egfrviii mcar-modiﬁed t-cell therapy cures mice with ... · egfrviii mcar-modiﬁed t-cell...

Cancer Therapy: Preclinical

EGFRvIII mCAR-Modified T-Cell Therapy Cures Mice withEstablished Intracerebral Glioma and Generates HostImmunity against Tumor-Antigen Loss

John H. Sampson1,2,3, Bryan D. Choi1,2,3, Luis Sanchez-Perez1,3, Carter M. Suryadevara1, David J. Snyder1,Catherine T. Flores1, Robert J. Schmittling1, Smita K. Nair1, Elizabeth A. Reap1, Pamela K. Norberg1,James E. Herndon II4, Chien-Tsun Kuan2, Richard A.Morgan5, Steven A. Rosenberg5, and Laura A. Johnson1,3

AbstractPurpose:Chimeric antigen receptor (CAR) transducedT cells represent a promising immune therapy that

has been shown to successfully treat cancers in mice and humans. However, CARs targeting antigens

expressed in both tumors and normal tissues have led to significant toxicity. Preclinical studies have been

limited by the use of xenograft models that do not adequately recapitulate the immune systemof a clinically

relevant host. A constitutively activated mutant of the naturally occurring epidermal growth factor receptor

(EGFRvIII) is antigenically identical in both human and mouse glioma, but is also completely absent from

any normal tissues.

Experimental Design: We developed a third-generation, EGFRvIII-specific murine CAR (mCAR), and

performed tests to determine its efficacy in a fully immunocompetent mouse model of malignant glioma.

Results:At elevateddoses, infusionwith EGFRvIIImCART cells led to cures in allmicewithbrain tumors.

In addition, antitumor efficacywas found to be dependent on lymphodepletive host conditioning. Selective

blockadewith EGFRvIII soluble peptide significantly abrogated the activity of EGFRvIIImCAR T cells in vitro

and in vivo, and may offer a novel strategy to enhance the safety profile for CAR-based therapy. Finally,

mCAR-treated, cured mice were resistant to rechallenge with EGFRvIIINEG tumors, suggesting generation of

host immunity against additional tumor antigens.

Conclusion:All together, these data support that third-generation, EGFRvIII-specificmCARs are effective

against gliomas in the brain and highlight the importance of syngeneic, immunocompetent models in the

preclinical evaluation of tumor immunotherapies. Clin Cancer Res; 20(4); 972–84. �2013 AACR.

IntroductionGlioblastoma is the most common brain tumor and also

the most deadly. Despite aggressive treatment, including

surgical resection, dose-intensive radiation, andmultimod-al chemotherapy, prognosis remains exceedingly grim witha median survival of less than 15 months from the time ofdiagnosis (1). Moreover, the effects of current therapies areoften nonspecific, causing significant collateral damage tohealthy cells and adjacent brain.

In contrast, immunotherapy offers an extremely preciseapproach with the potential to eliminate cancer specifi-cally while leaving normal tissues intact. Substantialevidence suggests that T cells in particular have the abilityto eradicate large, well-established tumors in mice andhumans (2–7). As such, a number of attempts have beenmade to establish large populations of tumor-reactive Tcells in vivo, mainly via adoptive transfer of expandedtumor-infiltrating lymphocytes (TIL) or autologous Tcells that have been transduced to express specific T-cellreceptors (TCR; refs. 2, 3, and 5). Although promising,these approaches have been limited by a number oftechnical and functional drawbacks. Although TILs aredifficult to isolate in most cancers, TCR-transduced T cellsrecognize only specific MHC alleles, restricting them to asubset of patients and making them vulnerable to MHCdownregulation by tumors (8).

Authors' Affiliations: 1Duke Brain Tumor Immunotherapy Program, Divi-sion of Neurosurgery, Department of Surgery; 2Department of Pathology;3The Preston Robert Tisch Brain Tumor Center at Duke, Duke UniversityMedical Center; 4Department of Biostatistics and Bioinformatics, DukeUniversity, Durham, North Carolina; and 5Surgery Branch, National CancerInstitute, National Institutes of Health, Bethesda, Maryland

J.H. Sampson, B.D. Choi, and L. Sanchez-Perez contributed equally to thiswork.

Current address for Laura A. Johnson: Translational Research Program,AbramsonFamilyCancerResearch Institute, PerelmanSchool ofMedicine,University of Pennsylvania, Philadelphia, PA.

Corresponding Authors: Laura A. Johnson, Translational Research Pro-gram, Abramson Family Cancer Research Institute, Department of Pathol-ogy and Laboratory Medicine, Perelman School of Medicine, 3400 CivicCenter Blvd, University of Pennsylvania, Philadelphia, PA 19104. Phone:215-746-1887; Fax: 215-573-8590; E-mail: [email protected]; and JohnH. Sampson, Duke Brain Tumor Immunotherapy Program, Division ofNeurosurgery, Department of Surgery, Duke University Medical Center,Box 3050, Durham, NC 27710. Phone: 919-684-9043; Fax: 919-419-1741;E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-13-0709

�2013 American Association for Cancer Research.

ClinicalCancer

Research

Clin Cancer Res; 20(4) February 15, 2014972

on November 24, 2020. © 2014 American Association for Cancer Research.clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst December 18, 2013; DOI: 10.1158/1078-0432.CCR-13-0709

http://clincancerres.aacrjournals.org/

To address these limitations, a versatile class of recep-tors known as chimeric antigen receptors (CAR) has beengenerated by combining the variable region of an anti-body with a T-cell signaling molecule, usually CD3 (9).Because their capacity for antigen recognition is derivedfrom antibody binding, CARs have the ability to mimicendogenous TCR-mediated activation without the draw-backs of classical MHC restriction. Moreover, althoughphysiologic TCRs are restricted by thymic selection, anti-body-redirected CARs can accommodate virtually infiniteantigenic diversity and operate at affinities even in thenanomolar range (10, 11).An additional advantage of the CAR platform is the

incorporation of costimulatory molecules such as CD28and 4-1BB into the CD3 signaling domain to improveT-cell expansion, survival, cytokine secretion, and tumorlysis (12–14). Clinical trials utilizing these second- andthird-generation CARs have now targeted a variety ofantigens and malignancies and have demonstrated theirremarkable potential (15–18). However, severe adverseevents have occurred when these CARs have been directedagainst antigens shared by normal tissues, such as ERBB2/HER2 (19). As such, the lack of truly tumor-specifictargets for CARs and a poor toxicity profile to daterepresent critical barriers to the safe and effective trans-lation of this promising therapy.EGFRvIII is a tumor-specific mutation of the epidermal

growth factor receptor that is absent from normal tissues,but commonly expressed on the surface of glioblastomasand other neoplasms (20). Functionally, EGFRvIII is aconstitutively active version of the wild-type receptor, con-ferring enhanced tumorgenicity (21, 22), invasiveness (23),and therapeutic resistance (24) to tumor cells. Because thismutation results in the translation of a unique extracellular

epitope, it is readily recognized by a number of previouslydescribed monoclonal antibodies (20); EGFRvIII thusrepresents an ideal target for CAR-based therapeuticdevelopment.

With few exceptions, the great majority of preclinicalstudies for CARs have been performed in vitro or in vivowith xenogeneic models wherein human T cells aretested against human tumors implanted into immune-compromised mice (25–30). This strategy is often theonly available option, due to the lack of immunocom-petent rodent models possessing surface molecules ofequivalent binding affinities and function to those foundin humans. Unfortunately, preclinical reports of gene-modified T cells in xenograft systems have not beenpredictive of dramatic toxicities that occurred upon trans-lation in early clinical trials (13, 19). In addition toinadequately assessing autoimmune toxicity, these xeno-graft models also do not permit realistic analyses ofparameters that may critically impact efficacy in humans,such as the influence of host-conditioning regimens,species-specific immunosuppressive factors, and thepotential generation or priming of endogenous immunity(26).

In this study, we directly address the limitations ofprevious immunecompromised models by generating amurine-derived, third-generation, EGFRvIII-specific CAR[EGFRvIII murine CAR (mCAR)] for evaluation in a fullyimmunocompetent mouse model of malignant glioma(31). In addition, we target a murine homologue of EGFR-vIII that demonstrates identical antibody-binding charac-teristics to the human EGFRvIII (32). Our results demon-strate that murine T cells transduced with the EGFRvIIImCAR (mCAR T cells) express IFN-g specifically in thepresence of target cells expressing the EGFRvIII mutation.Despite conventional notions of immune-privilege, treat-ment withmCAR T cells led to complete eradication of 3- to5-day established, syngeneic, EGFRvIII-expressing gliomaslocated subcutaneously and in the brain. Therapeutic effectswere shown to be dose dependent and required host lym-phodepletion before adoptive transfer for efficacy. We alsoshow the ability to block this mCAR T-cell function in vivowith systemic administration of EGFRvIII peptide. Finally,successfully cured mice did not develop tumors uponrechallenge with EGFRvIIINEG matched tumor, suggestingthat adoptive transfer in this setting may generate hostimmunity against novel tumor antigens, thereby circum-venting tumor-antigen loss and preventing tumor recur-rence (33).

Together, our findings support that third-generationCARs can be used to produce effective immune responsesagainst tumors in thebrain. In addition, these data highlightthe requirement of lymphodepletive host conditioning forefficacy, and also suggest that epitope spreading may con-tribute to favorable outcomes in the setting of CAR-medi-ated therapies. Ultimately, these studies lend credence tothe utilization of fully immunocompetent models in thesetting of adoptive transfer, with the potential to informcritical aspects of clinical trial design.

Translational RelevanceOur study describes a new model for testing chimeric

antigen receptor (CAR) gene-engineered T lymphocytesagainst solid brain tumors in vivo in a syngeneic,immune-intact mouse. Using a naturally arising murineglioma expressing the tumor-specific mutation EGFR-vIII, wewere able to treat thesemice by systemic adoptivetransfer of syngeneic T cells transduced with a thirdgeneration CAR encoding anti-EGFRvIII mAb scFvlinked to murine CD28:4-1BB:CD3z intracellulardomain. Intravenous delivery of these murine CAR Tcells was able to treat tumors in a lymphodepletive anddose-dependent fashion. Long-term "cured" animalswere able to reject rechallengewith an EGFRvIII-negativetumor, suggesting endogenous immune system contri-bution to immunity. This work describes a relevant invivo model for translational cancer research, namely,using immune-intact mice for adoptive T-cell–basedcancer therapy, and has direct clinical implications thathave previously been largely overlooked.

EGFRvIII mCARs for Malignant Glioma

www.aacrjournals.org Clin Cancer Res; 20(4) February 15, 2014 973




Materials and MethodsCell lines and media

The murine glioma parental cell line SMA560 (EGFR-vIIINEG), and its subline, SMA560vIII (EGFRvIIIPOS) havebeen previously described (31). Murine T cells werestimulated in complete RPMI-1640 media (R10; HEPES,Pen/Strep, NEAA, L-Glut, 2ME, RPMI-1640 plus 10%FBS) supplemented with 50 IU/mL rh interleukin (IL)-2 or 1 ng/mL mIL-7 for 2 days, followed by expansion in50 IU/mL IL-2.

Retrovirus design and T-cell transductionThe murine (m) 139 single-chain antibody variable frag-

ment (scFv)mCAR retrovirus was generated similarly to thepublished human (h) 139 scFv hCAR (34). The 139 scFvwas inserted in tandem with mCD8 trans-membrane,mCD28, m4-1BB, and mCD3z intracellular regions in theMSGV1 retroviral vector as previously described (13, 35).Retroviral supernatant was generated by transient cotrans-fection of HEK 293T cells by Lipofectamine 2000 Transfec-tion Reagent (Invitrogen), along with pCL-Eco helperplasmid (Imgenex). The resulting retroviral supernatantwas used to transduce murine splenocytes as previouslydescribed (36). Briefly, mouse splenocytes were collectedfrom VM/Dkmice, disaggregated, and passed through a 70-mm mesh filter to generate a single-cell suspension. Cellswere cultured in complete R10 mouse T-cell media supple-mented with 1 ng/mL rmIL-7 and activated on day 0 with2.5 mg/mL Concanavalin A (ConA). T cells were transducedon RetroNectin-coated plates (Takara Inc.) at a density of1 � 106/mL on day 2 after stimulation. Cells were thenmaintained at 1 to 2 � 106/mL in mouse T-cell media with50 IU/mL rhIL-2 for 4 to 7 days.

Flow cytometric analysisLive cells were gated on the lymphocyte population

by forward scatter/side scatter (FSC/SSC). TransducedT cells were stained for mCAR surface expression usinggoat–anti-human F(ab0)2-biotin primary and streptavi-din–phycoerythrin (SA–PE) secondary antibodies as pre-viously described (34), or using a PEPvIII-PE multimergenerated in the lab. Transduction efficiency was con-firmed by paired GFP-transduced controls. SMA560 andSMA560vIII tumors were stained for surface EGFRvIIIexpression using the L8A4 mAb (Duke University) witha goat–anti-mouse Ig-PE secondary antibody.

IFN-g ELISpot assayT-cell responses to EGFRvIIIPOS tumor target cells were

measured ex vivo by direct IFN-g ELISpot assay. Single-cellsuspensions of splenic lymphocytes from na€�ve or micetreated with GFP or EGFRvIII mCAR T cells in R10 mediumwere prepared by physical disruption. After red blood celllysis and 2 washes with R10 medium, lymphocytes wereisolated by density gradient centrifugation on LympholyteM (Cedarlane), washed 2 times, and resuspended in freshR10 medium. Splenic lymphocytes (2.5 � 105/well) wereadded to duplicate wells of ELISpot assay plates coated with

anti-mouse IFN-g monoclonal antibody (AN18) andallowed to settle for 1 hour. SMA560 or SMA560vIII tumortargets were subsequently added after 1 hour at an effector-to-tumor cell ratio of 1:1 and incubated overnight at 37�C.Plates were washed with PBS Tween-20, incubated withbiotinylated, anti-mouse IFN-g mAb [R4-6A2 (2 hours),avidin–peroxidase complex (1 hour), and substrate (3-amino-9-ethylcarbazole) for 4 minutes], separated by PBSwashes at room temperature. Plates were dried and shippedto Zellnet Consulting for spot enumeration by automatedanalysis with a Zeiss KS ELISpot system. This assay wasMIATA compliant.

Syngeneic murine model for brain tumorimmunotherapy

SMA560vIII cellswereharvestedwith trypsin,washedonetime in serum-containing medium, and washed 2 times inDulbecco’s PBS (DPBS). Cell pellets were resuspended inDPBS at the appropriate concentration of viable cells asdetermined by trypan blue dye exclusion, mixed with anequal volume of 10%methylcellulose in Zinc OptionMod-ifiedEaglemedium(MEM)and loaded into a250-mL syringewith an attached 25-gauge needle. In VM/Dkmice, the tip ofthe needle was positioned at bregma and 2 mm to the rightof the cranial midline suture and 4 mm below the surfaceof the cranium using a stereotactic frame (David KopfInstruments). Each mouse received 5 � 104 tumor cellsimplanted into the brain in a volume of 5 mL. Three to fivedays following implantation,micewere treatedwith 1� 105

to 1 � 107 mCAR or GFP control VM/Dk T cells deliveredsystemically via tail–vein injection. Absolute numbers ofT cells were used for calculations, rather than number ofmCAR positive T cells (mCAR surface expression rangedfrom 50% to 75%). In experiments evaluating efficacyagainst subcutaneous tumors, 5 � 105 SMA560vIII tumorcells were injected into the right flank, in a volume of 100 mLPBS, followed 3 days later by 5� 106 mCAR or GFP controlT cells delivered systemically via tail–vein injection. Whereindicated, mice were conditioned with 5 Gy total bodyirradiation (TBI) immediately before T-cell transfer. For eachin vivo experiment, groups consisted of 8 mice, unless statedotherwise. All experiments were repeated a minimum of2 times. According to Duke University and InstitutionalAnimal Care and Use Committee (IACUC) protocols, ani-mal survival was monitored, with mortality recorded ormice sacrificed upon reaching predetermined morbidityendpoints. For subcutaneous experiments, tumors weremeasured forwidth and length atwidest diameters to obtainsize in mm2, animals were sacrificed upon any single diam-eter reaching 20 mm or upon ulceration.

Soluble PEPvIII blockade in vitroThe effect of peptide blocking on EGFRvIII mCAR T

cells was measured utilizing the ELISpot assay describedabove. 24-well tissue-culture plate wells were coated with50 mg/mL of soluble EGFRvIII peptide, PEPvIII (LEEKKG-NYVVTDHC), or control. EGFRvIII mCAR or GFP controlT cells were added to the wells in presence or absence

Sampson et al.

Clin Cancer Res; 20(4) February 15, 2014 Clinical Cancer Research974




of 10 or 100 mg/mL of PEPvIII or irrelevant isocitratedehydrogenase (IDH) peptide (GWVKPIIIGHHAYGD-QYR). Cells (5 � 104/well) were incubated for 18 hoursand submitted for IFN-g ELISpot analysis.

Soluble PEPvIII blockade in vivoPEPvIII peptide was prepared as a stock solution in

dimethyl sulfoxide, then diluted 10� in PBS before inject-ing 100 mg in 100 mL via tail vein into mice that hadpreviously receivedmCAR or GFP control T cells. Injectionswere administered 2 to 4 hours after T-cell transfer, andrepeated 10 days later.

In vivo passaging of SMA560p tumorThe SMA560 tumor line was cultured in Zinc Option

MEM media for 1 week before implantation. At the timeof implantation, cells were harvested with trypsin,washed 1 time in serum-containing medium, washed 2times in DPBS, and adjusted to a concentration of 5 �106 cells/mL. One hundred microliters of cells weresubcutaneously injected into the right flank of VM/Dkmice and were allowed to grow and establish for 14 days.Tumors were then harvested and made into a single-cellsuspension by passing them through a 70 mm cell strain-er and submitted to EGFRvIII staining as describedabove.

Analysis ofCARpersistence inmurineperipheral bloodlymphocytesWhole blood obtained by retro-orbital bleeding was

analyzed for EGFRvIII mCAR T cells as follows: 50 mL ofblood was incubated with the antibody CD3-APC, andPEPvIII-PE multimer in 150 mL FACS buffer protected fromlight for 15 minutes at room temperature. Red blood cellswere lysed and cells were fixed using 1 mL 1� BD FACSLysing Solution (BD Bioscience, Cat. No. 349202) andincubated overnight at 4�C. Cells were then washed andresuspended in 2% paraformaldehyde and submitted toflow cytometry analysis. All samples were analyzed on aFACSCalibur flow cytometer (BD Biosciences). Absolutenumbers per microliter of blood were calculated usingFlowcount beads from Beckman Coulter according toman-ufacturer instructions.

Rechallenge with EGFRvIIINEG tumorEither tumor na€�ve mice, tumor na€�ve mice receiving T

cells, or mice that had been cured of previous subcutaneousSMA560vIII tumors following treatment with EGFRvIIImCAR T cells were rechallenged with 5 � 105 SMA560(EGFRvIIINEG) tumors injected subcutaneously into the left(contralateral) flank.

ResultsGeneration of a third-generation, EGFRvIII-specificmCARTo evaluate CAR T-cell function in a syngeneic, immu-

necompetent brain tumor model, we chose the murineglioma cell line, SMA560, derived from an astrocytoma

that arose spontaneously in a VM/Dk mouse (37), in itsparental form(SMA560)or stably transfectedwithEGFRvIII(SMA560vIII; Fig. 1A). This astrocytoma mirrors humangliomas phenotypically and morphologically, and secretesimmunosuppressive cytokines, including TGF-bs secretedby human gliomas (38). To develop an appropriate mCAR,we combined murine signaling domains for CD8TM,CD28, 4-1BB, and CD3z (35) with a scFv isolated from ahuman antibody clone specific for EGFRvIII, mAb 139,inserted into the MSGV1 retroviral backbone (ref. 13; Fig.1B). The selection of mAb 139 was based on observationsthat this clone, compared with 6 others, could be consis-tently expressed in retrovirally transduced cells and recog-nize EGFRvIII antigenwith high avidity and specificity (34).Also, because 139 possesses known specificity for the nat-urally occurring EGFRvIII epitope in both mouse andhuman tumors, its incorporation into preclinical modelsoffers translational potential as well as biologic principle.

Next, we sought to determine the expression of theEGFRvIII mCAR on the surface of retrovirally transducedsplenocytes isolated fromVM/Dkmice. The same retrovirusencoding GFP was used to transduce control T cells. Flowcytometric analysis following transduction revealed expres-sion of GFP (data not shown), and surface expression of theEGFRvIII mCAR at 58% (MFI 328 vs. negative control of<1%,MFI 8; Fig. 1C). Importantly, when tested for reactivityagainst tumor cell lines, GFP control T cells did not display adetectable response to EGFRvIIIPOS SMA560vIII by IFN-gELISpot, whereas T cells transduced with the EGFRvIIImCAR yielded a significant number of spot-forming units(P < 0.05; Fig. 1D), which was found to be dependent onEGFRvIII expression on target tumor cells.

To test receptor specificity in vitro, we leveraged apreviously described EGFRvIII-derived soluble peptide,PEPvIII, that corresponds with the extracellular antigenicepitope of the EGFRvIII tumor mutation (33). We pre-coated tissue culture plates with PEPvIII or a negative-control peptide, and added EGFRvIII mCAR or vectorcontrol T cells. Only EGFRvIII mCAR T cells plated onPEPvIII showed reactivity by producing IFN-g , and thisreactivity was blocked up to 75% in a concentration-specific manner by adding increasing amounts of solublePEPvIII (P < 0.01), but not by the nonspecific IDH controlpeptide (Fig. 1E).

mCAR T cells show efficacy against intracerebraltumors that is dependent upon lymphodepletivepreconditioning

Recent clinical trials of TCR- and CAR-engineered T-celltherapy have used systemic delivery to treat widely dissem-inated disease (2, 3, 5, 15, 17, 39, 40). Previous studies havedemonstrated that adoptively transferred T cells adminis-tered intravenously can even treat intracerebral tumormetastases (2, 15, 41); however, in contrast to tumorsarising from central nervous system (CNS) parenchyma,metastatic lesions tend to be associatedwith vascular borderzones at the gray and white matter junction (42) and maynot accurately recapitulate challenges associated with






tumors arising from tissues originating behind the blood–brain barrier. To overcome the criticisms of previously usedxenograft models (26), we selected an immune-intactmurinemodel with a syngeneic, invasive glioma to evaluatethe preclinical efficacy of EGFRvIII mCAR T cells in vivo.

Previous studies have shown that the survival and anti-tumor function of adoptively transferred T cells in vivo canbe dependent upon lymphodepletive host conditioning(43–45). However, the impact of this effect has been under-appreciated for CAR T cells, given the few available immu-necompetent animal models for preclinical testing. Usingour syngeneic murine system, we sought to determine ifEGFRvIII mCAR T cells with or without lymphodepletivehost conditioning with TBI could demonstrate efficacyagainst tumors in the brain.

To investigate the ability of EGFRvIII mCAR T cells tolocalize and treat gliomas in the CNS, VM/Dk mice wereimplanted intracerebrally with SMA560vIII gliomas. Threeto five days later,mice were treatedwith or without 5Gy TBIto induce lymphodepletion, and were then injected intra-venously with either control or EGFRvIII mCAR T cells andmonitored for survival endpoints. In the absence of TBI hostconditioning, mice treated with EGFRvIII mCAR T cellsdid not display any survival benefit over mice receivingnegative control GFP T cells (Fig. 2A). However, signif-icantly prolonged survival was observed using the samenumber of EGFRvIII mCAR T cells, compared with GFPT cells in the setting of 5 Gy TBI (P < 0.01; Fig. 2B). Thus,in our syngeneic, immunocompetent mouse model, lym-phodepletion is shown to be a critical factor in promotingthe efficacy of CAR-modified T cells against tumors in theCNS.

EGFRvIII mCAR T-cell activity is abrogated by solublepeptide

Perhaps the greatest drawback of CAR T-cell therapy hasbeen the lethal toxicity resulting from interaction between Tcells and cognate antigen coexpressed in both tumors andnormal tissues (19). Thus, a great need exists for novelstrategies to reduce the toxicity profile of CARs targetingshared antigens on nontumor cells. As proof-of-concept, wesought to determinewhether a short peptide correspondingto the antigenic epitope expressed on target cells could beapplied to competitively inhibit and reduce the functionalactivity of EGFRvIII mCAR T cells in vivo, similar to theeffects we saw in vitro.

To perform antigen-specific blockade in vivo in ourmodelsystem, VM/Dk mice bearing intracerebral SMA560vIIItumors were irradiated and treated with EGFRvIII mCAR

58%

A

139 scFv LTRLTR

CD8TM, CD28, 4-1BB, CD3

sasd

SMA560 SMA560vIII0

200

400

600

800

EGFRvIII mCAR

Vector control

P = 0.0220

IFN

-g S

FU

per

1×

10

6 c

ells

C

0

200

400

600

800

1,000

Coated PEPvIII

+ 1

0 m

g s

PE

P ID

H

+ 1

00

mg s

PE

P ID

H

0

200

400

600

800

1,000

Coated PEPvIII

+ 1

0 m

g s

PE

PvII

I

+ 1

00

mg

sP

EP

vII

I

P = 0.0004

P = 0.0054

IFN

-g S

FU

per

1.2

5×

10

4 c

ells

IFN

-g S

FU

per

1.2

5×

10

4 c

ells

B

D

E

EGFRvIII (L8A4)

100

80

60

40

20

0100 101 102 103 104

EGFRvlll multimer-PE

% o

f M

ax

SMA560 SMA560vlll100

80

60

% o

f M

ax

% o

f M

ax

40

20

0

100

80

60

40

20

0100 101 102 103 104 100 101 102 103 104

Figure 1. EGFRvIII mCAR retroviral gene delivery tomurine T cells confersantigen-specific activity. A, VM/Dk-derived glioma SMA560vIII showsEGFRvIII expression compared with SMA560 by surface antibodystaining (L8A4) and flow cytometry; gray represents negative stainingcontrol. B, the EGFRvIII mCAR MSGV1 retroviral vector contains thehuman anti-EGFRvIII scFv 139, in tandem with murine CD8TM, CD28,4-1BB, andCD3z intracellular regions. EGFRvIIImCARMSGV1 retroviruscontaining supernatant was generated by transient transfection of HEK293T cells with the corresponding retroviral vectors and the helper pCL-Eco plasmid. Supernatants harvested 48 hours after transfection wereused to transduce previously ConA-activated murine VM/Dksplenocytes. C, retroviral mCAR gene expression was evaluated by flowcyometric analysis 3 days after transduction; gray represents negativestaining control. D, antigen specificity of EGFRvIII mCAR transduced andvector control (MSGV1-GFP) T cells was evaluated against EGFRvIII-expressing SMA560 cells (SMA560vIII) or the parental EGFRvIIINEG line(SMA560) by IFN-g ELISpot assay. Statistical analysis was performedby 2-way ANOVA. Horizontal bar represents a statistical significance

of P < 0.05. E, soluble EGFRvIII peptide (PEPvIII) at 50 mg/mLwas used tocoat tissue culture plates before addition of EGFRvIII mCAR or GFPcontrol T cells. Cells were cultured 18 hours in the presence of increasingdosages of soluble PEPvIII or irrelevant peptide (IDH) and subjected toIFN-g ELISpot. Statistical analysis was performed by unpaired t testcomparing groups defined by EGFRvIII target expression. Horizontal barrepresents a statistical significance of P < 0.05. Experiments arerepresentative of 2 independent repeats with similar results.

Sampson et al.





T cells followed by PEPvIII via intravenous infusion.Using only 2 separate injections of 100 mg PEPvIII, given10 days apart, we observed a significant reduction inEGFRvIII mCAR T-cell activity in vivo, demonstrated bya reduction in overall survival compared with animalsreceiving EGFRvIII-targeted T cells alone (P < 0.05; Fig.2C). These data support that soluble cognate peptide maybe used to selectively inhibit mCAR T cells in vivo. Inaddition, this selective blockade also provides evidencesupporting the antigen specificity of the 139 scFv EGFR-vIII mCAR.

EGFRvIII mCAR T cells treat 3- to 5-day established,syngeneic subcutaneous, and intracranial tumors in adose-dependent manner

To evaluate if treatment effect was dose-dependant, werepeated the intracranial glioma treatment experiment,using increasing numbers of EGFRvIII mCAR T cells. T-celltreatment numbers were based upon total number of cellsandwere transduced 50% to 75%. At the lowest dose (0.7�105), there was no apparent impact on survival over controlT-cell–treated mice. However, there was a significant, dose-dependent improvement in survival inmice receiving 3.5�106 to 1.0� 107 EGFRvIII mCAR T cells, and at the highestdose, all animalswere curedof intracerebral tumorswithoutapparent toxicity (P<0.0002; Fig. 3A). Taken together, thesedata demonstrate that systemic administration of EGFRvIIImCAR T cells can potently eliminate EGFRvIII-expressingtumors, even in the immunologically privileged CNS.

To follow tumor growth characteristics in the context ofCAR treatment, VM/Dk mice were implanted subcutane-ously with SMA560vIII, and tumor growth was recorded tomeasure treatment effect. Tumors in mice lymphodepletedwith 5 Gy TBI before receiving GFP control T cells did notexhibit significantly altered tumor growth compared withuntreated mice. However, lymphodepleted mice receiving

A

B

0 10 20 300

20

40

60

80

100

60

7.0×106 vector control

0.7×105 EGFRvIII mCAR




Days after tumor implantation

Pe

rce

nta

ge

su

rviv

al (%

)

P = 0.0002

0 10 20 300

100

200

300

400


Untreated

5 Gy TBI

5×106 Vector control 5 Gy TBI

5×106

EGFRvIII mCAR 5 Gy TBI

*5/8 complete regression

P < 0.0001

Tum

or

siz

e (

mm

2)

Figure 3. Adoptive T-cell therapy with EGFRvIII mCAR T cells leads toregression and eradication of subcutaneous and intracerebral syngeneictumors in a dose-dependent fashion. VM/Dkmice (8mice per group) with3-day established (A) intracerebral, or (B) subcutaneous SMA560vIIItumors were irradiated (5 Gy TBI) before intravenous infusion withvarious doses of vector control (MSGV1-GFP) or EGFRvIII mCART cells as indicated. Mice were monitored for morbidity endpointsapproved by the Duke University IACUC and sacrificed when endpointswere met. Survival analysis was performed using the log-rank(Mantel Cox) test. Tumor growth was monitored in 2 dimensions.Statistical significance was determined at P < 0.05 using 2-wayANOVA. Experiments are representative of 3 independent repeatswith similar results.

0 5 10 15 20 25 300

20

40

60

80

100A

B

C

3.5×106 Vector control, no TBI

3.5×106

EGFRvIII mCAR, no TBI

Days

Per

cen

tag

e su

rviv

al (

%)

0 5 10 15 20 25 300

20

40

60

80

100

3.5×106 Vector control, 5 Gy TBI

3.5×106

EGFRvIII mCAR, 5 Gy TBI

P < 0.01

Days

Per

cen

tag

e su

rviv

al (

%)

0 5 10 15 20 250

20

40

60

80

100Vector control 5 Gy TBI

EGFRvIII mCAR 5 Gy TBIEGFRvIII mCAR PEPvIII 5 Gy TBI

P < 0.05

Days

Per

cen

tag

e su

rviv

al (

%)

Figure 2. EGFRvIII mCAR T-cell therapy against intracerebral tumorsrequires lymphodepletive host conditioning for efficacy, and can beabrogated by soluble peptide administration. To evaluate the impact oflymphopenia on the antitumor efficacy of EGFRvIII mCAR T cells, VM/Dkmice (8 per group) with 3-day established intracerebral SMA560vIIItumors were left untreated (A, no TBI) or irradiated (B, 5 Gy TBI) beforeintravenous infusion with 3.5 � 106 vector control (MSGV1-GFP) orEGFRvIII mCAR T cells. C, to evaluate whether a short peptidecorresponding to the antigenic epitope expressed on target cells couldbe applied to competitively inhibit and reduce the functional activity ofEGFRvIII mCAR T cells in vivo, mice with 3-day implanted SMA560vIIItreated as in (B) above, were injectedwith 100mgPEPvIII or saline 4 hoursfollowing mCAR T-cell infusion intravenously, followed by a secondinjection 10 days later. Mice were monitored for morbidity endpointsapproved by the Duke University IACUC and sacrificed when endpointswere met. Survival analysis was performed using the log-rank (MantelCox) test. Statistical significance was determined at P < 0.05.






EGFRvIII mCAR T cells exhibited compelling treatmenteffects with subcutaneous tumor growth completely elim-inated in 5 of 8 mice (P < 0.0001; Fig. 3B).

EGFRvIII mCAR T cells show long-term persistence invivo

T-cell persistence was evaluated in peripheral blood lym-phocytes (PBL) of 5-Gy–irradiated mCAR-treated micemore than 35 days, and mCAR T cells were detected at allcollection timepoints. Over time, the percentage of circu-lating PBL thatwere EGFRvIIImCAR-positive declined from4.5% after 7 days to 1.5% on day 35, with the steepestdecline occurring between days 14 and 28. (Fig. 4A). As thisis the same time period in which endogenous lymphocyteswould be expected to begin recovering from TBI in irradi-ated mice, we also assessed the total absolute number ofmCAR T cells in the periphery. The total numbers ofcirculating EGFRvIII mCAR T cells in mice did not changesignificantly between weeks 1 to 5 after infusion, maintain-ing approximately 15 to 25 per mL of blood volume (Fig.4B). The same trend was observed in mice receiving GFPcontrol T cells (data not shown). This suggested that theoverall number of transferred EGFRvIII mCAR T cells weremaintained in vivo over time, remaining relatively constant,whereas endogenous lymphocytes recovered and increasedin numbers in the periphery, reducing the percentage ofcirculating EGFRvIII mCAR T cells, but not the totalamount.

EGFRvIII mCAR treatment of mice with EGFRvIII-expressing tumors provides long-term protectionagainst rechallenge and EGFRvIII tumor-antigen loss

Substantial evidence supports that immune-based ther-apies have the ability to eradicate tumors based on theexpression of specific antigens (7). However, a limitationof current strategies is that they often result in tumor

escape owing to heterogeneous antigen expression; indeed,although EGFRvIII peptide vaccines have been shown tosuccessfully eliminate EGFRvIII-expressing tumor cells inmice and humans, outgrowth of EGFRvIIINEG cells ulti-mately leads to tumor recurrence (33).

One suggested mechanism by which T-cell–based thera-pies in particular may be able to protect against tumorimmune escape is through the triggering of epitope spread-ing, initiated by a localized endogenous immune responseby the release of inflammatory cytokines at the tumor. Theresulting recruitment of the endogenous immune system inthe face of tumor destruction may confer the ability togenerate new responses to tumor cells that do not expressthe original target antigen. To date, the ability of third-generation CAR T cells to effectively initiate this activityremains unknown, again owing to the relative dearth ofavailable preclinical models with fully competent immunesystems capable of mounting such a response.

We sought todeterminewhether treatmentwith EGFRvIIImCAR T cells could elicit de novo priming against additionaltumor antigens, which could then serve as protectionagainst tumor-antigen loss. Specifically, we wanted to testwhether mice that were previously treated with mCARscould acquire protective immunity against tumors that didnot contain the original targeted antigen of interest, in thiscase, EGFRvIII. To do this, we challenged mice—eithertumor na€�ve mice with no treatment, or those previouslygiven EGFRvIII mCAR T cells and cured of SMA560vIIItumor. In the control group, tumors grew rapidly and allreached humane size endpoints within 40 days. In contrast,previously cured mice were completely protected againstrechallenge with EGFRvIIINEG tumor (Fig. 5A).

The ability of EGFRvIIINEG glioma cell lines to expressEGFRvIII upon transfer in vivo has been recently demon-strated (46). We sought to rule out the possibility that thiswas the reason for protection against EGFRvIIINEG tumor

7 14 21 28 35

0

1

2

3

4

5

6A B

Days after adoptive transfer

% E

GF

RvIII m

CA

R T

cells

7 14 21 28 35

0

10

20

30

Days after adoptive transfer

EG

FR

vII

I m

CA

R T

ce

lls

pe

r mL

of

blo

od

Figure 4. Constant numbers ofEGFRvIII mCAR T cells persisted inPBL more than 5 weeks whileendogenous lymphocytesrecovered. Constant numbers ofEGFRvIII mCAR T cells persisted inPBL more than 4 weeks. VM/Dkmice irradiated with 5 Gy (N ¼ 7)had PBL sampled 1, 2, 4, and 5weeks after intravenous EGFRvIIImCAR T-cell transfer. PBL werestainedwithmAb against CD3, andPEPvIII-PE multimer was used toidentify EGFRvIII mCAR T cells.Absolute numbers per microliter ofblood were evaluated by flowcytometry using Flowcount beadsfrom Beckman Coulter accordingto manufacturer instructions.Percentages (A) and absolutenumbers (B) of EGFRvIII mCAR Tcells in blood were calculated.

Sampson et al.





challenge. To evaluate whether SMA560 upregulates EGFR-vIII in vivo, we implanted SMA560 tumors subcutaneouslyin mice. Fourteen days later, we excised the tumors, stained

them for EGFRvIII expression, and performed a flow cyto-metric analysis. Figure 5B shows that although the EGFRvIIIexpression did increase slightly, from 0.18% in vitro to

0 20 40 60 80

0

100

200

300

400

Naïve

Rechallenge


Tu

mo

r siz

e (

mm

2)

A

In vitro

SMA560 (EGFRvIIINeg)

Ex vivo

B

C

Tumor Naïve0.0

0.5

1.0

1.5

2.0

% E

GF

RvIII m

CA

R T

cells

Tumor Naïve0

5

10

15

20

25

EG

FR

vIII m

CA

R T

cells

per

µL o

f blo

od

Before After0

5

10

15

20

25

P = 0.0263

EG

FR

vIII m

CA

R T

cells

per

µL o

f blo

od

Before After0.0

0.5

1.0

1.5

2.0

P = 0.0125

% E

GF

RvIII m

CA

R T

cells

0.18%

Co

un

t

Co

un

t 1.20%

Tumor naïve Tumor curedD E

0 10 20 30 40

0

100

200

300

400


Untreated: tumor naïve

Vector control T cells: tumor naïve

EGFRvIII mCAR T cells: tumor naïve

EGFRvIII mCAR T cells: tumor cured

P < 0.0001Tu

mo

r siz

e (

mm

2)

EGFRvIII (L8A4)

300

200

100

100 101 102 103 104 100 101 102 103 1040

40

30

20

10

0

Figure 5. Treatment of mice bearing EGFRvIIIPOS tumors with mCARs provides long-term protection against EGFRvIIINEG tumor. A, to evaluate whethertreatment with EGFRvIII mCAR T cells could elicit de novo priming against antigen-loss variants, long-term survivors (greater than 60 days after treatment) ofsubcutaneous SMA560vIII tumors by EGFRvIII mCAR therapy were rechallenged with 5 � 105 EGFRvIIINEG SMA560 cells on the contralateral flankand their survival was monitored. B, SMA560 (EGFRvIIINEG) gliomas were maintained in vitro or passaged in a mouse subcutaneously for 14 days, thenharvested, and both were stained for EGFRvIII expression using L8A4 mAb and analyzed by flow cytometry. Total numbers and percentages of mCAR Tcells in blood were evaluated in (C) tumor-naïve mice, and (D) SMA 560vIII (EGFRvIIIPOS) tumor-cured mice before and 7 days after subcutaneous challengewithSMA560 (EGFRvIIINEG) tumors. Statistical significancewasdeterminedusing apaired t test. Horizontal bar represents a statistical significanceofP<0.05.C–E, tumor naïve VM/Dk mice were either left untreated or received vector control or EGFRvIII mCAR T cells 1 day after receiving a lymphodepletive-conditioning regimen by 5 Gy TBI. All mice were challenged subcutaneously with SMA560 (EGFRvIIINEG) tumors 28 days after adoptive T-cell transfer. Inaddition, VM/Dkmice that were previously inoculated with 5� 105 SMA560vIII (EGFRvIIIPOS) tumor cells andwere cured of tumor after receiving 5Gy TBI andEGFRvIIImCARTcells (tumor cured)were rechallenged subcutaneouslywith 5�105SMA560 (EGFRvIIINEG) tumor cells at the same time. E, tumor growthwasmonitored in 2 dimensions. Mice were examined for morbidity endpoints approved by the Duke University IACUC and sacrificed when endpoints were met.Experiment is representative of 2 independent repeats with similar results. Statistical significance of P < 0.0001 was determined using 2-way ANOVA.






1.20% post passaging in vivo, the majority (>98%) oftumor cells did not exhibit EGFRvIII expression, making itunlikely to be the cause of tumor rejection.

To further evaluate the possibility of EGFRvIII antigenexpression in vivo on SMA560, we looked for mCAR T-cellproliferation in vivo in response to SMA560 challenge. Weused groupsofmice that had received EGFRvIIImCART cellsin the absence of tumor, and mice that had received thesame T cells in the presence of SMA560vIII, and had suc-cessfully cleared the tumor. Both T-cell–injected tumorna€�veand cured mice were found to have circulating EGFRvIIImCAR T cells in their blood 5 weeks after injection, beforeSMA560 challenge (Fig. 5C and D). Both groups thenreceived the EGFRvIIINEG SMA560 tumor implanted subcu-taneously, andwere re-evaluated for circulatingmCARTcells7 days later. In the tumor na€�vemice, there was no change inabsolute numbers or percentages of EGFRvIII mCAR T cellsbefore and after tumor challenge. However, unexpectedly,the cured mice showed a significant (P < 0.05) increase inboth absolute numbers andpercentages of circulating EGFR-vIII mCAR T cells after the antigen-negative tumor rechal-lenge. This was surprising, as we found no significantincrease in EGFRvIII expression on the tumor in vivo, andbecause EGFRvIII mCAR T cells alone were not sufficient toprotect the mice against SMA560 rechallenge (Fig. 5Eand Table 1). Importantly, these data suggest that the mech-anism responsible for rejection of the rechallenge parentaltumor in vivo may have included EGFRvIII mCAR T cells;however, additional elements were necessary for antitumorresponse. These results show that the mCAR T cells havesome response to antigen-negative tumor rechallenge in vivo,whether by direct tumor interaction, or indirectly by anothergeneralized immune stimulation in vivo. Together, theseresults suggest that third-generation, mCAR-modified T cellsmay contribute to the development of host immunityagainst tumor-antigen escape.

DiscussionSeveral groups have demonstrated that T cells can be

successfully redirected by CARs to induce potent antitumorimmune responses; however, unexpected toxicity has beenobserved because of shared target antigen expression onnormal tissues (11, 19). Because of its exquisite tumor spec-

ificity, EGFRvIII has been previously shown to be a promisingtarget for gene-modified T-cell therapy immune responsesboth in vitro and in xenogeneic, immunodeficient mousemodels (25, 34, 47). Our findings in this study demonstratethat a third-generation, EGFRvIII-specific mCAR can be usedto treat 3- to 5-day established intracerebral glioma in a fullyimmunocompetent, syngeneic murine model.

With 3 notable exceptions, CD19 (48, 49), vascularendothelial growth factor receptor (VEGFR)-2 (48), andfibroblast activation protein (FAP) (50), previous pre-clinical studies of CAR-redirected T cells have been per-formed in vitro or have been limited to in vivo studies againstxenografts in immune-deficient mice (25, 28–30, 51).Xenograft models, while enabling important experimentsusing human tumors and reagents, suffer from numerousshortcomings that have been reviewed in detail elsewhere(26). Briefly, because these mice lack physiologic immunesystems, a number of signaling ligands, effector molecules,and other circulating factors are absent that could other-wise impact T-cell function, trafficking, and persistence.Similarly, human tumor cells generally do not grow inmouse tissues as they do in patients, owing to severalspecies-specific stromal interactions and autologous growthfactors thought to promote highly aggressive and invasivecharacteristics of human tumors in situ. In this study, ourutilization of an immunocompetent mouse model withtumors derived from a syngeneic, spontaneously arisingglioma, addresses many of these limitations.

The utilization of our immunocompetent mousemodel revealed a number of findings that had previouslybeen impossible to evaluate in a xenograft setting. Onefinding was the need for lymphodepletive host precon-ditioning to achieve antitumor efficacy using adoptivelytransferred mCAR T cells. We speculate this may bebecause of a number of factors, including making "space"for the engineered lymphocytes to expand in the peri-phery, lack of cytokine sinks for lymphoid survival fac-tors, removal of circulating regulatory cells, and possibleincreased Toll-like receptor signaling in immune cells,induced by damaging radiation (52, 53).

Total mCAR T-cell numbers in PBL were maintainedin mice for more than a 35-day period, although percent-age of mCAR T cells diminished over time because of

Table 1. Eradication of SMA560vIII by EGFRvIII 139 CAR T cells is required for immunity against SMA560(EGFRvIII-negative) parental tumors

GroupNumberof mice

SMA560vIIIchallenge

CART cells

SMA560vIIIcures

SMA560challenge

SMA560cures

Untreated, SMA560vIII naïve 8 � � N/A þ 0/8Untreated, SMA560vIII exposed 15 þ � 0/15 N/A N/AmCAR T cells, SMA560vIII naïve 4 � þ N/A þ 0/4mCAR T cells, SMA560vIII exposed 15 þ þ 10/15 þ 10/10

NOTE: Day 0, mice received SMA560vIII tumors subcutaneously; day 5, mice received 5 Gy TBI followed by T cells intravenously; day28, mice were challenged with SMA560 parental tumors.

Sampson et al.





reconstitution of endogenous lymphocytes (Fig. 4). It isnot known what proportion of T cells may have migratedinto tumor, tissue, or lymphoid organs, however, so thismay be an underestimate of total mCAR T-cell retention inthe VM/Dk model. It remains unknown what happens tothe cells in vivo following ACT. A number of factors may,and likely do contribute, including body size, metabolicprocesses, extent of tumor infiltration, and potential sup-pressive, or immune activation factors in the periphery orat the tumor site. Although total numbers of survivingmCAR T cells in our mouse model are unknown, it ispossible that each CAR T-cell could function to kill manytumor cells, effectively generating "serial killers," as hasbeen reported clinically for the CD19 CAR in patients withleukemia (17, 54), suggesting that clinically, the quality ofthe cell may be more important than the quantity.One of the biggest clinical problems with glioblastoma is

its infiltrativemicrometastases. Although glioblastoma rare-ly metastasizes outside the brain, it does recur generally inthe area around the initial tumor excision, indicating thattumor cells remain after gross total resection. One unex-pected finding in our study is that mice treated with EGFR-vIII mCARs developed long-lasting immunity against syn-geneic EGFRvIIINEG tumors, suggesting that this approachmay be superior to prior peptide vaccine approaches thathave been limited by antigen escape (33). A major mech-anism by which tumors have been shown to evade theimmune system is through the process of immune editing,whereby tumor cells are favorably selected for the ability toeither mutate or downregulate a targeted antigen of interest(55). These phenomenae likely contributed to observationsin our previous clinical studies of an EGFRvIII peptidevaccine, in which multiple patients achieved significantsurvival benefits but eventually had tumor recurrence owingto the outgrowth of EGFRvIII antigen-loss variant tumors(33). Unlike the EGFRvIII vaccine, whichwas found to elicita predominantly humoral immune response, EGFRvIIICARs directly mediate a potent, highly avid T-cell response.Importantly, T-cell immunity in particular is known to playcentral roles in the promotion of endogenous primingassociated with epitope spreading (56), and is also capableof protecting against antigen loss through direct bystandereffects (57).We speculate that the expanded immune response we see

may be occurring because of mCAR T-cell interaction withcognate target-bearing tumors, causing T-cell stimulation,and resulting in proliferation and production of multiplecytokines, including the inflammatory IL-1, IL-6, and IL-8,plus TNF-a and IFN-g . These inflammatory cytokines areknown to have direct effects on tumors, including upregu-lation of antigen expression on tumor cells, and in additioncan recruit endogenous innate immune cells, includingprofessional antigen-presenting cells (APC) such as den-dritic cells and macrophages to the site of tumor destruc-tion. The combination of mCAR T cells causing tumor celllysis, recruitment of APC, and an inflammatory milieu ofpro-Type 1 immune cytokines would create an optimalenvironment for the priming and activation of dendritic

cells in the presence of tumor antigen. Previously referred toas a cytokine storm (58, 59), this "perfect storm" may besufficient to cause activated APCs to engulf and presenttumor-specific antigens to na€�ve T cells in situ, or at thedraining lymph nodes, resulting in priming of endogenousT cells against novel tumor antigens. Although we cannotrule out a possible immune response against foreign anti-gens found on the tumors (i.e., fetal calf serum from in vitroculture), this would likely require the same immune prim-ing process in vivo.

The development of endogenous antitumor (non-EGFR-vIII) immunity in our mCAR-treated mice was somewhatunexpected, as our mice were lymphodepleted, althoughnot myeloablated with radiation (5 Gy) at the time of theirmCAR T-cell transfer. Even more surprising was the in vivoexpansion of EGFRvIII mCAR T cells observed upon rechal-lenge with an EGFRvIIINEG tumor. It is unknown whetherthis increase in mCAR T cells was the result of proliferationinduced by cognate antigen binding, or if another mecha-nism allowed for expansion of these cells in vivo, such asbystander proliferation as a result of IL-2 productionby other activated T cells. We also postulate that the lym-phodepleted periphery may have provided an ideal envi-ronment for adoptively-transferred mCAR T cells withendogenous tumor-reactive TCRs, or newly emergent na€�velymphocytes from the bone marrow with endogenous TCRrecognition of tumor antigens, to selectively expand. With-out competition for progrowth cytokines or suppression byT regulatory and other suppressive cells, activation ofendogenous tumor-specific TCRs would provide a selectiveadvantage to those cells to proliferate and expand into thelymphopenic periphery. Given that recent clinical studieshave associated favorable clinical responses with the abilityto produce broad immunity to previously untargeted anti-genic determinants (60), further study about the ability forCARs to elicit these secondary responses should be anemerging priority.

The remarkable cytotoxic potential of CAR-based T-celltherapy is perhaps most unfortunately illustrated by thedramatic toxicities observed in clinical trials of CARs target-ing antigens coexpressed in both tumors and normal tis-sues. Thus, the selectionof apotential target antigenmust becarefully considered to reduce the risk of life-threateningautoimmunity. To our knowledge, the EGFRvIII-specificCAR represents the first truly tumor-specific application ofthe CAR platform, and promises to circumvent toxicitiesthat have been previously observed for genetically engi-neered T cells targeting shared antigens, including MART-1(5), gp100 (2), CEA (40), CAIX (11), ERBB2 (19), andCD19 (18, 39). However, in situations where these antigenscontinue to be targets for CAR therapy, our data presentedhere support that strategies utilizing soluble peptide block-ade may be further developed to provide a novel antidotefor reducing undesired CAR-mediated T-cell activationin vivo.

Although not directly addressed in this study, our im-munocompetent murine model could prove useful in theevaluation of endogenous suppressive factors that have






emerged as major impediments to successful antitumorimmunetherapy. Many of these interactions are physio-logically complex and as such are not adequately modeledby in vitro systems or immune-deficient animal models ofhuman tumor xenograft. Molecular interactions such asthose observed between PD-1, CTLA-4, and their respec-tive ligands (61), as well as the impact of suppressiveregulatory T cells and TGF-b in patients with glioma (62)have been cited as promising therapeutic targets for thereversal of tumor-associated immune suppression. Studiesdesigned to address these and other questions in immuno-competent animal models may provide useful alternativesand will certainly gain relevance as CAR-mediated thera-pies progress through clinical trials. Here, we begin toprovide evidence that immune-replete, preclinical modelscan reveal critical details for therapy that may informclinical trial design.

Together, these data support that third-generationEGFRvIII CAR T cells have the ability to treat establishedsyngeneic gliomas in the brain and may in additionconfer host immunity against tumor-antigen loss variantsthrough epitope spreading. As shown here, preclinicalmodels with competent immune systems represent anopportunity to greatly advance our understanding ofadoptive transfer therapies. Although it will be importantto evaluate treatment at later timepoints following tu-mor implantation to better mirror the clinical situation,EGFRvIII-targeted CAR T cells may provide a highlyspecific, promising therapeutic candidate for patientswith tumors in the CNS and are now in phase I clinicaltrials at the NCI Surgery Branch (NCT01454596, www.clinicaltrials.gov).

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: J.H. Sampson, L. Sanchez-Perez, J.E. Herndon II,L.A. JohnsonDevelopment of methodology: J.H. Sampson, L. Sanchez-Perez, S.A.Rosenberg, L.A. JohnsonAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): J.H. Sampson, L. Sanchez-Perez, C.M. Suryade-vara, D.J. Snyder, C.T. Flores, R.J. Schmittling, C.-T. Kuan, L.A. JohnsonAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): J.H. Sampson, B.D. Choi, L. Sanchez-Perez, R.J. Schmittling, S.K. Nair, E.A. Reap, J.E. Herndon II, C.-T. Kuan, L.A.JohnsonWriting, review, and/or revision of the manuscript: J.H. Sampson, B.D.Choi, L. Sanchez-Perez, E.A. Reap, P.K. Norberg, R.A. Morgan, S.A. Rosen-berg, L.A. JohnsonAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): J.H. Sampson, C.M. Suryadevara,D.J. Snyder, C.T. Flores, R.J. Schmittling, S.K. Nair, E.A. Reap, P.K. Norberg,C.-T. Kuan, L.A. JohnsonStudy supervision: J.H. Sampson, L. Sanchez-Perez, L.A. Johnson

AcknowledgmentsThe authors thank Dr. D. Bigner for his kind gift of the VM/Dk mice and

L8A4 mAb.

Grant SupportThis work was supported by funding from the NIH Director’s Office

New Innovator Award DP2CA174501, in addition to funding from VoicesAgainst Brain Cancer and the American Brain Tumor Association toL.A. Johnson. This work was also supported by an NIH NCI grant1R01CA177476-01 to J.H. Sampson.

The costs of publication of this article were defrayed in part by, thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received March 15, 2013; revised November 4, 2013; accepted November26, 2013; published OnlineFirst December 18, 2013.

References1. Stupp R, Hegi ME, MasonWP, van den BentMJ, TaphoornMJ, Janzer

RC, et al. Effects of radiotherapy with concomitant and adjuvanttemozolomide versus radiotherapy alone on survival in glioblastomain a randomised phase III study: 5-year analysis of the EORTC-NCICtrial. Lancet Oncol 2009;10:459–66.

2. JohnsonLA,MorganRA,DudleyME,CassardL,YangJC,HughesMS,et al. Gene therapy with human and mouse T-cell receptors mediatescancer regression and targets normal tissues expressing cognateantigen. Blood 2009;114:535–46.

3. Robbins PF, Morgan RA, Feldman SA, Yang JC, Sherry RM, DudleyME, et al. Tumor regression in patients with metastatic synovial cellsarcoma and melanoma using genetically engineered lymphocytesreactive with NY-ESO-1. J Clin Oncol 2011;29:917–24.

4. Rosenberg SA. Cell transfer immunotherapy for metastatic solid can-cer—what clinicians need to know.Nat RevClinOncol 2011;8:577–85.

5. Morgan RA, Dudley ME, Wunderlich JR, Hughes MS, Yang JC, SherryRM, et al. Cancer regression in patients after transfer of geneticallyengineered lymphocytes. Science 2006;314:126–9.

6. Dudley ME, Wunderlich JR, Robbins PF, Yang JC, Hwu P, Schwart-zentruber DJ, et al. Cancer regression and autoimmunity in patientsafter clonal repopulation with antitumor lymphocytes. Science 2002;298:850–4.

7. Restifo NP, Dudley ME, Rosenberg SA. Adoptive immunotherapy forcancer: harnessing the T cell response. Nat Rev Immunol 2012;12:269–81.

8. Rosenberg SA, Yang JC, Robbins PF, Wunderlich JR, Hwu P, SherryRM, et al. Cell transfer therapy for cancer: lessons from sequential

treatments of a patient with metastatic melanoma. J Immunother2003;26:385–93.

9. Gross G, Gorochov G,Waks T, Eshhar Z. Generation of effector T cellsexpressing chimeric T cell receptor with antibody type-specificity.Transplant Proc 1989;21:127–30.

10. HollymanD, Stefanski J, PrzybylowskiM,Bartido S, Borquez-OjedaO,Taylor C, et al. Manufacturing validation of biologically functionalT cells targeted to CD19 antigen for autologous adoptive cell therapy.J Immunother 2009;32:169–80.

11. Lamers CH, Sleijfer S, Vulto AG, Kruit WH, Kliffen M, Debets R, et al.Treatment of metastatic renal cell carcinoma with autologous T-lym-phocytes genetically retargeted against carbonic anhydrase IX: firstclinical experience. J Clin Oncol 2006;24:e20–2.

12. Milone MC, Fish JD, Carpenito C, Carroll RG, Binder GK, Teachey D,et al. Chimeric receptors containing CD137 signal transductiondomains mediate enhanced survival of T cells and increased antileu-kemic efficacy in vivo. Mol Ther 2009;17:1453–64.

13. Zhao Y, Wang QJ, Yang S, Kochenderfer JN, Zheng Z, Zhong X, et al.A herceptin-based chimeric antigen receptor with modified signal-ing domains leads to enhanced survival of transduced T lymphocytesand antitumor activity. J Immunol 2009;183:5563–74.

14. Zhong XS, Matsushita M, Plotkin J, Riviere I, Sadelain M. Chimericantigen receptors combining 4-1BB and CD28 signaling domainsaugment PI3kinase/AKT/Bcl-XL activation and CD8þ T cell-mediatedtumor eradication. Mol Ther 2010;18:413–20.

15. Pule MA, Savoldo B, Myers GD, Rossig C, Russell HV, Dotti G, et al.Virus-specific T cells engineered to coexpress tumor-specific

Sampson et al.





receptors: persistence and antitumor activity in individuals with neu-roblastoma. Nat Med 2008;14:1264–70.

16. Kochenderfer JN, Dudley ME, Feldman SA, Wilson WH, Spaner DE,Maric I, et al. B-cell depletion and remissions of malignancy alongwith cytokine-associated toxicity in a clinical trial of anti-CD19 chi-meric-antigen-receptor-transduced T cells. Blood 2012;119:2709–20.

17. Porter DL, Levine BL, Kalos M, Bagg A, June CH. Chimeric antigenreceptor-modified T cells in chronic lymphoid leukemia. N Engl J Med2011;365:725–33.

18. Brentjens RJ, Riviere I, Park JH, Davila ML, Wang X, Stefanski J, et al.Safety and persistence of adoptively transferred autologous CD19-targetedTcells in patientswith relapsedor chemotherapy refractoryB-cell leukemias. Blood 2011;118:4817–28.

19. MorganRA,Yang JC,KitanoM,DudleyME, LaurencotCM,RosenbergSA.Case report of a serious adverse event following the administrationof T cells transduced with a chimeric antigen receptor recognizingERBB2. Mol Ther 2010;18:843–51.

20. Wikstrand CJ, Hale LP, Batra SK, Hill ML, Humphrey PA, Kurpad SN,et al. Monoclonal antibodies against EGFRvIII are tumor specific andreact with breast and lung carcinomas and malignant gliomas. CancerRes 1995;55:3140–8.

21. Al-Nedawi K, Meehan B, Micallef J, Lhotak V, May L, Guha A, et al.Intercellular transfer of the oncogenic receptor EGFRvIII by microve-sicles derived from tumour cells. Nat Cell Biol 2008;10:619–24.

22. IndaMM,BonaviaR,MukasaA,Narita Y, SahDW,VandenbergS, et al.Tumor heterogeneity is an active process maintained by a mutantEGFR-induced cytokine circuit in glioblastoma. Genes Dev 2010;24:1731–45.

23. Boockvar JA,KapitonovD,KapoorG,Schouten J,CounelisGJ,BoglerO, et al. Constitutive EGFR signaling confers a motile phenotype toneural stem cells. Mol Cell Neurosci 2003;24:1116–30.

24. LammeringG,Hewit TH,HolmesM, Valerie K,HawkinsW, LinPS, et al.Inhibitionof the type III epidermal growth factor receptor variantmutantreceptor by dominant-negative EGFR-CD533 enhances malignantglioma cell radiosensitivity. Clin Cancer Res 2004;10:6732–43.

25. OhnoM, Natsume A, Ichiro Iwami K, Iwamizu H, Noritake K, Ito D, et al.Retrovirally engineered T-cell-based immunotherapy targeting type IIIvariant epidermal growth factor receptor, a glioma-associated antigen.Cancer Sci 2010;101:2518–24.

26. Morgan RA. Human tumor xenografts: the good, the bad, and the ugly.Mol Ther 2012;20:882–4.

27. Dotti G, Savoldo B, Brenner M. Fifteen years of gene therapy based onchimeric antigen receptors: "are we nearly there yet?". HumGene Ther2009;20:1229–39.

28. Ahmed N, Salsman VS, Kew Y, Shaffer D, Powell S, Zhang YJ, et al.HER2-specific T cells target primary glioblastoma stem cells andinduce regression of autologous experimental tumors. Clin CancerRes 2010;16:474–85.

29. Balyasnikova IV, Ferguson SD, Sengupta S, Han Y, Lesniak MS.Mesenchymal stemcellsmodifiedwith a single-chain antibody againstEGFRvIII successfully inhibit the growth of human xenograft malignantglioma. PLoS ONE 2010;5:e9750.

30. Kahlon KS, Brown C, Cooper LJ, Raubitschek A, Forman SJ, JensenMC. Specific recognition and killing of glioblastoma multiforme byinterleukin 13-zetakine redirected cytolytic T cells. Cancer Res 2004;64:9160–6.

31. Sampson JH, Ashley DM, Archer GE, Fuchs HE, Dranoff G, Hale LP,et al. Characterization of a spontaneous murine astrocytoma andabrogation of its tumorigenicity by cytokine secretion. Neurosurgery1997;41:1365–72; discussion 72–3.

32. Sampson JH, Crotty LE, Lee S, Archer GE, Ashley DM, Wikstrand CJ,et al. Unarmed, tumor-specific monoclonal antibody effectively treatsbrain tumors. Proc Natl Acad Sci U S A 2000;97:7503–8.

33. Sampson JH, Heimberger AB, Archer GE, Aldape KD, Friedman AH,Friedman HS, et al. Immunologic escape after prolonged progression-free survival with epidermal growth factor receptor variant III peptidevaccination in patients with newly diagnosed glioblastoma. J ClinOncol 2010;28:4722–9.

34. Morgan RA, Johnson LA, Davis JL, Zheng Z, Woolard KD, Reap EA,et al. Recognition of glioma stem cells by genetically modified T cells

targeting EGFRvIII and development of adoptive cell therapy forglioma. Hum Gene Ther 2012;23:1043–53.

35. Chinnasamy D, Yu Z, Theoret MR, Zhao Y, Shrimali RK, Morgan RA,et al. Gene therapy using genetically modified lymphocytes targetingVEGFR-2 inhibits the growth of vascularized syngenic tumors in mice.J Clin Invest 2010;120:3953–68.

36. Kerkar SP, Sanchez-Perez L, Yang S, Borman ZA, Muranski P, Ji Y,et al. Genetic engineering of murine CD8þ and CD4þ T cells forpreclinical adoptive immunotherapy studies. J Immunother 2011;34:343–52.

37. Serano RD, PegramCN, Bigner DD. Tumorigenic cell culture lines froma spontaneous VM/Dk murine astrocytoma (SMA). Acta Neuropathol1980;51:53–64.

38. Heimberger AB, Crotty LE, Archer GE, McLendon RE, Friedman A,Dranoff G, et al. Bonemarrow-deriveddendritic cells pulsedwith tumorhomogenate induce immunity against syngeneic intracerebral glioma.J Neuroimmunol 2000;103:16–25.

39. Kochenderfer JN, Wilson WH, Janik JE, Dudley ME, Stetler-Ste-venson M, Feldman SA, et al. Eradication of B-lineage cells andregression of lymphoma in a patient treated with autologousT cells genetically engineered to recognize CD19. Blood 2010;116:4099–102.

40. Parkhurst MR, Yang JC, Langan RC, Dudley ME, Nathan DA, FeldmanSA, et al. T cells targeting carcinoembryonic antigen can mediateregression of metastatic colorectal cancer but induce severe transientcolitis. Mol Ther 2011;19:620–6.

41. Hong JJ, Rosenberg SA, Dudley ME, Yang JC, White DE, Butman JA,et al. Successful treatment of melanoma brain metastases with adop-tive cell therapy. Clin Cancer Res 2010;16:4892–8.

42. HwangTL, Close TP,Grego JM,BrannonWL,Gonzales F. Predilectionof brain metastasis in gray and white matter junction and vascularborder zones. Cancer 1996;77:1551–5.

43. Dummer W, Niethammer AG, Baccala R, Lawson BR, Wagner N,Reisfeld RA, et al. T cell homeostatic proliferation elicits effectiveantitumor autoimmunity. J Clin Invest 2002;110:185–92.

44. Wrzesinski C, Paulos CM, Kaiser A, Muranski P, Palmer DC, GattinoniL, et al. Increased intensity lymphodepletion enhances tumor treat-ment efficacy of adoptively transferred tumor-specific T cells. JImmunother 2010;33:1–7.

45. Gattinoni L, Finkelstein SE, Klebanoff CA, Antony PA, Palmer DC,Spiess PJ, et al. Removal of homeostatic cytokine sinks by lympho-depletion enhances the efficacy of adoptively transferred tumor-spe-cific CD8þ T cells. J Exp Med 2005;202:907–12.

46. Del Vecchio CA, Giacomini CP, Vogel H, Jensen KC, Florio T, Merlo A,et al. EGFRvIII gene rearrangement is an early event in glioblastomatumorigenesis and expression defines a hierarchy modulated by epi-genetic mechanisms. Oncogene 2013;32:2670–81.

47. Bullain SS, Sahin A, Szentirmai O, Sanchez C, Lin N, Baratta E, et al.Genetically engineered T cells to target EGFRvIII expressing glioblas-toma. J Neurooncol 2009;94:373–82.

48. Santos EB, Yeh R, Lee J, Nikhamin Y, Punzalan B, Punzalan B, et al.Sensitive in vivo imaging of T cells using a membrane-bound Gaussiaprinceps luciferase. Nat Med 2009;15:338–44.

49. Kochenderfer JN, Yu Z, Frasheri D, Restifo NP, Rosenberg SA. Adop-tive transfer of syngeneic T cells transduced with a chimeric antigenreceptor that recognizes murine CD19 can eradicate lymphoma andnormal B cells. Blood 2010;116:3875–86.

50. Tran E, Chinnasamy D, Yu Z, Morgan RA, Lee CC, Restifo NP, et al.Immune targeting of fibroblast activation protein triggers recognition ofmultipotent bone marrow stromal cells and cachexia. J Exp Med2013;210:1125–35.

51. Carpenito C, Milone MC, Hassan R, Simonet JC, Lakhal M, SuhoskiMM, et al. Control of large, established tumor xenografts with genet-ically retargeted human T cells containing CD28 and CD137 domains.Proc Natl Acad Sci U S A 2009;106:3360–5.

52. RosenbergSA,RestifoNP,Yang JC,MorganRA,DudleyME.Adoptivecell transfer: a clinical path to effective cancer immunotherapy.NatRevCancer 2008;8:299–308.

53. Paulos CM, Wrzesinski C, Kaiser A, Hinrichs CS, Chieppa M, CassardL, et al. Microbial translocation augments the function of adoptively






transferred self/tumor-specific CD8þ T cells via TLR4 signaling. J ClinInvest 2007;117:2197–204.

54. GruppSA, KalosM, Barrett D, AplencR, Porter DL, Rheingold SR, et al.Chimeric antigen receptor-modified T cells for acute lymphoid leuke-mia. N Engl J Med 2013;368:1509–18.

55. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integratingimmunity's roles in cancer suppression and promotion. Science 2011;331:1565–70.

56. Pilon SA, Kelly C, Wei WZ. Broadening of epitope recognition duringimmune rejection of ErbB-2-positive tumorprevents growth of ErbB-2-negative tumor. J Immunol 2003;170:1202–8.

57. Zhang B, Karrison T, Rowley DA, Schreiber H. IFN-g- and TNF-dependent bystander eradication of antigen-loss variants in estab-lished mouse cancers. J Clin Invest 2008;118:1398–404.

58. Hombach AA, Holzinger A, Abken H. The weal and woe of costi-mulation in the adoptive therapy of cancer with chimeric anti-

gen receptor (CAR)-redirected T cells. Curr Mol Med 2013;13:1079–88.

59. D'Elia RV, Harrison K, Oyston PC, Lukaszewski RA, Clark GC. Target-ing the "cytokine storm" for therapeutic benefit. Clin Vaccine Immunol2013;20:319–27.

60. Corbiere V, Chapiro J, Stroobant V, Ma W, Lurquin C, Lethe B,et al. Antigen spreading contributes to MAGE vaccination-induc-ed regression of melanoma metastases. Cancer Res 2011;71:1253–62.

61. Pardoll DM. The blockade of immune checkpoints in cancer immu-notherapy. Nat Rev Cancer 2012;12:252–64.

62. Fecci PE, Mitchell DA, Whitesides JF, Xie W, Friedman AH,Archer GE, et al. Increased regulatory T-cell fraction amidsta diminished CD4 compartment explains cellular immune de-fects in patients with malignant glioma. Cancer Res 2006;66:3294–302.

Sampson et al.





2014;20:972-984. Published OnlineFirst December 18, 2013.Clin Cancer Res John H. Sampson, Bryan D. Choi, Luis Sanchez-Perez, et al. against Tumor-Antigen LossEstablished Intracerebral Glioma and Generates Host Immunity EGFRvIII mCAR-Modified T-Cell Therapy Cures Mice with

Updated version

10.1158/1078-0432.CCR-13-0709doi:

Access the most recent version of this article at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/4/972.full#ref-list-1

This article cites 62 articles, 26 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/4/972.full#related-urls

This article has been cited by 25 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://clincancerres.aacrjournals.org/content/20/4/972To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-13-0709

http://clincancerres.aacrjournals.org/content/20/4/972.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/4/972.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/20/4/972


egfrviii mcar-modiﬁed t-cell therapy cures mice with ... · egfrviii mcar-modiﬁed t-cell...

Documents