oncogenic transformation can orchestrate immune evasion ... · alex miranda1, juan m. funes2, nilda...

Microenvironment and Immunology

Oncogenic Transformation Can OrchestrateImmune Evasion and Inflammation in HumanMesenchymal Stem Cells Independently ofExtrinsic Immune-Selective PressureAlex Miranda1, Juan M. Funes2, Nilda S�anchez1, Celia M. Limia1,3, M�onica Mesa1,Sergio A. Quezada2, Rolando P�erez1, and Joel de Le�on1

Abstract

Immune escape is a hallmark of cancer, but whether it reliesupon extrinsic immune-selective pressure or is inherently orches-trated by oncogenic pathways is unresolved. To address thisquestion, we took advantage of an in vitro model of sequentiallytransformed humanmesenchymal stem cells (hMSC). Neoplastictransformation in this model increased the natural immune-evasive properties of hMSC, both by reducing their immunoge-nicity and by increasing their capacity to inhibit mitogen-drivenT-cell proliferation. We also found that IFNg signaling was glob-ally affected in transformed hMSC. As a consequence, the naturalinhibitory effect of hMSCon T-cell proliferation switched from aninduciblemechanismdepending on IFNg signaling andmediated

by indoleamine 2,3-dioxygenase to a constitutivemechanism thatrelied upon IL1b involving both secreted and membrane-expressedmolecules. After transformation, increased IL1b expres-sion both sustained the immunosuppressive properties of hMSCand increased their tumorigenicity. Thus, in this model system,IL1b acted as intrinsic inflammatory mediator that exerted anautocrine influence on tumor growth by coordinately linkingimmune escape and tumorigenicity. Collectively, our findingsshow how oncogenes directly orchestrate inflammation andimmune escape to drive the multistep process of cancer progres-sion, independently of any need for immunoediting in the tumormicroenvironment. Cancer Res; 75(15); 3032–42. �2015 AACR.

IntroductionGenetic and epigenetic somatic lesions that sustain the hall-

marks of cancer (1) are selected by intrinsic and extrinsic pres-sures, including immune surveillance, which persistently operateonmalignant cells during cancer progression. The theory of cancerimmunoediting postulates that effector mechanisms of theimmune system exert a dual role, eliminating cancer cells andpromoting cancer progression through the selection of those cellswhose genetic alterations allow tumor to progress in immuno-competent host (2).

The molecular links between cancer cells and immune sys-tem components extend beyond immune surveillance evasion.Chronic inflammation can trigger the earliest stages of celltransformation (3), whereas intrinsic inflammatory programsprompted by genetic events influence genetic instability andenable the rest of cancer hallmarks, leading to neoplastic pro-

gression (1, 4).Molecular inflammatorymediators can contributeto tumor immune escape, mainly by recruiting regulatoryimmune cells that suppress protective tumor immunity (5).

However, whether immune-suppressive capacities of cancercells evolve along the multistep development of human tumorsindependently of immune-extrinsic pressure and cancer-intrinsicinflammation contributes to concert such immunosuppressivefeature are unsolved and clinically relevant issues. To addressthese questions, we took advantage of an in vitro experimentalmodel in which five-stepwise genetic modifications convert nor-mal humanmesenchymal stem cells (hMSC) into neoplastic cells(6). This is a valuable model because hMSC naturally exert aplethora of modulatory effects on immune cell functions (7, 8).These cells also contribute to the progression of certain tumorseither modifying the microenvironment (9, 10) or behaving astumor-initiating cells (11, 12). In fact, using a more simplisticmodel of transformation based on murine MSC, we have pre-liminary evidence that in vitro–induced neoplastic transformationmodifies the relationship betweenMSC and lymphocytes (13). Inthe human model, the induced sequential transformation gen-erates incomplete or 5 hits fully transformed MSC lines bysequential and accumulative infections with retroviruses encod-ing for hTERT (1 hit), HPV-16 E6 (p53 kd; 2 hits) and E7 (pRb kd;3 hits) proteins, SV40 small T antigen (c-myc activation; 4 hits),and H-RasV12 (H-RAS activation; 5 hits).

Materials and MethodsCell culture

Peripheral bloodmononuclear cells (PBMC) isolated by Ficoll-Paque Plus density gradient (cat. no. 17-1440-02; GE Healthcare

1Tumor Immunology Direction, Center of Molecular Immunology(CIM), Havana, Cuba. 2Cancer Institute, University College of London,London, United Kingdom. 3Tropical Medicine Institute, Havana, Cuba.

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

Current address for J.M. Funes: Spanish National Cancer Research Centre,Madrid, Spain.

CorrespondingAuthor: Joel de Le�on, Center of Molecular Immunology, 216 St. &15th Ave, Atabey, Playa, Havana 11600, Cuba. Phone: 537-214-3178; Fax: 537-272-0644; E-mail: joel.leon@infomed.sld.cu

doi: 10.1158/0008-5472.CAN-14-3276

�2015 American Association for Cancer Research.

CancerResearch

Cancer Res; 75(15) August 1, 20153032

Published OnlineFirst June 11, 2015; DOI: 10.1158/0008-5472.CAN-14-3276

Life Science) from buffy coats of healthy blood donors werecultured in RPMI-1640 medium (Gibco) supplementedwith 10% FCS, 2 mmol/L glutamine, 100 U/mL penicillin,100 mg/mL streptomycin, and 50 mmol/L b-mercaptoethanol.hMSC lines (7) were grown in DMEM medium (Gibco) sup-plemented with 10% FCS. In some experiments, 2 � 105 hMSCwere seeded in 6-well plates and cultured for 72 hours withmedium containing 1.5 ng/mL IFNg (cat. no. 554617; BDPharmingen) and/or 1.5 ng/mL TNFa (cat. no. 210-TA-050;R&D Systems), or 1:2 diluted supernatant collected from PBMCactivated with 5 mg/mL of Concanavalin A (ConA; cat. no.234567; Merck Millipore).

Proliferation assaysPBMC (106) labeled with 0.5 mmol/L of carboxyfluorescein

diacetate succinimidyl ester (CFSE; cat. no. C34554; Life Tech-nologies) were cocultured in 24-well plates with supernatantderived fromhMSCorwith different amounts of hMSCpretreatedor not with 25 mg/mL of mitomycin C (cat. no. M4287; Sigma).When required, Transwell chambers (0.4-mm pore size; BD Fal-con) were used. Lymphocytes were activated with ConA(5 mg/mL), harvested after 72 hours, and stained with phycoer-ythrin (PE)-conjugated anti-CD3 Ab (1:100 dilution, cat. no.12003842; eBioscience) before acquisition in a Gallios flowcytometer (Beckman Coulter).

FACSCell surface expression of HLA-ABC, PD-L1, and IL1R1 was

measured by direct immunofluorescence using FITC-conjugatedanti–HLA-ABC antibody (1:200 dilution; cat. no. 555552; BDPharmingen), APC-conjugated anti-CD274 (B7-H1, PD-L1) anti-body (1:100 dilution; cat. no. 329707; Biolegend), and APC-conjugated anti-hIL1R1 (1:100 dilution; cat. no. FAB269A; R&DSystem), respectively. For intracellular staining, cells were fixedand permeabilized with a Golgi Stop Kit according to the man-ufacturer's protocol (BD Pharmingen), and then stained with PE-conjugated anti-Rantes Ab (1:100 dilution; cat. no. 554732; BDPharmingen), PE-conjugated anti-IL1b Ab (1:100 dilution; cat.no. 554732; eBioscience), or isotype controls. Cell acquisitionwas analyzed with FlowJo 7.2.2 software. In each case, DMFIvalues were calculated as [MFI (relevant Ab) – MFI (isotypecontrol Ab)].

Western blottingCell lysates (30 mg protein) were fractionated on 10% to 15%

SDS-polyacrylamide gradient gels (Bio-Rad) and transferred toHybond-P membranes (Amersham Pharmacia). Membraneswere blocked with 5% milk in TBS containing Tween 0.1%and incubated with: mouse anti-IDO Ab (1:500 dilution; cat. no.05-840; Millipore); rabbit anti–HLA-ABC Ab (1:500 dilution; cat.no. SC-52810; Santa Cruz Technologies); rabbit anti-STAT1(1:1,000 dilution; cat. no. 9172; Cell Signaling); rabbit anti–p-STAT1 (Tyr701; 1:1,000 dilution; cat. no. 7649; Cell Signaling);mouse anti-IL1b (1:300 dilution; cat. no. 552289; BD Bio-sciences); rabbit anti–b-Actin Ab (1:1,000 dilution; cat. no.49671; Cell Signaling). Horse anti-mouse IgG (1:1,000 dilution;cat. no. 7076; Cell Signaling) or goat anti-rabbit IgG (1:1,000dilution; cat. no. 7074; Cell Signaling) conjugated to horseradishperoxidase was used for detection, and visualized using enhancedchemiluminescence detection system (Santa Cruz Technologies).

ELISAConcentration of IL2, TGFb1, PGE-2, and IL1b in culture

supernatants was measured by ELISA assays following the man-ufacturer's instructions: IL2 (cat. no. 88-7025; eBioscience), TGF-b1 (cat. no. DY1679; R&D Systems), PGE-2 (cat. no. KGE004B;R&D Systems), and IL1b (cat. no. 557953; BD Biosciences).

Tumorigenicity assaysFor anchorage-independent growth of hMSC, 104 cells were

transferred to 2 mL of culture medium containing 0.35% lowmelting point agarose (Sigma). Cells were seeded in triplicate in6-well plates containing a layer of solidified 0.6% agarose.Fresh medium was added every 3 days. Colonies were photo-graphed at �40 magnification on day 12. For in vivo tumorformation assay, 5 � 106 5 hits hMSC were injected subcuta-neously in both flanks of 8-week-old athymic nude mice(CENPALAB). Tumors were explanted and incubated for 30minutes at 37�C with 1 mg/mL of collagenase (cat. no. 9001-12-1; Worthington Biochemical Corporation), 0.5 mg/mL ofDNase I (cat. no. AMPD1; Sigma), and 0.1 mol/L of EDTA pH7.2 (Sigma). Tumors were homogenized to a single-cell sus-pension and placed into culture conditions. All animal han-dling and in vivo experiments were performed in accordancewith CIM's guidelines.

Statistical analysesStatistical significance (P < 0.05) was determined by an

unpaired two-tailed Student t test for single comparisons andone-way ANOVA with the Dunnett post-hoc test for multiplecomparisons. Statistical analyses and graphs were developedusing SPSS software and GraphPad Prism 5.0, respectively.

ResultsOncogenic transformation strengthens natural immune-evasive properties of hMSC

To assess the inherent contribution of oncogenic pathways onthe natural immune-evasive capabilities of hMSC, we focused ourattention on comparing hMSC from 2 to 5 hits and explored theeffect of transformation on cell immunogenicity and activeimmunosuppression.

We followed the transcription and protein expression of HLAclass I during the transformation process. Although HLA-A and Balleles are particularly sensitive to inactivation of pRB function(Fig. 1A), the mRNAs of each HLA alleles and the total HLA-ABCprotein levels in 2 and 5 hits cells were barely modified asconsequence of neoplastic transformation (Fig. 1A). However,HLA class I expression on the plasmatic membrane of fullytransformed cells was 50% reduced (Fig. 1B). During transfor-mation, a Gene Set Enrichment Analysis (GSEA) reveals thatoverall expression profile of genes involved in HLA class I anti-gen-presenting machinery (APM; Supplementary Table S1) wassignificantly downregulated (P < 0.001). Hierarchical clusteringanalysis identified a set of genes downregulated at the late stagesof transformation (Fig. 1C). Transcriptional repression of b2-microglobulin, PDIA3 (ERp57), ERAP2, and NPEPPS can impairmembrane expression of HLA-ABC on neoplastic hMSC (Supple-mentary Fig. S1).

Normal hMSC affect proliferation and effector functions of Tlymphocytes (14). We examined the inhibitory effect of hMSCover mitogen-triggered T-cell proliferation following stepwise

Intrinsic Inflammation and Immune Evasion in Cancer

www.aacrjournals.org Cancer Res; 75(15) August 1, 2015 3033

Figure 1.In vitro stepwise transformation strengthensnatural immune-evasive properties of hMSC.A, left, HLA class I mRNA alleles and totalHLA-ABC protein expression determined byqRT-PCR (�� , P < 0.01; �� , P < 0.001;n ¼ 3, mean � SEM) and Western blot assay(right; values are band intensities relativeto 2 hits). n.s., nonsignificant. B, cell surfaceexpression of total HLA class I moleculesevaluated by FACS analysis expressedas percentages of positive cells(� SEM), and DMFI values isotype controlsstaining are shown as filled histograms. C,hierarchical clustering analysis of Z scoredata and GSEA (P < 0.001) of genes related toHLA class I APM. The heatmap summarizes theaverage of three samples from each cell line.Underlined gene symbols representdifferentially expressed genes for 5 hits versus2 hits comparison. D, effect of sequentiallytransformed hMSC lines on T-cell proliferation.CFSE-stained PBMC were cocultured withhMSC lines (2 to 5 hits) at 1:25 (MSC:PBMC)ratio. The percentage of CD3þ proliferatingcells was evaluated by FACS after 72 hoursof stimulation with ConA.

Miranda et al.

Cancer Res; 75(15) August 1, 2015 Cancer Research3034

transformation. A progressive increase on the suppressive effect ofhMSC was observed at 1:25 hMSC:PBMC ratio. Maximum effectwas seen with 5 hits cells with up to 50% inhibition of CD3þ

lymphocytes proliferation comparing to stimulated PBMC-pos-itive controls (Fig. 1D). This effect was confirmed at differenthMSC:PBMC ratios using mitomycin C–pretreated hMSC cells(Supplementary Fig. S2). In addition, although normal hMSC donot affect T-cell activation (15), a reduction of IL2 secretion andCD25 expression occurs when PBMCwere cocultured with hMSCfollowing transformation events (Supplementary Fig. S3). Col-lectively, these data suggest that transformation can developimmune-evasive properties on hMSC, not as a consequence ofextrinsic immune pressure but as an intrinsic result of oncogenicpathways.

Insensitivity to IFNg signaling influences the immunogenicityand themechanismof T-cell suppressionon transformedhMSC

Normal hMSC affect T-cell function through the concertedaction of IDO and T-cell–specific chemokines, which are inducedby inflammatory cytokines on hMSCs (16, 17). We addressedwhether neoplastic transformation has the intrinsic capability toeither potentiate the natural mechanism of immune dysfunctionproduced by hMSC or induce an alternative mechanism.

When the supernatant collected from ConA-activated PBMCwas added to hMSC lines, a significant decrease in inducibleexpression of IDO (Supplementary Fig. S4A), CXCL9, CXCL10,and Rantes (Supplementary Fig. S4B) was detected on trans-formed cells. As described for normal hMSC (16), exogenousIFNg elicited IDO expression in 2 hits cells (Fig. 2A), while

Figure 2.Progressive oncogenic transformationshapes the response to inflammatorycytokines on hMSC. hMSC lines werecultured for 72 hours in IFNg and/orTNFa supplemented medium. A, IDOmRNA and protein levels weredetermined by qRT-PCR (left) andWestern blot assay (right),respectively. B, CXCL-9 (left) andCXCL-10 (right) mRNA assayed byqRT-PCR. Relative fold change of eachgene expression versus 2 hits cellscultured in control medium is shown(� , P < 0.05; �� , P < 0.01; �� , P < 0.001;n ¼ 3; mean � SEM for each culturecondition). n.s., nonsignificant. C,Western blot analysis of total HLAclass I proteins level. Relative bandintensities are shown for each hMSC linecompared with counterpart cellscultured in control medium. D,membrane expression of total HLAclass I molecules detected by FACS on2 and 5 hits cell lines, shown aspercentages of positive cells (�SEM)treated or not with IFNg and DMFIvalues. Filled histograms correspondto isotype controls staining.

concomitant signaling of both IFNg and/or TNFa inflammatorycytokines was required for chemokine induction (Fig. 2B). Con-versely, the inducible expression of IDO was reduced in 4 hitscells, as shown by Western blotting, whereas it was abrogated inthe fully transformed 5 hits hMSC. CXCL9 and CXCL0 mRNAlevels were also impaired in neoplastic hMSC (Fig. 2B). Therefore,immunosuppression of T cells mediated by IDO and induced inhMSC by inflammatory cytokines is not reinforced but abrogatedin transformed hMSC.

The apparent lack of sensitivity of neoplastic hMSC to IFNgcould reflect a balance between the disadvantage of reducing the

production of IDO, as a mediator of immunosuppression, or theadvantage of controlling the immunogenicity by limitingthe ability of IFNg signaling to upregulate the transcription ofthe APM. In fact, transcription of individual HLA class I allelesinduced by IFNg exposure is higher in partially than in fullytransformed cells (Supplementary Fig. S5). The total protein level(Fig. 2C) and membrane expression (Fig. 2D) of HLA-ABC wasunmodified on 5 hits hMSC treated with this cytokine.

Because transformed hMSC failed to elicit a range of IFNg-inducible genes, suggesting unresponsiveness to IFNg , we testedthe hypothesis that the IFNg signaling pathway is defective in

Figure 3.IFNg signaling pathway is affected onprogressively transformed hMSC.A, comparative intracellularphosphorylation and nuclearaccumulation of STAT1 on 2 and 5 hitshMSC lines cultured for 2 hours incontrol or IFNg added medium.p-STAT1-FITC and cell nucleicounterstained with Hoechst weredetected by immunofluorescence. B,Pearson colocalization coefficient forp-STAT1 and Hoechst was calculatedfrom three fields in two independentexperiments. C, Western blot analysisof STAT1 andp-STAT1 (Tyr701) on cellscultured for 1 hour in control or IFNgadded medium. D, STAT1 mRNAassayed by qRT-PCR on hMSC linescultured for 72 hours in control or IFNgadded medium. Fold changes arerelative to the expression of the genein each hMSC line versus 2 hits cellscultured in control medium (n ¼ 3,mean � SEM). E, hierarchicalclustering analysis of Z score dataand GSEA of IFNg signaling pathway(left; GSEA P < 0.001) and regulatorsof IFNg signaling (right; GSEA P >0.05). The heatmap summarizes theaverage of three samples from eachcell line. Underlined gene symbolsrepresent differentially expressedgenes for the 5 hits versus 2 hitscomparison.

Miranda et al.

Figure 4.The constitutive ability of neoplastic hMSC to inhibit T-cell proliferation partially depends on secreted factors. A, lymphocyte proliferation induced with ConAand assayed as CFSE dilution by FACS. CFSE-stained PBMC were cocultured with hMSC lines at 1:10 (MSC:PBMC) ratio in a Transwell system for 72 hours.Data are expressed as percentage of T-cell proliferation recovery (�SEM) relative to PBMC activated in the absence of hMSC. B, percentage of proliferatingT cells based on CFSE dilution by FACS. PBMC cultured in the supernatants of hMSC lines (1:2 diluted in fresh medium) were stimulated with ConA for 72 hours(n ¼ 3, mean � SEM). Dashed line, percentage of proliferating cells in the absence of hMSC supernatants. C, comparative amount of TGFb1, HGF, COX-2,and iNOSmRNAon2 and5hits hMSC lines analyzedby qRT-PCR. Fold changes are relative to the expression of eachgene versus 2 hits cells. D andE, concentration ofTGFb1 (D) and PGE-2 (E) determined in the supernatant of hMSC lines by ELISA (� , P < 0.05; �� , P < 0.01; �� , P < 0.001; n ¼ 3, mean � SEM). n.s., nonsignificant.

neoplastic hMSC. p-STAT1 homodimers rapidly accumulate inthe nucleus to mediate IFNg biologic responses (18). Weassessed the nuclear accumulation of p-STAT1 at single-celllevel by fluorescence microscopy. Unlike neoplastic 5 hits cells,rapid nuclear accumulation of p-STAT1 was detected in non-transformed 2 hits cells treated with IFNg (Fig. 3A and B).Coherently, IFNg-induced phosphorylation of STAT1 wasreduced in 5 hits cells (Fig. 3C). STAT1 is an IFNg-induciblegene (19), so we examined whether transformation abrogatesinducible transcription of STAT1. As shown in Fig. 3D, expres-sion of STAT1 mRNA is progressively reduced following trans-formation, reaching complete abrogation on fully transformed

hMSC cells. To characterize the IFNg receptor signaling further,we then compared the gene expression profiles of selected genesinvolved in this pathway (Supplementary Table S1). The GSEAindicates that transformation significantly reduced (P < 0.001)the overall expression of genes implicated in IFNg signaling(Fig. 3E, left), and four genes other than STAT1 were signifi-cantly reduced comparing 2 versus 5 hits cells. No significantchange was detected in the overall expression of negativeregulatory genes (P > 0.05; Fig. 3E, right). Altogether, thesedata highlight a deficit in IFNg signaling on transformed hMSC,which molds both immunogenicity and immunosuppressionmechanisms in neoplastic cells.

Figure 5.Neoplastic transformation of hMSCinduces the production of IL1b, whichsustains the inhibition of T-cellproliferation. A, expression of IL1bmRNA on each hMSC line relativeto 2 hits cells by qRT-PCR.n.s., nonsignificant. B, concentration ofIL1b determined by ELISA in thesupernatant of hMSC lines cultured for72 hours. C, FACS detection of CD3þ

proliferating cells. CFSE-stained PBMCwere cocultured with 5 hits cellspreviously infected with lentivirusencoding for scramble or IL1b shRNA.Proliferation was assessed after ConAstimulation for 72 hours. Dashed line,percentage of proliferating cells in theabsence of hMSC. D, comparativeexpression of TGFb1, HGF, and COX-2mRNA on 5 hits scramble versus 5 hitsIL1b shRNA–modified cells. E and F,concentration of TGFb (E) and PGE-2(F) in the culture supernatants of 5 hitshMSC lines. � , P < 0.05; �� , P < 0.01;�� , P < 0.001; each bar: n¼ 3, mean�SEM. G, percentage of proliferatingT cells based on CFSE dilution byFACS. PBMC were cultured in thesupernatant collected from thedifferent variants of 5 hits hMSC lines(1:2 diluted in fresh medium) and werestimulated to proliferate with ConA for72 hours (n¼ 3, mean� SEM). Dashedline, percentage of proliferating cells inthe absence of hMSC supernatants.H and I, cell surface expressionof PD-L1on each hMSC line (H) and on 5 hitshMSC modified with scramble or IL1bshRNA (I), detected by FACS.

Miranda et al.

Constitutive capacity of transformed hMSC to suppress T-cellproliferation depends on IL1b production

To address the mechanism underlying the increased inhibitorycapacity of transformed hMSC, we initially evaluated whether theinhibition of T-cell proliferation depends on cellular contact.Mitogen-driven proliferation assay was performed in Transwellchambers using 1:10hMSC:PBMCcells ratio to reach themaximalinhibitory capacity of each hMSC line (Fig. 4A, top). Partiallytransformed hMSC lost more than 80% of their suppressivecapacity, whereas neoplastic 5 hits cells retained about 50%inhibitionof lymphocyte proliferationwhen the Transwell systemis used (Fig. 4A). This indicates that transformed hMSC increasesthe dependency of secreted factors to suppress T-cell proliferation.To confirm this result, PBMC were activated in culture mediumcontaining 50% of supernatant collected from hMSC lines.Although supernatants from 4 hits cells significantly inhibited

T-cell proliferation comparing with 2 hits, fully transformed cellsexerted the highest suppressive effect (Fig. 4B). Hence, secretedfactors partially determine the constitutive suppressive effectpromoted by hMSC transformation.

To identify molecules involved in suppressing T-cell prolif-eration, we measured a set of molecules that are producedby normal hMSC playing a secondary role in the inhibition ofT-cell proliferation (14, 20). Unlike inducible nitric oxidesynthase (iNOS; Fig. 4C) and IDO (Fig. 2A) enzymes, a signif-icant upregulation in constitutive transcription of TGFb1, COX-2, and HGF genes was detected comparing 2 and 5 hits cells(Fig. 4C). Quantification of TGFb1 (Fig. 4D) and PGE2(Fig. 4E), secreted by hMSC cell lines, revealed a significantincrease of both molecules due to transformation. Becausethese molecules are downstream targets of the IL1b signalingpathway (21), we considered the possibility that IL1b could be

Figure 6.Tumorigenicity of in vitroprogressively transformed hMSCdepends on IL1b expression. A,comparative anchorage-independentgrowth capacity in soft agar and in vivotumor growth in athymic nude miceof 5 hits cells modified with scrambleshRNA or IL1b shRNA. B,representative photographs ofcolonies growing in soft agar. C, plot ofPearson correlation between thenumber of colonies and theintracellular levels of IL1b detected on5 hits IL1b shRNA hMSC. D, geneexpression microarray analysis of IL1b(top) and percentages of intracellularIL1b expressing cells (shown aspercentages of positive cells � SEM;bottom) on 5 hits hMSC line and cellsobtained from explanted tumors(Exp). The heatmaps show threesamples from 5 hits hMSC and fiveindividual explanted tumors.Underlined gene symbols denotedifferential expression of IL1b in 5 hitscell lines versus explanted tumors.Isotype controls signal in the FACSassay are shown in filled histograms.

orchestrating the mechanism of T-cell suppression exerted bytransformed hMSC.

IL1b mRNA (Fig. 5A) and intracellular protein levels (Sup-plementary Fig. S6A and S6B) increase on hMSC followingsequential transformation. Noteworthy, IL1b was only detectedin the supernatant of fully transformed hMSC (Fig. 5B), notassociated with cell lysis (Supplementary Fig. S7A) or apoptosis(Supplementary Fig. S7B). GSEA showed no significant enrich-ment of genes involved in IL1b signaling (Supplementary TableS1) owing to neoplastic transformation, though particulargenes were upregulated comparing 5 versus 2 hits cell lines(Supplementary Fig. S8A). Interesting, we found diminishedexpression of IL1 receptor 1 (IL1R1) on 5 hits cells (Supple-mentary Fig. S8B).

Because IL1b is only active once secreted (22), we focused onthe role of IL1b in the inhibition of T-cell proliferation by 5 hitshMSC. IL1b expression was knockeddown (KD) on 5 hits hMSCby infection with lentivirus encoding IL1b shRNA (Supplemen-tary Fig. S9A). Change in cells proliferation rate was not observed(Supplementary Fig. S9B). This manipulation restores T-cell pro-liferation, demonstrating adeterminant role of IL1b for neoplastichMSC-immunosuppressive property (Fig. 5C).

Neither the supernatant collected from mitogen-activatedPBMC nor recombinant IFNg recovered the inducible levels ofIDO (Supplementary Figs. S10A and S11A), CXCL9 and CXCL10(Supplementary Fig. S10B), STAT1 (Supplementary Fig. S11B),and HLA-ABC expression (Supplementary Fig. S11C) in IL1b KDhMSC. Therefore, IL1b seems tobe unrelated to the insensitivity toIFNg signaling caused by neoplastic transformation, whereas it isinvolved in the novelmechanism of immune suppression orches-trated during the acquisition of the neoplastic phenotype. In fact,IL1bKDhMSC have significantly reduced transcription of TGFb1,HGF, and COX-2 mRNA (Fig. 5D) as well as secreted levels ofTGFb1 (Fig. 5E) and PGE-2 (Fig. 5F) when compared with thescramble shRNA control. To ponder the contribution of suchsoluble factors to IL1b-concerted immunosuppression, PBMCwere mitogen activated in culture medium containing 50% ofculture supernatant collected from IL1b KD or scramble shRNAcell lines. Supernatant from IL1b KD cells exhibited only apartially reduced capacity to inhibit T-cell proliferation (Fig.5G), suggesting that beside secreted molecules, a complementarymechanism controlled by IL1b is mediating the suppression of Tlymphocytes proliferation.

Signaling through the programmed death-1 receptor (PD-1)induced by its ligands PD-L1 and PD-L2 modulates the T-cellresponse (23). Recent evidence demonstrates that IL1b can inducethe upregulation of PD-L1 (24). Fully transformed hMSC havesignificantly higher expression of PD-L1 (Fig. 5H), which isreduced as a consequence of IL1b KD (Fig. 5I), indicating thatIL1b orchestrates on transformed hMSC a mechanism of immu-nosuppression mediated by secreted and membrane-expressedmolecules.

IL1b secretion sustains the neoplastic phenotype acquired bymultistep transformation of hMSC

We further explored the role of IL1b signaling in sustaininghMSC tumorigenicity. Knockdown of IL1b in 5 hits cells dramat-ically reduces the number and size of colonies growing in softagarose (Fig. 6A and B). Correspondingly, only one out of sixtumors grew in athymic nudemice inoculated with IL1bKD5hitshMSC (Fig. 6A). The sensitivity of tumorigenic phenotype to IL1b

was demonstrated when 5 hits lines expressing different levelsIL-1b were obtained and tested in a colony formation assay. Wefounda strong correlationbetween the productionof IL1b and thenumber of growing colonies, revealing the sensitivity of neoplas-tic hMSC to an IL1b-mediated autocrine loop (Fig. 6C). Toaddress whether IL1b production is sustained in vivo, we com-pared IL1b expression on 5 hits hMSC and explanted tumors.Gene expression microarray analysis (Fig. 6D, top) and intracel-lular staining (Fig. 6D, bottom) demonstrate that in vivo growingcells preserve IL1b expression. Transcription of IL1b gene washighest on cells isolated from tumors, whereas explanted cellscultured in vitro have as much intracellular expression of IL1b as 5hits hMSC. It suggests that IL1b expressionwas prompted into thetumor microenvironment.

DiscussionImmunity exerts dual protective and tumor-progressive actions

on developing tumors. Escape of cancer cells from immunecontrol has been assumed as a direct consequence of immunepressure (25), involving reduced presentation of immunogenicantigens (26) and a plethora of active immune-suppressive strat-egies (27). Based on this hypothesis, immune effector mechan-isms select for expansion of those cancer cells possessing aneffective counterattack mechanism. However, taking advantageof a neoplastic transformation model created by in vitro–inducedsequential and accumulative genetic alterations on hMSC (6), weshowed that immune-evasive mechanisms can evolve along themultistep process of neoplastic transformation independently ofimmune-selective pressure. Progressive transformation reduceshMSC immunogenicity, evidenced as transcriptional downregu-lationof theAPMand reduced expressionofHLA class Imoleculesat the plasmatic membrane, one of the best characterized mech-anismused by tumors to impair antigen presentation and avoid T-cell recognition (28).

Stepwise transformation progressively increases the naturalinhibitory effect of hMSC on T-cell proliferation. The inhibitoryeffect switches from a mechanism inducible and regulated byextrinsic inflammatory signals to another that is constitutive andregulated by an autocrine inflammatory loop. Neoplastic hMSClose the sensitivity to the canonical IFNg signaling pathwayevidenced by altered STAT1 phosphorylation and migration tothe nucleus, associated with transcriptional impairment of IFNgsignaling–related genes. This benefits cancer progression, consid-ering the relevance of IFNg in cancer immunosurveillance (29). Infact, potential overexpression of APM and secretion of chemo-kines recruiting T cells (i.e., CXCL9, CXCL10, and RANTES)influencedby IFNg is dampened. Inparallel, halt of IFNg signalingabrogates the inducible expression of IDO, a mediator of thenatural suppressive capacity of hMSC (16). Hence, neoplastichMSC sustain the suppressive effect by gaining a constitutiveexpression of TGFb, HGF, PGE2, and PD-L1, dissociating thesuppressive effect from the extrinsic inflammatory influence.Along with upregulation of HLA-A and HLA-B alleles, inhibitionof pRb expression (3 hits cells) reduces the expression of PD-L1. Itsuggests that achievement of oncogenesis depends on circumvent-ing incomplete transformation stages that are highly sensitive toimmune attack during the multistep process of transformation.

It is well established that oncogene activation leads to aprotumoral intrinsic inflammatory program (30). In our model,wedetected increased transcription and intracellular expression of

Miranda et al.

IL1b following sequential transformation. Secretion of this cyto-kinewas only detected on fully transformed5hits cells, suggestingthat early transformation steps cooperate on transcription andtranslation of IL1b, whereas H-RAS–induced signaling activatesthe secretion of the mature cytokine. The link between RASpathway and the consequent secretion of IL1b, which, in turn,upregulates HIF1a via Cox2 has been described (31, 32). Corre-spondingly, an IL1b-sensitive increase in PGE2 secretion occurson 5 hits cells, whereas IL1b expression is upregulated on hMSCgrowing as tumors in in vivo hypoxic conditions. Furthermore,secreted IL1b induces an autocrine inflammatory loop on neo-plastic hMSC,which directs the suppression of T-cell proliferationby supporting the constitutive expression of immunosuppressivemolecules. In fact, either secreted or membrane-expressed sup-pressor factors prompted by neoplastic transformation on hMSCare sensitive to IL1b signaling (21, 24). Of note, although H-RASconstitutive activation favors IL1b secretion and p-STAT-1(Tyr701) inhibition, IL1b is not responsible for IFNg-haltedsensitivity induced by transformation on hMSC. It suggests thattransformation induces a signaling bias: reducing the extrinsicinflammation-induced immunogenicity and increasing theintrinsic inflammation-induced immune suppression.

Previous experimental models highlighted the relevance ofhost-derived IL1b on chemical carcinogenesis (33). Also, itscapacity to mobilize suppressive cells into the tumor microenvi-ronment has been seen in genetically manipulated animals (34).The relevance of IL1b for cancer progression in humans has beendemonstrated in melanoma (24) and non–small cell lung cancer(NSCLC) patients (35). The former study demonstrated that IL1bsecreted by tumors carrying the BRAFmutation V600Epromotes aPD-1 ligand–mediated suppressive effect in tumor-associatedfibroblasts. The latter showed that IL1b is elevated in NSCLCpatients and is crucial in lung carcinogenesis, involving theHIF1a–COX2 axis in cancer cells. We provide further evidenceof an IL1b-mediated autocrine inflammatory loop that can sup-port the malignant phenotype, both in vitro and in vivo, whichorchestrates direct cancer cell–mediated immunosuppressiveability. Initiation and maintenance of cancer phenotype basedon inflammatory loops have been also demonstrated for IL6 (36).In contrast with our results, it has been reported that colonyformation in soft agar is not affected in IL1b-deficient fibrosar-coma cells lines (37), whereas Raf but not RasV12-induced trans-formation onNIH/3T3 cells requires IL1 signaling (38).However,these models do not incorporate, as our does, the cooperative

effect of multistep genetic alterations in stem cells contributing toprogressive cancer-phenotype appearance.

This work demonstrates that oncogenic pathways, indepen-dently of immune-selective pressure, reduce the influence ofextrinsic inflammation in favor of intrinsic inflammation tosustain tumorigenicity and to orchestrate immunosuppression.However, whether a different oncogenic program favors an alter-native inflammatory loop or promotes a different suppressivemechanism in differentiated nonnaturally immune-suppressivecells (i.e. non-stem cells) needs further attention. Noteworthy,gene microarray's analysis of in vitro–transformed human fibro-blasts (39) showed progressive upregulation of IL1b (Supple-mentary Fig. S12). In terms of therapeutic relevance, IL1b block-ade appears as an alternative to interfere with both tumorigenicityand immunosuppressive effect of malignant tumors.

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

Authors' ContributionsConception and design: A. Miranda, R. P�erez, J. de Le�onDevelopment of methodology: A. Miranda, N. S�anchez, C.M. LimiaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A. Miranda, J.M. Funes, N. S�anchez, C.M. Limia,M. MesaAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A. Miranda, M. Mesa, R. P�erez, J. de Le�onWriting, review, and/or revision of the manuscript: A. Miranda, J.M. Funes,N. S�anchez, C.M. Limia, S.A. Quezada, R. P�erez, J. de Le�onAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): J.M. FunesStudy supervision: J. de Le�on

AcknowledgmentsThe authors thank Belinda S�anchez and Salvador Moncada for their useful

comments about this work, CIM's animal house technicians for excellentassistance, Janet Fernihough for editing, and Yasser Almeida, Julio Fernandez,and Xitlally Popa for technical help.

Grant SupportThis work was financially supported by CIM, complemented with BIF and

EMBO travel grants.The costs of publication of this articlewere defrayed inpart by the payment of

page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received November 7, 2014; revised April 20, 2015; accepted May 12, 2015;published OnlineFirst June 11, 2015.

References1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation.

Cell 2011;144:646–74.2. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integrating

immunity's roles in cancer suppression and promotion. Science 2011;331:1565–70.

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

4. Allavena P, Garlanda C, Borrello MG, Sica A, Mantovani A. Pathwaysconnecting inflammation and cancer. CurrOpinGenetDev 2008;18:3–10.

5. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer.Cell 2011;140:883–99.

6. Funes JM, QuinteroM, Henderson S, Martinez D, Qureshi U,Westwood C,et al. Transformation of human mesenchymal stem cells increases theirdependency on oxidative phosphorylation for energy production. ProcNatl Acad Sci U S A 2007;104:6223–8.

7. Nicola MD, Carlo-Stella C. Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenicstimuli. Transplantation 2002;99:3838–43.

8. Tse WT, Pendleton JD, Beyer WM, Egalka MC, Guinan EC. Suppression ofallogeneic T-cell proliferation by human marrow stromal cells: implica-tions in transplantation. Transplantation 2003;75:389–97.

9. Karnoub AE, Dash AB, Vo AP, Sullivan A, Brooks MW, Bell GW, et. al.Mesenchymal stem cells within tumour stroma promote breast cancermetastasis. Nature 2007;4:557–63.

10. Dwyer RM, Potter-Beirne SM, Harrington KA, Lowery AJ, Hennessy E,Murphy JM, et. al. Monocyte chemotactic protein-1 secreted by primarybreast tumors stimulates migration of mesenchymal stem cells. ClinCancer Res 2007;4:5020–7.

11. Gali�e M, Konstantinidou G, Peroni D, Scambi I, Marchini C, Lisi V, et al.Mesenchymal stem cells share molecular signature with mesenchymal

tumor cells and favor early tumor growth in syngeneic mice. Oncogene2008;27:2542–51.

12. Helman LJ, Meltzer P. Mechanisms of sarcoma development. Nat RevCancer 2003;3:685–94.

13. S�anchez N, Miranda A, Funes JM, Hevia G, P�erez R, de Le�on J. Oncogenictransformation tunes the cross-talk betweenmesenchymal stem cells and Tlymphocytes. Cell Immunol 2014;289:174–84.

14. Gebler A, Zabel O, Seliger B. The immunomodulatory capacity of mesen-chymal stem cells. Trends Mol Med 2012;18:128–34.

15. Ramasamy R, Tong CK, Seow HF, Vidyadaran S, Dazzi F. The immuno-suppressive effects of human bone marrow-derived mesenchymal stemcells target T-cell proliferation but not its effector function. Cell Immunol2008;251:131–6.

16. Ren G, Su J, Zhang L, Zhao X, LingW, L'huillie A, et. al. Species variation inthe mechanisms of mesenchymal stem cell-mediated immunosuppres-sion. Stem Cells 2009;27:1954–62.

17. KramperaM, Cosmi L, Angeli R, Pasini A, Liotta F, Andreini A, et al. Role forinterferon-gamma in the immunomodulatory activity of human bonemarrow mesenchymal stem cells. Stem Cells 2006;24:386–98.

18. Bach EA, Aguet M, Robert D, Schreiber RD. The IFN-g receptor: a paradigmfor cytokine receptor signaling. Annu Rev Immunol 1997;15:563–91.

19. Der SD, Zhou A, Williams BR, Silverman RH. Identification of genesdifferentially regulated by interferona,b, or g using oligonucleotide arrays.Proc Natl Acad Sci U S A 1998;95:15623–8.

20. Aggarwal S, Pittenger MF. Human mesenchymal stem cells modulateallogeneic immune cell responses. Blood 2005;105:1815–22.

21. Dinarello CA. Biologic basis for interleukin-1 in disease. Blood 1996;87:2095–147.

22. Dinarello CA. Immunological and inflammatory functions of the inter-leukin-1 family. Annu Rev Immunol 2009;27:519–50.

23. Nurieva XL, Dong C. Yin-Yang of costimulation: crucial controls ofimmune tolerance and function. Immunol Rev 2009;229:88–100.

24. Khalili JS, Liu S, Rodríguez-Cruz TG. Oncogenic BRAF (V600E) promotesstromal cell-mediated immunosuppression via induction of interleukin-1in melanoma. Clin Cancer Res 2012;18:5329–40.

25. Vesely MD, Kershaw MH, Schreiber RD, Smyth MJ. Natural innate andadaptive immunity to cancer. Annu Rev Immunol 2011;29:235–71.

26. HirokazuM. Cancer exome analysis reveals a T-cell-dependentmechanismof cancer immunoediting. Nature 2012;482:400–4.

27. Rabinovich GA, Gabrilovich D, Sotomayor EM. Immunosuppressive strat-egies that are mediated by tumor cells. Annu Rev Immunol 2007;25:267–96.

28. GarridoF, Cabrera T, AptsiauriN.Hard and soft lesions underlying theHLAClass I alterations in cancer cells: implications for immunotherapy. Int JCancer 2010;127:249–56.

29. Kaplan DH, Shankaran V, Dighe AS, Stockert E, Aguet M, Old LJ, et al.Demonstration of an interferon g-dependent tumor surveillance sys-tem in immunocompetent mice. Proc Natl Acad Sci U S A 1998;95:7556–61.

30. Grazia M, Degl D, Pierotti MA. Inflammation and cancer: the oncogene-driven connection. Cancer Lett 2008;267:262–70.

31. Kaluz S, Van Meir EG. At the crossroads of cancer and inflammation: Rasrewires an HIF-driven IL-1 autocrine loop. J Mol Med 2011;89:91–4.

32. Jung YJ, Isaacs JS, Lee S, Trepel J, Neckers L. IL-1bmediated up-regulation ofHIF-1a via NFkb/Cox2 pathway identifies HIF-1a as a critical link betweeninflammation and cancer. FASEB J 2003;17:2115–7.

33. Apte RN, Dotan S, Elkabets M, White MR, Reich E, Carmi Y, et. al.The involvement of IL-1 in tumorigenesis, tumor invasiveness, metas-tasis and tumor-host interactions. Cancer Metastasis Rev 2006;25:387–408.

34. Tu S, Bhagat G, Cui G, Takaishi S, Kurt-Jones EA, Rickman B, et al. Over-expression of interleukin-1b induces gastric inflammation and cancer andmobilizes myeloid-derived suppressor cells in mice. Cancer Cell 2008;14:408–19.

35. Wang L, Zhang L, Wu J. IL-1 b-mediated repression of microRNA-101 iscrucial for inflammation-promoted lung tumorigenesis. Cancer Res2014;74:4720–30.

36. Iliopoulos D, Hirsch H, Struhl K. An epigenetic switch involving NF-kappaB, Lin28, Let-7 MicroRNA, and IL6 links inflammation to celltransformation. Cell 2009;139:693–706.

37. Nazarenko I, Marhaba R, Reich E, Voronov E, Vitacolonna M, HildebrandD, et al. Tumorigenicity of IL-1a– and IL-1b–deficient fibrosarcoma cells.Neoplasia 2008;10:549–62.

38. Vale T, Ngo T, White M, Lipsky P. Raf-induced transformation requires aninterleukin 1 autocrine loop. Cancer Res 2001;61:602–7.

39. Collado M, Gil J, Efeyan A, Guerra C, Schuhmacher AJ, Barradas M, et al.Tumour biology: senescence in premalignant tumours. Nature 2005;436:642.

Miranda et al.

2015;75:3032-3042. Published OnlineFirst June 11, 2015.Cancer Res Alex Miranda, Juan M. Funes, Nilda Sánchez, et al. Extrinsic Immune-Selective PressureInflammation in Human Mesenchymal Stem Cells Independently of Oncogenic Transformation Can Orchestrate Immune Evasion and

Updated version

10.1158/0008-5472.CAN-14-3276doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2015/06/12/0008-5472.CAN-14-3276.DC1.html

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/75/15/3032.full.html#ref-list-1

This article cites 39 articles, 10 of which you can access for free at:

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

Subscriptions

Reprints and

.pubs@aacr.org

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

Permissions

.permissions@aacr.org

To request permission to re-use all or part of this article, contact the AACR Publications Department at

oncogenic transformation can orchestrate immune evasion ... · alex miranda1, juan m. funes2, nilda...

Documents

lugar y fecha de realizaciónestudio comparativo en la...

randomized convective cloud seeding experiment … ·...

jv miranda1 dissprt

factores asociados al rendimiento escolar y sus ... ·...

jose m. castro-perez1 2, ingeborg vanhees2, vincent … ·...

the prosolv manual - cbirt prosolv manual.pdf · amanda...

streptomyces development in colonies and soilsangel...

quadrilateral mesh generation using hierarchical...

prof: guillermo a sanchez1 caracterÍsticas de los programas...

relatório de estágio sandra...

ethansharygin ,herbertl.smith ,vitormiranda ,tombyrne...

a cmos morlet wavelet generator - radioengradioengineering,...

rodrigo b. sanchez1, josÉ marques jÚnior2, gener t....

administracion adilson perez1

adriana armellini tuchman , josé ramón … armellini...

movilizaciones católicas en tiempos de represión y...

comunicado 22 técniconovembro,...

seminário:qualidade na prática mediúnica projeto manoel...

reverse transcription-loop-mediated isothermal ... ·...

chilean society for cell biology xxxi annual meeting ... ·...