An ARC-Regulated IL1b/Cox-2/PGE2/ b-Catenin/ARC Circuit ...#25-072-CV, Corning). Human MSCs were isolated from bone marrow samples obtained from healthy subjects as described previously

Tumor Biology and Immunology

An ARC-Regulated IL1b/Cox-2/PGE2/b-Catenin/ARC Circuit Controls Leukemia–Microenvironment Interactions and ConfersDrug Resistance in AMLBing Z. Carter, Po Yee Mak, Xiangmeng Wang,Wenjing Tao, Vivian Ruvolo,Duncan Mak, Hong Mu, Jared K. Burks, and Michael Andreeff

Abstract

The apoptosis repressor with caspase recruitment domain(ARC) protein is a strong independent adverse prognosticmarker in acute myeloid leukemia (AML). We previouslyreported that ARC regulates leukemia–microenvironmentinteractions through the NFkB/IL1b signaling network.Malignant cells have been reported to release IL1b, whichinduces PGE2 synthesis in mesenchymal stromal cells(MSC), in turn activating b-catenin signaling and inducingthe cancer stem cell phenotype. Although Cox-2 and itsenzymatic product PGE2 play major roles in inflammationand cancer, the regulation and role of PGE2 in AML arelargely unknown. Here, we report that AML–MSC coculturesgreatly increase Cox-2 expression in MSC and PGE2 pro-duction in an ARC/IL1b–dependent manner. PGE2 inducedthe expression of b-catenin, which regulated ARC and aug-

mented chemoresistance in AML cells; inhibition ofb-catenin decreased ARC and sensitized AML cells to che-motherapy. NOD/SCIDIL2RgNull-3/GM/SF mice trans-planted with ARC-knockdown AML cells had significantlylower leukemia burden, lower serum levels of IL1b/PGE2,and lower tissue human ARC and b-catenin levels, pro-longed survival, and increased sensitivity to chemotherapythan controls. Collectively, we present a new mechanism ofaction of antiapoptotic ARC by which ARC regulates PGE2production in the tumor microenvironment and microen-vironment-mediated chemoresistance in AML.

Significance: The antiapoptotic protein ARC promotesAML aggressiveness by enabling detrimental cross-talk withbone marrow mesenchymal stromal cells.

IntroductionAcute myeloid leukemia (AML) is a highly heterogeneous

hematologicmalignancywith poor overall survival. Intrinsic char-acteristicsof leukemiacells suchasoverexpressionofantiapoptoticproteins and activation of survival signaling pathways, extrinsicmicroenvironmental factors such as alteration of growth factors/cytokines, and leukemia-niche interactions that further augmentthe changes in leukemia cells andmicroenvironments all supportleukemia cell and leukemia stem/progenitor cell growth, survival,and resistance to therapy. Thus, targeting the survival and resis-tance mechanisms that are common in AML should benefit manypatients with this disease. We previously reported that the apo-ptosis repressorwith caspase recruitmentdomain (ARC)protein isa strong independent, poor prognosis factor in AML. ARCprotects

leukemia cells from apoptosis induced by multiple agents (1, 2).We subsequently found that ARC induces IL1b expression in AMLcells, increases chemokineCCL2,CCL4,andCXCL12expression inmesenchymal stromal cells (MSC) in an ARC/IL1b–dependentmanner, and facilitates leukemia–microenvironment interactionsthrough a NFkB/IL1b signaling network (3).

Bone marrow–derived MSCs are an essential structural andfunctional component of the bone marrow microenvironment,and they are critical for hematopoiesis (4). Within the context ofleukemia, MSCs play an essential role in protecting leukemiacells from chemotherapeutic agents (5). IL1b was previouslyshown to promote apoptosis resistance in AML blasts (6), andthe IL1 receptor accessory protein (IL1RAP) is reportedly over-expressed in AML stem/progenitor cells (7). Furthermore, target-ing IL1RAP with a neutralizing antibody selectively killed AMLstem cells (8). These data support the critical role of bonemarrowIL1b signaling in AML cell survival. Furthermore, a recentstudy demonstrated that cancer cells release IL1b that stronglyinduces Cox-2/PGE2 in MSCs, which in turn activates b-cateninsignaling and induces a stem cell phenotype in epithelial cancercells (9).

The roles of Cox-2 and its enzymatic product PGE2 in inflam-mation and cancer are well documented. Cox-2–catalyzed pros-taglandin synthesis is regulated by NFkB/IL1b signaling (10–12),which is frequently upregulated in cancer cells. PGE2 signalsthrough EP1-EP4 receptors that regulate multiple signaling cas-cades including b-catenin (13). Cox-2 and PGE2 havebeen reported to enhance hematopoietic stem cell survival,

Section of Molecular Hematology and Therapy, Department of Leukemia, TheUniversity of Texas MD Anderson Cancer Center, Houston, Texas.

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

CorrespondingAuthors:MichaelAndreeff, TheUniversityof TexasMDAndersonCancer Center, 1515 Holcombe Boulevard, Unit 448, Houston, TX 77030-4009.Phone: 713-792-7261; Fax: 713-794-3063; E-mail: [email protected];and Bing Z. Carter, Phone: 713-794-4014; Fax: 713-794-1903; E-mail:[email protected]

doi: 10.1158/0008-5472.CAN-18-0921

�2019 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1165

on May 26, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 23, 2019; DOI: 10.1158/0008-5472.CAN-18-0921

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-18-0921&domain=pdf&date_stamp=2019-3-12

http://cancerres.aacrjournals.org/

self-renewal, engraftment, and homing (14). Treatment with thePGE2 analog dimethyl-PGE2 (dmPGE2) was found to upregulateCXCR4 through increasing HIF1a and enhance hematopoieticstem/progenitor cell homing and engraftment (15, 16).

Although PGE2 is known to possess potent tumor-stimulatingactivity (17), its role in leukemia is largely unknown. AML cell lineHL-60 was reported to express high levels of microsomal PGEsynthase 1 (mPGES-1), which is frequently induced concomi-tantly with Cox-2 by various proinflammatory stimuli andresponsible for increased PGE2 production (18). mPGES-1 wasundetectable in normal blood mononuclear cells and its inhibi-tion reduced PGE2 production and viability of HL-60 cells (19).PGE2 from MSCs was shown to activate PKA signaling andantagonize p53-induced cell death (20). Furthermore, Wnt/b-catenin signaling is constitutively active in AML (21) and isrequired for the development of AML stem cells (22). Expressionof b-catenin by AML cells portends enhanced clongenic capacitiesand poor disease prognosis (23).

We therefore hypothesized that ARC exerts its role in leukemia–stromal interactions bymodulating PGE2 levels. In this study, weinvestigated Cox-2 expression in MSCs cocultured with AML cellsin which ARC expression was genetically modified. We alsoinvestigated PGE2 levels in vitro in anAML–MSCcoculture system,in fresh bone marrow samples from patients with AML andnormal controls, and in vivo in immunodeficientmice xenograftedwith ARC knockdown (ARC KD) AML cells. We demonstrate that

bothCox-2 expression andPGE2 generation are ARC/IL1bdepen-dent and that ARC, regulated by b-catenin, is an integral compo-nent of an IL1b/PGE2/b-catenin circuit. Cox-2/PGE2, regulatedby ARC and induced by AML–MSC coculture contributes toMSC-mediated chemoprotection in AML.

Materials and MethodsCells, cell culture, and cell treatments

OCI-AML3 cells, provided by Dr. M. Minden (Ontario CancerInstitute, Toronto, Ontario, Canada) were validated by STR DNAfingerprinting using the AmpF_STR Identifier Kit according to themanufacturer's instructions (catalog number #4322288, AppliedBiosystems). The STR profiles were compared with known ATCCfingerprints, and to the Cell Line Integrated Molecular Authenti-cation database (CLIMA) version 0.1.200808 (http://bioinformatics.hsanmartino.it/clima/; ref. 24). The STR profile was identifiedas unique. Mycoplasma testing was performed using the PCRMycoplasma Detection Kit from Applied Biological Materials(catalog number #G238) per the manufacturer's instructions.Authenticated and Mycoplasma-free cells are stored under liquidnitrogen and are never kept in culture for >4 months. Primarysamples were acquired from patients with AML or normal con-trols after informed written consent following the institutionapproved protocol in accordance with Declaration of Helsinki.Patient characteristics are shown in Table 1. Mononuclear cells

Table 1. Patient characteristics

No. % Blast Source Treatments and responses Cytogenetics Mutations

1 76 PB Untreated 46,XY[20] NPM1, NRAS1, TP53, STAG2,DNMT3A, FLT3-ITD

2 58 PB Untreated Complex ASXL1, ETNK1, RUNX1, SETBP1, WT1, FLT3-ITD3 85 BM Treated with multiple agents Complex FLT3-ITD/D835, NA for other mutations4 90 PB Refractory/resistance AML 46,XY[20] RUNX1, WT1, FLT3-ITD/D8355 63 PB AML, progressed from MPN.

Treated with decitabineComplex NOTCH1, JAK, TP53

6 54 PB Refractory AML Complex RUNX1, TP537 73 PB Relapsed AML.

Resistance to multiple agentsComplex NRAS, TP53, FLT3-D835

8 74 PB Relapsed AML NA NA9 52 BM Untreated 46,XX[20] IDH1, WT1, DMNT3A10 55 BM Treated with 2016-1084

BAY and hydroxyurea47,XY,þ8[8]/46,XY[12] ASXL1, CREBBP, DNMT3A, IDH1,

JAK2, U2AF1, WT111 88 BM Treated with chemotherapy Complex PTPN11, CALR, CBL, FLT3-ITD/D835, WT1 STAT5A12 83 BM Treated with hydroxyurea 46,XY[20] TET2, NRAS, DNMT3A, U2AF13 68 BM Treated with chemotherapy Complex FLT3-D835, SH2B3, STAG214 81 BM Treated with hydroxyurea and cytarabine Complex CREBBP15 78 BM Treated with hydroxyurea Complex ASXL1, BRAF, DNMT3A, FLT3-D835,

NRAS, RUNX1, TET216 61 BM Treated with Hydroxyurea 46,XX[20] IDH2, NPM1 FLT3-ITD17 70 BM Newly diagnosed Complex JAK2, MPL, WT118 81 BM Transformed from MDS to AML.

Resistance to multiple therapiesComplex FLT3-ITD

19 40 BM Newly diagnosed Complex IDH1, TP5320 85 BM Relapsed AML 48,XY,þ13,þ13[20] RUNX1, ASXL121 96 PB Transformed from MDS to AML.

Resistance to multiple therapies46,XX,del(7)(q22q34)[20] TET2, WT1

22 87 PB Relapsed/refractory AML 46,XY[20] NPM1, MPL, NRAS, IDH2 FLT3-ITD23 88 PB Refractory AML Complex NA24 86 PB Newly diagnosed Pseudodiploid clone

46,XY,t(9;11)(p22;q23)[20]NA

25 93 PB Relapsed/refractory AML Complex FLT3-ITD, NA for other mutations26 85 PB Relapsed/refractory AML Complex NA27 94 PB Newly diagnosed 46,XX[20] FLT3-D835, NA for other mutations

Abbreviations: BM, bone marrow; NA, not available; PB, peripheral blood; WT, wild type.

Carter et al.

Cancer Res; 79(6) March 15, 2019 Cancer Research1166



http://bioinformatics.hsanmartino.it/clima/

http://bioinformatics.hsanmartino.it/clima/


were isolated from primary samples by density–gradient centri-fugation using Lymphocyte SeparationMedium (catalog number#25-072-CV, Corning). Human MSCs were isolated from bonemarrow samples obtained from healthy subjects as describedpreviously (25). Cell lines were cultured in RPMI1640 mediumand cells from primary samples and MSCs in a-MEM medium,both supplemented with 10% heat-inactivated FCS, 2 mmol/LL-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin.Cells were kept at 37�C in a humidified atmosphere of 5% CO2.ARCKDAML cells (2) andMSCs (26)were generated as describedpreviously. For coculture experiments, leukemia cells were addedto MSCs (4:1) that were plated the night before and cultured ina-MEM medium with supplements. AML cells, MSCs, or thecocultured cells were treated with IL1b (catalog number #200-01B)withorwithout IL1RA (catalog number #200-01RA) (Pepro-Tech), dmPGE2 (16,16-dimethyl-PGE2, a PGE2 analogue; cata-log number #14750, Cayman Chemical), Ara-C with or withoutCox-2 inhibitor celecoxib (catalog number #C-1502, LC Labora-tories), or b-catenin inhibitor C-82 (provided by PRISM Pharma;refs. 27, 28) with or without Ara-C.

Determination of PGE2 and human IL1bPGE2 in the supernatant of cultured cells, mouse serum, and

human bone marrow samples was determined by PGE2 ELISA(catalog number #KGE004B), and human IL1b in mouse serumand human bone marrow samples was determined by ELISAspecific for human IL1b (catalog number #DLB50; both fromR&D Systems) following the manufacturer's instructions.

Silencing of b-catenin expressionOCI-AML3 cells were electroporated with a scramble control or

the SMARTpoolON-TARGET plusCTNNB1 siRNAs (750 nmol/L;catalog number #L-003482-00-0005, Dharmacon) using anAmaxa apparatus (Solution T, catalog number #VCA-1002; pro-gram X-001; Lonza) following the manufacturer's instructions asdescribed previously (28).

RNA isolation and RT-PCRRNA isolation and TaqMan RT-PCR were carried out as

described previously (29). Primers sets used are: b-catenin(Hs00355049_m1), ARC (Hs00358724_g1), and ABL(Hs01104728_m1; all from Thermo Fisher Scientific). The abun-danceof each transcript relative to that of ABLwas calculatedusingthe 2�DCt method, where DCt is the mean Ct of the transcript ofinterestminus themeanCt of the transcript for ABL housekeepinggene.

ARC promoter-GFP reporter constructs and lentiviraltransduction

The promoter region of the ARC coding gene NOL3, identifiedusing the Eukaryotic Primer Database (https://epd.vital-it.ch/index.php; residues 67173446-67174045, RefSeq NC_000016.10, human chromosome 16, GRCh38.p7) has two LEF1/TCF4Emotifs (CTTTGTGC and GGCCAAAG) by searching PROMO3.0 (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB ¼ TF_8.3). We amplified by PCR and inserted thisregion between the Cla I and Nhe I sites in the lentivectorpCDH-CMV-MSC-EF1a-Puro (System Biosciences), to replacethe CMV promoter. We then inserted the open reading forcopGFP between the Nhe I and BamH1 sites such that the startcodon for the GFP moiety was in frame with the NOL3 startcodon. We mutated the LEF1/TCF4E motifs using a Quik-

Change II XL Site-Directed Mutagenesis Kit as directed by themanufacturer (Agilent), except that we used NEBStable cells(New England Biolabs) in lieu of XL10-Gold cells. Mutantswere identified by the presence of both Pml I and Sal I sites.Primers used for these constructions are listed in Supplemen-tary Table S1. Lentivirus was prepared by transfecting HEK293Tcells (ATCC) with an equimolar mix of reporter vector andpackaging plasmids psPAX2 and pMD2.G (gifts of DidierTrono, Addgene) using JetPrime transfection reagent as directedby the manufacturer (Polyplus). OCI-AML3 cells were trans-duced with the lentivirus as described previously (2).

Western blot analysisProtein levels were determined by Western blot analysis as

described previously (3) using the Odyssey Infrared ImagingSystem for signal detection and Odyssey software version 3.0for quantification (LI-COR Biosciences). Cytoplasmic andnuclear fractions were prepared as described previously (30).Antibodies against b-catenin (Cat#8480) and ARC (Cat#NBP2-41753) were purchased from Cell Signaling Technology andNovus, respectively. Histone H3 was used as loading control fornuclear fraction, a-tubulin for cytoplasm, and b-actin for totallysate.

Protein determination by flow cytometryAfter staining with Ghost Dye Violet 510 (catalog number

#13-0870-T500, Tonbo Biosciences), cells were washed and fixedwith 4% paraformaldehyde and permeabilized with 100%meth-anol, and then stained with Fc-blocker (catalog number #130-059-901,Miltenyi Biotec), followedwith Cox-2-PE (catalog num-ber #12282, 1:50, Cell Signaling Technology), CD90-PerCP/CyC5.5 (catalog number #328118), and CD45-Pacific Blue (cat-alog number #304029; BioLegend) in 5% BSA/PBS. The stainedcells were analyzed using a Gallios flow cytometer (BeckmanCoulter Life Sciences) and quantified using FlowJo analytic plat-form (BD Biosciences). Cox-2 levels in MSCs or AML cells wereexpressed as the geometric mean fluorescent intensity (MFI)difference of cells stained with Cox-2 antibody or with IgG inCD90þ or CD45þ cells, respectively.

CyTOF analysisAML patient samples were stained with a panel ofmetal-tagged

antibodies for cell surface markers and intracellular proteins(Supplementary Table S2) and subjected to CyTOF analysis asdescribed previously (3, 28, 31). Viable (cisplatin low) single cellswere gated with FlowJo software (v10, FlowJo LLC) and exportedas flow cytometry standard (FCS) data for subsequent analysis inCytofkit (32). RPhenoGraph was used for unsupervised subsetdetectionbased on cell surfacemarkers. t-SNE embedded FCSfileswere further analyzed in FlowJo and cell populations identified byRPhenoGraph were mimicked and gated on the t-SNE map forquantitation of intracellular marker expression. The expressionlevel of each protein in the desired cell compartments is expressedas ArcSinh-transformed data.

Apoptosis assayApoptosis was estimated in CD45þ cells by flow cytometry

after Annexin V–Cy5.5 staining in the presence of 7-amino-actinomycin D (7-AAD) using LSRII flow cytometry (BD Bios-ciences). Apoptosis in primary samples was assessed in CD45þ

andCD34þCD38� cells after the cells were incubatedwith CD34-

ARC in Leukemia and Microenvironment

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1167



https://epd.vital-it.ch/index.php

https://epd.vital-it.ch/index.php

http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB = TF_8.3




PE,CD38-PECy7, andCD45-APCH7antibodies (BDBiosciences)and expressed as specific apoptosis:

% of apoptosis in treated cells � %of apoptosis in untreated cells%of viable untreated cells

�100%

In vivo studyNOD/SCIDIL2RgNull-3/GM/SF (NSGS)mice (male, 3months

old) were injected with either vector control or ARC KD OCI-AML3 cells (1�106/mouse). Aweek later, control or ARCKDcell-injected mice (10/group) were either untreated or treated withAra-C (100mg/kg) via intraperitoneal injection, 3 times per weekfor approximately 5 and half weeks. Leukemia burden wasassessed by flow cytometric measurement of human CD45þ

(huCD45þ) cells in mouse periphery blood, bone marrow, andspleen. Serum IL1b and PGE2 levels were determined. Tissue sizeswere measured and protein expression in mouse tissues wasdetermined by Opal/TSA (tyramide signal amplification) multi-spectral IHC and Vectra imaging (see below). Mouse survival wasfollowed. All the mouse experiments were carried out followingprotocols approved by the Institution Animal Care and UseCommittee at MD Anderson Cancer Center.

Protein determination by Opal/TSA multiplex IHC and Vectramultispectral imaging

Paraffin-embedded mouse tissue slides were deparaffinizedand rehydrated followed by antigen retrieval with BiogenexLaboratories Antigen Citra RTUSE (catalog number#NC0359148, Thermo Fisher Scientific) at 95�C for 15 minutesand blocked with 3% BSA in TBS-T for additional 15 minutes.Tissueswere incubatedwith aprimary antibody followedbyHRP-conjugated IgG (broad spectrum) provided in SuperPicture Poly-mer Detection Kit (catalog number #878963, Thermo FisherScientific) for 15 minutes at room temperature and then labeledwith TSA plus reagents following the manufacturer's protocols(PerkinElmer). The above steps were repeated sequentially forstaining eachof the following proteins: ARC,CD45, andb-cateninall specific for human. For ARC detection, tissues were stainedovernight at 4�C with an anti-ARC antibody (1:4,000, catalognumber #38916S, Cell Signaling Technology) and for b-catenindetection, they were incubated 1 hour at room temperature withan anti–b-catenin antibody (1:1,000; catalog number #LS-B4123,LifeSpan BioSciences Inc.). For huCD45 detection, tissues werestained 2 hours at room temperature with an anti-huCD45antibody (1:2,000; catalog number #M0701, AgilentDako). Opalfluorescein reagents used for TSA were Opal 540 (catalog number#FP1494001KT), Opal 570 (catalog number #FP1488001KT),Opal 650 (catalog number #FP1296001KT; PerkinElmer) forCD45, ARC, and b-catenin; respectively. Opal fluorescent reagentwas diluted with 1� plus amplification diluent (1:150; catalognumber #FP1498, PerkinElmer). Each issue was incubated with200 mL diluted Opal reagent at room temperature for 5 minutesand washed with TBS-T twice, 3 minutes each. Finally, the tissueslides were counterstained with 2.5 mg/mLDAPI (catalog number#D21490, Thermo Fisher Scientific) at room temperature for 5minutes, sealed by anti-fade fluorescent mounting medium (cat-alog number #S3023, Dako), and stored at 4�C.

Multispectral images were acquired using Vectra 3.0 automatedquantitative pathology imaging system (CLS142338, PerkinEl-mer) with �20 objective and software version 3.012 (PerkinEl-

mer). Images were analyzed by InForm version 2.3 software(PerkinElmer). Briefly, a fluorophore library was created bymonospectral labelled tissue images to define spectral curves foreach fluorophore and counterstain to adjust background noiseand positive staining of biomarkers. InForm software providesobjective counting of cell population and biomarkers, allowstissue separation by machine training, and increases the accuracyof the statistical analysis. Algorithms were created for separatingtumor tissues and normal tissues and for evaluating variousprotein levels via specific antibodies. For protein quantification,huCD45þ cells were selected from multiple fields and multiplecells in eachfield. The scanfields for spleenswere selected from thecentral region of the tissue section to maximally cover the wholespleen avoiding themargins, and the fields for bonemarrowwereselected from the whole marrow compartments away from bonefragments. The expression level is expressed as mean opticaldensity (OD)/huCD45þ cell in each field.

Statistical analysesExperiments were conducted in triplicate unless otherwise

indicated. One-way ANOVA or Student t test was performed tocompare the differences between groups. Results were expressedas the mean � SEM. Correlation coefficient was determined byPearson correlation analysis. The combination index (CI),expressed as the mean of CI values at the 50%, 75%, and 90%effectivedoseswasdeterminedby theChou–Talalaymethod (33).CI < 1.0 was considered synergistic; ¼ 1.0 additive; and >1.0antagonistic. Mouse survival was estimated by the Kaplan–Meiermethod and analyzed using log-rank statistics. Statistical signif-icance was set at P < 0.05.

ResultsAML-MSC cocultures increase Cox-2 expression in MSCs in anIL1b- and ARC-dependent manner

MSCs were treated with IL1b or cocultured with OCI-AML3cells in the absence or presence of IL1R antagonist IL1RA (48hours). Cox-2 protein levels were determined by flow cytometryin both MSCs (CD90þ) and OCI-AML3 cells (CD45þ). Asexpected, IL1b greatly induced the expression of Cox-2 in MSCs,which was suppressed by IL1RA. Cocultures of MSCs with OCI-AML3 cells also significantly increased Cox-2 expression inMSCs,and similarly this induction was inhibited by cotreatment of cellswith IL1RA (Fig. 1A), suggesting that increased Cox-2 expressionin MSCs cocultured with AML cells is mediated through IL1b. Bycomparison,OCI-AML3 cells expressed lower levels of Cox-2 thanMSCs, and coculture of AML cells withMSCs also increased Cox-2expression in AML cells in an IL1b-dependent manner, but to alesser degree (Fig. 1A). Importantly, similar results were obtainedwhen primary AML patient samples (Table 1, samples 1–3) wereused (Fig. 1B) and IL1b, even at 25 pg/mL level, was sufficient tosignificantly increase Cox-2 in MSCs (P ¼ 0.013).

We next cocultured MSCs with vector control or ARC KD OCI-AML3 cells, vector control or ARCKDMSCswithOCI-AML3 cells,or vector control or ARC KDMSCs with vector control or ARC KDOCI-AML3 cells (48hours).We found that significantly less Cox-2proteinwas induced inMSCswhenARCwas knockeddowneitherin AML cells (P¼ 0.00025) orMSCs (P¼ 0.0083) compared withthe respective controls and that ARC KD in both MSCs and AMLcells resulted in the lowest level of Cox-2 expression (P ¼ 0.007,0.021, or 0.001, versus ARC KD inMSCs, ARC KD in AML cells, or

Carter et al.





90,000

60,000

30,000

0

75,000

50,000

25,000

0

60,000

40,000

20,000

0

75,000

50,000

25,000

0

30,000

20,000

10,000

0

n = 3P = 0.0085

P = 0.015

P = 0.00025P = 0.0083

P = 0.001

P = 0.021P = 0.007

P = 0.015

P = 0.001P = 0.009

P = 0.013

n = 3 n = 3

n = 3

P = 0.008

P = 0.0059

MSC

MSC MSC+IL1b MSC+IL1b MSC+IL1b MSC+IL1b MSC+AML MSC+AMLAML

OCI-AML3 MSC+OCI-AML3 MSC+OCI-AML3IL1RA

MSC+IL1b MSC+IL1bIL1RA

Cox

-2 (s

tain

ed-Ig

G) M

FIC

ox-2

(sta

ined

-IgG

) MFI

CD90+ MSC

CD90+ MSC

CD90+ MSC CD90+ MSCCD45+ AML CD45+ AML

CD90+ MSCCD45+ AML

CD45+ AML

CD45+ AML patientsAML patient samples n = 3

100 ng/mL100 pg/mL25 pg/mL 100 ng/mL 100 ng/mL

100 ng/mLIL1bIL1RA

CD90+ cells CD45+ cells

Cox-2 Cox-2

Cox-2 stainingCox-2 staining

IgG staining IgG staining

MSCMSC+IL1bMSC+IL1b/IL1RAMSC+OCI-AML3MSC+OCI-AML3+IL1RA

OCI-AML3OCI-AML3+MSCOCI-AML3+MSC+IL1RA

MSCs coculture withvec or ARC KD OCI-AML3

Vec or ARC KD MSCs coculture Vec or ARC KD MSCs coculture

OCI-AML3Vec VecARC KD ARC KD

ARC

with OCI-AML3 with vec or ARC KD OCI-AML3MSC

b-Actin

ARCb-Actin

Cox

-2 (s

tain

ing-

IgG

) MFI

Cox

-2 (s

tain

ed-Ig

G) M

FI

Cox

-2 (s

tain

ed-Ig

G) M

FI

MSC+ OCI- OCI-AML3 +OCI-AML3 OCI-AML3 OCI-AML3 OCI-AML3

OCI-AML3 +MSC+ OCI-AML3 Vec AML3 ARC KD MSC-Vec MSC-ARC KD

Vec + MSC- Vec + MSC-Vec

ARC KD +MSC-Vec MSC-ARCARC KD

ARC KD +

KD

A

B

C

Figure 1.

AML-MSC cocultures increase Cox-2 in MSCs in an IL1b- and ARC-dependent manner. A, Cox-2 expression in MSCs (CD90þ) and OCI-AML3 (CD45þ) cellscultured alone or cocultures, without or with IL1b (100 ng/mL) and/or ILIRA (100 ng/mL) treatments for 48 hours. Top, histogram of one representativeexperiment; bottom, results of three independent experiments with three different MSCs. B, Cox-2 expression in MSCs (CD90þ) and primary patient samples(CD45þ; n¼ 3, Table 1, samples 1–3) cultured alone or cocultures, without or with IL1b (25 pg/mL, 100 pg/mL, or 100 ng/mL) and/or ILIRA (100 ng/mL)treatments for 48 hours. The same MSC was used for all three cocultures. C, Cox-2 expression in MSC and OCI-AML3 cells by flow cytometry under MSCscocultured with control or ARC KDOCI-AML3 (left), control or ARC KDMSCs cocultured with OCI-AML3 (center), or control or ARC KDMSCs cocultured withcontrol or ARC KDOCI-AML3 (right) for 48 hours. Vec, vector control. Three independent experiments were performed.






vector control; respectively; Fig. 1C), indicating that Cox-2 expres-sion depends on ARC levels, which is in agreement with ourprevious finding that ARC regulates NFkB/IL1b signaling (3).

AML-MSC cocultures induce PGE2 in an IL1b- and ARC-dependent manner

As shown in Fig. 1, although Cox-2 is expressed in OCI-AML3cells and primary patient samples, AML cells express relativelylower Cox-2 than MSCs. We next determined PGE2 levels in theculture media of OCI-AML3 andMSCs in the absence or presenceof IL1b (48 hours) and found that compared with OCI-AML3cells,MSCs secreted >100-fold PGE2/cell. IL1b, at 1 to 100ng/mL,significantly induced PGE2 production in MSCs (6.5- to 9.2-fold;P < 0.0001), whereas at only 100 ng/mL, IL1b significantlyinduced PGE2 production in OCI-AML3 cells (2.5-fold; P ¼0.002; Fig. 2A), indicating that consistent with Cox-2 expression,MSCs are the major source of PGE2.

We then coculturedMSCs andOCI-AML3 cells for 48hours andfound markedly increased PGE2 levels in the medium of cocul-turedcells comparedwithMSCsorOCI-AML3culturedalone.Thisincreasewas suppressedby IL1 receptor antagonist IL1RAorCox-2inhibitor celecoxib (Fig. 2B).Wenext coculturedMSCswith vectorcontrol or ARC KD OCI-AML3 cells for 48 hours and found thatknockingdownARC inOCI-AML3cells significantlydecreased thesecretedPGE2 in the coculture system (P<0.0001; Fig. 2C, left). Inaddition, we found that in cocultured vector control or ARC KDMSCs with OCI-AML3 cells, knockdown of ARC in MSCs signif-icantly reducedPGE2production(P¼0.00081;Fig.2C, right).Ourdata suggest that, as with Cox-2 expression, AML-MSC cocultureinduces the production of PGE2 in an ARC- and IL1b-dependentmanner. Finally, coculture cells from peripheral blood samples ofprimary AML patients (n ¼ 5, Table 1, samples 4–8) with MSCs(48hours)markedlyincreasedsecretedPGE2,whereasPGE2levelswere significantly lower when patient cells were cocultured withARC KD MSCs (P ¼ 0.0027; Fig. 2D).

PGE2 treatment or coculture with MSCs induces b-catenin andARC and augments chemoresistance in AML cells

We previously showed that coculture of AML cells with MSCsincreases b-catenin (27) and ARC expression in AML (2). Todemonstrate that this effect is partially mediated through PGE2signaling, OCI-AML3 cells were treated with a PGE2 analogue,dmPGE2 or cocultured with MSCs without or with Cox-2 inhib-itor celecoxib. dmPGE2 induced protein levels of b-catenin andARC (Fig. 3A). As expected, coculture with MSCs increasedb-catenin and ARC in FACS-sorted OCI-AML3 (CD45þCD90�)cells after coculture (Fig 3A), and this increase was completelysuppressed by celecoxib (Fig. 3A). Importantly, both PGE2and cocultures primarily increased nuclear b-catenin in AMLcells (Fig. 3B). To examine the biological relevance of increasedPGE2 production in the leukemia–MSC system, we treatedOCI-AML3 cells with Ara-C, the commonly used chemothera-peutic agent for AML therapy with or without dmPGE2 or MSCs.As shown in Fig. 3C, Ara-C induced time and dose-dependentapoptosis in OCI-AML3 cells, which was significantly suppressedby MSC coculture and by dmPGE2 cotreatment in a dose-dependent manner. Blocking PGE2 production with celecoxib(200 nmol/L) significantly reduced the protective effect of MSCs.Our data indicate that increased PGE2 production in leukemia–MSC coculture system activates b-catenin, increases ARC expres-sion, and contributes to MSC-mediated chemoprotection.

ARC expression in AML cells is regulated by b-cateninWe have demonstrated that in AML-MSC cocultures, ARC

regulates IL1b in AML cells, induces Cox-2 in MSCs, andincreases secreted PGE2, which in turn induces the expressionof b-catenin in AML cells. Interestingly, in addition to b-catenin,PGE2 also increased ARC level. Once activated, b-catenin istranslocated to the nucleus where it interacts with DNA-bindingproteins T-cell factors (TCF)/lymphoid enhancer factors (LEF)that recognize and bind to specific sequencemotifs in promotersand enhancers of target genes (34). Promoter analysis demon-strated that the ARC promoter region contains two binding sitesfor TCF4 (CTTTGTG and GCCAAAG)/LEF1 (CTTTGTGC andGGCCAAAG), suggesting that ARC may be regulated by b-cate-nin. To test that, we silenced b-catenin in AML cells by siRNAs. Asshown in Fig. 4A, inhibition of b-catenin suppressed ARCexpression both at the RNA and protein levels, suggesting thatb-catenin regulates ARC expression transcriptionally, and thatthe PGE2-induced ARC increase is mediated through PGE2/b-catenin signaling.

We next generated an ARC promoter–driven GFP reporter andtransduced the construct intoOCI-AML3 cells. Silencing b-catenindecreased the level of GFP reporter in OCI-AML3 cells (Fig. 4B).We also mutated LEF1/TCF4Emotifs in the ARC promoter regionto generate an ARCmutant promoter–driven GFP reporter. Figure4C shows that dmPGE2 markedly increased, and inhibition ofb-catenin with a specific inhibitor C-82 decreased GFP levels inOCI-AML3 cells transduced with ARC promoter–driven GFP.dmPGE2 or C-82 had only minimal effects on GFP levels in cellstransduced with ARC-mutant promoter-driven GFP, further sup-porting that PGE2/b-catenin signaling transcriptionally regulatesARC expression.

Inhibition of ARC in AML decreases IL1b and PGE2, exhibitsantileukemia activity, and sensitizes AMLcells to chemotherapyin vivo in a human xenograft NSGS model

To assess the biological relevance of ARC/IL1b/PGE2/b-cateninsignaling in AML in vivo, we injected NSGS mice with vectorcontrol or ARC KD OCI-AML3 cells. Mice were left untreated ortreatedwithAra-C andperipheral blood and tissueswere collected24 or 25 days after leukemia cell injection (Fig. 5A). We firstdetermined the levels of IL1b and PGE2 inmouse serum by ELISAand found a statistically significant decrease in serum IL1b andPGE2 in mice transplanted with ARC KD compared with controlAML cells (Fig. 5B). In addition, mice with ARC KDAML cells hadlower leukemia burden in all tissues examined. Specifically,significantly less huCD45þ cells were found in mouse bonemarrows and spleens (Fig. 5C), along with markedly reducedliver and spleen size (Fig. 5D) compared with mice injected withcontrol cells.

We next determined the expression levels of ARC and b-cateninin huCD45þ cells in mouse spleen and bonemarrow using Opal/TSA IHC and Vectra imaging that allows simultaneous measure-ment of multiple markers within a single tissue section. Figure 5Eshows the individual or mixed fluorescent images of mousespleen stained with huCD45, ARC, b-catenin, and DAPI (left)and quantification of ARC and b-catenin protein levels (right) inthe spleens of all four experimental groups. We analyzed 35 fieldsin control cell-injected mice (7,164–10,873 cells/field, total319,431 cells), 26 fields in control cell-injected mice with Ara-C treatment (2,635–4,777 cells/field, total 99,727 cells), 17 fieldsin ARCKD cell-injectedmice (610–5,021 cells/fields, total 37,206

Carter et al.





cells), and 22 fields in ARC KD cell-injected mice with Ara-Ctreatment (248–3,927 cells/field, total 45,431 cells); respectively.As expected, mice injected with ARC KD OCI-AML3 cells with orwithout Ara-C treatment expressed drastically lower levels of ARCprotein inhuman spleenCD45þ cells (P<0.0001) comparedwith

those injected with control cells with or without treatment.Interestingly, their spleen huCD45þ cells also had significantlylower levels of b-catenin (P < 0.0001 with Ara-C treatment, P <0.01 without Ara-C treatment), supporting the existence of anARC/IL1b/PGE/b-catenin/ARC regulatory loop. Similarly, mice

100,0001,000,000

100,000

10,000

1,000

100

10

1

10,000

1,000

100

10

1

n = 3

n = 3n = 3

P = 0.002

P = 0.0001 P = 0.0005P < 0.0001

P < 0.0001P < 0.0001P < 0.0001

OCI-AML3

MSC48 h

48 h

48 h

IL1b (ng/mL) IL1b (ng/mL)0 1 10 100 0 1 10 100

PGE2

(pg/

mill

ion

cells

)PG

E2 (p

g/m

L)

PGE2

(pg/

mL)

PGE2

(pg/

mill

ion

cells

)

PGE2

(pg/

mL)

6,000

4,000

2,000

0MSC OCI-AML3 co-cul co-cul MSC co-cul

+IL1RA +Celecoxib

2,500

2,000

1,500

1,000

500

0

2,500

2,000

1,500

1,000

500

0

n = 3n = 3

P = 0.00006

P = 0.00081

Coculture of MSCs with

Coculture of AML patient samples Coculture of AML patient samples

Coculture of control or ARC KD MSCscontrol or ARC KD OCI-AML3 with OCI-AML3

MSC

MSC Vec

MSC ARC KD

OCI-AML3 V

ec

OCI-AML3 A

RC KD

OCI-AML3 p

arental

OCI-AML3 p

arental

OCI-AML3 +

MSC-Vec

OCI-AML3 +

MSC-A

RC KD

MSC + OCI-A

ML3 Vec

MSC + OCI-A

ML3 ARC K

D

MSC

with MSCs with ARC KD or control MSCs20,000

15,000

10,000

5,000

0

20,000

15,000

10,000

5,000

0AML AML + MSC MSC MSC-Vec MSC-ARC KD

48 h

n = 5 n = 5 P = 0.0027

PGE2

(pg/

mL)

PGE2

(pg/

mL)

# 4# 5# 6# 7# 8Mean

# 4# 5# 6# 7# 8Mean

A

C

D

B

Figure 2.

AML-MSC cocultures increase secreted PGE2 in an IL1b- and ARC-dependent manner. A,OCI-AML3 cells or MSCswere cultured in the absence or presence of IL1b(1–100 ng/mL) for 48 hours. B,OCI-AML3 and MSCs were cultured alone or cocultured (co-cul) without or with IL1RA (100 ng/mL) or celecoxib (200 nmol/L) for48 hours. C,MSCswere cocultured with vector control or ARC KDOCI-AML3 (left) and vector control or ARC KDMSCs were cocultured with OCI-AML3 cells(right) for 48 hours. D, Cells from AML patient samples (n¼ 5, Table 1, samples 4–8) were cocultured with MSCs (left) and with ARC KD or control MSCs (right)for 48 hours. PGE2 levels in the supernatant were determined by ELISA. Vec, vector. Cell line experiments were done in triplicates. Three MSCs were used forA and B.






injected with ARC KD OCI-AML3 cells with or without Ara-Ctreatment expressed significantly lower levels of ARC and b-cate-nin in bonemarrow huCD45þ cells compared with those injectedwith control cells without treatment. However, in bone marrow,Ara-C–treated groups had lower ARC and b-catenin levels com-pared with their respective untreated controls, especially for bonemarrow cells obtained from control cell–injected mice (Supple-mentary Fig. S1). The reason for this difference is unclear. We didobserve that in bone marrow samples, cells from Ara-C–treatedmice did not attach to slides well compared with cells fromuntreatedmice, which was not observed in spleen samples. Lowerleukemia burden in mice with ARC KD AML cells was furtherdemonstrated by Opal/TSA staining of huCD45 in spleen andbone marrow and Vectra imaging (Fig. 5E; SupplementaryFig. S1).

Furthermore, both ARC inhibition, and the chemotherapeuticdrug Ara-C, significantly lowered leukemia burden in peripheralblood assessed by flow cytometry of huCD45þ cells (Fig. 5F) andprolonged survival (Fig. 5G). Mice with ARC KD AML cells livedlonger (median survival, 33 days; P ¼ 0.0006), similar to micewith control cells treated with Ara-C (34 days, P ¼ 0.0003)

compared with those with control cells not receiving treatment(27 days; Fig. 5G). Importantly, inhibition of ARC sensitized AMLto Ara-C: mice injected with ARC KD cells treated with Ara-Cshowed the longest median survival time (40.5 days; P¼ 0.0006,0.0004, and 0.0012 vs. control, control þ Ara-C, and ARC KD;respectively; Fig. 5G).

IL1b/PGE2/b-catenin/ARC cascade and targeting in primaryAML samples

To assess potential roles of IL1b and PGE2 in AML and the AMLbone marrow microenvironment, we collected bone marrowfrom patients with AML (n ¼ 8, Table 1, samples 9-16) andnormal controls (n ¼ 4) and determined IL1b and PGE2 levels:both IL1b and PGE2 levels were significantly higher in bonemarrow from patients with AML compared with normal controls(Fig. 6A).We then determinedb-catenin andARCprotein levels inAML patient bone marrow cells (Table 1, samples 17–20) byCyTOF mass cytometry in various cell compartments andobserved strong correlations of b-catenin and ARC levels, in bulkas well as in CD34þCD38þ and CD34þCD38� stem/progenitorcells (Fig. 6B). Furthermore, inhibition of b-catenin by C-82

A OCI-AML3

dmPGE2, mmol/L

dmPGE2, mmol/L

b-Catenin

b-Catenin

a-Tubulin

Histone H3

48 h

Ara-C (mmol/L)

0.00

60

40

20

0

60

40

20

00.50 1.00

Ara-C only

Ara-C + COXAra-C + Celecoxib-COX

CD

45+ A

nn

V+ /

7AA

D+

(%)

CD

45+ A

nn

V+ /

7AA

D+

(%)

Ara-C + 1 mmol/L PGE2 Ara-C + 4 mmol/L PGE2

Ara-C only

Ara-C + COXAra-C + Celecoxib-COX

Ara-C + 1 mmol/L PGE2 Ara-C + 4 mmol/L PGE2

2.00 4.00 0.00 0.50 1.00 2.00 4.00

Ara-C (mmol/L)

72 h

b-Actin

Celecoxib, nmol/L

ARC

200

CD90

OCI-AML3Cytosol

n = 3

Nucleus

CD

45

CD45+CD90–Leuk

CD45 PE+

AP

4+ +

+ +4 4

1MSC

MSC

B

C

Figure 3.

AML cells treated with dmPGE2 orcocultured with MSCs expressincreased b-catenin and ARC andare more resistance tochemotherapy. A,OCI-AML3 cellswere treated with dmPGE2 (1 or 4mmol/L) or cocultured with MSCswithout or with Cox-2 inhibitorCelecoxib (200 nmol/L) for 48hours. b-Catenin and ARC proteinlevels were determined byWesternblot analysis. For cocultureexperiments, OCI-AML3 cells wereFACS-sorted after coculture, andWestern blot analysis was doneusing the lysate from the sortedCD45þCD90� cells as shown in thehistogram. B,OCI-AML3 cells weretreated with dmPGE2 (4 mmol/L) orcocultured with MSCs for 48 hours.Cytosolic and nuclear b-cateninlevels were determined byWesternblot analysis. C,OCI-AML cells weretreated with Ara-C with or withoutdmPGE2 (1 or 4 mmol/L) or withMSC coculture in the absence orpresence of Cox-2 inhibitorcelecoxib (200 nmol/L) for 48 and72 hours. Apoptosis was assessedby flow cytometry. COX, coculture.

Carter et al.





b-CateninARC

24 h

Rel

ativ

e to

co

ntr

ol

(AB

L)

1.2

0.8

0.4

0.0

A

ARC promoter-drivenGFP

B

Scramble CTNNB siRNA

siRNA750 nmol/L

ARC

b-Catenin

b-Actin

48 h

CON CTNNB1

24 h

CON CTNNB1

OCI-AML3

MW

M

C ARC promoter-drivenGFP

ARC mutant promoter-driven GFP

GFP-cells

GFP-cells

Control

2

4

dmPGE2(mmol/L)

Scramble

CTNNBsiRNA

2

4

dmPGE2(mmol/L)

Control

GFP-cells

100 101 102 103 104

GFP 100 101 102 103 104

GFP

0.25

0.50

C-82(mmol/L)

C-82(mmol/L)

100 101 102 103 104

Comp-FITC-A:: GFP100 101 102 103 104

Comp-FITC-A:: GFP

100 101 102 103 104

GFP

0.25

0.50

GFP-cells GFP-cells

ControlControl

n = 3P < 0.001

n = 3 48 h

ARC WT

ARC Mutant

C-82 (mmol/L)

0.25 0.50Control42

dmPGE2 (mmol/L)

ControlScramble CTNNB siRNA

Rel

ativ

e to

th

e re

spec

tive

con

tro

l (G

FP

+ M

FI)

3

2

1

0Rel

ativ

e to

th

e co

ntr

ol

(GF

P+

MF

I)

1.2

0.8

0.4

0.0

24 h

Figure 4.

b-Catenin regulates ARC expression in AML cells. A,OCI-AML3 cells were transfected with a control siRNA (scramble) or SMARTpool siRNAs against b-cateningene (CTNNB1 siRNA; 750 nmol/L) by Amaxa electroporation. The RNA (24 hours) and protein (24 and 48 hours) levels of b-catenin and ARCwere determinedby RT-PCR andWestern blot analysis, respectively, after transfection. CON, scramble control. MWM, molecular weight marker. B,OCI-AML3 cells transducedwith ARC promoter–driven GFP reporter gene were transfected with scramble control or b-catenin siRNAs, and GFP levels were determined 24 hours aftertransfection. C,OCI-AML3 cells transduced with ARC promoter (WT)- or ARC-mutant promoter-driven GFP reporter gene were treated with dmPGE2 or C-82 for48 hours. B and C, Histograms (top) are representative results from one experiment and bar grafts (bottom) are results of triplicates. GFP levels were determinedby flow cytometry. Untransduced OCI-AML3 cells were used as a negative control for GFP.






ARC KD orcontrol OCI-AML3 cells

TissuePB harvest END

IL1bP < 0.0001 PGE2

P < 0.0001

IL1b

(pg

/mL

)

400

300

200

100

0

7 2425 44 49Day 0Ara-C, three times a week

Control ARC KD No cells Control ARC KD No cells

PG

E2

(pg

/mL

) 2,000

1,500

1,000

500

0

A B

P = 0.025 P = 0.0012

Control ARC KD Control ARC KD

Liver

Spleen

C

hu

CD

45+

cells

(%

), B

M

30

20

10

0

hC

D45

+ ce

lls (

%),

sple

en

100

50

5

Control ARC KD

ARC levels in huCD45+ cells within mouse spleen

Vector - AraC (n = 35)Vector + AraC (n = 26)

ARC KD - AraC (n = 17)ARC KD + AraC (n = 22)

Vector - AraC (n = 35)

Vector + AraC (n = 26)

ARC KD - AraC (n = 17)

ARC KD + AraC (n = 22)

b-Catenin levels in huCD45+ cells within mouse spleen

P < 0.0001

P < 0.0001

Pro

tein

inte

nsi

ty (

OD

/cel

l) 6

5

4

3

2

1

0

Control Control + Ara-C ARC KD ARC KD + Ara-C

Spleen

P < 0.01

P < 0.0001

Pro

tein

inte

nsi

ty (

OD

/cel

l) 18

16

14

12

10

8

6

CD45

ARC

Merged

DAPI

b-Catenin

Days after transplantation

10 20 30 40 50

Per

cen

t su

rviv

al

100

80

60

40

20

0

D

E

hu

CD

45+

cells

(%

) in

PB 5

4

3

2

1

0

-1

P < 0.0001P < 0.0001

P < 0.0001

Control Control+Ara C

ARC KD ARC KD+Ara C

CD45, Opal 540; ARC, Opal 570; b-catenin, Opal 650; DAPI, nucleus

F GGroup Median (days)

40.5

Ara-C

ControlP = 0.0003

ARC KD/Ara-C

ARC KD 33

34

27

P = 0.0012

P = 0.0006

P = 0.0004

P = 0.0006

Ara-C treatment

Figure 5.

Inhibition of ARC in AML cells reduces serum IL1b and PGE2 levels, decreases leukemia burden in various tissues, prolongs survivals, and sensitizes to Ara-C inNSGSmice.A, In vivo experiment scheme. B, IL1b and PGE2 levels in serum of mice harboring control or ARC KDOCI-AML3 cells determined by ELISA. C, huCD45positivity in the bone marrow (BM) and spleen of mice harboring control or ARC KD OCI-AML3 cells determined by flow cytometry. D, Liver and spleen frommiceharboring control or ARC KD OCI-AML3 cells. E, Expression of ARC and b-catenin in spleen huCD45þ cells of mice harboring control or ARC KD OCI-AML3 cellswith or without Ara-C treatment, determined by Opal/TSAmultiplex IHC staining and Vectra multispectral imaging analysis. Left, huCD45 (yellow), ARC (red),b-catenin (cyan), DAPI (blue), and merged imaging (objective lens�20; scale bar, 50 mm). Fluorophores for CD45, ARC, and b-catenin are Opal 540, Opal 570,and Opal 650, respectively. Right, the quantitation of ARC and b-catenin expression in huCD45þ cells of mouse spleen: 35 fields in control cell-injected mice(7,164–10,873 cells/field; total 319,431 cells), 26 fields in control cell-injected mice with Ara-C treatment (2,635–4,777 cells/field; total 99,727 cells), 17 fields inARC KD cell-injected mice (610–5,021 cells/fields; total 37,206 cells), and 22 fields in ARC KD cell-injectedmice with Ara-C treatment (248–3,927 cells/field; total45,431 cells), respectively. F, huCD45 positivity in peripheral blood of mice harboring control or ARC KDOCI-AML3 cells with or without Ara-C treatment by flowcytometry.G, Survival of mice injected with control or ARC KD OCI-AML3 cells without or with Ara-C treatment. PB, peripheral blood.

Carter et al.





Color keyand histogram

01_4401256_BM70Rphenograph

cluster mean heat map n = 4CD45+

R2 = 0.95

P = 0.0001

0 5 10ARC

15 20

0 5 10ARC

15 20

CD34+CD38+

R2 = 0.87P = 0.0021

6

4

2

0

b-C

aten

in

6

4

2

0

b-C

aten

in

CD34+CD38-R2 = 0.96

P = 0.00046

6

4

2

0

b-C

aten

in

0 5 10ARC

15 20

CD

45C

D38

CD

90T

im3

CD

11b

CD

123

CD

34

561108279121143

Clusters

n = 2, C-82 0.5 mmol/L, 48 h

CD34+CD38-CD34+CD38+CD45+

Alone Coculture Alone Coculture Alone Coculture

Con C-82 Con C-82 Con C-82 Con C-82 Con C-82 Con C-82Low High

P < 0.0001

P = 0.013

60

40

20

0BM

IL1b

(p

g/m

L)

A

NBM AML-BM

NBM AML-BMn = 4 n = 8

4,000

3,000

2,000

1,000

0BM

PG

E2

(pg

/mL

)

CD34+CD38-, 48 h

n = 4

Alone CI = 0.52 ± 0.01cocx CI = 0.79 ± 0.21

Alone CI = 0.82 ± 0.06cocx CI = 0.77 ± 0.01

n = 5

Bulk, 48 hC82 + Ara-C-aloneC82 + Ara-C-cocxC82-aloneC82-cocxAra-C-aloneAra-C-cocx

100

75

50

25

0

Sp

ecif

ic a

po

pto

sis

(%)

CD

34+ C

D38

-

100

75

50

25

0

Sp

ecif

ic a

po

pto

sis

(%)

CD

45+

D

E

0 0.125 0.25 0.5 1 2 C-82, mmol/L0 0.125 0.25 0.5 1 2 Ara-C, mmol/L

0 0.125 0.25 0.5 1 2 C-82, mmol/L0 0.125 0.25 0.5 1 2 Ara-C, mmol/L

C

B

CD34+CD38+CD34+CD38-

CD45+

Figure 6.

PGE2/b-catenin/ARC cascade and targeting in primary AML samples and the proposed mechanism of ARC action. A, IL1b and PGE2 levels in bone marrow (BM)samples from patients with AML (n¼ 8; Table 1, samples 9–16) and normal controls (NBM; n¼ 4) by ELISA. B, Correlation of ARC and b-catenin protein levels,determined by CyTOF in various bone marrow cell populations of patients with AML (n¼ 4; Table 1, samples 17–20). C, Levels of ARC protein, determined byCyTOF in AML patient samples (n¼ 2; Table 1, samples 21 and 22) treated with b-catenin inhibitor C-82 (0.5 mmol/L) without or with MSC coculture for 48 hours.Protein levels determined by CyTOF are expressed as ArcSinh-transformed counts. D, AML patient samples were treated with C-82, Ara-C, or both for 48 hours.Apoptosis was determined in blasts (n¼ 5; Table 1, samples 23–27) and CD34þCD38� (n¼ 4; Table 1, samples 23–26) cells. cocx, coculture. E, Proposedmechanism of action. ARC, regulated by b-catenin, mediates leukemia stromal interaction through ARC-IL1b/Cox-2/PGE2/b-catenin circuit.






decreased ARC and MSC coculture-induced ARC expression inCD45þ blasts and CD34þCD38þ and CD34þCD38� stem/pro-genitor cells from primary samples (Fig. 6C; n ¼ 2, Table 1,samples 21, 22). To demonstrate whether targeting the IL1b/PGE2/b-catenin/ARC cascade sensitizes to chemotherapy, wetreated primary AML cells with C-82 and Ara-C (Table 1, samples23-27): C-82 and Ara-C synergistically induced apoptosis in bulk(n ¼ 5) and CD34þCD38� cells (n ¼ 4) even when cells werecocultured with MSCs (CI < 1; Fig. 6D).

DiscussionARC was first identified as an antiapoptotic protein. We

reported that ARC is a strong adverse prognostic marker in AMLand protects leukemia cells from therapeutic agent-induced celldeath (1, 2) consistent with its antiapoptotic property. Interest-ingly, we also observed that ARC plays a role in leukemia–MSCinteractions. On the basis of our previous finding that ARCmodulates leukemia–microenvironment interactions by regulat-ing NFkB/ILb signaling (3), we here demonstrate that ARC reg-ulates IL1b in AML,which then induces Cox-2 expression inMSCsand PGE2 production, which in turn activates b-catenin andincreases ARC in AML, while ARC itself is regulated by b-catenin.The data support the concept that in addition to directly suppres-sing apoptosis, ARC, regulated by b-catenin, mediates leukemia–stromal interactions through the ARC-IL1b/Cox-2/PGE2/b-cate-nin circuit. Furthermore,ARC inMSCsalsoaffectsCox-2 levels andPGE2 production in AML-MSC coculture. We previously reportedthat ARC inMSCs induces nuclear and phospho-NFkB (3). NFkBis known to regulate Cox-2–catalyzed PGE2 synthesis (10–12).Therefore, ARC is an integral component of and a regulator ofleukemia–microenvironment interactions (Fig. 6E).

PGE2 was reported to suppress apoptosis in multiple cell typesthrough various mechanisms such as PI3K/AKT signaling orregulating survivin expression (35–37). In this study, we foundthat PGE2 protects AML cells from chemotherapy and stronglyinduces the expression of antiapoptotic ARC protein in AML cells,a new mechanism of PGE2-mediated antiapoptotic effect. Fur-thermore,wedemonstrated thatMSCs confer AMLdrug resistancein a PGE2-dependent manner. Our data reveal that AML cells,microenvironmental factors, and leukemia–stromal interactionscooperatively support leukemia cells through the novel ARC/IL1b/Cox-2/PGE2/b-catenin circuit reported here.

We discovered that ARC is transcriptionally regulated by b-cate-nin. PGE2 treatment or coculture with MSCs induces b-catenin inAML in agreementwith our previousfinding that leukemia cells inthemarrow had higher b-catenin levels than paired leukemia cellsin circulation (27), suggesting that the bone marrow microenvi-ronment induces b-catenin levels in AML cells at least in partthrough stromal cell secreted PGE2. Furthermore, using a xeno-graft model of human AML in NSGS mice, we demonstrated thatinhibition of ARC by shRNA significantly decreased leukemiaburden and serum IL1b/PGE2 levels, prolonged survival, andsensitized to chemotherapy. Interestingly, in addition todecreased ARC levels, low b-catenin levels in huCD45þ cells werealso detected in mouse tissues injected with ARC KD AML cellscompared with controls, supporting the ARC/IL1b/Cox/PGE2/b-catenin regulatory circuit in vivo.

We demonstrate that both IL1b and PGE2 levels are signifi-cantly higher in bone marrow from AML patients compared withnormal controls and that the levels of bonemarrow b-catenin and

ARC strongly correlate in AML bulk and stem/progenitor cells,supporting the regulatory cascade in primary AML. It is technicallychallenging to establish ARC knockdown or overexpression inprimary AML cells and currently no specific ARC inhibitor isavailable. However, targeting the components of the regulatorycircuit can be explored. We previously reported that inhibition ofb-catenin by its selective inhibitor PRI-724 (C-82 prodrug) hasantileukemia activity in vivo in immunodeficient mice engraftedeither with an AML cell line or AML PDX cells and sensitized totyrosine kinase inhibitor in FLT3 mutated AML (27). We hereshow that C-82 decreases ARC and sensitizes to Ara-C in bulk andCD34þCD38� AML cells even when they were cocultured withMSCs.

Evasion of immunity in hematologic malignancies has beenincreasingly recognized in recent years (38). Wnt/b-catenin andCox-2/PGE2 signaling are both implicated in immunosuppres-sion (39). As ARC is a component of the IL1b/Cox-2/PGE2/b-cate-nin cascade that connectsb-cateninand IL1b andgivenourfindingthathighARCexpression is a strongadversepredictorof survival inAML, we envision that higher ARC levels in AML may also con-tribute to a more inflammatory and immunosuppressive micro-environment, a hypothesis that is currently under investigation.

In conclusion, we demonstrate that ARC is an integral com-ponent of a novel IL1b/Cox-2/PGE2/b-catenin circuit, plays acritical role in leukemia growth and the leukemia microenviron-ment, and may serve as a potential novel therapeutic target inAML. We are currently trying to develop inhibitors targeting ARC.The finding that ARC functions as a survival modulator throughmultiple mechanisms explains its strong adverse effect in AML.This sets ARC apart from other antiapoptotic proteins.

Disclosure of Potential Conflicts of InterestJ.K. Burks is a consultant/advisory boardmember for Fluidigm. No potential

conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: B.Z. Carter, V. Ruvolo, M. AndreeffDevelopment of methodology: P.Y. Mak, W. Tao, V. Ruvolo, J.K. BurksAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): P.Y. Mak, X. Wang, W. Tao, V. Ruvolo, D. Mak,H. Mu, J.K. BurksAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): B.Z. Carter, P.Y. Mak, X. Wang, W. Tao, H. Mu,J.K. Burks, M. AndreeffWriting, review, and/or revision of the manuscript: B.Z. Carter, V. Ruvolo,M. AndreeffAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D. Mak, H. Mu, M. AndreeffStudy supervision: B.Z. Carter, M. Andreeff

AcknowledgmentsWe thankDrs. NumsenHail for editorial support and IvoVeletic for technical

assistance. This workwas supported in part by grants from theUniversity CancerFoundationvia the InstitutionalResearchGrantprogramatMDAnderson toB.Z.Carter and from theNIH (P01CA055164), Cancer Prevention Research Instituteof Texas (CPRIT,RP121010), andby thePaul andMaryHaasChair inGenetics toM. Andreeff andMDAnderson's Cancer Center Support Grant CA016672 (FlowCytometry and Cellular Image Facility and Characterized Cell Line core).

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

Received March 26, 2018; revised September 17, 2018; accepted January 16,2019; published first January 23, 2019.

Carter et al.





References1. Carter BZ, Qiu YH, Zhang N, Coombes KR, Mak DH, Thomas DA, et al.

Expression of ARC (apoptosis repressor with caspase recruitment domain),an antiapoptotic protein, is strongly prognostic in AML. Blood 2011;117:780–7.

2. Mak PY, Mak DH, Mu H, Shi Y, Ruvolo P, Ruvolo V, et al. Apoptosisrepressor with caspase recruitment domain is regulated byMAPK/PI3K andconfers drug resistance and survival advantage to AML. Apoptosis 2014;19:698–707.

3. Carter BZ, Mak PY, Chen Y, Mak DH,MuH, Jacamo R, et al. Anti-apoptoticARC protein confers chemoresistance by controlling leukemia-microenvi-ronment interactions through a NFkappaB/IL1beta signaling network.Oncotarget 2016;7:20054–67.

4. Morrison SJ, Spradling AC. Stem cells and niches: mechanisms thatpromote stem cell maintenance throughout life. Cell 2008;132:598–611.

5. Konopleva M, Tabe Y, Zeng Z, Andreeff M. Therapeutic targeting ofmicroenvironmental interactions in leukemia: mechanisms andapproaches. Drug Resist Updat 2009;12:103–13.

6. Turzanski J, Grundy M, Russell NH, Pallis M. Interleukin-1beta maintainsan apoptosis-resistant phenotype in the blast cells of acute myeloidleukaemia via multiple pathways. Leukemia 2004;18:1662–70.

7. Barreyro L, Will B, Bartholdy B, Zhou L, Todorova TI, Stanley RF, et al.Overexpression of IL-1 receptor accessory protein in stem and progen-itor cells and outcome correlation in AML and MDS. Blood 2012;120:1290–8.

8. AskmyrM, AgerstamH,HansenN,Gordon S, Arvanitakis A, RisslerM, et al.Selective killing of candidate AML stem cells by antibody targeting ofIL1RAP. Blood 2013;121:3709–13.

9. Li HJ, Reinhardt F, Herschman HR, Weinberg RA. Cancer-stimulatedmesenchymal stem cells create a carcinoma stem cell niche via prostaglan-din E2 signaling. Cancer Discov 2012;2:840–55.

10. Ackerman WEt, Summerfield TL, Vandre DD, Robinson JM, Kniss DA.Nuclear factor-kappa B regulates inducible prostaglandin E synthaseexpression in human amnion mesenchymal cells. Biol Reprod 2008;78:68–76.

11. Kaltschmidt B, Linker RA, Deng J, Kaltschmidt C. Cyclooxygenase-2 is aneuronal target gene of NF-kappaB. BMC Mol Biol 2002;3:16.

12. Yamamoto K, Arakawa T, Ueda N, Yamamoto S. Transcriptional roles ofnuclear factor kappa B and nuclear factor-interleukin-6 in the tumornecrosis factor alpha-dependent induction of cyclooxygenase-2 inMC3T3-E1 cells. J Biol Chem 1995;270:31315–20.

13. Kalinski P. Regulation of immune responses by prostaglandin E2.J Immunol 2012;188:21–8.

14. Pelus LM, Hoggatt J. Pleiotropic effects of prostaglandin E2 in hemato-poiesis; prostaglandin E2 and other eicosanoids regulate hematopoieticstem and progenitor cell function. Prostaglandins Other Lipid Mediat2011;96:3–9.

15. Broxmeyer HE, Pelus LM. Inhibition of DPP4/CD26 and dmPGE(2)treatment enhances engraftment of mouse bone marrow hematopoieticstem cells. Blood Cells Mol Dis 2014;53:34–8.

16. Speth JM, Hoggatt J, Singh P, Pelus LM. Pharmacologic increase in HIF1al-pha enhances hematopoietic stem and progenitor homing and engraft-ment. Blood 2014;123:203–7.

17. Nakanishi M, Rosenberg DW. Multifaceted roles of PGE2 in inflammationand cancer. Semin Immunopathol 2013;35:123–37.

18. MurakamiM,NarabaH, Tanioka T, SemmyoN,Nakatani Y, Kojima F, et al.Regulation of prostaglandin E2 biosynthesis by inducible membrane-associated prostaglandin E2 synthase that acts in concert with cyclooxy-genase-2. J Biol Chem 2000;275:32783–92.

19. Li Y, Yin S, Nie D, Xie S, Ma L, Wang X, et al. MK886 inhibits theproliferation of HL-60 leukemia cells by suppressing the expression ofmPGES-1 and reducing prostaglandin E2 synthesis. Int JHematol 2011;94:472–8.

20. Naderi EH, Skah S, Ugland H, Myklebost O, Sandnes DL, Torgersen ML,et al. Bonemarrow stroma-derived PGE2 protects BCP-ALL cells fromDNA

damage-induced p53 accumulation and cell death. Mol Cancer 2015;14:14.

21. SimonM,Grandage VL, LinchDC, Khwaja A. Constitutive activation of theWnt/beta-catenin signalling pathway in acute myeloid leukaemia. Onco-gene 2005;24:2410–20.

22. Wang Y, Krivtsov AV, Sinha AU, North TE, Goessling W, Feng Z, et al. TheWnt/beta-catenin pathway is required for the development of leukemiastem cells in AML. Science 2010;327:1650–3.

23. Ysebaert L, Chicanne G, Demur C, De Toni F, Prade-Houdellier N, Rui-davets JB, et al. Expression of beta-catenin by acute myeloid leukemia cellspredicts enhanced clonogenic capacities and poor prognosis. Leukemia2006;20:1211–6.

24. Romano P, Manniello A, Aresu O, Armento M, Cesaro M, Parodi B. CellLine Data Base: structure and recent improvements towards molecularauthentication of human cell lines. Nucleic Acids Res 2009;37:D925–32.

25. Studeny M, Marini FC, Champlin RE, Zompetta C, Fidler IJ, Andreeff M.Bone marrow-derived mesenchymal stem cells as vehicles for interferon-beta delivery into tumors. Cancer Res 2002;62:3603–8.

26. Mak PY, Mak DH, Ruvolo V, Jacamo R, Kornblau SM, Kantarjian H, et al.Apoptosis repressor with caspase recruitment domain modulates secondmitochondrial-derived activator of caspases mimetic-induced cell deaththrough BIRC2/MAP3K14 signalling in acute myeloid leukaemia. Br JHaematol 2014;167:376–84.

27. Jiang X,Mak PY,MuH, TaoW,MakD, Kornblau S, et al.Disruption ofWnt/beta-catenin exerts anti-leukemia activity and synergizes with FLT3 inhi-bition in FLT3-mutant acute myeloid leukemia. Clin Cancer Res 2018;24:2417–29.

28. Zhou H, Mak PY, Mu H, Mak DH, Zeng Z, Cortes J, et al. Combinedinhibition of beta-catenin and Bcr-Abl synergistically targets tyrosinekinase inhibitor-resistant blast crisis chronic myeloid leukemia blasts andprogenitors in vitro and in vivo. Leukemia 2017;31:2065–74.

29. Carter BZ, Mak PY, Mu H, Zhou H, Mak DH, Schober W, et al. Combinedtargetingof BCL-2 andBCR-ABL tyrosine kinase eradicates chronicmyeloidleukemia stem cells. Sci Transl Med 2016;8:355ra117.

30. Zhang B, Li M, McDonald T, Holyoake TL, Moon RT, Campana D, et al.Microenvironmental protection of CML stem and progenitor cells fromtyrosine kinase inhibitors through N-cadherin and Wnt-beta-catenin sig-naling. Blood 2013;121:1824–38.

31. Han L,Qiu P, Zeng Z, Jorgensen JL,MakDH, Burks JK, et al. Single-cellmasscytometry reveals intracellular survival/proliferative signaling in FLT3-ITD-mutated AML stem/progenitor cells. Cytometry A 2015;87:346–56.

32. Chen H, Lau MC, Wong MT, Newell EW, Poidinger M, Chen J. Cytofkit: abioconductor package for an integrated mass cytometry data analysispipeline. PLoS Comput Biol 2016;12:e1005112.

33. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: thecombined effects of multiple drugs or enzyme inhibitors. Adv EnzymeRegul 1984;22:27–55.

34. Cadigan KM, Waterman ML. TCF/LEFs and Wnt signaling in the nucleus.Cold Spring Harb Perspect Biol 2012;4.

35. Baratelli F, Krysan K, Heuze-Vourc'h N, Zhu L, Escuadro B, Sharma S, et al.PGE2 confers survivin-dependent apoptosis resistance in human mono-cyte-derived dendritic cells. J Leukoc Biol 2005;78:555–64.

36. Fernandez-Martinez A, Molla B, Mayoral R, Bosca L, Casado M, Martin-Sanz P. Cyclo-oxygenase 2 expression impairs serum-withdrawal-inducedapoptosis in liver cells. Biochem J 2006;398:371–80.

37. George RJ, Sturmoski MA, Anant S, Houchen CW. EP4 mediates PGE2dependent cell survival through the PI3 kinase/AKT pathway.Prostaglandins Other Lipid Mediat 2007;83:112–20.

38. AndersenMH. The targeting of immunosuppressive mechanisms in hema-tological malignancies. Leukemia 2014;28:1784–92.

39. Prima V, Kaliberova LN, Kaliberov S, Curiel DT, Kusmartsev S. COX2/mPGES1/PGE2 pathway regulates PD-L1 expression in tumor-associatedmacrophages andmyeloid-derived suppressor cells. Proc Natl Acad Sci U SA 2017;114:1117–22.






2019;79:1165-1177. Published OnlineFirst January 23, 2019.Cancer Res Bing Z. Carter, Po Yee Mak, Xiangmeng Wang, et al. Drug Resistance in AML

Microenvironment Interactions and Confers−Controls Leukemia -Catenin/ARC Circuitβ/Cox-2/PGE2/βAn ARC-Regulated IL1

Updated version

10.1158/0008-5472.CAN-18-0921doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2019/01/23/0008-5472.CAN-18-0921.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/79/6/1165.full#ref-list-1

This article cites 38 articles, 15 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

[email protected]

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

Permissions

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

.http://cancerres.aacrjournals.org/content/79/6/1165To 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-18-0921

http://cancerres.aacrjournals.org/content/suppl/2019/01/23/0008-5472.CAN-18-0921.DC1

http://cancerres.aacrjournals.org/content/79/6/1165.full#ref-list-1

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/79/6/1165


Documents

An ARC-Regulated IL1b/Cox-2/PGE2/ b-Catenin/ARC Circuit ...#25-072-CV, Corning). Human MSCs were isolated from bone marrow samples obtained from healthy subjects as described previously