Pericytes Promote Malignant Ovarian Cancer Progression in … · Biology of Human Tumors Pericytes Promote Malignant Ovarian Cancer Progression in Mice and Predict Poor Prognosis

Biology of Human Tumors

Pericytes Promote Malignant Ovarian CancerProgression in Mice and Predict Poor Prognosis inSerous Ovarian Cancer PatientsDevbarnaSinha1,2, LynnChong1, JoshyGeorge3,HolgerSchl€uter1,4, SusannM€onchgesang1,Stuart Mills1, Jason Li2, Christopher Parish5, David Bowtell3,4,6,7, for theAustralian Ovarian Cancer Study Group, and Pritinder Kaur1,4,8

Abstract

Purpose: The aim of this study was to investigate the role ofpericytes in regulating malignant ovarian cancer progression.

Experimental Design: The pericyte mRNA signature was usedto interrogate ovarian cancer patient datasets to determine itsprognostic value for recurrence and mortality. Xenograft modelsof ovarian cancer were used to determine if co-injection withpericytes affected tumor growth rate and metastasis, whereas co-culture models were utilized to investigate the direct effect ofpericytes on ovarian cancer cells. Pericyte markers were used tostain patient tissue samples to ascertain their use in prognosis.

Results: Interrogation of two serous ovarian cancer patientdatasets [the Australian Ovarian Cancer Study, n ¼ 215; and theNCI TCGA (The Cancer Genome Atlas), n ¼ 408] showed that ahigh pericyte score is highly predictive for poor patient prognosis.Co-injection of ovarian cancer (OVCAR-5 & -8) cells with

pericytes in a xenograft model resulted in accelerated ovariantumor growth, and aggressive metastases, without altering tumorvasculature. Pericyte co-culture in vitro promoted ovarian cancercell proliferation and invasion. High aSMA protein levels inpatient tissue microarrays were correlated with more aggressivedisease and earlier recurrence.

Conclusions: High pericyte score provides the best means todate of identifying patients with ovarian cancer at high risk ofrapid relapse and mortality (mean progression-free survival time<9months). The stroma contains rare yet extremely potent locallyresident mesenchymal stem cells—a subset of "cancer-associatedfibroblasts" that promote aggressive tumor growth andmetastaticdissemination, underlying the prognostic capacity of a highpericyte score to strongly predict earlier relapse and mortality.Clin Cancer Res; 22(7); 1813–24. �2015 AACR.

IntroductionOvarian cancer is the most life-threatening gynecologic cancer,

with very high rates of recurrence and mortality following diag-nosis. Statistics from the United States (NCI) and United King-dom(CRUK)—countrieswith someof the highest rates of ovariancancer—show that although diagnosis at stage 1 at a young age isassociated with approximately 90% survival, the overwhelmingmajority of patients with ovarian cancer (�85%) are diagnosed atadvanced stages of disease (3 and 4), after metastatic dissemina-

tion (1). Despite aggressive surgical intervention combined withchemotherapeutic and platinum/paclitaxel treatment to elimi-nate residual cancer, 60% to 70% of late-stage ovarian cancerpatients relapsewith recurrence, dyingwithin2 years of treatment.The predominant diagnosis of ovarian cancer at advanced stages isattributed largely to asymptomatic spread of disease in the peri-toneal cavity combined with symptoms confused with otherinnocuous gastrointestinal effects (e.g., bloating, discomfort,indigestion, and pelvic pain). Thus, an increased understandingof the critical biological changes underlying the development andprogression of ovarian cancer, aside fromnotable genetic changes,is critical to the success of earlier diagnosis and design of noveltherapeutic interventions to reduce highmortality rates associatedwith ovarian cancer.

The tumormicroenvironment (TME) or cancer-associated stro-ma, including cancer-associated fibroblasts (CAF), bonemarrow–derived mesenchymal stem cells (BM-MSC), endothelial cells,pericytes, and immune components plus the growth factors andextracellular matrix proteins they produce, regulates tumor celldissemination and metastases. Ovarian cancer is classified intodistinct histopathological subtypes, that is, high-grade and low-grade serous, endometroid, clear cell, and mucinous and trans-formed cells with low malignant potential (LMP) reminiscent ofthe anatomyof origin (2).High-grade serous ovarian cancer, oftendiagnosed after metastatic spread into the abdominal cavity andomentum (1), is distinguished by a highly mitotic, stratifiedepithelium, with reactive stroma (2–4) making it an ideal modelto study epithelial–stromal interactions in the TME.

1Epithelial Stem Cell Biology Laboratory, Peter MacCallum CancerCentre, Melbourne, Victoria, Australia. 2Bioinformatics Core Facility,Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia.3Cancer Genetics & Genomics Laboratory, Peter MacCallum CancerCentre, Melbourne, Victoria, Australia. 4Sir Peter MacCallum Depart-ment of Oncology, University of Melbourne, Parkville, Victoria,Australia. 5The John Curtin School of Medical Research, AustralianNational University, Canberra, Australia. 6Department of Pathology,University of Melbourne, Parkville,Victoria, Australia. 7Department ofBiochemistry and Molecular Biology, University of Melbourne, Park-ville, Victoria, Australia. 8Department of Anatomy and Neuroscience,University of Melbourne, Parkville, Victoria, Australia.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Pritinder Kaur, Hudson Institute of Medical Research,27-31 Wright Street, Clayton, Melbourne, Victoria 3168, Australia. Phone: 61-3-9594-4398; Fax: 61-3-9594-7114; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-15-1931

�2015 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 1813

on June 18, 2020. © 2016 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 20, 2015; DOI: 10.1158/1078-0432.CCR-15-1931

http://clincancerres.aacrjournals.org/

Stromal signatures strongly predict relapse in ovarian, colorec-tal, pancreatic, and breast cancers (5–9), contributing to theprevailing view that the TME is pro-tumorigenic. Although suf-ficient evidence exists to support the notion that recruitment of acooperative stroma is essential for malignant progression, it isequally plausible that the influx of some stromal elements is thebody's attempt to limit cancer spread through fibroblastic encap-sulation typical of many tumor types. Notably, the TME has beenimplicated in contributing at least partially to resistance againstcancer therapeutic reagents (10), but also in enhancing therapeu-tic efficacy depending on context [see ref. (11) for review].

It is, therefore, vital that the functional heterogeneity anddiverse origins of the tumor stroma are more fully mapped. It iswell known that CAFs isolated from cancers promote tumorgrowth, invasiveness, and metastasis of many cancers, notablybreast, prostate, and pancreatic carcinomas (12–14) comparedwith normal fibroblasts. However, the postulated origins of CAFsinclude many normal stromal cells, including tissue residentmyofibroblasts, activated adipocytes, and BM-MSCs (13, 15)—the latter representing perhaps the best defined source of CAFs.However, BM-MSCs that are home todeveloping tumors inducingincreased metastases comprise approximately 20% of CAFs(16–20), leaving some 50% to 80% of CAFs that are not BM-derived and may arise from locally resident fibroblasts andpresumably MSC-like populations such as pericytes that formthe focus of this study.

Pericytes are best known for regulating endothelial cell prolif-eration, differentiation, and microvascular perfusion/permeabil-ity through paracrine regulators such as TGF-b and vasoactiveagents (21, 22), and are identified as a-smooth muscle actin(a-SMA)–positive, contractile cells located abluminally in micro-vessels. It is increasingly evident that pericytes andBM-MSCs sharemany phenotypic and functional attributes, including multiline-age differentiation capacity, andmay have a pro-proliferative rolein organ growth, repair, and regeneration (23–25). Notably, wehave previously demonstrated that pericytes promote normalepithelial cell proliferation and regeneration in the absence ofangiogenesis, most likely through the secretion of the LAMA5isoform of laminin (26). In the context of cancer, targeting both

endothelial cells and pericytes through kinase inhibitors of VEGF,which promotes blood vessel growth, and PDGF-B, which pro-motes proliferation and survival of pericytes (27), improved theefficacy of anti-cancer therapies in animalmodels (28–31), attrib-uted to destabilizing microvascular structure (27). Althoughsubsequent studies claimed to show unaffected tumor growthfollowing pericyte ablation, complete pericyte knockdown wasnot achieved by treatment with AX102—an inhibitor of PDGF-Bsignaling (32), or in PDGFBret/retmice harboring amutation in thePDGF-B retention motif (33) with only partial decrease in tumorvasculature and pericyte number. A maximal 50% reduction inpericyte number is reached in PDGFBret/ret mice (34). In contrast,comprehensive knockdown of pericytes via NG2-promoter–driv-en thymidine kinase results in tumor hypoxia, leading to epithe-lial–mesenchymal transition (EMT) and increased metastaticlung dissemination inmousemodels of breast cancer,melanoma,and renal cell carcinoma (35). In fact, retaining pericytes withintumor blood vessel walls may limit metastatic spread throughleaky blood vessels (36).

Thus, arguments to support both tumor growth–limiting andmetastasis-promoting roles have been made for pericytes andmost likely depend on the nature of the experimentalmodel used.In this study, we show that the molecular signature of pericytes ishighly predictive for patient relapse and mortality in high-gradeserous ovarian cancer patients, demonstrating our ability toidentify with a very high degree of certainty those patients whodie in less than 9 months, despite aggressive treatment at diag-nosis, suggesting a potent pro-tumorigenic/pro-metastatic role forpericytes in ovarian cancer progression. Consistent with thisclinical correlate, we experimentally demonstrate that MSC-likepericytes are potent stimulators of both poorly and highly tumor-igenic ovarian cancer cell lines when introduced into the TME in axenograft model, accelerating tumor growth rates and earliermetastasis in aggressive ovarian cancer cells, but also inducingmetastasis in nonmetastatic cell lines without affecting tumorvasculature. These influences in malignant ovarian cancer cellproliferation, migration, and invasion were also demonstrable invitro using co-culture models, further indicating a novel and as yetunappreciated function for pericytes in malignant progression.

Materials and MethodsIn silico analyses

Bioinformatics analyses for predicting prognosis based on thepericyte signature and gene set enrichment are described in theSupplementary data.

Cell cultureCD45�VLA-1bri pericytes and CD45�VLA-1dim fibroblasts were

isolated from human neonatal foreskin, as described (26) andexpanded in culture up to passage 4 (p4) (24). Fibroblasts weremaintained in DMEM with 10% FCS, and pericytes in EGM–2(Lonza; # CC-4147);OVCAR-5 andOVCAR-8 cells obtained fromNCI were authenticated using short tandem repeat markersto confirm cell identity against the Genome Project Database(Wellcome Trust Sanger Institute) and were maintained in RPMI1640 (Invitrogen; #11875) with 10% FCS, 25 mmol/L HEPESbuffer, 1% penicillin-streptomycin, and 1.5% Diflucan, andHEK293T cells in DMEM with 10% FCS, 1% L-glutamate (2mmol/L), and 1% sodium pyruvate (1 mmol/L) without anti-biotics for lentiviral production. OVCAR-5 cells and pericytes

Translational Relevance

There is an urgent clinical need to identify those serousovarian cancer patients at greater risk of earlier recurrence,relapse, and mortality at diagnosis despite similar treatment.Although the predictive capacity of tumor-associated stromagene expression is well recognized in these patients, we reportthat this can be highly refined into those that succumb torelapse and mortality in less than 9 months versus those thatrespond better to current treatment modalities (surviving forup to 29 months) on the basis of stromal heterogeneity—specifically by using the molecular signature of human stemcell–like pericytes rather than fibroblasts. Although ovariancancer tends to be diagnosed at advanced stages and thuscarries inherently poor prognosis, it would be of great benefitto select those patients at significantly higher risk for recur-rence andmortality formore aggressive or alternate treatmentsto increase their chances of survival.

Sinha et al.

Clin Cancer Res; 22(7) April 1, 2016 Clinical Cancer Research1814




were transduced with GFP-luciferase, as described in the Supple-mentary data. All human tissue experimentation was approvedby the Peter MacCallum Human Research Ethics Committee(#03/44).

AnimalsSix to eight week old female nude Balb/c mice (WEHI), housed

in a pathogen-free 12-hour light–dark environment, fed ad libitumwere used for tumorigenicity assays. All experimentation wasapproved by the Peter MacCallum Animal Research EthicsCommittee (# E394).

Tumorigenicity assayOVCAR-5 cells (8� 106)� 10% fibroblasts or pericytes in 100

mL of 1:1 sterile PBS and standard Matrigel (BD Biosciences) wereinjected subcutaneously (s.c.) into the flanks of mice. Five micewere injected/group; all experiments performed in triplicate.

Immunostaining and morphometric analyses, invasion assaysAll procedures and antibodies used are described in the

Supplementary data.

Luciferase imagingMetastatic spread and pericyte survival in vivo were monitored

using the Xenogen Real-Time Imaging System. D-luciferin (GoldBiotechnology Inc.), a substrate for the luciferase enzyme, wasinjected s.c. at 150mg/g bodyweight in PBS.Micewere allowed freemovement for 6 to 8minutes, anaesthetized with isofluorane, andimaged within 10 to12 minutes of luciferin injection. Biolumi-nescent imaging (BLI)was repeated every 7 days to trackmetastaticspreaduntil an experimental or ethical endpointwas reached, fromd14 when primary tumors were palpable. At the endpoint, as theluminescent signal from primary tumors was saturated potentiallymasking signal from smaller metastatic nodules, primary tumorswere carefully excised after sacrifice, and the peritoneum openedsurgically and imaged with increased exposure times to improvevisualization of metastases. Metastatic burden was quantified andorgans harvested for histology/GFP staining.

Vascular permeability assayA volume of 100 mL of 10 mg/mL FITC-dextran (2,000,000

MW,Sigma)was injected into the tail vein anhour before sacrifice.Tumors were collected, snap-frozen in liquid nitrogen, and cryo-sections co-stained with the endothelial marker CD31 for fluo-rescence microscopy analysis.

Tissue microarraysPatient tissue microarrays (TMA) consisting of 4-mm cores of

formalin-fixed, paraffin-embedded, high-grade serous ovariancancer biopsy tissues were obtained from the Australian OvarianCancer Study (AOCS) approved by the AOCS review board.

In-vitro proliferation assaysCo-culture proliferation assays were performed in a 6-well

format. GFPþOVCAR-5 (2 � 104) cells alone or with 2 � 104

p4 pericytes or p4 fibroblasts were mixed and seeded in either 1%or 10% FBS epidermalization medium. Plates were incubated at37�C, in 5% CO2 for 24, 48, and 72 hours, fixed in 4% parafor-maldehyde (w/v), and immunostained for GFP to determine thenumber of OVCAR-5 cells over time.

Statistical analysisData analyses were performed using the Prism 6.0 (Graphpad

software) or R software. Pooled data were represented as mean�SD, unless otherwise indicated.

ResultsThe transcriptional profile of pericytes predicts significantlyearlier relapse and mortality in high-grade serous ovariancancer patients

TheAOCSGroup showed that high-grade serous ovarian cancerpatients with a stromal signature had a poor clinical outcome (5),as reported for breast cancer patients (6). Previously we demon-strated that MSC-like pericytes had high potency in increasingepithelial proliferative capacity compared with fibroblasts (26),so we used the molecular signature of these two distinct stromalcell types (26), to compare their predictive capacity for clinicaloutcome in the AOCSovarian cancer patient dataset in silico, usingthe AOCS ovarian cancer stromal signature as a reference (5).Notably, the pericyte-specific signature or high pericyte score wasa potent predictor of rapid relapse and mortality (P ¼ 0.00067;Kaplan–Meier plots Fig. 1A), identifying those patients with amean progression-free survival (PFS) time of 9 months or lessversus those with a low pericyte score (mean PFS time of 29months) despite similar treatment, as compared with the AOCSovarian cancer stromal signature (P ¼ 0.0011; ref. 5) and thenormal fibroblast signature (P ¼ 0.01).

Analysis of genes co-expressed by laser-capturemicro-dissectedovarian CAFs in the AOCS study and normal pericytes revealed146 genes, including well-known pericyte markers (PDGFRb,ACTA2, RGS5, CALD1, MCAM, and ANGPT1) (26), growthfactors, adhesion ligands and receptors (FGF/FGFRs, TenascinC, LAMA3, LAMA5, CSPG-4, the VLA-1, VLA-3 and VLA-7 integ-rins), BMPs, and Notch pathway signaling genes (SupplementaryTable S1), linked to CAFs and ovarian cancer progression (4).Importantly, minimal overlap was detected with the Gene Ontol-ogy classification angiogenesis signature (GO Angiogenesis—GO:0001525) with our pericyte signature, given that pericytesstabilize tumor vasculature, with only two common genes (angio-poietin 1 and 2). Moreover, no overlap was present between therecently described angiogenic signatures (37), suggesting that thesignificantly earlier relapse observed in patients with a highpericyte score was unrelated to angiogenesis. An interrogation ofthe NCI TCGA (The Cancer Genome Atlas) ovarian cancer patientdataset further confirmed the ability of the pericyte signatureto predict decreased survival in a group of 408 patients(Fig. 1B; P ¼ 0.008), leading us to investigate whether pericytescould promote ovarian tumor growth experimentally withoutaffecting angiogenesis.

Pericytes accelerate ovarian tumor growth in vivoOVCAR-5 cells derived from a serous ovarian cancer patient

with metastatic disease before treatment with anti-cancer agents(38) were used as an ovarian cancer model. GFP-luciferaseþ

OVCAR-5 cells resuspended in Matrigel were injected s.c.in nude mice, either alone (OVCAR-5) or with pericytes(OVCAR-5þP) or fibroblasts (OVCAR-5þF), at a 10:1 tumor:stromal cell ratio. Pericyte co-injection consistently led toaccelerated tumor growth compared with OVCAR-5 orOVCAR-5þF (Fig. 1C; P < 0.0001; n ¼ 5 independent experi-ments), and increased endpoint tumor volumes (and mass) at

Pericytes Promote Metastasis and Predict Relapse

www.aacrjournals.org Clin Cancer Res; 22(7) April 1, 2016 1815




day 35 (Fig. 1D; P < 0.0001); OVCAR-5þP tumors reached200 mm3 4 to 5 days earlier. Furthermore, a dose-dependenteffect was observed when the proportion of pericytes wasincreased from 10% to 50%, keeping the number of OVCAR-5 cells constant, with greater endpoint tumor volumes thanOVCAR-5 controls in both the 10:1 and 1:1 pericyte co-injectedgroups (P < 0.05 and P < 0.01, respectively; Fig. 1E).

Dual staining for epithelial-specific EpCam and Ki67 (withanti-human specific antibodies) revealed a direct increase inOVCAR-5 cell proliferation, i.e., number of EpCamþ/Ki67þ cellscomparedwithOVCAR-5–only controls andOVCAR-5þF tumors(P < 0.0001 and P < 0.01 respectively; Fig. 1F). No difference inapoptotic index was observed between experimental groups atday 11 or day 35 by staining for cleaved caspase-3 (CC3; Sup-plementary Fig. S1A–S1H), excluding decreased apoptosis inincreasing tumor size. Moreover, in vitro experiments showedthat pericyte co-cultured GFPþOVCAR-5 cells displayed increasedproliferation compared with controls within 72 hours in both 1%

(P¼ 0.0349) and 10% serum (P¼ 0.0328), not seen in fibroblastco-cultures (Supplementary Fig. S2A–S2C).

Injected pericytes persist but do not proliferate or contribute toangiogenesis in OVCAR-5 tumors

BLI of xenografts generated with unlabeled OVCAR-5 cellsand GFP-luciferaseþ–tagged pericytes permitted pericyte-trackingin developing tumors. Whilst control animals (OVCAR-5 cellsalone) gave no signal despite luciferin injection (Fig. 2A),pericytes persisted within co-injected tumors at all time pointsanalyzed (Fig. 2B). Histological analyses revealed single GFPþ

pericytes in OVCAR-5þP tumors, declining in numbers over time(Fig. 2C). Notably, GFPþ pericytes were Ki67-negative at all timepoints, indicating that they did not proliferate during tumorigen-esis (Fig. 2C).

We next addressed whether co-injected pericytes acceleratedtumor growth by increasing or stabilizing tumor vasculature. Thearea of CD31þ blood vessels in tumors remained unaltered both

CB

Pro

babi

lity

of s

urvi

val

50 150100

0.0

0

.2

0.4

0

.6

0.8

1

.0

2000

High pericyte score

A

Time to relapse (months)log-rank test P value: 0.00067

High-grade serous ovarian cancer samples stratified into two groups based on the expression of pericyte genes

Low pericyte score

6050403020100

0.0

0.2

0.4

1.0

0.8

0.6

D E F


TCGAAOCS

Pro

babi

lity

of s

urvi

val High pericyte score

Low pericyte score

00 5 10 15 20

Days25 30 35 40

500

1,000

Da

y 3

5 t

um

or

vo

lum

e (

mm

3)

Tu

mo

r v

olu

me

(m

m3)

Da

y 3

5 t

um

or

vo

lum

e (

mm

3)

1,500OVCAR-5OVCAR-5+FOVCAR-5+P

OVCAR-5

OVC

AR-5

OVC

AR-5

OVC

AR-5

+FO

VCAR-5

+P

OVC

AR-5

+P (1

0:1)

OVC

AR-5

+P (1

:1)

00

500

1,000

1,500

0

500

1,000

1,500

ns

ns

50

100

150

200

#E

pc

am

+K

i67+

ce

lls

pe

r fi

eld

(da

y 3

5)

OVCAR-5+F OVCAR-5+P

Figure 1.The pericyte-specific gene signature predicts poor prognosis in patients with ovarian cancer, and co-injection of OVCAR-5 ovarian cancer cells with pericytesaccelerates tumor cell proliferation and tumor volume. Kaplan–Meier curves showing a significantly poorer PFS rate among 215 high-grade serous ovarian cancerpatients with a high score of pericyte-specific genes in the AOCS dataset (A) and the NCI TCGA patient dataset (B). C, nonlinear regression fit of tumor volumesagainst time generated from the injection of 8� 106 OVCAR-5 cells alone or co-injected at a ratio of 10:1 with pericytes (OVCAR-5þP) or fibroblasts (OVCAR-5þF).Data represented as mean tumor volume � SD of 26 mice per group from 5 independent experiments. Repeated measure data for each time point werecompared using two-way ANOVA. D, quantification of endpoint tumor volumes at day 35 represented as mean tumor volume � SEM of 5 independentexperiments, calculated from data shown in A. E, quantification of endpoint tumor volume at day 35, demonstrating the dose effect of increasing thenumber of pericytes on OVCAR-5 tumor growth, i.e., injection of 5 � 106 OVCAR-5 cells with or without pericytes at a ratio of 10:1 and 1:1. Data are shownas mean tumor volume � SEM of 10 mice per group from 2 independent experiments. Statistical analysis in E and F performed using one-way ANOVA.F, quantification of dual immunofluorescent staining with anti-human specific antibodies to the proliferation marker Ki67 and the epithelial marker EpCamof ovarian tumors generated by OVCAR-5 cells alone or co-injected with pericytes—OVCAR-5þP and fibroblasts, OVCAR-5þF showing a significant increasein the number of Ki67þ/EpCamþ tumor cells in the OVCAR-5þP group. Data are shown as mean� SEM from 3 tumors per group from 2 independent experiments.Statistical analyses performed by one-way ANOVA. � , P < 0.05; �� , P < 0.01; �� , P < 0.0001, ns, not significant.

Sinha et al.





in the center and in the edges of the OVCAR-5þP tumors com-paredwith controls (Fig. 3A andB) at day 11 (Fig. 3C;P¼0.1879),confirmed further by measuring microvessel density (MVD)(Fig. 3D; P ¼ 0.8910). The CD31þ blood vessel area (Fig. 3E;P ¼ 0.2021) and MVD remained unaffected at day 35 (Fig. 3F;P ¼ 0.7790).

Furthermore, we could not find any differences in the aSMAþ

pericyte coverage index (MPI) of CD34þ microvessels betweenOVCAR-5þP and OVCAR-5 controls (Fig. 3G: P ¼ 0.5321),indicating that pericyte inclusion did not alter the structuralstability of OVCAR-5 tumor vasculature.

Similar analyses of microvessels in clinical samples, i.e., TMAsof serous ovarian cancer patients, demonstrated that CD34þ

expression in tumor microvessels did not correlate with time torelapse or survival (Fig. 3H and I), providing independent veri-fication that poor prognosis predicted by the pericyte signaturehad minimal overlap with the angiogenic signature.

Finally, vascular permeability determined by injecting FITC-conjugated dextran into tumor-bearing mice an hour before

sacrifice followed by analysis of tumor cryosections co-stainedfor CD31 and FITC-dextran showed no differences betweencontrol and pericyte co-injected OVCAR-5 tumors with minimalFITC-dextran signal outside the vessels at day 11 and day 35 (Fig.3J). These data were consistent with the observation that GFP-tagged pericytes did not associate with CD34þ blood vessels (Fig.2D) in OVCAR-5 xenografts, but were "stroma-associated." Thesedata strongly suggest that the tumor-promoting actionof pericytesis not mediated by affecting tumor angiogenesis directly orindirectly.

Pericytes promote aggressive invasion in OVCAR-5 cells in vitroand in vivo

At harvest pericyte co-injected xenografts appearedmacroscop-ically different with indistinct tumor margins indicative ofoutgrowths. Histological analysis confirmed the presence ofinvasive nodules of cells at the tumor edges as early as day 6in the OVCAR-5þP group compared with controls (Fig. 4A)

Figure 2.Co-injected human pericytes survive but do not proliferate in ovarian tumors in vivo. A and B, representative BLI images of mice with unlabeled OVCAR-5 tumorsor OVCAR-5þGFP-luciferase tagged pericytes imaged at days 6–35. BLI imaging conducted in 3 mice/group/time point in 2 replicate experiments. C, dualimmunofluorescent staining for GFPþ pericytes (green) and Ki67þ proliferating cells (red) in days 6–35 pericyte co-injected tumors, showing decline inpericyte numbers over time and absence of Ki67þGFPþ pericytes. Images are representative of three random fields from 3 tumors per experimental groupfrom 2 independent experiments. D, dual immunofluorescent staining for CD34 and GFPþ pericytes, illustrating that injected pericytes do not incorporate intohost blood vessels. (P ¼ pericytes; BV ¼ blood vessels). Immunostaining is representative of multiple sections per mouse. Scale bar ¼ 25 mm.






GFP-immunostaining showed clear encapsulation with GFP�

stromal cells in control tumors, while invasive nodules ofGFPþOVCAR-5 cells were present at the tumor margins ofOVCAR-5þP tumors (Fig. 4B).

Moreover, in vitro Boyden chambermigration assays confirmedthat co-culture of OVCAR-5 cells with pericytes increased bothmigration (2–3� fold; P < 0.05; data not shown) and invasionthroughMatrigel and an8-mmfiltermembrane (Fig. 4C;P<0.05),while fibroblasts had no significant effect (Fig. 4C).

Pericytes promote aggressive ovarian cancer metastases todistant organs in OVCAR-5 and OVCAR-8 cells

These data led us to examine whether pericytes facilitatedmetastases in xenografts—BLI analysis of GFP-lucifera-seþOVCAR-5 tumors in vivo tracked at regular intervals revealedmetastatic spread of ovarian cancer cells to the peritoneal cavity asearly as day 21 in OVCAR-5þP mice (Fig. 4D). By day 28,metastases associated with the intestine, liver, and lung weredetected in these mice (Fig. 4E), whereas control mice

(OVCAR-5 and OVCAR-5þF injected) were completely free ofmetastases (Fig. 4E and F). Moreover, a dose-dependent effect onmetastatic burden was demonstrable at day 28—increasing theOVCAR-5 cell:pericyte ratio from 10:1 to 1:1 resulted in increasedmetastases to distant organs (Fig. 4G, P < 0.05), achieving strongstatistical significanceoverOVCAR-5 controls (Fig. 4G;P<0.001).

By day 42, extensive local metastases were evident throughoutthe peritoneal cavity associated with the upper and lower gastro-intestinal tracts in both control and OVCAR-5þP groups macro-scopically (Supplementary Fig. S3A–S3D), and on GFPþ stainingof tissue sections (Supplementary Fig. S3J), confirming that theywere derived fromGFPþOVCAR-5 cells. However, more extensivemetastases were evident in distant organs such as the liver, spleen,kidney, and lungmacroscopically (Supplementary Fig. S3A–S3D;day 42) and by GFP-immunostaining (Supplementary Fig. S3E–S3I; day 35), in the pericyte–co-injected group only.

We next tested whether pericytes could affect the less aggressiveOVCAR-8 cell line derived from an early-stage cisplatin-treatedpatient reported to form noninvasive tumors with long periods of

Figure 3.Tumor-promoting activity of pericytes is independent of angiogenesis. A and B, CD31 staining of OVCAR-5 and OVCAR-5þP tumors shows no difference at thetumor center or tumor edge at day 35. Scale bar ¼ 100 mm. C and E, quantification of combined tumor center and edge CD31 staining at day 11 or day 35; andD and F, MVD per field at day 11 and day 35. G, quantification of MPI in day 11 tumors. Data shown as mean � SEM from 3 random fields from 5 tumors/group,from 2 independent experiments. H, CD34 staining of representative patient TMAs from early-, late-, or no-relapse groups. Scale bar ¼ 200 mm. I, quantificationof CD34 staining from patient biopsy samples versus relapse (n ¼ 7 patients/relapse group; 3 fields/patient). Statistical differences analyzed by one-wayANOVA. J, FITC-dextran and CD31 (red) staining at day 11 and day 35 in control OVCAR-5 and OVCAR-5þP tumors. Scale bar ¼ 20 mm.

Sinha et al.





latency (39, 40). Co-injection of OVCAR-8 cells with pericytes(10:1 ratio) into nude mice resulted in a 15-day decrease inlatency of tumor formation, accelerated tumor growth (Supple-mentary Fig. S4A), and larger tumor volumes (SupplementaryFig. S4B: P < 0.0001). Notably, while GFP-luciferaseþOVCAR-8cells did not yield metastases by themselves, pericyte co-injectionled to OVCAR-8 metastasis to distal organs, i.e., liver, lung,bladder, kidney, in addition to the GI tract, peritoneum andomentum (Supplementary Fig. S4C–S4I: GFPþ immunostain-ing). These data clearly demonstrate the potent ability of pericytesto confer malignancy on nonmetastatic ovarian cancer cells.

Interestingly, bioinformatic analyses of gene expression enrich-ment in the AOCS high-grade serous ovarian cancer patientsrevealed that early-relapse patients identified by a high pericytescore displayed upregulation of molecular pathways, involvingmatrix degradation, ECM remodeling, negative regulation ofcell adhesion, invasion, and migration, compared with thosepatients with late relapse, using two independent methods(i.e., enrichment analysis of GO terms or KEGG pathways amongoverexpressed genes or using Gene Set Enrichment Analysis;Supplementary Table S2), providing a clinical correlate for ourexperimental findings.

Day

21

OVCAR-5 OVCAR-5+P

E

D

F

Day

28

- pr

imar

y tu

mor

exc

ised

bef

ore

imag

ing

OVCAR-5+FOVCAR-5

BA

C

OVCAR-5+POVCAR-5OVCAR-5+POVCAR-5

GF

P

H&

E

GOVCAR-5

0

5

Mea

n nu

mbe

r of

met

asta

stic

foci

per

mou

se (

day

28)

10

15

0

1

2

3

Fol

d in

vasi

on

4P < 0.05

OVCAR-5OVCAR-5

+P (1

0:1)

OVCAR-5+P

(1:1

)

OVCAR-5+F OVCAR-5+P

Figure 4.Pericytes promote invasion and metastasis of OVCAR-5 cells. A, H&E staining of invasive nodules in OVCAR-5þP tumor edges at day 11 compared with the smoothmargins of controls. B, GFP staining showing singleGFPþOVCAR-5 cells at the edgeofOVCAR-5þP tumors (blue arrow) comparedwith controls encapsulatedwithina GFP� stromal lining (red arrows). Scale bar ¼ 100 mm. Images in A and B are representative of 12 to 15 tumors per experimental group from 3 independentexperiments. C, quantification of Transwell invasion through Matrigel towards pericytes or fibroblasts normalized to OVCAR-5 control. Mean � SD from 3independent experiments. D, representative BLI images of nude mice carrying control GFP-luciferaseþOVCAR-5 and GFP-luciferaseþOVCAR-5þP co-injectedtumors at day 21, indicating the position of primary tumor. E, BLI at day 28 after sacrifice and surgical excision of primary tumor and exposing organs in nudemice, indicatingmetastasis on co-injection of pericytes. F, BLI at day 28 in nudemice injectedwith OVCARþF after sacrifice and removal of primary tumors revealingabsence of metastases. Images in D–F are representative of 10 mice per experimental group from 2 independent experiments. G, quantification of metastaticburden calculated as the mean number of bioluminescent metastatic foci at the same exposure time after primary tumor removal in nude mice injected withOVCAR-5 (control), 10:1, or 1:1 OVCAR-5:P cells. Data are mean � SEM of 10 mice/group from 2 independent experiments. Statistical analysis performedby one-way ANOVA. � , P < 0.05; �� , P < 0.0001.






Day

11

Day

35

Aα S

MA

OVCAR-5+POVCAR-5 B C

D E

αSM

A/C

D34

(da

y 11

)F

G

αSM

A

Late relapseEarly relapse

H

JProgression-free survival (αSMA)

Pro

babi

lity

of s

urvi

val

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

1.0

I


4020 60 800

High αSMA expressionLow αSMA expression

Pro

babi

lity

of s

urvi

val


4020 60 800

Overall survival (αSMA)High αSMA expressionLow αSMA expression

αSMA

Rat

io (

stai

ned

area

: tot

al a

rea)

0.0

0.1

0.2

0.3

Pericyte scoreP value: 0.0078

20–1–2 31

K

No relapse

Num

ber

of s

trom

a-as

soci

ated

α SM

A c

ells

/mm

+3

Num

ber

of s

trom

a-as

soci

ated

αSM

A c

ells

/mm

+3

Num

ber

of v

esse

l-ass

ocia

ted

αS

MA

cel

ls/m

m+

3

Figure 5.Pericytes increase recruitment of aSMAþ cells to ovarian tumors in mice; increase in aSMAþ cells correlates with early relapse in patients with ovarian cancer.A, aSMA staining in OVCAR-5 and OVCAR-5þP tumors at day 11 and day 35. Quantification of total aSMAþ cells at day 11 (B) and day 35 (C). Quantificationof vessel-associated (D) or stroma-associated aSMAþ cells (E) in OVCAR-5 and OVCAR-5þP tumors. Mean� SEM from 3 independent experiments. F, illustrationof vessel-associated (arrows) and stroma-associated (arrowheads) aSMAþ cells by co-staining for CD34. Scale bar ¼ 100 mm. G, immunostaining ofrepresentative AOCS patient TMAs for aSMAþ cells. Scale bar ¼ 200 mm. H, quantification of stroma-associated aSMAþ cells in early-, late-, and no-relapseAOCS patient TMAs (n ¼ 7 patients/relapse group; 3 fields/patient). I and J, Kaplan–Meier curves correlating aSMA protein expression and progression-free (I) oroverall (J) survival in AOCS serous ovarian cancer patients. K, scatter plot of correlation between expression levels of aSMA and pericyte score from 105AOCS serous ovarian cancer patients.

Sinha et al.





Pericytes increase recruitment of hostaSMAþ cells to the TMEatsites unrelated to angiogenesis in experimental tumors—also afeature of early-relapse patients with ovarian cancer

An increase in the proportion of the "stromal" compartment oftumors—cells and acellular matrix—is prognostic for poor sur-vival in patientswith advanced ovarian cancer (41).We, therefore,immunostained OVCAR-5 xenografts for the stromal cell markeraSMA (Fig. 5A), revealing higher numbers of aSMAþ cells in theOVCAR-5þP group at day 11 (P < 0.001) and day 35 (P < 0.05)comparedwith controls (Fig. 5B andC). Althoughnoquantitativedifference in the percentage of vessel-associated aSMAþ cells wasobserved (P ¼ 0.5762; Fig. 5D), a significant increase in stroma-associated aSMAþ cells in OVCAR-5þP tumors (P < 0.0001; Fig.5E) was evident, evaluated by co-staining for aSMAþ cells andCD34þ blood vessels (Fig. 5F) via immunofluorescence. Notably,the absence of detectableaSMAþ/Ki67þ cells in xenografts even atday 11 (Supplementary Fig. S5A) suggests that this was probablythe result of increased recruitment of host aSMAþ cells to xeno-grafts, not proliferation.

The accumulation ofaSMAþ stromal cells was then analyzed inAOCS-patient TMAs and correlated with clinical outcome; early-relapse ovarian cancer patients (mean PFS time ¼ 8.98 months)showed a significant increase in aSMAþ staining compared withlate-relapse patients (mean PFS time ¼ 28.45 months; Fig. 5G).Closer inspection of the aSMAþ sections for blood vessels in 7patient TMAs per early-, late-, and no-relapse group revealedthat this was attributed to increased numbers of tumor stroma–associated aSMAþ cells in early-relapse patients (Fig. 5H;P < 0.001), as observed for theOVCAR-5þP experimental tumors,with no significant differences in vessel-associated aSMAþ cells(P ¼ 0.4902; data not shown). Thus, our experimental ovariancancermodel stronglymimics the clinical situationwith commonbiological features.

Given that BM-MSCs can be recruited to the TME,we co-stainedfor aSMA and the murine BM-MSC markers CD73 or Sca-1 in allxenografts. Interestingly, only the pericyte co-injected OVCAR-5tumors contained CD73þ and Sca-1þ populations with only asmall proportion of aSMAþ cells co-expressing these markers(Supplementary Fig. S5B and S5C). Since CAF-derived CXCL12has been strongly implicated in recruiting BM-MSCs to tumorsand driving metastatic spread (42), we immunostained for thischemokine (Supplementary Fig. S5D), not detecting it in thestroma of any experimental tumors, despite abundant CXCL12expression in OVCAR-5 cells, as reported previously for otherovarian cancer cell lines (43) in all xenografts not correlated withmetastasis, suggesting a role for alternate signaling pathways ininducing metastasis, while not excluding a role for CXCL12 inpromoting ovarian cancer tumor growth by increasing angiogen-esis, as reported previously (43).

Greater aSMA levels predict earlier relapse in serous ovariancancer patients

We next sought to determine if a single pericyte marker at theprotein level could be prognostic at diagnosis. TMAs from AOCSserous ovarian cancer patients were immunostained for aSMAand PDGFRb (and CD34 control), and their expression levelsquantitated for individual patients morphometrically and corre-lated with time to relapse. The levels of CD34 or PDGFRbexpression were not predictive for early-relapse (P ¼ 0.1342,n ¼ 112 patients; and P ¼ 0.1861, n ¼ 102 patients, respectively;Supplementary Fig. S6Aand S6B); however, higher levels ofaSMA

correlated significantly with early relapse for both PFS(P ¼ 0.03691, n ¼ 105 patients; Fig. 5I) and overall survival(P ¼ 0.006652; Fig. 5J). Consistent with this, a significant corre-lation was obtained betweenaSMA expression levels and pericytescore (Fig. 5K; P ¼ 0.0078), but not CD34 (P ¼ 0.8412) orPDGFRb (P ¼ 0.3761; Supplementary Fig. S6C and S6D).

DiscussionPericytes are widely known to regulate microvascular function,

including structural stability, limiting hypoxia, and blood–brainbarrier permeability. In the context of cancers, killing pericytesdestabilizes tumor vasculature, resulting in tumor regression (27),or causes hypoxia, inducing EMT and increased metastatic dis-semination in various cancers (35). Our data demonstrate thatplacing pericytes in the tumor stroma of OVCAR-5 and -8 ovariancancer cells while leaving the tumor vasculature intact results inaccelerated tumor expansion via increased cell proliferation,shortening the latency of OVCAR-8 tumors by 15 days. Moreover,pericytes induced invasion and metastatic spread in nonmeta-static OVCAR-8 cells—a core clinical feature of aggressive serousovarian cancer (1, 44, 45), and faster, distal spread of OVCAR-5cells comparedwith controls thatmetastasized only locallywithinthe peritoneal cavity to the gastrointestinal tract. These datademonstrate that normalMSC-like pericytes placed in close prox-imity to ovarian cancer cells drive malignant conversion, whilenormal fibroblasts do not affect tumor growth or metastasis, asreported previously. Notably, this was observed despite the use ofheterologous, that is, non-ovarian stromal cells (primarily due tothe difficulties in obtaining human ovarian tissue in sufficientquantities and at regular frequencies to undertake adequateexperimentation), indicating sufficient conservation of functionexists in MSC-like pericytes, despite being tissue of origin, con-sistent with published data (24). Indeed, current transcriptionalprofiling work in our laboratory comparing adult and neonatalpericytes from male and female donors and from different ana-tomical sites reveals minimal differences in mRNA expressionprofiles. We speculate that pericytes are a more potent stromalstem-cell–like population than fibroblasts, whose involvement isa harbinger for poor clinical outcome in patients.

Consistent with this notion, we demonstrated that the pericytesignature had strong clinical relevance for high-grade serousovarian cancer patients—outperforming the stromal signaturederived from ovarian cancer patient stroma (5) in predictingsignificantly earlier patient relapse, despite similar treatment inboth the AOCS (n¼ 215) and theNCI TCGApatient datasets (n¼408). The early-relapse patient group expressed gene sets enrichedfor biological processes clearly increased experimentally by peri-cytes such as invasion and migration that are key features ofaggressive metastatic disease, that is, cell motility, negative regu-lation of cell adhesion, and EMT. In contrast, the inability ofnormal fibroblasts to promote malignant ovarian cancer progres-sion was correlated well with their signature performing relativelypoorly as a predictor of early patient relapse. These data illustratethe need to understand the nature of stromal heterogeneity inboth normal and cancerous tissues. The ability of tumor cells toattract specific subtypes of stromal cells may facilitate tumorprogression to a malignant state. Presumably, the process ofpericyte association and dissociation from blood vessels duringtissue remodeling in wound healing and cancer requires tightmolecular regulation. The contribution of pericytes to malignant






progression has remained unappreciated, masked by the fact thatthe markers used to detect "CAFs" or BM-MSCs are also co-expressed by pericytes, e.g., aSMA, MCAM/CD146, and CD73.A further potentially confounding factor is the low incidence atwhich these cells may exist in the TME—like most stem cellpopulations, a large number is not required to effect significantchange.

Apart from their ability to affect tumor cell proliferation andinvasive capacity directly in co-culture Transwell assays in vitro,also mirrored in xenografts in vivo, the most striking feature of thepericyte co-injected tumors was the recruitment of host aSMAþ

yet Ki67� cells that formed a nonvascular network between theOVCAR-5 tumor cells in early day 11 xenografts. The correlationwith increased aSMAþ cell numbers in TMAs from early-relapsepatients suggested that this was a critical functional component ofthe TME, indicative of tumor progression, leading to the findingthat aSMA protein levels yield prognostic significance in a largesample of patient TMAs. However, the combined pericyte signa-ture at the mRNA level was much more effective at predictingrelapse (P ¼ 0.00067) than the level of aSMA staining (P ¼0.03691). High aSMA protein has also been reported to be ofprognostic value in colorectal cancer (7) and at themRNA level inpancreatic adenocarcinoma (8, 46). In contrast, recent studies inmurine models of pancreatic adenocarcinoma provide evidencein favor ofaSMAþ cells having a role in limiting tumor growth andmetastasis by either suppressing immune surveillance (47) orperhaps decreasing tumor angiogenesis (48). Notably, in pancre-atic adenocarcinoma patients low aSMA levels were associatedwith poorer survival (47). These studies further substantiate theneed to examine the intratumoral heterogeneity of aSMAþ stro-mal subsets and examine their role in epithelial cancers ofdifferent tissue origins (e.g., ovarian vs. pancreas).

Interestingly, fibroblast co-injected OVCAR-5 tumors did notshow a sustained increase in the number ofaSMAþ cells (data notshown), correlating with unchanged tumor growth, absence ofinvasive cells at tumor margins, and absence of metastasis.Whereas the influx of higher aSMAþ cell numbers in both exper-imental tumors and early-relapse patient TMAsmaybe an attemptby the host to limit tumor growth and therefore a red herring, itremains possible that their recruitment or perhaps a subtypetherein is required formetastatic spread. Anothermajor differencein pericyte co-injected OVCAR-5 tumors was the recruitment ofhost Sca-1þ/CD73þ BM-MSCs not observed inOVCAR-5 controlsor fibroblast co-injected tumors—given their widely reported rolein cancer cell dissemination, their recruitment by pericytes maywell contribute to metastasis.

The inability of the normal pericyte marker PDGFRb to subsetserous ovarian cancer patients for the probability of relapsesuggests that aSMA and PDGFRb do not identify pericytes exclu-sively in cancer and are expressed by other stromal cells in theTME. Consistent with this, PDGFRb expression was observed inthe tumor stroma in addition to its classic perivascular localiza-tion in patient TMAs. Attempts to define a single pericyte markerto predict poor prognosis in patients with ovarian cancer atdiagnosis were only partially successful. Poor correlation betweenhigh pericyte score and PDGFRb expression levels in TMAs beliedits inability to predict relapse with a high degree of certainty.CD34 served as a negative control, given poor correlation betweenMVD and early versus late relapse. Thus, although angiogenesis isobviously critical for tumor development, it is not relevant tomalignant progression at advanced stages ofmalignancy predom-

inant in the patients analyzed here. AlthoughaSMAprotein levelsachieved reasonable significance levels for predicting relapse(P ¼ 0.03691), it is likely that a number of pericyte markersmight be required to identify patients at greater risk of relapse.

An obvious target of further work is to understand the processby which pericytes become dissociated from blood vesselsduring physiological tissue remodeling. We speculate that cyto-kines used by endothelial cells to attract pericytes to newlyforming blood vessels such as PDGF-B may also be synthesizedby tumor cells—indeed, overexpression of PDGF-B in squamouscarcinomamodels promotes tumor cell proliferation and acts asa chemoattractant and activator for mesenchymal cells (49).However, metastases were not observed in this model, suggest-ing that this single factor is unlikely to cause malignant pro-gression. Certainly the mRNAs co-expressed by pericytes andearly-relapse ovarian cancer patients point to a coordinateregulation of genes enriched in processes essential for tissueremodeling.

These data represent a paradigm shift in the current thinkingabout the contribution of pericytes to the TME while providingan effective means of identifying those patients that are atsignificantly greater risk of earlier relapse and mortality.Undoubtedly, this brings a further level of complexity totherapeutic approaches aimed at inhibiting angiogenesis, butprovides new opportunities to develop effective strategiesagainst stem-cell–like pericytes in the TME, given that anti-pericyte reagents not only exist, but are in clinical use in theguise of anti-angiogenic therapies. The in vitro invasion datasuggest that pericytes secrete soluble factors that induce tumorcell dissemination forming the basis for identifying specificproteins that promote malignant progression that could alsoserve as biomarkers for ovarian cancer, particularly early-stagedisease, given that experimentally, pericyte involvement in theTME results in the induction of metastases in the poorlytumorigenic and nonmetastatic OVCAR-8 cells. Perhaps thegreatest barrier to translating the significance of our findingsto early diagnosis and thereby increasing the chances of overallpatient survival is the lack of early-stage ovarian cancer patientdatabases combining transcriptional and proteomic profilingwith clinical outcome following diagnosis. The collation ofpatient cancer proteomic analysis being undertaken by the NCICPTAC initiative is eagerly anticipated, given the corroborationof our findings between the AOCS and TCGA patient datasets.

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

Authors' ContributionsConception and design: D. Sinha, L. Chong, H. Schl€uter, D. Bowtell, P. KaurDevelopment of methodology: D. Sinha, L. Chong, H. Schl€uter, S. Mills, J. Li,C. Parish, P. KaurAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D. Sinha, L. Chong, H. Schl€uter, S. Mills, D. Bowtell,P. KaurAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D. Sinha, L. Chong, J. George, H. Schl€uter,S. M€onchgesang, J. Li, P. KaurWriting, review, and/or revision of themanuscript:D. Sinha, S. M€onchgesang,J. Li, C. Parish, D. Bowtell, P. KaurAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D. Sinha, L. Chong, H. Schl€uter, P. KaurStudy supervision: H. Schl€uter, S. Mills, C. Parish, P. Kaur

Sinha et al.





Other (obtained funding for the project and oversaw its execution andpublication): P. Kaur

AcknowledgmentsThe authors thank Prof. Robin Anderson, Drs. Nick Clemons, Clare Slaney,

and Izhak Haviv for valuable discussions and technical advice, and Prof. StevenStacker and Prof. Ruth Ganss for critical reading of the manuscript.

Grant SupportThis work was supported by the CASS Foundation, Cancer Council of

Victoria # 807184 and NHMRC # 1025874 grants to P. Kaur and US Army

Medical Research and Materiel Command Grant DAMD17-01-1-0729, theCancer Council Tasmania, the 618 Cancer Foundation of Western Australia,NHMRC # 400413 and Cancer Australia # 1004673 to D. Bowtell. D. Sinhawas supported by an International HDR PhD scholarship from ANU,Canberra.

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 August 13, 2015; revised October 28, 2015; accepted October 30,2015; published OnlineFirst November 20, 2015.

References1. Khan S, Taylor JL, Rinker-Schaeffer CW. Disrupting ovarian cancer meta-

static colonization: insights from metastasis suppressor studies. J Oncol2010;2010:286925.

2. Bast RC Jr., Hennessy B, Mills GB. The biology of ovarian cancer: newopportunities for translation. Nat Rev Cancer 2009;9:415–28.

3. RosenDG, YangG, LiuG,Mercado-Uribe I, Chang B, Xiao XS, et al.Ovariancancer: pathology, biology, and disease models. Front Biosci 2009;14:2089–102.

4. Schauer IG, Sood AK, Mok S, Liu J. Cancer-associated fibroblasts and theirputative role in potentiating the initiation and development of epithelialovarian cancer. Neoplasia 2011;13:393–405.

5. Tothill RW, Tinker AV, George J, Brown R, Fox SB, Lade S, et al. Novelmolecular subtypes of serous and endometrioid ovarian cancer linked toclinical outcome. Clin Cancer Res 2008;14:5198–208.

6. Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, et al.Stromal gene expression predicts clinical outcome in breast cancer. NatMed 2008;14:518–27.

7. Tsujino T, Seshimo I, Yamamoto H, Ngan CY, Ezumi K, Takemasa I, et al.Stromal myofibroblasts predict disease recurrence for colorectal cancer.Clin Cancer Res 2007;13:2082–90.

8. Fujita H, Ohuchida K, Mizumoto K, Nakata K, Yu J, Kayashima T,et al. alpha-smooth muscle actin expressing stroma promotesan aggressive tumor biology in pancreatic ductal adenocarcinoma.Pancreas 2010.

9. Calon A, Lonardo E, Berenguer-Llergo A, Espinet E, Hernando-MomblonaX, Iglesias M, et al. Stromal gene expression defines poor-prognosis sub-types in colorectal cancer. Nat Genet 2015;47:320–9.

10. Frame MC, Serrels A. FAK to the rescue: activated stroma promotes a "safehaven" for BRAF-mutant melanoma cells by inducing FAK signaling.Cancer Cell 2015;27:429–31.

11. KlemmF, Joyce JA.Microenvironmental regulation of therapeutic responsein cancer. Trends Cell Biol 2015;25:198–213.

12. Olumi AF, Grossfeld GD, Hayward SW, Carroll PR, Tlsty TD, Cunha GR.Carcinoma-associated fibroblasts direct tumor progression of initiatedhuman prostatic epithelium. Cancer Res 1999;59:5002–11.

13. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer 2006;6:392–401.

14. Pietras K, Ostman A. Hallmarks of cancer: interactions with the tumorstroma. Exp Cell Res 2010;316:1324–31.

15. Cirri P, Chiarugi P. Cancer associated fibroblasts: the dark side of the coin.Am J Cancer Res 2011;1:482–97.

16. Studeny M, Marini FC, Dembinski JL, Zompetta C, Cabreira-Hansen M,Bekele BN, et al. Mesenchymal stem cells: potential precursors for tumorstroma and targeted-delivery vehicles for anticancer agents. J Natl CancerInst 2004;96:1593–603.

17. Mishra PJ, Mishra PJ, Humeniuk R, Medina DJ, Alexe G, Mesirov JP, et al.Carcinoma-associated fibroblast-like differentiation of human mesenchy-mal stem cells. Cancer Res 2008;68:4331–9.

18. Hung SC, Deng WP, Yang WK, Liu RS, Lee CC, Su TC, et al. Mesenchymalstem cell targeting of microscopic tumors and tumor stroma developmentmonitored bynoninvasive in vivo positron emission tomography imaging.Clin Cancer Res 2005;11:7749–56.

19. 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;449:557–63.

20. Quante M, Tu SP, Tomita H, Gonda T, Wang SS, Takashi S, et al. Bonemarrow-derived myofibroblasts contribute to the mesenchymal stem cellniche and promote tumor growth. Cancer Cell 2011;19:257–72.

21. Shepro D, Morel NM. Pericyte physiology. FASEB J 1993;7:1031–8.22. Armulik A, Abramsson A, Betsholtz C. Endothelial/pericyte interactions.

Circ Res 2005;97:512–23.23. Chen H, Yang WW, Wen QT, Xu L, Chen M. TGF-beta induces fibroblast

activation protein expression; fibroblast activation protein expressionincreases the proliferation, adhesion, and migration of HO-8910PM[corrected]. Exp Mol Pathol 2009;87:189–94.

24. Crisan M, Yap S, Casteilla L, Chen CW, Corselli M, Park TS, et al. Aperivascular origin for mesenchymal stem cells in multiple human organs.Cell Stem Cell 2008;3:301–13.

25. Feng J, Mantesso A, De Bari C, Nishiyama A, Sharpe PT. Dual origin ofmesenchymal stem cells contributing to organ growth and repair. ProcNatlAcad Sci U S A 2011;108:6503–8.

26. Paquet-Fifield S, Schluter H, Li A, Aitken T, Gangatirkar P, BlashkiD, et al. Arole for pericytes as microenvironmental regulators of human skin tissueregeneration. J Clin Invest 2009;119:2795–806.

27. Bergers G, Song S, Meyer-Morse N, Bergsland E, Hanahan D. Benefits oftargeting both pericytes and endothelial cells in the tumor vasculature withkinase inhibitors. J Clin Invest 2003;111:1287–95.

28. Druker BJ. STI571 (Gleevec) as a paradigm for cancer therapy. Trends MolMed 2002;8:S14–8.

29. Erber R, Thurnher A, Katsen AD, Groth G, Kerger H, Hammes HP, et al.Combined inhibition of VEGF and PDGF signaling enforces tumor vesselregression by interfering with pericyte-mediated endothelial cell survivalmechanisms. FASEB J 2004;18:338–40.

30. Kuhnert F, Tam BY, Sennino B, Gray JT, Yuan J, Jocson A, et al. Solublereceptor-mediated selective inhibitionof VEGFR andPDGFRbeta signalingduring physiologic and tumor angiogenesis. Proc Natl Acad Sci U S A2008;105:10185–90.

31. Maciag PC, Seavey MM, Pan ZK, Ferrone S, Paterson Y. Cancer immuno-therapy targeting the highmolecular weightmelanoma-associated antigenprotein results in a broad antitumor response and reduction of pericytes inthe tumor vasculature. Cancer Res 2008;68:8066–75.

32. Sennino B, Kuhnert F, Tabruyn SP, Mancuso MR, Hu-Lowe DD, Kuo CJ,et al. Cellular source and amount of vascular endothelial growth factor andplatelet-derived growth factor in tumors determine response to angiogen-esis inhibitors. Cancer Res 2009;69:4527–36.

33. NisanciogluMH, Betsholtz C, Genove G. The absence of pericytes does notincrease the sensitivity of tumor vasculature to vascular endothelial growthfactor-A blockade. Cancer Res 2010;70:5109–15.

34. Lindblom P, Gerhardt H, Liebner S, Abramsson A, Enge M, Hellstrom M,et al. Endothelial PDGF-B retention is required for proper investment ofpericytes in the microvessel wall. Genes Dev 2003;17:1835–40.

35. CookeVG, LeBleuVS,KeskinD,KhanZ,O'Connell JT, TengY, et al. Pericytedepletion results in hypoxia-associated epithelial-to-mesenchymal transi-tion and metastasis mediated by met signaling pathway. Cancer Cell2012;21:66–81.

36. Xian X, Hakansson J, Stahlberg A, Lindblom P, Betsholtz C, Gerhardt H,et al. Pericytes limit tumor cell metastasis. J Clin Invest 2006;116:642–51.

37. Bentink S, Haibe-Kains B, Risch T, Fan JB, Hirsch MS, Holton K, et al.Angiogenic mRNA and microRNA gene expression signature predicts anovel subtype of serous ovarian cancer. PLoS One 2012;7:e30269.






38. Hamilton TC, YoungRC,Ozols RF. Experimentalmodel systems of ovariancancer: applications to the design and evaluation of new treatmentapproaches. Semin Oncol 1984;11:285–98.

39. Hamilton TC, Young RC, McKoy WM, Grotzinger KR, Green JA, Chu EW,et al. Characterization of a human ovarian carcinoma cell line (NIH:OVCAR-3) with androgen and estrogen receptors. Cancer Res 1983;43:5379–89.

40. Godwin AK, Meister A, O'Dwyer PJ, Huang CS, Hamilton TC, AndersonME. High resistance to cisplatin in human ovarian cancer cell lines isassociated with marked increase of glutathione synthesis. Proc Natl AcadSci U S A 1992;89:3070–4.

41. Labiche A, Heutte N, Herlin P, Chasle J, Gauduchon P, Elie N. Stromalcompartment as a survival prognostic factor in advanced ovarian carcino-ma. Int J Gynecol Cancer 2010;20:28–33.

42. Cojoc M, Peitzsch C, Trautmann F, Polishchuk L, Telegeev GD, DubrovskaA. Emerging targets in cancer management: role of the CXCL12/CXCR4axis. Onco Targets Ther 2013;6:1347–61.

43. Kulbe H, Thompson R, Wilson JL, Robinson S, Hagemann T, Fatah R, et al.The inflammatory cytokine tumor necrosis factor-alpha generates an

autocrine tumor-promoting network in epithelial ovarian cancer cells.Cancer Res 2007;67:585–92.

44. Lee SJ, Bae JH, Lee AW, Tong SY, Park YG, Park JS. Clinical characteristics ofmetastatic tumors to the ovaries. J Korean Med Sci 2009;24:114–9.

45. Lengyel E. Ovarian cancer development and metastasis. Am J Pathol2010;177:1053–64.

46. Infante JR, Matsubayashi H, Sato N, Tonascia J, Klein AP, Riall TA, et al.Peritumoral fibroblast SPARC expression and patient outcomewith resect-able pancreatic adenocarcinoma. J Clin Oncol 2007;25:319–25.

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

48. Rhim AD, Oberstein PE, Thomas DH, Mirek ET, Palermo CF, Sastra SA,et al. Stromal elements act to restrain, rather than support, pancreatic ductaladenocarcinoma. Cancer Cell 2014;25:735–47.

49. Lederle W, Stark HJ, Skobe M, Fusenig NE, Mueller MM. Platelet-derivedgrowth factor-BB controls epithelial tumor phenotype by differentialgrowth factor regulation in stromal cells. Am J Pathol 2006;169:1767–83.


Sinha et al.




2016;22:1813-1824. Published OnlineFirst November 20, 2015.Clin Cancer Res Devbarna Sinha, Lynn Chong, Joshy George, et al. and Predict Poor Prognosis in Serous Ovarian Cancer PatientsPericytes Promote Malignant Ovarian Cancer Progression in Mice

Updated version

10.1158/1078-0432.CCR-15-1931doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2015/11/20/1078-0432.CCR-15-1931.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/22/7/1813.full#ref-list-1

This article cites 48 articles, 16 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://clincancerres.aacrjournals.org/content/22/7/1813To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-15-1931

http://clincancerres.aacrjournals.org/content/suppl/2015/11/20/1078-0432.CCR-15-1931.DC1

http://clincancerres.aacrjournals.org/content/22/7/1813.full#ref-list-1

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/22/7/1813


Documents

Pericytes Promote Malignant Ovarian Cancer Progression in … · Biology of Human Tumors Pericytes Promote Malignant Ovarian Cancer Progression in Mice and Predict Poor Prognosis