PRECLINICAL STUDY

CXCL13–CXCR5 co-expression regulates epithelialto mesenchymal transition of breast cancer cells during lymphnode metastasis

Subir Biswas • Suman Sengupta • Sougata Roy Chowdhury • Samir Jana •

Gunjan Mandal • Palash Kumar Mandal • Nipun Saha • Vivek Malhotra •

Arnab Gupta • Dmitry V. Kuprash • Arindam Bhattacharyya

Received: 8 November 2013 / Accepted: 4 December 2013

� Springer Science+Business Media New York 2013

Abstract We investigated the expression of –CXC che-

mokine ligand 13 (CXCL13) and its receptor –CXC che-

mokine receptor 5 (CXCR5) in 98 breast cancer (BC) patients

with infiltrating duct carcinoma, out of which 56 were found

lymph node metastasis (LNM) positive. Interestingly, co-

expression of CXCL13 and CXCR5 showed a significant

correlation with LNM. Since, epithelial to mesenchymal

transition (EMT) is highly associated with metastasis we

investigated EMT-inducing potential of CXCL13 in BC cell

lines. In CXCL13-stimulated BC cells, expression of various

mesenchymal markers (Vimentin, N-cadherin), EMT regu-

lators (Snail, Slug), and matrix metalloproteinase-9 (MMP9)

was increased, whereas the expression of epithelial marker

E-cadherin was found to be decreased. In addition, expres-

sion of receptor activator of nuclear factor kappa-B ligan-

d (RANKL), which is known to regulate MMP9 expression

via Src activation, was also significantly increased after

CXCL13 stimulation. Using specific protein kinase inhibi-

tors, we confirmed that CXCL13 stimulated EMT and MMP9

expression via RANKL–Src axis in BC cell lines. To further

validate this observation, we examined gene expression

patterns in primary breast tumors and detected significantly

higher expression of various mesenchymal markers and

regulators in CXCL13–CXCR5 co-expressing patients.

Therefore, this study showed the EMT-inducing potential of

CXCL13 as well as demonstrated the prognostic value of

CXCL13–CXCR5 co-expression in primary BC. Moreover,

CXCL13–CXCR5–RANKL–Src axis may present a thera-

peutic target in LNM positive BC patients.

Keywords Lymph node metastasis � Epithelial to

mesenchymal transition � CXCL13 � CXCR5 � RANKL �Breast cancer

Abbreviations

ANOVA Analysis of variance

AP Alkaline phosphatase

BC Breast cancer

BCIP 5-Bromo-4-chloro-30-indolyphosphate

BSA Bovine serum albumin

CXCL13 –CXC chemokine ligand 13

CXCR5 –CXC chemokine receptor 5

DAB 3-30-Diaminobenzidine

DAPI 40,6-Diamidino-2-phenylindole

Suman Sengupta, Sougata Roy Chowdhury, and Samir Jana

contributed equally to this work.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10549-013-2811-8) contains supplementarymaterial, which is available to authorized users.

S. Biswas � S. Sengupta � S. Jana � G. Mandal �A. Bhattacharyya (&)

Immunology Laboratory, Department of Zoology, University of

Calcutta, 35, Ballygunge Circular Road, Kolkata 700019,

West Bengal, India

e-mail: arindam19@yahoo.com

S. Roy Chowdhury

Materials Science Centre, Indian Institute of Technology

Kharagpur, Kharagpur, India

P. K. Mandal

Department of Pathology, North Bengal Medical College,

Darjeeling, India

N. Saha � V. Malhotra � A. Gupta

Department of Surgical Oncology, Saroj Gupta Cancer Centre

and Research Institute, Kolkata, India

D. V. Kuprash

Laboratory of Immunoregulation, Engelhardt Institute of

Molecular Biology, Russian Academy of Sciences, Moscow,

Russia

123

Breast Cancer Res Treat

DOI 10.1007/s10549-013-2811-8

DMEM Dulbecco’s modified Eagle’s medium

DPX Distyrene plasticizer and xylene

E-cadh E-cadherin

ECM Extracellular matrix

EMT Epithelial to mesenchymal transition

ER Estrogen receptor

FAK Focal adhesion kinase

FBS Foetal bovine serum

FITC Fluorescein isothiocyanate

HRP Horse-radish-peroxidase

IDC Infiltrating duct carcinoma

IHC Immunohistochemistry

LNM Lymph node metastasis

MMLV Moloney murine leukemia virus

MMP2 Matrix metalloproteinase-2

MMP9 Matrix metalloproteinase-9

MRM Modified radical mastectomy

NBT Nitro-blue tetrazolium

N-cadh N-cadherin

p Probability

PBS Phosphate buffered saline

PR Progesterone receptor

p-Src Phosphorylated-Src

PVDF Polyvinylidine difluoride

RANKL Receptor activator of nuclear factor

kappa-B ligand

RIPA Radio-immunoprecipitation assay

SDS-PAGE Sodium dodecyl sulfate-polyacrylamide gel

electrophoresis

SGCC & RI Saroj Gupta Cancer Centre and Research

Institute

SD Standard deviation

t-Src Total-Src

TNF Tumor necrosis factor

WB Western blot

Introduction

Chemokines are low molecular weight (8–10 kDa) che-

motactic cytokines, classified into four highly conserved

groups—CXC, CC, C, CX3C—based on the pattern of

their N-terminal cysteines [1]. They signal through binding

to their corresponding G-protein coupled receptors on tar-

get cells [1–5]. The primary functions of CXC-chemokines

are chemoattraction and activation of leukocytes in diverse

immunological responses [1]. Surprisingly, increasing

evidence suggests that CXC-chemokine ligands and their

corresponding receptors also appear to play significant role

in neoplastic transformation, cancer cell migration, inva-

sion, and metastasis [6–9].

To date, human cancers have been demonstrated to

utilize a complex chemokine network that influences tumor

cell growth, survival, migration, angiogenesis, and metas-

tasis [9–13]. Metastatic cancer cells first infiltrate into the

nearby lymph nodes, thus, LNM is directly associated with

poor prognosis of cancer [14]. Expression of CXCR4 has

been established as a prognostic marker for many cancer

types [15–20]. Increasing evidence shows that CXCR4 and

CXCL12 may play a critical role in the organ specific

metastasis in various cancer types, particularly in BC [12,

13, 15–20]. Other CXC chemokines, such as CXCL1,

CXCL8, and CXCR2, act as autocrine growth factors for

melanoma and other cancers [21–23]. CXCL8, which is a

key factor in BC invasion and angiogenesis, is associated

with tumor size and ER status [24].

B cell chemoattractant CXCL13 and its receptor

CXCR5 have gained interest after some initial findings

with different cancers including BC [25–29]. Overexpres-

sion of CXCL13, both in the tumor tissues and in the

peripheral blood of BC patients, has already been reported

[27]. Moreover, CXCL13-stimulated migration and inva-

sion are PI3Kp110a, Src, and FAK dependent [30]. Based

on this information, we focused on the prognostic value of

CXCL13–CXCR5 axis by comparing expression status of

the receptor-ligand pair with different clinicopathological

features. Significant correlation between their co-expres-

sion and LNM led us to perform additional in vitro

experiments.

EMT is an essential process highly associated with cancer

metastasis [31]. It increases the chance of metastasis and

invasion in different cancer types including BC [32]. During

EMT and tumor invasion, epithelial markers such as E-cadh,

cytokeratins are down-regulated, whereas mesenchymal

markers such as Vimentin, N-cadh are up-regulated [33, 34].

Thus, loss of E-cadh could be considered as a hallmark of

cancer metastasis [34]. Some transcription factors are also

found to regulate EMT, such as Snail, Slug, Twist, etc., [35]

and tumor cells undergoing EMT are known to secrete

MMPs, which degrade the structural components of the

ECM to facilitate tumor cell migration [1].

RANKL, a member of TNF-family, is essential for

lactating mammary gland development during pregnancy

[36]. RANKL increases c-Src phosphorylation and thereby

activates c-Src-Akt and c-Src-ERK signaling pathways and

it is associated with increased migration of BC cells [37].

Furthermore, stimulation with RANKL increases MMP9

expression [38]. In our study, we made an attempt to

establish the relationship between CXCL13–CXCR5 axis

and RANKL during EMT induction in vitro. This study is

novel because it highlights the prognostic value of

CXCL13 and CXCR5 co-expression, rather than sole

expression of either, in support for an autocrine signaling

by the tumor itself.

Breast Cancer Res Treat

123

Materials and methods

Patients, tumor tissues and normal breast tissues

Tumor samples and their corresponding normal tissues were

collected from 98 BC patients who underwent MRM surgery

at SGCC & RI, Kolkata, India. Permission was obtained from

Institutional Ethics Committee of SGCC & RI, and informed

consents were taken from patients. Grading and staging of

tumor samples were done according to UICC TNM classifi-

cation (Table 1; ESM_1). Freshly operated tissues were

immediately processed for RNA extraction using TRIzol

reagent (Invitrogen, CA, US) and rest were used for protein

extraction using RIPA buffer containing protease inhibitor

cocktail (Cell Signaling Technology, Inc., US).

Cell lines and culture

The BC cell lines MDA-MB-231 and T-47D were obtained

from NCCS, Pune, India. MDA-MB-231 is known to be an ER/

PR negative cell line, whereas T-47D is known to be an ER/PR

positive one [39]. They were cultured in DMEM (Gibco,

Invitrogen, CA, US), supplemented with 10 % FBS (Gibco,

Invitrogen, CA, US), 100 U ml-1 penicillin and streptomycin

(Gibco, Invitrogen, CA, US), and maintained in a cell culture

incubator at 37 �C with 5 % CO2. For CXCR5 overexpression

and vector control, cloned ORF of CXCR5 (OriGene Tech-

nologies, Inc., US) and empty PCMV6XL4 vector (OriGene

Technologies, Inc., US) was transfected, respectively, using

Lipofectamine 2000TM (Invitrogen, CA, US). Transfection

was done 24 h before CXCL13 treatment. Both the transfected

and untransfected cell lines were cultured for 12 h in low FBS

(2 %) medium before stimulation with purified recombinant

CXCL13 (PeproTech) at laboratory optimized concentration

(50 ng-mL). CXCL13-stimulated cells were incubated for 24 h

in the presence or absence of anti-CXCR5 antibody (Epito-

mics, CA, US) and small molecule inhibitors of PI3Kp110a(PI-103; Echelon Biosciences, Inc., US; 3 lM), Src (SU6656;

Sigma Aldrich, US; 5 lM). The different treatment conditions

that were used are Control, Vector control transfected (VC Tr),

Ligand treated (LT), Vector control transfected ? Ligand

treated (VC Tr ? LT), Receptor overexpressed (ROE), and

Receptor overexpressed ? Ligand treated cells (ROE ? LT),

where ligand means recombinant CXCL13, vector control

means empty pCMV6-XL4 vector (ESM_2).

Reverse transcription PCR

The reverse transcription reaction from 2 lg total RNA was

done in 20 ll reaction volume using random hexamer

(Thermo Scientific) and MMLV reverse trancriptase (Epi-

centre Biotechnologies). PCR amplifications were done for 35

cycles on ABi 2720 Thermal cycler using 500 ng of first

strand c-DNA. Primer sequences are listed in Table 2. Neg-

ative controls without c-DNA for all primer pairs were taken.

18s r-RNA was taken as internal control in all PCR reactions.

All the PCR experiments were performed at least thrice.

Real-time PCR

PCR amplifications were performed in 40 cycles on ABi

StepOnePlusTM using 100 ng of first strand c-DNA. Rela-

tive mRNA expression quantities (DDCt) were obtained by

normalization against 18s r-RNA. All the PCR experiments

Table 1 Clinicopathological characteristics of patients with expres-

sion grades of CXCL13 and CXCR5

Characteristics Number of patients per group

Total 98

Age (median and range) in years 47 (24–75)

pT status

pT1 18

pT2 46

pT3 26

pT4 8

pN status

pN0 42

pN1 15

pN2 24

pN3 17

M status

M0 83

M1 8

MX 7

Stage

I 18

II 39

III 33

IV 8

Tumor differentiation grade

Well (I) 14

Moderate (II) 45

Poor (III) 39

CXCL13 expression grade

0 26

I 33

II 29

III 10

CXCR5 expression grade

0 37

I 26

II 27

III 8

Breast Cancer Res Treat

123

were performed at least thrice and the average values were

considered for analyses.

Western blotting

Samples of lysates containing 60 lg of denatured total

protein were resolved and the blots were transferred on to

PVDF membrane, incubated overnight at 4 �C with pri-

mary antibodies: anti-b-actin (Santa Cruz Biotechnology,

CA, US), anti-CXCR5, anti-N-cadh, anti-E-cadh, anti-

Vimentin, anti-RANKL, anti-Src, anti-p-Src (Epitomics,

CA, US), anti-Snail, anti-Slug, and anti-CXCL13 (Abcam).

After incubating with AP-conjugated secondary antibodies,

bands were developed using the developer solution i.e.,

NBT and BCIP with proper ratio in AP buffer.

Gelatin zymography

Gelatin zymography was performed in 10 % SDS-PAGE,

containing 0.1 % gelatin under non-reducing conditions to

determine the enzymatic activity of MMP2 and MMP9 in

control as well as ROE ? LT MDA-MB-231 and T-47D

cells. Native proteins were electrophoresed followed by

washing the gels twice in 2.5 % Triton X-100 each for

30 min at room temperature. The gels were then incubated

overnight at 37 �C in substrate buffer containing 50 mM

Tris–Hcl pH 7.4, 5 mM CaCl2, 1 lM ZnCl2 and stained with

0.5 % Coomassie G250 in 30 % Ethanol, 10 % glacial acetic

acid for 30 min. The gels were then destained by destaining

solution containing 30 % ethanol and 10 % acetic acid.

Wound healing

Confluent ROE ? LT and control cells in 6-well plates

were scratch-wounded using sterile micro-tip, just prior to

CXCL13 treatment and incubated for 24 h in the medium

with 2 % FBS. Wounded monolayer cells were washed

thrice by PBS. The speed of wound closure was photo-

graphed after 24 h with a phase contrast microscope (Zeiss

Aixovert 40C) at 5X. Each experiment was performed in

triplicates and repeated thrice.

Immunofluorescence

8 9 104 Cells were seeded on sterile cover slips in 6-well

plates. Transfected cells were stimulated 12 h after trans-

fection with CXCL13 (PeproTech; 50 ng/mL) for 24 h.

Both control and treated cells were fixed with 4 % para-

formaldehyde in PBS for 20 min at room temperature,

followed by permeabilization using 0.2 % Triton X-100 for

10 min and blocked for 60 min with PBS containing 3 %

BSA. Subsequently, cells were incubated with anti-

Vimentin or anti-Snail antibodies overnight at 4 �C, and

then washed in PBS. Cells were then incubated with FITC

conjugated anti-rabbit secondary antibody (Abcam) for

60 min at room temperature. Nuclei were counterstained

with DAPI (1 lg/mL) for 5 min and mounted on glass

slides. Finally, slides were examined under Fluorescence

microscope (Olympus BX51).

Immunohistochemistry

Immunohistochemistry for CXCR5, CXCL13, and RANKL

was performed on formalin-fixed, paraffin-embedded tissue

sections of 3–5 lM. The sections were de-paraffinized,

rehydrated, blocked, and incubated overnight with primary

antibodies at 4 �C. HRP-conjugated anti-rabbit secondary

antibodies were added at 1:250 dilutions. The slides were

developed using DAB chromogen and counterstained with

hematoxylin. The protein levels were scored as 0, 1, 2, and 3

Table 2 Summary of PCR primers

Name Forward primer (50 ? 30) Reverse primer (50 ? 30) Product size (bp)

CXCL13 GAGGCAGATGGAACTTGAGC CTGGGGATCTTCGAATGCTA 158

CXCL13 TGTGGACCCTCAAGCTGAATGGAT TTGCCTCTTGACAGGCTCAAGTTC 565

CXCR5 GCACCTCCCATCCTAATCATC CTAAGCTGATGGAGTGTGTTCT 103

CXCR5 TACCTGCAAGCTGAATGGCTCTCT GGTGCTCAGCTTCTGGCTTTGTTT 706

N-cadherin CAGTGCAGTCTTATCGAAGG GGCTCACTGCTCTCATATTG 105

E-cadherin CCCGGGACAACGTTTATT GTCGTTACGAGTCACTTCAG 116

Slug CTCTCCTCTTTCCGGATACT GCTTGGACTGTAGTCTTTCC 131

Snail TATTTCAGCCTCCTGTTTGG GAATAGTTCTGGGAGACACATC 146

Vimentin CTCGTCACCTTCGTGAATAC GATTAGTTTCCCTCAGGTTCAG 153

MMP2 AGAACCTCAGGGAGAGTAAG GGAAGCAAACCTCGAACA 106

MMP9 TGGGCTACGTGACCTATGA CCCTTTCCTCCAGAACAGAATAC 152

RANKL CATCCCATCTGGTTCCCATAAA CTCTGTAGCTAGGTCTCCTGAA 181

18s r-RNA GTAACCCGTTGAACCCCATT CCATCCAATCGGTAGTAGCG 151

Breast Cancer Res Treat

123

for negative, weak, moderate, and high expressions,

respectively, after three independent observations.

Statistical analyses

Fisher’s exact test was used to analyze association between

gene expression and clinicopathological characteristics of

the patients. Kolmogorov–Smirnov test was used to iden-

tify distribution of variables. Mann–Whitney U test was

performed to examine differences in fold change of gene

expressions between healthy tissues and tumor samples

(CXCL13–CXCR5 co-expressed and non-co-expressed).

One-way ANOVA (Bonferroni correction) was performed

to assess the level of significance among paired datasets in

terms of fold change of mRNA and protein expressions in

cell lines. Statistical analysis was performed using SPSS

Statistics 17.0 and OriginPro8. All data are presented as

mean ± SD, and p value of \0.05 was considered statis-

tically significant.

Results

Co-expression of CXCR5 and CXCL13 in the tumor

tissues of LNM positive BC patients

Expression of both CXCL13 and CXCR5 was assessed in

primary breast tumor samples from 98 BC patients with IDC

and associated autologous healthy breast tissues (Table 1;

ESM_1). We found that co-expression of the chemokine

ligand CXCL13 and its receptor CXCR5 was significantly

(p = 0.00002, Fisher’s exact test) associated with both stage

and LNM, but not with tumor size, tumor differentiation grade,

or menopausal status of the patient. The correlation between

LNM and co-expression of CXCL13–CXCR5 was especially

evident (Fig. 1a; Table 3). Figure 1b shows representative

IHC data for samples with various levels of CXCR5 and

CXCL13 expression. Our data imply possible existence of an

autocrine loop between CXCR5 and its ligand CXCL13 in

LNM positive BC patients.

Regulation of EMT markers expression by CXCL13–

CXCR5 axis

With increasing reports on the function of chemokine ligands

and receptors in the cancer development and progression, we

aimed to look at the CXCL13–CXCR5 signaling that might

help EMT. We, therefore, introduced the components of the

CXCL13–CXCR5 signaling axis into both ER/PR positive

and ER/PR negative cells and performed conventional as

well as real-time PCR analyses of CXCR5, N-cadh,

Vimentin, Snail, Slug, and E-cadh in these cells (Fig. 2a, b).

In both cell lines, we found increased mRNA levels of

N-cadh, Vimentin, Slug, and Snail only when both the

CXCL13 ligand and its receptor were present (ROE ? LT

cells), whereas E-cadh was observed significantly (p \ 0.05)

down-regulated (5.8 fold) in ROE ? LT T-47D cells

(Fig. 2a, b). A significant fold increase (p \ 0.001) in

CXCR5 mRNA level in both ROE and ROE ? LT cells

confirmed the functionality of pCMV6–XL4–CXCR5 con-

struct in both cell types (Fig. 2b). The RT-PCR results were

Fig. 1 Association between CXCL13 and CXCR5 co-expression and

lymph node metastasis. a CXCL13–CXCR5 co-expression within

breast tumors of LNM positive and LNM negative patients. The

number of patients: LNM positive, 56; LNM negative, 42 (see

Table 3). b Representative IHC pictures of BC samples showing high

(i and v) or moderate (ii and vi) or weak (iii and vii) or negative (iv

and viii) expression of CXCR5 and CXCL13, respectively

Table 3 Expression status of CXCL13 and CXCR5 markers in LNM

positive and LNM negative BC patients

Characteristics LNM negative

(pN0)

LNM positive

(pN1–3)CXCL13 CXCR5

2 2 17 0

1 2 17 3

2 1 2 7

1 1 6 46

Breast Cancer Res Treat

123

further confirmed by WB analysis (Fig. 2c, d) and immu-

nofluorescence, which also demonstrated increased expres-

sions of Vimentin and Snail upon CXCL13 stimulation

(Fig. 3a, b). We also observed that the shape of ROE ? LT

cells appeared to be more elongated than that of the control

cells of both cell types (Fig. 3). Altogether, these results

clearly indicated that CXCL13–CXCR5 induced EMT-

related changes in both ER/PR positive and negative BC cell

lines.

RANKL-dependent regulation of MMP9

The matrix metalloproteinase family proteins MMP2 and

MMP9 have long been associated with breast cancer EMT

and metastasis. We found increased mRNA expression of

MMP9 in ROE ? LT cells, whereas MMP2 expression

remained almost unchanged (Fig. 2a). The mRNA level of

MMP9 was significantly (p \ 0.01) increased (15.5 and

26.0-fold for MDA-MB-231 and T-47D cells, respectively)

in ROE ? LT cells as compared to control cells (Fig. 2b).

Moreover, significantly (p \ 0.05) increased activity of

MMP9 was observed for ROE ? LT cells in the zymogram

(Fig. 4). It has already been reported that RANKL regu-

lates Src activity and MMP9 expression. Therefore, we

performed expression analysis of RANKL and found that

ROE ? LT cells have significantly (p \ 0.01) increased

(26.0-fold, p = 0.0000013 and 29.5-fold, p = 0.0000042

for MDA-MB-231 and T-47D cells, respectively) (Fig. 2b).

Increased RANKL protein expression in CXCL13-stimu-

lated ROE ? LT cells was confirmed by WB analysis

(Fig. 2c, d). These data suggested that EMT-related chan-

ges in CXCL13-stimulated BC cells involve RANKL

Fig. 2 CXCL13 treated MDA-MB-231 and T-47D cells showed

EMT and up-regulation of MMP9 and RANKL. MDA-MB-231 and

T-47D cells were transfected with either CXCR5 or empty vector and

stimulated with (50 ng/mL) or without recombinant CXCL13.

a mRNA levels of Vimentin, Slug, Snail, E-cadh, N-cadh, CXCR5,

RANKL, MMP2, and MMP9 were assessed by conventional RT-

PCR, followed by agarose gel electrophoresis. 18s r-RNA was used as

internal control. b Quantitative real-time RT-PCR was performed for

the same set of genes. Fold changes are represented as relative values

normalized with 18s r-RNA and quantified. Both conventional and

real-time PCR showed up-regulation of Vimentin, Slug, Snail,

N-cadh, RANKL, and MMP9 in ROE ? LT cells. Fold increase

was as follows: Vimentin, 4.3-fold, p \ 0.05 and 11.4-fold, p \ 0.01

for MDA-MB-231, and T-47D cells, respectively; N-cadh, 25.4-fold,

p \ 0.01 and 34.5-fold, p \ 0.01 in MDA-MB-231, and T-47D cells,

respectively; Snail, 3.5-fold, p \ 0.05 and 11.3-fold, p \ 0.01 for

MDA-MB-231, and T-47D cells, respectively; Slug, 2.7-fold,

p \ 0.05 and 5.3-fold, p \ 0.05 for MDA-MB-231, and T-47D cells,

respectively; and MMP9, 15.5 and 26.0-fold for MDA-MB-231, and

T-47D cells, respectively. E-cadh was down-regulated in ROE ? LT

T-47D cells. Up-regulated CXCR5 mRNA expression was observed

in both ROE and ROE ? LT cells of both types. c Protein levels of

Vimentin, Slug, Snail, N-cadh, E-cadh, RANKL, and CXCR5 were

evaluated by immunoblotting. b-actin was used as loading control.

Up-regulation of Vimentin, Slug, Snail, N-cadh, and RANKL was

found in ROE ? LT MDA-MB-231 AND T-47D cells. E-cadh was

down-regulated in ROE ? LT T-47D cells. Up-regulated CXCR5

protein expression was found in ROE and ROE ? LT cells of both

BC cell types. d WB band densitometries were done and relative

intensities for those having expressions in control cell were calculated

and are shown in bar graphs. Results are representative of three

independent experiments performed in triplicate and are represented

as mean ± SD. One-way ANOVA (Bonferroni correction) was

performed, where significance level stands for * p B 0.05,

** p B 0.01, *** p B 0.001

Breast Cancer Res Treat

123

induction, which might as well be capable of regulating

MMP9 and its enzyme activity.

Increased migration and Src activation by CXCL13–

CXCR5 activation

Cellular migration is one of the important steps of cancer

metastasis and is known to be associated with EMT. We

performed wound healing assay in control and ROE ? LT

cells and found significantly (p \ 0.05 for T-47D and

p \ 0.001 for MDA-MB-231) increased wound closure after

24 h in ROE ? LT cells of both cell types, compared to

control cells (Fig. 5a, b). This suggested that CXCL13–

CXCR5-induced EMT is associated with increased cellular

motility and migration. Active (phosphorylated) Src i.e., p-Src

level was found to be significantly (p \ 0.05) increased in

CXCL13-stimulated (ROE ? LT) cells, whereas t-Src level

remained almost unchanged (Fig. 5c, d). It confirmed that

CXCL13–CXCR5-mediated signaling proceeds with phos-

phorylation of Src.

Src kinase-dependent CXCL13–CXCR5 signaling

and activation of EMT markers

We used pharmacological inhibitors of PI3Kp110a and

Src, i.e., PI-103 and SU6656, respectively, to check whe-

ther CXCL13-induced EMT in BC cell lines follows the

same pathway and compared with attenuation of CXCR5

Fig. 3 CXCL13-induced up-regulation of Vimentin and Snail and

change in morphology of a MDA-MB-231 and b T-47D cells.

Expression of Vimentin and Snail was monitored under fluorescence

microscope using FITC tagged specific antibodies. It shows increased

expression of both Vimentin and Snail in the ROE ? LT cells.

Moreover, shape of the ROE ? LT cells was found to be more

elongated than the control cells. Each image shown is representative

of 20 random fields observed. Indicated scale bars signify 10 lm

distance and photographs were taken at 1009 zoom. Results are

representative of three independent experiments performed in

triplicate

Fig. 4 ROE ? LT cells showed increased MMP9 enzyme activity in

gelatin zymography. Enzymatic activities of MMP9 and MMP2 were

evaluated for control and ROE ? LT cells of both BC cell types.

Relative intensities of the bands are shown as bar graphs. Increased

MMP9 activity was found for ROE ? LT cells. Results are

representative of three independent experiments performed in tripli-

cate and are represented as mean ± SD. One-way ANOVA (Bonfer-

roni correction) was performed, where significance level stands for

* p B 0.05, ** p B 0.01, *** p B 0.001

Breast Cancer Res Treat

123

function (by treating the ROE cells with anti-CXCR5-

antibody prior to CXCL13 treatment). Results showed that

inhibition of both Src and PI3Kp110a successfully atten-

uated EMT in CXCL13-stimulated BC cells by down-

regulating EMT positive markers (Fig. 6). Similar results

were obtained after treating these cells with anti-CXCR5

antibody (Fig. 6). RANKL level remained unchanged

except for the treatment of anti-CXCR5-antibody (Fig. 6).

Functional inhibition of CXCR5 by anti-CXCR5-antibody

treatment also confirmed the surface expression of the

receptor in the ROE ? LT cells.

The mRNA expression of RANKL was significantly

decreased (p \ 0.05) only in anti-CXCR5-antibody treated

ROE ? LT cells (13.3-fold and 11.7-fold for MDA-MB-231

and T-47D cells, respectively) compared to untreated

ROE ? LT cells (Fig. 6b). E-cadh level was modestly

increased in treated ROE ? LT T-47D cells (Fig. 6b ii).

None of the treatments showed significant (p [ 0.05) effect

in the expression of CXCR5 mRNA (Fig. 6a, b). WB vali-

dated the results at the protein level (Fig. 6c, d). Thus, the

inhibition analyses suggested a PI3Kp110a- and Src-inde-

pendent regulation of RANKL which mean that the RANKL

molecule may lie upstream in the signaling axis. Further,

attenuation of CXCR5 decreased RANKL expression. Thus,

the observed pattern of gene and protein expression is con-

sistent with CXCL13-stimulated expression of RANKL,

which, in turn, increases Src and PI3Kp110a activity leading

to up-regulation of MMP9, Vimentin, N-cadh, Slug, Snail

and down-regulation of E-cadh, resulted in EMT and deg-

radation of ECM.

Expression of RANKL, different EMT markers

and regulators, MMPs in tumor tissues of BC patients

Finally, to validate our findings, we performed expression

analysis of EMT markers, regulators, RANKL, and MMPs

categorically in co-expressing and non-co-expressing patient

samples. Kolmogorov–Smirnov test identified that mRNA

expression of CXCR5, CXCL13, Vimentin, Slug, Snail,

N-cadh, E-cadh, RANKL, MMP2, and MMP9 as variables not

following the normal distribution. Therefore, non-parametric

Mann–Whitney U test was performed to study the association

between co-expression of CXCL13–CXCR5 and mRNA

level expressions of the above mentioned genes. The p values,

derived from Mann–Whitney U test, are represented as scat-

tered plot (Fig. 7) to understand the significance level of fold

change of the relevant genes in terms of their mRNA

expression between CXCL13–CXCR5 co-expressed and non-

co-expressed (expression of one or none) tissues sampled

from tumor as well as healthy areas of breast of BC patients.

The analyses revealed that change in mRNA expression rel-

ative to the autologous healthy tissues were different for

CXCL13–CXCR5 co-expressing and non-co-expressing

tumor tissues. As Fig. 7 shows, statistically significant over-

expression of CXCR5 (p = 0.000000028) and N-cadh (p =

0.0000026) was only observed in co-expressing samples.

Fig. 5 ROE ? LT cells showed increased cellular migration and Src

phosphorylation. a To examine relative migration of control and

ROE ? LT cells, wounds were made in ROE ? LT MDA-MB-231

and T-47D cell layers just prior to CXCL13 treatment. Wound closure

was observed after 24 h. b Linear width of the wound was measured

and relative migration ratio is shown as bar graphs. c p-Src and t-Src

in control and ROE ? LT cells were evaluated by immunoblotting. b-

actin was used as loading control. d Band densitometries are shown in

bar graphs. Results are representative of three independent experi-

ments performed in triplicate and are represented as mean ± SD.

One-way ANOVA (Bonferroni correction) was performed, where

significance level stands for * p B 0.05, *** p B 0.001

Breast Cancer Res Treat

123

Moreover, relative expressions of CXCL13, Vimentin, Snail,

RANKL (ESM_3), and MMP9 in co-expressing samples are

much higher than that of non-co-expressing samples. How-

ever, E-cadh, Slug, and MMP2 were found to be overex-

pressed in both types of tumor samples.

Discussion

Breast remains one of the major cancer sites in women

worldwide [40]. BC is usually not fatal in its early stages,

but becomes so due to its metastatic spread. We still lack

comprehensive understanding of molecular and cellular

mechanisms underlying the process of metastasis and the

idea of BC as an intrinsically metastatic disease is being

debated [41]. Statistical reports have shown that BC

occurrence in India is rapidly rising, affecting a significant

percentage of women [42]. Fifty percent of the total BC

population in India is reported to be diagnosed at either of

stage-III or -IV [43]. Therefore, it is utmost important to

find new prognostic markers and treatment for BC patients

with advanced stage.

Among the various cellular changes that accompany

malignant transformation, EMT is considered the most

important one that helps cancer cells to become more

motile [31–35]. Increasing evidence suggests that multiple

Fig. 6 Involvement of RANKL, Src, PI3Kp110a in CXCL13-

induced EMT and MMP9 expression. Prior to CXCL13 treatment,

ROE ? LT cells were treated with inhibitors of Src and PI3Kp110a,

i.e., SU6656 and PI-103, respectively, or with anti-CXCR5 mono-

clonal antibody. a mRNA levels of Vimentin, Slug, Snail, E-cadh,

N-cadh, CXCR5, RANKL, and MMP9 were assessed by conventional

RT-PCR, followed by agarose gel electrophoresis. 18s r-RNA was

used as internal control. b Quantitative real-time RT-PCR was

performed for these mRNAs. Fold changes are represented as relative

values (2-DDCt) for SU6656, anti-CXCR5-antibody, PI-103-treated

ROE ? LT cells normalized with internal control and quantified.

Fold changes in MDA-MB-231 were as follows: Vimentin decreased

4.4, 3.7, 4.0-fold, respectively; N-cadh decreased 22.5, 31.8, 36.3-

fold, respectively; Snail decreased 16.1, 17.2, 16.7-fold, respectively;

Slug decreased 2.2, 8.33, 3.9-fold, respectively; and MMP9 decreased

7.7, 12.5, 20.0-fold, respectively. For T-47D cells, fold changes were

as follows: Vimentin decreased 24.9, 37.3, 38.5-fold, respectively;

N-cadh decreased 17.3, 38.6, 34.2-fold, respectively; Snail decreased

26.0, 38.5, 35.1-fold, respectively; Slug decreased 3.1, 5.7, 8.2-fold,

respectively; and MMP9 was decreased 5.9, 37.3, 16.5-fold, respec-

tively. E-cadh expression increased significantly in all the treatment

conditions for T-47D cells. RANKL expression, however, decreased

significantly only upon anti-CXCR5 treatment. No treatment had any

effect on CXCR5 mRNA level. c Expressions of Vimentin, Slug,

Snail, N-cadh, E-cadh, RANKL, and CXCR5 were evaluated by

immunoblotting. b-actin was used as loading control. Vimentin, Slug,

Snail, and N-cadh expressions were decreased in SU6656, PI-103,

anti-CXCR5-antibody treated ROE ? LT cells. Inhibitor/anti-

CXCR5-antibody treated ROE ? LT T-47D cells have decreased

E-cadh expression. RANKL expression was only found to be down-

regulated in anti-CXCR5-antibody treated ROE ? LT cells. d WB

band densitometries shown in bar graphs. Results are representative

of three independent experiments performed in triplicate and are

represented as mean ± SD. One-way ANOVA (Bonferroni correc-

tion) was performed, where significance level stands for * p B 0.05,

** p B 0.01, *** p B 0.001

Breast Cancer Res Treat

123

chemokines and their corresponding receptors are involved

in BC progression and play significant roles in metastatic

process [29]. CXCR5, a well-known G-protein-coupled

chemokine receptor of hematopoietic cells, was recently

reported to be expressed in colon, prostate, breast carci-

noma tissues, and in neuroblastoma cells [25–28, 44–46].

There are also literature reports on CXCL13 overexpres-

sion in BC [27] as well as molecular data on CXCL13–

CXCR5 signaling in prostate cancer cell invasion [30, 44,

45]. Therefore, a BC study correlating patient samples and

cell lines has been desirable and indeed, we showed sig-

nificant co-expression of chemokine receptor CXCR5 and

its ligand CXCL13 in primary tumor tissues of LNM

positive BC patients. The interaction between CXCR5 and

CXCL13 has been widely studied for the development of

lymph nodes and Peyers’ patches [47]. Our results suggest

that a similar autocrine signaling within primary BC tumor

tissue might support LNM. EMT-inducing potential of

CXCL13–CXCR5 signaling in BC cells is demonstrated by

elevated expression of RANKL, Snail, Slug, Vimentin,

N-cadh, MMP2, and MMP9 which correlated significantly

with co-expression of CXCL13 and CXCR5 within the

primary breast tumors. Another important hallmark of

metastatic cancer is the loss of E-cadh [34]. E-cadh gene

expression is down-regulated by EMT regulators, Slug, and

Snail [48] via recruitment of G9a methyl-transferase,

methylation of histone H3 on lysine 9 and subsequent DNA

methylation at E-cadh promoter [49]. Snail expression is

regulated by ER, which, therefore, acts as an EMT

inhibitor and maintains the epithelial nature of normal

breast cells [50]. MDA-MB-231 cells lack ER and thus

demonstrate constitutive E-cadh promoter methylation and

no E-cadh expression.

Earlier reports showed that RANKL increases the

migration of BC cells by activating c-Src/Akt and c-Src/

ERK signaling [37]. Our observations that RANKL

expression was sensitive to CXCR5 modulation, (either by

CXCL13 or by anti-CXCR5), but not to specific Src or

PI3Kp110a inhibition, confirmed that in BC cells, RANKL

acts upstream of Src. CXCL13-mediated Src activation

appears to activate at least three distinct signaling pathways

related to cell survival, invasion, and growth [51]. Src

mediates phosphorylation of caspase-8 at Tyr-380, which,

in turn, stimulates Src phosphorylation at Tyr-416 and Src

has been found to remain active upon phosphorylation and

binding caspase-8 at phosphorylated Tyr-380, which effi-

ciently induced EMT and promoted tumor cell metastasis

[52]. We showed here that similar molecular machinery

operates in BC cells where stimulation with CXCL13 also

induces EMT, which as well depends on Src/PI3Kp110a.

In summary, CXCL13 has stimulated BC cells to

overexpress RANKL, which might be responsible for

activation of Src. Activated Src induced cell migration

pathway through activation of PI3Kp110a along with

EMT. EMT proceeded through regulation of different

mesenchymal and epithelial markers as well as EMT reg-

ulators. RANKL might also be responsible for increased

MMP9 expression in CXCL13-stimulated cells.

Fig. 7 Scattered plot showing

statistical correlation between

CXCL13–CXCR5 co-

expression and EMT markers,

EMT regulators, RANKL, and

MMPs. Scattered plot showing

relative gene expressions (tumor

tissues/healthy tissues) of both

CXCL13–CXCR5 co-

expressing and non-co-

expressing tumor tissues.

Mann–Whitney U test-derived

p values from CXCL13–

CXCR5 co-expressing samples

are plotted against X-axis and

from non-co-expressing samples

are plotted against Y-axis.

Region of p value [ 0.05,

between 0.05 and 0.01, and

\0.01 are represented as

insignificant, significant, and

overexpressed, respectively

Breast Cancer Res Treat

123

From the clinical point of view, it reports that autocrine

CXCL13–CXCR5 signaling exists in LNM positive BC

patients. Co-expression of CXCR5 and CXCL13 could be

used as a poor-prognosis marker for BC in the future,

although validation of this test requires future investiga-

tion. Finally, this study indicates CXCR5 attenuation as a

possible therapeutic strategy for advanced metastatic BC.

Acknowledgments Financial supports were made by Department of

Science and Technology, Govt. of India (Sanction No.-INT/RFBR/P-

82), Council of Scientific and Industrial Research (37/1455/10/EMR-

II), and Russian Foundation for Basic Research, Russian Federation

(Sanction No.-10-04-92657) for the work and Council of Scientific and

Industrial Research—JRF/NET Fellowship Grant [No. 9/028(842)/

2011-EMR-I] to Subir Biswas. We thank the patients and their families

who participated in this study and also thank the nursing staffs of

SGCC & RI. We are thankful to Shravasti Roy, for her help in

acquiring histopathological information of the patients, Prakriti Roy

and Mayuri Nath for their help in sample collection, Kayum Alam for

his help in acquisition of real-time PCR data, Subhadip Kundu and

Soumya Chatterjee for initial project design, Soham Mitra, Tarun

Keswani, Shauryabrota Dalui for their help in running the experiments.

Conflict of interest The authors declared that they do not have any

conflict of interest.

References

1. Balkwill F (2004) Cancer and the chemokine network. Nat Rev

Cancer 4:540–550

2. Gupta SK, Lysko PG, Pillarisetti K, Ohlstein E, Stadel JM (1998)

Chemokine receptors in human endothelial cells. Functional

expression of CXCR4 and its transcriptional regulation by

inflammatory cytokines. J Biol Chem 273:4282–4287

3. Zlotnik A, Yoshie O (2000) Chemokines: a new classification

system and their role in immunity. Immunity 12:121–127

4. Patel J, Channon KM, McNeill E (2013) The downstream regu-

lation of chemokine receptor signaling: implications for athero-

sclerosis. Mediators Inflamm. doi:10.1155/2013/459520

5. Latek D, Modzelewska A, Trzaskowski B, Palczewski K, Filipek

S (2012) G protein-coupled receptors-recent advances. Acta

Biochim Pol 59:515–529

6. Zlontik A (2004) Chemokines in neoplastic progression. Semin

Cancer Biol 14:181–185

7. Longo-Imedio MI, Longo N, Trevino I, Lazaro P, Sanchez-

Mateos P (2005) Clinical significance of CXCR3 and CXCR4

expression in primary melanoma. Int J Cancer 117:861–865

8. Murphy C, McGurk M, Pettigrew J et al (2005) Nonapical and

cytoplasmic expression of interleukin-8, CXCR1, and CXCR2

correlates with cell proliferation and microvessel density in

prostate cancer. Clin Cancer Res 11:4117–4127

9. Keeley EC, Mehrad B, Strieter RM (2010) CXC chemokines in

cancer angiogenesis and metastases. Adv Cancer Res 106:91–111

10. Belperio JA, Keane MP, Arenberg DA et al (2000) CXC che-

mokines in angiogenesis. J Leukoc Biol 68:1–8

11. Vicari AP, Caux C (2002) Chemokines in cancer. Cytokine

Growth Factor Rev 13:143–154

12. Murakami T, Cardones AR, Hwang ST (2004) Chemokine

receptors and melanoma metastasis. J Dermatol Sci 36:71–78

13. Tanaka T, Bai Z, Srinoulprasert Y, Yang BG, Hayasaka H, Mi-

yasaka M (2005) Chemokines in tumor progression and metas-

tasis. Cancer Sci 96:317–322

14. Shayan R, Achen MG, Stacker SA (2006) Lymphatic vessels in

cancer metastasis: bridging the gaps. Carcinogenesis

27:1729–1738

15. Li YM, Pan Y, Wei Y et al (2004) Up-regulation of CXCR4 is

essential for HER2-mediated tumor metastasis. Cancer Cell

6:459–469

16. Bachelder RE, Wendt MA, Mercurio AM (2002) Vascular

endothelial growth factor promotes breast carcinoma invasion in

an autocrine manner by regulating the chemokine receptor

CXCR4. Cancer Res 62:7203–7206

17. Muller A, Homey B, Soto H et al (2001) Involvement of che-

mokine receptors in breast cancer. Nature 410:50–56

18. Wiseman BS, Werb Z (2002) Stromal effects on mammary gland

development and breast cancer. Science 296:1046–1049

19. Kim M, Koh YJ, Kim KE et al (2010) CXCR4 signaling regulates

metastasis of chemoresistant melanoma cells by a lymphatic

metastatic niche. Cancer Res 70:10411–10421

20. Rhodes LV, Short SP, Neel NF et al (2011) Chemokine receptor

CXCR4 mediates estrogen-independent tumorigenesis, metasta-

sis, and resistance to endocrine therapy in human breast cancer.

Cancer Res 71:603–613

21. Dhawan P, Richmond A (2002) Role of CXCL1 in tumorigenesis

of melanoma. J Leukoc Biol 72:9–18

22. Zhou Y, Zhang J, Liu Q et al (2005) The chemokine GRO-alpha

(CXCL1) confers increased tumorigenicity to glioma cells. Car-

cinogenesis 26:2058–2068

23. Singh S, Singh AP, Sharma B, Owen LB, Singh RK (2010)

CXCL8 and its cognate receptors in melanoma progression and

metastasis. Future Oncol 6:111–116

24. Lin Y, Huang R, Chen L, Li S, Shi Q, Jordan C, Huang RP (2004)

Identification of interleukin-8 as estrogen receptor-regulated

factor involved in breast cancer invasion and angiogenesis by

protein arrays. Int J Cancer 109:507–515

25. Meijer J, Zeelenberg IS, Sipos B, Roos E (2006) The CXCR5

chemokine receptor is expressed by carcinoma cells and promotes

growth of colon carcinoma in the liver. Cancer Res

66:9576–9582

26. Burkle A, Niedermeier M, Schmitt-Graff A, Wierda WG, Keating

MJ, Burger JA (2007) Overexpression of the CXCR5 chemokine

receptor, and its ligand, CXCL13 in B-cell chronic lymphocytic

leukemia. Blood 110:3316–3325

27. Panse J, Friedrichs K, Marx A et al (2008) Chemokine CXCL13

is overexpressed in the tumor tissue and in the peripheral blood of

breast cancer patients. Br J Cancer 99:930–938

28. Airoldi I, Cocco C, Morandi F, Prigione I, Pistoia V (2008)

CXCR5 may be involved in the attraction of human metastatic

neuroblastoma cells to the bone marrow. Cancer Immunol Im-

munother 57:541–548

29. Razmkhah M, Jaberipour M, Safaei A, Talei AR, Erfani N,

Ghaderi A (2012) Chemokine and chemokine receptors: a com-

parative study between metastatic and non-metastatic lymph

nodes in breast cancer patients. Eur Cytokine Netw 23:72–77

30. El Haibi CP, Sharma PK, Singh R, Johnson PR, Suttles J, Singh

S, Lillard JW Jr (2010) PI3Kp110-, Src-, FAK-dependent and

DOCK2-independent migration and invasion of CXCL13-stimu-

lated prostate cancer cells. Mol Cancer 9:85

31. Guarino M (2007) Epithelial-mesenchymal transition and tumor

invasion. Int J Biochem Cell Biol 39:2153–2160

32. Markiewicz A, Ahrends T, Wełnicka-Jaskiewicz M et al (2012)

Expression of epithelial to mesenchymal transition-related

markers in lymph node metastases as a surrogate for primary

tumor metastatic potential in breast cancer. J Transl Med 10:226

33. Thiery JP, Sleeman JP (2006) Complex networks orchestrate

epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol

7:131–142

Breast Cancer Res Treat

123

34. Cavallaro U, Christofori G (2004) Cell adhesion and signalling

by cadherins and Ig-CAMs in cancer. Nat Rev Cancer 4:118–132

35. Thiery JP, Acloque H, Huang RY, Nieto MA (2009) Epithelial-

mesenchymal transitions in development and disease. Cell

139:871–890

36. Fata JE, Kong YY, Li J et al (2000) The osteoclast differentiation

factor osteoprotegerin-ligand is essential for mammary gland

development. Cell 103:41–50

37. Zhang L, Teng Y, Zhang Y et al (2012) C-Src-mediated RANKL-

induced breast cancer cell migration by activation of the ERK and

Akt pathway. Oncol Lett 3:395–400

38. Armstrong AP, Miller RE, Jones JC, Zhang J, Keller ET, Dougall

WC (2008) RANKL acts directly on RANK-expressing prostate

tumor cells and mediates migration and expression of tumor

metastatic genes. Prostate 68:92–104

39. Neve RM, Chin K, Fridlyand J et al (2006) A collection of breast

cancer cell lines for the study of functionally distinct cancer

subtypes. Cancer Cell 10:515–527

40. Jemal A, Siegel R, Xu J, Ward E (2010) Cancer statistics, 2010.

CA Cancer J Clin 60:277–300

41. Weigelt B, Peterse JL, van’t Veer LJ (2005) Breast cancer

metastasis: markers and models. Nat Rev Cancer 5:591–602

42. Sen U, Sankaranarayanan R, Mandal S, Ramanakumar AV,

Parkin DM, Siddiqi M (2002) Cancer patterns in eastern India:

the first report of the Kolkata cancer registry. Int J Cancer

100:86–91

43. http://www.breastcancerindia.net/bc/statistics/stati.htm. Accessed

15 Jun 2013

44. Singh S, Singh R, Singh UP et al (2009) Clinical and biological

significance of CXCR5 expressed by prostate cancer specimens

and cell lines. Int J Cancer 125:2288–2295

45. Singh S, Singh R, Sharma PK et al (2009) Serum CXCL13

positively correlates with prostatic disease, prostate-specific

antigen and mediates prostate cancer cell invasion, integrin

clustering and cell adhesion. Cancer Lett 283:29–35

46. Del Grosso F, Coco S, Scaruffi P et al (2011) Role of CXCL13–

CXCR5 crosstalk between malignant neuroblastoma cells and

Schwannian stromal cells in neuroblastic tumors. Mol Cancer Res

9:815–823

47. Ebisuno Y, Tanaka T, Kanemitsu N et al (2003) Cutting edge: the

B cell chemokine CXC chemokine ligand 13/B lymphocyte

chemoattractant is expressed in the high endothelial venules of

lymph nodes and Peyer’s patches and affects B cell trafficking

across high endothelial venules. J Immunol 171:1642–1646

48. Peinado H, Portillo F, Cano A (2004) Transcriptional regulation

of cadherins during development and carcinogenesis. Int J Dev

Biol 48:365–375

49. Dong C, Wu Y, Yao J et al (2012) G9a interacts with Snail and is

critical for Snail-mediated E-cadherin repression in human breast

cancer. J Clin Invest 122:1469–1486

50. Fujita N, Jaye DL, Kajita M, Geigerman C, Moreno CS, Wade

PA (2003) MTA3, a Mi-2/NuRD complex subunit, regulates an

invasive growth pathway in breast cancer. Cell 113:207–219

51. El-Haibi CP, Singh R, Gupta P, Sharma PK, Greenleaf KN, Singh

S, Lillard JW Jr (2012) Antibody microarray analysis of signaling

networks regulated by Cxcl13 and Cxcr5 in prostate cancer.

J Proteomics Bioinform 5:177–184

52. Zhao Y, Li X, Sun X, Zhang Y, Ren H (2012) EMT phenotype is

induced by increased Src kinase activity via Src-mediated cas-

pase-8 phosphorylation. Cell Physiol Biochem 29:341–352

Breast Cancer Res Treat

123