NKG2D-Based CAR T Cells and Radiotherapy Exert ......an MHC-independent manner (7). CAR T-cell therapy has led to encouraging clinical responses in hematologic malignancies (8, 9)

Translational Science

NKG2D-Based CAR T Cells and RadiotherapyExert Synergistic Efficacy in GlioblastomaTobias Weiss1, Michael Weller1, Matthias Guckenberger2,Charles L. Sentman3, and Patrick Roth1

Abstract

Chimeric antigen receptor (CAR) T-cell therapy is an emergingimmunotherapy against several malignancies including glioblas-toma, the most common and most aggressive malignant primarybrain tumor in adults. The challenges in solid tumor immuno-therapy comprise heterogenously expressed tumor target antigensand restricted trafficking of CAR T cells to and impaired long-termpersistence at the tumor site, aswell as theunaddressed integrationof CAR T-cell therapy into conventional anticancer treatments.Weaddressed these questions using aNKG2D-based chimeric antigenreceptor construct (chNKG2D) in fully immunocompetent ortho-topic glioblastomamousemodels. ChNKG2D T cells demonstrat-ed high IFNg production and cytolytic activity in vitro. Uponsystemic administration in vivo, chNKG2D T cells migrated to thetumor site in the brain, did not induce adverse events, prolongedsurvival, and cured a fraction of glioma-bearing mice. Survivingmice were protected long-term against tumor rechallenge. Mech-

anistically, this was not solely the result of a classical immunememory response, but rather involved local persistence ofchNKG2D T cells. A subtherapeutic dose of local radiotherapy incombination with chNKG2D T-cell treatment resulted in syner-gistic activity in two independent syngeneicmouse gliomamodelsby promoting migration of CAR T cells to the tumor site andincreased effector functions.We thus provide preclinical proof-of-concept of NKG2D CAR T-cell activity in mouse glioma modelsand demonstrate efficacy, long-term persistence, and synergisticactivity in combination with radiotherapy, providing a rationaleto translate this immunotherapeutic strategy to human gliomapatients.

Significance: These findings provide evidence for synergy ofconventional anticancer therapy andCAR T cells and heralds futurestudies for other treatment combinations. Cancer Res; 78(4); 1031–43.�2017 AACR.

IntroductionGlioblastoma is the most common malignant primary brain

tumor in adults (1). It remains one of the most challengingcancers and has still a poor prognosis despite multimodal treat-ment regimens comprising surgery, radiotherapy, and chemo-therapy with temozolomide (2, 3). Therefore, novel treatmentmodalities are urgently needed. Adoptive immunotherapy withgenetically engineered T cells that express a chimeric antigenreceptor (CAR) is an emerging treatment strategy that may alsohold promise for neoplasms in the central nervous system (CNS;ref. 4).

The design of CARs, which consist of an extracellular tumorantigen–binding domain linked to hinge, transmembrane, andintracellular signaling domains (5, 6), allows customized T-cellengineering and an efficient antitumor response of bulk T cells inan MHC-independent manner (7). CAR T-cell therapy has led to

encouraging clinical responses in hematologic malignancies(8, 9) and is also studied in solid tumors including glioblastoma(10). However, in solid tumors, there are several challenges thathamper the efficacy of CAR T-cell therapy, which need to beaddressed, such as the identification of homogeneously expressedtumor-associated or tumor-specific target antigens, the migrationof CAR T cells to the tumor site, and the immunosuppressivemicroenvironment that may impede the function and persistenceof CAR T cells (11).

CAR T-cell strategies that are currently explored against glio-blastoma target single tumor antigens such as EGFR variant III(EGFRvIII; ref. 12), erythropoietin-producing hepatocellular car-cinoma A2 (EphA2; ref. 13), Her2 (14, 15), or IL13 receptorsubunit alpha 2 (IL13Ra2; ref. 16). These targets are nonhomo-geneously expressed and susceptible to antigen escape (17).

Weandothers have assessed the importance of thenatural killergroup 2-member D (NKG2D) system in glioblastoma (18–22),which has unique features such as the promiscuous bindingproperties of the NKG2D receptor to multiple tumor-associatedNKG2D ligands and the inducibility of these ligands on the tumorcell surface by chemotherapy and radiotherapy (23). NKG2D-based CAR T cells elegantly use the favorable properties of theNKG2D system. The NKG2D CAR design comprises the full-length NKG2D protein fused to CD3z and it associates withDNAX-activation protein 10 (DAP10) at the cell surface. ThisNKG2D–CD3z–DAP10 complex functionally acts as a second-generation CAR, which provides a T-cell activation signal throughCD3z and costimulation through DAP10 (24, 25). NKG2D CART cells have never been tested against intracranially growingtumors such as gliomas.

1Department of Neurology and Brain Tumor Center, University Hospital Zurichand University of Zurich, Zurich, Switzerland. 2Department of Radiation Oncol-ogy, University Hospital Zurich, University of Zurich, Zurich, Switzerland. 3Centerfor Synthetic Immunity and Department of Microbiology & Immunology, GeiselSchool of Medicine, Hanover, New Hampshire.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Patrick Roth, Department of Neurology, UniversityHospital Zurich, Frauenklinikstrasse 26, 8091 Zurich, Switzerland. Phone: 41-44-255-5511; Fax: 41-44-255-4380; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-17-1788

�2017 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1031

on January 21, 2021. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst December 8, 2017; DOI: 10.1158/0008-5472.CAN-17-1788

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-17-1788&domain=pdf&date_stamp=2018-2-1

http://cancerres.aacrjournals.org/

Furthermore, the combination of CAR T-cell therapy withconventional treatment regimens has never been examined, butthe inducibility of NKG2D ligands by various stressors provides astrong rationale for this approach.

Here, we investigated NKG2D-based CAR T cells in orthotopic,syngeneic gliomamodels and addressed the questions of efficacy,trafficking, persistence, and combination with conventional anti-cancer therapy in fully immunocompetent hosts.

Materials and MethodsCell lines

SMA glioma cell lines were obtained from Dr. D. Bigner (DukeUniversity Medical Center, Durham, NC) and GL-261 cells wereobtained from the National Cancer Institute (Frederick, MD).EL-4 cells were obtained from the ATCC. All cell lines weremaintained in DMEM (Invitrogen), containing 2 mmol/LL-glutamine (Gibco Life Technologies), and 10% FCS (BiochromKG) and regularly tested negative for mycoplasma by PCR. Cellswere authenticated routinely at the Leibniz Institute DSMZ-German Collection ofMicroorganisms and Cell Cultures by shorttandem repeat analysis, previously in 2013.

CAR design and generation of CAR T cellsThe design of the murine NKG2D-based CAR (chNKG2D) and

the corresponding control construct that overexpresses wild-type-NKG2D without the intracellular CD3z domain (wtNKG2D) hasbeen described (25). CAR T cells were generated by retroviraltransduction (25). In short, syngeneic splenocytes from C57BL/6or VM/Dk mice were activated with 1 mg/mL Concanavalin A(Sigma-Aldrich) for 18–20 hours and retrovirally transducedwithchNKG2D or wtNKG2D. The cells weremaintained in RPMI1640(Gibco Life Technologies) supplemented with 10% FCS,10 mmol/L HEPES, 2 mmol/L L-glutamine, 1 mmol/L pyru-vate, 0.1 mmol/L nonessential amino acids (all from Gibco),50 mmol/L 2-mercaptoethanol (Sigma-Aldrich), and 25 IU/mLrecombinant murine IL2 (PeproTech) for 6–8 days and subse-quently used for experiments.

Antibodies and flow cytometryThe following mAbs were used for flow cytometry: anti-CD4-

AF700, anti-CD8-APC, anti-IFNg-BV421, anti-CD45.1-AF488,and anti-CD45.2-APC (Biolegend). For blocking of FC receptors,samples were preincubated with anti-mouse CD16/CD32(Biolegend). As controls, we used isotype-matched antibodiesfrom Sigma-Aldrich. Acquisition was performed on a BDFACSVerse Analyzer (BD Biosciences) and data were analyzedwith FlowJo (TreeStar).

Cytotoxicity assayGlioma cells as target cells were cocultured for 4 hours with

chNKG2D or wtNKG2D T cells at different effector: target ratios.Glioma cell lysis was assessed by a flow cytometry–based assay(18). Specific lysis was expressed as percentage of death of labeledtarget cells. For blocking experiments, transduced T cells werepreincubated for 1 hour at 4�C with anti-NKG2D or isotypecontrol from eBioscience.

IHCStaining of brain cryosections from tumor-bearing mice has

been described in detail (26). Anti-CD45.1 and anti-CD45.2

antibodies for IHC were obtained from Novus Biologicals. TheHistofine Simple Stain Mouse MAX PO was obtained fromNichirei and used as secondary antibody system.

Real-time PCRRNA isolation and cDNA preparation were performed as

described previously (27). Gene expression was measured in aQuantStudio 6 Flex Real-Time PCR System (Applied Biosystems)with SYBR Green from Thermo Fisher Scientific. The followingprimers (Microsynth AG) were used: chNKG2D (28): forward 50-GGCGTCGACACCATGAGAGCAAAATTCAGCAGGAG-30, reverse50-GGCGCTCGAGTTACACCGCCCTTTTCATGCAGAT-30, mouseHPRT1: forward 50- TTGCTGACCTGCTGGATTAC-30, reverse 50-TTTATGTCCCCCGTTGACTG-30, respectively. The conditionswere40 cycles at 95�C/15 seconds and 60 �C/1 minute. Relativequantification of gene expression was calculated with the DDCt

method (29) for relative quantification compared with the house-keeping gene HPRT1.

IFNg ELISPOT assayIndividual spleen samples (N¼3)were obtained fromNKG2D

CAR T-cell–treated long-term survivors 8 months after initialtreatment or na€�ve control mice. Untouched T cells were isolatedusing Pan T Cell Isolation Kit II (Miltenyi Biotec) and 2 � 106

T cells were cocultured with GL-261 cells for 24 hours. Thenumber of IFNg-secreting cells was measured using mouse IFNgELISPOT Ready-SET-Go from eBioscience and spot formingcells (SFC) were counted using AID EliSpot Reader classic(AID GmbH).

IFNg ELISANKG2D CAR T-cell–treated long-term survivors, 12 months

after initial treatment, or na€�ve control mice were (re)challengedwith intracranial implantation of GL-261 cells. Three days aftertumor implantation, mice were euthanized, splenocytes wereisolated, and brain-resident CD4þ or CD8þ T cells were FACSsorted. Subsequently, 5 � 104 splenocytes from long-term sur-viving mice or na€�ve control mice were cocultured with 2.5� 104

GL-261 or EL-4 cells and 5 � 103 FACS-sorted brain-residentCD4þ or CD8þ T cells were cocultured with 2.5 � 103 GL-261 orEL-4 cells. After 72 hours, cell-free conditioned media wasassessed for IFNg by ELISA from Thermo Fisher Scientific.

Animal experimentsAll experiments were done in accordance with the Institutional

Animal Care and Use Committee of the cantonal veterinary office(ZH006/15) and according to guidelines of the Swiss federal lawon animal protection. Wild-type C57BL/6 CD45.2 mice werepurchased from Charles River Laboratories and C57BL/6 CD45.1mice were obtained from the Jackson Laboratory. VM/Dk micewere bred in pathogen-free facilities at the University of Zurich.Mice of 6 to 12 weeks of age were used in all experiments. Forintracranial tumor implantation, GL-261 cells (2 � 104) werestereotactically implanted into the right striatum at day 0. Micewere observed daily and sacrificed as indicated or when devel-oping neurologic symptoms. Where indicated, mice received 5 �106 chNKG2DorwtNKG2DT cells intravenously at days 5, 7, and10 after tumor implantation. For local administration, up to 2 �106 transduced T cells were injected intratumorally at day 5 aftertumor implantation. If indicated, local cranial radiotherapywith asingle dose of 4 Gy was applied at day 7 after tumor implantation

Weiss et al.

Cancer Res; 78(4) February 15, 2018 Cancer Research1032




using a Gulmay 200 kV X-ray unit at 1 Gy/min. Long-termsurviving mice were rechallenged 5 to 8 months after the initialtumor inoculation with another tumor implantation in the con-tralateral hemisphere without any further treatment. For isolationof tumor-infiltrating immune cells at indicated time points, weperfused mice with ice-cold PBS to remove all circulating leuko-cytes from the CNS and isolated the brains. Subsequently, tumorcells were separated from myelin and red blood cells using aPercoll gradient suspension (Sigma-Aldrich). Cells were washedwith PBS and stained with Zombie Aqua Fixable Viability Kit(Biolegend) and fluorochrome-conjugated antibodies specific tocell surface markers for flow cytometry as indicated.

MRIMRIwas performedwith a 4.7 T imager (Bruker Biospin) at day

15 after tumor implantation. Coronal T2-weighted images wereacquired using Paravision 6.0 (Bruker BioSpin). If indicated,mean and SD of the tumor volume in mm3 from 5 mice/groupwere calculated using the formula (length � width � depth)/2.

Fluorescence molecular tomographyTransduced T cells were labeled with CellBriteTM NIR790

(Biotium) and administered to tumor-bearing mice as indi-cated. For fluorescence molecular tomography (FMT) imagingat indicated time points after T-cell injection, mice wereanesthetized by gas anesthesia, depilated at the tumor region,placed in a FMT4000 system (PerkinElmer) and scanned withthe 790 nm laser channel. Image analysis was performed usingTrueQuant 3.1 (PerkinElmer). A region of interest of equal sizewas placed above the tumor region and fluorescence intensitywas automatically calculated by the software.

Statistical analysisData are presented as means and SD. Experiments were repeat-

ed at least three times, if not indicated differently. Statisticalanalyses were performed in GraphPad Prism using multipletwo-tailed Student t tests and correction formultiple comparisonsusing the Holm–Sidak method. Kaplan–Meier survival analysiswas performed to assess survival differences among the treatmentgroups and P values were calculated with the log-rank test.Significance was concluded at �P < 0.05 and ��P < 0.01.

ResultsChNKG2D-transduced T cells lyse glioma cells in a NKG2D-dependent manner

We generated murine chNKG2D CAR T cells or wtNKG2Doverexpressing T cells by retroviral transduction of splenocytesderived from C57BL/6 or VM/Dkmice. NKG2D cell surface levelsin splenocytes transduced with chNKG2D or wtNKG2D wereequivalent (Supplementary Fig. S1). The differential biologicaleffects exerted by these cells can therefore be attributed specificallyto the chimeric construct. To determine the cytolytic activity ofchNKG2D- or wtNKG2D-transduced T cells, we used these cells aseffector cells and different syngeneic murine glioma cell lines astarget cells in cytotoxicity assays. ChNKG2D T cells had a signif-icantly higher specific cytolytic activity against all murine gliomacell lines than wtNKG2D T cells (Fig. 1A–D, top). Inhibition ofNKG2D signaling using a blocking antibody abrogated theenhanced cytolysis, confirming the NKG2D dependency ofchNKG2D T cells (Fig. 1A–D, bottom). Because cytokines are

important effector molecules of CAR T-cell function, we assessedthe T-cell–specific IFNg production of chNKG2D or wtNKG2DT cells cocultured with syngeneic glioma cells by intracellularcytokine staining. Both CD4þ and CD8þ T cells produced moreIFNg after transduction with chNKG2D compared withwtNKG2D when the T cells were cocultured with syngeneicglioma cells (Fig. 2A–D).

NKG2D-based CAR T cells home to orthotopic gliomas aftersystemic administration

The concept of CNS immune privilege has been refined by thedemonstrationof an intact afferent armofCNS-related immunity,which includes the discovery of classical lymphatic vessels in themeninges (30). However, the efferent arm is restricted, i.e., leu-kocyte migration to the brain (31) and the immunosuppressivemicroenvironment of solid tumors such as gliomas is anotherchallenge for access of CAR T cells to the tumor site. To determinewhether NKG2D-based CAR T cells reach orthotopically growinggliomas after systemic administration, we labeled chNKG2D Tcells with a near-infrared dye that allows in vivo tracking by FMT.After a single intravenous injection of 5 � 106 chNKG2D T cells,we detected an FMT signal at the orthotopic tumor site thatincreased over several days (Fig. 3A). Labeled wtNKG2D T cellsalso reached the tumor site (Supplementary Fig. S2). To corrob-orate this finding, we injected chNKG2D-transduced T cells gen-erated from splenocytes of CD45.1þ donor mice into CD45.2þ

glioma-bearing animals. After tumor explantation and isolationof tumor-infiltrating immune cells, we detected a CD45.1þ

chNKG2D T-cell population by flow cytometry (Fig. 3B; Supple-mentary Fig. S3). In addition, the infiltration of intracranialtumors with CD45.1þ chNKG2D T cells was verified by IHC (Fig.3C). Compared with direct intratumoral injection, which servedas a control, we detected fewer CD45.1þ cells after intravenousinjection, indicating that only a fractionof the administered T cellsmigrates to and stays at the tumor site (Fig. 3B and C).

NKG2D-based CAR T cells prolong survival of syngeneicorthotopic glioma-bearing mice

To explore the efficacy of NKG2D-based CAR T-cell therapyagainst experimental gliomas in vivo, we intravenously injectedchNKG2D or wtNKG2D T cells on days 5, 7, and 10 afterorthotopic tumor implantation. This resulted in a significantlyprolonged survival of GL-261 glioma-bearing mice and cured22%of the animals as confirmedbyMRI and long-term follow-up(Fig. 4A). Systemic administration of CAR T cells may be associ-ated with on-target off-tumor activity and thus local or systemictoxicity.Wedid not observeweight loss as an indirect indicator fortoxicity following repetitive systemic administrationof chNKG2Dor wtNKG2D T cells (Fig. 4B). According to the mouse geneexpression database and the BioGPS gene expression database,NKG2D ligands are generally expressed at low levels in normaltissues with the highest expression of RAE-1 in liver, hematopoi-etic system, and reproductive system(Supplementary Fig. S4AandS4B). We observed no changes in liver enzyme amounts in sera ofmice treated repetitively with chNKG2D or wtNKG2D T cells.Furthermore, we did not observe differences in peripheral bloodcounts (Fig. 4B).

To investigate the potential treatment effect and tolerabilityin the setting of an unrestricted migration to the tumor site, weinjected chNKG2D T cells directly intratumorally in glioma-bearing mice. This application route significantly prolonged

Radiotherapy Augments NKG2D CAR T Cells against Glioblastoma

www.aacrjournals.org Cancer Res; 78(4) February 15, 2018 1033




the survival after a single injection of chNKG2D T cells in adose-dependent manner and increased the fraction of curedmice compared with intravenous CAR T-cell administration(Fig. 4C).

Glioma-bearingmice surviving after chNKG2DT-cell treatmentare long-term protected against tumor rechallenge

An impressive feature of cancer immunotherapy is the potentialfor a long-lasting treatment effect (32).We therefore determined a

Figure 1.

NKG2D CAR T cells lyse glioma cells ina NKG2D-dependent manner. Themurine glioma cell lines GL-261 (A),SMA-497 (B), SMA-540 (C), orSMA-560 (D) were used as target cellsin 4-hour cytolysis assays. T cells fromC57BL/6 mice (for GL-261) or VM/Dkmice (for SMA-497, SMA-540, andSMA-560) were transduced withchNKG2D (CH) or wtNKG2D (WT) andused as effector cells at variouseffector:target (E:T) ratios(corresponding upper graphs for eachcell line). Blocking anti-NKG2D orisotype control antibodies were usedto preincubate chNKG2D or wtNKG2DT cells for 1 hour before using them aseffector cells at an effector:target ratioof 40:1 (corresponding lower graphsfor each cell line). Data are presentedasmean� SD (� , P <0.05; �� , P <0.01).

Weiss et al.





potential long-term immune protection in glioma-bearing micethat had survived following chNKG2D CAR T-cell treatment. Tothis end, GL-261 glioma cells were implanted into the contralat-

eral hemisphere 5 to 8 months after the first implantation. Noadditional treatment was administered. All of these mice thatsurvived after the initial CAR T-cell treatment also survived the

Figure 2.

NKG2D-based CAR T cells produce high levels of IFNg upon coculture with glioma cells. The murine glioma cell lines GL-261 (A), SMA-497 (B), SMA-540 (C),or SMA-560 (D) were cocultured with chNKG2D- or wtNKG2D-expressing T cells. After 6 hours of coculture, IFNg levels were determined in CD4þ

and CD8þ T cells by flow cytometry. Bar plots show mean and SD from two independent experiments (� , P < 0.05; �� , P < 0.01). FACS plots show data fromone representative experiment and numbers indicate percentage of IFNg-positive T cells.






Figure 3.

NKG2D-based CAR T cells migrate to intracranially growing gliomas after systemic administration. A, ChNKG2D T cells were labeled with CellBrite NIR790.Subsequently, 106-labeled cells (þctrl) or unlabeled cells (PBS, �ctrl, respectively) were injected intracranially in the brain of a nontumor-bearing mouseand the signal from labeled T cells was detected at the tumor site by FMT (left). Tumor-bearing mice were treated with a single intravenous injectionof 5� 106 chNKG2D T cells at day 5 after tumor cell implantation. The near-infrared signal was acquired at the tumor site by FMT at the indicated time points afterinjection (right). Color scales indicate signal intensities. B, A total of 5 x 106 CD45.1þ chNKG2D T cells were injected either intravenously or intratumorally ata single time point in CD45.2þ tumor-bearing mice at day 5 after tumor implantation. Three days later, tumor-infiltrating immune cells were isolated from thetumor-bearing hemisphere and CD45.1þ and CD45.2þ cells were detected by flow cytometry. C, The animalswere treated as inB and CD45.1þ cells were detected byIHC after removal of the brains at day 10. CD45.1þ T cells are stained in brown. Spleen sections from CD45.1þ mice were used as positive control. Onerepresentative image out of three different mice per group is shown.

Weiss et al.





Figure 4.

NKG2D-based CAR T-cell treatment confers a survival benefit in syngeneic orthotopic glioma-bearing mice. A, GL-261 tumor-bearing mice were treatedintravenously with 5 � 106 chNKG2D- or wtNKG2D-expressing T cells on days 5, 7, and 10 after tumor implantation. Survival data are presented asKaplan–Meier plots (left). Representative T2w MRI images of mice receiving wtNKG2D (top) or chNKG2D T cells (bottom) at day 15 after tumor implantationare shown (right). White arrow, tumor. B, Body weight was determined every other day (top left), liver enzymes (upper right) wereassessed at day 16, and peripheral blood counts were assessed at day 14 of GL-261 tumor-bearing mice treated intravenously with 5 � 106 chNKG2D orwtNKG2D T cells at days 5, 7, and 10 after tumor-implantation. C, GL-261 tumor-bearing mice were treated with a single intratumoral injection of 2 � 104, 2 � 105,or 2 � 106 chNKG2D T cells or PBS control. Survival data are presented as Kaplan–Meier plots (left). P values were calculated with log-rank test (� , P < 0.05;�� , P < 0.01). Representative T2w MRI images of mice receiving wtNKG2D (top) or 2 � 106 chNKG2D T cells (bottom) at day 15 after tumor implantationare shown (right). White arrow, tumor.







Weiss et al.




tumor rechallenge, and at day 18 after tumor implantation, notumormass was detectable byMRI (Fig. 5A and B). In contrast, allna€�ve mice had large tumors and finally had to be euthanized. Toinvestigate the mechanism underlying this long-term protection,we rechallenged the long-term survivors a second time andisolated lymphocytes from the brain and from cervical, axillary,and inguinal lymph nodes 3 days after this second tumor rechal-lenge. A prominent CD8þ T-cell population was detected withinthe rechallenged hemisphere of chNKG2D T-cell–treated long-term survivors but not in na€�ve mice (Fig. 5C). The brain-residentCD8þ T-cell population predominantly secreted IFNg in responsetoGL-261 glioma cells, as demonstrated by FACS-sorting of brain-residentCD4þ andCD8þT cells from long-term survivors or na€�vetumor-bearing mice 3 days after tumor rechallenge and ex vivostimulation with GL-261 cells or MHC-matched but non-gliomaEL-4 cells (Supplementary Fig. S5). There was no difference inintermediate and high CD44-expressing CD4þ and CD8þ T cellsbetween lymph nodes of rechallenged chNKG2D T-cell–treatedlong-term survivors and those of na€�ve tumor-bearing mice(Fig. 5D). This suggests a local long-term protection independentfrom a peripheral memory response. In line with these findings,we noticed an equivalent tumor cell cytolysis and IFNg produc-tion after coculture of GL-261 cells with T cells isolated fromlymph nodes of rechallenged chNKG2D T-cell–treated long-termsurvivors or T cells from na€�ve tumor-bearing mice. We detectedchNKG2D transcripts in brains from long-term survivors but notfrom na€�ve mice and not in peripheral lymph nodes or spleensfrom long-term survivors or na€�ve mice (Fig. 5D), which suggestslocal persistence of NKG2D CAR T cells in the CNS.

Irradiation increases the therapeutic activity of NKG2D-basedCAR T cells against gliomas

There is an intense discussion on the implementation ofimmunotherapy into conventional treatment regimens. Radio-therapy is part of the standard treatment of gliomas (33).We havepreviously observed an upregulation of NKG2D ligands on theglioma cell surface upon irradiation (23). This built the rationaleto combine radiotherapy with NKG2D-based CAR T-cell therapy.A single cranial irradiation with 4 Gy at day 7 after tumorimplantation had no effect on survival when given alone butfurther increased the activity of NKG2D-based CAR T cells againstorthotopic gliomas as demonstrated by synergistically prolongedsurvival in two independent glioma models, reduced tumorvolumemeasured byMRI, and an increased fraction of long-termsurviving mice in the SMA-560 glioma model (Fig. 6A–D). Weidentified two underlyingmechanisms for this synergistic activity.

Low-dose preirradiation of glioma cells and subsequent coculturewith chNKG2D- or wtNKG2D-expressing T cells resulted in anincreased cytolysis and IFNg production in vitro (Fig. 7A). Alsoin vivo, local tumor irradiation increased the IFNg expression oftumor-infiltrating chNKG2D T cells (Fig. 7B; SupplementaryFig. S6A). This points towards a direct tumor-cell–related effectof irradiation that boosts NKG2D CAR T-cell activity, which canbe attributed due to the induction of NKG2DL (SupplementaryFig. S6B; ref. 23). Moreover, we observed an indirect migration-related mechanism by tracking fluorescently labeled chNKG2DCAR T cells after intravenous injection. Here, we noticed anincreased accumulation of CAR T cells upon irradiation(Fig. 7C), suggesting that local subtherapeutic irradiation pro-motes the migration of CAR T cells to the tumor site.

DiscussionCART-cell therapy is an emerging immunotherapy that is under

development against several malignancies including glioblasto-ma. However, the potentially immunosuppressive microenviron-ment and intraparenchymal location of solid tumors representchallenges that require a systematic development of strategiesboosting the antitumor efficacy of CAR T cells against thesetumors (34). Furthermore, it remains to be determined howimmunotherapy with CAR T cells can be integrated at its bestinto conventional cancer therapies.

We addressed these questions in orthotopic, fully immuno-competent, preclinical glioblastomamodels. Except for one report(35), all preclinical studies published so far have assessed CAR Tcells against experimental gliomas in a xenograft setting.However,only a syngeneic setting, which involves an intact immune systemand a physiologic tumor microenvironment, allows for appro-priately assessing the challenges of CAR T-cell therapy againstsolid tumors, including potential off-tumor toxicities.

In contrast to other single target antigens for CAR T-cell therapythat are currently being investigated against glioblastoma, theNKG2DCAR elegantly targets multiple tumor-associated ligands.This may decrease the probability of tumor immune escape fromCAR T-cell treatment due to antigen loss (17).

We demonstrate strong antitumor activity of NKG2D-basedCAR T cells against glioma cells in vitro in a syngeneic setting(Figs. 1 and 2). Furthermore, in vivo tracking of CAR T cells verifieshomingofCART cells to the intracerebral tumor site after systemicCAR T-cell administration (Fig. 3). The target antigens of theNKG2D CAR are not exclusively expressed on tumor cells but canbe induced also in nontumoral tissues by unphysiologic cell stress

Figure 5.Glioma-bearing mice surviving after NKG2D CAR T-cell treatment are long-term protected against tumor rechallenge. Long-term surviving mice cured bysystemic (i.v.) or intratumoral (i.t.) administration of chNKG2D T cells were rechallenged 6 months after the initial tumor implantation with GL-261 cells in thecontralateral hemisphere. As a control, GL-261 cells were inoculated into na€�ve mice. A, Kaplan–Meier survival curves for the three cohorts are indicated.B, T2w MRI scans at day 18 after tumor implantation are shown. White arrow, tumor. The top panel represents images from na€�ve control mice and the bottompanel represents images from rechallenged chNKG2D T-cell long-term survivors following tumor implantation. C, Long-term surviving mice received asecond tumor rechallenge (2 months after the first rechallenge) and 3 days after tumor (re)implantation, tumor-infiltrating CD4þ and CD8þ T cells wereisolated and analyzed by flow cytometry. An individual plot of tumor-infiltrating lymphocytes from onemouse is shown on the left and a diagram depicting themeanand SD of three mice is shown on the right (� , P < 0.05; �� , P < 0.01). D, Same setup as in C and isolation of cervical, axillary, and inguinal lymph nodes at day 3 afterthe second tumor rechallenge. Effector (Teff) and memory (Tmem) T cells were separated using flow cytometry and CD44 expression levels. Populationswith intermediate (effector T cells) or high (memory T cells) CD44 expression in the CD4þ or CD8þ T-cell compartments are indicated (left). Theisolated lymph node–derived T cells were used as effector cells in a 4-hour immune cell lysis assay or IFNg ELISPOT with fresh GL-261 cells as target cells(right top panel). Real-time PCR for chNKG2D was performed after RNA isolation and cDNA preparation from brains, spleens, and peripheral lymph nodesisolated 3 days after tumor challenge of na€�ve control mice or 3 days after second rechallenge of long-term surviving mice 8 months after initial NKG2DCAR T-cell treatment (right bottom panel; n.d., nondetectable). E:T, effector:target ratio.






Figure 6.

Irradiation and systemic NKG2D-based CAR T-cell treatment act synergistically against experimental gliomas. GL-261 (A and C) or SMA-560 (B and D)tumor-bearing mice were treated intravenously with 5 � 106 chNKG2D (CH) or wtNKG2D (WT) T cells on days 5, 7, and 10 after tumor implantation eitheralone or with a single local irradiation with 4 Gy at day 7 after tumor implantation. A and B, Kaplan–Meier curves are shown and P values were calculated withlog-rank test (�� , P < 0.01 for WT vs. CH; �� , P < 0.001 for WT þ IR vs. CHþIR). C, Representative MRI scans at day 15 of three animals of each of the fourcohorts of GL-261 tumor-bearing mice are shown (left). White arrow, tumor. Tumor volumes (right) were calculated on the basis of MRI analyses using theformula H � W � L/2 (� , P < 0.05 and �� , P < 0.01 for WT vs. CH or WTþIR vs. CHþ IR; þ, P < 0.05 for CH vs. CHþIR). D, Same setting as in C but for SMA-560tumor-bearing mice.

Weiss et al.





Figure 7.

Irradiation-mediated boosting of CAR T-cell activity relies on direct effects on tumor cells as well as increased CAR T-cell migration. A, GL-261 cells,preirradiated (þIR) 24 hours prior to the assay with 4 Gy or not, were used as target cells in cytotoxicity assays using chNKG2D (CH) or wtNKG2D (WT)T cells as effector cells at the indicated effector:target (E:T) ratios (left). After 4 hours of coculture at a ratio of 40:1, IFNg levels were determined by intracellularcytokine staining andflowcytometric assessment (right). Mean and SDare shown (� ,P<0.05; �� ,P<0.01 forWTvs. CHandþ,P<0.05;þþ,P<0.01 forWT/CHvs.WT/CH þ IR). B, 5 � 106 CD45.1þ wtNKG2D or chNKG2D were i.v. injected on days 5, 7, 10 after implantation of GL-261 tumors in CD45.2þ mice. In addition, mice wereirradiated (IR) at day 7 with a single local dose of 4 Gy or not. At day 12 after tumor implantation, brain-infiltrating immune cells were isolated and IFNg expressionassessed inCD45.1þ cells by flowcytometry. Mean and SD from threemice are shownwith �� ,P<0.01 forWTvs. CHandWTþ IR vs. CHþIR andþ,P<0.05; forWT/CHvs. WT/CH þ IR. C, ChNKG2D T cells were labeled with CellBrite NIR790. Subsequently, 5 � 106 labeled cells were intravenously injected on days 5, 7, and10 after tumor implantation. In addition, mice were irradiated at day 7 with a single local dose of 4 Gy or not. The near-infrared signal was acquired at the tumor-siteby FMT at the following time points: T1 ¼ 24 hours prior to irradiation, T2 ¼ 24 hours post-irradiation, T3 ¼ 144 hours post-irradiation. Representative images areshown on the left and a quantification of the detected signal is shown on the right (� , P < 0.05; �� , P < 0.01).






or inflammation (36–38). AlthoughNKG2D ligand expression byother tissuesmay result in on-target off-tumor toxicity, we did notobserve any signs of toxicity upon systemic CAR T-cell adminis-tration (Fig. 4B) at the level of body weight, liver function tests orperipheral blood counts. In line with these findings, a first-in-human phase I trial assessing NKG2D-based CAR T cells againsthematologic malignancies has completed enrollment of a firstcohort of patients without treatment-related safety issues (39).

Treatment of glioma-bearing mice with NKG2D CAR T cellsresulted in a significant proportion of surviving mice, which werelong-termprotected against tumor rechallenge in the contralateralhemisphere (Fig. 5A and B). In contrast to previous studies (28,35), this may not only be the result of a classical memory T-cellresponse, but probably it may also involve long-term persistenceof NKG2D CAR T cells in the brain as demonstrated by detectionof the chNKG2D transcripts in the brain tissue of long-termsurviving mice 8 months after initial CAR T-cell treatment (Fig.5D). It remains anopenquestionwhyNKG2DCART cells seem topersist in the CNS but not in extracerebral organs. Possibly, theCNS as an immune-privileged site provides a niche for homeo-static proliferation of tissue-resident T cells, which allows CAR T-cell persistence.

Future glioblastoma treatment regimens may rely on combi-nation therapies and may implement novel immunotherapeuticstrategies such as CAR T-cell administration. The combination ofimmunotherapy with conventional anticancer treatments such asradiotherapy may result in synergistic activity (40, 41). However,the combination of CART-cell therapywith radiotherapy has onlybeen investigated with total body irradiation used as a myeloa-blative host-conditioning regimen before administration of CART cells (35). Our data demonstrate synergistic antitumor activityof local tumor irradiation andNKG2DCART-cell therapy (Fig. 6).Mechanistically, this combination produces stronger CAR T-cellactivity upon recognition of irradiated tumor cells and animproved trafficking of intravenously injected CAR T cells to thetumor site (Fig. 7).

In summary, this study highlights the potential of NKG2DCART cells as a promising therapeutic approach against glioblastoma.

Long-term tumor control due to persistence of these cells at thetumor site and synergistic activity in combination with radiother-apy suggest that NKG2DCAR T cells represent a novel therapeuticoption that warrants clinical evaluation in glioma patients.

Disclosure of Potential Conflicts of InterestC.L. Sentman reports receiving a commercial research grant fromCelyad and

Celdara Medical, has ownership interest (including patents) in Celyad, is aconsultant/advisory board member for Celyad and Celdara Medical, and hasprovided expert testimony for Celyad and CeldaraMedical. P. Roth has receivedspeakers bureau honoraria from BMS, Novartis, and Novocure and is a con-sultant/advisory board member for MSD, Virometix, Roche, Covagen, andMolecular Partners. No potential conflicts of interest were disclosed by theother authors.

Authors' ContributionsConception and design: T. Weiss, M. Weller, C.L. Sentman, P. RothDevelopment of methodology: T. Weiss, P. RothAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): T. WeissAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): T. Weiss, M. Weller, C.L. Sentman, P. RothWriting, review, and/or revision of the manuscript: T. Weiss, M. Weller,M. Guckenberger, C.L. Sentman, P. RothAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): T. Weiss, M. Weller, M. Guckenberger, P. RothStudy supervision: M. Weller, P. Roth

AcknowledgmentsThis study was supported by grants from the Swiss National Science Foun-

dation (310030_170027 to P. Roth), the Gertrud-Hagmann Foundation, andthe Swiss Cancer League (KFS-3478-08-2014 to P. Roth), and "Hochspeziali-sierte Medizin Zurich" (HSM-2 to P. Roth, M. Guckenberger, and M. Weller) aswell as the Detas Foundation to M. Weller. We thank Professor Christian M€unzfor providing access to the ELISpot Reader.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received June 15, 2017; revised October 25, 2017; accepted November 29,2017; published OnlineFirst December 8, 2017.

References1. Ostrom QT, Gittleman H, Fulop J, Liu M, Blanda R, Kromer C, et al.

CBTRUS statistical report: primary brain and central nervous systemtumors diagnosed in the United States in 2008–2012. Neuro Oncol 2015;17:iv1–iv62.

2. Weller M, van den BentM, Tonn JC, Stupp R, PreusserM, Cohen-Jonathan-Moyal E, et al. European association for neuro-oncology (EANO) guidelineon the diagnosis and treatment of adult astrocytic and oligodendroglialgliomas. Lancet Oncol 2017;18:e315–e29.

3. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ,et al. Radiotherapy plus concomitant and adjuvant temozolomide forglioblastoma. N Engl J Med 2005;352:987–96.

4. Johnson LA, June CH. Driving gene-engineered T cell immunotherapy ofcancer. Cell Res 2017;27:38–58.

5. Gacerez AT, Arellano B, Sentman CL. How chimeric antigen receptordesign affects adoptive T cell therapy. J Cell Physiol 2016;231:2590–8.

6. Sadelain M, Brentjens R, Riviere I. The basic principles of chimeric antigenreceptor design. Cancer Discov 2013;3:388–98.

7. Lim WA, June CH. The principles of engineering immune cells to treatcancer. Cell 2017;168:724–40.

8. Grupp SA, Kalos M, Barrett D, Aplenc R, Porter DL, Rheingold SR, et al.Chimeric antigen receptor-modified T cells for acute lymphoid leukemia.N Engl J Med 2013;368:1509–18.

9. Porter DL, Levine BL, KalosM, Bagg A, June CH. Chimeric antigen receptor-modified T cells in chronic lymphoid leukemia. N Engl J Med 2011;365:725–33.

10. Brown CE, Alizadeh D, Starr R, Weng L, Wagner JR, Naranjo A, et al.Regression of glioblastoma after chimeric antigen receptor T-cell therapy.N Engl J Med 2016;375:2561–9.

11. GilhamDE,Debets R, PuleM,Hawkins RE, AbkenH. CAR-T cells and solidtumors: tuning T cells to challenge an inveterate foe. Trends Mol Med2012;18:377–84.

12. Johnson LA, Scholler J, Ohkuri T, Kosaka A, Patel PR, McGettigan SE, et al.Rational development and characterization of humanized anti-EGFRvariant III chimeric antigen receptor T cells for glioblastoma. Sci TranslMed 2015;7:275ra22.

13. Chow KK, Naik S, Kakarla S, Brawley VS, Shaffer DR, Yi Z, et al. T cellsredirected to EphA2 for the immunotherapy of glioblastoma. Mol Therapy2013;21:629–37.

14. AhmedN, Brawley V, HegdeM, Bielamowicz K,Wakefield A, Ghazi A, et al.Autologous HER2 CMV bispecific CAR T cells are safe and demonstrateclinical benefit for glioblastoma in a Phase I trial. J ImmunoThera Cancer2015;3:O11.

15. Zhang C, Burger MC, Jennewein L, Genssler S, Schonfeld K, Zeiner P, et al.ErbB2/HER2-specific NK cells for targeted therapy of glioblastoma. J NatlCancer Inst 2016;108. doi: 10.1093/jnci/djv375.

Weiss et al.





16. Brown CE, Badie B, Barish ME, Weng L, Ostberg JR, Chang WC, et al.Bioactivity and safety of IL13Ralpha2-redirected chimeric antigen receptorCD8þ T cells in patients with recurrent glioblastoma. Clin Cancer Res2015;21:4062–72.

17. Hegde M, Mukherjee M, Grada Z, Pignata A, Landi D, Navai SA, et al.Tandem CAR T cells targeting HER2 and IL13Ralpha2 mitigate tumorantigen escape. J Clin Invest 2016;126:3036–52.

18. CodoP,WellerM,Meister G, Szabo E, Steinle A,WolterM, et al.MicroRNA-mediated down-regulation of NKG2D ligands contributes to gliomaimmune escape. Oncotarget 2014;5:7651–62.

19. Crane CA, Austgen K, Haberthur K, Hofmann C, Moyes KW, Avanesyan L,et al. Immune evasion mediated by tumor-derived lactate dehydrogenaseinduction of NKG2D ligands on myeloid cells in glioblastoma patients.Proc Nat Acad Sci U S A 2014;111:12823–8.

20. Crane CA, Han SJ, Barry JJ, Ahn BJ, Lanier LL, Parsa AT. TGF-beta down-regulates the activating receptor NKG2D on NK cells and CD8þ T cells inglioma patients. Neuro Oncol 2010;12:7–13.

21. Eisele G,Wischhusen J,MittelbronnM,MeyermannR,Waldhauer I, SteinleA, et al. TGF-beta and metalloproteinases differentially suppress NKG2Dligand surface expression on malignant glioma cells. Brain 2006;129:2416–25.

22. Friese MA, Platten M, Lutz SZ, Naumann U, Aulwurm S, Bischof F, et al.MICA/NKG2D-mediated immunogene therapy of experimental gliomas.Cancer Res 2003;63:8996–9006.

23. Weiss T, Schneider H, Silginer M, Steinle A, Pruschy M, Poli�c B, et al.NKG2D-dependent antitumor effects of chemotherapy and radiotherapyagainst glioblastoma. Clin Cancer Res 2017. doi: 10.1158/1078-0432.CCR-17-1766.

24. Demoulin B, Cook WJ, Murad J, Graber DJ, Sentman ML, Lonez C, et al.Exploiting natural killer group 2D receptors for CAR T-cell therapy. FutureOncol 2017;13:1593–605.

25. Zhang T, Lemoi BA, Sentman CL. Chimeric NK-receptor-bearing T cellsmediate antitumor immunotherapy. Blood 2005;106:1544–51.

26. Schneider H, Szabo E, Machado RA, Broggini-Tenzer A, Walter A, LobellM, et al. Novel TIE-2 inhibitor BAY-826 displays in vivo efficacy inexperimental syngeneic murine glioma models. J Neurochem 2017;140:170–82.

27. Silginer M, Burghardt I, Gramatzki D, Bunse L, Leske H, Rushing EJ, et al.The aryl hydrocarbon receptor links integrin signaling to the TGF-betapathway. Oncogene 2016;35:3260–71.

28. Barber A, Zhang T, Sentman CL. Immunotherapy with chimeric NKG2Dreceptors leads to long-term tumor-free survival and development of

host antitumor immunity in murine ovarian cancer. J Immunol 2008;180:72–8.

29. Livak KJ, Schmittgen TD. Analysis of relative gene expression data usingreal-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods2001;25:402–8.

30. Louveau A, Smirnov I, Keyes TJ, Eccles JD, Rouhani SJ, Peske JD, et al.Structural and functional features of central nervous system lymphaticvessels. Nature 2015;523:337–41.

31. Harris MG, Hulseberg P, Ling C, Karman J, Clarkson BD, Harding JS, et al.Immune privilege of the CNS is not the consequence of limited antigensampling. Sci Rep 2014;4:4422.

32. Schadendorf D, Hodi FS, Robert C, Weber JS, Margolin K, Hamid O, et al.Pooled analysis of long-term survival data from phase II and phase III trialsof ipilimumab in unresectable or metastatic melanoma. J Clin Oncol2015;33:1889–94.

33. Weller M, van den Bent M, Hopkins K, Tonn JC, Stupp R, Falini A, et al.EANO guideline for the diagnosis and treatment of anaplastic gliomas andglioblastoma. Lancet Oncol 2014;15:e395–e403.

34. Newick K, O'Brien S, Moon E, Albelda SM. CAR T cell therapy for solidtumors. Annu Rev Med 2017;68:139–52.

35. Sampson JH, Choi BD, Sanchez-Perez L, Suryadevara CM, Snyder DJ,Flores CT, et al. EGFRvIII mCAR-modified T-cell therapy cures mice withestablished intracerebral glioma and generates host immunity againsttumor-antigen loss. Clin Cancer Res 2014;20:972–84.

36. HuangWC, EasomNJ, Tang XZ, Gill US, Singh H, Robertson F, et al. T cellsinfiltrating diseased liver express ligands for the NKG2D stress surveillancesystem. J Immunol 2017;198:1172–82.

37. Strid J, Roberts SJ, Filler RB, Lewis JM, Kwong BY, Schpero W, et al. Acuteupregulation of anNKG2D ligand promotes rapid reorganization of a localimmune compartment with pleiotropic effects on carcinogenesis. NatImmunol 2008;9:146–54.

38. Strid J, Sobolev O, Zafirova B, Polic B, Hayday A. The intraepithelial T cellresponse to NKG2D-ligands links lymphoid stress surveillance to atopy.Science 2011;334:1293–7.

39. Nikiforow S, Murad JM, Daley H, Negren H, Reder J, Sentman C, et al.A first-in-human phase I trial of NKG2D chimeric antigen receptor-Tcells in AML/MDS and multiple myeloma. J Clin Oncol 2016;34.

40. Melero I, Berman DM, Aznar MA, Korman AJ, Perez Gracia JL, Haanen J.Evolving synergistic combinations of targeted immunotherapies to combatcancer. Nat Rev Cancer 2015;15:457–72.

41. Weichselbaum RR, Liang H, Deng L, Fu YX. Radiotherapy and immuno-therapy: A beneficial liaison? Nat Rev Clin Oncol 2017;14:365–79.






2018;78:1031-1043. Published OnlineFirst December 8, 2017.Cancer Res Tobias Weiss, Michael Weller, Matthias Guckenberger, et al. Efficacy in GlioblastomaNKG2D-Based CAR T Cells and Radiotherapy Exert Synergistic

Updated version

10.1158/0008-5472.CAN-17-1788doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2017/12/08/0008-5472.CAN-17-1788.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/78/4/1031.full#ref-list-1

This article cites 38 articles, 11 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/78/4/1031.full#related-urls

This article has been cited by 13 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/78/4/1031To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-17-1788

http://cancerres.aacrjournals.org/content/suppl/2017/12/08/0008-5472.CAN-17-1788.DC1

http://cancerres.aacrjournals.org/content/78/4/1031.full#ref-list-1

http://cancerres.aacrjournals.org/content/78/4/1031.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/78/4/1031


Documents

NKG2D-Based CAR T Cells and Radiotherapy Exert ......an MHC-independent manner (7). CAR T-cell therapy has led to encouraging clinical responses in hematologic malignancies (8, 9)