Role of CXC chemokine ligand 13 in oral squamous cell carcinoma associated osteolysis in athymic mice

Oral squamous cell carcinomas (OSCC) are malignant tumors with a potent activity of local bone invasion; however, the molecular mechanisms of tumor osteolysis are unclear. In this study, we identified high level expression of chemokine ligand, CXCL13 and RANK ligand (RANKL) in OSCC cells (SCC1, SCC12 and SCC14a). OSCC cell‐conditioned media (20%) induced osteoclast differentiation which was inhibited by OPG in peripheral blood monocyte cultures indicating that OSCC cells produce soluble RANKL. Recombinant hCXCL13 (10 ng/ml) significantly enhanced RANKL‐stimulated osteoclast differentiation in these cultures. Trans‐well migration assay identified that CXCL13 induces chemotaxis of peripheral blood monocytes in vitro which was inhibited by addition of anti‐CXCR5 receptor antibody. Zymogram analysis of conditioned media from OSCC cells revealed matrix metalloproteinase‐9 (MMP‐9) activity. Interestingly, CXCL13 treatment to OSCC cells induced CXCR5 and MMP‐9 expression suggesting an autocrine regulatory function in OSCC cells. To examine the OSCC tumor cell bone invasion/osteolysis, we established an in vivo model for OSCC by subcutaneous injection of OSCC cells onto the surface of calvaria in NCr‐nu/nu athymic mice, which developed tumors in 4–5 weeks. μCT analysis revealed numerous osteolytic lesions in calvaria from OSCC tumor‐bearing mice. Histochemical staining of calvarial sections from these mice revealed a significant increase in the numbers of TRAP‐positive osteoclasts at the tumor‐bone interface. Immunohistochemical analysis confirmed CXCL13 and MMP‐9 expression in tumor cells. Thus, our data implicate a functional role for CXCL13 in bone invasion and may be a potential therapeutic target to prevent osteolysis associated with OSCC tumors in vivo.

Head and neck squamous cell carcinoma (HNSCC) are the most common malignant neoplasms, with a prevalence estimated to be greater than 40,000 cases annually in the US. Oral squamous cell carcinoma (OSCC), which contributes to 40% of all HNSCC is associated with mucosal surfaces of the oral cavity and oropharynx. 1 The etiology of OSCC involves both a genetic predisposition and exposure to environmental carcinogens such as tobacco, alcohol, chronic inflammation and viral infection. 2 Curcumin has been shown to suppress HNSCC growth both in vitro and in vivo, 3 and Fas-mediated apoptosis in HNSCC is a proven and efficient therapeutic approach in a xenograft animal model. 4 Furthermore, tamoxifen inhibition of OSCC cell growth in vitro 5 and a role for human longevity assurance gene 1 (LASS1) and C 18 -ceramide in chemotherapy induced cell death in HNSCC have been reported. 6 Malignant HNSCC tumors are known to have a potent activity of local bone invasion; however, the molecular mechanisms of tumor-associated osteolysis are unclear.
The osteoclast is hematopoietic in origin and is the boneresorbing cell derived from monocyte/macrophage lineage. Tumor necrosis factor (TNF) family member, RANK ligand (RANKL), which is expressed on marrow stromal/osteoblast cells in response to several osteotropic factors, is critical for osteoclast precursor differentiation to form multinucleated osteoclasts, which resorb bone. 7 Osteoclast activity is controlled by local factors produced in the bone microenvironment. In addition, the osteoclast is an autocrine/paracrine, intracrine regulatory cell that produces factors such as IL-6, annexin II, TGF-beta and OIP-1/hSca, which influence its own formation and activity. Matrix metalloproteinase-9 (MMP-9), a type IV collagenase is highly expressed in osteoclast cells and plays an important role in degradation of the extracellular matrix. 8 Osteoclast activation plays an important role in several malignancies including oral cancers invasion of bone and subsequent metastasis. 9 Further, studies using a murine mandibular bone invasion model for OSCC demonstrated mRNA expression of cytokines associated with osteoclast activation such as IL-6, TNF-a and PTHrP in tumor tissue as well as high bone resorption. 9 Also, conditioned media from OSCC cells derived from patients with bone involvement stimulated osteoclast differentiation in vitro. 10 Chemokines are a superfamily of small, cytokine-like proteins that selectively attract and activate different cell types. 11 CXC chemokines are known to promote angiogenesis 12 and have a characteristic heparin-binding domain. Chemokines interact with seven-transmembrane-domain glycoprotein receptors coupled to the G protein signaling pathway. 11 In several studies, tumor cells were shown to express functionally active chemokine receptors which regulate cellular functions and metastasis. 13 HNSCC has been reported to predominantly expressed chemokine receptors such as CCR7 and CXCR5; however, CXCR4 expression is low or undetectable. 14 CXCL13 (BCA-1) which binds monogamously to the CXCR5 receptor and is involved in B-cell chemotaxis and is induced under inflammatory conditions. 15 Microarray analysis for gene expression profiling in OSCC identified gene signatures which include chemokine (CXC motif) ligand-13 and matrix-metalloproteinases (MMPs) that are highly relevant to OSCC development and progression. 16 However, a functional role for CXCL13 in HNSCC tumor cell invasion and osteolysis is unknown. In this study, we showed CXCL13 expression and an autocrine regulation of MMP-9 production in tumor cells. We further show CXCL13 and RANKL expression in OSCC cells support osteoclastogenesis. We developed an in vivo model for OSCC by subcutaneous injection of SCC14a cells onto the surface of calvaria in NCr-nu/nu athymic mice which showed osteolytic lesions. Our data implicate CXCL13 a potential therapeutic target to prevent OSCC tumor-associated osteolysis in vivo.

Cell lines and culture
Human OSCC-derived cell lines SCC1, SCC12 and SCC14a were generously provided by Dr. Thomas E. Carey (University of Michigan, Ann Arbor, MI). The OSCC, normal human fibroblast cell lines, WI-38 and IMR-90 (ATCC, Manassas, VA) were maintained in Dulbecco's Modified Eagle Medium containing 10% fetal bovine serum (FBS) and supplemented with L-glutamine, penicillin and streptomycin. The RWPE-1 (normal human prostatic epithelial cells) (ATCC, Manassas, VA) cells were grown in keratinocyte serum free medium (K-SFM) containing 50 lg/ml bovine pituitary extract and 5 ng/ml epidermal growth factor. All cells were incubated at 37 C in 5% CO 2.

Osteoclast differentiation assay
Human peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood using Histopaque 1077 (Sigma, St Louis, MO) and resuspended in a-MEM with 10% FBS as described. 18 Briefly, 15 ml of whole blood was mixed with 15 ml of warm (37 C) a-MEM, layered over 15 ml of histopaque and centrifuged (1,500g, 30 min) at room temperature. The cell layer on top of the histopaque was collected, resuspended in 10 ml of a-MEM and centrifuged. The mononuclear cells collected were plated in 96-well plates at 6 Â 10 5 cells per well in 0.2 ml of medium (a-MEM, pH 7.4, containing 10% FBS) supplemented with 10 ng/ml hM-CSF and presence of 100 ng/ml hRANKL or OSCC cell conditioned media (CM). The cells were re-fed twice weekly by semidepletion (half of the medium withdrawn and replaced with fresh medium). At the end of culture period (10 days), the cells were fixed with 2% glutaraldehyde in PBS and stained for tartrate-resistant acid phosphatase (TRAP) activity using a histochemical staining kit (Sigma, St. Louis, MO) as described. 19 TRAP-positive multinucleated cells containing 3 or more nuclei were scored as osteoclast cells under a microscope.

Chemotaxis assay
Chemotaxis of freshly isolated PBMC as described above was assessed in a 24-well plate containing polycarbonate membrane inserts with an 8-lm pore size (Costar Corning, NY) as described. 20 Briefly, cell suspensions (4.5 Â 10 3 ) of monocytes in culture medium were added to the upper chemotaxis chambers overlying wells containing medium with or without recombinant human CXCL13 (0-10 ng/ml). After 1 hr incubation at 37 C in 5% CO 2 , cells that migrated into the lower chamber were counted using a hemocytometer.

Western blot analysis
OSCC cells were seeded in 6-well plates at a density of 5 Â 10 5 cells per ml of DMEM containing 10% FBS and cultured for 48 hr in the presence or absence of CXCL13 (10 ng/ml). The cells were lysed in RIPA buffer containing 1Â protease inhibitor cocktail (Sigma, St. Louis, MO). Protein content of the samples was measured using the BCA reagent (Pierce, Madison, WI). Protein (20 lg) samples were then subjected to SDS-PAGE using 12% Tris-HCl gels and blot transferred on to a nitrocellulose membrane. Blocking was performed with 5% nonfat dry milk in TBST buffer (50 mM Tris, pH 7.2, 150 mM NaCl; 0.1%Tween 20) for 1 hr. The membrane was incubated with a primary anti-MMP-9 and anti-CXCR5 antibody at 1:1,000 dilution overnight at 4 C. The blots were then incubated for 1 hr with horseradish peroxidase conjugated secondary antibody and developed using an ECL system (Pierce, Madison, WI). b-actin expression levels in all the samples were used to normalize loading variations. The band intensity was quantified by densitometric analysis using the NIH ImageJ Program.

Gelatin zymography for MMP-9 activity
The conditioned media obtained from OSCC cells stimulated with and without CXCL13 (10 ng/ml) for 48 hr were analyzed for MMP-9 activity by gelatin substrate gel electrophoresis as described earlier. 21 Serum free cell cultured media samples were applied without reduction to a 10% polyacrylamide gel containing 0.1% gelatin. After electrophoresis, the gels were washed in washing buffer (50 mM Tris-HCl, pH 7.5; 5 mM CaCl 2 , 1 lM ZnCl 2 , and 2.5% Triton X-100) for 30 min at room temperature, and then incubated overnight at 37 C in the above buffer without Triton X-100. The gels were stained with a solution containing 0.1% Coomassie Brilliant Blue R-250. Formation of clear zone against the blue background on the polyacrylamide gels indicated the gelatinolytic activity, which was quantified by densitometric analysis using the NIH ImageJ Program.
CXCL13 shRNA knock-down CXCL13 expression in SCC14a cells was knock-down using SureSilencing shRNA plasmid obtained from SABiosciecnes (Frederick, MD). CXCL13 targeting sequence in the plasmid was 5 0 -ATCCCTAGACGCTTCATTGAT-3 0 and the scrambled sequence in the control shRNA plasmid was 5 0 -GGAATCTCAT TCGATGCATAC-3 0 . SCC14a cells were stably transfected with shRNA containing vectors by lipofectamine and clonal cell lines were established by selection with neomycin (800 lg/ml). shRNA suppression of CXCL13 production and mRNA expres-sion were confirmed by ELISA and real-time RT-PCR as described.

In vivo model for OSCC
Athymic NCr-nu/nu male mice, aged 4-6 weeks (NCI, Frederick, MD) were used to develop an in vivo model for OSCC tumor cell invasion into bone and osteolysis. Under sterile conditions, 7 Â 10 6 OSCC cells in phosphate buffered saline (PBS) were injected subcutaneously (n ¼ 10) overlaying the calvaria and PBS alone injected were as served control group (n ¼ 8). Tumor development over calvaria was monitored weekly using vernier calipers. Animals were sacrificed when the tumor reached 2,000 mm 3 . At the end of experimental period, the animals were sacrificed and calvaria were collected for lCT analysis. Tumor were surgically removed and fixed in formalin for histological analysis.

Micro-computed tomography (lCT) imaging
Calvaria were surgically removed from PBS treated control, SCC14a, SCC12 tumors-bearing athymic mice were fixed in 70% ethanol and scanned using a Skyscan 1072 lCT instrument (Skyscan, Antwerp, Belgium). lCT-Analyser software (from SkyScan) was used to analyze the structure of the sample using the global segmentation method. Two-dimensional images were used to generate three-dimensional reconstructions with the software supplied with the instrument.

Histologic analysis
Formalin-fixed SCC14a tumor specimens collected from athymic mice were processed for paraffin sectioning. Serial 5lm sections were cut on a modified Leica RM 2155 rotary microtome (Leica Microsystems, Ontario, Canada) and stained with hematoxylin and eosin. Immunocytochemical staining of the sections were performed by incubation of serial sections with anti-CXCR5, anti-CXCL13 or anti-MMP-9 antibody in 5% BSA overnight followed by HRP labeled secondary antibody and DAB staining. Specimens treated with non-specific IgG served as control.
To perform histochemical staining, tumor-bearing calvaria were decalcified in 0.5 M EDTA (pH 7.4) for a 1-3 week period and processed for paraffin embedding. Serial 4 to 6-lm sections of paraffin embedded calvaria were stained for TRAP activity. Histomorphometric analysis of TRAP-positive osteoclast numbers at the tumor-bone interface was performed with OsteoMeasure version 2.2 software.

Statistical analysis
Results are represented as mean 6 SE of at least 3 independent experiments. Differences between experimental groups

Cancer Cell Biology
Pandruvada et al.
were analyzed by ANOVA using Graphpad Prism software. Values were considered significantly different for *p < 0.05.

CXCL13 and CXCR5 expression in OSCC cells
Chemokines are implicated in tumor progression and metastasis. 22 Recent evidence indicated down regulation of CXCL5 expression inhibits HNSCC development and bone invasion. 23 In this study, we examined CXCL13 and CXCR5 expression in OSCC derived cell lines (SCC1, SCC12 and SCC14a). ELISA analysis of conditioned media (CM) obtained from OSCC cell lines revealed CXCL13 at significantly high (38 to 118 pg/ml) concentration. In contrast, normal epithelial and fibroblast cells produced low levels (18 to 25 pg/ml) of CXCL13 (Fig. 1a). We further examined CXCL13 and CXCR5 receptor mRNA expression in OSCC cells by real-time RT-PCR analysis. CXCL13 mRNA expression was abundant in SCC1 and SCC14a; but at very low lev-els in SCC12 cells. CXCR5 mRNA is expressed at high levels in all the OSCC cell lines analyzed (Fig. 1b). We further examined whether CXCL13 stimulates CXCR5 mRNA expression in OSCC cells. OSCC cells were stimulated with different concentration of CXCL13 (0-25 ng/ml) for 48 hr. Real-time RT-PCR analysis of total RNA isolated from these cells identified a dose-dependent stimulation of CXCR5 mRNA expression (Fig. 1c). Western blot analysis of total cell lysates further confirmed CXCL13 stimulation of CXCR5 expression in OSCC cells (Fig. 1d). Thus, our results implicate CXCL13/CXCR5 expression in OSCC and suggest an autocrine regulatory role for CXCL13 for enhancing CXCR5 receptor expression in OSCC cells.

OSCC cells support osteoclast formation
Tumor cells activate bone resorbing osteoclasts, thereby facilitating osteolytic process and bone invasion. 24 RANKL is a critical osteoclastogenic factor in the bone microenvironment. 7 To determine the potential of OSCC cells to stimulate osteoclast formation, CM collected from these cells was tested at 1 and 20% concentrations for its capacity to stimulate osteoclast formation in peripheral blood mononuclear cell (PBMC) cultures as described in the methods. Treatment of PBMC with SCC14a cell CM at a 20% concentration in the presence of M-CSF (10 ng/ml) significantly increased (190 6 23) TRAP positive multinucleated cell formation when compared with control cultures treated with M-CSF alone (8 6 2) in the absence of RANKL. Osteoclasts formed in PBMC cultures stimulated with RANKL (100 ng/ml) and M-CSF (10 ng/ml) served as a positive control (321 6 32) (Fig. 2a). Osteoprotegerin (OPG) is a decoy receptor of RANKL which inhibits RANKL-stimulated osteoclast differentiation 25 and addition of OPG significantly inhibited osteoclast formation when compared with OSCC-CM alone in these cultures (Fig. 2b). These data suggest that the OSCC cells produce significant levels of RANKL, thereby stimulating osteoclast differentiation. Real-time RT-PCR analysis further confirmed RANKL mRNA expression in OSCC cells (Fig. 2c).

CXCL13 stimulates MMP-9 expression in OSCC cells
Tumor cell production of MMP-9 facilitates bone invasion. 26 Therefore, we further examined MMP-9 expression and the potential of CXCL13 to up-regulate MMP-9 expression in OSCC cells. Zymogram analysis of CM obtained from OSCC cell lines (SCC14a, SCC1 and SCC12) stimulated with CXCL13 for 48 hr demonstrated a significant increase in the levels of MMP-9 activity (Fig. 3a). Total RNA isolated from OSCC cells treated with recombinant CXCL13 (0-25 ng/ml) for a 48 hr period were subjected to real-time RT-PCR analysis for MMP-9 mRNA expression. As shown in Figure  3b, OSCC cells stimulated with CXCL13 demonstrated a dose-dependent increase in the levels of MMP-9 mRNA expression. Western blot analysis of total cell lysates obtained from these cells further confirmed CXCL13 stimulation of MMP-9 expression in a dose-dependent manner (Fig. 3c). Further, OSCC cells stimulated with CXCL13 in the presence of anti-CXCR5 antibody showed a significant decrease in the levels of MMP-9 expression (Fig. 3d). These data indicate that the CXCL13/CXCR5 axis play an important role in bone invasion through up-regulation of MMP-9 expression in OSCC cells.

CXCL13 promotes chemotactic recruitment of preosteoclasts
Chemokines are chemotactic and modulate cellular functions in bone remodeling. Moreover, abundant levels of CXCR5, but very low levels of CXCL13 expression in differentiating osteoclasts have been reported. 27 We, therefore, examined a potential chemotactic role for CXCL13 expression in OSCC tumor cells in chemotaxis for human peripheral bloodderived osteoclast progenitor cells using a transwell system as described in methods. Recombinant CXCL13 protein dosedependently stimulated chemotaxis of peripheral blood monocytes in 1 hr at 10 ng/ml concentration (126 6 15 vs. 37 6 6 in control) (Fig. 4a). This migration was significantly inhibited in the presence of anti-CXCR5 receptor antibody (Fig. 4b). These results indicate that the CXCL13/CXCR5 axis may be critical in chemo-attraction of preosteoclast cells which facilitates osteoclast differentiation.

OSCC tumor invasion/osteolysis in vivo
To further examine the molecular mechanisms associated with OSCC tumor cell invasion and osteolysis of bone, we established an in vivo model for OSCC by subcutaneous injection of OSCC cells (SCC14a, SCC1 and SCC12) onto the surface of calvaria (7 Â 10 6 cells in PBS) in NCr-nu/nu athymic mice. Tumors appeared by day 7 and thereafter continued to grow for 4-5 weeks. The tumor size was measured at regular intervals with vernier calipers. All the OSCC cell lines demonstrated tumor development; however, SCC14a cells with high tumor growth when compared with SCC1 and SCC12 cell lines (Fig. 5a) and athymic mice with vehicle (PBS) control and SCC14a tumor-bearing mice are shown in Figure 5b. After tumors reached 2,000 mm 3 , the mice were sacrificed and the isolated calvaria were subjected to lCT and histological analysis. lCT imaging analysis revealed numerous osteolytic lesions in calvaria isolated from OSCC tumor bearing mice. The increased rate of osteolysis in SCC14a tumor bearing mice compared to SCC12 could be related to high level production of CXCL13 (118 pg/ml) in these cells (Fig. 5c). Histologic analysis of calvaria derived from tumor-bearing mice revealed tumor cell invasion into the bone matrix. Histochemical staining for TRAP activity identified a significant increase in the numbers of TRAP-positive osteoclasts at the tumor-bone interface when compared with control mice. (Figs. 6a-d). The sections were analyzed by immunohistochemical staining to further confirm the expression of CXCL13/CXCR5 in tumor cells in vivo. As shown in Figure 6e, tumor cells showed positive immunocytochemical staining with antibodies against CXCL13 and CXCR5. Tumor cells also stained positive for MMP-9 expression. In contrast, a control IgG did not stain. To further determine the potential of CXCL13 in OSCC tumor osteolysis, we developed SCC14a cells stably transfected with vectors containing control scrambled shRNA and shRNA against CXCL13 by lipofectamine and clonal cell lines were established by selection with neomycin (800 lg/ml). shRNA suppression of CXCL13 production and mRNA expression were confirmed by ELISA and real-time RT-PCR, respectively (Figs. 5e and 5d). As shown in Figure 5f, calvaria from SCC14a cells knock-down of CXCL13 expression demonstrated a significant decrease (52%) in tumor osteolysis when compared with control scrambled shRNA transfected cells; however, no change in tumor size. Further, histologic analysis of calvaria from SCC14a cells shRNA knock-down for CXCL13 tumor bearing mice showed inhibition of tumor invasion into bone matrix and significant decrease in osteoclasts number at tumor-bone interface when compared with control scrambled shRNA transfected cells (Figs. 6f and 6g). Taken together, our results implicate CXCL13 play an important role in OSCC tumor invasion and osteolysis in vivo.

Discussion
Chemokines have a profound influence on bone remodeling in pathologic conditions. Recent studies indicated that down regulation of CXCL5 inhibits SCC carcinogenesis and that CXCL8 modulates cellular proliferation and migration involved in HNSCC tumor progression. 28 Studies also implicated chemokine CXCL12 (SDF-1) and its receptor CXCR4 in migration and homing of multiple myeloma cells to the bone environment. 20 SDF-1/CXCR4 signaling has been shown to be involved in lymph node metastasis in OSCC. 29 Therefore, it is necessary to define a pathologic role for tumor-derived factors in bone invasion and the osteolysis process. In this study, we identified CXCL13/CXCR5 receptor expression and an autocrine regulation of CXCR5 receptor expression in OSCC cells. Furthermore, autocrine/paracrine activation of chemokine receptor-7 (CCR7) in HNSCC has been shown to mediate cell survival and metastasis. 30 Inflammatory cytokines such as IL-1b have been shown to induce CXCL13 production in differentiated osteoblasts 31 and human osteoblasts have been shown to express functional CXCR3 and 5 receptors. 32 Also, CXCL12 (SDF-1) and CXCL13 (BCA-1) chemokines have been shown to directly modulate proliferation and type I collagen expression in osteoblastic cells in osteoarthritis. 33 Therefore, high levels of CXCL13 production by OSCC cells could modulate osteoclast activity and gene expression in the tumor-bone environment. NF-jB has been shown to be constitutively activated in SCC and inactivation of NF-jB has been shown to suppress a malignant phenotype. 34 In this study, we show OSCC production of soluble RANKL and that CXCL13 chemotaxis of preosteoclast cells stimulated osteoclast differentiation. Similarly, SDF-1 binding to the CXCR4 receptor on human osteoclast precursor cells has been shown to promote chemotactic recruitment, development and survival of osteoclasts. 20 However, cocultures of human OSCC cells (BHY and HSC-2 cells) with primary osteoblasts indicated suppression of OPG expression which is implicated in osteoclast differentiation. 35 Thus, it is possible that OSCC cells may inhibit OPG expression in the tumor-bone environment to enhance osteoclastogenesis and osteolytic activity. Conversely, bone derived SDF-1 from osteoblastic cells has been shown to stimulate IL-6 release from oral cancer cells which promotes osteoclastogenesis. 36 Previously it has been reported that IL-8 in saliva and IL-6 in serum are potential biomarkers for oral cancers. 37 MMPs have been implicated in tumor invasion and metastasis of carcinomas. IL-8 secreted by OSCC has been shown to contribute to tumor cell invasion through modulation of MMP-7 expression in tumor cells. 38 Furthermore, MMP-7 produced by osteoclasts at the prostate tumor-bone interface in a rodent model has been shown to promote osteolysis through solubilization of RANKL. 39 Therefore, our findings that OSCC produces soluble RANKL which stimulated osteoclast differentiation and that CXCL13 induces high levels of MMP-9 expression in tumor cells suggests that CXCL13 plays an important role in OSCC bone invasion/osteolysis. We found no significant change in the levels of MMP-9 mRNA expression in CXCL13 treated cells in the presence of actinomycin D (data not shown), which suggests transcriptional regulation of MMP-9 expression. However, we cannot exclude the possibility that other factors produced by OSCC cells such as PTHrP and TNF-a may potentiate the actions of CXCL13 to stimulate osteoclast formation. Cytokines such as TNF-a and TGF-b1 have been shown to promote oral cancer invasion through up-regulation of MMP-9 expression. 40 The in vivo model that we developed of OSCC over calvaria in athymic mice mimicked tumor invasion and osteolysis by activated osteoclasts and abundant expression of CXCL13 and MMP-9 by tumor cells in the bone environment. Furthermore, increased rate of osteolysis observed with SCC14a cells which produce high levels of CXCL13 when compared with SCC12 could implicate CXCL13 as a prognostic marker of OSCC tumor growth/osteolysis. Also, our findings that shRNA knock-down of CXCL13 expression in OSCC cells results in a significant decrease in tumor and in calvaria from control mice treated with PBS alone were counted using a micrometer scale and expressed per mm 2 (*p < 0.05) (E) Immuno histochemical analysis of CXCR5, CXCL13 and MMP-9 expression in SCC14a tumors from athymic mice. Immunostaining with antibodies specific to CXCR5, CXCL13 and MMP-9 was performed and a rabbit non-specific IgG served as control. (f) Histochemical staining of calvaria isolated from control scrambled shRNA, CXCL13 shRNA knock-down SCC14a cell tumor-bearing athymic mice were performed with H&E and TRAP activity staining for osteoclasts. (g) Osteoclasts numbers in the control and CXCL13 shRNA knock-down tumor-bone interface were counted using a micrometer scale and expressed per mm 2 (*p < 0.05). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Cancer Cell Biology
Pandruvada et al.
osteolysis further suggests CXCL13 as a therapeutic target for OSCC tumor osteolysis in vivo. The calvaria model of OSCC offers advantages for further unraveling the molecular mechanisms associated with osteoclast activation at the tumorbone interface in a stage-specific manner. Thus, our results implicate a functional role for CXCL13 in OSCC tumor inva-sion into bone and may be a potential therapeutic target for prevention of osteolysis associated with OSCC in vivo.