico V, Rossig C, Pule M, Dotti G, et al. (2006) Addition of the CD28 signaling domain to chimeric T-cell receptors enhances chimeric T-cell resistance to T regulatory cells. Leukemia 20: 1819–1828.
- 36. Brinkrolf P, Landmeier S, Altvater B, Chen C, Pscherer S, et al. (2009) A high proportion of bone marrow T cells with regulatory phenotype (CD4+CD25hiFoxP3+) in Ewing sarcoma patients is associated with metastatic disease. Int J Cancer 125: 879–886.
- 37. Viola A, Lanzavecchia A (1996) T cell activation determined by T cell receptor number and tunable thresholds. Science 273: 104–106.
- 38. Thomas S, Xue SA, Bangham C, Jakobsen BK, Morris E, et al. (2011) Human T cells expressing affinity matured TCR display accelerated responses but fail to recognise low density of MHC/peptide antigen. Blood. doi:10.1182/blood-2010-12-326736.
- 39. Cooper LJ, Kalos M, Lewinsohn DA, Riddell SR, Greenberg PD (2000) Transfer of specificity for human immunodeficiency virus type 1 into primary human T lymphocytes by introduction of T-cell receptor genes. J Virol 74: 8207–8212.
- 40. Van Valen F, Fulda S, Truckenbrod B, Eckervogt V, Sonnemann J, et al. (2000) Apoptotic responsiveness of the Ewing's sarcoma family of tumours to tumour necrosis factor-related apoptosis-inducing ligand (TRAIL). Int J Cancer 88: 252–259.
- 41. Sun Q, Burton RL, Lucas KG (2002) Cytokine production and cytolytic mechanism of CD4(+) cytotoxic T lymphocytes in ex vivo expanded therapeutic Epstein-Barr virus-specific T-cell cultures. Blood 99: 3302–3309.
- 42. de Hooge AS, Berghuis D, Santos SJ, Mooiman E, Romeo S, et al. (2007) Expression of cellular FLICE inhibitory protein, caspase-8, and protease inhibitor-9 in Ewing sarcoma and implications for susceptibility to cytotoxic pathways. Clin Cancer Res 13: 206–214.
- 43. Verneris MR, Karami M, Baker J, Jayaswal A, Negrin RS (2004) Role of NKG2D signaling in the cytotoxicity of activated and expanded CD8+ T cells. Blood 103: 3065–3072.
- 44. Shibuya A, Campbell D, Hannum C, Yssel H, Franz-Bacon K, et al. (1996) DNAM-1, a novel adhesion molecule involved in the cytolytic function of T lymphocytes. Immunity 4: 573–581.
- 45. Lakshmikanth T, Burke S, Ali TH, Kimpfler S, Ursini F, et al. (2009) NCRs and DNAM-1 mediate NK cell recognition and lysis of human and mouse melanoma cell lines in vitro and in vivo. J Clin Invest 119: 1251–1263.
- 46. Mitsiades N, Poulaki V, Kotoula V, Leone A, Tsokos M (1998) Fas ligand is present in tumors of the Ewing's sarcoma family and is cleaved into a soluble form by a metalloproteinase. Am J Pathol 153: 1947–1956.
- 47. Sorkin A, Von Zastrow M (2002) Signal transduction and endocytosis: close encounters of many kinds. Nat Rev Mol Cell Biol 3: 600–614.
- 48. Murphy JE, Padilla BE, Hasdemir B, Cottrell GS, Bunnett NW (2009) Endosomes: a legitimate platform for the signaling train. Proc Natl Acad Sci U S A 106: 17615–17622.
- 49. Zhao Y, Moon E, Carpenito C, Paulos CM, Liu X, et al. (2010) Multiple injections of electroporated autologous T cells expressing a chimeric antigen receptor mediate regression of human disseminated tumor. Cancer Res 70: 9053–9061.
- 50. Meidenbauer N, Marienhagen J, Laumer M, Vogl S, Heymann J, et al. (2003) Survival and tumor localization of adoptively transferred Melan-A-specific T cells in melanoma patients. J Immunol 170: 2161–2169.
- 51. Mitchell MS, Darrah D, Yeung D, Halpern S, Wallace A, et al. (2002) Phase I trial of adoptive immunotherapy with cytolytic T lymphocytes immunized against a tyrosinase epitope. J Clin Oncol 20: 1075–1086.