Osteoblast-secreted factors enhance the expression of dysadherin and CCL2-dependent migration of renal carcinoma cells


  • Yvonne Schüler,

    1. Section for Transplantation Immunology and Immunohematology, Center for Medical Research, University of Tübingen, Germany
    Search for more papers by this author
    • The authors declare no competing financial interests.

  • Cornelia Lee-Thedieck,

    1. Department of New Materials and Biosystems, Max Planck Institute for Metals Research, Stuttgart, Germany
    Search for more papers by this author
    • The authors declare no competing financial interests.

  • Konstanze Geiger,

    1. Section for Transplantation Immunology and Immunohematology, Center for Medical Research, University of Tübingen, Germany
    Search for more papers by this author
  • Tatjana Kaiser,

    1. Department of Gynecology and Obstetrics, University of Tübingen, Germany
    Search for more papers by this author
  • Yoshinori Ino,

    1. Division of Pathology, National Cancer Center Research Institute, Tokyo, Japan
    Search for more papers by this author
  • Wilhelm K. Aicher,

    1. Center for Regenerative Biology and Medicine, University of Tübingen, Germany
    2. Department of Orthopedic Surgery, Center for Medical Research, University of Tübingen, Germany
    Search for more papers by this author
  • Gerd Klein

    Corresponding author
    1. Section for Transplantation Immunology and Immunohematology, Center for Medical Research, University of Tübingen, Germany
    • Center for Medical Research, Section for Transplantation Immunology and Immunohematology, Waldhörnlestrasse 22, 72072 Tübingen, Germany
    Search for more papers by this author
    • Tel: +49-7071-29-84465, Fax: +49-7071-29-5567


Renal cell carcinoma (RCC) frequently metastasizes to the bone marrow. These metastases are characterized by extensive osteolytic lesions. The mechanism, however, by which RCC cells metastasize to bone marrow remains poorly understood. To unravel the role of bone marrow cells in this context, we performed cell adhesion and migration assays using human RCC cell lines to analyze the influence of resident bone marrow cells on renal tumor cells. The strongest adhesion of RCC cells was observed to osteoblasts. Moreover, conditioned medium of osteoblasts (OB-CM) significantly increased RCC cell migration. By gene expression analysis dysadherin was identified as a transcript whose expression could be elevated more than twofold in RCC cells when exposed to OB-CM. Suppression of dysadherin expression in RCC cells by siRNA reduced their ability to migrate in the presence of OB-CM. Furthermore, the RCC cells secreted high amounts of the chemokine CCL2 when tumor cells migrated under the influence of osteoblast-secreted factors. CCL2 neutralization strongly reduced the migratory ability of the RCC cells. Silencing the expression of dysadherin in RCC cells resulted in a twofold reduction of CCL2 protein expression indicating a dysadherin-dependent expression of the chemokine. Taken together, our data show that osteoblasts are the major cell type of the bone marrow that affect RCC cells by secreting factors that increase the expression of dysadherin and CCL2 in the tumor cells leading to enhanced cell migration. These data suggest an osteoblast-induced autocrine mechanism for a facilitated homing of RCC cells to the bone marrow.

Renal cell carcinoma (RCC), the most common malignant tumor arising from the kidney, accounts for 3% of adult malignancies. Metastases to bone occur in 35–40% of RCC, which are characterized by excessive osteolytic lesions leading to severe bone pain and skeletal complications including pathological fractures and spinal cord compression.1 These complications result from the ability of RCC cells to disrupt the interplay between osteoblasts and osteoclasts which regulates bone remodeling leading to pathological bone loss. The mechanism leading to bone tropism remains unknown. However, emerging evidence exists that the interaction with the bone marrow microenvironment may play a crucial role in the homing process of disseminated tumor cells.2, 3

Cancer cell metastasis to distant organs is a multistep process that involves the detachment of cancer cells from the primary tumor, entering the circulation and survival therein, extravasation and immigration and growing in a target organ.4 In prostate cancer metastasis to bone it has been shown that these tumor cells preferentially interact with specific cells of the bone marrow including osteoblasts and bone marrow endothelial cells.5 These interactions could contribute to the colonization of bone by prostate tumor cells.

In our study, dysadherin proved to be an important factor in the interaction between RCC cells and osteoblasts. Dysadherin, also known as FXYD Domain Containing Ion Transport Regulator 5 (FXYD5) or Related to Ion Channel (RIC), was cloned and characterized as a cancer-associated membrane glycoprotein.6 It belongs to the FXYD family of single span type I membrane proteins, which interact with and modulate properties of the Na+, K+-ATPase.7, 8 Previous studies indicated that dysadherin is strongly expressed in several human tumor types such as colon, pancreatic or breast tumors, whereas only a limited number of normal cell types such as endothelial cells or T-lymphocytes express dysadherin.6, 9–12 Dysadherin can down-regulate E-cadherin which results in decreased cell adhesion, increased cell motility and metastasis.6, 12 Furthermore, dysadherin was also shown to be able to promote invasion and metastasis of cancer cells that do not express E-cadherin.13

A prominent modulator of metastatic growth in the bone microenvironment is CCL2, the chemokine (C-C motif) ligand 2, also known as monocyte chemoattractant protein-1 (MCP-1). CCL2 belongs to the CC chemokine family. Beside its function as a chemoattractant for monocytes, memory T-cells and natural killer cells, CCL2 has also been reported to be implicated in the regulation of cancer cell growth, angiogenesis and metastasis of different tumors such as prostate cancer, breast, colon and RCC.14–16 Moreover, especially in prostate cancer, CCL2 has been identified as a prominent modulator of metastatic growth in the bone microenvironment. In animal models, it could be shown that inhibition of CCL2 decreased marcrophage infiltration, osteoclast function and inhibited prostate cancer growth in bone.17–19

To date, little is known about the dissemination of RCC cells to the bone. Therefore, the aim of our study was to characterize the molecular and cellular interactions of RCC cells with cells of the human bone marrow microenvironment to clarify whether the bone marrow cells can attract RCC cells and whether the bone marrow microenvironment is a favorable site for RCC cell homing. Therefore we investigated (i) whether different bone marrow cell types influenced the adhesion and migration of RCC cells, (ii) which factors were differentially expressed in the RCC cells and (iii) which factors were secreted by RCC cells after exposure to bone marrow cells.

Material and Methods

Cell culture and reagents

Human renal carcinoma cell lines A-498 and ACHN as well as osteoblastic cell lines MG-63 and CAL72 were purchased from ATCC (Manassas, VA) and cultured in RPMI 1640 medium (Invitrogen, Karlsruhe, Germany) supplemented with 10% (v/v) fetal calf serum (PAA Laboratories, Cölbe, Germany). The established human bone marrow stromal cell lines L87/4, L88/5 and HS-5 and the osteoclast precursor cell line GCT (ATCC) were cultivated in the supplier's recommended growth media.20, 21

Human bone specimens were obtained after written consent from the patients and approval by the local ethics committee. Primary human osteoblasts (pOB) were isolated from bone waste after endoprosthesis surgery according to an established protocol.22 The cells were maintained in DMEM (4.5 g/L glucose, L-Gln; Invitrogen) supplemented with 20% fetal calf serum, 2% minimum essential medium vitamin solution, 1% fungizone, 50 μg/mL ascorbic acid and 1.4 mM β-glycerophosphate (Sigma Aldrich, Taufkirchen, Germany).

Osteoclast precursors were isolated from buffy coat cell preparations and generated according to an established protocol.23 Tartrate-resistant acidic phosphatase (TRAP) staining of mature osteoclasts was performed after 17 days of culture according to manufacturer's instruction (Sigma Aldrich). Primary bone marrow stromal cells (CFU-Fs) were obtained from bone marrow aspirates and grown in RPMI 1640 medium supplemented with 10% (v/v) fetal calf serum and 10% horse serum. All cell cultures were performed at 37°C and 5% CO2. Conditioned media were prepared from all bone marrow cell types except mature osteoclasts by cultivating these cells at 90% confluence in serum-free medium for 24 hr. Conditioned media from primary osteoblasts were designated as pOB-CM and conditioned media from the osteoblastic cell line MG-63 as OB-CM, respectively. Serum-free medium was used as control medium.


The monoclonal antibody against human dysadherin (clone NCC-M53) has been described earlier.10 The monoclonal anti-CCL2 and anti-GM-CSF antibodies and an IgG control antibody were obtained from R&D Systems (Wiesbaden, Germany). The neutralizing anti-human interleukin-6 antibody was purchased from Immunotools (Friesoythe, Germany). The antibody W6/32.HL, which recognizes an epitope of the heavy chain of MHC class I antigens, was used as another control antibody.24

Cell-cell adhesion assay

Osteoblastic cell lines MG-63 and CAL72, primary human osteoblasts or bone marrow stromal cells were seeded in 48-well plates and grown for less than 24 hr. 1 × 106 RCC cells were stained with the fluorescent dye BCECF-AM (2′, 7′-bis (2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester; Sigma Aldrich) for 15 min. The confluent monolayer of the osteoblasts or stromal cells was washed twice and 1 × 105 labeled RCC cells in control medium were allowed to attach to the bone cells for 30 min. After washing, the fluorescence of the attached tumor cells was quantified with a fluorometer (Fluoroskan Ascent; Labsystems, Thermo Scientific, Dreieich, Germany). The percentage of adherent tumor cells was calculated as follows:

equation image

The fluorescence of 1 × 105 labeled RCC cells alone was used as positive control and bone marrow cells without adding RCC cells were taken as negative control. The assays were carried out in triplicate. After quantification, RCC cells attached to bone cells were visualized under a light microscope (Zeiss Axiovert, Göttingen, Germany).

Cell migration assay

Cell migration was investigated by wound healing and transwell assays. In a wound healing assay, 2.5 × 105 RCC cells were seeded in 12-well plates (Corning, NY). After 24 hr, the wells were scraped with a pipette tip and washed with serum-free medium. The remaining cells were incubated with 600 μL conditioned medium of osteoblasts, osteoclasts and stromal cells, respectively, in triplicates. Each generated wound was imaged at the same position using a 10× objective lens at 0, 24 and 48 hr. Additionally tumor cells were incubated with 5 mM hydroxyurea (Sigma Aldrich) to exclude cell proliferation.

The transwell migration assay was performed as follows: RCC cells at a density of 1 × 104 in control medium were plated in the upper chamber of 24-well culture plates with 8 μm polycarbonate membranes (Corning) that were coated with 0.05% gelatine (Biochrom, Berlin, Germany). The lower chamber was filled with control medium or with conditioned medium of osteoblastic cells. After incubation for 24 hr, remaining cells inside the inserts were removed and membranes were fixed and stained with 0.5% crystal violet. The number of cells migrated through the membrane was counted in three random areas under a light microscope. Migration assays were performed in the absence or presence of neutralizing antibodies against dysadherin, CCL2, GM-CSF and interleukin-6 or with an isotype control.


Total RNA from A-498 cells cultured in OB-CM and control medium was isolated with the RNeasy Mini Kit according to the manufacturer's protocol (Qiagen, Hilden, Germany). Further purification was achieved by DNaseI digestion (Qiagen). Using the TrueLabeling-AMP 2.0 kit (SABiosciences, Frederick, MD), the RNA was reversely transcribed into cDNA and converted into biotin-labeled cRNA using biotin-16-UTP (Roche, Mannheim, Germany) by in vitro transcription. Prior to hybridization, the cRNA probes were purified with the ArrayGrade cRNA cleanup kit (SABiosciences, Frederick). The purified cRNA probes were hybridized to the pretreated Oligo GEArray Microarrays Human Tumor Metastasis (OHS-028) and Human Osteogenesis (OHS-026) (SABiosciences, Frederick) which covers 113 genes on each array. After washing, array spots binding cRNA were detected using alkaline phosphatase-conjugated streptavidin and CDP-Star as the chemiluminescent substrate. Chemiluminescence was detected with the CCD camera system DIANA III (Raytest, Straubenhardt, Germany). Data were analyzed using the GEArray Expression Analysis Suite software.

Quantitative RT-PCR analysis

Total RNA was isolated from A-498 cells after 24 hr cultivation in OB-CM and control medium. RNA samples were reversely transcribed using oligo-dT primers and the SuperScript III First-Strand Synthesis System (Invitrogen). Quantitative PCR was performed in a LightCycler (Roche) using SYBR Green I as the detection system (Light Cycler® FastStartDNA Masterplus SYBR Green I; Roche). Primer sequences detecting human dysadherin were: FXYD5 for: 5′-CCT CTG GTC GCC TGT GTC-3′ and FXYD5 rev: 5′-TGG GCT GGA GTT CTG TGT AG-3′. Pyruvate dehydrogenase (PDH) was used as reference gene and primer sequences were PDH for: 5′-GGT ATG GAT GAG GAG CTG GA-3′ and PDH rev: 5′-CTT CCA CAG CCC TCG ACT AA-3′. The results were analyzed by using the LCDA Software Version 3.5.28.

Preparation of cell lysates

Cells were washed with ice-cold PBS and lysed on ice for 1 hr using lysis buffer containing 0.5% Triton X-100, 2 mM CaCl2 in PBS, pH 7.4 and a complete mini EDTA-free protease inhibitor cocktail (Roche). Lysed cells were scraped off and treated with ultrasound. The insoluble cell debris was removed by centrifugation. Protein concentrations were determined photometrically using the Pierce BCA™ protein assay kit (Thermo Scientific, Rockford, IL).

Immunoblot analysis

Cell lysates were separated on 10% SDS-polyacrylamide gels. Separated proteins were blotted onto PVDF membranes by semi-dry blotting. Membranes were incubated with anti-dysadherin monoclonal antibody NCC-M53 (1:500) and anti-vinculin monoclonal antibody (1:2,000; Cell Signaling Technology, Danvers, MA) overnight at 4°C followed by an incubation with secondary antibody conjugated to alkaline phosphatase (1:400; DAKO, Glostrup, Denmark). Signals were visualized by using the NBT/BCIP detection kit (Sigma Aldrich). For quantitative determination of dysadherin expression in RCC cells infrared-labeled secondary antibodies were used and signals were measured using the Odyssey system (LI-COR Biosciences, Bad Homburg, Germany).

si-RNA knockdown of dysadherin

RCC cells plated in 24-well plates were transfected with small interfering RNA (siRNA) designed against human dysadherin (FXYD5/HSS147589, FXYD5/HSS147590, FXYD5/HSS147591). Two scrambled oligonucleotides (MED_GC, HIGH_GC) and a oligonucleotid against human GAPDH served as control siRNAs. All siRNAs were obtained from Invitrogen. Transfection of siRNAs was carried out using Lipofectamine™ RNAiMAX according to manufacturer's instructions (Invitrogen). Cells transfected with siRNAs were used 24 hr post-transfection for migration assays and protein isolation.

Cytokine array

The relative expression levels of cytokines secreted by RCC cells after cultivation in OB-CM were determined with a human chemokine antibody array according to manufacturer's instructions (R&D Systems). Signals were detected by enhanced chemiluminescence. The x-fold changes among the different samples were analyzed using ImageJ software (ImageJ 1.42q, Wayne Rasband National Institutes of Health). For each spot, the net density gray level was determined by subtracting the background level from the total raw density gray levels. The data after background subtraction were normalized according to positive control densities.

Quantification of human MCP-1/CCL2 by enzyme-linked immunosorbent assay (ELISA)

In a wound healing assay, the RCC cell line A-498, either un-transfected or transfected with dysadherin siRNA and scrambeled siRNA, was cultivated in control medium or OB-CM for 24 hr. All culture supernatants of A-498 cells as well as OB-CM alone were analyzed by ELISA for CCL2 expression according to manufacturer's instructions (R&D Systems).

Statistical analysis

All values are expressed as mean ± standard derivation (SD). For the determination of statistical significance, parametric t-tests and analysis of variance tests were performed using GraphPad Prism 4 software (Version 4.02). Differences were considered to be significant for p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***) with increasing degrees of significance.


RCC cells differentially adhere to cells of the human bone marrow

To analyze the interaction of RCC cells with bone marrow resident cell types, we performed cell adhesion assays with cultivated osteoblasts, osteoclasts and bone marrow stromal cells. Osteoblasts or bone marrow stromal cells were cultivated in a monolayer and fluorescently labeled renal carcinoma cells were allowed to attach. A-498 cells exhibited a significantly stronger adhesion of around 50% to primary osteoblasts and the osteoblastic cell lines MG-63 and CAL72 compared to ∼ 35% adhesion to bone marrow stromal cells (Fig. 1a). When RCC cells were allowed to adhere to mature osteoclasts, these cells exhibited an anti-adhesive effect on the tumor cells and were only surrounded by RCC cells (Fig. 1b). These results indicate that the RCC cells analyzed preferentially adhere to osteoblasts.

Figure 1.

Cell adhesion of RCC cells to bone marrow-derived cell types. (a) Cell attachment of A-498 cells to the osteoblastic cell lines MG-63, CAL72, two primary osteoblast cultures (pOB-1 and -2) and three bone marrow stromal cell lines was quantified. The RCC cell line A-498 showed a stronger attachment to osteoblastic cells compared to the bone marrow stromal cells. Values represent the mean of three independent experiments (**p < 0.01; *p < 0.05). (b) No cell binding of RCC cells to mature osteoclasts was observed after TRAP staining. Scale bars: 500 and 100 μm, respectively.

Osteoblast-conditioned medium promotes migration of RCCs

Conditioned media of primary osteoblasts, the osteoblastic cell line MG-63, osteoclast-like cells (GCT) and stromal cells (CFU-Fs and L88/5) were used in wound healing assays. Conditioned media of osteoblasts (primary osteoblasts as well as the osteoblastic cell line MG-63) strongly enhanced the migration of the RCC cell line A-498 when compared to the conditioned media of osteoclast-like cells (GCT), stromal cells (L88/5 and CFU-F) and control medium (unconditioned medium) after 24 hr (Fig. 2a). The addition of the proliferation inhibitor hydroxyurea revealed that the observed enhanced migratory effect was due to cell migration and not to cell proliferation (Fig. 2b). Transwell migration assays were performed to confirm the migration enhancing effect of osteoblast-conditioned medium on A-498 cells (Fig. 2c). A significant increase in cell migration could be observed when osteoblast-conditioned medium was in the lower chamber compared to the control medium. A similar result was observed for the RCC cell line ACHN (Fig. 2c).

Figure 2.

Osteoblast-secreted factors strongly influence cell migration of RCC cells. (a) The migratory ability of the A-498 cells in conditioned medium (CM) from different bone marrow-derived cell types are shown in these representative micrographs. A strongly enhanced migration of A-498 cells was observed in OB-CM and pOB-CM compared to CM from osteoclasts (GCT), stromal cells (L88/5 and CFU-F) or control medium (N = 3 independent experiments). (b) A wound healing assay with A-498 cells after incubation with OB-CM in the presence of 5 mM hydroxyurea showed that wound closure was not due to cell proliferation. (c) Cell migration of A-498 and ACHN tumor cells towards control medium or OB-CM and pOB-CM was analyzed in transwell migration assays. Osteoblast-conditioned media significantly enhanced RCC cell migration (N = 3 independent experiments; bars, SD; ***p < 0.001 for OB-CM, **p < 0.01 for pOB-CM versus control).

Increased dysadherin expression in RCC cells cultured in osteoblast-conditioned medium

In short-term adhesion assays, RCC cells exhibited the strongest attachment to osteoblasts, but factors secreted by osteoblasts also increased the migration of RCC cells. To investigate the influence of osteoblast-secreted factors on RCC cell gene expression, wound healing assays with A-498 cells in presence of OB-conditioned medium or control medium were performed for 24 hr. RNA isolated from these cells was used for two pathway-specific microarrays. Gene expression profiles revealed a prominent change in mRNA expression of dysadherin in the presence of OB-conditioned medium (Fig. 3a). These results were confirmed by quantitative RT-PCR analysis and revealed a significant increase of dysadherin expression in A-498 cells after cultivation in OB-conditioned medium (Fig. 3b). Western blot analyses showed that the amount of the dysadherin protein was clearly increased when A-498 cells were cultivated in OB-conditioned medium compared to non-conditioned medium (Fig. 3c). For quantification infrared-labeled secondary antibodies were used. The signals obtained for dysadherin were normalized to vinculin expression. Compared to a Jurkat cell lysate which was used as a positive control, untreated A-498 cells expressed dysadherin at a lower level. However, when A-498 cells were cultivated in OB-conditioned or pOB-conditioned medium dysadherin expression was strongly enhanced (Fig. 3d).

Figure 3.

Osteoblast-secreted factors increase the expression of dysadherin in RCC cells. (a) Microarray data of A-498 cells after cultivation in OB-CM or control medium showed elevated signals for dysadherin expression under the influence of the OB-media (Exp. 1–3: independent experiments). (b) The data obtained by the microarrays were confirmed by real-time qRT-PCR (N = 3 independent experiments, *p = 0.0108). (c) Immunoblot analyses of dysadherin expression in lysates of A-498 cells after cultivation in control medium or OB-CM. Different amounts of total protein were loaded and a Jurkat cell lysate was used as a positive control for dysadherin expression. Vinculin was used as loading control. (d) Quantitative immunoblot analyses of dysadherin revealed a strongly enhanced expression of dysadherin in A-498 RCC cells after incubation with OB-CM and pOB-CM.

Knockdown of dysadherin expression drastically reduces RCC cell migration in vitro

To investigate the role of dysadherin in RCC cell migration, adherent A-498 cells were analyzed in wound healing assays in the presence of the anti-dysadherin antibody NCC-M53. When the antibody was added to the OB-conditioned medium, the RCC cells exhibited a drastically impaired cell migration compared to IgG control or to untreated OB-conditioned medium (Fig. 4a). In addition to the antibody treatment, siRNA knockdown of dysadherin in A-498 cells was applied. Western blot analysis of the siRNA transfected cells exhibited a clear knockdown of dysadherin expression with all tested siRNAs compared to scrambled RNA-transfected control cells (Fig. 4b). A drastic reduction in RCC cell migration in osteoblast-conditioned medium was achieved after siRNA treatment which was comparable to the migration observed in the control medium (Fig. 4c).

Figure 4.

Inhibition and knockdown of dysadherin in RCC cells impair cell migration. (a) The micrographs show wound closure of A-498 cells in cell migration assays in control medium or OB-CM in the absence or presence of antibodies against dysadherin (10 μg/mL) or IgG control antibody (10 μg/mL). The anti-dysadherin antibody impaired RCC cell migration (N = 3 independent experiments). (b) Immunoblot of A-498 lysates after knockdown of dysadherin by siRNA transfection. Transfected siRNAs are indicated at the top. Vinculin was used as loading control. (c) Cell migration of A-498 cells is impaired after knockdown of dysadherin. Representative micrographs of A-498 cells after siRNA transfection when cultivated in OB-CM (N = 3 independent experiments).

Expression of tumor cell-derived factors in the presence of osteoblast-conditioned medium

Potential soluble factors that could be involved in the enhanced RCC cell migration in the presence of OB-conditioned medium were analyzed with a human cytokine antibody array. Wound healing assays were performed for 24 hr with the RCC cell line A-498 in OB-conditioned medium and control medium. Then both supernatants were collected and analyzed for cytokine expression. In addition, OB-conditioned medium was analyzed in parallel. A drastic increase in GM-CSF, IL-6 and CCL2 could be detected in the medium of A-498 cells grown in the presence of OB-conditioned medium compared to control medium (Fig. 5a). The two cytokines GM-CSF and IL-6 as well as the chemokine CCL2 were not detected in osteoblast-conditioned medium indicating that they were produced by the tumor cells. As CCL2 showed the strongest induction in A-498 cells after OB-CM treatment, we focused on CCL2 and analyzed its role in RCC cell migration. A similar enhanced secretion of CCL2 was obtained for the RCC cell line ACHN (Fig. 5b). In transwell migration assays, a neutralizing antibody against human CCL2 was able to significantly inhibit RCC cell migration under the influence of osteoblast-conditioned medium (Fig. 5c). Consistent results were also seen in the wound healing assay (Fig. 5d). In contrast, neutralizing antibodies against human GM-CSF or IL-6 did not significantly inhibit tumor cell migration (data not shown).

Figure 5.

Expression of tumor cell-derived chemokines and cytokines in the presence of osteoblast-conditioned medium. (a) Cytokines secreted by RCC cell lines were analyzed using a cytokine antibody array. A wound healing assay was performed with A-498 cells in control medium or OB-CM. Arrays were probed with supernatants from both migration assays and with OB-CM alone. In response to secreted osteoblastic factors the RCC cell line A-498 showed elevated signals for CCL2 (16-fold), GM-CSF (7-fold) and IL-6 (3-fold) compared to cultivation in control medium. (b) This representative picture of an cytokine array indicates the relative CCL2 expression level secreted by RCC cells which was ∼16-fold enhanced in A-498 cells (∼ 24-fold in ACHN cells) after incubation with OB-CM. (c) Transwell migration assay was performed with A-498 cells that were incubated in control medium, OB-CM, OB-CM with anti-CCL2 antibody (10 μg/mL) and OB-CM with IgG control (10 μg/mL). Inhibition of CCL2 resulted in decreased migration of A-498 cells in OB-CM (N = 3 independent experiments, bars: SD). (d) Representative micrographs of a wound healing assay using A-498 cells that were incubated either in control medium, OB-CM, OB-CM supplemented with anti-CCL2 antibody or IgG control. Neutralization of CCL2 reduced A-498 cell migration in OB-CM (N =3 independent experiments).

Effect of enhanced dysadherin expression on CCL2 secretion by RCC cells

Using an ELISA assay, we detected a ∼ 3 fold elevated CCL2 expression of around 6 ng/mL, when cells were in contact with osteoblastic factors for 24 hr compared to control medium, whereas primary osteoblasts alone produced CCL2 at a much lower level of less than 1 ng/mL (Fig. 6a). Dysadherin knockdown by siRNA treatment caused a ∼2-fold decrease in the level of CCL2 protein secretion by A-498 cells being in contact with osteoblastic factors indicating that dysadherin expression seems to be involved in CCL2 secretion by these cells (Fig. 6b).

Figure 6.

Effect of dysadherin on CCL2 expression in RCC cells. (a) The level of CCL2 protein secreted by A-498 cells was determined by ELISA. A-498 cells cultivated in OB-CM showed a ∼3-fold higher CCL2 expression compared to incubation in control medium or OB-CM alone [Columns, mean (N = 3 independent experiments); bars: SD; **p < 0.01; *p < 0.05]. (b) A-498 cells were transfected with dysadherin siRNA or scrambled siRNA and supernatants were analyzed for CCL2 protein expression. Down-regulation of dysadherin resulted in a significantly decreased CCL2 expression [Columns, mean (N = 3 independent experiments); bars: SD; *p < 0.05].


Human osteoblasts secrete so far unidentified factors which enhance the expression of dysadherin and chemokine (C-C motif) ligand 2 (CCL2) in RCC cell lines thereby stimulating the migration of these tumor cells. Silencing dysadherin as well as neutralizing CCL2 in RCC cells resulted in reduced cell migration in response to osteoblast-secreted factors. Moreover, dysadherin knockdown in renal tumor cells significantly reduced their CCL2 secretion.

When RCC metastasizes to bone, it becomes highly resistant to radiation therapy and chemotherapy, whereas progress is seen after surgery or in targeted therapies including multikinase inhibitors.25–28 However, to evaluate an effective treatment or even prevent RCC bone metastasis, it is essential to understand the unique role of the bone marrow microenvironment in this process and the mechanisms leading to skeletal metastasis of RCC.

In our study, we demonstrate that osteoblasts are the major cell type of the human bone marrow with an influence on RCC cells in terms of cell adhesion and migration. Similar studies dealing with prostate cancer also revealed a preferred interaction of cancer cells with osteoblasts.5, 29 Furthermore, OB-conditioned medium strongly influenced RCC cell migration compared to unconditioned medium or conditioned media of other bone marrow-derived cell types such as osteoclasts or stromal cells. Direct cell-cell interactions between osteoblasts and tumor cells do not seem to be necessary for the enhanced migration. The performed gene expression analysis revealed an increased expression of dysadherin in the RCC cells strongly migrating in response to OB-conditioned medium. Suppressing dysadherin in RCC cells by knockdown or antibody treatment reduced the enhanced migration toward osteoblast-conditioned medium indicating that dysadherin up-regulation in tumor cells is directly induced by osteoblast-secreted factors. Dysadherin has been described to down-regulate the cell adhesion molecule E-cadherin on tumor cells, which results in diminished cell adhesion, increased cell motility and metastasis.6, 12 However, human renal proximal tubular cells which give rise to most clear cell RCC do not express E-cadherin,30 and the A-498 RCC cell line used in our study is also devoid of E-cadherin expression.31 Dysadherin was found to be strongly expressed in several human tumors such as colon, pancreatic or breast carcinoma.6, 9–12 In the normal kidney, dysadherin is expressed in the cortex together with other FXYD family members suggesting to play a role in regulating ion flux in normal homeostasis. However, the form of dysadherin found in normal kidney (∼24 kDa) has only a minimal contingent of carbohydrate side chains, whereas in many tumor cells dysadherin is heavily O-glycosylated (Mw ∼50–55 kDa).7, 11 To our knowledge, there are no reports so far describing dysadherin expression in primary RCC tumors or RCC metastasis. Our study showed that in different RCC cell lines dysadherin revealed a molecular weight ranging from 50 to 55 kDa. Our data also indicate that enhanced dysadherin expression in response to osteoblastic factors is involved in RCC cell migration, possibly by down-regulating other cell adhesion molecules or by up-regulating chemokines such as CCL2 as described by Nam et al.13

We also analyzed secreted cytokines that were produced by migrating RCC cells after exposure to OB-conditioned medium compared to unconditioned medium. In response to OB-secreted factors, the RCC cells secreted increased levels of CCL2, GM-CSF and interleukin-6. CCL2 showed the strongest increase in secretion levels. Both IL-6 and CCL2 have been described to attract, differentiate and activate osteoclasts, suggesting an involvement of these cytokines in the formation of osteolytic lesions of bone metastasis.32–34

Neutralizing antibodies against CCL2 strongly inhibited the enhanced RCC cell migration in OB-conditioned medium. A functional autocrine loop can be observed in renal tumor cells in which soluble osteoblastic factors enhance RCC migration by an increased dysadherin expression and CCL2 secretion in renal tumor cells. In contrast, human liver myofibroblasts secreting CCL2 act on hepatoma cells in a paracrine manner promoting migration and invasion.35 CCL2 has also been identified as a prominent modulator of metastatic growth of prostate cancer in the bone marrow microenvironment, and over-expression with resultant promotion of tumor growth has been observed in several tumors such as melanoma, ovarian, lung or colon carcinomas and leukemias.16 Elevated levels of CCL2 have been identified in serum of RCC patients compared to healthy individuals which correlated with disease stages.36 In breast cancer, over-expression of CCL2 is also frequently associated with advanced tumor stages and metastatic relapse. Here CCL2+ breast tumor cells engage CCR2+ monocytic stromal cells to facilitate colonization to lung and bone.37 There is only one study showing that the CCL2 mRNA transcript is mostly affected by dysadherin knockdown in breast cancer cells which may play a role in mediating the prometastatic effect of dysadherin.13 Dysadherin was shown to up-regulate CCL2 expression in part through activation of the NF-κB signaling pathway leading to paracrine and autocrine effects of CCL2.38 Knockdown of dysadherin in the RCC cell line A-498 significantly reduced the CCL2 protein secretion indicating that CCL2 secretion is affected by dysadherin.

In summary, we show that osteoblast-secreted factors strongly stimulated RCC cell migration by enhancing the expression of dysadherin and the chemokine CCL2. This could be an important step in the colonization of renal tumor cells in the bone marrow to find suitable niches and to form bone micro-metastases. Further studies using an animal model as recently described by Strube et al.39 are certainly necessary to investigate the role of dysadherin and CCL2 in metastasis of RCC cells to bone in vivo in more detail. However, our present findings provide new insights in the biology of RCC bone metastasis and suggest new approaches to control RCC tumor bone metastasis formation.


We are grateful to Dr. Andrew Benest (University of Leeds, UK) for critically reading the article. We thank Dr. Bernd Rolauffs (Center for Traumatology, BGU Hospital Tübingen) for his assistance in obtaining the bone specimens. This work was supported by a grant of the IZKF program of the Medical Faculty of the University of Tübingen (grant no. 1686-0-0) and by a stipend (Graduate School GK794) to Yvonne Schüler by the Deutsche Forschungsgemeinschaft.