targetingfibroblastactivationproteinintumorstromawith ......matrix (7–11); and (iii) exerting an...

Research Article

Targeting Fibroblast Activation Protein in Tumor StromawithChimeric Antigen Receptor T Cells Can Inhibit Tumor Growthand Augment Host Immunity without Severe Toxicity

Liang-Chuan S. Wang1, Albert Lo4, John Scholler2, Jing Sun1, Rajrupa S. Majumdar4, Veena Kapoor1,Michael Antzis2, Cody E. Cotner1, Laura A. Johnson2, Amy C. Durham5, Charalambos C. Solomides6,Carl H. June2,3, Ellen Pur�e4, and Steven M. Albelda1

AbstractThe majority of chimeric antigen receptor (CAR) T-cell research has focused on attacking cancer cells.

Here, we show that targeting the tumor-promoting, nontransformed stromal cells using CAR T cells may offerseveral advantages. We developed a retroviral CAR construct specific for the mouse fibroblast activationprotein (FAP), comprising a single-chain Fv FAP [monoclonal antibody (mAb) 73.3] with the CD8a hinge andtransmembrane regions, and the human CD3z and 4-1BB activation domains. The transduced muFAP-CARmouse T cells secreted IFN-g and killed FAP-expressing 3T3 target cells specifically. Adoptively transferred73.3-FAP-CAR mouse T cells selectively reduced FAPhi stromal cells and inhibited the growth of multipletypes of subcutaneously transplanted tumors in wild-type, but not FAP-null immune-competent syngeneicmice. The antitumor effects could be augmented by multiple injections of the CAR T cells, by using CART cells with a deficiency in diacylglycerol kinase, or by combination with a vaccine. A major mechanism ofaction of the muFAP-CAR T cells was the augmentation of the endogenous CD8þ T-cell antitumor responses.Off-tumor toxicity in our models was minimal following muFAP-CAR T-cell therapy. In summary, inhibitingtumor growth by targeting tumor stroma with adoptively transferred CAR T cells directed to FAP can be safeand effective, suggesting that further clinical development of anti-human FAP-CAR is warranted. CancerImmunol Res; 2(2); 154–66. �2013 AACR.

IntroductionOne approach to adoptive T-cell therapy has been to trans-

fect patient-derived blood lymphocytes with chimeric antigenreceptor (CAR) genes that combine the effector functions ofT lymphocytes with the high specificity of single-chain anti-body fragments (scFv) that recognize predefined surface anti-gens in a non-MHC–restricted manner (1, 2). With recentmodifications and improvements, CAR T-cell therapy is show-ing great promise in the clinic (3–5).

To minimize on-target/off-tumor toxicity, target antigensshould be expressed at high levels on the surface of tumorcomponents with minimal or no expression on "essential"normal tissues. To date, potential targets have largely beenantigens on cancer cells, such as CD19, ErbB2, CEA, or GD2(3–6). We postulate that noncancer cell components of thetumor may also be attractive targets for CAR-based immu-notherapy. First, the genetic instability of tumor cells con-tributes to their immune escape, whereas stromal cells aremore genetically stable. Second, the tumor stroma contri-butes to tumor growth and resistance to therapy by (i)forming a physical barrier to tumor-targeting agents; (ii)supporting tumor cell growth, invasion, and angiogenesisthrough the production of growth factors, chemokines, andmatrix (7–11); and (iii) exerting an immunosuppressiveinfluence by secreting factors that attract immunosuppres-sive cells, by producing factors that regulate T-cell functionsand modulate the phenotype of myeloid cells (7, 12), and byexpressing inhibitory surface molecules PD-L1 and PD-L2(13). Finally, the mechanisms by which the cancer stromasupports tumorigenesis are shared among various stromalcell types, so that therapies targeting such mechanisms arelikely to have therapeutic implications across a broad spec-trum of cancers.

One attractive stromal cell target is the fibroblast activa-tion protein (FAP), a transmembrane serine protease highlyexpressed in the cancer-associated stromal cells (CASC) of

Authors' Affiliations: 1Thoracic Oncology Research Laboratory; 2Abram-son Family Cancer Research Institute; 3Department of Pathology andLaboratory Medicine, Perelman School of Medicine; Departments of 4Ani-mal Biology and 5Pathobiology, School of Veterinary Medicine, Universityof Pennsylvania; and 6Department of Pathology, Thomas Jefferson Uni-versity, Philadelphia, Pennsylvania

Note: Supplementary data for this article are available at Cancer Immu-nology Research Online (http://cancerimmunolres.aacrjournals.org/).

L.-C.S. Wang and A. Lo contributed equally to this work.

Corresponding Authors: Steven M. Albelda, Abramson Research Center,Room 1016B, University of Pennsylvania, 3615 Civic Center Boulevard,Philadelphia, PA19104.Phone: 215-573-9933; Fax: 215-573-4469; E-mail:[email protected]; and Ellen Pur�e, Department of Animal Biol-ogy, University of Pennsylvania, 3800 Spruce Street, 216E Vet, Philadel-phia, PA 19104. Phone: 215-573-9406; Fax: 215-573-4469; E-mail:[email protected]

doi: 10.1158/2326-6066.CIR-13-0027

�2013 American Association for Cancer Research.

CancerImmunology

Research

Cancer Immunol Res; 2(2) February 2014154

on August 19, 2020. © 2014 American Association for Cancer Research. cancerimmunolres.aacrjournals.org Downloaded from

Published OnlineFirst November 12, 2013; DOI: 10.1158/2326-6066.CIR-13-0027

http://cancerimmunolres.aacrjournals.org/

virtually all epithelial cancers (14–18). FAP is also express-ed during embryonic development, in tissues of healingwounds, and in chronic inflammatory and fibrotic condi-tions such as liver cirrhosis and idiopathic pulmonaryfibrosis (19–22). However, FAP has not been detected byimmunohistochemistry in benign tumors nor in most nor-mal quiescent adult stromal cells (23–25). FAPþ cells havebeen studied using Bac transgenic mice containing the FAPpromoter driving the GFP or the human diphtheria toxin(DT) receptor to conditionally eliminate FAPþ cells (26, 27).Although the elimination of FAPþ cells caused the decreasedgrowth of immunogenic OVA-expressing Lewis lung carci-noma tumors and augmented the efficacy of an OVA vaccine,the complete ablation of FAPþ cells was associated withanemia and cachexia, thought to result from the loss ofFAPþ stromal cells in the bone marrow and muscle. SomeFAP expression was also seen in the pancreas.Therapeutically, FAPþ cells have been successfully tar-

geted using vaccines and immunoconjugate therapy result-ing in some tumor growth inhibition without obvious toxicityor effects on wound healing (28–30). However, given theincreased efficacy of adoptive T-cell transfer, we and othershave hypothesized that anti-FAP-CAR T cells may provemore effective. Since the submission of this manuscript,three contrasting articles have been published describingthe use of anti-FAP-CAR T cells. Tran and colleagues, usinganti-FAP-CAR mouse T cells [comprising the scFv from theFAP-5 monoclonal antibody (mAb; ref. 30) that targetshuman and mouse FAP] in combination with total bodyirradiation and interleukin-2 (IL-2) injections, showed limit-ed antitumor efficacy that was associated with severe bonemarrow toxicity and cachexia (31). In contrast, Kakarla andcolleagues showed antitumor efficacy without toxicity usinganti-FAP-CAR human T cells [comprising the scFV from theM036 mAb (32) that targets both human and mouse FAP] inan immunodeficient mouse model of human lung cancer(33). Similarly, Schuberth and colleagues showed a survivaladvantage in an intraperitoneal mesothelioma xenograftmodel after administering FAP-CAR human T cells [usinga scFv from the human-specific F19 mAb (18)] together withFAP-expressing human mesothelioma cells (34). In this study,the CAR-mediated on-target/off-tumor toxicity could not beevaluated because the F19 anti-human FAP antibody doesnot cross-react with the mouse FAP.Given the recent data that FAP-expressing cells may play an

important role in regulating antitumor immunity (35), wedeveloped a system to study advanced-generation FAP-CART-cell therapy in wild-type (WT)mice with fully intact immunesystems. Using a different anti-FAP scFv (mAb 73.3) linked toboth human and murine cytoplasmic domains, we generatedmuFAP-CAR mouse T cells that elicited antitumor efficacymediated by the activation of endogenous immune responsesin multiple tumor models. The anti-FAP-CAR T cells demon-strated enhanced antitumor response when combined with atumor vaccine. Our FAP-CAR constructs induced minimaltoxicity with no anemia or weight loss, suggesting that furtherclinical development of anti-human FAP-CAR T cells may befeasible.

Materials and MethodsCell lines

Mouse AE17.ova mesothelioma cells (expressing chickenovalbumin) were provided by Delia Nelson (University ofWestern Australia, Crawley, WA, Australia; ref. 36). TC1 lungcancer cells were derived from mouse lung epithelial cellsimmortalized with human papillomavirus (HPV) HPV-16 E6and E7 and transformed with the c-Ha-ras oncogene (37). Themouse LKR cell line was derived from an explant of a pul-monary tumor from an activated K-rasG12D–mutant mousegenerated in the laboratory of Tyler Jacks at MIT (38). Mouse4T1 mammary carcinoma cells, CT26 colon cancer cells, and3T3Balb/C cells were purchased from the American TypeCulture Collection. Mouse FAP-expressing 3T3BALB/C(3T3.FAP) cells were created by lentiviral transduction of theFAP-3T3 parental line with murine FAP (data not shown). Allcell lines used were checked for Mycoplasma; except for theprotein expression of the transgenes, no additional authen-ticationwas performed. See SupplementaryMethods formoredetails.

AntibodiesSpecific antibodies used are listed in the Supplementary

Methods.

Generation of 73.3 hybridomaFAP-null mice (39) were immunized and twice boosted with

FAP-expressing 3T3 cells intraperitoneally. Three days after thefinal boost, splenocytes were harvested and fused to myelomacells. Hybridoma supernatants were screened for mAbs thatreacted specifically with 3T3.FAP cells and activated primaryWT fibroblasts, but not the parental 3T3 cells or the activatedprimary fibroblasts from FAP-null mice. mAb 73.3 with thisreactivity profile was purified and characterized by ELISA andby immunoblotting as specific for mouse FAP based on thereactivity with the purified recombinant mouse FAP extracel-lular domain (rFAP-ECD), its reactivity with a single proteinwith an apparent molecular weight of approximately 95 kDain the protein extracts of FAPþ but not FAP� cells and tissues,and a lack of reactivity with cells expressing human FAP[human FAP transduced 3T3, human foreskin fibroblasts, orprimary human tumor-associated stromal cell (data notshown)]. Biacore studies (see Supplementary Methods)showed that the affinity of mAb 73.3 for purified rFAP-ECDproteinwas less than 1 nmol/L and estimated to be between 0.1and 0.3 nmol/L (data not shown).

Generation of anti-muFAP-CAR constructsTotal RNA from 73.3 hybridoma cells was isolated and

reverse transcribed into cDNA. Variable heavy (VH) and light(VL) chains of the 73.3 immunoglobulin were PCR-amplifiedand sequenced (Fig. 1A and B). The VH and VL sequences werefused with a CAR construct that is being evaluated in clinicaltrials [CD8a hinge, CD8a transmembrane domain, and twohuman intracellular signaling domains (ICD) derived from4-1BB and CD3z; ref. 40]. This CAR was then inserted intoan internal ribosome entry site (IRES)-containing retroviral

FAP-Redirected CAR T Cells

www.aacrjournals.org Cancer Immunol Res; 2(2) February 2014 155




MigR1 vector (Fig. 1C) that also expresses GFP for trackingpurposes (41). A fully mouse construct of FAP-CAR and FAP-CAR-m28z was also created by coupling the same 73.3 scFvwith the murine CD3z chain and the murine CD28 ICD. Thisconstruct was inserted into another retroviral vector MSGV(Fig. 1D; ref. 42). Infective particles were generated fromthe supernatants of 293T cells transfected with the retro-viral vector plasmid and helper plasmids using Lipofectamine2000 (Invitrogen), as described previously (42, 43). Twocontrol FAP-CAR constructs, one with human CD3z andthe other with mouse CD3z ICDs, were also generated toevaluate the function of the costimulatory domains (seeSupplementary Methods).

Isolation, transduction, and expansion of primarymouse T lymphocytes

Primarymurine splenic T cells were isolated and transducedas previously described (43). See Supplementary Methods.

Antigen- or antibody-coated beadsrFAP-ECD (see Supplementary Methods), bovine serum

albumin (BSA; Fisher Scientific), or anti-CD3e/anti-CD28 anti-bodies (eBioscience) were chemically cross-linked to tosylac-tivated 4.5 mm Dynabeads (Invitrogen; #140-13) as per themanufacturers' instructions.

ImmunoblottingFAP-CAR–transduced T cells were incubated either with

BSA or FAP-ECD–coated beads (at 2:1 bead to T-cell ratio),or with anti-CD3e antibody for 10 minutes. Lysates werethen prepared and immunoblotted for phosphorylatedextracellular signal–regulated kinase (ERK), phosphorylatedAKT, phosphorylated inhibitor of IkB kinase (IKK)-a/b, orb-actin.

Cytotoxicity and IFN-g ELISAParental 3T3 and 3T3.FAP cells were transduced with lucif-

erase as described previously (44). T cells and target 3T3 cellswere cocultured at the indicated ratios, in triplicate, in 96-wellround-bottomed plates. After 18 hours, the culture superna-tants were collected for IFN-g analysis using an ELISA (mouseIFN-g ; BD OpEIA). Cytotoxicity of transduced T cells wasdetermined by detecting the remaining luciferase activity fromthe cell lysate using a previously described assay (43).

CART-cell transfer intomice bearing established tumorsMice were injected subcutaneously with 2 � 106 AE17.ova

(C57BL/6 mice), 1 � 106 TC1 (C57BL/6 mice), 2 � 106 LKR(C57BL/6 crossed with 129 pf/j), 0.5 � 106 4T1 (BALB/c mice),or 1 � 106 CT26 (BALB/c mice) tumor cells into the dorsal–lateral flank. Mice bearing established tumors (100–150 mm3)were randomly assigned to receive either FAP-CAR T cells orMigR1-transduced T cells or remained untreated (minimum,5 mice per group, each experiment repeated at least once). Atotal of 1 � 107 T cells was administered through the tail vein.Body weight and tumor size were measured by electronic scalesand calipers, respectively. At the end of the experiment, tumorsand spleens were harvested for flow-cytometric analyses.

Flow-cytometric analysesTumors were harvested 3 and 8 days after adoptive transfer

of FAP-CAR T cells to analyze intratumoral cells by flowcytometry as described previously (42). Cell acquisition wasperformed on LSR-II using FACSDiva software (BD Biosci-ence). Data were analyzed using FlowJo (TreeStar).

Statistical analysesFor flank-tumor studies comparing two groups, the Student

t test was used. For comparisons of more than two groups, we

AVH

BVL

QVQLKESGGLVQPGGSLKLSCAASGFTFSSYGMSWVRQTADKRLELVATTNNNGGVTYYPDSVKGRFTISRDNAKNTLYLQMSSLQSEDTAMYYCARYGYYAMDYWGQGISVTVSS

DVLMTQTPLWLPVSLGDQASISCRSSQSIVHSNGNTYLEWYLQKPGQSPKLLIYKVSNRFSGVPDRFSGSGSGTDFTVKISRVEAEDLGVYYCFGGSHVPYTFGGGTKLEIK

CRetroviral

constructs

with

human ICDs

DRetroviral

construct

with

mouse ICDs

73.3-hBBz (FAP-CAR vector)IRES MigR1

backboneVH VL 4-1BB CD3ζ GFPGFPCD8α hinge CD8α TMCD8α leader

73.3-hzVH VL CD3ζCD8α hinge CD8α TMCD8α leader

73.3-m28z (FAP-CAR-m28z vector)MSGVbackbone

VH VL CD28 CD3ζCD8α hinge CD28 TMCD8α leader

73.3-mzMSGVbackbone

VH VL CD3ζCD8α hingeCD8α leader CD8α TM

IRES MigR1backbone

GFPGFP

Figure 1. Structure of FAP-CAR.Total RNA of 73.3 hybridoma wasextracted, reverse transcribed, andthe cDNA PCR-amplified andinserted into a cloning vector toobtain the sequence of the variabledomains of immunoglobulinG (IgG)heavy (A) and light (B) chains. Theanti-muFAP-CAR consists of theanti-muFAP scFv, CD8a hinge,and transmembrane (TM) domain,plus 4-1BB and CD3z ICDs, andwas cloned into the MigR1retroviral vector for transduction ofprimary mouse T cells (C). A fullymouse FAP-CAR construct wasalso synthesized,which consists ofthe anti-muFAP scFv, CD8a hinge,andCD28 transmembrane domain,plus CD28 and CD3z ICDs ofmouse origin, and was cloned intothe MSGV retroviral vector totransduce primary mouse T cells(D). Two first-generation FAP-CARconstructs with a human CD3z (C)or a mouse CD3z (D) signalingdomain were also generated.

Wang et al.

Cancer Immunol Res; 2(2) February 2014 Cancer Immunology Research156




used one-way ANOVA with appropriate post hoc testing. Dif-ferences were considered significant when P value was lessthan 0.05. Data are presented as mean � SEM.

ResultsIn vitro evaluation of mouse FAP-CAR T cellsOur primary retroviral CAR construct (containing the scFv

from anti-murine FAP antibody 73.3 coupled to the humanCD3z and 4-1BB cytoplasmic domains that we have usedpreviously in murine models; ref. 42) and a control virusexpressing only GFP (Fig. 1) were used to transduce acti-vated mouse T cells resulting in greater than 60% of the Tcells expressing GFP (MigR1) or GFP plus FAP-CAR (Fig. 2A).To verify functionality, mouse T cells expressing FAP-CAR

were stimulated for 18 hourswith beads coatedwith either BSA(negative control), or recombinant FAP protein, or anti-CD3/anti-CD28 antibodies (positive control). The FAP-coated beadsactivated FAP-CAR T cells, as shown by increased CD69expression above that of the negative control (Fig. 2B).To further evaluate intracellular signaling, lysates from

bead-stimulated T cells were electrophoresed and immuno-blotted. In comparison with BSA-coated beads, FAP-coatedbeads induced the phosphorylation of AKT, ERK, and IKK-a/bin FAP-CAR T cells (Fig. 2C).

To assess effector functions, transduced mouse T cells werecocultured with 3T3 fibroblasts (which do not express FAP) orwith 3T3 fibroblasts transduced to express mouse FAP (3T3.FAP; data not shown). After 18 hours, T cells expressing theFAP-CAR construct (but not the control GFP construct) effec-tively killed 3T3.FAP fibroblasts (Fig. 2D) and secreted IFN-g(Fig. 2E) in a dose-dependent manner, but had no effect onparental 3T3 cells.

Injection of mouse FAP-CAR T cells reduces tumorgrowth in a FAP-specific fashion

We next explored the capability of FAP-CAR mouse Tcells to inhibit tumor growth using three different tumorlines that do not express FAP: AE17.ova mesothelioma cells,TC1, and LKR lung cancer cells. Cells were injected into theflanks of syngeneic mice and allowed to form establishedtumors. The tumors had an easily detectable number ofmouse FAP-expressing cells with the majority of the FAPþ

cells being CD45�/CD90þ stromal cells (�3% of total tumorcells), and only a small minority being CD45þ hematopoie-tic cells (�0.2% of total tumor cells; Table 1 and Supple-mentary Fig. S1).

When tumors reached approximately 100 to 150 mm3 (7–14 days after tumor cell inoculation), 107 T cells wereinjected intravenously and the tumors were measured with

MigR1 FAP-CARA

BBSA FAP

Anti-CD3Anti-CD28

CD8

CD4

CD69

pAKT

pERK

β-Actin

pIKK-α

C BSA FAP Anti-CD3

D

E

0

10

20

30

40

50

60

70

% C

yto

lys

is

Effector:target ratio

MigR1:3T3

FAP-CAR:3T3

MigR1:3T3.FAP

FAP-CAR:3T3.FAP

0

200

400

600

800

1,000

IFN

-γ (p

g/m

L)

Effector:target ratio

MigR1:3T3

FAP-CAR:3T3

MigR1:3T3.FAP

FAP-CAR:3T3.FAP

*

**

*

**

*

* * * *

Figure 2. Ex vivo assessment ofmouse CAR T cells redirectedagainst FAP and signaling inFAP-CAR T cells. A, retroviraltransduced mouse T cellsexpressed GFP (MigR1) or GFPand anti-mFAP-CAR. B,upregulation of CD69 on CAR(GFP)-positive CD8 and CD4 Tcells was determined followingstimulation with BSA- or FAP-coated beads for 18 hours. T cellswere stimulated with anti-CD3/anti-CD28 beads as positivecontrol. C, FAP-CAR T cells wereexposed to either BSA- or FAP-coated beads for 10 minutes. Celllysates were then prepared andimmunoblotted for phospho-ERK(pERK), phospho-AKT (pAKT), andphospho-IKK-a/b (pIKK-a/b).Anti-CD3e antibody was usedas a positive control for T-cellactivation, and b-actin wasimmunoblotted for equal loading.To determine target-specificcytolytic activity (D) and IFN-gproduction (E) of FAP-CAR T cells,various effector:target ratios ofMigR1 and FAP-CAR T cells werereacted with parental 3T3 or 3T3.FAP fibroblasts for 18 hours.�, Statistical significance betweenFAP-CAR–treated 3T3.FAPgroup versus the other threegroups; P < 0.05.






calipers. FAP-CAR T cells, but not MigR1 T cells, significantly(P < 0.05) reduced the growth of TC1 tumors (Fig. 3A), LKRtumors (Fig. 3B), and AE17.ova tumors (Fig. 3C) by 35% to50%.

To confirm specificity, we inoculated AE17.ova cells intoFAP-null C57BL/6 mice and treated the tumors as describ-ed above. In contrast with the effect on AE17.ova tumors

in WT C57BL/6 mice (Fig. 3C), FAP-CAR T cells had noeffect on the growth of AE17.ova tumors in FAP-null mice(Fig. 3D).

Given the differences between our efficacy data and thoseof Tran and colleagues' data (31), we also treated two ofthe same tumor lines, CT26 and 4T1, they reported. In con-trast to their findings, our FAP-CAR construct induced a

Table 1. Depletion of FAPþ cells in flank tumors after FAP-CAR treatment

AE17.ova TC1 LKR

Untreated FAP-CAR P Untreated FAP-CAR P Untreated FAP-CAR P

CD45-CD90þ 1.86 1.15 0.001a 1.72 0.60 0.04a 2.68 1.84 0.009a

CD45þ 0.12 0.08 0.157 0.10 0.04 0.066 0.28 0.15 0.035a

Average tumorvolume, mm3

627 324 0.03a 379 140 0.002a 410 260 0.02a

NOTE: The values above indicate the averages of percentage FAPþ cells per total tumor population. Tumors were harvested 7–9 daysafter T-cell infusion, and therewere 5mice in eachgroup. Studentpaired t testwasperformed to evaluateFAPþcell depletion, aswell aschange in tumor volume, after adoptive T-cell therapy.aStatistical significance between untreated and FAP-CAR–treated samples; P < 0.05.

Figure 3. Antitumor activities of FAP-CAR T cells in mice bearing flank tumors. Syngeneic mice bearing (A) TC1, (B) LKR, (C) AE17.ova, (E) CT26, and (F)4T1 tumors were injected intravenously with 10 million FAP-CAR or MigR1 T cells when the tumors reached approximately 100 to 150 mm3. Tumormeasurements followed. D, to test the target specificity of FAP-CAR T cells, AE17.ova tumor cells were also injected into FAP-null C57BL/6 mice. FAP-CART cells were given 7 days later. �, Statistical significance between untreated, MigR1-, and FAP-CAR–treated samples; P < 0.05.

Wang et al.





significant reduction in tumor size (Fig. 3E and F), althoughthe changes were smaller than those seen in Fig. 3A–D.

Effect of the injection of mouse FAP-CAR T cells on FAPþ

cellsToevaluate the effect of theT cells on the FAPþ stromal cells,

we harvested tumors 7 to 9 days after T-cell infusion andanalyzed the dissociated cells by flow cytometry. As shownin Table 1, at this time point, the FAPþ/CD45�/CD90þ andFAPþ/CD45þ populations were decreased by about 50% incomparison with those in the untreated group, whereas theamount of FAPþ cells remaining in the MigR1 group wassimilar to that in the untreated controls (data not shown).We characterized this depletion in more detail using

the AE17.ova model and evaluated the FAPþ cells 3 days afterT-cell transfer. At this earlier time point, we saw largerdecreases in FAPþCD90þ stromal cells (a reduction of 82%)and FAPþCD45þ leukocytes (a reduction of 56%; Supplemen-tary Fig. S2A). Moreover, in control animals, we could identifyboth low- and high-FAP–expressing cells in both the CD45�

CD90þ and CD45þ populations (Supplementary Fig. S1).Whenwe gated on these specific populations, we noted that the FAP-CAR T cells selectively depleted the FAP-high–expressing cells,with little effect on FAP-low–expressing cells (SupplementaryFig. S2B and S2C).

Kinetics of FAP-CAR T-cell persistenceWe assessed the number of intratumoral FAP-CAR T cells in

the AE17.ova model at 3, 7, and 10 days after adoptive transferand found that the number peaked at day 3 after injection anddiminished at the 7 and 10 day time points by approximately65% (Supplementary Fig. S3A).To determine whether this rapid loss of T cells was a

consequence of abnormal function of the human 4-1BB andCD3z cytoplasmic domains within the mouse T cells, weengineered a second construct by inserting our scFv anti-FAP73.3 antibody fragment into a fully murine CAR containingthe murine CD28 cytoplasmic domain and the murine CD3zchain (73.3m28z; Fig. 1D). In vitro, this construct showed simi-lar cytotoxicity and IFN-g release when reacted with FAP-expressing fibroblasts compared with the human version ofthe same CAR (73.3-hBBz; Supplementary Fig. S3B and S3C).To confirm the importance of the costimulatory cyto-plasmic domains, we synthesized two additional FAP-CARconstructs that lack these costimulatory domains, i.e., 73.3-human CD3z and 73.3-mouseCD3z (Fig. 1C and D). Neitherthe mouse nor the human CD3z construct showed significantcytolytic activity or IFN-g production when reacted with3T3mFAP (Supplementary Fig. S3B and S3C).After the cells were injected into mice bearing AE17.ova

tumors, we found that the trafficking and persistence of thetwo types of second-generation FAP-CAR T cells, 73.3-hBBzand 73.3-m28z, were similar (Supplementary Fig. S3D) as weretheir antitumor efficacy (Supplementary Fig. S3E). These datashow that, compared with human CARs injected into immu-nodeficient mice, mouse CAR T cells injected into syngeneichosts persisted for a short time, despite the presence of thehuman or mouse costimulatory cytoplasmic domains.

Approaches to enhance FAP-CAR T-cell therapyBecause our mouse CAR T cells persisted for only a short

period of time in vivo, we hypothesized that giving a secondinfusion of FAP-CAR T cells might enhance therapeutic effi-cacy. AE17.ova tumor cells were injected into the flanks ofC57BL/6 mice. When the tumors reached approximately 100mm3, a first dose of FAP-CAR T cells was given intravenously.One week later, we randomly divided the FAP-CAR–treatedanimals into two groups, one treated with an additional doseof control MigR1 T cells (Single dose; Fig. 4A) and one treatedwith a second dose of FAP-CAR T cells (Double dose; Fig. 4A).At 2 weeks, tumors in mice given two doses of FAP-CAR T cellswere significantly smaller (P < 0.05) than those in mice givenonly one dose of FAP-CAR T cells.

Another, not mutually exclusive, explanation for the lack oftumor eradication might be the suboptimal CAR signaling inmouse T cells and/or functional suppression in the tumormicroenvironment. As we recently reported that mesothelin-targeted CAR mouse T cells deficient in the inhibitory enzymediacylglycerol kinase-z (DGKz) had enhanced effector func-tions in vitro and in vivo and increased persistence (43), wecompared the efficacy of comparably transduced FAP-CARsplenic T cells isolated from WT C57BL/6 versus DGKz-nullmice. DGKz-deficient FAP-CAR T cells were more efficient inlysing 3T3.FAP cells (Supplementary Fig. S4A) and in secretingIFN-g (Supplementary Fig. S4B) with retention of specificity invitro. The DGKz-deficient FAP-CAR T cells were also moreeffective (P < 0.05 on day 11) after being injected into AE17.ova-bearing mice (Fig. 4B). The increased efficacy was associatedwith greater persistence of theDGKz-knockout comparedwithWT FAP-CAR T cells (GFPþ cells; Supplementary Fig. S4C).Thus, the enhanced antitumor efficacy was likely due to bothincreased T-cell activity and to increased persistence.

Role of the acquired immune system in the efficacy ofFAP-CAR T cells

To evaluate the role of the acquired immune system inthe FAP-CAR T cell–mediated antitumor response, weinjected AE17.ova tumor cells into the flanks of either WTC57BL/6 or immunodeficient NSG mice, followed by oneinjection of 107 FAP-CAR T cells. AE17.ova tumors grewmore rapidly in NSG than in WT mice (Fig. 5A vs. B), reflec-ting the endogenous antitumor activity in WT mice thatwas lost in the NSG mice. In contrast to their efficacy inWT mice (Fig. 5A), the mouse FAP-CAR T cells had noantitumor effects on AE17.ova tumors in the immunode-ficient NSG mice (Fig. 5B). This loss in activity was not due tothe loss of FAP expression in the NSG tumor microenviron-ment, as we confirmed that AE17.ova tumors had a similarlevel of FAP expression in the NSG mice as in the immune-competent C57BL/6 mice (data not shown).

To further explore this issue, we used TC1 tumor cellsexpressing the viral oncoprotein HPV-E7 and AE17.ovacells expressing chicken ovalbumin to evaluate the impact ofFAPþ cell depletion on the endogenous antitumor immunityby E7- or ova-specific tetramer staining of the infiltratinglymphocytes 8 days after adoptive transfer. We found a sta-tistically significant (P ¼ 0.02) increase in the percentage of






total CD8þT cells within the tumors of FAP-CAR–treatedmicecompared with control- or MigR1-T cell-treated mice (Fig. 5Cand E). Tumors from the FAP-CAR–treated mice had a signif-icant (P¼ 0.015) increase in the numbers of E7-specific T cells(Fig. 5D) or ova-specific T cells within the TC-1 and AE17.ovatumors, respectively (Fig. 5F).

To determine the mechanisms of this immune response,we repeated this experiment in AE17.ova tumor-bearingmice and analyzed the endogenous (non-GFP–expressing)T cells 3 and 8 days after T-cell injection. Consistent with theprevious findings by Kraman and colleagues (26), who used agenetic approach to ablate FAPþ cells, the number ofintratumoral T cells was similar between all three groups3 days after adoptive transfer (Fig. 6A). However, at this timepoint, the number of CD4þ T cells producing TNF-a wassignificantly higher in the FAP-CAR T-cell group comparedwith untreated and control T cell–treated groups (Fig. 6B,black bars), whereas there was no difference in the num-bers of activated CD69þ and 4-1BBþ T cells, nor in IFN-g–producing CD8 T cells (Fig. 6C–E, black bars). Eight daysafter treatment with FAP-CAR T cells, the number of T cellswas higher in tumors treated with FAP-CAR T cells (asabove) compared with the two control groups (Fig. 6F). Atthis time point, the numbers of CD69þ and IFN-gþ CD8þ T

cells were increased (Fig. 6C and E, gray bars), whereasthe numbers of TNF-producing T cells and 4-1BBþ–expres-sing T cells were similar among the groups (Fig. 6B and D,gray bars). Together, these results showed that depletionof FAPþ cells in tumors might enhance antitumor immu-nity by initially activating endogenous T cells, followed byincreasing intratumor T-cell infiltration at a later time point.

Augmentation of the efficacy of FAP-CAR T cells bycombination with an antitumor vaccine

Given the effects of FAPþ tumor cell depletion on antitumorimmunity, we hypothesized that combining a tumor vaccinewith FAP-CAR T-cell administrationmight enhance antitumorefficacy compared with either approach alone. HPV-E7–expressing TC1 tumor cells were injected subcutaneouslyinto C57BL/6 mice; when the tumors reached approximately200 mm3, saline or one subcutaneous dose of a vaccine con-sisting of 109 plaque–forming units (pfu) of an adenovirus-expressing HPV-E7 (Ad.E7; black arrow) was administeredto boost the endogenous T-cell response against E7-expres-sing cells. Four days after injection with saline or the Ad.E7vaccine, FAP-CAR T cells (107 cells; gray arrow) were givenintravenously. Both the Ad.E7 cancer vaccine and the FAP-CAR T cells alone had only modest effects on these large

BA

0

300

600

900

1,200

1,500

1,800

151050

Tu

mo

r v

olu

me

(m

m3)

Tu

mo

r v

olu

me

(m

m3)

Tu

mo

r v

olu

me

(m

m3)

Time after first T-cell infusion (d) Time after T-cell infusion (d)

Untreated

Single dose

Double dose

0

100

200

300

400

500

121086420

Untreated

WT FAP-CAR

DGKz KO FAP-CAR

0

500

1,000

1,500

2,000

3020100

Time after tumor cell inoculation (d)

UntreatedAd.E7FAP-CARCombo

C

Figure 4. Enhanced therapeutic responses of FAP-CAR T cells. Mice with AE17.ova flank tumors were injected intravenously with FAP-CAR T cells whentumors were approximately 100 mm3. A, the overall efficacy of FAP-CAR T cells was enhanced when a second dose of FAP-CAR T cells was givena week later. The gray arrow indicates the injection time of the second dose of FAP-CAR T cells. B, efficacy could also be enhanced after injectionof FAP-CAR T cells lacking the negative intracellular regulator DGKz. �, Statistical significance between untreated and FAP-CAR–treatedsamples; P < 0.05. #, Statistical significance between single dose FAP-CAR–treated group versus double dose group or DGKz knockout (KO) FAP-CAR–treated group. C, FAP-CAR T cells enhance efficacy of cancer vaccine. TC1 tumor cells were inoculated into the right flanks of C57BL/6mice. When tumors reached 200 mm3, one dose of Ad.E7 (109 pfu) was given to the mice contralaterally to their flank tumors (black arrow). FAP-CART cells (10 million cells) were given 4 days later (gray arrow). Tumor measurements followed. The values are expressed as the mean � SEM (n ¼ 5).�, Significant difference between untreated and the combo groups (P < 0.05).

Wang et al.





established tumors (Fig. 4C). However, the combination reg-imen induced tumor regressions and suppressed tumorgrowth for up to 2 weeks before further tumor progression.

ToxicityBecause FAP is an endogenous protein and toxicity (espe-

cially weight loss and anemia) was recently reported afterdepletion of FAPþ cells either by genetic ablation (26, 27) orFAP-CAR T-cell administration (31), we assessed off-tumor/on-target adverse effects after administration of our FAP-CART cells. In contrast with these reports, we observed no clinicaltoxicity or anemia in any of the FAP-CAR T-cell studiesdescribed above. The body weight of tumor-bearing miceremained constant or increased throughout each experiment(Supplementary Fig. S5).To further evaluate toxicity, necropsies were performed and

visceral organs (heart, lungs, pancreas, liver, spleen, kidneys,skeletal muscle, and bone marrow) were harvested, sectioned,stained, and analyzed in a blinded fashion 8 days after T-cellinjection in mice treated with one dose of WT FAP-CAR T cells

and 8 days after a second dose ofWT-FAP-CART cells from themice from the experiment shown in Fig. 4A. When comparedwith control tumor-bearing mice, no abnormalities wereobserved in the mice given WT FAP-CAR T cells. Specifically,this included a lack of bone marrow hypoplasia (Supplemen-tary Fig. S6A–S6C) as reported by Tran and colleagues (31), orany change in skeletal muscle (data not shown).

We also preformed necropsies on mice 8 days after injec-tion of the hyperactive DGKz-deficient FAP-CAR T cells fromthe experiment depicted in Fig. 4B. No abnormalities werenoted, except in the pancreatic sections that showed somemild focal perivascular and peri-islet lymphocytic infiltra-tion (Supplementary Fig. S6F). These changes were not seenin mice injected with WT FAP-CAR T cells (SupplementaryFig. S6E).

DiscussionIn this study, we investigated the antitumor efficacy and

safety of CAR-transduced T cells targeted to cells expressing

Figure 5. Adaptive immune response plays a key role in FAP-CAR T cell–induced antitumor response. AE17.ova tumors were injected into bothC57BL/6 (A) and NSG mice (B). When tumors reached approximately 100 to 125 mm3, a single dose (10 million) of FAP-CAR T cells wasadoptively transferred through the tail vein into mice. Tumor measurements followed. �, Statistical significance between untreated and FAP-CAR–treated samples; P < 0.05. FAP-CAR–induced infiltration of antigen-specific CD8 T cells into tumors. TC1 (C and D) and AE17.ova (E and F) tumors wereharvested 8 days after adoptive transfer of FAP-CAR T cells in mice. Tumors were digested and made into single-cell suspensions. Cells werestained with fluorochrome-conjugated tetramer loaded with E7- or SIINFEKEL(ova)-peptide, together with anti-CD8 antibody to determine thepercentage of tumor-specific CD8 T cells in tumors. �, Statistical significance between untreated, MigR1-, and FAP-CAR–treated samples; P < 0.05. TIL,tumor-infiltrating lymphocytes.






FAP, which is highly upregulated in the tumor stroma. Giventhat CASCs may have a major immune-modulating effect onboth innate and acquired immunity (13, 26), we used fullyimmune-competent mice, without bone marrow ablation, sothat we could evaluate the role of the acquired immune systemin FAP-CAR T cell–mediated antitumor response.

We demonstrate in multiple mouse models with establish-ed mesothelioma and lung cancer that our mouse FAP-CAR T cells exhibited antigen-specific cytotoxicity againstFAPþ stromal cells and markedly reduced the rare subsetof FAPþ/CD45þ/F4/80þ myeloid cells and the more pre-valent FAPþ/CD90þ stromal cells (Table 1). A single treat-ment with the FAP-CAR T cells resulted in approximately80% depletion of the FAPhi stromal cells 3 days after treat-ment of FAP-CAR T cells (Supplementary Fig. S2A), leavingwhite blood cells and FAPlo cells relatively unaffected (Sup-plementary Fig. S2B and S2C). The depletion of FAPþ cellswas associated with a significant inhibition (35%–50%) oftumor growth compared with untreated and vector control–transduced CAR T-cell (MigR1)-treated tumors (Fig. 3A–C, E,and F). Importantly, the antitumor activity of the FAP-CART cells was lost in the FAP-null mice (Fig. 3D) indicatingthat the antitumor activity is dependent on FAP expressionon host-derived cells.

The antitumor efficacy of the FAP-CAR T cells was also lostin immunodeficient mice (Fig. 5B), indicating the importanceof the acquired immune system in these tumor models. To

delineate this effect, we evaluated the endogenous T cellswithin the tumors at 3 and 8 days after CAR T-cell infusion.At the earlier time point, we did not see an increase in T-cellinfiltration or CD8 T-cell activation. However, there was anincrease in the number of TNF-a-producing CD4þ T cells (Fig.6B). Similarly, Kraman and colleagues showed that there wasno difference in the numbers of CD4þ or CD8þ cells in thetumors 48 hours after genetically ablating FAPþ cells with DT(26), but they did note an increase in the levels of TNF-amRNA.In contrast, at the later time point, we observed an increasein total CD8þ T cells as well as antigen-specific CD8þ T cells inboth the AE17.ova and E7-positive TC1 tumors. More IFN-g–producing CD8þ T cells, as well as CD69þ T cells, were alsofound at this time point (Fig. 6C and E). We postulate that theFAP-CAR T cells enter the tumors, deplete the FAPþ cells, andby an as yet unknown mechanism activate the endogenousCD4þ T cells to produce TNF-a. The high levels of TNF-amay induce tumor cell apoptosis and a temporary tumorvasculature shut down that may limit early infiltration byendogenous T cells. It seems that by 8 days after FAP-CART-cell treatment, activated CD8þ cells enter the tumor andfunction to further limit tumor growth. It should be notedthat the tumors in our mouse models are relatively immuno-genic with few fibroblasts, and therefore the contribution ofthe immune-mediated mechanisms may be more promi-nent than the contribution of the nonimmune-mediatedmechanisms (i.e., alterations in matrix and/or angiogenesis)

0

10

20

30

40

50

FAPCARMigR1Untreated

CD

3 T

ILs (

% t

ota

l)

0

5

10

15

FAPCARMigR1Untreated

CD

3 T

ILs (

% t

ota

l)

A

F

*

0

0.5

1

1.5

2

2.5

TN

F+

(%

CD

4 T

ILs)

0

2

4

6

8

IFN

-γ+ (

% C

D8 T

ILs)

0

1

2

CD

69

+ C

D3 T

ILs (

% t

ota

l)

0

1

2

3

44-1

BB

+ C

D3 T

ILs (

% t

ota

l)

B

C

D

E

*

* *

Untreated MigR1 FAP-CAR

TNF

IFN-γ

4-1BB

CD69

Day 3

Day 8

Day 3

Day 8

Day 3

Day 8

Day 3

Day 8

Day 3




Day 8

Figure 6. Activation and tumor infiltration of endogenous T cells following treatment with FAP-CAR T cells. AE17.ova tumor-bearing mice were injectedintravenously with 10 million FAP-CAR or MigR1 T cells when tumors reached approximately 100 mm3. At 3 and 8 days following adoptive transfer,tumors were harvested and digested to check for the frequencies of CD3þ tumor-infiltrating (A) total T cells 3 days after transfer; (B) TNF-producingcells; (C) IFN-g–producing cells; (D) 4-1BBþ cells; (E) CD69þ T cells; and (F) total T cells 8 days after transfer. �, Statistical significance betweenuntreated, MigR1-, and FAP-CAR–treated samples; P < 0.05.

Wang et al.





in nonimmunogenic and fibroblast-rich tumors. Preliminarystudies using more desmoplastic, nonimmunogenic mousetumor models and human xenografts support this idea.In our tumormodels, the FAP-CART cells elicited significant

but temporary inhibition of tumor growth. One explanation forthe transient effect is the short persistence of themurineCARTcells; the number of intratumoral murine FAP-CAR T cellsrapidly decreased with time (Supplementary Fig. S3A). This islikely due to a number of well-known intrinsic differencesbetween mouse and human T cells. After expanding humanCART cells, the lymphocytes are "rested down" before injectionand are less sensitive to immediate activation-induced celldeath (AICD). In themouse system, the injected cells are highlyactivated (as required for retroviral transduction). Althoughthe humanT cells persist and proliferate forweeks in tumors inimmunodeficient mice (3, 44), the transduced mouse T cellshave short lifespan and undergo AICD (Supplementary Fig. S7).There is relatively little information about the persistence ofmouse CAR T cells, but data from this study and our previouswork with mesothelin-CAR mouse T cells (43) are consistentwith the results reported by Peng and colleagues (45) using invitro expanded pmel-1 T cells that showed similar shortpersistence of mouse T cells in mice. In their study, pmel-1T cells were undetectable in the peripheral blood by day 9 afteradoptive transfer and by day 13 at the tumor site. To ensurethat the short lifespan was not caused by the human CD3z andhuman 4-1BB activation domains, we constructed a fullymouse FAP-CAR comprised the 73.3 scFv coupled with themouse CD3z chain and mouse CD28 domain (Fig. 1D). Wefound virtually equal efficacy in killing, cytokine production,persistence, and antitumor activity between the T cells expres-sing both constructs (Supplementary Fig. S3B–S3E). We alsofound that both CARs were equally susceptible to AICD(Supplementary Fig. S7B). We determined whether our FAP-CARmouse T cellsmight be enhanced if we preconditioned thehost by inducing lymphodepletion (46, 47), but no increase inefficacy was observed inmice irradiated before injection of ourFAP-CAR T cells (data not shown). We have not yet testedwhether the administration of IL-2 might enhance theirefficacy.Given the short persistence of our FAP-CAR T cells (Sup-

plementary Fig. S3A), a second dose of T cells was admin-istered 1 week after the first injection and showed addedefficacy (Fig. 4A), which clearly indicated that enhancedpersistence would augment efficacy. Our recent observation(43) indicates that blocking a T cell-intrinsic negative regu-latory mechanism (upregulation of the enzyme DGK) aug-ments the killing ability and persistence of murine CART cells. Similar to our observations with mouse CAR-T cellstargeted to mesothelin (42), FAP-CAR T cells deficient inDGKz had enhanced ability to kill FAP-expressing cellsin vitro and were clearly more efficacious in vivo (Fig. 4B).Our data suggest that it will be advantageous to optimizeboth the persistence and potency of the T cells.As in any cancer therapy where the target is not completely

tumor specific, potential toxicity is a major concern. In pre-vious studies using a vaccine against FAP and with immuno-conjugate therapy (28–30),murine tumor growthwas inhibited

without obvious toxicity. Some of these studies also performedanalyses and showed that there was no significant inhibition ofwound healing (28, 29). In a phase I human trial, trace levels of131I-labeled humanized anti-FAP antibody (sibrotuzamab)showed excellent tumor targeting with no detectable uptakein normal organs and no major toxicity at any dose level (48).However, as described above, the authors of two recent articlesreported that ablation of FAPþ cells (using a genetic approachin ref. 27 or using FAP-CAR T cells in ref. 31) led to significantcachexia (weight loss) and anemia. We used flow cytometryto assess the percentage of FAPþ cells in the pancreas,lungs, tumors, and bone marrow collected from mice bearingdifferent flank tumors. Despite different mouse strains used ineach mouse tumor model, we found less than 0.1% of FAPþ

stromal cells in the lungs and bonemarrow (data not shown). Incontrast and consistent with results reported by Roberts andcolleagues (27), 3% to 5% of the dissociated cells from thepancreas expressed FAP (data not shown). However, when wecompared FAP expression on those pancreatic stromal cellswith that on the tumor-associated stromal cells isolated fromthe samehosts, we found that theCASCsexpressedhigher levelsof FAP (Supplementary Fig. S8).

In contrast with these reports (27, 31) and consistent withthe findings of Kakarla and colleagues (33) and Schuberth andcolleagues (34), administration of one dose or even twodoses ofour FAP-CAR T cells did not cause any weight loss (Supple-mentary Fig. S5), nor did they cause any decrease in hematocritor increase in serum amylase (data not shown). Furthermore,detailed histologic analyses of necropsy samples showed nomicroscopic abnormalities, and specifically no damage to themuscle, the pancreas (Supplementary Fig. S6D and S6E), or thebone marrow (Supplementary Fig. S6A–S6C).

The reasons for these differences in toxicity are notcompletely understood. It is possible that the total bodyirradiation before adoptive T-cell transfer used by Tran andcolleagues (31) somehow contributed to the side effects.Although our main construct had the 4-1BB costimulatorymolecule rather than CD28, in vitro studies showed no differ-ences in their killing ability or IFN production (SupplementaryFig. S3B and S3C) or in vivo activity (Supplementary Fig. S3E).A likely explanation relates to the different properties of thescFv antibodies used. The affinities of the scFv antibodies formouse FAP are similar. The affinity of the FAP-5 antibody formouse FAP was reported to be 0.6 nmol/L (31). Our biacoremeasurement showed that the affinity of the 73.3 mAb formouse FAP protein was less than 1 nmol/L (0.1–0.3 nmol/L).However, we found that FAP-5 and 73.3 reacted with distinctepitopes on FAP in antibody competition experiments (Sup-plementary Fig. S9), consistent with the specificity of the 73.3mAb for murine FAP compared with that of FAP-5 for anepitope shared by murine and human FAP. The distinctepitope specificity seems to allow the FAP-CAR T cells expres-sing the 73.3 scFv to efficiently eliminate the FAPhi cells, whilesparing cells expressing lower levels of FAP (SupplementaryFig. S2), such as cells in the pancreas, and presumably cells inthe bone marrow and muscle. The position of the bindingepitope of a scFv in a CAR T cell has previously been shown todetermine the efficacy of activation (49). Data from the studies






conducted by Roberts and colleagues and Tran and colleaguessuggest that complete ablation of FAPþ cells may cause suchside effects as fatigue and anemia (27, 31). However, our datashow that it is possible to partially deplete tumor-associatedFAPþ cells while retaining antitumor efficacy, but withouteliciting severe side effects. The only instance in which weobserved any histologic abnormalities was in the pancreas(Supplementary Fig. S6F) but only when we used the hyper-active DGKz-deleted cells. Data from Kakarla and colleaguessupport this idea (33).

Further studies will be needed to determine why we and twoother groups (33, 34) induced significant antitumor efficacy,whereas Tran and colleagues (31) did not. Although the con-structs are all slightly different, we showed antitumor activitywith CARs containing either human or mouse cytoplasmicdomains (Supplementary Fig. S3E). Our data suggest thatendogenous immune responses may be important for theactivity of our CARs. Because Tran and colleagues (31) usedtotal body irradiation, they may have lost some of that effect.Alternatively, irradiation might have impacted the tumorstroma, thus limiting the effect or antistromal therapy. Fur-thermore, the amount of stroma was very low in the smallmouse tumors studied by Tran and colleagues (ref. 31; data notshown). In any case, we were able to show clear antitumorefficacy albeit to a lesser extent in two of the tumor cell linesused by Tran and colleagues (ref. 31; Fig. 3A–3C, E, and F).

The optimal utility of FAP-CAR T cells may be in combina-tion with other therapies, including immuno- and chemother-apy, in which stromal disruption could enhance drug delivery.Kakarla and colleagues (33) showed augmented efficacy bycombining FAP-CAR T cells with CAR T cells targeting EphA2on tumor cells. In the current study, we combined FAP-CAR Tcells with an Ad.E7 cancer vaccine that elicits E7-specificadaptive immune response against TC1 cells. As expected,neither the cancer vaccine nor the FAP-CAR T cells aloneworked well on large, established tumors (Fig. 4C); however, incombination, there was an additive or synergistic effect againstTC1 tumors.

In summary, we showed that FAP-CAR T cells reduce tumorgrowth in vivo in an antigen-specific manner. Targeting tumorstromal cells with CAR T cells augmented antitumor immu-nity. Although we did not eradicate tumors completely in ouranimal models, we believe that the efficacy of FAP-CAR T cellscould be enhanced by improving the persistence of T cells,generating more highly active T cells, administering multipledoses of FAP-CAR-T cells, and by combining with other typesof immuno- or chemotherapy. We did not observe toxicityunder the dosing conditions tested and in otherwise healthytumor-bearing animals.

We believe that data from Karkala and colleagues (33),Schuberth and colleagues (34), and in this report support thedevelopment of anti-human FAP-CAR for potential translationto the clinic. Potential toxicities, such as anemia and/orcachexia, should they occur, are tolerable and similar to thoseseen in most conventional chemotherapy treatments. Impor-tantly, our current clinical approach using biodegradablemRNA-transduced CAR T cells (50) will allow us to determinewhether these toxicities may only be temporary. It may well bethat the FAP-elimination approach will best be used in com-binationwith other therapies andmay only need to be given forshort periods of time.

Disclosure of Potential Conflicts of InterestC.H. June has received commercial research support from Novartis. No

potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: L.-C.S. Wang, A. Lo, J. Scholler, J. Sun, E. Pur�e, S.M.AlbeldaDevelopment of methodology: L.-C.S. Wang, J. Scholler, J. Sun, R.S. Majumdar,L.A. Johnson, A.C. Durham, E. Pur�eAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): L.-C.S.Wang, A. Lo, J. Scholler, V. Kapoor, C.E. Cotner,A.C. Durham, E. Pur�eAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): L.-C.S. Wang, A. Lo, J. Scholler, C.E. Cotner, A.C.Durham, C.C. Solomides, E. Pur�e, S.M. AlbeldaWriting, review, and/or revision of the manuscript: L.-C.S. Wang, A. Lo,J. Scholler, C.H June, E. Pur�e, S.M. AlbeldaAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): V. Kapoor, M. AntzisStudy supervision: C.H June, E. Pur�e, S.M. Albelda

AcknowledgmentsThe authors thank Jennifer Whealdon, James Shecter, Geetha Muthuku-

maran, and Danielle Qing for their technical support. The authors also thankDr. Gary Koretzky for providing DGKz knockout mice, Dr. Wayne Hancock forproviding Thy1.1 congenic C57BL/6 mice, Boehringer Ingelheim for providingFAPLacZ knockin mice, and Jonathan Cheng for providing FAP-ECD–producingHEK293 cells. This study made use of the Wistar Institute Cancer Center FlowCytometry Core facility, the Protein Production, Libraries &Molecular ScreeningFacility (thanks to Dr. David C. Schultz, Director), and infrastructure support tothe Wistar Institute from the Commonwealth of Pennsylvania.

Grant SupportThis work was supported by National Cancer Institute (NCI) grants P01 CA

66726-07 (to S.M. Albelda and C.H June), R01 CA 141144 and RO1 CA 172921 (to S.M. Albelda and E. Pur�e) and a "Clinic and Laboratory Integration Program"Grantfrom the Cancer Research Institute (to E. Pur�e and S.M. Albelda). L.-C.S. Wangwas supported by the Mesothelioma Applied Research Foundation (MARF)award and A. Lowas sponsored by a START fellowship from the Cancer ResearchInstitute.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received March 24, 2013; revised November 5, 2013; accepted November 5,2013; published OnlineFirst November 12, 2013.

References1. June CH. Principles of adoptive T cell cancer therapy. J Clin Invest

2007;117:1204–12.2. June CH. Adoptive T cell therapy for cancer in the clinic. J Clin Invest

2007;117:1466–76.3. Porter D, LevineB, KalosM, BaggA, JuneCH. Delayed tumor lysis and

CLL remission from chimeric antigen receptor-modified T cells. NewEngl J Med 2011;365:725–33.

4. Kalos M, Levine B, Porter D, Katz S, Grupp SA, Bagg A, et al. T cellswith chimeric antigen receptors have potent antitumor effects and canestablish memory in patients with advanced leukemia. Sci Trans Med2011;3:95ra73.

5. 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.

Wang et al.





6. Jena B, Dotti G, Cooper LJ. Redirecting T-cell specificity by intro-ducing a tumor-specific chimeric antigen receptor. Blood 2010;116:1035–44.

7. Bhowmick NA, Neilson EG, Moses HL. Stromal fibroblasts in cancerinitiation and progression. Nature 2004;432:332–7.

8. Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delaunay T,Naeem R, et al. Stromal fibroblasts present in invasive human breastcarcinomas promote tumor growth and angiogenesis through elevat-ed SDF-1/CXCL12 secretion. Cell 2005;121:335–48.

9. Santos AM, Jung J, Aziz N, Kissil JL, Pure E. Targeting fibroblastactivation protein inhibits tumor stromagenesis and growth in mice. JClin Invest 2009;119:3613–25.

10. Ito TK, Ishii G, Chiba H, Ochiai A. The VEGF angiogenic switch offibroblasts is regulated by MMP-7 from cancer cells. Oncogene2007;26:7194–203.

11. Zhang Y, Tang H, Cai J, Zhang T, Guo J, Feng D, et al. Ovarian cancer-associated fibroblasts contribute to epithelial ovarian carcinomametastasis bypromoting angiogenesis, lymphangiogenesis and tumorcell invasion. Cancer Lett 2011;303:47–55.

12. Fridlender ZV, Sun J, Kim S, Kapoor V, Cheng G, Ling L, et al.Polarization of tumor-associated neutrophil phenotype by TGF-beta:"N1" versus "N2" TAN. Cancer Cell 2009;16:183–194.

13. Nazareth MR, Broderick L, Simpson-Abelson MR, Kelleher RJ Jr,Yokota SJ, Bankert RB. Characterization of human lung tumor-asso-ciated fibroblasts and their ability to modulate the activation of tumor-associated T cells. J Immunol 2007;178:5552–62.

14. Cohen SJ, Alpaugh RK, Palazzo I, Meropol NJ, Rogatko A, Xu Z, et al.Fibroblast activation protein and its relationship to clinical outcome inpancreatic adenocarcinoma. Pancreas 2008;37:154–8.

15. Goscinsky MA, Suo Z, Florenes VA, Vlatkovic L, Nesland JM,Giercksky KE. FAP-alpha and uPA show different expression patternsin premalignant and malignant esophageal lesions. Ultrastruct Pathol2008;32:89–96.

16. Henry LR, Lee HO, Lee JS, Klein-Szanto A, Watts P, Ross EA, et al.Clinical implications of fibroblast activation protein in patients withcolon cancer. Clin Cancer Res 2007;13:1736–41.

17. Scanlan MJ, Mohan Raj BK, Calvo B, Garin-Chesa P, Sanz-MoncasiMP, Healey JH, et al. Molecular cloning of fibroblast activtion proteina,a member of the serine protease family selectively expressed instromal fibroblasts of epithelial cancers. Proc Natl Acad Sci U S A1994;91:5657–61.

18. Garin-Chesa P, Old LJ, RettigWJ. Cell surface glycoprotein of reactivestromal fibroblasts as a potential antibody target in human epithelialcancers. Proc Natl Acad Sci U S A 1990;87:7235–9.

19. Niedermeyer J,Garin-ChesaP,KrizM,HilbergF,Mueller E,BambergerU, et al. Expression of the fibroblast activation protein during mouseembryo development. Int J Dev Biol 2001;45:445–7.

20. Mathew S, Scanlan MJ, Mohan Raj BK, Murty VV, Garin-Chesa P, OldLJ, et al. The gene for fibroblast activation protein [alpha] (FAP), aputative cell surface-bound serine protease expressed in cancerstroma and wound healing, maps to chromosome band 2q23. Geno-mics 1995;25:335–7.

21. Wang XM, Yao TW, Nadvi NA, Osborne B, McCaughan GW, GorrellMD. Fibroblast activation protein and chronic liver disease. FrontBiosci 2008;13:3168–80.

22. Acharya PS, Zukas A, Chandan V, Katzenstein AL, Pur�e E. Fibro-blast activation protein: a serine protease expressed at the remo-deling interface in idiopathic pulmonary fibrosis. Hum Pathol 2006;37:352–60.

23. Rettig WJ, Garin-Chesa P, Beresford HR, Oettgen HF, Melamed MR,Old LJ. Cell-surface glycoproteins of human sarcomas: differentialexpression in normal and malignant tissues and cultured cells. ProcNatl Acad Sci U S A 1988;85:3110–4.

24. Huber MA, Kraut N, Park JE, Schubert RD, Rettig WJ, Peter RU, et al.Fibroblast activation protein: differential expression and serine prote-ase activity in reactive stromal fibroblasts ofmelanocytic skin tumors. JInvest Dermatol 2003;120:182–8.

25. Aertgeerts K, Levin I, Shi L, Snell GP, Jennings A, Prasad GS, et al.Structural and kinetic analysis of the substrate specificity of humanfibroblast activation protein alpha. J Biol Chem 2005;280:19441–4.

26. Kraman M, Bambrough PJ, Arnold JN, Roberts EW, Magiera L, JonesJO, et al. Suppression of antitumor immunity by stromal cells expres-sing fibroblast activation protein-alpha. Science 2010;330:827–30.

27. Roberts EW, Deonarine A, Jones JO, Denton AE, Feig C, Lyons SK,et al. Depletion of stromal cells expressing fibroblast activation pro-tein-a from skeletal muscle and bone marrow results in cachexia andanemia. J Exp Med 2013;210:1137–51.

28. Loeffler M, Kruger JA, Niethammer AG, Reisfeld RA. Targeting tumor-associated fibroblasts improves cancer chemotherapy by increasingintratumoral drug uptake. J Clin Invest 2006;116:1955–62.

29. Lee J, Fassnacht M, Nair S, Boczkowski D, Gilboa E. Tumor immu-notherapy targeting fibroblast activation protein, a product expressedin tumor-associated fibroblasts. Cancer Res 2005;65:11156–63.

30. Ostermann E, Garin-Chesa P, Heider KH, Kalat M, Lamche H, PuriC, et al. Effective immunoconjugate therapy in cancer modelstargeting a serine protease of tumor fibroblasts. Clin Cancer Res2008;14:4584–92.

31. 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.

32. Lee PP, Yee C, Savage PA, Fong L, Brockstedt D, Weber JS, et al.Characterization of circulating T cells specific for tumor-associatedantigens in melanoma patients. Nat Med 1999;5:677–85.

33. Kakarla S, Chow KKh, Mata M, Shaffer DR, Song XT, Wu MF, et al.Antitumor effects of chimeric receptor engineered human T cellsdirected to tumor stroma. Mol Ther 2013;21:1611–20.

34. Schuberth PC, Hagedorn C, Jenesen SM, Gulati P, van den Broek M,Mischo A, et al. Treatment of malignant pleural mesothelioma byfibroblast activation protein-specific re-directed T cells. J Transl Med2013;11:187.

35. Liu R, Li H, Liu L, Yu J, Ren X. Fibroblast activation protein: a potentialtherapeutic target in cancer. Cancer Biol Ther 2012;13:123–9.

36. Jackaman C, Bundell CS, Kinnear BF, Smith AM, Fillion P, van HagenD, et al. IL-2 intratumoral immunotherapy enhances CD8þ T cells thatmediate destruction of tumor cells and tumor-associated vasculature:a novel mechanism for IL-2. J Immunol 2003;171:5051–63.

37. Lin KY, Guarnieri FG, Staveley-O'Carroll KF, Levitsky HI, August JT,Pardoll DM, et al. Treatment of established tumorswith a novel vaccinethat enhances major histocompatibility class II presentation of tumorantigen. Cancer Res 1996;56:21–6.

38. Johnson L,Mercer K, GreenbaumD, Bronson RT, Crowley D, TuvesonDA, et al. Somatic activation of the K-ras oncogene causes early onsetlung cancer in mice. Nature 2001;410:1111–6.

39. Niedermeyer J, Kriz M, Garin-Chesa P, Bamberger U, Lenter MC, ParkJ, et al. Targeted disruption of mouse fibroblast activation protein. MolCell Biol 2000;20:1089–94.

40. 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.

41. Pear WS, Miller JP, Xu L, Pui JC, Soffer B, Quackenbush RC, et al.Efficient and rapid induction of a chronic myelogenous leukemia-likemyeloproliferative disease in mice receiving P210 bcr/abl-transducedbone marrow. Blood 1998;92:3780–92.

42. Morgan RA, Johnson LA, Davis JL, Zheng Z, Woolard KD, Reap EA,et al. Recognition of glioma stem cells by genetically modified T cellstargeting EGFRvIII and development of adoptive cell therapy forglioma. Hum Gene Ther 2012;23:1043–53.

43. Riese MJ, Wang L-CS, Moon EK, Ranganathan A, Joshi RP, June CH,et al. Enhanced effector responses in activated CD8þ T cells deficientin diacylglycerol kinases. Cancer Res 2013;73:3566–77.

44. Moon EK, Carpenito C, Sun J, Wang L-CS, Kapoor V, Predina J, et al.Expression of a functional CCR2 receptor enhances tumor localizationand tumor eradication by retargeted human T cells expressing amesothelin-specific chimeric antibody receptor. Clin Cancer Res2011;17:4719–30.

45. Peng W, Liu C, Xu C, Lou Y, Chen J, Yang Y, et al. PD-1 blockadeenhances T-cell migration to tumors by elevating IFN-g induciblechemokines. Cancer Res 2012;72:5209–18.






46. Chinnasamy D, Tran E, Yu Z, Morgan RA, Restifo NP, Rosenberg SA.Simultaneous targeting of tumor antigens and the tumor vasculatureusing T lymphocyte transfer synergize to induce regression of estab-lished tumors in mice. Cancer Res 2013;73:3371–80.

47. Muranski P, Boni A, Wrzesinski C, Citrin DE, Rosenberg SA, ChildsR, et al. Increased intensity lymphodepletion and adoptive immu-notherapy-how far can we go? Nat Clin Pract Oncol 2006;3:668–81.

48. Scott AM, Wiseman G, Welt S, Adjei A, Lee FT, Hopkins W, et al. Aphase I dose-escalation study of sibrotuzumab in patients with

advanced or metastatic fibroblast activation protein-positive cancer.Clin Cancer Res 2003;9:1639–47.

49. Hombach AA, Schildgen V, Heuser C, Finnern R, GilhamDE, Abken H.T cell activation by antibody-like immunoreceptors: the position of thebinding epitope within the target molecule determines the efficiency ofactivation of redirected T cells. J Immunol 2007;178:4650–7.

50. Zhao Y, Moon E, Carpenito C, Paulos CM, Liu X, Brennan AL, et al.Multiple injections of electroporated autologous T cells expressing achimeric antigen receptor mediate regression of human disseminatedtumor. Cancer Res 2010;70:9053–61.

Wang et al.





2014;2:154-166. Published OnlineFirst November 12, 2013.Cancer Immunol Res Liang-Chuan S. Wang, Albert Lo, John Scholler, et al. Augment Host Immunity without Severe ToxicityChimeric Antigen Receptor T Cells Can Inhibit Tumor Growth and Targeting Fibroblast Activation Protein in Tumor Stroma with

Updated version

10.1158/2326-6066.CIR-13-0027doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerimmunolres.aacrjournals.org/content/suppl/2013/11/13/2326-6066.CIR-13-0027.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerimmunolres.aacrjournals.org/content/2/2/154.full#ref-list-1

This article cites 50 articles, 23 of which you can access for free at:

Citing articles

http://cancerimmunolres.aacrjournals.org/content/2/2/154.full#related-urls

This article has been cited by 19 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

Permissions

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

.http://cancerimmunolres.aacrjournals.org/content/2/2/154To request permission to re-use all or part of this article, use this link



http://cancerimmunolres.aacrjournals.org/lookup/doi/10.1158/2326-6066.CIR-13-0027

http://cancerimmunolres.aacrjournals.org/content/suppl/2013/11/13/2326-6066.CIR-13-0027.DC1

http://cancerimmunolres.aacrjournals.org/content/2/2/154.full#ref-list-1

http://cancerimmunolres.aacrjournals.org/content/2/2/154.full#related-urls

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

mailto:[email protected]

http://cancerimmunolres.aacrjournals.org/content/2/2/154


targetingfibroblastactivationproteinintumorstromawith ......matrix (7–11); and (iii) exerting an...

Documents