FAP Promotes Immunosuppression by Cancer-Associated ... · 4/7/2016 · Green Master kit (TaKaRa)...

Microenvironment and Immunology

FAP Promotes Immunosuppression byCancer-Associated Fibroblasts in the TumorMicroenvironment via STAT3–CCL2 SignalingXuguang Yang1,2, Yuli Lin1,2, Yinghong Shi2,3, Bingji Li1,2, Weiren Liu2,3,Wei Yin1,2,Yongjun Dang2,4, Yiwei Chu1,2, Jia Fan2,3, and Rui He1,2

Abstract

Cancer-associated fibroblasts (CAF) are components of thetumor microenvironment whose contributions to malignantprogression are not fully understood. Here, we show that thefibroblast activation protein (FAP) triggers induction of a CAFsubset with an inflammatory phenotype directed by STAT3activation and inflammation-associated expression signaturemarked by CCL2 upregulation. Enforcing FAP expression innormal fibroblasts was sufficient to endow them with aninflammatory phenotype similar to FAPþCAFs. We identifiedFAP as a persistent activator of fibroblastic STAT3 througha uPAR-dependent FAK–Src–JAK2 signaling pathway. In amurine liver tumor model, we found that FAPþCAFs werea major source of CCL2 and that fibroblastic STAT3–CCL2signaling in this setting promoted tumor growth by enhancing

recruitment of myeloid-derived suppressor cells (MDSC). TheCCL2 receptor CCR2 was expressed on circulating MDSCs intumor-bearing subjects and FAPþCAF-mediated tumor pro-motion and MDSC recruitment was abrogated in Ccr2-defi-cient mice. Clinically, we observed a positive correlationbetween stromal expression of FAP, p-STAT3, and CCL2 inhuman intrahepatic cholangiocarcinoma, a highly aggressiveliver cancer with dense desmoplastic stroma, where elevatedlevels of stromal FAP predicted a poor survival outcome. Takentogether, our results showed how FAP–STAT3–CCL2 signalingin CAFs was sufficient to program an inflammatory compo-nent of the tumor microenvironment, which may have par-ticular significance in desmoplasia-associated cancers. CancerRes; 76(14); 1–12. �2016 AACR.

IntroductionCancer-associated fibroblasts (CAF) are the most prominent

stromal components and play important roles in modulating thetumor microenvironment and influencing the behavior of tumorcells primarily by releasing proteolytic enzymes, growth factorsand cytokines (1, 2).Despite extensive studies showing the tumor-promoting role of CAFs, recent studies suggested an unexpectedtumor-suppressing role of CAFs, as depletion of CAFs by targetingmyofibroblasts or interfering with hedgehog signaling accelerates

pancreatic cancer in mouse models (3, 4). This paradox could bedue to the heterogeneity of CAFs, in which different CAF subsetsmay play opposing roles in tumor development. Thus, the under-standing of the unique pathophysiological functions of differentCAF subsets and their roles in tumor development would behelpful to develop a more specific anticancer strategy targeting atumor-promoting CAF subset.

Although CAFs have traditionally been considered as a-SMAþ

myofibroblasts (5), several other markers have been used toidentify and investigate CAFs, including fibroblast activationprotein-a (FAP), fibroblast-specific protein 1 (FSP1), and plate-let-derived growth factor receptor a (PDGFRa; refs. 6–8).However, these markers do not mark all CAFs, and some ofthem are not uniquely expressed in CAFs. For example, FSP1and PDGFRa are also expressed in normal fibroblasts or othercell types (9, 10). In contrast, FAP is selectively expressed byCAFs in most of human epithelial cancers as well as by reactivestromal fibroblasts under some inflammatory conditions, suchas liver cirrhosis (11, 12). FAP is originally identified as amembrane-bound serine protease implicated in extracellularmatrix remodeling (11). Accumulating data demonstrated thattargeting FAP inhibited tumor growth in mouse models; how-ever, most of studies have focused on its influence on tumor cellbiology and angiogenesis by using xenograft models of humantumors in immune-deficient mice (13, 14). Interestingly, arecent study demonstrated that FAP-expressing stromal cellscould mediate immunosuppression, as ablation of these cellsenhanced T cell–mediated killing of tumor cells in an immu-nogenic, transplanted tumor model (15). Further study showed

1Departmentof ImmunologyandKeyLaboratoryofMedicalMolecularVirologyofMinistries ofEducationandHealth, School ofBasicMedicalSciences, Fudan University, Shanghai, China. 2Shanghai Medical Col-lege and Fudan University, Shanghai. 3Key Laboratory of Carcinogen-esis andCancer Invasion ofMinistryof Education, Department of LiverSurgery, Liver Cancer Institute, ZhongshanHospital, FudanUniversity,Shanghai,China. 4DepartmentofBiochemistryandMolecular Biology,School of Basic Medical Sciences, Fudan University, Shanghai, China.

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

X. Yang, Y. Lin, and Y. Shi contributed equally to this article.

Corresponding Authors: Rui He, Department of Immunology, School of BasicMedical Sciences, Shanghai Medical College, Fudan University 130 Dong AnRoad, Shanghai 200032, China. Phone: 86-21-54237419; Fax: 86-021-54237419;E-mail: [email protected]; and Jia Fan, Department of Liver Surgery, Zhong-shan Hospital, 180 FengLin Road, Shanghai 200032, China. Phone: 86-21-64041990; Fax: 86-21-64037181; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-15-2973

�2016 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org OF1

Research. on November 19, 2020. © 2016 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 23, 2016; DOI: 10.1158/0008-5472.CAN-15-2973

http://cancerres.aacrjournals.org/

that FAP-expressing CAFs exhibited upregulated expression ofproinflammatory genes and may be responsible for tumorimmune evasion in a mouse model of pancreatic cancer(16). A previous study also showed that CAFs isolated fromseveral types of solid tumor were inflammatory and orches-trated tumor-promoting inflammation (8). Despite the fact thatFAP is selectively expressed in CAFs and often used as a CAFmarker, it remains largely unknown whether FAP could induceinflammatory CAFs and how FAP-expressing CAFs mediatetumor-promoting inflammation and immunosuppression.

In this study, we compared the inflammatory phenotypeof FAPþCAFs and FAP�CAFs derived from mouse tumors andinvestigated the signaling pathway mediating FAP-induced in-flammatory CAFs and the mechanism by which fibroblasticFAP contributed to CAF-mediated tumor growth. Lastly, weevaluated the clinical relevance of FAP in intrahepatic cholangio-caicinoma (ICC), a highly aggressive and demosplastic humanprimary liver tumor.

Materials and MethodsMice and cell lines

Female C57BL/6 WT mice were purchased from the ChineseAcademyof Sciences (Shanghai, China) andB6.129S4-Ccr2tm1Ifc/J(Ccr2�/�) mice from The Jackson Laboratory. All mice were keptand bred in a specific pathogen-free environment in the animalfacility at Fudan University Shanghai Medical College. All animalexperiments were approved by the Animal Care and Use Com-mittee at Fudan University (Shanghai, China). The murine celllines of hepatoma Hepa1-6, Lewis lung carcinoma (LLC), andmelanoma B16-F10 were obtained from the Chinese Academy ofScience. All cell lines, obtained between 2013 and 2014, wereauthenticated bymorphology andbiologic behavior and tested toexclude mycoplasma contamination before experiments. Cellswere cultured less than 3 months after resuscitation.

Preparation of fibroblasts and immunofluorescenceHepa1-6 tumor tissues were digested by 1 mg/mL collagenase

IV (Sigma) to isolate CAFs. Normal fibroblasts (NF) were isolatedfrom skin tissues of age-matched C57BL/6 mouse. All fibroblastsused for the in vitro study were within two passages. CAFs werefixed in 4% formaldehyde followed by permeabilization with0.5% Triton X-100. Cells then were treated with primary anti-bodies against FAP (Abcam), a-SMA (Dako), and vimentin(Abcam) followed by species-specific secondary antibodies con-jugated with AlexaFluor 488 or 594 labeled (Invitrogen).

Coimmunoprecipitation and Western blottingFor coimmunoprecipitation (Co-IP), fap gene (NM_007986.2)

was cloned into pCMV-Tag2B-flag vector, and plaur gene(NM_011113.3) into PCDNA3.1-HA vector. Primary antibodiesagainst uPAR (R&D Systems), FAP (Abgent), HA (Santa CruzBiotechnology) were used. The rabbit normal IgG antibodies(Santa Cruz Biotechnology) were added as a control. Anti-FLAGM2 Affinity Gel (Sigma) was used for immunoprecipitation. ForWestern blotting, the following primary antibodies were used:anti-mouse uPAR, FAP, HA, and flagM2 (Sigma), rabbit anti-totalFAK, STAT3, p65, c-Src, JAK2, rabbit anti-phospho-FAK (Tyr576/577), Phospho-STAT3 (Tyr705), Phospho-p65 (Ser536), Phos-pho-c-Src (Tyr416), Phospho-JAK2 (Tyr1007), GAPDH (all1:2,000 dilution; Cell Signaling Technology). Inhibitors of

P38, PI3K, ERK, FAK, c-Src, and JAK2 were purchased fromSelleck Chemicals, and PT100 inhibitor (Talabostat) fromMedchemexpress.

Lentivirus vector preparation and fibroblast transductionThe fap gene was constructed in lentiviral vector pCDH-EF1-

MCS-T2A-copGFP (System Biosciences), and lentivirus was gen-erated according to the manufacturer's protocol. NFs were trans-duced by FAP-PCDH-copGFP or PCDH-GFP lentivirus. Thesequences for si-Fap, si-Stat3, si-Ccl2 and si-Plaur were designedand synthesized by GenePharma, and the sequences wereshown in Supplementary Table S1. SiRNAs were transfected bylipofectamine iMAX (Invitrogen). Short-hairpin RNA (shRNA)specifically targeting mouse Ccl2, Stat3, and Fap or scrambledsequences were constructed into pSIH1-H1-copGFP vector (Sys-tem Biosciences). ShRNA-Lentivirus targeting these genes weregenerated and were transduced into CAFs or fibroblasts.

STAT3 luciferase reporter activityThree copies of a STAT3 consensus binding site were insert-

ed into pGL4 vector to construct STAT3 luciferase reportervector. Six mg pGL4M-STAT3 and 3 mg of pRL-TK-renilla-luciferase plasmid (Promega) were cotransfected by lipofec-tamine LTX. Luciferase activity was determined using a dualluciferase assay kit (Promega). Relative luciferase activity wascalculated as the ratio of firefly luciferase activity to Renillaluciferase activity.

Syngeneic tumor modelVarious fibroblasts (all used in 2�105 each, except for

FAPþCAFs used in 105 each) were subcutaneously coinjectedwith Hepa1-6 (1�106) into C57BL/6 mice or Ccr2�/� mice. Thetumor volume was measured as described previously (13).

Flow cytometryThe following antibodies were used: fluorochrome-labeled

anti-mouse CD45 (30-F11), CD31 (390), PDGFR (53-6.7), CD4(GK1.5), CD8 (APA5), CD11b (M1/70), F4/80 (BM8), IFNg(AN-18), Gr-1(RB6-8C5; all from eBioscience); Ly6G (1A8) andLy6C (HK1.4; both from BD), CCR2 (475301) and uPAR(MAB531; both from R&D Systems), FAP (Abgent). For isolationof CAFs and analysis of different cell populations, single-cell sus-pensions prepared from Hepa1-6 tumors were further separatedby FACS into: CAFs (PDGFRaþCD45�EpCAM�CD31�); FAPþ

CAFs (FAPþPDGFRaþCD45�EpCAM�CD31�); endothelial cells(CD31þCD45�EpCAM�); leukocytes (CD45þEpCAM�CD31�)and tumor cells (EpCAMþCD45�CD31�). Samples were acquiredor sorted by CyAn orMoflo (BeckmanCoulter) and analyzedwithSummit 5.2 (Beckman Coulter).

RNA isolation and quantitative real-time PCRTotal RNA was isolated using TRIzol reagent (Invitrogen) and

transcribed into cDNA by PrimeScript RT Master Mix (TaKaRa).quantitative PCR (qPCR) was performed using the Power SYBRGreen Master kit (TaKaRa) with Applied Biosystems 7500. Therelative expression of target gene was calculated using the 2DC(t)

method. Fold change of target gene expression were calculatedby normalization to that in control group injected with Hepa1-6 alone. The primer sequences of all genes for PCR are shown inSupplementary Table S1.

Yang et al.

Cancer Res; 76(14) July 15, 2016 Cancer ResearchOF2




Chromatin immunoprecipitationFibroblasts were fixed in 1% formaldehyde to cross-link DNA–

protein complexes. The fixed cells were lysed, sonicated andcleared with Protein G-beads (Sigma) followed by incubationwith control IgG or STAT3 antibodies. Input samples wereremoved from the lysate before antibody incubation. Anti-body–protein–DNA complexes were immunoprecipitated byprotein G beads. Genomic DNA was purified and subjectedto PCR togetherwith input samples. The primers usedwere shownin Supplementary Table S1.

Transwell assayAliquots of 105 Gr-1þCD11bþ cells were added to the upper

chambers with 5-mm pore inserts (Corning). Supernatants fromvarious fibroblasts were added to the lower chambers with orwithout CCL2 (10 ng/mL, peprotech), neutralizing anti-CCL2mAb (500 ng/mL, eBioscience). Cells that had migrated to thelower chambers were counted by CyAn (Beckman Coulter) after4-hour incubation.

ImmunohistochemistryAfter antigen retrieval, the slides were incubated with the

antibodies against FAP (SAB2900181, Sigma), CCL2 (ab9669,Abcam), p-STAT3 (Tyr705; D3A7, Cell Signaling Technology)followed by secondary antibody incubation (GK500705, GeneTech). Integrated optical density of all the positive staining in eachphotograph was calculated, and its ratio to total area of eachphotograph was considered as FAP, CCL2, and p-STAT3 density,and was categorized as either "Low" or "High" based on themedian value.

Statistical AnalysisTwo-tailed Student t tests or ANOVA followed by a Bonferroni

correction were performed using GraphPad Prism 6 to determinethe statistical significance of difference between two groups orMultiple-group comparisons. Survival was determined by theKaplan–Meier method and survival curves between differentgroups were calculated by log-rank test. A Pearson correlationwas used to evaluate the association of the staining density.Statistical significance was defined as �, P < 0.05; ��, P < 0.01;��, P < 0.001; ns ¼ no significance.

ResultsFAP is required for a subset of inflammatory CAFs

The purity of CAFs was verified by expression of fibroblast-specific protein vimentin and lack of expression of markers ofother major stromal cell types (Supplementary Fig. S1A and B).Consistent with previous studies (8, 15), CAFs were inflam-matory characterized by markedly upregulated expression ofgenes related with inflammation and immune cell chemotaxis,among which Ccl2 was highly expressed compared with NFs(Supplementary Fig. S2A). As expected, FAP was detected inCAFs, but not NFs (Fig. 1A and data not shown). Interestingly,although FAP expression was completely overlapped witha-SMA that was expressed in the majority of CAFs, only asubset of a-SMAþCAFs expressed FAP (Fig. 1A), suggesting thatFAPþCAFs could serve some distinct function. To assess thecontribution of FAP to the inflammatory phenotype of CAFs,we FACS sorted two subsets of CAFs based on FAP expression.FAPþCAFs, which constituted about 6.7% of total CAFs,

expressed much higher levels of most of inflammatory genesthan FAP�CAFs (Fig. 1B and C). Consistently, Ccl2 is thehighest expressed inflammatory gene in FAPþCAFs (Fig. 1C).We next examined the influence of FAP on the activation ofSTAT3 and NF-kB, two important transcriptional factorsthat are closely involved in tumor-associated inflammation(17, 18). Although both STAT3 and NF-kB were activated inCAFs (Supplementary Fig. S2B), FAPþCAFs had higher ratio ofp-STAT3/STAT3 levels, but similar p-p65/p65 levels comparedwith FAP�CAFs (Fig. 1D), indicating that FAP could activateSTAT3. These findings were further confirmed by the findingsthat siRNA FAP knockdown in CAFs markedly reduced theexpression of inflammatory genes and p-STAT3 levels, but onlyslightly reduced p-p65 levels (Supplementary Fig. S2A and B).Similar results were obtained in CAFs, in which STAT3 wasknocked down (Supplementary Fig. S2A), indicating thatFAP-mediated inflammatory CAFs was dependent on STAT3.Collectively, these results suggest that FAPþCAFs represent asubset of inflammatory CAFs characterized by STAT3 activationand upregulated expression of inflammatory genes, particularlyCcl2.

FAP induces inflammatory fibroblasts by activating STAT3through the FAK–c-Src–JAK2–STAT3 pathway, which isdependent on uPAR

We next investigated the molecular mechanism underlyingFAP-induced inflammatory CAFs. Although FAP with shortintracellular domain is unlikely to initiate the activation ofthe signaling pathway itself, it was reported to form close asso-ciation with uPAR (19). Unlike FAP, uPAR was expressed inboth NFs and CAFs at similar levels (Fig. 2A). The interactionbetween FAP and uPAR in CAFs was confirmed by Co-IP(Fig. 2B). uPAR is reported to frequently cause activation ofFAK, an important upstream signaling molecule that drivesvarious tumor-promoting signaling pathways (20, 21). Wefound that FAPþCAFs had significantly higher ratio of p-FAK/FAK levels than FAP�CAFs (Fig. 2C). Similarly, FAP knockdowngreatly reduced p-FAK levels in CAFs (Supplementary Fig. S2B).The paralleled decreased levels of p-FAK and p-STAT3 followinglack of FAP suggested a possible role of FAK in FAP-inducedSTAT3 activation. Furthermore, the FAK inhibitor markedly re-duced p-STAT3 levels in CAFs (Fig. 2F). To elucidate the specificsignaling pathway initiated by FAK to activate STAT3, CAFs weretreated with inhibitors of several downstream signaling mole-cules of FAK. The inhibition of c-Src, but not ERK, AKT or p38,strikingly reduced p-STAT3 levels and CCL2 secretion by CAFs(Fig. 2D and E), indicating c-Src as the major downstreammolecule of FAK. c-Src could activate STAT3 directly or indi-rectly through JAK2 (22, 23). We demonstrated that CAFs hadconcomitantly elevated levels of p-c-Src, p-JAK2 and p-STAT3,which were markedly decreased in FAP�CAFs or FAK inhibitor-treated CAFs (Fig. 2F). Furthermore, c-Src inhibitor reducedlevels of p-JAK2 and p-STAT3, while JAK2 inhibitor onlyreduced the levels of p-STAT3, but not p-c-Src (Fig. 2F). Theinhibitor of JAK2 and c-Src comparably reduced CCL2 levels(Fig. 2E). These data suggested that c-Src-induced STAT3 acti-vation is, at least partly, dependent on JAK2.

We further demonstrated that FAP expression was sufficientto endow NFs with the inflammatory phenotype by engineer-ing FAP expression by transducing NFs with FAP lentivirus. TheeGFP-positive fibroblasts were FACS sorted, and FAP expression

FAP via STAT3–CCL2 Promote Tumor Immunosuppression

www.aacrjournals.org Cancer Res; 76(14) July 15, 2016 OF3




was confirmed by Western blotting (Supplementary Fig. S3Aand B). We found that FAP expression greatly increased levelsof p-FAK, p-c-Src, p-JAK2, and p-STAT3 (Fig. 3A), with no effecton the expression of uPAR and a-SMA in NFs (data not shown).A strikingly increased luciferase activity in FAP-expressing fibro-blasts (denoted as FAP-Fbs) that were transfected with a STAT3reporter vector and a control vector was detected comparedwith similarly treated control fibroblasts (Fig. 3B), emphasizingthe ability of FAP to induce STAT3 transcriptional activity.Similarly to CAFs, FAP-Fbs had markedly upregulated expres-sion of inflammatory genes, particularly Ccl2 (Fig. 3C), butonly had mildly increased expression of genes that are relat-ed to angiogenesis and collagen synthesis (SupplementaryTable S2), suggesting that FAP preferably induced inflamma-tory CAFs. Moreover, STAT3 knockdown completely abrogatedupregulation of inflammatory gene expression in FAP-Fbs(Fig. 3C), with no effect on control fibroblasts (SupplementaryFig. S4). These data reinforce that FAP is an activator of fibro-blastic STAT3, which in turn mediates upregulated inflamma-tory gene expression. Furthermore, we demonstrated that FAP-induced inflammatory fibroblasts were dependent on uPAR,as uPAR knockdown almost reduced the phosphorylated levelsof FAK, c-Src, JAK2, and STAT3 as well as the expression ofinflammatory genes in FAP-Fbs to the baseline levels (Fig. 3Aand C), with no effect on control fibroblasts (SupplementaryFig S4). Given that FAP shares the highest similarity with

dipeptidyl peptidase IV (DDPIV), and inhibition of DDP activ-ity of FAP was reported to attenuate tumor growth (13, 24), weinvestigated the role of the DDP enzymatic activity in FAP-induced inflammatory CAFs. Addition of PT-100, an inhibitorof DDPIV (25), did not affect p-STAT3 levels and CCL2 secre-tion in CAFs or FAP-Fbs (Fig. 3D and E), suggesting that DDPactivity might not be required by FAP to induce inflammatoryCAFs. Collectively, these results demonstrate that FAP inducesinflammatory CAFs by activating STAT3 through the uPAR–FAK–c-Src–JAK2 pathway in a uPAR-dependent way.

Fibroblastic FAP expression is critical for the ability of CAFsto promote tumor growth and immunosuppression in aSTAT3-dependent way

To determine the role of FAP in CAF-mediated tumor-pro-moting inflammation, we first compared the tumor-promotingeffect of CAFs, FAPþCAFs and FAP�CAFs by coinjection oftumor cells and fibroblasts into immunocompetent syngeneicWT mice. Coinjection with CAF or FAPþCAFs comparablypromoted Hepa1-6 tumor growth, while FAP�CAFs failedto (Fig. 4A), suggesting a critical role of FAP mediatingtumor-promotion of CAFs. To further elucidate the underlyingmechanism, we next examined immune cell composition intumor sites. Abundant presence of myeloid-origin cells, partic-ularly myeloid-derived suppressor cells (MDSC), includingPMN-MDSCs and M-MDSCs, is a hallmark of tumor-promoting

Figure 1.

FAPþCAFs represent a subset of inflammatory CAFs characterized by STAT3 activation and highly upregulated Ccl2 expression. A, representative images ofimmunofluorescence staining of cultured CAFswitha-SMA (red) and FAP (green) antibodies and counterstainedwith DAPI (blue; magnification,�200). Scale bars,50 mm. B, FAP-positive (FAPpos) and FAP-negative (FAPneg) CAFs were sorted from total CAFs. C, qRT-PCR analysis of inflammatory gene expression. D,representativeWestern blots showing FAP and the total and phosphorylated STAT3 and p65, and densitometry quantification. Data, mean� SEM. n¼ 4. Results arerepresentative of at least three independent experiments. ns, not significant.

Yang et al.





inflammation, and implicated in directly promoting tumorprogression primarily by their potent immunosuppressive activ-ities (26, 27). We found that coinjected with CAFs or FAPþCAFs,but not FAP�CAFs, significantly increased frequencies of PMN-MDSCs and M-MDSCs in Hepa1-6 tumors (Fig. 4B). Signifi-cantly increased frequencies of macrophages were also found(Fig. 4B), which is consistent with a previous study (8). Accord-ingly, coinjection with CAFs or FAPþCAFs, but not FAP�CAFs,significantly increased the expression of genes that are related toMDSCs andmacrophages (Fig. 4C). In contrast, coinjection withCAFs or FAPþCAFs, but not FAP�CAFs, significantly decreasedfrequencies of IFNgþCD8þT cells (Fig. 4D). Coinjection of CAFsor FAPþCAFs comparably influenced tumor infiltration ofMDSCs and IFNgþT cells (Fig. 4B and D), although FAPþCAFsseemed to cause higher expression of S100a8 and Arg-1(Fig. 4C). Similar results were also obtained from the coinjec-tion of CAFs in which STAT3 or FAP were knocked down byshRNA (Fig. 4A–D; Supplementary Fig. S5A–D).

We also investigated whether the tumor-promoting effect ofFAPþCAFs could be due to its direct ability to promote the cancerstemness of Hepa1-6 cells or the function of immune cells. Wedemonstrated that FAP expression did not affect the stemness-

supporting ability of CAFs, as supernatants from CAFs andFAPþCAFs comparably upregulated stemness-related gene expres-sion and the proliferation of Hepa1-6 cells (SupplementaryFig. S6A–B). Furthermore, supernatants from FAPþCAFs hadsimilar ability to upregulate the expression of immunosuppres-sive genes of MDSCs, but were slightly more effective in enhanc-ing the ability of MDSCs to inhibit T-cell proliferation comparedwith those fromCAFs (Supplementary Fig. S6C–D). Supernatantsfrom CAFs and FAPþCAFs had similar ability to inhibit Hepa1-6cells killing by CD8þ T cells, which was comparable with MDSCs(Supplementary Fig. S6E). These data suggest that FAP may notbe a major contributor to the ability of CAFs to regulate thefunction of MDSCs and CD8þT cells. These data together sug-gest that the FAP–STAT3 axis is responsible for the ability ofCAFs to mediate tumor immunosuppression primarily by pro-moting infiltration of MDSCs, leading to impaired antitumorT-cell immunity in the tumor microenvironment.

CCL2 derived from FAPþCAFsmediates their ability to promotetumor growth and MDSC infiltration

We further confirmed the critical role of the FAP–STAT3 axisin CAF-mediated tumor growth and immunosuppression by

Figure 2.

FAP is responsible for inflammatory phenotype of CAFs by activating the uPAR–FAK–c-Src–JAK2–STAT3 pathway. A, representative flow cytometric analysisof uPAR expression. B, CoIP assays to analyze the interaction between FAP and uPAR in CAFs. Normal rabbit/goat IgG antibodies were served as anegative control. C, representative Western blots showing the total and phosphorylated FAK, and densitometry quantification. D, representative Westernblots showing the effect of different inhibitors of downstream molecules of FAK on STAT3 activation in CAFs. E, ELISA analysis of CCL2 concentrations in theculture of CAFs treated with DMSO or different inhibitors. �� , P < 0.01; �� , P < 0.001 versus CAFs treated with DMSO. ns, not significant. F, representativeWestern blots showing the phosphorylation of c-Src, JAK2, and STAT3 in CAFs and FAPnegCAFs, and the effect of inhibitors of FAK, c-Src, and JAK2 on thephosphorylation of c-Src, JAK2, and STAT3 in CAFs. DMSO was used as a vehicle control. In C and E, data represented as mean � SEM. Results arerepresentative of at least three independent experiments.






showing that forced FAP expression endowed NFs with tumor-promoting activities similar to FAPþCAFs (Fig. 5A–D), and STAT3knockdown completely abrogated the tumor-promoting effect(Fig. 5A–D). We next attempted to identify the critical target geneof STAT3 that is responsible for the tumor-promoting activities ofFAPþCAF. We chose to focus on CCL2 because it is the mostabundant gene expressed in FAPþCAFs, and a well-known che-moattractant for leukocytes (28). We first demonstrated thatfibroblastic STAT3 activated by FAP could bind to Ccl2 promoterusing chromatin immunoprecipitation (ChIP) assay, as a signif-icantly increased loading of STAT3 onto theCcl2promoter (�190/�3) was detected in FAP-Fbs compared with control fibroblasts(Fig. 5E), indicating that STAT3 could directly regulate CCL2transcription. To determine the relative contribution of CAFs toCCL2 expression within tumors, we analyzed CCL2 gene expres-sion of different cell populations sorted from Hepa1-6 tumors.CAFs, particularly FAPþCAFs, expressed much higher Ccl2 genelevels than other cell populations (Fig. 5F). Furthermore, CAFs,

including FAPþCAFs and FAP�CAFs, secreted much more CCL2than several tumor cells lines (Fig. 5G). These data identified FAPþ

CAFs as the major cell source of CCL2 within tumors. Moreimportantly, specific CCL2 knockdown in FAP-Fbs dramaticallyimpaired their ability to promote tumor growth, completelyabrogated the increases in the infiltration of MDSCs and macro-phages and related gene expression, and restored infiltratingantitumor IFNgþ T cells in Hepa1-6 tumor (Fig. 5A–D). Theseresults were similar to those from STAT3 knockdown in FAP-Fbs(Fig. 5A–D). A previous study showed that CAF-derived CCL2could support the stemness of breast cancer cells (29); however,the cancer stemness-promoting ability of FAPþCAFs seemed to beCCL2- independent, as CCL2 neutralization did not affect stem-ness-related gene expression and proliferation of Hepa1-6 cells(Supplementary Fig S5A–B). Collectively, these data demonstratethat fibroblastic CCL2 expression is under STAT3 control andcausally linked to the ability of FAP-expressing CAFs to promotetumor growth and immunosuppressive microenvironment.

Figure 3.

FAP activates fibroblastic STAT3 in a uPAR-dependent way. A, representative Western blots showing the phosphorylation of FAK, c-Src, JAK2, andSTAT3. B, luciferase assays using the STAT3 reporter vector. C, qRT-PCR analysis of inflammatory gene expression. Insets, Western blots showingthe effect of uPAR siRNA. D and E, representative Western blots showing the total and phosphorylated STAT3 (D) and CCL2 concentrations (E). CAFsand FAP-Fbs were treated with or without PT-100 (5 nmol/L), the inhibitor of DDPIV. Data, mean � SEM. Results are representative of at leastthree independent experiments. ns, not significant.

Yang et al.





CCR2 is required for tumor-promoting activities ofFAP-expressing fibroblasts by mediating tumor recruitmentof MDSCs

CCL2 mediates tissue recruitment of inflammatory cells byinteracting with its receptor CCR2 (28, 30). We found that CCR2was expressed in both circulating and tumor-infiltrating MDSCsand tumor-infiltrating macrophages, with higher levels in circu-lating MDSCs (Fig. 6A). We sorted Gr-1þCD11bþ cells thatinclude PMN-MDSCs and M-MDSCs from bone marrow ofHepa1-6 tumor-bearingmice to examine theirmigration to super-natants from various fibroblast cultures. Supernatants from CAFsor FAP-Fbs attracted significantly more Gr-1þCD11bþ cells thanthose from control fibroblasts, which was greatly impaired byknockdown of FAP or STAT3, while adding exogenous CCL2could rescue Gr-1þCD11bþ cell chemotaxis (Fig. 6B). In contrast,neutralizing anti-CCL2 antibody or CCL2 knockdown completelyblocked themigrationofGr-1þCD11bþ cells toward supernatantsfrom FAP-Fbs (Fig. 6B). Moreover, Ccr2�/� Gr-1þCD11bþ cellsexhibited much less chemotactic activity toward supernatantsfrom FAP-Fbs compared with WT controls (Fig. 6C). Collectively,these results demonstrate that FAP-expressing CAFs-derivedCCL2

mediates migration of Gr-1þCD11bþ cells in a CCR2-dependentway. To further determine whether the tumor-promoting activi-ties of FAP-expressing fibroblasts was dependent on CCR2,FAP-Fbs or control fibroblasts with Hepa1-6 cells were coinjectedinto WT or Ccr2�/� mice. FAP-Fbs failed to promote tumorgrowth in Ccr2�/� mice, which was accompanied by significantdecreases in tumor infiltration of MDSCs and macrophagesand related gene expression and significantly increased infiltratingIFNgþ T cells (Fig. 6D–G). Notably, evenHepa1-6 cells coinjectedwith control fibroblasts grew much more slowly accompaniedby less infiltratingMDSCs andmacrophages inCcr2�/�mice thanthose in WT mice (Fig. 6D–G), further emphasizing a criticalrole of CCR2 for tumor growth and MDSC recruitment. Collec-tively, these data demonstrate that CCR2-dependent recruit-ment of MDSCs in response to CCL2 produced by FAPþCAFsis important to promote tumor growth.

FAP correlates with p-STAT3 and CCL2 expression in tumorstroma and predicts poor prognosis of ICC patients

To investigate the relevance of ourfindings tohuman tumor,westudied ICC, a highly aggressive human primary liver tumor,

Figure 4.

The lack of FAPþCAFs or STAT3 knockdown greatly impairs the ability of CAFs to promote tumor growth and MDSC infiltration. A, Hepa1-6 cells were injectedalone or coinjected with various fibroblasts subcutaneously into WT mice, respectively. Insets, Western blots showing the effect of STAT3 shRNA. Tumorgrowth was measured at the indicated time points. B, representative flow cytometry data and averaged percentages of PMN-MDSCs, M-MDSCs, andmacrophages. C, qRT-PCR analysis of tumor tissues for MDSCs and macrophage-related gene expression. D, representative flow cytometry data and averagedpercentages of IFNgþCD8þ T cells and IFNgþCD4þ T cells. Data, mean � SEM. n ¼ 6–8. Results are representative of two independent experiments.






because desmoplasia/fibrosis, as a hallmark of ICC, plays anactive role in ICC progression (31). To investigate whether theFAP–STAT3–CCL2 axis in CAFs was associated with ICC char-acteristics and prognosis, we evaluated the relationship betweenexpression of FAP, p-STAT3, and CCL2 in tumor stroma of hu-man ICC tissue microarrays. Varied expression of FAP, p-STAT3,and CCL2 was found in tumor stroma, which exhibited a signif-icant positive correlation between each other (Fig. 7A). We nextcorrelated relevant clinical data of ICC patients (SupplementaryTable S3) with tumor stromal FAP levels to determine whetherincreased stromal FAP was associated with ICC progression.Stromal expression of FAP was positively associated with tumorsize and lymphonudus metastasis, but not tumor differentiation(Supplementary Table S4), suggesting that FAP may not directlyinfluence tumor cells. More importantly, ICC patients with highlevels of stromal FAP had significantly lower survival rates, andhigher cumulative recurrence rates than those with low levels ofstromal FAP (Fig. 7B). Consistently, high levels of p-STAT3 orCCL2 in ICC tumors were negatively associated with ICC patient

survival (Supplementary Fig S7A and B). These data collectivelyindicate the adverse predictive role of FAP in the clinic outcome ofICC patients.

DiscussionOur study offers a molecular and cellular mechanism by

which FAP endows CAFs with the ability to mediate tumor-promoting inflammation and immunosuppression. Specifical-ly, FAP induces inflammatory CAFs by activating STAT3through the uPAR–FAK–c-Src–JAK2 pathway, and FAPþCAFsare the major cell source of CCL2, which promotes the tumorrecruitment of MDSCs and immunosuppression in a CCR2-dependent way, leading to tumor growth (Fig. 7C).

The role of STAT3 in cancer has largely been focused on tumorcells and myeloid cells (18, 32); however, we showed robustSTAT3 activation in CAFs, which is consistent with several studies(29, 33), and identify FAP as a persistent activator of fibroblasticSTAT3, which in turn mediates inflammatory properties of

Figure 5.

FAP–STAT3–CCL2 axis is critical for the ability of fibroblasts to promote tumor growth and MDSC infiltration. A, Hepa1-6 cells were injected alone orcoinjected with various fibroblasts subcutaneously into WT mice, respectively. Insets, qRT-PCR showing the effect of CCL2 shRNA. Tumor growth wasmeasured at the indicated time points. B, the averaged percentages of PMN-MDSCs, M-MDSCs, and macrophages. C, qRT-PCR analysis of tumortissues for MDSCs and macrophages-related gene expression. D, the averaged percentages of IFNgþCD8þ T cells and IFNgþCD4þ T cells. E, ChIP assaysshowing STAT3 binding to the CCL2 promoter. Ccl5 promoter (�120/�1) was used as a positive control and Ccl2 promoter (�1379/�1245) that isknown for NF-kB binding as a negative control. Conventional PCR (left) and qRT-PCR (right) were performed to amplify Ccl2 promoter region usinganti-STAT3 antibody immunoprecipitated and input DNA fragments, respectively. n ¼ 3. F, qRT-PCR analysis of CCL2 gene expression in different celltypes sorted from tumor tissues. n ¼ 3. G, CCL2 concentrations in cell cultures. Data, mean � SEM. n ¼ 8. Results are representative of at least threeindependent experiments. ns, not significant.

Yang et al.





FAPþCAFs. We further reveal that FAP-induced STAT3 activationrequires the interaction between FAP and uPAR to activate FAK–c-Src–JAK2 signaling. Both uPAR and FAK are known to partic-ipate in tissue remodeling and cancer development (21, 34). FAKwas recently demonstrated to regulate skin fibrosis throughactivation of ERK–CCL2 signaling. However, our study identifiedc-Src, but not ERK, as the major downstream molecule of FAKto activate JAK2–STAT3 signaling, thereby leading to increas-ed expression of inflammatory genes, particularly Ccl2, inFAPþCAFs. Thus, our findings point out the important role ofuPAR–FAK signaling for FAPþCAFs to acquire the inflammatoryphenotype through STAT3 activation.

Moreover, we demonstrate that inflammatory FAPþCAFsplay a critical role in CAF-mediated tumor-promoting inflam-mation in a STAT3-dependent way. Interestingly, Erez andcolleagues reported that NF-kB is required by CAF-mediatedtumor-promoting inflammation (8). We detected NF-kB acti-vation in CAFs, which, however, was not affected by FAPexpression, suggesting a minor role of FAP in fibroblastic

NF-kB activation. Notably, CAFs used by the study of Erez andcolleagues were isolated from an incipient mouse skin cancerand defined by PDGFR-a that is extensively expressed by allfibroblast, while CAFs used by our study were isolated fromestablished subcutaneous tumors and defined by FAP that wasmore restrictively expressed. Thus, it is likely that NF-kB andSTAT3 play important roles in a different subset of CAFs, whichcould make their respective contribution to CAF-mediatedtumor-promoting inflammation at different stages of tumordevelopment.

Another important finding of our study is uncovering amolecular and cellular mechanism underlying FAPþCAF-medi-ated tumor immunosuppression. We demonstrate thatFAPþCAFs were able to promote infiltration of MDSCs, whichcould skew a more immunosuppressive tumor microenviron-ment to antagonize antitumor IFNgþT-cell immunity. Addi-tionally, increased tumor-infiltrating macrophages were alsoobserved in tumors coinjected with FAPþCAFs. This could bedue to increased infiltrating MDSCs, as MDSCs were reported to

Figure 6.

FAP-expressing fibroblasts fail to promote tumor growth and tumor infiltration of MDSCs in Ccr2�/� mice. A, representative flow cytometric analysis ofCCR2 expression. B, Transwell assays of Gr-1þCD11bþ cell chemotaxis toward supernatants with or without CCL2 protein or anti-CCL2 mAb. C, thechemotaxis activity of WT or Ccr2�/� Gr-1þCD11bþ cell toward supernatants from Ctrl-Fbs and FAP-Fbs. n ¼ 3. D, Hepa1-6 cells were coinjected withCtrl-Fbs or FAP-Fbs subcutaneously into WT or Ccr2�/� mice. Tumor growth was measured at the indicated time points. n ¼ 7. E, the averagedpercentages of PMN-MDSCs, M-MDSCs, and macrophages. F, qRT-PCR analysis of tumor tissues for MDSCs and macrophages-related gene expression.G, the averaged percentages of IFNgþCD8þ T cells and IFNgþCD4þ T cells. Data, mean � SEM; n ¼ 7. Results are representative of at least threeindependent experiments.






preferentially differentiate into macrophages in tumor sites(31). However, we still could not exclude other mechanismsby which MDSCs or macrophages contribute to tumor growth.For example, MDSCs and macrophages were recently reportedto promote cancer stemness (35, 36). We also confirmed it byshowing that coculture of MSDCs enhanced the expressionof stemness-related genes of Hepa1-6 cells and their prolifer-ation. Thus, it is likely that the tumor-promoting effect ofFAPþCAFs is partly mediated by attracting more MDSCsor macrophages to tumor sites, where they enhance stemnessof cancer cells and/or promote expansion of cancer stem cells.Furthermore, we identified FAPþCAFs as the major source ofCCL2 within tumors, and Ccl2 as the major target geneof STAT3 in FAPþCAFs to mediate their tumor-promotingactivities by attracting MDSCs to tumor sites in a CCR2-depen-dent way. This is also supported by a recent study showingthat CCL2 mediates the recruitment of CCR2-expressingGr-1þinflammatory monocytes to promote lung metastasis(37). Although CCL2-derived from tumor cells was also report-ed to mediate T-cell migration (28, 38), including regulatory

T cells, another important immunosuppressive cells, coinjec-tion of FAPþCAFs or fibroblastic knockdown of CCL2 had noeffect on tumor-infiltrating regulatory T cells (data not shown),and even caused decreased IFNgþT cell infiltration. Given thattyrosine nitrated CCL2 attracted myeloid cells but not T cellsand tyrosine nitration was reported in protumorigenic fibro-blasts (39, 40), this could explain why CCL2 derived fromFAPþCAFs was effective in attracting MDSCs, but not T cells, totumor sites. CXCL1/2 was reported to be important for therecruitment of MDSCs to sites of inflammation or cancer(41, 42). Despite mildly upregulated CXCL1/2 expression inFAPþCAFs, we found that tumor cells, but not CAFs, were themajor source of CXCL1/2 (unpublished data, Lin and collea-gues), which is also consistent with a recent study showing thattumor cell–derived CXCL1/2 mediates the tumor recruitmentof CD11bþGr-1þmyeloid cells (42).

Recently, FAPþ cells were reported to express different func-tional proteins mediating essential physiological functions inskeletal muscle and bone marrow (43). Combined with ourfinding that FAPþCAFs mediate tumor via CCL2-induced MDSC

Figure 7.

Stromal expression of FAP, CCL2, and p-STAT3 is positively correlated in human ICC, and FAP predicts poor prognosis of ICC patients. A, representativeIHC staining of FAP, p-STAT3, and CCL2 in human ICC samples (magnification, �100). Scale bars, 200 mm. Pearson correlations between stromalexpression of FAP, p-STAT3, and CCL2. n ¼ 196. B, Kaplan–Meier survival or recurrence analysis of human ICC tissue microarray based on FAP expressionlevels in tumor stroma. n ¼ 196. C, schematic diagram of the cellular and molecular mechanism by which FAP induces inflammatory CAFs topromote tumor growth.

Yang et al.





recruitment, it suggests that the function of FAPþ cells could betissue specific. Furthermore, recent studies suggested that thesystemic ablation of FAPþ cells may lead to unexpected sideeffects, such as bone toxicity and cachexia (43, 44). Thus, ourfindingsmay provide some new insight into a specificmechanismthat mediates the function of FAPþ cells in different tissues underphysiologic or pathologic conditions, which could be importantto design safer andmore specific therapeutic strategy against FAPþ

cells in tumor or other diseases.In summary, our present study provides mechanistic insights

into the causal link between FAP and a subset of inflamma-tory CAFs that mediate tumor-promoting inflammationand immunosuppression, and highlight the important roleof STAT3–CCL2 signaling in the tumor-promoting effect ofinflammatory FAPþCAFs primarily by skewing the immuno-suppressive microenvironment through the promotion of tu-mor recruitment of MDSCs. Therefore, our study suggests thepossibility of specific targeting FAP and its downstream factorsto re-educate the inflammatory CAFs for the treatment ofcertain types of cancer that are associated with inflammationand desmoplasia.

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

Authors' ContributionsConception and design: X. Yang, Y. Lin, Y. Shi J. Fan, R. HeDevelopment of methodology: B. LiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): X. Yang, Y. Lin, Y. Shi, B. Li, W. Liu, W. Yin, R. HeAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, com-putational analysis): X. Yang, Y. Lin, Y. Shi, B. Li, W. Liu, W. Yin, J. Fan, R. HeWriting, review, and/or revision of the manuscript: Y. Lin, R. HeAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): Y. Dang, Y. Chu, R. HeStudy supervision: R. He

AcknowledgmentsThe authors thankDr. Linrong Lu (ZhejiangUniversity) for critical reading of

the manuscript.

Grand SupportThis work is supported by NSFC grants 813220437, 81471555 (R. He.)

and 81272389 (Y. Shi), National Key Sci-Tech Project 2012ZX10002011-002(J. Fan).

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 October 27, 2015; revisedMarch 9, 2016; accepted March 23, 2016;published OnlineFirst May 23, 2016.

References1. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer 2006;6:

392–401.2. Ohlund D, Elyada E, Tuveson D. Fibroblast heterogeneity in the cancer

wound. J Exp Med 2014;211:1503–23.3. RhimAD,Oberstein PE, ThomasDH,Mirek ET, PalermoCF, Sastra SA, et al.

Stromal elements act to restrain, rather than support, pancreatic ductaladenocarcinoma. Cancer Cell 2014;25:735–47.

4. Ozdemir BC, Pentcheva-Hoang T, Carstens JL, Zheng X, Wu CC, SimpsonTR, et al. Depletion of carcinoma-associated fibroblasts and fibrosisinduces immunosuppression and accelerates pancreas cancer with reducedsurvival. Cancer Cell 2014;25:719–34.

5. Desmouliere A,GuyotC,Gabbiani G. The stroma reactionmyofibroblast: akey player in the control of tumor cell behavior. Int J Dev Biol 2004;48:509–17.

6. Santos AM, Jung J, Aziz N, Kissil JL, Pure E. Targeting fibroblast activationprotein inhibits tumor stromagenesis and growth in mice. J Clin Invest2009;119:3613–25.

7. Trimboli AJ, Cantemir-Stone CZ, Li F, Wallace JA, Merchant A, Creasap N,et al. Pten in stromal fibroblasts suppresses mammary epithelial tumours.Nature 2009;461:1084–91.

8. Erez N, Truitt M, Olson P, Arron ST, Hanahan D. Cancer-associatedfibroblasts are activated in incipient neoplasia to orchestrate tumor-pro-moting inflammation in an NF-kappaB-dependent manner. Cancer Cell2010;17:135–47.

9. Cortez E, Roswall P, Pietras K. Functional subsets ofmesenchymal cell typesin the tumor microenvironment. Semin Cancer Biol 2014;25:3–9.

10. Osterreicher CH, Penz-Osterreicher M, Grivennikov SI, GumaM, KoltsovaEK, Datz C, et al. Fibroblast-specific protein 1 identifies an inflammatorysubpopulation of macrophages in the liver. Proc Natl Acad Sci U S A2011;108:308–13.

11. Scanlan MJ, Raj BK, Calvo B, Garin-Chesa P, Sanz-Moncasi MP, HealeyJH, et al. Molecular cloning of fibroblast activation protein alpha, amember of the serine protease family selectively expressed in stromalfibroblasts of epithelial cancers. Proc Natl Acad Sci U S A 1994;91:5657–61.

12. LevyMT,McCaughanGW, Abbott CA, Park JE, CunninghamAM,Muller E,et al. Fibroblast activation protein: a cell surface dipeptidyl peptidaseand gelatinase expressed by stellate cells at the tissue remodelling inter-face in human cirrhosis. Hepatology 1999;29:1768–78.

13. Cheng JD, Dunbrack RL Jr., Valianou M, Rogatko A, Alpaugh RK, WeinerLM. Promotion of tumor growth bymurine fibroblast activation protein, aserine protease, in an animal model. Cancer Res 2002;62:4767–72.

14. Lo A, Wang LCS, Scholler J, Monslow J, Avery D, Newick K, et al. Tumor-promoting desmoplasia is disrupted by depleting FAP-expressing stromalcells. Cancer Res 2015;75:2800–10.

15. Kraman M, Bambrough PJ, Arnold JN, Roberts EW, Magiera L, Jones JO,et al. Suppression of antitumor immunity by stromal cells expressingfibroblast activation protein-alpha. Science 2010;330:827–30.

16. Feig C, Jones JO, Kraman M, Wells RJ, Deonarine A, Chan DS, et al.Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblastssynergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proc NatlAcad Sci U S A 2013;110:20212–7.

17. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation.Nature 2008;454:436–44.

18. Yu H, Pardoll D, Jove R. STATs in cancer inflammation and immunity:a leading role for STAT3. Nat Rev Cancer 2009;9:798–809.

19. Artym VV, Kindzelskii AL, Chen WT, Petty HR. Molecular proximity ofseprase and the urokinase-type plasminogen activator receptor on malig-nant melanoma cell membranes: dependence on beta(1) integrins andthe cytoskeleton. Carcinogenesis 2002;23:1593–601.

20. Sulzmaier FJ, Jean C, Schlaepfer DD. FAK in cancer: mechanistic findingsand clinical applications. Nat Rev Cancer 2014;14:598–610.

21. Smith HW, Marshall CJ. Regulation of cell signalling by uPAR. Nat RevMol Cell Bio 2010;11:23–36.

22. Bowman T, Broome MA, Sinibaldi D, Wharton W, Pledger WJ, Sedivy JM,et al. Stat3-mediatedMyc expression is required for Src transformation andPDGF-induced mitogenesis. Proc Natl Acad Sci U S A 2001;98:7319–24.

23. Bai KJ, Chen BC, Pai HC, Weng CM, Yu CC, Hsu MJ, et al. Thrombin-induced CCN2 expression in human lung fibroblasts requires the c-Src/JAK2/STAT3 pathway. J Leukoc Biol 2013;93:101–12.

24. Christiansen VJ, Jackson KW, Lee KN, Downs TD, McKee PA. Targetinginhibition of fibroblast activation protein-alpha and prolyl oligopepti-dase activities on cells common to metastatic tumor microenvironments.Neoplasia 2013;15:348–58.

25. Adams S, Miller GT, JessonMI, Watanabe T, Jones B, Wallner BP. PT-100, asmall molecule dipeptidyl peptidase inhibitor, has potent antitumoreffects and augments antibody-mediated cytotoxicity via a novel immunemechanism. Cancer Res 2004;64:5471–80.






26. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated regulationof myeloid cells by tumours. Nat Rev Immunol 2012;12:253–68.

27. Ostrand-Rosenberg S, Sinha P. Myeloid-derived suppressor cells: linkinginflammation and cancer. J Immunol 2009;182:4499–506.

28. Deshmane SL, Kremlev S, Amini S, Sawaya BE. Monocyte chemoattrac-tant protein-1 (MCP-1): an overview. J Interferon Cytokine Res 2009;29:313–26.

29. Tsuyada A, Chow A, Wu J, Somlo G, Chu P, Loera S, et al. CCL2 mediatescross-talk between cancer cells and stromal fibroblasts that regulates breastcancer stem cells. Cancer Res 2012;72:2768–79.

30. Shi C, Pamer EG. Monocyte recruitment during infection and inflamma-tion. Nat Rev Immunol 2011;11:762–74.

31. Sirica AE, Campbell DJ, Dumur CI. Cancer-associated fibroblasts inintrahepatic cholangiocarcinoma. Curr Opin Gastroenterol 2011;27:276–84.

32. Deng J, Liu Y, Lee H, Herrmann A, Zhang W, Zhang C, et al. S1PR1-STAT3signaling is crucial for myeloid cell colonization at future metastatic sites.Cancer Cell 2012;21:642–54.

33. Wang L, Yi T, KortylewskiM, Pardoll DM, ZengD, YuH. IL-17 can promotetumor growth through an IL-6-Stat3 signaling pathway. J Exp Med 2009;206:1457–64.

34. Wong VW, Rustad KC, Akaishi S, SorkinM, Glotzbach JP, JanuszykM, et al.Focal adhesion kinase links mechanical force to skin fibrosis via inflam-matory signaling. Nat Med 2012;18:148–52.

35. Cui TX, Kryczek I, Zhao L, Zhao E, Kuick R, RohMH, et al. Myeloid-derivedsuppressor cells enhance stemness of cancer cells by inducing micro-RNA101 and suppressing the corepressor CtBP2. Immunity 2013;39:611–21.

36. Wan S, Zhao E, Kryczek I, Vatan L, Sadovskaya A, Ludema G, et al. Tumor-associated macrophages produce interleukin 6 and signal via STAT3 to

promote expansion of human hepatocellular carcinoma stem cells.Gastroenterology 2014;147:1393–404.

37. Qian BZ, Li J, Zhang H, Kitamura T, Zhang J, Campion LR, et al. CCL2recruits inflammatory monocytes to facilitate breast-tumour metastasis.Nature 2011;475:222–5.

38. Brown CE, Vishwanath RP, Aguilar B, Starr R, Najbauer J, Aboody KS, et al.Tumor-derived chemokine MCP-1/CCL2 is sufficient for mediating tumortropism of adoptively transferred T cells. J Immunol 2007;179:3332–41.

39. MolonB,Ugel S,Del Pozzo F, Soldani C, Zilio S, AvellaD, et al. Chemokinenitration prevents intratumoral infiltration of antigen-specific T cells. J ExpMed 2011;208:1949–62.

40. Augsten M, Sjoberg E, Frings O, Vorrink SU, Frijhoff J, Olsson E, et al.Cancer-associatedfibroblasts expressingCXCL14Rely uponNOS1-derivednitric oxide signaling for their tumor-supporting properties. Cancer Res2014;74:2999–3010.

41. Sander LE, Sackett SD, Dierssen U, Beraza N, Linke RP, Muller M, et al.Hepatic acute-phase proteins control innate immune responses duringinfection by promoting myeloid-derived suppressor cell function. J ExpMed 2010;207:1453–64.

42. Acharyya S, Oskarsson T, Vanharanta S,Malladi S, Kim J,Morris PG, et al. ACXCL1 paracrine network links cancer chemoresistance and metastasis.Cell 2012;150:165–78.

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

44. Tran E,ChinnasamyD, YuZ,MorganRA, LeeCC, RestifoNP, et al. Immunetargeting of fibroblast activation protein triggers recognition of multi-potent bone marrow stromal cells and cachexia. J Exp Med 2013;210:1125–35.


Yang et al.




Published OnlineFirst May 23, 2016.Cancer Res Xuguang Yang, Yuli Lin, Yinghong Shi, et al. Signaling

CCL2−Fibroblasts in the Tumor Microenvironment via STAT3 FAP Promotes Immunosuppression by Cancer-Associated

Updated version

10.1158/0008-5472.CAN-15-2973doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2016/05/21/0008-5472.CAN-15-2973.DC1

Access the most recent supplemental material at:

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

Subscriptions

Reprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/early/2016/07/04/0008-5472.CAN-15-2973To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-15-2973

http://cancerres.aacrjournals.org/content/suppl/2016/05/21/0008-5472.CAN-15-2973.DC1

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/early/2016/07/04/0008-5472.CAN-15-2973


FAP Promotes Immunosuppression by Cancer-Associated ... · 4/7/2016 · Green Master kit (TaKaRa)...

Documents

Transcript of FAP Promotes Immunosuppression by Cancer-Associated ... · 4/7/2016 · Green Master kit (TaKaRa)...