Interaction between CXCR4 and CCL20 pathways regulates tumor growth

Abstract

The chemokine receptor CXCR4 and its ligand CXCL12 is overexpressed in the majority of tumors and is critically involved in the development and metastasis of these tumors. CXCR4 is expressed in malignant tumor cells whereas its ligand SDF-1 (CXCL12) is expressed mainly by cancer associated fibroblasts (CAF). Similarly to CXCR4, the chemokine CCL20 is overexpressed in variety of tumors; however its role and regulation in tumors is not fully clear. Here, we show that the chemokine receptor CXCR4 stimulates the production of the chemokine CCL20 and that CCL20 stimulates the proliferation and adhesion to collagen of various tumor cells. Furthermore, overexpression of CCL20 in tumor cells promotes growth and adhesion in vitro and increased tumor growth and invasiveness in vivo. Moreover, neutralizing antibodies to CCL20 inhibit the in vivo growth of tumors that either overexpress CXCR4 or CCL20 or naturally express CCL20. These results reveal a role for CCL20 in CXCR4-dependent and -independent tumor growth and suggest a therapeutic potential for CCL20 and CCR6 antagonists in the treatment of CXCR4- and CCL20-dependent malignancies.

Figure 1. CXCR4 up-regulates CCL20 production by prostate cancer cell line PC3.

(A–C) Wild-type PC3 cells, CXCR4-transduced PC3 cell line and single cell clone with stable overexpression of CXCR4 (PC3-CXCR4.5) were stained for the control (IgG2a-PE, purple) and CXCR4 antibodies (IgG2a-12G5, green) and analyzed using FACS. PC3 and PC3-CXCR4.5 cells were stimulated with CXCL12 at concentrations of 50, 250 and 1,000 ng/ml for 48 hours, harvested and viable cells were counted using PI staining and FACS analysis. PC3 and PC3-CXCR4.5 cells were (5×10⁶/mice) were injected subcutaneously into SCID/beige mice. 60 days following the injection, animals were sacrificed, tumor size (cm²) and tumor weight (g) were measured. Data is presented as mean±SE from 5 mice. (D) PC3 and PC3-CXCR4.5 cells were stimulated with CXCL12 at concentrations 5 and 500 ng/ml for 24 hours, total RNA was extracted, reverse-transcribed and subjected to quantitative PCR for CCL20. PCR analysis was carried out in triplicates. (E) PC3 and PC3-CXCR4.5 cells were stimulated with various concentrations of CXCL12 (5, 50, 25, 500, and 1,000 ng/ml) for 48 hours and CCL20 secretion was assessed by ELISA. The results represent the average of triplicates±SD (** P<0.05). (F) PC3-CXCR4.5 cells were incubated with CXCL12 at concentration of 500 ng/ml. At the indicated time points CCL20 was assessed in extra-cellular (culture medium) and intra-cellular (whole cell lysate) fractions using ELISA method. (G) In order to inhibit CXCR4 signaling, PC3 and PC3-CXCR4.5 cells were cultured with anti-CXCL12 antibodies or pertussis toxin alone or in combination with CXCL12 during 48 hours, and CCL20 secretion was assessed by ELISA. CCL20 secretion in PC3 and PC3-CXCR4.5 cells was also inhibited using JAK-2 inhibitor AG-490 at 1 µm/ml. and the MEK inhibitor- PD98059 (20 µM). The results represent the average of triplicates±SD (** P<0.05).

Figure 2. Regulation of CCL20 expression and function.

(A–B) CXCR4 expression levels of PC3-CXCR4.5 cells 48 hours following the transfection with control non-specific siRNA and specific anti-CXCR4 siRNA. The cells were stained for the control and CXCR4 antibodies and evaluated by FACS. Semi-quantitative RT-PCR analysis of CXCR4 and CCL20 mRNA of the siRNA-transfected PC3-CXCR4.5 cells 48 hours following the transfection. β-actin confirmed comparable loading of RT-PCR products in each lane. Control and CXCR4 siRNA-transfected PC3-CXCR4.5 cells at 48 hours post-transfection were incubated with CXCL12 500 ng/ml for an additional 48 hours. CCL20 secretion was assessed by ELISA. The results represent the average of triplicates±SD (** P<0.05). (C) CCR6 expression in prostate cancer cell lines PC3, LaNCAP, 22Rv1 and DU145 was evaluated by FACS and PCR whereas CCL20 expresion was evaluated by PCR and ELISA assays. Purple line represents mouse IgG control antibody, green line represents staining with CCR6 monoclonal antibody. (D) PC3 and PC3-CXCR4 cells were incubated with various concentration of CCL20 for 6 days. Following 3 days of incubation, the medium with or without CCL20 was renewed. On day 6, the cells were harvested and viable cells were counted using PI staining and FACS analysis. In addition, in order to determine proliferation of PC3 cells, the cells were labeled with BrdU (10 µM) during the last 16 hours of incubation and processed for BrdU detection using specific anti-BrdU FITC-conjugated antibody and FACS analysis. Data is presented as mean±SD from triplicates (** P<0.05). Data is representative of two separate experiments. (E–F) PC3 and PC3-CXCR4.5 cells either untreated or treated with various concentrations of CCL20 were placed on collagen I- or fibronectin-coated plates (10 µg/ml) for 30 minutes. Non-adherent cells were washed twice with cold PBS. Adherent cells were collected in 300 µl FACS buffer with 5 mM EDTA and counted by FACS. Data is presented as mean±SD from triplicates (** P<0.05). In addition, PC3 cells (E) that demonstrated increased adhesion to collagen I and fibronectin in response to stimulation with CCL20, were co-incubated with CCL20 and PTX (100 ng/ml) and were allowed to adhere to collagen I- and fibronectin-coated plates.

Figure 3. Regulation of CCL20 expression and function in various tumor cells.

(A) CCL20 mRNA and protein expression in CCL20-transfected PC3 single-cell clones tested by semi-quantitative RT-PCR and ELISA. (B) PC3-CCL20 clones were seeded at 2×10⁴ cells/1 ml per well into a 24-well plate and incubated for 6 days. On day 6, the cells were harvested and viable cells were counted using PI staining and FACS analysis. Data is presented as mean±SD from triplicates (** P<0.05). (C) PC3-CCL20 single-cell clones were grown to confluence, harvested, resuspended in their conditioned culture medium and allowed to adhere to collagen I- and fibronectin-coated plates for 30 minutes. Non-adherent cells were washed, and adherent cells were collected in 300 µl FACS buffer with 5 mM EDTA and counted by FACS. Data is presented as mean±SD from triplicates (** P<0.05). (D) Leukemic cell lines NB4 and HL60, primary human leukemic blasts, HT-29 cells, and normal human keratinocytes were incubated with various concentrations of CXCL12 for 48 hours. CCL20 secretion to culture medium was assessed using ELISA method. (E) CCR6 mRNA expression in leukemic cell lines NB4 and HL60 and colon cancer HT-29 cells assessed by semi-quantitative RT-PCR. β-actin confirmed comparable loading of RT-PCR products in each lane. (F) HL60 and HT-29 cells either untreated or treated with various concentrations of CCL20 were placed on collagen I coated plates (10 µg/ml) for 30 minutes. Non-adherent cells were washed twice with cold PBS. Adherent cells were collected in 300 µl FACS buffer with 5 mM EDTA and counted by FACS. Data is presented as mean±SD from triplicates (** P<0.05). Data is representative of three separate experiments.

Figure 4. CCL20 regulates CXCR4 dependent and independent growth of tumor cells.

(A) Effect of CCL20 stable expression on prostate tumor growth. PC3-CCL20.30, PC3-CCL20.10 and PC3-mock transfected cells (5×10⁶/mouse) were injected subcutaneously into SCID/beige mice. 48 days or 75 days following the injection, animals were sacrificed, and xenograft tumors generated by PC3-mock and PC3-CCL20.30 injected cells were purified . Tumor size (cm²) was measured twice a week using caliper. Results are representative of three independent experiments with five mice in each group. Data is presented as mean±SE from five mice. (B) H&E staining of paraffin-embedded tumor tissue sections derived from PC3-mock and PC3-CCL20.30 tumors on day 48. Black arrows signify non-invasive borders of PC3-mock tumor (B,a), small blood vessel in PC3-mock tumor (B,b), aberrant blood vessels in PC3-CCL20.30 tumor (B, c,d), original magnification of ×200 is shown. (C) Vessel functionality (ΔSo₂) was measured by fMRI. Functionality of the vasculature was tested during inhalation of air-CO₂ and carbogen (95% oxygen+5% CO₂) in mice implanted with PC3-mock cells or with PC3-CCL20.30 cells. ΔSo₂ values from PC3-mock cells and PC3-CCL20.30 are shown. The mean±SD values of ΔSo₂ were calculated from whole tumor, and normalized to contra-lateral muscle, pooling data from 9 mice from the PC3-CCL20.30 group and 5 mice from the PC3-mock group (four slices/mouse; p<0.001). (D) Adhesion of PC3-CCL20.30 cells to collagen I. PC3-CCL20.30 cells either umstimulated or stimulated with 50 ng/ml of CCL20 with or without co-incubation with neutralizing anti-CCL20 antibodies (10 µg/ml) were placed on collagen I-coated plates (10 µg/ml) for 30 minutes. Non-adherent cells were washed twice with cold PBS. Adherent cells were collected in 300 µl FACS buffer with 5 mM EDTA and counted by FACS. Data is presented as mean±SD from triplicates (** P<0.016). (E) PC3-CCL20.30 cells (5×10⁶/mouse) were injected subcutaneously into SCID/beige mice. Twenty-four hours after the cell injection, mice started to get subcutaneous injections of anti-human CCL20 antibodies or isotype control antibodies, 20 µg of antibody per injection, three times a week, during four weeks. Tumor size (cm²) was measured once a week using caliper. Results are representative of two independent experiments with ten mice in each group. Data is presented as mean±SE from ten mice. (F) On day 64 following the cell injection, the experiment was terminated, animals were sacrificed; subcutaneous tumors were removed, measured and weighted. Data is presented as mean±SE from ten mice in each group (** P<0.0002). (G) PC3-CXCR4.5 cells (5×10⁶/mouse) were injected subcutaneously into SCID/beige mice. Twenty-four hours after the cell injection, mice started to get subcutaneous injections of anti-human CCL20 antibodies or isotype control antibodies, 20 µg of antibody per injection, three times a week, during four weeks. Tumor size (cm²) was measured once a week using caliper. On day 55 following cell injection, the experiment was terminated; animals were sacrificed, subcutaneous tumors were removed, measured and weighted. Data is presented as mean±SE from ten mice in each group (** P<0.0027). (H) HT-29 cells (2×10⁶/mouse) were injected subcutaneously into nude mice. Twenty-four hours after cell injection, mice started to get subcutaneous injections of anti-human CCL20 antibodies or isotype control antibodies, 20 µg of antibody per injection, five times a week, during two weeks. On day 17 following cell injection, the experiment was terminated, animals were sacrificed; subcutaneous tumors were removed, measured and weighted. Data is presented as mean±SE from ten mice in each group (** P<0.0002).

Figure 5. CCL20, CCR6 and CXCR4 expression in prostate cancer cell lines, in primary prostate tumor tissue and in normal prostate tissue.

(A) Expression of CCL20 and CCR6 in human prostate cancer tissue and in human normal prostate tissue. Immunohistostaining of prostate cancer and normal specimens using the polyclonal antibody for CCL20 and the monoclonal antibody 140706 for CCR6. Original magnification of ×400 is shown. CCL20 and CCR6 expression was observed in endothelial and fibromuscular cells of prostate samples (signed with black arrows). A commercially available array including 52 samples (CA2) from patients with prostate cancer was stained for CCL20 (A′) and CCR6 (A″). Expression was scored on two levels: low or negative expression, and high expression. Statistical analysis of the immunohistochemical staining was performed using two-tailed Mann-Whitney test. (B) Comnmon expression pattern of CCL20 and CXCR4 in human prostate cancer. Immunohistostaining of prostate cancer specimens using the polyclonal antibody for CCL20 and the monoclonal antibody 12G5 for CXCR4. Original magnification of ×400 is shown. CCL20 and CXCR4 expression was observed in endothelial and fibromuscular cells of prostate samples. (B′) A commercially available array including 48 samples (CA3) from patients with prostate cancer was stained for CCL20 and CXCR4. Expression was scored on three levels: low, intermediate, and high expression. Samples that demonstrated the same scoring in CCL20 and CXCR4 expression levels were classified as having similar expression pattern. Samples where scores in CCL20 and CXCR4 expression differed in one level or more were classified as different expression patterns.