Chemokines play the key role in initiating immune responses by regulating the attraction and homing of immune cells to the lymphoid and nonlymphoid tissues. CXCL14 is a chemokine that in tumors may act as chemoattractant for monocytes and dendritic cells (DC), which may modulate antitumor immune responses in certain cancers. In this study, we investigated the mechanisms of loss of CXCL14 in prostate cancer cells. Cell treatment with the demethylating agent 5-aza-2-deoxycytidine resulted in the recovery of CXCL14 mRNA and protein expression. Hypermethylated CpG island sequences encompassing the CXCL14 gene promoter were identified. The restoration of CXCL14 by 5-aza-2-deoxycytidine treatment had functional impact, based on the DC chemoattractant activity of conditioned medium from drug-treated cells. Conversely, CXCL14 removal from conditioned media by affinity chromatography abolished its chemotactic properties, confirming that functionally active CXCL14 was generated in prostate cancer cells by relieving its transcriptional silencing with 5-aza-2-deoxycytidine. Our findings offer the first direct evidence for epigenetic regulation of chemokine expression in tumor cells. Cancer Res; 70(11); 4394–401. ©2010 AACR.

Chemokines are the superfamily of proinflammatory polypeptide signaling molecules that selectively attract and activate different cell types in lymphoid and nonlymphoid tissues (1, 2). Chemokines and their receptors are involved in regulating many pathophysiologic conditions such as oncogenesis, infection, allergy, and autoimmunity by modulating cellular attraction, proliferation, angiogenesis, as well as tumor cell growth and spreading (2, 3).

A chemokine, known as CXCL14, was initially named BRAK because of its isolation from the human breast and kidney derived cells (4). Constitutive expression of CXCL14 was observed in a variety of epithelia, including the basal keratinocytes and dermal fibroblasts of skin (5). Using quantitative reverse transcription-PCR (RT-PCR), several groups of investigators independently showed that CXCL14 mRNA and protein ubiquitously expressed in normal tissues but are absent in tumor cell lines and in primary tumors (4, 69).

The potential biological functions of CXCL14 are still under investigation. To date, it was reported that CXCL14 might play a role in the trafficking of natural killer cells to the sites of inflammation or malignancy (10). This chemokine might control the epidermal recruitment of circulating CD14+ dendritic cell (DC) precursors and promote their in situ differentiation into functional DC (11). CXCL14 inhibited angiogenesis stimulated by multiple angiogenic factors (7). Recent data suggest that CXCL14 chemoattracts both activated monocytes (12) and immature DC both in vitro and in vivo (7, 8). In addition to being a potent DC chemoattractant, CXCL14 also increased DC maturation and their functional abilities, which were associated with increased activity of NF-κB (8). The demonstration of a high-affinity binding site for CXCL14 on immature DC is an important finding, which opens new opportunities for the identification and characterization of CXCL14 receptor (7).

DC are known as antigen-presenting cells detected in immature state in virtually every tissue, where they capture antigens followed by maturation and migration to secondary lymphoid organs to activate naive antigen-specific T cells. The capacity of immature DC to migrate into the tumor site in search for antigens is a key to the successful induction of the antitumor immune responses (13). Their absence in many malignant tissues is in agreement with the deficiency of effective antitumor immune responses in cancer patients (14, 15). The fact that CXCL14, a potent DC chemokine, is downregulated or absent in the malignant tissues suggests that CXCL14 may have a unique role in the tumor recognition by the immune system. Because DC could bind CXCL14 with high affinity (7), migrate to CXCL14-expressing tissues, and be activated by CXCL14 (7, 8), it was suggested that the downregulation of CXCL14 expression in tumor tissues might represent a new mechanism of tumor evasion, which allows tumor cells to escape recognition by antigen-presenting cells (8).

Although, we and others have reported the loss of CXCL14 expression in tumor tissues and tumor cell lines, the mechanism of this phenomenon is not clear (68). A significant decrease of CXCL14 mRNA in many human tumor cell lines makes it likely that the suppression of CXCL14 expression occurs at the transcriptional rather than the translational level (4, 6, 8, 9). This transcriptional downregulation of CXCL14 expression may result from genetic alterations and/or epigenetic changes (promoter hypermethylation). Promoter hypermethylation deserves a special attention because it could silence a variety of tumor suppressor genes in several malignant neoplasms and was able to decrease tumor immunogenicity (1619). The role of the epigenetic regulation of chemokine expression in tumor cells has not been yet investigated.

The main goal of the present work was to evaluate the primary mechanisms of the downregulation of CXCL14 expression in prostate cancer. Our results show that coincubation of CXCL14-negative prostate adenocarcinoma cells with a demethylating agent results in the restoration of CXCL14 mRNA and protein expression. Furthermore, demethylation of CpG islands in the promoter region of the CXCL14 gene in tumor cells restores the expression of functional CXCL14 as seen by the restoration of attraction of human DC to the tumor cells. These data point toward promoter hypermethylation as a possible mechanism of the CXCL14 gene suppression in prostate cancer. Using methylation-specific PCR (MSP) and bisulfite-sequencing PCR (BSP), we found that DNA from prostate cancer cells contains hypermethylated CpG island sequences encompassing the transcriptional regulatory region of the CXCL14 gene. Thus, failure to express CXCL14 in tumor cells is a consequence of transcriptional “silencing.” These data are the first demonstration of the epigenetic regulation of chemokine expression in tumor cells, which open new opportunities for pharmacologic regulation of chemokine expression in the tumor microenvironment.

Tumor cell lines and tissues

Prostate cancer cell lines PC-3 and DU145 were obtained from the American Type Culture Collection (ATCC) and cultured in complete RPMI 1640 supplemented with 2 mmol/L l-glutamine, 50 μg/mL gentamicin sulfate, 10 mmol/L HEPES, 10% FCS, and 10 mmol/L nonessential amino acids (Life Technologies). The PZ-HPV-7 cell line was derived from normal epithelial cells from the peripheral zone of the human prostate (ATCC). Formalin-fixed, paraffin-embedded tissue sections of normal prostate (n = 7), benign prostatic hyperplasia (n = 7), and primary prostate adenocarcinoma (n = 10) were from the Tissue Bank, University of Pittsburgh Medical Center. In addition, paired malignant glands of adenocarcinoma and benign prostatic glands were collected from serial frozen sections by needle microdissection technique. Tissue sections (5-μm) were mounted on silicon-coatedslides, stained with H&E, and reviewed by a pathologist to confirm the presence of tumor and histologically normal tissue. Then, the areas of carcinoma and benign tissue were circled and carefully dissected and scraped with a sterile 22-gauge needle. Collected tissues were processed for RNA isolation and for further analysis.

PC3 and DU145 cells were treated with a demethylating agent 5-aza-2′-deoxycytidine (5-aza-dC; Sigma) to determine whether DNA methylation influences a gene expression. Cells were split to low density (1 × 106 cells per T-75 flask) 24 hours before the treatment. Cells were treated with 10 and 25 μmol/L 5-aza-dC for 6 days. CXCL14 protein and mRNA was determined before and after the treatment by immunocytochemistry and RT-PCR, respectively.

Human DC cultures

Human DC cultures were initiated from CD14+ monocytes as described earlier (20). Briefly, peripheral blood mononuclear cells (PBMC) were separated using Histopaque (1.007 g/mL; Sigma) density centrifugation. PBMC were washed and plated at 1 × 107 cells per well in 2 mL of AIM-V medium (Life Technologies) in six-well plates. After a 1-hour incubation at 37°C in a humidified 5% CO2 atmosphere, nonadherent cells were removed. Adherent monocytes were cultured with recombinant human granulocyte macrophage colony-stimulating factor (1,000 U/mL; PeproTech) and human recombinant interleukin-4 (1,000 U/mL; PeproTech) in complete RPMI 1640 for 6 to 7 days and immature DC were used in chemotaxis studies.

Chemotaxis assay

DC migration was evaluated using a 48-well Transwell system (5 μm pore size, Corning Costar) with DC placed in the upper chamber (1 × 106 cells/mL, 100 μL). Assay medium, chemokines, or conditioned medium from tumor cell cultures were added to the bottom chamber (600 μL). Tumor cells were seeded at 1 × 106 in 4 mL of assay medium. Forty-eight hours later, cell-free supernatant was collected. The following groups were used for DC attraction in a bottom chamber: (a) RPMI 1640 containing 1% fetal bovine serum (FBS; assay medium); (b) assay medium containing 5-aza-dC; (c) assay medium with 5-aza-dC after passing through the protein A agarose bound with anti-CXCL14 antibody (Aviva systems; assay media 1, 2, and 3 were used to measure spontaneous DC migration); (d) recombinant murine DC chemokine MIP-1α (20 ng/mL; PeproTech), a positive control; (e) cell-free conditioned media collected from PC3 cells nontreated or (f) treated with 25 μmol/L 5-aza-dC; and (g) CXCL14-free conditioned media (filtrate fraction after passing through the agarose A/anti-CXCL14) collected from PC3 cells treated with 25 μmol/L 5-aza-dC. Migration of DC was assessed after 4 hours of incubation at 37°C, 5% CO2. After incubation, cells transmigrated thought the membrane were collected and acquired on FACScan (Becton Dickinson) for 1 minute for enumeration. Data are reported as the mean numbers of transmigrated cells from triplicate wells.

Elimination of CXCL14 protein from tumor cell supernatant by affinity chromatography

To eliminate CXCL14 protein from the culture medium of PC3 cells treated with 25 μmol/L 5-aza-dC, cell-free culture supernatant was incubated with anti-CXCL14 antibody (Aviva systems) overnight at 4°C to form immune complexes. Immune complexes were added to the protein A agarose beads, placed to the Pierce spin cup columns (Pierce). The columns were placed inside of microcentrifuge tubes and centrifuged for 1 minute at 3,000 g. Filtrate fractions were collected and used for chemotaxis assay. The same procedure was performed with a medium containing 5-aza-dC to be used in chemotaxis assay as a negative control.

Immunohistochemistry/immunocytochemistry

Expression of CXCL14 protein in tissue sections and prostrate cancer cells was determined with anti-CXCL14 monoclonal antibodies (10 μg/mL; R&D Systems). Formalin-fixed and paraffin-embedded tissue blocks were cut into 5-μm sections, dewaxed, and rehydrated. Tumor cells were cytospinned (100 g, 5 minutes) and air dried. Antigens were retrieved by boiling in 1 mmol/L EDTA/NaOH solution (pH 8.0) in a microwave oven. The primary antibodies were applied for 16 hours followed by 30-minute incubation with biotinylated secondary horse anti-mouse antibodies (Jackson ImmunoResearch Laboratories). The color reaction was developed using the avidin-biotin peroxidase technique (Vector Laboratories, Inc.). Tyramide (1:100; Vector Laboratories, Inc.) was applied to amplify the intensity of staining. For a semiquantitative analysis of CXCL14 protein expression, “−” was graded for no expression, “+”was graded for <2% expression, “++” was graded for 10% to 20%, “+++” was graded for >20% to 50%, and “++++” was graded for >50% expression.

Reverse transcription-PCR

RNA was extracted from prostate cancer cell lines and normal prostatic epithelium using Trizol LS reagent (Molecular Research Center) following the supplier's instructions. First-strand complementary DNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Life Technologies) and oligo d(T)16 (Roche Applied Science) from 1 μg of total RNA. PCR was performed using specific primers for CXCL14 (forward, 5′-TCCGGTCAGCATGAGGCTCC-3′, and reverse, 5′-CACCCTATTCTTCGTAGACC-3′; 313 bp). PCR cycles involved incubation at 94°C for 2 minutes; then 94°C for 30 seconds, 61°C for 30 seconds, and 72°C for 1 minute for 40 cycles; and a final extension at 72° for 7 minutes. Primers for β-actin transcript were used as an internal control (323 bp).

The PCR products were analyzed in 1.2% agarose gel containing ethidium bromide. A DNA ladder II (GeneChoice, Inc.) was used as a size marker. For quantitation of the levels of CXCL14 mRNA, the PCR product bands were analyzed using UN-SCAN-IT gel software (Silk Scientific). The bands corresponding to CXCL14 were quantified by pixel density using β-actin as a quantitative control. The PCR index was calculated as sample pixel total/β actin pixel total ratio and expressed as relative percentage.

Bisulfite modification, MSP, and BSP

The state of the CXCL14 gene promoter methylation in prostate cancer cells was evaluated by MSP and BSP. Genomic DNA for deamination, and subsequent MSP and BSP analysis was extracted from tumor cells using QIAamp DNA Mini kit (QIAGEN, Inc.). CpGenome universal unmethylated and methylated DNA sets were used as a control for MSP and BSP (Chemicon International, Inc.). After extraction, 1 μg of DNA from tumor cells and control DNA was used for bisulfite treatment with the EZ DNA Methylation kit (Zymo Research) following the supplier's instructions. MSP was performed to analyze the change in the methylation status of the 5′ CpG islands of the CXCL14 gene during 5-aza-dC treatment in prostate cancer cells. MSP primers selective for unmethylated CXCL14 promoter (forward, 5′-GAGTTTGTTTGTTGTGAGGGTA-3′, and reverse, 5′-ACCAAAAACCTCATACTAACC-3′; 289 bp) and for methylated CXCL14 promoter (forward, 5′-TTAATGAGTTCGTTCGTTGCGAG-3′, and reverse, 5′-ACCAAAAACCTCATACTAACC-3′; 294 bp) were designed. PCR was performed in a reaction mix containing 500 ng bisulfite-converted genomic DNA. Amplifications were run on the DNA thermal cycler 480 (Perkin-Elmer), hot started with an initial denaturation step at 95°C for 7 minutes; the PCR profile were 40 cycles at 95°C for 30 seconds, 62°C to 64°C (depending on Tm values of individual primers) for 30 seconds, and 72°C for 40 seconds; final extension was performed at 72°C for 7 minutes. PCR products were loaded on 1.2% agarose gel.

To determine the methylation pattern of 5′CpG islands of the CXCL14 gene promoter in prostate cancer cells, bisulfite-treated DNA was then subjected to BSP using bisulfite-sequencing PCR primers. Primers for bisulfite sequencing were selected using the Methprimer software program, which identifies the location and structure of CpG islands. The primers BSP-F (5′-GGTTGGGAAGGTTTTTTTTT-3′) and BSP-R (5′-ACCCAACTCTACTCRACTTTCT-3′) were used to amplify CpG islands in the promoter region with the expected product 270 bp in length. Hot-start PCR was carried out for amplifying the CpG islands of the CXCL14 gene. The PCR product was purified, sequenced by the ABI 3100 DNA sequencer (Applied Biosystems), and CpG sites were analyzed for methylation. The methylation level of each CpG site on CXCL14 was estimated by comparing the height of the guanidine peak with the height of the adenine peak on the chromatograms. A single guanidine peak was considered to represent 100% methylation. In the partial methylation, the percentage of methylation was estimated as the ratio of the peak heights of the guanidine-to-guanidine plus adenine signals.

Statistical analysis

For a single comparison of two groups, the Student's t test was used after the evaluation for normality. If data distribution was not normal, a Mann-Whitney rank sum test was performed. One-way ANOVA was used for the comparison of more than two groups after the evaluation for normality. For all statistical analysis, the level of significance was set at a probability of 0.05 to be considered significant.

Downregulation of CXCL14 protein and mRNA expression in human prostate adenocarcinoma

Immunohistochemical analysis of the CXCL14 protein in prostate adenocarcinoma tissues revealed a loss of chemokine expression when compared with the nonmalignant tissues (Fig. 1A). Figure 1A shows that normal prostate (n = 7) and benign prostatic hyperplasia tissues (n = 7) are strongly positive for CXCL14, whereas prostate adenocarcinoma tissues (n = 10) are negative for CXCL14 staining. Comparative analysis of the number of CXCL14+ cells showed that the level of expression of CXCL14 protein in primary prostate adenocarcinoma was between 0% and 2%, whereas for normal prostate tissues, it was 10% to 20%. Finally for benign prostatic hyperplasia, the level of CXCL14 expression were reached up to 60%. Next, we evaluated the expression of CXCL14 protein in PC3 and DU145 human prostate cancer cell lines. Human PBMC–derived monocytes served as a negative control. Human monocytes stimulated with 0.5 μg/mL lipopolysaccharide (LPS; Sigma) were used as a positive control (6). First, human PBMC were incubated for 1 hour at 37°C; nonadherent cells were removed; and adherent monocytes were stimulated with LPS for 6 hours. We found that PC3 cells were CXCL14 negative, whereas DU145 expressed CXCL14 protein (Fig. 1B).

Figure 1.

CXCL14 expression in prostate cancer specimens and cell lines. A, 5-μm sections of primary tumor and control tissues were stained with anti-CXCL14 antibodies. BPH, benign prostatic hyperplasia. B, tumor cells were cytospined onto the microscopic slides and stained with anti-CXCL14 antibodies. A and B, positive staining is developed as a red-brown color. The representative immunohistochemical data from the analysis of 7 to 10 specimens in each group are shown in. C, total RNA was extracted from DU145 and PC3 cells. Lanes 2 and 5, DU145; lanes 3 and 6, PC3. Normal prostate epithelial cells served as a positive control (lanes 1 and 4). RT-PCR was carried out to evaluate the expression of CXCL14 (lanes 1–3) and β-actin (lanes 4–6) mRNA (313 and 323 bp). Lane M, markers. Results from a representative experiment are shown (n = 5). D, total RNA was extracted from paired malignant glands of adenocarcinoma (T1–T5) and adjunct histologically benign prostatic glands (N1–N5). Expression of CXCL14 mRNA was assessed by RT-PCR. E, normal prostate epithelial cells served as a control. All evaluated specimens showed significantly reduced level of CXCL14 expression in prostate cancer tissues when compared with the normal adjunct areas.

Figure 1.

CXCL14 expression in prostate cancer specimens and cell lines. A, 5-μm sections of primary tumor and control tissues were stained with anti-CXCL14 antibodies. BPH, benign prostatic hyperplasia. B, tumor cells were cytospined onto the microscopic slides and stained with anti-CXCL14 antibodies. A and B, positive staining is developed as a red-brown color. The representative immunohistochemical data from the analysis of 7 to 10 specimens in each group are shown in. C, total RNA was extracted from DU145 and PC3 cells. Lanes 2 and 5, DU145; lanes 3 and 6, PC3. Normal prostate epithelial cells served as a positive control (lanes 1 and 4). RT-PCR was carried out to evaluate the expression of CXCL14 (lanes 1–3) and β-actin (lanes 4–6) mRNA (313 and 323 bp). Lane M, markers. Results from a representative experiment are shown (n = 5). D, total RNA was extracted from paired malignant glands of adenocarcinoma (T1–T5) and adjunct histologically benign prostatic glands (N1–N5). Expression of CXCL14 mRNA was assessed by RT-PCR. E, normal prostate epithelial cells served as a control. All evaluated specimens showed significantly reduced level of CXCL14 expression in prostate cancer tissues when compared with the normal adjunct areas.

Close modal

The analysis of CXCL14 mRNA expression in prostate cancer cell lines confirmed the downregulation of CXCL14 expression in PC3 cells, whereas normal prostate epithelial and DU145 cells expressed CXCL14 mRNA (Fig. 1C). The ratios of CXCL14 to β-actin were 1.02 ± 0.21, 0.38 ± 0.05, and 0.04 ± 0.02 for normal prostate epithelial, DU145, and PC3 cells, respectively (Fig. 1C). Prostate cancer cells obtained from primary human tumor specimens (n = 7) by needle microdissection technique showed low or no CXCL14 mRNA expression, whereas adjunct normal prostate cells expressed markedly higher levels of CXCL14 mRNA (Fig. 1D).

Thus, CXCL14 expression is low or lost in primary prostate cancer specimens and in some prostate cancer cell lines, including PC3 cells.

The CXCL14 gene promoter methylation in prostate cancer cells

Next, we focused on the mechanisms of this phenomenon and tested whether the expression of CXCL14 in prostate cancer cells is regulated by the epigenetic mechanisms, i.e., hypermethylation. The DNA methylation patterns in the CpG islands of the CXCL14 gene were determined by MSP before and after treatment with the demethylating agent 5-aza-dC. First, genomic DNA of PC3 cells nontreated and treated with 25 μmol/L 5-aza-dC were subjected to bisulfite modification. Then, using modified DNA, PCR was carried out to detect unmethylated and methylated DNA by specific primer sets designed for unmethylated and methylated DNA, respectively. PCR primers were designed to amplify CXCL14 promoter region with CpG islands. Consequently, CXCL14 promoter methylation was found in PC3 cells: nontreated PC3 cells displayed the methylated alleles, whereas treatment of tumor cells with 25 μmol/L 5-aza-dC induced the appearance of the unmethylated alleles of the CXCL14 gene (Fig. 2A). Universal unmethylated DNA was fully unmethylated, displaying a single band in the MSP corresponding to unmethylated allele, whereas universal methylated DNA was fully methylated, respectively (Fig. 2A). In addition, the methylation status of 22 CpG sites in the CpG island covering the promoter region of the CXCL14 gene was determined by bisulfite sequencing PCR in PC3 cells. The BSP verified the reliability of the MSP results. Representative sequencing, presented in Fig. 2B, shows that most CpG sites are highly (85–100%) methylated and only 3 (+255, +273, and +277) of 22 CpG sites were methylated at 50% to 75%. Thus, our data directly show the state of hypermethylation of CpG islands in the CXCL14 gene promoter region in PC3 cells and its reversibility by demethylating agents.

Figure 2.

Evaluation of the methylation status of CpG sites in CpG islands of the promoter region of the CXCL14 gene in PC3 cell line by (A) MSP and (B) BSP. A, genomic DNA isolated from PC3 cells was subjected to bisulfite modification, followed by MSP with methylation-specific primers (M) and nonmethylation specific primers (U). Lanes: PC3, modified DNA isolated from cultured PC3 cell line; PC3 + AZA, modified DNA isolated from PC3 cells treated with 25 μmol/L of 5-aza-dC. AZA, 5-aza-dC. CpGenome universal unmethylated (control DNA/U) and methylated (control DNA/M) DNA sets were used as controls for MSP. B, BSP analysis of the CXCL14 gene in PC3 cells, universal methylated, and universal unmethylated DNA sets. Representative sequencing results show the methylation status in the promoter region of CXCL14 in DNA from PC3 cells, which do not express CXCL14. Methylated CpG sites at the level of 100% (•), 85% to 90% (○), and 50% to 75% (

graphic
) methylation. Results from one representative experiment are shown (n = 7).

Figure 2.

Evaluation of the methylation status of CpG sites in CpG islands of the promoter region of the CXCL14 gene in PC3 cell line by (A) MSP and (B) BSP. A, genomic DNA isolated from PC3 cells was subjected to bisulfite modification, followed by MSP with methylation-specific primers (M) and nonmethylation specific primers (U). Lanes: PC3, modified DNA isolated from cultured PC3 cell line; PC3 + AZA, modified DNA isolated from PC3 cells treated with 25 μmol/L of 5-aza-dC. AZA, 5-aza-dC. CpGenome universal unmethylated (control DNA/U) and methylated (control DNA/M) DNA sets were used as controls for MSP. B, BSP analysis of the CXCL14 gene in PC3 cells, universal methylated, and universal unmethylated DNA sets. Representative sequencing results show the methylation status in the promoter region of CXCL14 in DNA from PC3 cells, which do not express CXCL14. Methylated CpG sites at the level of 100% (•), 85% to 90% (○), and 50% to 75% (

graphic
) methylation. Results from one representative experiment are shown (n = 7).

Close modal

To confirm that blockage of hypermethylation of CpG islands in the CXCL14 gene promoter region in PC3 cells has a biological significance, we next determined whether 5-aza-dC alters the expression of CXCL14 mRNA and protein in these cells.

Blockage of hypermethylation with 5-aza-dC restores the expression of CXCL14 protein and mRNA in PC3 cells

Next, we tested whether 5-aza-dC can change the expression of CXCL14 at mRNA and protein levels in PC3 cells, and whether it correlates with the status of the CXCL14 gene promoter methylation. Figure 3A and B show that the treatment of PC3 cells with the demethylating agent 5-aza-dC resulted in the restoration of CXCL14 protein (Fig. 3A) and mRNA (Fig. 3B) levels in these cells. For example, the ratio of CXCL14 to β actin was increased from 0.26 ± 0.04 in nontreated tumor cells to 0.95 ± 0.06 in PC3 cells treated with 25 μmol/L 5-aza-dC (Fig. 3B). Thus, treatment of PC3 cells with 5-aza-dC results in the reappearance of CXCL14 expression in tumor cells. This raises the next question about the functional significance and activity of CXCL14 chemokine after the restoration of its expression in tumor cells by hypermethylation blockage.

Figure 3.

Demethylating agent 5-aza-dC restores the expression of (A) CXCL14 protein and (B) CXCL14 mRNA in PC3 cells. A, nontreated PC3 cells and PC3 cells treated with 25 μmol/L 5-aza-dC were stained with anti-CXCL14 antibodies. Positive staining was determined in PC3 cells treated with 5-aza-dC and is shown in red color. B, total RNA was extracted from PC3 cells before and after treatment with 5-aza-dC. RT-PCR was carried out to evaluate the expression of CXCL14 and β-actin mRNA (313 and 323 bp). Lanes: M, markers; 1, normal prostate epithelial cells as a positive control; 2, nontreated PC3 cells; 3 and 4, PC3 cells treated with 10 and 25 μmol/L of 5-aza-dC, respectively. AZA, 5-aza-dC. The results of a representative experiment are shown (n = 3).

Figure 3.

Demethylating agent 5-aza-dC restores the expression of (A) CXCL14 protein and (B) CXCL14 mRNA in PC3 cells. A, nontreated PC3 cells and PC3 cells treated with 25 μmol/L 5-aza-dC were stained with anti-CXCL14 antibodies. Positive staining was determined in PC3 cells treated with 5-aza-dC and is shown in red color. B, total RNA was extracted from PC3 cells before and after treatment with 5-aza-dC. RT-PCR was carried out to evaluate the expression of CXCL14 and β-actin mRNA (313 and 323 bp). Lanes: M, markers; 1, normal prostate epithelial cells as a positive control; 2, nontreated PC3 cells; 3 and 4, PC3 cells treated with 10 and 25 μmol/L of 5-aza-dC, respectively. AZA, 5-aza-dC. The results of a representative experiment are shown (n = 3).

Close modal

5-aza-dC restores the expression of functional CXCL14 in prostate cancer cells

Recently, we and others showed that CXCL14 is a potent DC chemoattractant in vitro and in vivo (7, 8). Here, we proved that the restoration of CXCL14 expression in PC3 cells with the demethylating agent 5-aza-dC has the biological significance. We have determined if control and 5-aza-dC–treated tumor cells attracted DC and if chemoattraction is due to the production of CXCL14 chemokine. First, we showed that the conditioned medium from intact PC3 cell cultures did not attract immature human DC (Fig. 4). However, attraction of DC was significantly increased by the supernatants of PC3 cells pretreated with 25 μmol/L of 5-aza-dC. For example, in the presence of cell-free conditioned media collected from PC3 cells treated with a demethylating agent, migration of DC reached 1,467 ± 12 versus 941 ± 14 cells per minute transmigrated DC in control wells (P < 0.05; Fig. 4). Elimination of the CXCL14 protein from the tumor cell supernatants by affinity chromatography resulted in the loss of chemoattractive properties of DC toward tumor-conditioned media, suggesting that chemoattraction of DC by 5-aza-dC–treated tumor cells was mediated by CXCL14 chemokine (Fig. 4). A comparable chemoattraction of DC (1,405 ± 5 cells/min) was also detected in the presence of known DC chemoattractant MIP-1α (20 ng/mL) used as a positive control.

Figure 4.

Demethylating agent 5-aza-dC restores the expression of functional CXCL14 in PC3 cells. Conditioned media from PC3 cells nontreated, treated with 25 μmol/L of 5-aza-dC (PC3 + Aza), and treated with condition medium from PC3 + Aza group after CXCL14 depletion were tested for their ability to attract DC in migration assay. Removal of CXCL14 from the medium was performed by binding with specific anti-CXCL14 antibodies using affinity chromatography. AZA, 5-aza-dC. As a control for spontaneous migration, the groups used were as follows: (a) Assay medium (RPMI + 1% FBS), (b) Assay medium + 5-aza-dC, and (c) Assay medium containing 5-aza-dC after passing through the protein A agarose bound with anti-CXCL14 antibody. These three control groups showed the same levels of chemoattractive activity for DC and were combined together. The results of a representative experiment are shown; columns, mean; bars, SEM. Three independent experiments have shown similar results. *, P < 0.05 versus control (t test); #, P < 0.05 versus PC3 + 5-aza-dC (t test).

Figure 4.

Demethylating agent 5-aza-dC restores the expression of functional CXCL14 in PC3 cells. Conditioned media from PC3 cells nontreated, treated with 25 μmol/L of 5-aza-dC (PC3 + Aza), and treated with condition medium from PC3 + Aza group after CXCL14 depletion were tested for their ability to attract DC in migration assay. Removal of CXCL14 from the medium was performed by binding with specific anti-CXCL14 antibodies using affinity chromatography. AZA, 5-aza-dC. As a control for spontaneous migration, the groups used were as follows: (a) Assay medium (RPMI + 1% FBS), (b) Assay medium + 5-aza-dC, and (c) Assay medium containing 5-aza-dC after passing through the protein A agarose bound with anti-CXCL14 antibody. These three control groups showed the same levels of chemoattractive activity for DC and were combined together. The results of a representative experiment are shown; columns, mean; bars, SEM. Three independent experiments have shown similar results. *, P < 0.05 versus control (t test); #, P < 0.05 versus PC3 + 5-aza-dC (t test).

Close modal

Thus, our data revealed that treatment of PC3 cells with 5-aza-dC resulted in the reappearance of functional CXCL14 in tumor cells. Altogether, these results show for the first time that epigenetic mechanisms are responsible for lowering the expression of CXCL14 chemokine in prostate cancer, which could be normalized pharmacologically using hypermethylation inhibitors.

CXCL14 is a new chemokine with an unknown receptor that attracts immature DC both in vitro and in vivo. In addition to its chemoattractive properties, CXCL14 is able to activate DC maturation and function (7, 8). DC are known to play a key role in initiating antitumor immune responses (21) and infiltration of tumors by DC is of great importance for antitumor immunity (15, 2224).

The presence of DC, macrophages, and lymphocytes in solid tumors is regulated through the local production of chemokines by tumor and stromal cells (1, 8, 25, 26). DCs do not migrate to the tumors that do not express CXCL14 in vitro and in vivo (7, 8, 27), suggesting that CXCL14 may have a special role in host-tumor interaction and tumor recognition by the immune cells. It is conceivable that DC might be constitutively recruited to the CXCL14-expressing tissues, allowing antigen-presenting cells to leave circulation and enter these tissues in the absence of inflammation. On the other hand, the loss of CXCL14 expression in malignant tissues explains the decreased DC infiltration of tumors, thus allowing the tumor to escape immune control mechanisms.

We showed here that CXCL14 mRNA and protein expression were downregulated in prostate adenocarcinoma tissues and cells. The loss of CXCL14 mRNA expression in various tumor tissues and tumor cell lines has been reported (4, 6, 8, 9). It has been also shown that the transduction of tumor cell lines with the CXCL14 gene resulted in inhibition of tumor growth in vivo (28). Our recent data suggest that the expression of CXCL14 in tumor tissue results in the inhibition of tumor growth by attracting and activating DC, and inducing the antitumor immune responses.

The mechanisms of regulation of CXCL14 expression in cancer are not known. Recently, Peterson and colleagues (29) investigated the posttranslational pathway that may be involved in the loss of CXCL14 in prostate cancer. The CXCL14 construct was cloned into adenoviral vector and transduced into LNCaP prostate cancer cells. Using this approach to create cells overexpressing CXCL14, the authors reported that CXCL14 protein was not detectable in tumor cells due to its degradation by ubiquitin-mediated proteolysis. We have recently showed that retroviral transduction of CXCL14-negative human head and neck squamous cell carcinoma cell lines with the CXCL14 gene resulted in the reappearance of the chemokine in tumor cells and increased DC attraction in vitro as well as significant tumor infiltration by DC in vivo in chimeric animal models (8). It is possible that the way of gene delivery determines the expression of transgene and is important for the generation of stable high-titer producer cell lines and/or protein stability.

The main goal of this work was to evaluate the primary mechanisms of the downregulation of CXCL14 expression in nonmodified prostate cancer cells on the transcriptional level. Among the mechanisms of gene silencing, hypermethylation of CpG islands located in the promoter regions of tumor suppressor genes has been recognized as being a molecular hallmark of human cancers (16, 3037). Regulatory genes that are commonly hypermethylated in malignant cells include the RB1 gene in retinoblastoma (38); RARβ2, E-caderin, and RTVP-1 in gastric (39) and prostate cancer (40, 41); and the p15 gene in leukemias and colorectal cancer (38, 42). However, it was unknown whether expression of chemokines in tumor cells may be regulated by the same mechanism. Here, we examined whether the loss of CXCL14 expression in the PC3 cell line was due to an epigenetic alteration. We showed that the demethylating reagent 5-aza-dC is effective in inducing CXCL14 mRNA and protein expression in PC3 cells in a dose-dependent manner. Furthermore, treatment of PC3 cells with 5-aza-dC resulted in the restoration of expression of functional CXCL14 in tumor cells because DC displayed chemoattractive activity toward conditioned media from PC3 cells treated with 5-aza-2-deoxycytidine, but not from nontreated tumor cells. To verify that chemoattraction is due to CXCL14, we showed that the elimination of CXCL14 from tumor cell supernatant by affinity chromatography resulted in the loss of DC chemoattraction. These data directly suggest a role of epigenetic regulation of CXCL14 expression in PC3 cells.

To confirm this finding, we focused on the molecular mechanisms of downregulation of CXCL14 expression in PC3 cells. We speculated that CpG islands in the promoter region of the CXCL14 gene in tumor cells may be hypermethylated and we investigated the effect of 5-aza-dC on the demethylation of the CXCL14 promoter region by MSP. Our data revealed that the expression of CXCL14 in PC3 cells is affected by methylation-induced silencing. The potent inhibitor of DNA methylation, 5-aza-dC, restored the CXCL14 gene expression in tumor cells. Furthermore, the BSP verified the reliability of the MSP results. There was a good concordance between MSP status and bisulfate sequences, limiting the potential detection of false positives by the MSP.

Thus, our results implicate the epigenetic mechanisms extinguishing the CXCL14 gene expression in PC3 cells. Importantly, promoter hypermethylation is a reversible mechanism of gene silencing and therefore restoration of expression of lost chemokines in tumor cells may serve as a powerful alternative therapeutic approach to treating different cancers.

No potential conflicts of interest were disclosed.

Grant Support: DOD PC050252 (M.R. Shurin).

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

1
McColl
SR
. 
Chemokines and dendritic cells: a crucial alliance
.
Immunol Cell Biol
2002
;
80
:
489
96
.
2
Moser
B
,
Loetscher
P
. 
Lymphocyte traffic control by chemokines
.
Nat Immunol
2001
;
2
:
123
8
.
3
Payne
AS
,
Cornelius
LA
. 
The role of chemokines in melanoma tumor growth and metastasis
.
J Invest Dermatol
2002
;
118
:
915
22
.
4
Hromas
R
,
Broxmeyer
HE
,
Kim
C
, et al
. 
Cloning of BRAK, a novel divergent CXC chemokine preferentially expressed in normal versus malignant cells
.
Biochem Biophys Res Commun
1999
;
255
:
703
6
.
5
Meuter
S
,
Moser
B
. 
Constitutive expression of CXCL14 in healthy human and murine epithelial tissues
.
Cytokine
2008
;
44
:
248
55
.
6
Frederick
MJ
,
Henderson
Y
,
Xu
X
, et al
. 
In vivo expression of the novel CXC chemokine BRAK in normal and cancerous human tissue
.
Am J Pathol
2000
;
156
:
1937
50
.
7
Shellenberger
TD
,
Wang
M
,
Gujrati
M
, et al
. 
BRAK/CXCL14 is a potent inhibitor of angiogenesis and a chemotactic factor for immature dendritic cells
.
Cancer Res
2004
;
64
:
8262
70
.
8
Shurin
GV
,
Ferris
R
,
Tourkova
IL
, et al
. 
Loss of new chemokine CXCL14 in tumor tissue is associated with low infiltration by dendritic cells (DC), while restoration of human CXCL14 expression in tumor cells causes attraction of DC both in vitro and in vivo
.
J Immunol
2005
;
174
:
5490
8
.
9
Schwarze
SR
,
DePrimo
SE
,
Grabert
LM
,
Fu
VX
,
Brooks
JD
,
Jarrard
DF
. 
Novel pathways associated with bypassing cellular senescence in human prostate epithelial cells
.
J Biol Chem
2002
;
277
:
14877
83
.
10
Starnes
T
,
Rasila
KK
,
Robertson
MJ
, et al
. 
The chemokine CXCL14 (BRAK) stimulates activated NK cell migration: implications for the downregulation of CXCL14 in malignancy
.
Exp Hematol
2006
;
34
:
1101
5
.
11
Schaerli
P
,
Willimann
K
,
Ebert
LM
,
Walz
A
,
Moser
B
. 
Cutaneous CXCL14 targets blood precursors to epidermal niches for Langerhans cell differentiation
.
Immunity
2005
;
23
:
331
42
.
12
Kurth
I
,
Willimann
K
,
Schaerli
P
,
Hunziker
T
,
Clark-Lewis
I
,
Moser
B
. 
Monocyte selectivity and tissue localization suggests a role for breast and kidney-expressed chemokine (BRAK) in macrophage development
.
J Exp Med
2001
;
194
:
855
61
.
13
Gutkin
DW
. 
Tumor Infiltrating Dendritic Cells: the Pathologist's Perspective
. In:
Salter
MSaR
, editor.
Dendritic Cells in Cancer
.
New York
:
Springer
; 
2009
.
14
Iwamoto
M
,
Shinohara
H
,
Miyamoto
A
, et al
. 
Prognostic value of tumor-infiltrating dendritic cells expressing CD83 in human breast carcinomas
.
Int J Cancer
2003
;
104
:
92
7
.
15
Shurin
GV
,
Yurkovetsky
ZR
,
Shurin
MR
. 
Tumor-induced dendritic cell dysfunction
. In:
Ochoa
A
, editor.
Mechanisms of Tumor Escape from the Immune Response
.
London and New York
:
Taylor and Francis
; 
2003
, p.
112
38
.
16
Baylin
SB
. 
Mechanisms underlying epigenetically mediated gene silencing in cancer
.
Semin Cancer Biol
2002
;
12
:
331
7
.
17
Baylin
SB
,
Herman
JG
. 
DNA hypermethylation in tumorigenesis: epigenetics joins genetics
.
Trends Genet
2000
;
16
:
168
74
.
18
Grady
WM
,
Markowitz
SD
. 
Genetic and epigenetic alterations in colon cancer
.
Annu Rev Genomics Hum Genet
2002
;
3
:
101
28
.
19
Serrano
A
,
Tanzarella
S
,
Lionello
I
, et al
. 
Rexpression of HLA class I antigens and restoration of antigen-specific CTL response in melanoma cells following 5-aza-2′-deoxycytidine treatment
.
Int J Cancer
2001
;
94
:
243
51
.
20
Shurin
MR
. 
Preparation of human dendritic cells for tumor vaccination
.
Methods Mol Biol
2003
;
215
:
437
62
.
21
Banchereau
J
,
Briere
F
,
Caux
C
, et al
. 
Immunobiology of dendritic cells
.
Annu Rev Immunol
2000
;
18
:
767
811
.
22
Troy
AJ
,
Summers
KL
,
Davidson
PJ
,
Atkinson
CH
,
Hart
DN
. 
Minimal recruitment and activation of dendritic cells within renal cell carcinoma
.
Clin Cancer Res
1998
;
4
:
585
93
.
23
Triozzi
PL
,
Khurram
R
,
Aldrich
WA
,
Walker
MJ
,
Kim
JA
,
Jaynes
S
. 
Intratumoral injection of dendritic cells derived in vitro in patients with metastatic cancer
.
Cancer
2000
;
89
:
2646
54
.
24
Ishigami
S
,
Natsugoe
S
,
Matsumoto
M
, et al
. 
Clinical implications of intratumoral dendritic cell infiltration in esophageal squamous cell carcinoma
.
Oncol Rep
2003
;
10
:
1237
40
.
25
Giovarelli
M
,
Cappello
P
,
Forni
G
, et al
. 
Tumor rejection and immune memory elicited by locally released LEC chemokine are associated with an impressive recruitment of APCs, lymphocytes, and granulocytes
.
J Immunol
2000
;
164
:
3200
6
.
26
Fushimi
T
,
Kojima
A
,
Moore
MA
,
Crystal
RG
. 
Macrophage inflammatory protein 3α transgene attracts dendritic cells to established murine tumors and suppresses tumor growth
.
J Clin Invest
2000
;
105
:
1383
93
.
27
Cao
X
,
Zhang
W
,
Wan
T
, et al
. 
Molecular cloning and characterization of a novel CXC chemokine macrophage inflammatory protein-2 γ chemoattractant for human neutrophils and dendritic cells
.
J Immunol
2000
;
165
:
2588
95
.
28
Schwarze
SR
,
Luo
J
,
Isaacs
WB
,
Jarrard
DF
. 
Modulation of CXCL14 (BRAK) expression in prostate cancer
.
Prostate
2005
;
64
:
67
74
.
29
Peterson
FC
,
Thorpe
JA
,
Harder
AG
,
Volkman
BF
,
Schwarze
SR
. 
Structural determinants involved in the regulation of CXCL14/BRAK expression by the 26 S proteasome
.
J Mol Biol
2006
;
363
:
813
22
.
30
Lee
MN
,
Tseng
RC
,
Hsu
HS
, et al
. 
Epigenetic inactivation of the chromosomal stability control genes BRCA1, BRCA2, and XRCC5 in non-small cell lung cancer
.
Clin Cancer Res
2007
;
13
:
832
8
.
31
Kuzmin
I
,
Gillespie
JW
,
Protopopov
A
, et al
. 
The RASSF1A tumor suppressor gene is inactivated in prostate tumors and suppresses growth of prostate carcinoma cells
.
Cancer Res
2002
;
62
:
3498
502
.
32
Ren
C
,
Li
L
,
Yang
G
, et al
. 
RTVP-1, a tumor suppressor inactivated by methylation in prostate cancer
.
Cancer Res
2004
;
64
:
969
76
.
33
Lin
J
,
Lai
M
,
Huang
Q
,
Ma
Y
,
Cui
J
,
Ruan
W
. 
Methylation patterns of IGFBP7 in colon cancer cell lines are associated with levels of gene expression
.
J Pathol
2007
;
212
:
83
90
.
34
Park
JY
,
Zheng
W
,
Kim
D
, et al
. 
Candidate tumor suppressor gene SLC5A8 is frequently down-regulated by promoter hypermethylation in prostate tumor
.
Cancer Detect Prev
2007
;
31
:
359
65
.
35
Maruyama
R
,
Toyooka
S
,
Toyooka
KO
, et al
. 
Aberrant promoter methylation profile of bladder cancer and its relationship to clinicopathological features
.
Cancer Res
2001
;
61
:
8659
63
.
36
Zochbauer-Muller
S
,
Fong
KM
,
Virmani
AK
,
Geradts
J
,
Gazdar
AF
,
Minna
JD
. 
Aberrant promoter methylation of multiple genes in non-small cell lung cancers
.
Cancer Res
2001
;
61
:
249
55
.
37
Toyooka
S
,
Toyooka
KO
,
Maruyama
R
, et al
. 
DNA methylation profiles of lung tumors
.
Mol Cancer Ther
2001
;
1
:
61
7
.
38
Cheng
JC
,
Matsen
CB
,
Gonzales
FA
, et al
. 
Inhibition of DNA methylation and reactivation of silenced genes by zebularine
.
J Natl Cancer Inst
2003
;
95
:
399
409
.
39
Barber
M
,
Murrell
A
,
Ito
Y
, et al
. 
Mechanisms and sequelae of E-cadherin silencing in hereditary diffuse gastric cancer
.
J Pathol
2008
;
216
:
295
306
.
40
Graff
JR
,
Herman
JG
,
Lapidus
RG
, et al
. 
E-cadherin expression is silenced by DNA hypermethylation in human breast and prostate carcinomas
.
Cancer Res
1995
;
55
:
5195
9
.
41
Lin
X
,
Asgari
K
,
Putzi
MJ
, et al
. 
Reversal of GSTP1 CpG island hypermethylation and reactivation of pi-class glutathione S-transferase (GSTP1) expression in human prostate cancer cells by treatment with procainamide
.
Cancer Res
2001
;
61
:
8611
6
.
42
Toyota
M
,
Ohe-Toyota
M
,
Ahuja
N
,
Issa
JP
. 
Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype
.
Proc Natl Acad Sci U S A
2000
;
97
:
710
5
.