role of cxcl13-cxcr5 crosstalk between malignant ... · neuroblastoma is a stroma-poor (sp)...

Angiogenesis, Metastasis, and the Cellular Microenvironment

Role of CXCL13-CXCR5 Crosstalk Between MalignantNeuroblastoma Cells and Schwannian StromalCells in Neuroblastic Tumors

Federica Del Grosso1, Simona Coco1, Paola Scaruffi1, Sara Stigliani1, Francesca Valdora6,Roberto Benelli2, Sandra Salvi3, Simona Boccardo3, Mauro Truini3, Michela Croce4, Silvano Ferrini4,Luca Longo1,5, and Gian Paolo Tonini1

AbstractNeuroblastoma is a stroma-poor (SP) aggressive pediatric cancer belonging to neuroblastic tumors, also

including ganglioneuroblastoma and ganglioneuroma, two stroma-rich (SR) less aggressive tumors. Our previousgene-expression profiling analysis showed a different CXCL13 mRNA expression between SP and SR tumors.Therefore, we studied 13 SP and 13 SR tumors by reverse transcription quantitative real-time PCR (RT-qPCR)and we found thatCXCR5bwas more expressed in SP than in SR andCXCL13was predominantly expressed in SRtumors. Then, we isolated neuroblastic and Schwannian stromal cells by laser capture microdissection and wefound that malignant neuroblasts express CXCR5b mRNA, whereas Schwannian stromal cells express CXCL13.Immunohistochemistry confirmed that stroma expresses CXCL13 but not CXCR5. To better understand the roleof CXCL13 and CXCR5 in neuroblastic tumors we studied 11 neuroblastoma cell lines and we detected aheterogeneous expression of CXCL13 and CXCR5b. Interestingly, we found that only CXCR5b splice variant wasexpressed in both tumors and neuroblastoma lines, whereas CXCR5awas never detected. Moreover, we found thatneuroblastoma cells expressing CXCR5 receptor migrate toward a source of recombinant CXCL13. Lastly,neuroblastoma cells induced to glial cell differentiation expressed CXCL13 mRNA and protein. The chemokinereleased in the culture medium was able to stimulate chemotaxis of LA1–5S neuroblastoma cells. Collectively, ourdata suggest that CXCL13 produced by stromal cells may contribute to the generation of an environment in whichthe malignant neuroblasts are retained, thus limiting the possible development of metastases in patients with SRtumor. Mol Cancer Res; 9(7); 815–23. ’2011 AACR.

Introduction

Neuroblastic tumors are a group of pediatric cancers thatonset as localized or disseminated disease (1). Patients withlocalized tumor have good prognosis with a 3-years overallsurvival (OS) of 90% to 95%, whereas patients older thanone year with a disseminated disease have a worse prognosis

with a 3-years OS ranging between 30% and 35%. Thisdramatic behavior of the metastatic disease is mainly due tothe capacity of neuroblastoma cells to metastasize at bone,bone marrow, liver, and lymph nodes. Particularly, inpatients with disseminated disease, the normal bone marrowcell population is almost completely substituted by malig-nant neuroblasts that find a favorable environment for theirproliferation.Neuroblastic tumors show a quite heterogeneous histol-

ogy with different tumor histotypes including: neuroblas-toma predominantly composed of small roundundifferentiated or poorly differentiated neuroblastic cells,with few or absent Schwannian stromal cells, ganglioneur-oblastoma intermixed (GNB), and ganglioneuroma (GN)mostly composed of Schwannian stromal cells in which somenests or very few neuroblastic cells are present (2). Theabsence of stromal cellsmakes the tumormore aggressive andpatients have a worse outcome. On the contrary, the abun-dance of Schwannian stromal cells, as observed in GNB andGN, is associated with a less aggressive tumor and a localizeddisease with a more favorable prognosis (1, 3).Several evidences support a nonmalignant origin of

Schwannian stromal cells (4, 5), which are supposed to

Authors' Affiliation: 1Translational Oncopathology, 2Department ofImmunology, 3Department of Diagnostic Technologies, 4Laboratory ofImmunological Therapy, 5Italian Neuroblastoma Foundation, NationalCancer Research Institute; and 6Department of Oncology and Genetics,University of Genoa, Genoa, Italy

Notes: Supplementary data for this article are available at MolecularCancer Research Online (http://mcr.aacrjournals.org/).

Current address for P. Scaruffi: Center of Physiopathology of HumanReproduction, Dept. Obstetrics and Gynecology, "San Martino" Hospital,Genoa, 16132, Italy.

Corresponding Author: Gian Paolo Tonini, Translational Oncopathol-ogy, National Cancer Research Institute (IST), L.go R. Benzi, 10-16132,Genoa, Italy. Phone: 39-010-5737487/381; Fax: 39-010-5737487;E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-10-0367

’2011 American Association for Cancer Research.

MolecularCancer

Research

www.aacrjournals.org 815

on October 4, 2020. © 2011 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst June 3, 2011; DOI: 10.1158/1541-7786.MCR-10-0367

http://mcr.aacrjournals.org/

control tumor growth by secreting soluble factors thatinfluence cell proliferation, differentiation, and angiogen-esis. Indeed, various molecules have been found expressedby Schwannian stromal cells (6, 7), although their role incontrolling neuroblastic cells growth and dissemination isnot completely understood. In several tumors apart neuro-blastic tumors, the presence of stromal cells is associatedwith a better disease outcome and the contiguity betweenstroma and tumor cells has an important role in cancergrowth and invasion.Recently, we studied a series of neuroblastic tumors by

gene-expression profiling analysis and we observed differentlevels of CXCL13 transcripts in microdissected neuroblasticcells with respect to Schwannian stromal cells (5, 8). Ourresults showed that this chemokine was more expressed instromal cells than in malignant neuroblasts, suggesting afunctional role of CXCL13 in the interaction betweenstroma and neuroblastic cells.Chemokines are a large family of molecules that, together

with their receptors, are involved in the regulation ofchemotaxis, cell proliferation, cell migration, and severalother crucial cell–cell interactions (9). Moreover, chemo-kines play a major role in tumor growth, angiogenesis, andinvasion (10, 11) in several cancers and neoplastic cells mayexpress chemokine receptors (12). Over the last years,CCR1, CCR5, CCR6, CCR9, CXCR1, CXCR2, CXCR4,CXCR5, and CXCR6 were found expressed in neuroblas-toma and it has been shown that CXCR4 is virtuallyexpressed in all human neuroblastoma cell lines (13–16).Starting from our observation we further investigated the

expression of CXCL13 and its receptor CXCR5 in neuro-blastic primary tumors and human neuroblastoma cell lines.CXCL13 and CXCR5 mRNAs were differently expressedbetween Schwannian stromal cells and neuroblastic cellssuggesting a cell–cell crosstalk via CXCL13-CXCR5 axis.This crosstalk may contribute to retain neuroblastic cellswithin stroma-rich tumors and possibly to inhibit the malig-nant cells dissemination. Our results add new informationabout theCXCL13–CXCR5 axis in neuroblastic tumors andin the crosstalk between neuroblastic and stromal cells.

Materials and Methods

Tumor samples, neuroblastoma cell lines, and totalRNA isolationTumor specimens were collected from 26 patients at

onset of disease, who were diagnosed with a primaryneuroblastic tumor and referred to the Gaslini Children'sHospital. The study was approved by Ethics committee ofthe Gaslini Children's Hospital and informed consent wasobtained by all children's legal guardians. According to theInternational Neuroblastoma Pathology Committee (2), 13samples were from SP tumors with at least 80% of neuro-blastic cells, and 13 were from SR with Schwannian stromalcells ranging from 80% to 90% (Supplementary Table 1).Eleven neuroblastoma cell lines (ACN, GI-CA-N, GI-LI-N, GI-ME-N, IMR-5, IMR-32, LA1–5S, SH-SY5Y, SK-N-BE(2), SK-N-BE(2)c, SK-N-SH) and the Raji B-lym-

phocytic cell line were cultured at 37�C and 5% CO2 inRPMI 1640, supplemented with L-glutamine, penicilline/streptomycin, nonessential amino acids, and 10% FBS(Lonza). Cells were removed from substrate with PBS/EDTA. All cell lines were tested for mycoplasma contam-ination and authenticated (Supplementary Table 2). TotalRNA was isolated from neuroblastoma cell lines by Per-fectPure RNA Cell Kit (5Prime), and from tumor tissuesby PerfectPure Tissue RNA Cell Kit (5Prime), and treatedwith RNase-free DNase I. RNA integrity and quantifica-tion were checked by RNA 6000 Nano LabChip kit and2100 BioAnalyzer instrument (Agilent Technologies).

Isolation of neuroblastic and Schwannian stromal cellsby laser capture microdissection and total RNAisolationAbout 800 cells were laser microdissected from 3 SP and

3 SR frozen tumors to obtain pure cell populations ofneuroblastic cells and Schwannian stromal cells, as pre-viously described (8). Total RNA from laser capture micro-dissection-derived material was extracted by PicoPure RNAisolation kit (Arcturus Engineering), including a DNasetreatment, RNA quality control and quantification whereasconducted by RNA 6000 Pico LabChip kit and the 2100BioAnalyzer instrument (Agilent Technologies).

Reverse-transcription quantitative real-time PCRRNA (1 mg) from neuroblastoma cell lines and tumors

was reverse transcribed using 20 pmoles of random hex-amers (Eppendorf) and 200 U of SuperScript II enzyme(Invitrogen Life Technologies). RNA (50 ng) from micro-dissected cells was amplified and reverse transcribed byWT-Ovation RNA Amplification System kit (NuGEN Tech-nologies). CXCL13, CXCR5, CXCR5a, S100A6, and 18SrRNA transcripts were quantified by TaqMan Gene Expres-sion Assays (Applied Biosystems). An assay specific forCXCR5b isoform was designed by PrimerDesign Ltd.18S rRNA was used as reference gene. Any amplifiedproduct with a quantification cycle (Cq) higher than 36cycle was considered undetectable. According to the Mini-mum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) (17, 18), a checklistincluding technical details is submitted as Supplementarydata (MIQE checklist). The relative amount of each tran-script was determined using the equation 2�dCq, where dCq¼ (Cqtarget gene–Cq18SrRNA).

CXCR5 detection in neuroblastoma cells by flowcytometryNeuroblastoma cell suspensions (3 � 105 cells/tube) and

Raji cell line suspension (1.5� 105 cells/tube) were washedwith PBS and incubated with 2.5 mg/ml monoclonal mouseanti-human CXCR5/Blr-1 primary antibody (R&D Sys-tems) in 2% BSA/PBS for 30 minutes on ice. Afterincubation, cells were washed twice with 2% BSA/PBSand 50 ml of goat anti-mouse IgG2b-PE secondary antibody(Becton Dickinson) was added. Cell suspensions wereincubated for 30 minutes on ice. Percentage of positive

Del Grosso et al.

Mol Cancer Res; 9(7) July 2011 Molecular Cancer Research816




cells and mean fluorescence intensity of stained cells wereanalyzed using a FACScan (Becton Dickinson). The back-ground levels were set using cells stained with secondaryantibody alone and viable cells were discriminated byevaluating the level of propidium incorporation.

CXCR5 and CXCL13 detection in neuroblastoma cellsby immunofluorescenceNeuroblastoma cells were directly fixed in a methanol

fixing solution and were dropped on microscope glass slidesand dried at room temperature (RT). Slides were incubated30 minutes with 10% goat serum in PBS. Primary mono-clonal anti-human CXCR5/Blr-1 mouse antibody wasdiluted 1:40 with PBS 1.5% normal goat serum, appliedon slides at a concentration of 12 mg/ml and incubated 1hour at RT in a humidified chamber. After washing, slideswere incubated 40 minutes at RT in a humidified chamberwith goat anti-mouse-IgG2b-FITC conjugated secondaryantibody (Santa Cruz Biotechnology, Inc.) diluted to aconcentration of 1.5 mg/ml with PBS 3% goat normalserum. To identify the expression of CXCL13, acetonepermeabilized cells were incubated 1 hour at room tem-perature in a humidified chamber with primary anti-humanCXCL13 goat antibody diluted 1:50 with PBS 1.5%normal donkey serum, applied on slides at 2 mg/ml. Afterwashing, slides were incubated 45 minutes at RT in ahumidified chamber with donkey anti-goat IgG(HþL)TRITC conjugated secondary antibody (Jackson Immu-noresearch Europe Ltd) diluted 1:50 with PBS 3% donkeynormal serum. Staining without primary antibodies wascarried out as negative control. Slides were mounted withmounting medium and inspected under the microscopewith a fluorescence lamp.

CXCR5 detection in primary tumor byimmunofluorescenceFrozen sections of 4 mm from 5 SP and 5 SR tumors were

fixed in 4% paraformaldehyde for 2 minutes. Sections wereincubated with Ultra V Block (Thermo Fisher Scientific,LabVision Inc.) for 8 minutes at RT to block unspecificbinding of the antibody. Primary monoclonal anti-humanCXCR5/Blr-1 mouse antibody was diluted 1:40 with PBSand incubated 1 hour at RT in a humidified chamber. Afterwashes three-time, slides were incubated for 45 minutes atRT in a humidified chamber with goat anti-mouse IgG(HþL) Alexa Fluor 488 Conjugate (Molecular Probe, Invi-trogen, Inc.) diluted 1:100 in PBS and counterstained with40, 6 diamidino 2 phenylindole (DAPI). Staining withoutprimary antibody was carried out as negative control.

CXCR5 and CXCL13 detection in primary tumors andneuroblastoma cells by immunohistochemistryImmunohistochemistry analysis was conducted using

polymeric complex technique and automated immunostai-ner Benchmark XT (VentanaMedical System). Sections of 5mmfrom the same tumor samples were air dried and fixed (10minutes with acetone and 5–10 seconds in methanol at RT)and methanol fixed neuroblastoma cell lines (SH-SY5Y and

LA1–5S) were incubated for 30 minutes at 37�C withmonoclonal anti-human CXCR5/BLR1 mouse antibody(20 mg/ml for neuroblastoma cell lines; 10 mg/ml for tumorsections) and anti-human CXCL13/BLC/BCA-1 goat anti-body (0.8 mg/ml for tumor sections) (R&D Systems, Inc)and anti-human GFAP (Glial fibrillary acidic protein) poli-clonal rabbit antibody (Ventana Medical System). Anti-CXCR5 antibody was detected by UltraViewRed detectionkit (Ventana Medical System) based on polymeric alkalinephosphatase detection system. Staining with anti-CXCL13antibody was detected by LSABþSystem HRP Kit (DAKOCytomation, Glostrup, Denmark). Fixed cells and sectionswere counterstained with Gill modified hematoxylin(Ventana Medical System), dehydrated, and mounted.Staining without primary antibody was carried out asnegative control. Staining of Raji cells and frozen orparaffin-embedded from lymph nodes and spleen sampleswas carried out as positive control.

Glial differentiation assaysN-type (SH-SY5Y), I-type (SK-N-BE(2)c) and S-type

(LA1–5S) neuroblastoma cell lines were grown by adding10 mmol/L 50-Bromo-20-deoxyuridine (BrdU) (SigmaAldrich, Inc.) to cell culture medium to induce glial differ-entiation (19). Treatment was maintained for 3 weeks byrenewing media three times per week. During the first andthe second week of treatment, cell proliferation was notaffected by the treatment and cells were split twice. Culturemedia of treated cells were collected every three days for3 weeks and then concentrated from 40 ml to 200 ml withcentrifugal filter devices Centricon Plus-70 (Millipore,Inc.). Total RNA was extracted from the cells after 1, 2,and 3 weeks of treatment with BrdU. Pictures of differ-entiated cells have been taken after 3 weeks of treatment.

CXCL13 detection in neuroblastoma cell lineconditioned medium by ELISACXCL13 protein concentration was detected in neuro-

blastoma cell line conditioned medium by QuantikineHuman CXCL13 Immunoassay (R&D Systems, Inc.),following the manufacturer's recommendations. The assayemploys the quantitative sandwich enzyme immunoassaytechnique. A monoclonal antibody for CXCL13 wasprecoated onto a microplate. Human recombinantCXCL13 (hrCXCL13) standards, saliva, untreated, andBrdU differentiated neuroblastoma cells conditionedmedia were pipetted into the wells and CXCL13 wasbound by the immobilized antibody. After washing, anenzyme-linked monoclonal antibody specific for CXCL13was added. After incubation and extensive washing, asubstrate solution was added and the amount of CXCL13was quantified by a colorimetric measurement on a stan-dard curve of serial dilution of hrCXCL13.

Neuroblastoma cell migration assayCellmigration assays were conducted on SH-SY5Y, ACN,

and LA1–5S cells inmicrochambers (NeuroProbesGaithers-burg) as previously described (20). Cells were extensively

CXCL13-CXCR5 Crosstalk in Neuroblastoma

www.aacrjournals.org Mol Cancer Res; 9(7) July 2011 817




washed with PBS, resuspended in serum-free medium, andplaced in the upper compartment. The two compartments ofthe chamber were separated by 5 mmpore size polycarbonatefilters coated with 0.01% gelatine. The hrCXCL13 (100 or200 ng/ml) and concentrated serum-free conditioned med-ium fromBrdU-differentiated neuroblastoma cells were usedas chemoattractants in the lower chamber. To obtain specificneutralization of hrCXCL13 bioactivity a blocking antihu-man CXCL13/BLC/BCA-1 goat antibody (R&D Systems,Inc) was added to hrCXCL13 and to BrdU differentiatedneuroblastoma cells conditioned medium in the lowerchamber at 3.5 mg/ml and 0.5 mg/ml, respectively. Tospecifically neutralize CXCR5 receptor on neuroblastomacell surface, cells were previously incubated 30 minutes with15 mg/ml blocking monoclonal antihuman CXCR5/BLR1mouse antibody (R&D Systems, Inc). After 6 hours ofincubation at 37�C in 5% CO2, the filters were recovered,the cells on the upper surface were mechanically removed,and the cells on the lower surface were fixed and stained. Themigration was measured by counting the remaining cells onthe filter lower surface. Experiments were done in sixreplicates and repeated three times.

Results

CXCL13 and CXCR5b mRNA expression in wholetumor, isolated neuroblastic and Schwannian stromalcells, and neuroblastoma cell linesWe carried out RT-qPCR of 26 tumor samples to assess

the expression of CXCL13 and CXCR5b mRNA in neuro-blastic tumors. We firstly observed that SP tumors showedhigher levels of CXCR5b mRNA than SR tumors. Con-versely, CXCL13 was more expressed in SR tumors (Fig. 1).Noteworthy, RT-qPCR analysis showedCXCR5b transcriptvariant (accession number 032966; ref. 21) in neuroblastictumors. In our analysis we used primers for both CXCR5aand CXCR5b variants but only the CXCRb transcript wasdetected whereas CXCR5a was not detectable in any sam-ples (data not shown).Because neuroblastic tumors show tissue heterogeneity

and nonmalignant cells may infiltrate the tumor, we isolatedeither neuroblastic or Schwannian stromal cells from 6neuroblastic tumors (3 SP and 3 SR) by laser capturemicrodissection. Next, we extracted total RNA from bothpurified cell populations and we carried out RT-qPCR for

5,000

10,000

1,000

100

10

1

10,000

1,000

100

10

1

0.1

0.01

0.001

0.0001

4,500

4,000

3,500

3,000

MN

EM

NE

CXCL13 CXCL13

CXCR5b

P = 0.1

P = 0.002

SP SR

SP SR

CXCR5b

2,500

2,000

1,500

1,000

500

0

2036

_SP

2216

_SP

2277

_SP

1636

_SP

2220

_SP

1548

_SP

1678

_SP

1618

_SP

1958

_SP

2122

_SP

1644

_SP

1634

_SP

1630

_SR

1533

_SR

1644

_SR

1761

_SR

1689

_SR

1811

_SR

1797

_SR

1590

_SR

1525

_SR

2088

_SR

1669

_SR

1627

_SR

1589

_SR

2497

_SP

2036

_SP

2216

_SP

2277

_SP

1636

_SP

2220

_SP

1548

_SP

1678

_SP

1618

_SP

1958

_SP

2122

_SP

1644

_SP

1634

_SP

1630

_SR

1533

_SR

1644

_SR

1761

_SR

1689

_SR

1811

_SR

1797

_SR

1590

_SR

1525

_SR

2088

_SR

1669

_SR

1627

_SR

1589

_SR

2497

_SP

1,600

1,400

1,200

1,000

800

600

400

200

0

Figure 1. CXCL13 and CXCR5b expression in tumors. CXCL13 and CXCR5b expression by RT-qPCR in 26 Neuroblastic tumors: 13 SP and 13 SR tumors.Left, the level of CXCL13 and CXCR5 in SP and SR are shown, whereas right, shows the box plot representation; P value is reported. Most of SR tumorsexpressCXCL13 and very few expressCXCR5; on the contrary, SP tumors expressCXCR5 at various amount.CXCL13 is very low or undetectable expressedin SP tumors. (MNE, Mean normalized expression).

Del Grosso et al.





CXCL13 andCXCR5b.We observed expression ofCXCL13in the stromal cell components, as opposed to very low orundetectable levels observed in isolated neuroblastic cells(Fig. 2), supporting our results obtained by whole tumorsanalysis. On the contrary, we definitely detected CXCR5bexpression in the isolated neuroblastic cells but not instromal cells (Fig. 2). Afterward, we investigated the expres-sion of CXCL13 and CXCR5bmRNA in 11 neuroblastomacell lines and we detected a variable expression of CXCL13and CXCR5b. Among neuroblastoma cell lines, 3/11 (27%)express CXCL13 (GI-LI-N, IMR5, SK-N-SH) and 9/11(82%) express CXCR5b (GI-LI-N, GI-ME-N, IMR-32,IMR5, LA-1–5S, SH-SY5Y, SK-N-BE(2), SK-N-BE(2)c,SK-N-SH). GI-LI-N, IMR5, SK-N-SH cells express bothCXCL13 and CXCR5b and ACN and GI-CA-N cells do notexpress either CXCL13 or CXCR5b (data not shown).

CXCR5 and CXCL13 protein expression inneuroblastic tumors and neuroblastoma cell linesTo show that Schwannian stromal cells of SR tumors

express CXCL13 but not CXCR5 we carried out immu-nohistochemistry analysis on tumor tissue. As shown inFig. 3C, Schwannian stromal cells strongly expressCXCL13, whereas CXCR5 is not expressed in the sametissue (Fig. 3D). Because CXCR5b mRNA was foundvariable expressed we employed immunofluorescence tech-nique to have a better sensibility to detect CXCR5 expres-sion. CXCR5 protein was observed in neuroblastic cells(Fig. 3E and 3F), but not in SR tumors (data not shown).

Next, we analyzed SH-SY5Y and LA1–5S neuroblastomacell lines by either immunofluorescence (Fig. 3G and 3H) orimmunohistochemistry (Fig. 3I and Fig. 3J) andwe observeda mild surface expression of CXCR5 receptor in about 20%of SH-SY5Y and 30% of LA1–5S neuroblastoma cells,respectively (Fig. 3G and Fig. 3H). Immunofluorescenceanalysis also showed some cytoplasmic staining, possiblyrelated to a CXCR5 internalization. Immunofluorescenceanalysis was also carried out in the other neuroblastoma celllines with similar results and we confirmed CXCR5 expres-sion in SK-N-BE(2), SK-N-BE(2)c and SK-N-SH cells byFACS analysis (data not shown). Although IMR-5, SK-N-SH and GI-LI-N, expressed CXCL13 mRNA, only IMR-5showed CXCL13 cytoplasmatic expression but the secretedprotein was not detected (data not shown).

Induction of neuroblastoma cells to glial differentiationand CXCL13 productionHuman neuroblastoma cell lines retain a bilineage poten-

tial andmay differentiate toward a neuroblast or stromal/glialphenotype on different conditions (19). Because secretedCXCL13 was not detectable by ELISA in the conditionatedmedium of any neuroblastoma cell line (data not shown), wetreated SH-SY5Y, SK-N-BE(2)c and LA1–5S cells withBrdU for 3 weeks to evaluate whether neuroblastoma cellsproduce CXCL13 after glial differentiation induction. Fol-lowing BrdU treatmentmost of SH-SY5Y and SK-N-BE(2)ccells acquired an evident flat, large and more adherent cellmorphology resembling Schwannian stromal cells (Fig. 4A).Amorphologic change was also observed in LA1–5S cells at aless extent. The induction towards glial cell lineage wasshown by increase of S100A6mRNA expression in all 3 celllines (Fig. 4A). In addition, a variable de novo CXCL13mRNA induction was detected in neuroblastoma cell linesreaching a maximum at 3 weeks in BrdU-differentiated SK-N-BE(2)c cells (Fig. 4B). The secretion of CXCL13 proteinin the culture medium was remarkable in differentiated SK-N-BE(2)c and LA1–5S (Fig. 4C).

Triggering of neuroblastoma cells migration byCXCL13We tested whether the expression of CXCR5 could be

sufficient to trigger neuroblastoma cell migration in responseto CXCL13 stimulation by chemotaxis assays. Figure 5Ashows that the CXCR5-positive SH-SY5Y and LA1–5S cellssignificantlymigrated in presence of hrCXCL13 chemokine,whereas the CXRC5-negative ACN cells failed to migrate(P¼ not significant, data not shown). LA1–5S cell migrationwas inhibited by either blocking the receptor or the chemo-kine activity, following addition of specific antibodies againstCXCR5 or hrCXCL13 (Fig. 5B). LA1–5S cell migrationwas also observed when the conditioned medium of neuro-blastoma cells differentiated to the glial cell lineage were usedas chemoattractants, thus further substantiating the possi-bility that cell-released CXCL13 induces neuroblastoma cellmigration (Fig. 5C). This result shows that neuroblastomacells induced to glial cell lineage and acquiring a Schwannian-likemorphology secreteCXCL13 that stimulates chemotaxis

3.5E-05

3.0E-05

2.5E-05

2.0E-05

3.0E-06

4.0E-06

2.0E-06

1.5E-05

1.0E-05

5.0E-06

0.0E+00

1.0E-06

5.0E-06

6.0E-06

0.0E+00

1589_SS 1591_SS 1913_SS 2570_Nb 1919_Nb 2182_Nb

1589_SS 1591_SS 1913_SS 2570_Nb 1919_Nb 2182_Nb

CXCL13

NM

EN

ME

CXCR5b

Figure 2. CXCL13 and CXCR5b expression in isolated neuroblastic andSchwannian stromal cells. CXCL13 and CXCR5b expression by RT-qPCRin neuroblastic (Nb) and Schwannian stromal (SS) cell populations isolatedby Laser Capture Microdissection. Isolated SS express CXCL13 that isalmost undetectable in Nb cells. The receptor was found expressed inisolated neuroblastic cells of 2750_Nb sample and at very low levels in theremaining Nb samples but not detected in SS cells. (MNE, Meannormalized expression).






in neuroblasts expressing CXCR5. Moreover, migrationof LA1–5S neuroblastoma cells was significantly suppressedby the addition of a neutralizing anti-CXCL13 antibody.On the contrary, CXCR5-negative ACN cells showed a

more limited migration than LA1–5S in response to theconditioned medium of Schwannian-like cells and theirmigration was not inhibited by anti-CXCL13 antibody(Fig. 5C).

E F

G H

I L

C D

A B

Figure 3. CXCL13 and CXCR5protein expression in tumors andneuroblastoma cell lines. CXCR5and CXCL13 expression in GNBintermixed stroma-rich andneuroblastoma stroma-poortumors and in neuroblastomacells detected byimmunohistochemistry (A–D andI–L) and immunofluorescence(E–H). A shows hematoxylin andeosin staining of GNB intermixedstroma-rich; in B Schwannianstromal cells are markedly stainedby GFAP (Glial fibrillary acidicprotein) antibody (Cell MarqueInc.); the same cells are positivefor CXCL13 (C) but negative forCXCR5 (D). Detectable CXCR5protein expression is observed inneuroblastoma stroma-poor byimmunofluorescence (E and F;magnification 40� and 100�).CXCR5 expression is clearlydetected in SH-SY-5Y and LA1–5S cells by immunofluorescence(G and H) andimmunohistochemistry (I and L).Fluorescent positive signals areevident on the cell membrane andin the cytoplasm of some cells,indicating the internalization ofCXCR5 (G and H). The arrows in Iand L indicate the CXCR5 positiveedge of SH-SY-5Y and LA1–5Scells. An AxioImager M1microscope with light andfluorescence lamp (Zeiss Inc.) wasused for cells and tumor tissueobservation. Images werecaptured by Image capturesystem AxioVision Release 4.6.

Del Grosso et al.





Discussion

Chemokines play several roles in both physiologic andpathologic conditions.Hereby,we showthatCXCL13and itsreceptor CXCR5 are involved in the relation between neuro-blastic and Schwannian stromal cells of neuroblastic tumors, aheterogeneous groupof pediatric cancers. Several chemokineshave been found expressed in neuroblastoma; in particular,the role of the CXCR4-CXCL12 axis in neuroblastomacells proliferation, survival, and bone marrow disseminationis still controversial (13, 14, 22). Airoldi and colleagues (15)showed that neuroblastoma cells express CXCR5 and theyhypothesized that malignant cells are able to migrate even inthe bone marrow in response to CXCL13 stimulation.We confirm that neuroblastoma cells express CXCR5

receptor and we add a detailed analysis of CXCL13 andCXCR5 expression in neuroblastic tumors.Our results showthat only malignant neuroblasts express CXCR5 protein andwe show, for the first time, that CXCL13 is also secreted bySchwannian stromal cells of neuroblastic tumors.

Neuroblastic tumors stroma-rich tumors are usuallycomposed of abundant stroma and less amount of neuro-blastic cells; these tumors onset as a localized mass and rarelyshow metastatic disease. Patients with stroma-rich tumorare at low risk of relapse after complete surgical tumorresection and they usually have a good outcome. Because weobserved that CXCL13 is expressed by Schwannian stromalcells, and neuroblastic cells express CXCR5 we hypothesizethat CXCL13-CXCR5 axis may contribute to retain malig-nant CXCR5-positive neuroblasts in stroma-rich tumorslimiting their metastatic spreading. Such hypothesis isstrongly supported by the ability of neuroblastoma cellsto migrate toward a source of hrCXCL13. Moreover, ourdata indicate that the CXCR5bmRNA variant encodes for afunctionally active CXCR5b protein, which is capable tospecifically respond to the CXCL13 stimulus.The role of Schwannian stromal cells in neuroblastic

tumors has been widely debated. Ambros and colleagues(4) suggested that these cells secrete factors that limit theproliferation of malignant neuroblasts and Chienski and

SH-SY5Y SK-N-BE(2)c LA1-5S

SH-SY5Y+ 10 µmol/L BrdU

1.0E-04S100A6

CXCL13

CXCL13 in NB conditioned medium

100.0

75.0

50.0C

XC

L13

(pg/

mL)

25.0

0.0

8.0E-05

6.0E-05

MN

E

4.0E-05

2.0E-05

0.0E+00

2.5E-07

2.0E-07

3.0E-07

B

C

A

1.5E-07

MN

E

1.0E-07

5.0E-08

0.0E+00

SH-SY5Y

SH-SY5Y

SH-SY5Y

Brd

U

SK-N-B

E(2)c

SK-N-B

E(2)c

Brd

U

LA1-

5S

LA1-

5S B

rdU

SH-SY5Y

2W

SH-SY5Y

1W

SH-SY5Y

3W

SK-N-B

E(2)c

SK-N-B

E(2)c

1W

SK-N-B

E(2)c

2W

SK-N-B

E(2)c

3W

LA1-

5S

LA1-

5S 1

W

LA1-

5S 2

W

LA1-

5S 3

W

SH-SY5Y

SH-SY5Y

2W

SH-SY5Y

1W

SH-SY5Y

3W

SK-N-B

E(2)c

SK-N-B

E(2)c

1W

SK-N-B

E(2)c

2W

SK-N-B

E(2)c

3W

LA1-

5S

LA1-

5S 1

W

LA1-

5S 2

W

LA1-

5S 3

W

saliv

a

SK-N-BE(2)c+ 10 µmol/L BrdU

LA1-5S+ 10 µmol/L BrdU

Figure 4. Effects of treatment with 50-bromo-20-deoxyuridine (BrdU) on morphologic features and S100A6 mRNA expression (A) and CXCL13 mRNAexpression (B) and detection of secreted CXCL13 protein (C) in different neuroblastoma (NB) cell lines. SH-SY5Y, SK-N-BE(2)c, and LA1–5S cells belonging toNeuroblastic, Intermediate, and Substrate-adherent neuroblastoma subtypes, respectively, were grown in presence of 10 mmol/L BrdU to induce glialdifferentiation and obtain a model for Schwannian-like cell. A shows the effect of BrdU on morphological features after 3 weeks of treatment: large flatteningand subadherent cell shape is clearly shown in SH-SY5Y and SK-N-BE(2)c cells but less evident in LA1–5S cells. Morphologic changes correlate with theinduction of S100A6 (calcyclin) by RT-qPCR in SH-SY5Y, SK-N-BE(2)c, and LA1–5S cells after treatment with 10 mmol/L BrdU for 1, 2, or 3 weeks, indicatingthe induction of cells toward glial-cell lineage. A corresponding CXCL13 mRNA expression is observed in SK-N-BE(2)c and LA1–5S cell lines but not in SH-SY5Y (B). C shows the amount of secreted CXCL13 protein in cell conditioned media collected for 3 weeks. Detectable secretion of the chemokine in LA1–5Sand SK-N-BE(2)c conditioned media is depending on treatment with BrdU. (W, week; MNE, Mean normalized expression).






colleagues (7) indicated that Schwannian stromal cellsrelease other factors that modulate angiogenesis. Here,we show that CXCL13 is produced by Schwannian stromalcells adding new information about the factors released bystroma of neuroblastic tumors and we suggest a relevant roleof the CXCL13/CXCR5b axis in the crosstalk betweenneuroblasts and stromal cells. The novel discovery thatSchwannian stromal cells express CXCL13 supports theprevious findings that cells of neural origin are able toproduce CXCL13 (23, 24).As an in vitro model to test the possible involvement of

CXCL13 in the crosstalk betweenmalignant neuroblasts andSchwannian stromal cells we induced glial cell differentiationof neuroblastoma cells. Neuroblastoma cell lines treated withBrdU induce immature neuroblasts to a more mature phe-notype with a large and subadherent Schwannian-like mor-phology (19). After BrdU treatment, neuroblastoma cellsstop to proliferate and produce calcyclin (25), a glial, andSchwannian cells marker. We show for the first time, thatneuroblastoma cells, which possess a bilineage potential (19),express CXCL13 only after induction to a glial phenotype.Furthermore, we showed that LA1–5S neuroblastoma cellsare able to migrate under the effect of the conditioned

medium of BrdU differentiated neuroblastoma cells and thatthis effect is blocked by CXCL13-neutralizing antibodies.Overall data indicate that differentiated neuroblasts pheno-typically similar to Schwannian stromal cell are able to secreteCXCL13 and to exert a chemotactic effect on CXCR5-positive neuroblastoma cells, whereas CXCR5-negative neu-roblastoma cells showed limited migration possible due to aslight effect of other chemotactic factors produced by Schwan-nian-like cells. Collectively, our results strongly indicate aclose relation between Schwannian stromal cells and neuro-blastic cells mediated via a CXCL13–CXCR5 interaction.This observation is enforced by the finding that anti-CXCL13antibody significantly inhibits the migration of CXCR5-positive neuroblastoma cells in response to the conditionedmedium. Hence, in addition to the other factors that exerttheir effects on neuroblast survival and differentiation, and ontumor angiogenesis (4, 6, 7), our finding that Schwannian-like cells are able to secrete a chemotactic factor acting onneuroblastic cell migration adds insight on a broad role ofSchwannian stromal cells in stroma-rich neuroblastic tumors.In conclusion, ourmodel suggests thatCXCL13 produced

by Schwannian stroma may contribute to limit dissemina-tion of CXCR5-expressing neuroblasts outside the SR

700

600

500

400

Mig

rate

d ce

lls/fi

eld

Mig

rate

d ce

lls/fi

eld

Mig

rate

d ce

lls/fi

eld

60

50

40

30

20

10

0

605040302010

0

60

80

100

40

20

0

Mig

rate

d ce

lls/fi

eld

300

200

100

0SFM CXCL13

100 ng/mLCXCL13

200 ng/mL

SFM CM CMCXCL13200 ng/mL

+ anti-CXCR5

+ anti-CXCL13

+ anti-CXCL13

+ anti-CXCL13

SFM CXCL13100 ng/mL

P < 0.007

P < 0.0001

P < 0.0001

P < 0.0001P < 0.0001

P < 0.0001

P = ns

P < 0.0001

SH-SY5YA

B C

LA1-5S

LA1-5SLA1-5S

ACN

P < 0.0003

CXCL13200 ng/mL

Figure 5. CXCL-13 mediates neuroblastoma cells migration. A shows the number of migrating SH-SY5Y and LA1–5S neuroblastoma cells after6 hours of incubation at two different doses (100 ng/ml, 200 ng/ml) of hrCXCL13 added in the lower chamber. A low number of cells migrate in the presence ofserum free medium (SFM). B shows that CXCR5 positive LA1–5S migration is specifically mediated by the CXCR5/CXCL13 interaction by inhibitingboth the receptor and the chemokine with anti-CXCR5 (15 ng/ml) and anti-CXCL13 (3.5 ng/ml) blocking antibody, respectively. C showsCXCR5-positive LA1–5S and CXCR5-negative ACN neuroblastoma cells chemotaxis in the presence of concentrated serum-free conditioned-medium ofBrdU-differentiated neuroblastoma cells secreting CXCL13. Migration of LA1–5S cells is significantly reduced by anti-CXCL13 (0.5 ng/ml) blockingantibody added to the conditioned medium. ACN cells showed limited migration that was not influenced by anti-CXCL13 antibody. Statistical analysiswas carried out by unpaired two-tailed t-test. (X-axis shows the different chemoattractants, y-axis the number of cells migrated/field. ns: not significant).

Del Grosso et al.





tumors, whereas CXCR5-positive neuroblastic cells ofstroma-poor tumors more easily disseminate to CXCL13-producing distal sites, such as the bonemarrow. Althoughwecannot exclude the influence of CXCL13 released fromdistalbone marrow sites on neuroblastona cells, this effect appearsunlikely in the case of localized tumors. Indeed, very lowconcentrations of CXCL13 are usually present in the sys-temic circulation, whereas the Schwannian stromal cells andmalignant neuroblasts are in close contact thus facilitatinglocal CXCL13-CXCR5 interaction. Obviously, this modelcannot be the only reason for the different biological andclinical behavior of SR and SP tumors, which is predomi-nantly related to the different genetic and molecular char-acteristics observed in localized and metastatic tumors (26).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are grateful to surgeons, clinicians, pathologists of the Associazione ItalianaEmatologia Oncologia Pediatrica (AIEOP) for providing tumor samples, toDr. Alessandro De Ambrosis for the helpful discussions and to Silvia De Luca(Fondazione Italiana per la Lotta al Neuroblastoma) for language revision.

Grant Support

The work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC,Project IG 5331, 2008), Fondazione Italiana per la Lotta al Neuroblastoma,Ministero dell’Universit�a, Ricerca Scientifica e Tecnologica, Italy. M. Croce andL. Longo are supported by F. Italiana per la Lotta al Neuroblastoma, F. Valdora is afellow of the University of Genoa.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 11, 2010; revised April 26, 2011; accepted May 25, 2011;published OnlineFirst June 3, 2011.

References1. Maris JM, Hogarty MD, Bagatell R, Cohn SL. Neuroblastoma. Lancet

2007;369:2106–20.2. Shimada H, Ambros IM, Dehner LP, Hata J, Joshi VV, Roald B.

Terminology and morphologic criteria of neuroblastic tumors: recom-mendations by the International Neuroblastoma Pathology Commit-tee. Cancer 1999;86:349–63.

3. Shimada H, Umehara S, Monobe Y, Hachitanda Y, Nakagawa A, GotoS, et al. International neuroblastoma pathology classification forprognostic evaluation of patients with peripheral neuroblastic tumors.Cancer 2001;92:2451–61.

4. Ambros IM, Zellner A, Roald B, Amann G, Ladenstein R, Printz D, et al.Role of ploidy, chromosome 1p, and Schwann cells in the maturationof neuroblastoma. N Engl J Med 1996;334:1505–11.

5. Coco S, Defferrari R, Scaruffi P, Cavazzana A, Di Cristofano C,Longo L, et al. Genome analysis and gene expression profiling ofneuroblastoma and ganglioneuroblastoma reveal differencesbetween neuroblastic and Schwannian stromal cells. J Pathol 2005;207:346–57.

6. Huang D, Rutkowski JL, Brodeur GM, Chou PM, Kwiatkowski JL,Babbo A, et al. Schwann cell-conditioned medium inhibits angiogen-esis. Cancer Res 2000;60:5966–71.

7. Chlenski A, Liu S, Crawford SE, Volpert OV, DeVries GH, Evange-lista A, et al. SPARC is a key schwannian-derived inhibitor control-ling neuroblastoma tumor angiogenesis. Cancer Res 2002;62:7357–63.

8. Albino D, Scaruffi P, Moretti S, Coco S, Truini M, Di Cristofano C, et al.Identification of low intratumoral gene expression heterogeneity inneuroblastic tumors by genome-wide expression analysis and gametheory. Cancer 2008;113:1412–22.

9. Bonecchi R, Galliera E, Borroni EM, Corsi MM, Locati M, Mantovani A.Chemokines and chemokine receptors: an overview [review]. FrontBiosci 2009;14:540–51.

10. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer2004;4:540–50.

11. Savino Mantovani A, Locati M, Zammataro L, Allavena P, Bonecchi R.The chemokine system in cancer biology and therapy [review]. Cyto-kine Growth Factor Rev 2010;21:27–39.

12. Slettenaar VIF, Wilson JL. The chemokine network: a target in cancerbiology? Adv Drug Deliv Rev 2006;58:962–74.

13. Gross N, Meier R. Chemokines in neuroectodermal cancers: Thecrucial growth signal from the soil. Semin Cancer Biol 2009;19:103–10.

14. Raffaghello L, Cocco C, Corrias MV, Airoldi I, Pistoia V. Chemokines inneuroectodermal tumour progression and metastasis. Semin CancerBiol 2009;19:97–102.

15. Airoldi I, Cocco C, Morandi F, Prigione I, Pistoia V. CXCR5 may beinvolved in the attraction of human metastatic neuroblastoma cells tothe bone marrow. Cancer Immunol Immunother 2008;57:541–8.

16. Meier R, M€uhlethaler-Mottet A, Flahaut M, Coulon A, Fusco C,Louache F, et al. The chemokine receptor CXCR4 strongly promotesneuroblastoma primary tumour and metastatic growth, but not inva-sion. PLoS ONE 2007;2:e1016.

17. Lefever S, Hellemans J, Pattyn F, Przybylski DR, Taylor C, Geurts R,et al. RDML: structured language and reporting guidelines for real-time quantitative PCR data. Nucleic Acids Res 2009;37:2065–9.

18. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M,et al. The MIQE Guidelines: Minimum Information for publication ofQuantitative Real-TimePCRexperiments. Clin Chem2009;55:611–22.

19. Acosta S, Lavarino C, Paris R, Garcia I, de Torres C, Rodríguez E, et al.Comprehensive characterization of neuroblastoma cell lines subtypesreveals bilineage potential similar to neural crest stem cells. BMC DevBiol 2009;9:12.

20. Albini A, Benelli R, Giunciuglio D, Cai T, Mariani G, Ferrini S, et al.Identification of a novel domain of HIV Tat involved in monocytechemotaxis. Biol Chem 1998;273:15895–900.

21. Barella L, Loetscher M, Tobler A, Baggiolini M, Moser B. Sequencevariation of a novel heptahelical leucocyte receptor through alternativetranscript formation. Biochem J 1995;309:773–9.

22. Airoldi I, Raffaghello L, Piovan E, Cocco C, Carlini B, Amadori A, et al.CXCL12 does not attract CXCR4þ human metastatic neuroblastomacells: clinical implications. Clin Cancer Res 2006;12:77–82.

23. van de Pavert SA, Olivier BJ, Goverse G, Vondenhoff MF, Greuter M,Beke P, et al. Chemokine CXCL13 is essential for lymph node initiationand is induced by retinoic acid and neuronal stimulation Nat Immunol2009;10:1193–9.

24. Ramesh G, Borda JT, Gill A, Ribka EP, Morici LA, Mottram P, et al.Possible role of glial cells in the onset and progression of Lymeneuroborreliosis. J Neuroinflammation 2009;6:23.

25. Tonini GP, Casalaro A, Cara A, Di Martino D. Inducible expression ofcalcyclin, a gene with strong homology to S-100 protein, duringneuroblastoma cell differentiation and its prevalent expression inSchwann-like cell lines. Cancer Res 1991;51:1733–7.

26. Maris JM. Recent Advances in Neuroblastoma. N Engl J Med 2010;362:2202–11.






2011;9:815-823. Published OnlineFirst June 3, 2011.Mol Cancer Res Federica Del Grosso, Simona Coco, Paola Scaruffi, et al. Neuroblastic TumorsNeuroblastoma Cells and Schwannian Stromal Cells in Role of CXCL13-CXCR5 Crosstalk Between Malignant

Updated version

10.1158/1541-7786.MCR-10-0367doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2011/06/03/1541-7786.MCR-10-0367.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/9/7/815.full#ref-list-1

This article cites 26 articles, 6 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://mcr.aacrjournals.org/content/9/7/815To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-10-0367

http://mcr.aacrjournals.org/content/suppl/2011/06/03/1541-7786.MCR-10-0367.DC1

http://mcr.aacrjournals.org/content/9/7/815.full#ref-list-1

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/9/7/815


role of cxcl13-cxcr5 crosstalk between malignant ... · neuroblastoma is a stroma-poor (sp)...

Documents