Neutrophils Promote the Malignant Glioma Phenotype through S100A4 · Neutrophils Promote the Malignant Glioma Phenotype through S100A4. JiLiang1,YujiPiao1,LindsayHolmes2,GregoryN.Fuller3,VerleneHenry2,NingyiTiao1,andJohnF.deGroot1

Cancer Therapy: Preclinical

Neutrophils Promote the Malignant Glioma Phenotypethrough S100A4

Ji Liang1, Yuji Piao1, Lindsay Holmes2, Gregory N. Fuller3, Verlene Henry2, Ningyi Tiao1, and John F. de Groot1

AbstractPurpose: Antiangiogenic therapy is effective in blocking vascular permeability, inhibiting vascular

proliferation, and slowing tumor growth, but studies in multiple cancer types have shown that tumors

eventually acquire resistance to blockade of blood vessel growth. Currently, the mechanisms by which this

resistance occurs are not well understood.

Experimental Design: In this study, we evaluated the effects of neutrophils on glioma biology both in

vitro and in vivo and determined target genes by which neutrophils promote the malignant glioma

phenotype during anti-VEGF therapy.

Results:We found that an increase in neutrophil infiltration into tumors is significantly correlated with

glioma grade and in glioblastomawith acquired resistance to anti-VEGF therapy. Our data demonstrate that

neutrophils and their conditionmedia increased theproliferation rate of glioblastoma-initiating cells (GIC).

In addition, neutrophils significantly increased GICs Transwell migration compared with controls. Con-

sistent with this behavior, coculture with neutrophils promoted GICs to adopt morphologic and gene

expression changes consistent with a mesenchymal signature. Neutrophil-promoting tumor progression

could be blocked by S100A4 downregulation in vitro and in vivo. Furthermore, S100A4 depletion increased

the effectiveness of anti-VEGF therapy in glioma.

Conclusions: Collectively, these data suggest that increased recruitment of neutrophils during anti-

VEGF therapy promotes glioma progression and may promote treatment resistance. Tumor progression

with mesenchymal characteristics is partly mediated by S100A4, the expression of which is increased by

neutrophil infiltration. Targeting granulocytes and S100A4 may be effective approaches to inhibit the

glioma malignant phenotype and diminish antiangiogenic therapy resistance. Clin Cancer Res; 20(1);

187–98. �2013 AACR.

IntroductionGliomas are the most common type of primary brain

tumors in adults and classified into four grades by theWorldHealth Organization (WHO) according to their degree ofmalignancy and histologic features (1). For patients withhigh-grade gliomas (HGG), such as glioblastoma, the prog-nosis remains poorwith a short survival time (1, 2). Patientswith low-grade tumors have a comparative longer termsurvival, but nearly all low-grade tumors progress tohigh-grademalignancy (3). Identifying the key contributorsand molecular pathways that result in tumor progression isthe focus of intense investigation because inhibition of

malignant progression of glioma has the potential to sig-nificantly extend patient survival.

Over the past two decades, advances made in moleculartechnologies such as microarray technologies and genomesequencing have made it possible to evaluate the molec-ular and genetic changes in malignant brain tumors.Recent gene expression profiling studies have revealedthree to four molecular subclasses of HGGs based ondifferential gene expression (4, 5). One tumor subtype ischaracterized by expression of neural progenitor markers(proneural), which is enriched in low-grade gliomas(LGG; ref. 6), is associated with longer survival, whereasmalignant glioma with a mesenchymal signature areaggressive, commonly coincide with disease recurrence,resistant to chemotherapy, and is associated with a shortersurvival time. In addition, resistance to chemotherapy isalso seen in gliomas that were originally defined as pro-neural and later have increased mesenchymal genesexpression, suggesting a shift to the mesenchymal subtypeover time (4, 5). This shift from proneural to mesenchy-mal transition (PMT) in glioblastoma may drive its aggres-sive behavior and eventually cause chemotherapy failure(7). Currently, it is unknown what drives this PMT inglioblastoma.

Authors' Affiliations: Departments of 1Neuro-Oncology, 2Neurosurgery,and 3Pathology, The University of Texas MD Anderson Cancer Center,Houston, Texas

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

Corresponding Author: John F. de Groot, Department of Neuro-Oncol-ogy, The University of Texas MD Anderson Cancer Center, Unit 431, 1515Holcombe Boulevard, Houston, TX 77030-4009. Phone: 713-792-7255;Fax: 713-794-4999; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-13-1279

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 187

on August 28, 2021. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 15, 2013; DOI: 10.1158/1078-0432.CCR-13-1279

http://clincancerres.aacrjournals.org/

Recent evidence provided by our group and others pointto tumor microenvironment components as potential par-ticipants in the generation of tumor malignancy and che-moresistance (7–11). The tumor microenvironment con-tains largepopulations of cancer-related inflammatory cells,including tumor-associated neutrophils (TAN). In gliomas,increased infiltration of neutrophils has been observed inHGGs compared with LGGs (12). Resistance to chemother-apy is also associated with increased infiltration of neutro-phils. Our previous work shows that glioblastoma resis-tance to anti-VEGF therapy is associated with myeloid cellinfiltration and mesenchymal transition, indicating a pos-itive correlation between them (7). Similar findings havebeen reported for other tumors (13, 14), suggesting thattargeting TAN-promoting tumor progression can be a uni-versal method to overcome drug resistance and tumorrecurrence. Shojaei and colleagues reported that inherentanti-VEGF refractoriness is associatedwith infiltration of thetumor tissue by CD11bþGr1þ myeloid cells. Combininganti-VEGF treatment with antimyeloid cell therapy inhibitsgrowthof refractory tumorsmore effectively than anti-VEGFalone (15, 16). Recent work by Acharyya and colleaguesshowed that inhibition of neutrophil recruitment augmentsthe efficacy of chemotherapy against breast tumors andparticularly against lungmetastasis (8). It has been reportedthat blocking TAN infiltration into tumor tissues can influ-ent several key aspects of tumor progression, includingtumor growth (17–19), angiogenesis (16, 20, 21), andmetastasis (8, 11). Identifying the mechanisms by whichTANs promote tumor progression and the genetic charac-teristics associated with these phenotypic changes isessential to identify the key regulators of glioblastoma

progression and thus guide the development of new ther-apies to overcome resistance.

In this study,wemakeuse of an in vitro coculturemodel toinvestigate the interaction between glioma cells and neu-trophil progenitor cells, especially how neutrophil influ-ence glioma phenotypes and signaling pathways. We foundthat coculture of neutrophil and glioma stem cells increasesthe expression of S100A4 in glioma cells, which was alsoupregulated in anti-VEGF–resistant tumors. Downregulat-ing neutrophil-promoting expression of S100A4 can miti-gate the neutrophil-mediated malignant phenotype in vitroand in vivo. Furthermore, S100A4 depletion prolongs thesurvival of anti-VEGF treated animals and partially reducesglioblastoma resistance to anti-VEGF therapy.

Materials and MethodsBrain tumor tissue microarray

Ahumanglioma tissuemicroarray (TMA)was constructedusing formalin-fixed, paraffin-embedded archival tissueblocks as described previously (22). The TMA includedsamples from 232 primary brain tumors of varying gradestaken from sites of the most phenotypically representativetumor regions. The array contained 96 glioblastoma tumors,13 gliosarcoma tumors, 12 anaplastic mixed oligoastrocy-toma tumors, 32 anaplastic astrocytoma tumors, 24 ana-plastic oligodendroglioma tumors, 29 oligodendrogliomatumors, 11 mixed oligoastrocytoma tumors, and four low-grade astrocytomas. Normal brain tissue samples wereincluded in the array as a negative control. Expression levelsof myeloperoxidase-positive (MPO) were evaluated by astandard indirect immunoperoxidase procedure as previous-ly described (7). Mayer’s hematoxylin nuclear staining wasused as a counterstain. An intensity score was assigned toeach sample that represented the average numbers of thepositive staining on an arbitrary scale of 0 to 5. Statisticalanalysis was done using a Kruskal–Wallis analysis of ranks.

Cell cultureGlioma stem cell line GIC23, GIC11, GIC2, and GIC20

were obtained from Dr. Howard Colman (Department ofNeuro-Oncology, MD Anderson Cancer Center, Houston,TX). Glioblastoma-initiating cells (GIC) were maintained insuspension in Dulbecco’s Modified Eagle Medium (DMEM)containing EGF, basic fibroblast growth factor (bFGF), andB27 (Invitrogen) at 37�C in5%CO2atmosphere.Neutrophilprogenitor cell line CRL-11422 (AACR) was cultured inIscove’s modified Dulbecco’s medium with 4 mmol/L L-glutamine adjusted to contain 1.5 g/L sodium bicarbonatecontaining 10 ng/mL murine granulocyte macrophage col-ony-stimulating factor (GM-CSF), 80% and heat-inactivatedhorse serum, 20%.Before collecting conditionedmedia, cellswere changed to serum-free and GM-CSF–free media. Con-ditioned media were collected after 24 hours incubation.

Cell-cycle analysisFor cell-cycle analysis, asynchronous GIC cells were

fixed with ethanol and stained with 50 mg/mL propidiumiodide (PI) containing 0.1 mg/mL RNaseA (both from

Translational RelevanceTumor progression is affected by a wide variety of

components within the tumor microenvironment, andthe importance of neutrophils in glioma has yet to befully characterized. Although recent work, includingours, has overwhelmingly shown that neutrophils pro-mote tumor progression, some studies suggest that neu-trophils can be "polarized" to both protumor and anti-tumor phenotypes by the tumor microenvironment.Therefore, therapeutic targeting these cells may be chal-lenging. Amore straightforward approach is to target thespecific neutrophil-promoting factors on tumor cells.Here, for the first time, we show that S100A4 is upre-gulated on tumor cells by neutrophils. Downregulationof S100A4 can mitigate the neutrophil-mediated mes-enchymal phenotype in vitro and in vivo. Furthermore,S100A4 depletion prolongs the survival of anti-VEGFtreated animals and partially reduces glioblastoma resis-tance to anti-VEGF therapy. Therefore, the combinationof an S100A4 inhibitor with antiangiogenic therapymay delay glioma progression and anti-VEGF therapyresistance.

Liang et al.

Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research188




Sigma-Aldrich). Cells were analyzed by flow cytometry todetermine sub-G1 (apoptosis), G1, S, and G2–M cell-cycledistribution (FACStar Plus Flow Cytometer; Becton-Dickinson).

Immunoblot analysisCells were lysed in an ice-cold lysis buffer containing 50

mmol/L Tris-Cl, pH 7.5, 100 mmol/L NaCl, 1 mmol/LEDTA, 1% TritonX-100, 1 mmol/L phenylmethylsulfonyl-fluoride (PMSF), 1 mg/mL leupeptin, and 1 m/mL pepstainA. The protein concentration in the supernatant was deter-mined using a BCA protein assay (Pierce). Samples weresubjected to 8% to 12% SDS-PAGE, and the separatedproteins were electrophoretically transferred to nitrocellu-lose membranes. Blots were incubated with the primaryantibody against cyclin D1 (1:1,000; CST), cyclin D2(1:1,000; CST), c-myc (1:1,000; CST), tubulin (1:3,000;Sigma), Ykl-40 (1:1,000; SantaCruzBiotechnology),matrixmetalloproteinase (MMP)2 (1:1,000, Chemicon-Milli-pore), nestin (1:1,000; Abcam) or S100a4 (1:1,000;Abcam). The membranes were then incubated with horse-radish peroxidase-linked secondary anti-rabbit or anti-mouse antibodies (Bio-Rad).

Invasion assayMatrigel Basement Membrane Matrix (BD Labware) was

used to perform the in vitro cell invasion assays. Cells werepretreatedwith bevacizumab for 72hours. Transwell insertsfor 24-well plates were coated with diluted Matrigel, andcells were added in triplicate to the Transwells. Serum-freemedium was added to the bottom of the plate. Cells wereallowed to invade for 24 hours at 37�C. The filters were thenfixed and stained with 0.1% crystal violet in 20%methanol.The invasive cells were visualized using bright-field micros-copy. Transwellmembraneswere incubatedwith 2%deoxy-cholic acid for 20 minutes, and the absorbance at 595 nmwas recorded.

Microarray and Ingenuity pathway analysisAffymetrixGeneChipHumanGenomeHG-U133Plus 2.0

Arrays (Affymetrix) were used for expression profiling. Thelist of genes was overlaid onto a global molecular networkdeveloped from information contained in the IngenuityPathwayAnalysis (IPA) knowledgebase (IPKB). For networkanalysis, IPA computed a score [P score ¼ �log (P value)]according to the fit of the set of supplied genes and a list ofbiologic functions stored in the IPKB. The score takes intoaccount the number of genes in the network and the size ofthe network to approximate how relevant this network is tothe original list of genes. A score more than 1.3 (P < 0.05)indicates a significant change in the gene network. Thenetwork identified is presented as a graph indicating themolecular relationships between genes/gene products.

ImmunofluorescenceImmunofluorescence analysis was done as previously

described with minor modifications (23). Briefly, formal-dehyde-fixed cells were permeabilized with Triton X-100

0.1% in PBS, and blockedwith 5% serumdiluted in PBS-gel(0.2% gelatin in PBS) for 30 minutes. The primary anti-bodies were incubated in blocking solution overnight at4�C. Immunostaining was performed using the primaryantibody against Ykl-40 (1:50; Santa Cruz Biotechnology),CD31 (1:50; Abcam) and ly6B.2 (1:50; AbD Serotec).Coverslips were mounted using ProLong antifade reagent(Invitrogen). The images were acquired with an inverteddeconvolution microscope. Images were taken with a ZeissAxioskop 40 microscope equipped with AxioVision Rel.4.2software.

Animal xenograftsFor in vivo experiments, GIC cells (3 � 105) with or

without CRL11422 (9� 105) were implanted intracraniallyinto 4- to 6-week-old female nude mice (12 mice pergroup). The mice were euthanized at 3, 6, 9, 11 weeks, andtheir brains were removed and processed for analysis. Allexperiments were approved by the Institutional AnimalCare and Use Committee of The University of Texas MDAnderson Cancer Center. Tumor volume analysis was doneusing an unpaired two-tailed Student test and groups werecompared using the log-rank test. P < 0.05 was determinedto be significant.

ImmunohistochemistryParaffin sections from xenografts were used for immu-

nohistochemical analysis. The slides were deparaffinizedand subjected to graded rehydration. After blocking in 5%serum and an antigen retrieval step (citrate buffer, pH 6.0),the slides were incubated with the primary antibodiesovernight at 4�C. After washing in PBS with Tween 20,primary antibody reactions were detected using the VEC-TASTAIN ABC Kit (Vector Laboratories) with the respectivesecondary antibody.

TransfectionCells were plated at a density of 3� 105 per six-well plate 3

hours before transfection. Transfection was carried out usingHyFect reagents according to the vendor’s instructions. Trans-fected cultures were selected with puromycin (5 mg/mL) for10 to 14 days. At that time, antibiotic-resistant colonies werepicked, pooled, and expanded for further analysis underselective conditions. The pGIPZ control was generated withcontrol oligonucleotide GCTTCTAACACCGGAGGTCTT.pGIPZ S100a4 shRNA was generated with TGCTCAGCAT-CAAGCACGT and TGAGCTTGAACTTGTCACC.

The Cancer Genome Atlas analysisS100A4 gene expression values were determined among

glioblastoma datasets in TheCancerGenomeAtlas (TCGA),accessed through the cBio Cancer Genomics Portal (http://cbioportal.org) supplied byMemorial Sloan Kettering Can-cer Center (website accessed on April 25, 2013). The levelsof expression of S100A4 were evaluated in the differentTCGA glioblastoma subtypes (5). For each glioma subtype,patients were divided into two groups on the basis of ongene expression (Z score < or � 2). Kaplan–Meier curves

Neutrophils Promote Mesenchymal Phenotype through S100A4

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 189




were generated using these data. The difference in survivalbetween patients with high and low expression of S100A4was calculated for each subtype.

ResultsNeutrophil infiltration is correlated with glioma gradeand tumor progression

A human glioma TMA was used to evaluate for thepresence of neutrophil infiltration using immunohisto-chemical staining for MPO. The numbers of MPO-positivecells for each tumor sample were calculated from grade 2, 3,and4 glioma. Eighty-six percent of all glioma samples in our

TMA (n¼232) had some evidence of neutrophil infiltration(Fig. 1A). Grade 4 glioblastoma specimens had the highestnumbers of infiltration neutrophil [number of neutrophil >10/200� HPF (high-powered field; Fig. 1A) and Supple-mentary Table S1]. The level of neutrophil infiltration wassignificantly positively correlated with glioma grade. Toconfirm the level of neutrophil infiltration in murine mod-els of glioma, we evaluated the number of neutrophilsinfiltrated into tumor xenografts injected into the brain ofnude mice. The number of murine neutrophils character-ized by themarker Ly-6B.2 (24, 25) increased during tumorgrowth and progression (Fig. 1B).

5 wk

4 wk

3 wk

2 wk

Ly6B.2/Hoechst

GB

GS

AMOA

OL

AO

AA

MOA

LGA

NB

BA

Figure 1. Levels of neutrophilinfiltration into gliomas correlatewith tumor grades and tumorprogression. A, neutrophilinfiltration into different gradeglioma was detected byimmunohistochemical stainingusing an anti-MPO antibody(�200). Grade 4: GB, glioblastoma;GS, gliosarcoma; grade 3:AMOA, anaplastic mixedoligoastrocytoma; AO, anaplasticoligodendroglioma; AA, anaplasticastrocytoma; grade 2: OL,oligodendroglioma; MOA, mixedoligoastrocytoma; LGA, low-gradeastrocytoma; NB, normal brain. B,neutrophil infiltration was detectedby immunostaining of Ly6B.2 intumors collected from GICsxenograft nude mice at 2, 3,4, and 5 weeks (�200).

Liang et al.





Neutrophils promote the proliferation of glioma stemcellsTo identify the effects of neutrophils on tumor cells, we

performed in vitro coculture of glioma stem cells (GIC) withneutrophil progenitor cells (MPRO cells, CRL11422;ATCC). After 2 days of coculture, the number of GFP-positive GIC11 cells was significantly increased comparedwith controls (Fig. 2A). Conditioned media collected fromMPRO cells (24-hour serum-free culture medium) alsoincreased the numbers of GICs (Fig. 2B), suggesting thatsecreted factors from neutrophils may be partially respon-sible for the tumor proliferating effects.Further evaluation by PI staining identified that the

increase in cell number resulted from both a decrease inGIC apoptosis (Fig. 2C) and an increase in cell proliferation(Fig. 2D). Western blot analysis data demonstrated that

conditioned media upregulated the expression of prolifer-ation-related proteins cyclin D1, cyclin D2, and c-myc (Fig.2E). In addition, conditioned media increased the phos-phorylation of Akt and Erk (Supplementary Fig. S1A).Blocking Akt and Erk pathway activation by their inhibitorsdecreased the expression of cyclinD1 and c-myc induced byconditioned media (Supplementary Fig. S1B). These dataindicate that, at least inour in vitromodel, factors secretedbyneutrophils promote GIC proliferation.

Neutrophils promote a mesenchymal phenotypeWe next investigated the effect of neutrophils on the

ability of GIC invasion. Matrigel Transwell assays showedan increase in Transwell migration/invasion for GICs whencocultured with MPRO cells compared with controls (Fig.3A). GIC cells in culture demonstrated a spindle-shaped

A

B

C D

****

** *

****

**

Cyclin D1

Cyclin D2

C-myc

Tubulin

_ + _ +

GIC11 GIC23CME

2.5

2.0

1.5

1.0

0.5

0.0

GIC

11

GIC

11 +

MPR

OGIC

11

GIC

11 +

MPR

O

Fo

ld c

han

ge o

f

cell n

um

bers

1.5

1.0

0.5

0.0

Fo

ld c

han

ge o

f

cell n

um

bers

1.5

1.0

0.5

0.0

8

6

4

2

0

Fo

ld c

han

ge o

f

cell n

um

bers

25

Perc

en

tag

es o

f

G2--

M+

S

Perc

en

tag

es o

f

ap

op

tosis

20

15

10

5

0

GIC

11

GIC

11 +

CM

GIC

23 +

CM

GIC

23

GIC

11

GIC

11 +

CM

GIC

23 +

CM

GIC

23

GIC

23 +

CM

GIC

23

GIC

11 +

CM

GIC

11

Figure 2. Neutrophils promote theproliferation of GICs. A,morphology of GFP-positive GICswith and without coculture withMPRO cells (at 1:1 ratio) for 3 days(�100). Bar graph demonstratingthe fold change of GFP-expressingcell numbers analyzed by flowcytometry. B, fold changes of cellnumbers after 3 days incubation ofGIC11 or GIC23 with serum-freeconditioned media (CM) collectedfrom MPRO cells. Cell-cycleanalysis using PI staining showedthat CM decreased thepercentages of cells in sub-G1

stages (apoptotic cells, C) andincreased that in G2–MþS stages(growing cells, D). E, expression ofcyclin D1, cyclin D2, C-myc, andtubulin were detected by Westernblot analyses in GICs incubatedwith or without CM. �, P < 0.05;��, P < 0.01, Student t test.






morphology, which is a typical morphology for mesenchy-mal cells (26), when cocultured with MPRO cells (Fig. 3B).Consistent with a more mesenchymal phenotype, Westernblot analyses showed increased expression ofmesenchymalmarkers YKL-40 andMMP2, and less expression of nestin, aproneural marker following coculture (Fig. 3C). To confirmthe effects of neutrophils on GICs and to investigate theunderlying mechanisms by which neutrophils promote amesenchymal glioma phenotype, we performed geneexpression analysis of GIC23 treatedwith orwithoutMPROcells. IPA analysis (Supplementary Fig. S2) showed a sig-nificant shift toward a mesenchymal gene signature andincreased cell mobility in the coculture group comparedwith the control group. These data indicate that neutrophilspromote a mesenchymal shift in GICs. In addition, geneexpression analysis demonstrated that neutrophils alsopromote an increase in genes related to cell division pro-cesses, consistent with the results presented in Fig. 2.

Neutrophils promote a malignant glioma phenotypethrough S100A4

In our previous work, we found that glioblastoma resis-tance to anti-VEGF therapy was associated with neutrophilinfiltration and a mesenchymal phenotype (7). Geneexpression profiling data from these experiments also iden-tified (Fig. 4A) multiple genes that were upregulated duringthe development of resistance. Several of these genesmaybeinvolved in the regulation of neutrophil-mediated mesen-chymal transformation.One of the candidate genes S100A4was found to be highly expressed in GIC with a mesenchy-

mal phenotype (Fig. 4B, GIC2 and GIC20), whereas noexpression was observed in GICs of the proneural subclass(Fig. 4B, GIC11 andGIC23). To identify the role of S100A4,we developed a stable GIC23 cell line with a specific shRNAto S100A4 (Fig. 4C).

Consistentwith the gene expression analysis,GICs’ cocul-ture with neutrophils increased S100A4 expression, whichwas not observed in the S100A4 knockdown cell line.Although down regulation of S100A4 did not affect neu-trophil-mediated proliferation (Fig. 4E), neutrophil pro-motion of tumor cell migration (Fig. 4F) and YKL-40expression (Fig. 4G)were inhibitedby S100A4knockdown.To evaluate the impact of S100A4 on tumor growth in vivo,we then coinjected neutrophils with control shRNA orS100A4 shRNA GIC23 into nude mice. Consistent with thein vitro data, neutrophils’ coinjection increased humanS100A4 expression in tumor tissue, whereas almost nohumanS100A4 stainingwas observed in the S100A4knock-down group (Supplementary Fig. S3A). Our data showedthat coinjection of neutrophils with control GIC23 cellsenhanced tumor growth, an effect that was significantlyblunted by using cells with S100A4 knockdown (Fig. 5A).Increased YKL-40 expression was also observed in controlcells injected concurrently with neutrophils, but not inS100A4 knockdown cells with neutrophils (Fig. 5B). Final-ly, neutrophil coinjection increased vascularization in con-trol GICs group but not in the S100A4 knockdown group(Fig. 5C). Taken together, these data indicate that S100A4was partially responsible for neutrophil-mediated gliomaprogression.

GIC11 GIC11 + MPRO

GIC

11

GIC

11 +

MPR

O

GIC

23

GIC

23 +

MPR

O

GIC23 GIC23 + MPRO

OD

590 n

m

0

0.2

0.4

0.6

0

0.1

0.2

0.3

GIC11 GIC11 + MPROB

MPRO_ + _ +

YKL-40

Nestin

GIC11 GIC23

MMP2

Tubulin

A

C

OD

590 n

m

****

Figure 3. Neutrophils promote amesenchymal phenotype of GICsin vitro. A, Matrigel invasion assayfor GIC11 (left) and GIC23 (right)cells. GICs (3 � 105) were allowedto invade for 24 hours in serum-freemedium with or without MPROcells (3 � 105). Graphs representabsorbance at 590 nm afterincubation of the membranes withdeoxycholic acid. Pictures shownare the most representative formof the three independentexperiments. �, P < 0.05;��, P < 0.01, Student t test. B,morphology of GIC colonies withand without MPRO cells' (smallround cells) coculture (cellnumbers at 1:1 ratio; �200).C, expression of Ykl-40, MMP2,nestin, and tubulin were detectedby Western blot analyses in GICswhen cocultured with or withoutMPRO cells. GFP-positive GICswere sorted for Western blotanalyses after 1:1 coculture withMPRO cells for 3 days.

Liang et al.





S100A4 depletion increases the efficacy of anti-VEGFtherapy in gliomasBecause S100A4 depletion may block neutrophil-medi-

ated mesenchymal transformation, which is a prominentfeature of anti-VEGF–resistant tumors, we next investigatedwhether S100A4 depletion would improve the efficacy ofanti-VEGF therapy. Our survival study in animals demon-strated that mice treated with bevacizumab had a longersurvival time than the no treatment group [P¼ 0.0044, log-rank (Mantel–Cox) test; Fig. 6A], and S100A4 depletionfurther prolonged survival [P ¼ 0.0137, log-rank (Mantel–Cox) test; Fig. 6A]. Immunostaining data showed that thespecific shRNAdramatically inhibited S100A4expression intumor cells (Supplementary Fig. S3B), but this inhibition

did not affect the number of infiltrated neutrophils duringbevacizumab treatment (Supplementary Fig. S3C). Asreported previously (7, 27), we found that the angiogenesismarker CD31 was markedly inhibited in both control cellsand S100A4-depleted tumors following 4 weeks of bevaci-zumab treatment (Fig. 6B, 4 week), indicating that thebevacizumab-treated tumor had not escaped from therapyat that time point. However, at 6 weeks, tumors werebevacizumab-resistant and demonstrated an increase inCD31 staining, which was not observed in the S100A4depletion group (Fig. 6B, 6 week). Interestingly, YKL-40staining was increased in bevacizumab-treated tumors at asearly as 4 weeks, which was also inhibited by S100A4downregulation at 4 weeks and 6 weeks (Fig. 6C). These

MPRO_ +

ControlshRNA

S100A4shRNA

A B

E F

G

PN11 23 2 20

S100A4

Tubulin

MPRO − + − +

S100A4

GFP

MES

ControlshRNA

S100A4 shRNA

C

D

ID Fold changeCA9 2.247

CEBPD 2.982CHI3L1 2.134

COL15A1 3.514COL3A1 2.537COL4A2 2.109

CRISPLD2 2.747CXCL1 2.847HEY1 3.921

IFITM1 3.154IL8 2.474

MMP3 2.772MMP9 2.163S100A4 2.420

SERPINA3 3.221SERPINA5 2.150

TGFBI 2.928THY1 3.394TIMP2 2.011

* *

S100A4

Tubulin

S100A4shS100A4 − − +

− + −GIC23

YKL-40

_ + _ + MPRO

Tubulin

shRNA shS100A4

Figure 4. Neutrophils promote aS100A4-mediated malignantglioma phenotype. A, increasedexpression of mesenchymal-related markers were observedin microarray analysis forbevacizumab-resistant micetumors. B, S100A4 expression inGICs of proneural (PN) andmesenchymal (MES) subtypes.C, transient transfection ofS100A4overexpression and shRNAplasmids in GIC23. D, S100A4expression in GIC23 cells stablytransfected with either ControlshRNA or S100A4 shRNA whencocultured with MPRO cells. E,flow cytometry analysis of GFP-Control shRNA or GFP-S100A4shRNA GIC23 cells whencocultured with or without MPROcells (cell numbers at 1:1 ratio). F,Matrigel invasion assay for GIC23Control shRNA and S100A4shRNA cells. GICs (3 � 105) wereallowed to invade for 24 hours inserum-freemediumwith orwithoutMPRO cells (3 � 105). Graphsrepresent absorbance at 590 nmafter incubation of the membraneswith deoxycholic acid. Picturesshown are the most representativeform of the three independentexperiments. �, P < 0.05, Student ttest. G, expression of Ykl-40 andtubulin were detected by Westernblot analyses in GIC23 whencocultured with or without MPROcells. GFP-positive GICs weresorted for Western blot analysesafter 1:1 coculture with MPRO cellfor 3 days.






data suggest that the mesenchymal transformation of glio-mas was regulated by S100A4 and may be responsible forthe increase in angiogenesis during tumor progression onanti-VEGF therapy.

DiscussionGlioblastoma is recognized as the most common and

lethal form of brain cancer. Current treatments for glioblas-toma include temozolomide paired with radiotherapy, andto more recently the use of antiangiogenic agents (28).However, anti-VEGF agents have not significantly extendedthe life expectancies of patients diagnosed with glioblasto-ma (29, 30), which has partially been attributed to drugresistance. The major aim of this study is to determine themechanisms underlying glioblastoma resistance to antian-giogenic therapy. Although there is strong preclinical evi-

dence supporting the efficacy of antiangiogenic therapy,tumors eventually adopt a highly resistant phenotype. Inthis study, we demonstrate that neutrophils may play acritical role in tumor-escape from antiangiogenic therapy.Consistent with previous studies (7), we observed anincrease in neutrophil infiltration into gliomas, which maybe partially driven by tumor hypoxia possibly exacerbatedby prolonged anti-VEGF therapy (7, 31). The extent ofneutrophil infiltration was positively correlated with glio-ma grade. Our studies provide evidence that neutrophilspromote the glioma malignant phenotype and increase theexpression of mesenchymal markers in vitro and in vivo.

Neutrophils have been reported to play an important rolein stimulating tumor angiogenesis (32, 33). In patients withmyxofibrosarcomas, increased numbers of neutrophilswere associated with increased intratumoral microvesseldensity (34). In vivo studies using various tumor models

A

B

Co

ntr

ol

sh

RN

AS

10

0A

4 s

hR

NA

MPRO - +C

on

tro

l s

hR

NA

S

10

0A

4 s

hR

NA

MPRO - +

C

Co

ntr

ol

sh

RN

A

S100A

4 s

hR

NA

MPRO - +

* *

** **

**

Ykl-40/Hoechst

CD31

Figure 5. Neutrophils promote tumorprogression through S100A4 in miceglioma xenograft model. A, representativewhole mounts of hematoxylin and eosin(H&E)–stained brains containing gliomatumors from GIC23 Control shRNA cellsalone, GIC23 Control shRNA cellscoinjected with MPRO cells, GIC23S100A4 shRNA cells alone, and GIC23S100A4 shRNA cells coinjected withMPRO cells in nude mice (left). Tumorgrowth from the earlier four groups (right).Data are shown as averages � SD, n ¼ 9mice per group. ��,P < 0.01; P values weredetermined by t test. B, representativelight microscopy images (left) showingimmunostaining of Ykl-40 (green) in GICgliomas. Images were taken at �200. Bargraph (right) demonstrating the meanstaining of mesenchymal marker Ykl-40.��, P < 0.01, Student t test. C,representative light microscopy images(left) showing immunostaining ofangiogenesis marker CD31 (green) inGIC gliomas (�100). Bar graph (right)demonstrating the mean vascular densityin GIC gliomas. Data are shown asaverages � SD; n ¼ 9 fields taken from atleast 3 mice per group at each time point.�, P < 0.05, Student t test. Cell nuclei werecounterstained with Hoechst (blue).

Liang et al.





of angiogenesis demonstrated that neutrophils promoteneovascularization (20, 21, 35). In our proneural gliomastem-cell xenograft model, we found increased infiltrationof neutrophils and expression of mesenchymal marker

following 4 weeks of anti-VEGF treatment, with enhance-ment of angiogenesis at later stages of tumor escape. Mes-enchymal tumors are highly vascularized, consistent withthe findings that the mesenchymal gene expression

A

B

C

P = 0.0137

6 wk

*

6 wk

**

4 wk

* *

4 wk

* **

D

Days

Figure 6. Inhibition of S100A4increases the efficacy of anti-VEGFtherapy in gliomas. A, Kaplan–Meier survival curves for GIC11Control shRNA and S100A4shRNA cell-injected mice treatedwith or without bevacizumab (Bev,10 mg/kg). GIC11 S100A4 shRNAcell-injected mice treated with Bevhave a longer survival durationcompared with treated GIC11Control shRNA cell–injected mice.[P ¼ 0.0137, long-rank(Mantel–Cox) test]. Bar graphdemonstrating the mean stainingof CD31 (B) and Ykl-40 (C) at 4 and6 weeks. Data are shown asaverages � SD; n ¼ 9 fields takenfrom at least 3 mice per group ateach time point. �, P < 0.05;��, P < 0.01; P values weredetermined by t test. D, modelshowing how anti-VEGF treatmentfailure is promoted by neutrophilsthrough S100A4.






signature contains multiple proangiogenic genes (4). Ourdata suggest that neutrophils may promote angiogenesis byinducing a shift of gliomas fromproneural tomesenchymalsubtype (Fig. 6D). Although activated myeloid cells such asM2-skewedmacrophages can directly secrete proangiogenicfactors (36–39), TANs may also regulate tumor angiogen-esis in an indirect way, such as through shifting stem cellstoward a mesenchymal phenotype.

In the present study, we found that a mesenchymaltransformation in glioma stem cells was promoted byneutrophils through upregulation of S100A4. S100A4belongs to the S100 family of proteins that contain twoCa2þ-binding sites, including a canonical EF-hand motif(40). S100A4 is involved in the regulation of awide range ofbiologic effects, including cell motility, survival, differenti-ation, and contractility (41). Several of the S100A4 effectsresemble processes that occur during epithelial–mesenchy-mal transition (EMT), and a direct link between S100A4expression and EMT has been suggested in a number oforgans (42–44). In human patients with cancer, increasedexpression of S100A4 is positively associated with anincreased incidence of metastasis, invasiveness, aggres-siveness, andaworse prognosis (45–50).Ourmesenchymalglioblastoma stem cells expressed S100A4, whereas GICswith a proneural signature did not. Likewise, a query of theTCGA identified 5.8% of the glioblastoma samples (206cases) overexpressed S100A4 and most of these samples(58.6%) were of the mesenchymal subtype and associatedwith a shorter overall survival (P ¼ 0.003964). Conversely,14.3% proneural glioblastoma samples have downregula-tion of S100A4 and longer survival (P ¼ 0.009221). Con-sistent with our results that neutrophils can increase theexpression of S100A4 in vitro, S100A4 and CD64, a humanmonocyte marker, are significantly coexpressed in glioblas-toma patient samples from TCGA. Cumulatively, these datasuggest that S100A4 may be a key contributor of mesen-chymal transition in glioblastoma progression, and theincreased expression of S100A4 can be at least partiallypromoted by an increase in neutrophil infiltration.

Currently, it is difficult to discern the exact mechanismsunderlying S100A4-mediated mesenchymal transforma-tion because S100A4 can function both intra- and extra-cellularly. It is likely that the intracellular and the extracel-lular effects involve distinct mechanisms. S100A4 has beenreported to interact with multiple proteins that are knownto play a role in tumor progression and metastasis, such asMMP-9 (51). Iwamoto and colleagues showed that patients

with glioblastoma with longitudinal increases in MMP-9had a shorter survival (52). Studies in endothelial cells andtumors demonstrate that S100A4 expression can be pro-moted by TGF-b1 (43, 53), which is well known to induceEMT. Consistently, we have observed an increase in theTGF-b pathway in anti-VEGF–resistant tumors (7). Furtherstudies are needed to identify the molecular mechanism bywhich S100A4 regulate the mesenchymal phenotype.

Although TANs play a critical role in tumor progression,therapeutic targeting of this cell type in cancer will bechallenging. Neutrophils are critical mediators of hostdefense against infection, and depletion of these cells couldresult in dangerous levels of immunosuppression. Howev-er, we found that downregulation of S100A4 inhibitedneutrophil-promoting tumor progression independent ofthe infiltration of neutrophils. Our findings provide apossible alternative strategy to target the specific neutro-phil-activated regulator on tumor cells, S100A4. Therefore,the combination of S100A4 inhibitor with standard anti-angiogenesis therapy may inhibit glioma progression andanti-VEGF therapy resistance.

Disclosure of Potential Conflicts of InterestJ.F. de Groot received commercial research grant from EMD Serono,

Sanofi Aventis, AstraZeneca, Eli Lilly, and Deciphera Pharmaceuticals andis a consultant/advisory board member of Genentech and Novartis. Nopotential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: J. Liang, Y. Piao, J.F. de GrootDevelopment of methodology: J. Liang, J.F. de GrootAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): V. Henry, J.F. de GrootAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): J. Liang, Y. Piao, G.N. Fuller, J.F. de GrootWriting, review, and/or revision of themanuscript: J. Liang, G.N. Fuller,J.F. de GrootAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): N. Tiao, J.F. de GrootStudy supervision: L. Holmes, J.F. de Groot

Grant SupportThis work was supported in part by an American Society of Clinical

Oncology (ASCO) Career Development Award and a grant from the MarthaG. Williams Brain Tumor Research Fund (both to J. F. de Groot). The MDAnderson Cancer Center is supported in part by a Core grant (CA16672).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received May 13, 2013; revised October 23, 2013; accepted October 24,2013; published OnlineFirst November 15, 2013.

References1. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A,

et al. The 2007 WHO classification of tumours of the central nervoussystem. Acta Neuropathol 2007;114:97–109.

2. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, TaphoornMJ, et al. Radiotherapy plus concomitant and adjuvant temozolomidefor glioblastoma. N Engl J Med 2005;352:987–96.

3. Louis DN.Molecular pathology ofmalignant gliomas. Annu Rev Pathol2006;1:97–117.

4. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD,et al. Molecular subclasses of high-grade glioma predict prognosis,delineate a pattern of disease progression, and resemble stages inneurogenesis. Cancer Cell 2006;9:157–73.

5. VerhaakRG,HoadleyKA, PurdomE,WangV,QiY,WilkersonMD, et al.Integrated genomic analysis identifies clinically relevant subtypes ofglioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR,and NF1. Cancer Cell 2010;17:98–110.

Liang et al.





6. Cooper LA,GutmanDA, LongQ, JohnsonBA,Cholleti SR,KurcT, et al.The proneural molecular signature is enriched in oligodendrogliomasand predicts improved survival among diffuse gliomas. PLoS ONE2010;5:e12548.

7. Piao Y, Liang J, Holmes L, Zurita AJ, Henry V, Heymach JV, et al.Glioblastoma resistance to anti-VEGF therapy is associated withmyeloid cell infiltration, stem cell accumulation, and a mesenchymalphenotype. Neuro Oncol 2012;14:1379–92.

8. Acharyya S, Oskarsson T, Vanharanta S, Malladi S, Kim J, Morris PG,et al. A CXCL1 paracrine network links cancer chemoresistance andmetastasis. Cell 2012;150:165–78.

9. DeNardo DG, Brennan DJ, Rexhepaj E, Ruffell B, Shiao SL, MaddenSF, et al. Leukocyte complexity predicts breast cancer survival andfunctionally regulates response to chemotherapy. Cancer Discov2011;1:54–67.

10. Shree T, Olson OC, Elie BT, Kester JC, Garfall AL, Simpson K, et al.Macrophages and cathepsin proteases blunt chemotherapeuticresponse in breast cancer. Genes Dev 2011;25:2465–79.

11. Van den Eynden GG,Majeed AW, IllemannM, Vermeulen PB, Bird NC,Hoyer-Hansen G, et al. The multifaceted role of the microenvironmentin liver metastasis: biology and clinical implications. Cancer Res 2013;73:2031–43.

12. Fossati G, Ricevuti G, Edwards SW, Walker C, Dalton A, Rossi ML.Neutrophil infiltration into human gliomas. Acta Neuropathol 1999;98:349–54.

13. Nielsen BS, Timshel S, Kjeldsen L, SehestedM, Pyke C, Borregaard N,et al. 92 kDa type IV collagenase (MMP-9) is expressed in neutrophilsand macrophages but not in malignant epithelial cells in human coloncancer. Int J Cancer 1996;65:57–62.

14. Bellocq A, AntoineM, Flahault A, Philippe C, Crestani B, Bernaudin JF,et al. Neutrophil alveolitis in bronchioloalveolar carcinoma: inductionby tumor-derived interleukin-8 and relation to clinical outcome. Am JPathol 1998;152:83–92.

15. Shojaei F,Wu X, Qu X, KowanetzM, Yu L, TanM, et al. G-CSF-initiatedmyeloid cell mobilization and angiogenesis mediate tumor refractori-ness to anti-VEGF therapy in mousemodels. Proc Natl Acad Sci U S A2009;106:6742–7.

16. Shojaei F, Wu X, Zhong C, Yu L, Liang XH, Yao J, et al. Bv8 regulatesmyeloid-cell-dependent tumour angiogenesis. Nature 2007;450:825–31.

17. Huang S, Mills L, Mian B, Tellez C, McCarty M, Yang XD, et al. Fullyhumanized neutralizing antibodies to interleukin-8 (ABX-IL8) inhibitangiogenesis, tumor growth, andmetastasis of humanmelanoma. AmJ Pathol 2002;161:125–34.

18. Fridlender ZG, Sun J, Kim S, Kapoor V, Cheng G, Ling L, et al.Polarization of tumor-associated neutrophil phenotype by TGF-beta:"N1" versus "N2" TAN. Cancer Cell 2009;16:183–94.

19. Jablonska J, Leschner S, Westphal K, Lienenklaus S, Weiss S. Neu-trophils responsive to endogenous IFN-beta regulate tumor angio-genesis and growth in a mouse tumor model. J Clin Invest 2010;120:1151–64.

20. Nozawa H, Chiu C, Hanahan D. Infiltrating neutrophils mediate theinitial angiogenic switch in a mouse model of multistage carcinogen-esis. Proc Natl Acad Sci U S A 2006;103:12493–8.

21. Benelli R, Morini M, Carrozzino F, Ferrari N, Minghelli S, Santi L, et al.Neutrophils as a key cellular target for angiostatin: implications forregulation of angiogenesis and inflammation. FASEBJ2002;16:267–9.

22. Liu TJ,WangM,BreauRL,HendersonY, El-Naggar AK,SteckKD, et al.Apoptosis induction by E2F-1 via adenoviral-mediated gene transferresults in growth suppression of head and neck squamous cell car-cinoma cell lines. Cancer Gene Ther 1999;6:163–71.

23. de Groot J, Liang J, Kong LY, Wei J, Piao Y, Fuller G, et al. Modulatingantiangiogenic resistance by inhibiting the signal transducer andactivator of transcription 3 pathway in glioblastoma. Oncotarget 2012;3:1036–48.

24. Hirsch S, Gordon S. Polymorphic expression of a neutrophil differen-tiation antigen revealed bymonoclonal antibody 7/4. Immunogenetics1983;18:229–39.

25. McDonald JU, Cortini A, Rosas M, Fossati-Jimack L, Ling GS,Lewis KJ, et al. In vivo functional analysis and genetic modifi-

cation of in vitro-derived mouse neutrophils. FASEB J 2011;25:1972–82.

26. Thiery JP, Sleeman JP. Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol 2006;7:131–42.

27. Lucio-Eterovic AK, Piao Y, de Groot JF. Mediators of glioblastomaresistance and invasion during antivascular endothelial growth factortherapy. Clin Cancer Res 2009;15:4589–99.

28. Weller M. Novel diagnostic and therapeutic approaches to malignantglioma. Swiss Med Wkly 2011;141:w13210.

29. Gilbert MR, Dignam J, Won M, Blumenthal DT, Vogelbaum MA,Kenneth D, et al. RTOG 0825: Phase III double-blind placebo-con-trolled trial evaluating bevacizumab (Bev) in patients (Pts) with newlydiagnosed glioblastoma (GBM). J Clin Oncol 31: 2013 (suppl; abstr 1).

30. Armstrong TS, Won M, Wefel JS, Gilbert MR, Pugh SL, Brachman D,et al. Comparative impact of treatment on patient reported out-comes (PROs) in patients with glioblastoma (GBM) enrolled in ROTG0825. J Clin Oncol 31: 2013 (suppl; abstr 2003).

31. Du R, Lu KV, Petritsch C, Liu P, Ganss R, Passegue E, et al. HIF1alphainduces the recruitment of bone marrow-derived vascular modulatorycells to regulate tumor angiogenesis and invasion. Cancer Cell2008;13:206–20.

32. Tazzyman S, Lewis CE, Murdoch C. Neutrophils: key mediators oftumour angiogenesis. Int J Exp Pathol 2009;90:222–31.

33. Gregory AD, Houghton AM. Tumor-associated neutrophils: new tar-gets for cancer therapy. Cancer Res 2011;71:2411–6.

34. Mentzel T, Brown LF, Dvorak HF, Kuhnen C, Stiller KJ, Katenkamp D,et al. The association between tumour progression and vascularity inmyxofibrosarcomaandmyxoid/round cell liposarcoma. VirchowsArch2001;438:13–22.

35. Van Coillie E, Van Aelst I, Wuyts A, Vercauteren R, Devos R, De Wolf-Peeters C, et al. Tumor angiogenesis induced by granulocyte chemo-tactic protein-2 as a countercurrent principle. Am J Pathol 2001;159:1405–14.

36. McCourt M, Wang JH, Sookhai S, Redmond HP. Proinflammatorymediators stimulate neutrophil-directed angiogenesis. Arch Surg1999;134:1325–31.

37. Coussens LM, Tinkle CL, Hanahan D, Werb Z. MMP-9 supplied bybone marrow-derived cells contributes to skin carcinogenesis. Cell2000;103:481–90.

38. Bergers G, Brekken R, McMahon G, Vu TH, Itoh T, Tamaki K, et al.Matrix metalloproteinase-9 triggers the angiogenic switch during car-cinogenesis. Nat Cell Biol 2000;2:737–44.

39. Ardi VC,KupriyanovaTA, Deryugina EI,Quigley JP.Humanneutrophilsuniquely release TIMP-free MMP-9 to provide a potent catalyticstimulator of angiogenesis. Proc Natl Acad Sci U S A 2007;104:20262–7.

40. Heizmann CW, Cox JA. New perspectives on S100 proteins: a multi-functional Ca(2þ)-, Zn(2þ)- and Cu(2þ)-binding protein family. Bio-metals 1998;11:383–97.

41. Schneider M, Hansen JL, Sheikh SP. S100A4: a common mediator ofepithelial-mesenchymal transition, fibrosis and regeneration in dis-eases? J Mol Med 2008;86:507–22.

42. Iwano M, Plieth D, Danoff TM, Xue C, Okada H, Neilson EG. Evidencethat fibroblasts derive from epithelium during tissue fibrosis. J ClinInvest 2002;110:341–50.

43. Zeisberg EM, Potenta S, Xie L, Zeisberg M, Kalluri R. Discovery ofendothelial to mesenchymal transition as a source for carcinoma-associated fibroblasts. Cancer Res 2007;67:10123–8.

44. Okada H, Danoff TM, Kalluri R, Neilson EG. Early role of Fsp1 inepithelial-mesenchymal transformation. Am J Physiol 1997;273:F563–74.

45. Wang YY, Ye ZY, Zhao ZS, Tao HQ, Chu YQ. High-level expression ofS100A4 correlates with lymph node metastasis and poor prognosis inpatients with gastric cancer. Ann Surg Oncol 2010;17:89–97.

46. Yonemura Y, Endou Y, Kimura K, Fushida S, Bandou E, Taniguchi K,et al. Inverse expression of S100A4 and E-cadherin is associated withmetastatic potential in gastric cancer. Clin Cancer Res 2000;6:4234–42.

47. Kimura K, Endo Y, Yonemura Y, Heizmann CW, Schafer BW, Wata-nabe Y, et al. Clinical significance of S100A4 and E-cadherin-related






adhesion molecules in non-small cell lung cancer. Int J Oncol 2000;16:1125–31.

48. Wang HY, Zhang JY, Cui JT, Tan XH, Li WM, Gu J, et al. Expressionstatus of S100A14 and S100A4 correlates with metastatic potentialand clinical outcome in colorectal cancer after surgery. Oncolo Rep2010;23:45–52.

49. ZouM, Al-Baradie RS, Al-Hindi H, Farid NR, Shi Y. S100A4 (Mts1) geneoverexpression is associated with invasion andmetastasis of papillarythyroid carcinoma. Br J Cancer 2005;93:1277–84.

50. Ismail NI, Kaur G, Hashim H, Hassan MS. S100A4 overexpressionproves to be independent marker for breast cancer progression.Cancer Cell Int 2008;8:12.

51. Saleem M, Kweon MH, Johnson JJ, Adhami VM, Elcheva I, KhanN, et al. S100A4 accelerates tumorigenesis and invasion ofhuman prostate cancer through the transcriptional regulation ofmatrix metalloproteinase 9. Proc Natl Acad Sci U S A 2006;103:14825–30.

52. Iwamoto FM, Hottinger AF, Karimi S, Riedel E, Dantis J, Jahdi M,et al. Longitudinal prospective study of matrix metalloproteinase-9as a serum marker in gliomas. J Neuro Oncol 2011;105:607–12.

53. Zeisberg M, Hanai J, Sugimoto H, Mammoto T, Charytan D, StrutzF, et al. BMP-7 counteracts TGF-beta1-induced epithelial-to-mes-enchymal transition and reverses chronic renal injury. Nat Med2003;9:964–8.

Liang et al.





2014;20:187-198. Published OnlineFirst November 15, 2013.Clin Cancer Res Ji Liang, Yuji Piao, Lindsay Holmes, et al. S100A4Neutrophils Promote the Malignant Glioma Phenotype through

Updated version

10.1158/1078-0432.CCR-13-1279doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2013/11/15/1078-0432.CCR-13-1279.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/1/187.full#ref-list-1

This article cites 51 articles, 11 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/1/187.full#related-urls

This article has been cited by 7 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://clincancerres.aacrjournals.org/content/20/1/187To 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-13-1279

http://clincancerres.aacrjournals.org/content/suppl/2013/11/15/1078-0432.CCR-13-1279.DC1

http://clincancerres.aacrjournals.org/content/20/1/187.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/1/187.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/20/1/187


Documents

Neutrophils Promote the Malignant Glioma Phenotype through S100A4 · Neutrophils Promote the Malignant Glioma Phenotype through S100A4. JiLiang1,YujiPiao1,LindsayHolmes2,GregoryN.Fuller3,VerleneHenry2,NingyiTiao1,andJohnF.deGroot1