...the metastatic potential of mammary carcinoma cells |...

来自 : 发布时间：2024-05-20

Epigenetic silencing of CXCL12 increases the metastatic potential of mammary carcinoma cells AbstractExpression of the chemokine receptor CXCR4 has been linked with increased metastasis and decreased clinical prognosis in breast cancer. The current paradigm dictates that CXCR4 fosters carcinoma cell metastasis along a chemotactic gradient to organs expressing the ligand CXCL12. The present study asked if alterations in autocrine CXCR4 signaling via dysregulation of CXCL12 in mammary carcinoma cells modulated their metastatic potential. While CXCR4 was consistently detected, expression of CXCL12 characteristic of human mammary epithelium was silenced by promoter hypermethylation in breast cancer cell lines and primary mammary tumors. Stable re-expression of functional CXCL12 in ligand null cells increased orthotopic primary tumor growth in the mammary fat-pad model of tumorigenesis. Those data parallel increased carcinoma cell proliferation measured in vitro with little-to-no-impact on apoptosis. Moreover, re-expression of autocrine CXCL12 markedly reduced metastatic lung invasion assessed using in vivo bioluminescence imaging following tail vein injection. Consistent with those data, decreased metastasis reflected diminished intracellular calcium signaling and chemotactic migration in response to exogenous CXCL12 independent of changes in CXCR4 expression. Together these data suggest that an elevated migratory signaling response to ectopic CXCL12 contributes to the metastatic potential of CXCR4-expressing mammary carcinoma cells, subsequent to epigenetic silencing of autocrine CXCL12. IntroductionChemokines are immune regulatory molecules known to direct the trafficking of immune cells via binding to their cognate receptor on the surface of target cells (Rot and von Andrian, 2004). Chemokines expressed at sites of infection have been classified as proinflammatory mediators and regulate the directed migration of immune cells to those areas. Several additional constitutively expressed chemokines play key homeostatic roles in immunity including hematopoiesis, progenitor cell recruitment and leukocyte recirculation (Gazitt, 2004). The constitutive chemokine ligand CXCL12 and its cognate receptor CXCR4 are evolutionarily conserved and essential for life in mice (Nagasawa et al., 1996a, 1996b). Homeostatic roles for CXCL12 signaling through CXCR4 in various physiological compartments have been described, including B-cell development, gastrointestinal vascularization, heart and cerebral development, as well as mucosal epithelial wound healing (Nagasawa et al., 1996a; Tachibana et al., 1998; Zou et al., 1998; Heidemann et al., 2003; Smith et al., 2005; Moyer et al., 2007). Thus, CXCL12 and CXCR4 are widely expressed throughout the body and serve diverse physiological roles depending on the tissue or cell type.Metastatic cells of solid tumors hijack the chemokine signaling network leading to colonization of ectopic tissues with those carcinomas, with CXCR4 and CXCL12 receiving the most attention in regards to metastasis of colon, breast, kidney and lung carcinomas (Muller et al., 2001; Schrader et al., 2002; Phillips et al., 2003; Zeelenberg et al., 2003). This research has lead to the current model wherein CXCR4 expression on metastatic cells enables those cells to home to organs abundantly expressing CXCL12. Accordingly, mammary carcinoma cells metastasize primarily to those tissues expressing highest tissue levels of CXCL12 including bone, liver and lung (Muller et al., 2001). Studies suggest that CXCR4 expression on tumor cells correlates with increased metastasis and decreased clinical prognosis (Salvucci et al., 2006). Moreover, blocking CXCR4 inhibits breast cancer cells from metastasizing to organs expressing CXCL12 (Muller et al., 2001; Smith et al., 2004). Although the mechanisms modulating CXCR4 expression in carcinoma cells is incompletely understood, increased expression on metastatic cancer cells seems to play a critical role in cancer tumorigenesis. In contrast to CXCR4, the role for autocrine expression of its ligand CXCL12 in carcinogenesis remains to be established (Wendt et al., 2006).We and others have previously observed that several normal epithelial tissues express not only CXCR4 but also the ligand CXCL12 (Shirozu et al., 1995; Wendt et al., 2006). Genes such as CXCL12 with a high occurrence of CpG dinucleotides in the promoter region are subject to transcriptional silencing via cytosine methylation (Herman and Baylin, 2003). The mechanism of DNA methylation and gene silencing in cancer has become more clearly delineated and may occur early in the progression to metaplasia, preceding even frank pathology (Baylin and Chen, 2005). As detailed herein, absence of native CXCL12 expression in mammary carcinoma cell lines and primary tumors was dependent upon promoter hypermethylation consistent with our prior report in primary colonic carcinoma and cultured cell lines (Wendt et al., 2006). Moreover, reestablishment of CXCL12 production in CXCR4-expressing mammary carcinoma cells congruently increased proliferation and primary tumor growth and decreased chemotaxis and metastasis, suggesting that those primary tumor cells that silence CXCL12 are at a selective advantage for metastasis. These data present a mechanism of CXCR4 signaling dysregulation in breast cancer that significantly expands upon the model that elevated receptor expression drives tumor metastasis.ResultsLoss of CXCL12 expression in mammary carcinomaIn agreement with the epithelial cells lining the human gastrointestinal lumen (Agace et al., 2000; Smith et al., 2005; Wendt et al., 2006) CXCL12 was strongly and consistently expressed by epithelial cells lining mammary ducts (Figure 1a). In contrast, immunohistochemical analysis of primary mammary carcinomas revealed CXCL12 was minimally and variably expressed by tumor cells (Figure 1a). Reverse transcription (RT)鈥揚CR and immunofluorescence microscopy was used to determine if dysregulated CXCL12 expression was paralleled in a panel of mammary carcinoma cell lines encompassing highly aggressive cells such as the MDA-MB-231 and MDA-MB-435s, less aggressive lines such as the MCF-7 and MDA-MB-134 and, the presumably nontransformed, immortalized lines MCF-10A and MCF-12A. Consistent with our data from primary tumors, CXCL12 transcript expression was undetectable in six of eight in vitro breast cancer cell lines (Figure 1b). Notably, CXCL12 mRNA expression was restricted to the less aggressive carcinoma lines MCF-7 and MDA-MB-134. However, CXCL12 protein was not uniformly immunostained in all MCF7 cells (Figure 1c) and was insufficient to produce detectable levels of secreted protein assayed using enzyme-linked immunosorbent assay (not shown). Thus, CXCL12 protein expression in breast cancer cell lines differs from the consistent expression we noted in normal in vivo mammary epithelium and parallels more closely the variable expression noted in primary tumors (Figure 1a).Figure 1Constitutive CXCL12 expression in normal epithelium is absent in mammary carcinoma. (a) CXCL12 is constitutively expressed in normal in vivo mammary epithelial cells (upper panels). In contrast, CXCL12 expression by primary mammary tumor cells (lower panels) was variable and reduced. (b) Cultured mammary epithelial cell lines lack expression of CXCL12 assayed by RT鈥揚CR. CXCR4 mRNA expression was consistently observed. (c) CXCL12 protein expression was variable in MCF-7 cells as shown by immunofluorescence microscopy. (d) CXCR4 protein was consistently expressed in mammary carcinoma cell lines. Data are representative of three independent analyses.Full size imageIn contrast, mRNA (Figure 1b) and protein (Figure 1d) expression of CXCR4 was readily detectable in a majority of the cell lines examined. The lack of both CXCR4 and CXCL12 expression by the MDA-MB-435s cell line may reflect their clarification not as mammary carcinoma cells but as melanoma cells (Ellison et al., 2002). Thus, the expression profile acquired during mammary epithelial transformation was characteristically similar to highly migratory leukocytes in that CXCR4 expression predominates over little if any CXCL12 expression (Kimura et al., 2003).CXCL12 is silenced by promoter hypermethylation in mammary carcinomaTwo primary DNA methyltransferase enzymes responsible for gene silencing in cancer are DNMT1 and DNMT3b (Rhee et al., 2002), both of which we have previously shown to be involved in CXCL12 silencing (Wendt et al., 2006). Immunoblot and RT鈥揚CR analyses in a panel of mammary carcinoma cell lines indicated that each of the cell lines that lacked expression of CXCL12 readily expressed either or both enzymes (Figure 2a). In contrast, very little DNMT1 protein was detected in the MCF7 cell line in which CXCL12 transcript expression was maintained (Figure 2a). To determine if CXCL12 expression was aberrantly silenced by DNA methylation in these mammary carcinoma cell lines, we sought to restore CXCL12 expression by inhibition of DNMT function. Treatment of MDA-MB-231 cells with 5-aza-2鈥?deoxycytidine (5-aza) resulted in a dose- and time-dependent restoration of CXCL12 expression, with little-to-no-impact on CXCR4 expression (Figure 2b).Figure 2CXCL12 is silenced by DNA hypermethylation in mammary carcinoma cell lines. (a) DNA methyltransferase enzymes (DNMT1 and 3b) are readily expressed by mammary carcinoma cell lines as shown using immunoblot analysis (upper panels) and RT鈥揚CR (lower panels). (b) Treatment of MDA-MB-231 cells with 5-aza-2鈥?deoxycytidine (5-aza) reestablished CXCL12 mRNA expression after 3, 4, 5 and 6 days (D3鈥?) (upper panels) and using a range of 5-aza concentrations (lower panels) relative to the nonstimulated (NS) control cells. Data are representative of three independent analyses.Full size imageMethylation-specific PCR (MSP) was used to define hypermethylation of the CXCL12 proximal promoter and revealed promoter methylation in the MDA-MB-231, MCF10A, MCF12A, MDA-MB-453, MDA-MB-435s and BT-549 cell lines that correspondingly lacked expression of the gene (Figure 3a). In contrast, the MCF7 and MDA-MB-134 cell lines expressing CXCL12 were largely unmethylated. Further, the MCF10A and MCF12A cell lines derived from nontransformed mammary epithelial tissue and which lacked detectable CXCL12 transcripts possessed methylated CXCL12 and expressed DNMT1 and DNMT3b. To verify that methylation was a characteristic of transformation and not just an artifact of cell culture, we next used our diagnostic MSP primers to assess CXCL12 methylation in primary mammary carcinoma. As shown in Figure 3b, representative archived mammary biopsy tissues analysed using MSP demonstrated that CXCL12 promoter methylation was detected in 5 of 15 individual primary tumors. Further, methylation was detected in separate patient tumors, including a matched metastatic and primary tumor (Figure 3b, P2). A separate primary and matched normal mammary tissue was unmethylated (Figure 3b, P1). Cumulatively, MSP analysis indicated that 40% of 15 separate primary tumors examined exhibited aberrant methylation of CXCL12.Figure 3The CXCL12 proximal promoter is hypermethylated in primary and metastatic mammary carcinoma. (a) DNA hypermethylation of the CXCL12 promoter was detected using methylation-specific PCR (MSP) in all mammary carcinoma cell lines examined. MSP-specific primers located within the CXCL12 proximal promoter specifically amplified methylated (m) or unmethylated (u) alleles. (b) CXCL12 promoter hypermethylation was detectable in 40% of primary tumors examined (n=15 primary mammary tumor samples). Shown are four samples positive for methylation, lanes 1, 3, 4 and 5; and two negative for methylation, lanes 2 and 6. CXCL12 promoter methylation was detected in primary tumor (T) and metastatic (M) tissue, but not normal (N) mammary tissue from two separate patient samples (P1 and P2). (c) CpG dinucleotides were specifically methylated in MDA-MB-231 and MCF10A cell lines lacking expression of the gene. Open and filled lollipops represent unmethylated and methylated CpG dinucleotides, respectively, and no lollipop represents indeterminate methylation status. Data in panels (a) and (c) are representative of 2鈥? separate analyses.Full size imageTo more specifically examine the methylation status of the CXCL12 promoter in mammary carcinoma, we conducted bisulfite sequencing PCR (BSSP). These data confirmed that MDA-MB-231 and MCF10A cell lines that do not express the CXCL12 transcript or protein are heavily methylated while the MCF7 cell line which retained CXCL12 expression has few, if any, methylated CpG dinucleotides in the proximal promoter (Figure 3c). Notably, several Sp1 sites previously identified to be critical for CXCL12 promoter activity were heavily methylated in cells lacking CXCL12 expression (Garcia-Moruja et al., 2005). Importantly, methylation of Sp1 binding sites has previously been shown to block binding of this transcription factor and inhibit transcriptional activity (Clark et al., 1997). Together these data strongly suggest that CXCL12 was transcriptionally silenced by DNA methylation in mammary carcinoma.Re-expression of functional CXCL12 in MDA-MB-231 cellsTo define the functional significance of CXCL12 silencing in mammary carcinoma, we reestablished CXCL12 expression in MDA-MB-231 cells stably expressing firefly luciferase (231-luc). As shown in Figure 4a, 231-luc parental cells as well as double-transfected cell lines stably expressing either CXCL12 (231-luc-CXCL12) or eGFP (231-luc-eGFP) expressed comparable levels of luciferase. Only those cells stably re-expressing CXCL12 produced detectable amounts of protein (Figure 4b). Finally, CXCR4-dependent chemotaxis of receptor-expressing U937 monocytes was induced solely by conditioned medium from 231-luc-CXCL12 cell lines (Figure 4c).Figure 4Re-expression of functional CXCL12 in MDA-MB-231 cells. (a) MDA-MB-231 cells were stably transfected with luciferase (231-luc). These cells were further stably transfected with CXCL12 (231-luc-CXCL12) or, as a control, eGFP (231-luc-eGFP). The parental and each of the double-stable cell lines expressed comparable levels of luciferase. (b) ELISA and immunoblot (Inset) analyses demonstrated the production of either CXCL12 or eGFP in each of the respective cell lines. (c) CXCL12 secreted into serum-free conditioned medium by 231-luc-CXCL12 cells (CXCL12)-stimulated increased chemotaxis of U937 monocytes when compared to 231-luc-eGFP-conditioned media (eGFP) or nonconditioned media alone (NS). Chemotaxis induced by CXCL12-conditioned medium was blocked by pretreatment of U937 cells with the CXCR4 inhibitor AMD3100. Asterisk (*) indicates statistical significant difference in chemotaxis induced by either eGFP or nonconditioned media compared to CXCL12-conditioned medium. Asterisks (**) indicate statistically significant difference in chemotaxis of U937 cells treated with AMD3100 compared to the untreated control (P猢?/span>0.05).Full size imageCXCL12 re-expression increased orthotopic primary mammary tumor growthWe first sought to use those constructed cell lines to define the impact of autocrine CXCL12 expression on in vivo tumor growth upon orthotopic xenograft into the mammary fat-pad of severe combined immunodeficient (SCID) mice. Injection of similar numbers of 231-luc-eGFP (L-eGFP) or 231-luc-CXCL12 (L-CXCL12) cells into the mammary fat-pad, as verified upon imaging 鈭?/span>30鈥塵in postinjection (T0), resulted in the formation of primary tumors after 5 weeks (Figure 5a). Tumors resulting from the CXCL12 re-expressing cells were markedly larger than those established from the control eGFP cells (Figure 5a). Longitudinal in vivo tracking of tumor growth showed immediate and sustained growth of CXCL12-expressing cells where eGFP control cells lagged for 1 week before tumor growth was noted (Figure 5b). Moreover, the greater tumor size formed by CXCL12-expressing cells and observed using biophotonic imaging was verified by increased wet weight of those 5-week-old tumors (Figure 5c). Metastasis of those tumors was not observed in the initial 5 weeks of growth or in 10 weeks following primary tumor excision (data not shown). Thus, these data demonstrate that constitutive CXCL12 signaling leads to increased primary tumor growth in MDA-MB-231 cells.Figure 5CXCL12 re-expression in mammary carcinoma cells increases primary tumor growth. (a) Representative bioluminescent images showing equal engraftment at time zero (T0) resulting in increased orthotopic tumor formation in CXCL12-expressing cells after 5 weeks (5wk). (b) Averaged area flux readings normalized to T0 postinjection (eGFP n=8; CXCL12 n=7) showed that CXCL12-expressing tumors grew more rapidly than control eGFP tumors. Asterisks (* and **) indicate statistically significant difference between mean luminescence readings (P猢?/span>0.05). (c) Wet weight of primary mammary fat-pad tumors formed by L-CXCL12 cells was significantly greater than L-eGFP control cells 5 weeks postinjection (P猢?/span>0.05).Full size imageCXCL12 increases mammary carcinoma cell proliferationGiven the increased primary tumor growth observed, we next sought to define changes in mammary carcinoma cell proliferation upon autocrine CXCL12 expression (Orimo et al., 2005; Ueda et al., 2006). Propidium iodide staining demonstrated that re-expression of CXCL12 in two separate 231-luc-CXCL12 transfectant clones (L-CXCL12#1 and L-CXCL12#2) lead to an increase in the percentage of cells in S and G2/M phase of the cell cycle on day 3 and day 5 relative to control 231-luc cells (Figures 6a and b). In agreement with the changes in cell cycle, thymidine incorporation was similarly increased in those CXCL12 re-expressing cells compared to eGFP control cells (Figure 6c). Increased cellular proliferation induced by CXCL12 re-expression in mammary carcinoma cells is in stark contrast to our previous observations in intestinal epithelial cell lines (Smith et al., 2005). Moreover, the apoptotic response we had previously described in HT29 colorectal cancer cells following re-expression of CXCL12 (Wendt et al., 2006) was not observed in mammary carcinoma cells following reestablishment of ligand expression (Figure 6d). These data highlight differences in the pathophysiological impact of CXCL12 silencing between mammary and colorectal carcinoma.Figure 6CXCL12 re-expression in MDA-MB-231 cells increases cellular proliferation. (a) Percentages of cells in S and G2/M are similar at day 1 (D1) but increased at day 3 and 5(D3 and D5) in the CXCL12-expressing lines compared to control. (b) The percentage of cells in S and G2/M phase (% dividing cells) is increased in two CXCL12-expressing lines (#1 and #2) as compared to control 231-luc cells at 5 days. (c) Increased cumulative DNA synthesis assessed by 3H-thymidine incorporation by cells re-expressing CXCL12. Values are mean卤s.d. from a representative of three independent experiments completed in triplicate. (d) Apoptosis was unchanged in MDA-MB231 cells following re-expression of CXCL12. HT29 colonic carcinoma cells expressing CXCL12 or eGFP served as a positive control for caspase activation. Values are mean卤s.d. of triplicate samples from a representative of two independent experiments.Full size imageEndogenous CXCL12 expression decreased MDA-MB-231 lung metastasisIt has been observed that murine CXCL12 regulates the metastasis of xenografted CXCR4-expressing human carcinoma cells, including MDA-MB-231 cells (Muller et al., 2001). Therefore, we next sought to test the hypothesis that mammary carcinoma cells in which CXCL12 expression has been epigenetically silenced more actively respond to endocrine ligand, increasing the potential of those cells to invade organs expressing that chemokine. To this end we used our luciferase-tagged MDA-MB-231 cell clones in which CXCL12 had been re-expressed to track in vivo lung metastasis of mammary carcinoma cells upon injection into the lateral tail vein. As shown in Figure 7a, 231-luc-eGFP and 231-luc-CXCL12 cell lines were detected in the lungs of SCID mice 鈭?/span>30鈥塵in after injection. We observed that within 1 week there was a severe death phase for injected CXCL12-expressing and eGFP control cells in which luminescence was minimally, if at all, detectable in both groups of animals (Figure 7a). Within 3 weeks eGFP cells began forming multiple lung metastases resulting in drastically larger photon readings compared to those cells re-expressing CXCL12 (Figures 7a and b). In contrast, the CXCL12-expressing cells formed much fewer metastases reflected in a delayed tumor growth phase (Figures 7a and b). These data suggest that fewer of the CXCL12-expressing cells actively invaded the lung and successfully established metastases. Furthermore, survival time of mice bearing CXCL12 re-expressing xenografts was increased compared with eGFP control tumors (Figure 7c).Figure 7CXCL12 re-expression in MDA-MB-231 cells inhibits their lung metastasis. (a) 231-luc-eGFP or 231-luc-CXCL12 cells (5 脳 105) were injected into the lateral tail vein of SCID mice. Images taken at time of injection (T0) and weekly thereafter (1鈥? weeks) confirmed metastatic tumor formation in the lungs of engrafted mice. Shown are longitudinal images of representative mice from each group. (b) Luminescence readings normalized to T0 postinjection value demonstrated that cells re-expressing CXCL12 less efficiently metastasized relative to eGFP control cells. Asterisk (*) indicates statistically significant difference in luminescence (P猢?/span>0.05). (c) Mice injected with 231-luc-CXCL12 cells (CXCL12) survived longer than mice injected with control 231-luc-eGFP cells (eGFP). (d) Representative ex vivo lung tissue from mice injected with either 231-luc-CXCL12 (CXCL12) or 231-luc-eGFP cells (eGFP).Full size imageAs shown in Figure 7d, ex vivo examination of tumor-bearing lungs 8 weeks postinjection revealed that 231-luc-CXCL12 cells formed fewer macro-metastases than did the control 231-luc-eGFP cells. At 5 weeks and longer the size of metastasized tumors may have yielded inaccurate bioluminescence readings (Figure 7b) due to necrotic inner regions of established and larger tumors (Jenkins et al., 2005). After 11 weeks CXCL12-expressing cells that successfully metastasized to and invaded the lungs established comparably sized tumors as the control cells (Figure 7d). Taken together, these data suggest that reestablishment of autocrine CXCL12 expression in mammary carcinoma cells inhibits their ability to actively metastasize and invade ectopic tissues.Endogenous CXCL12 expression decreases mammary carcinoma cell responsiveness to exogenous stimulation independent of CXCR4 downregulationAs a potential mechanism for the observed disparity in homing to ectopic sources of CXCL12 during in vivo lung invasion, we hypothesized that mammary carcinoma cells in which expression of CXCL12 has been silenced possess increased responsiveness to exogenous ligand stimulation. Intracellular calcium flux linked to chemotactic migration is an established measure of CXCR4 signaling following ligand engagement (Hesselgesser et al., 1998). In support of our hypothesis, 231-luc cells exhibited a robust intracellular calcium flux in response to added CXCL12 (Figure 8a). In contrast, both CXCL12 re-expressing cell lines had a reduced response to exogenous ligand at each of the concentrations examined (Figures 8a and b). Further, decreased calcium flux observed in CXCL12 re-expressing cells correlated with a diminished in vitro chemotactic response to exogenous recombinant CXCL12 (Figure 8c). Consistent with our prior analyses (Smith et al., 2005) calcium flux in 231-luc cells was dependent upon CXCR4 signaling as pretreatment with either CXCR4 antagonist AMD3100 or pertussis toxin markedly decreased the response (data not shown). Our data demonstrating decreased calcium flux and decreased breast cancer cell chemotaxis in response to CXCL12 is supported by previous data in B cells (Brauweiler et al., 2007).Figure 8Lack of endogenous CXCL12 increases mammary carcinoma cell responsiveness to exogenous CXCL12. (a) Two separate cell lines re-expressing CXCL12 (#1 and #2) demonstrated decreased intracellular calcium flux relative to control (231-luc) cells, upon stimulation with exogenous CXCL12. (b) Maximum calcium flux confirmed decreased responsiveness of CXCL12-re-expressing cells following stimulation with indicated doses of exogenous CXCL12. Asterisk (*) indicates statistical significance between 231-luc calcium flux and indicated CXCL12-expressing cell line (P猢?/span>0.05). (c) Mammary carcinoma cells re-expressing CXCL12 have decreased chemotactic responsiveness to exogenous recombinant CXCL12 relative to control eGFP cells. (d) RT鈥揚CR analysis showing similar CXCR4 transcript levels in wild-type MDA-MB-231 cells (WT), eGFP (eGFP) and three independent CXCL12 re-expressing clones (#1, #2, #3). (e) Immunoblot analysis showing maintenance of total protein levels of CXCR4 in L-eGFP and L-CXCL12#1 cells. (f) Flow cytometric analysis verified similar levels of cell surface CXCR4 in control and CXCL12 re-expressing cells.Full size imageAs ligand binding is known to cause receptor internalization and transiently decrease cell surface levels of CXCR4 (Amara et al., 1997), we next sought to define receptor levels in our CXCL12 re-expressing cell lines. CXCR4 mRNA levels were similar between three separate CXCL12 re-expressing cells, control eGFP and wild-type MDA-MB-231 cell lines (Figure 8d). Similarly, comparable levels of total (Figure 8e) and cell surface (Figure 8f) CXCR4 protein were noted between 231-luc-CXCL12 and 231-luc-eGFP control cells over 3 days in conditioned medium, allowing for prolonged exposure to endogenous CXCL12 in those cells. Together, these data support the notion that carcinoma cells in which endogenous CXCL12 expression has been silenced are hyperresponsive to foreign sources of CXCL12, likely contributing to an increase in metastatic potential independent of CXCR4 expression regulation.DiscussionThe current paradigm suggests that increased CXCR4 expression dictates the ability of carcinoma cells to metastasize to organs such as the bone marrow and liver that express high amounts of CXCL12 (Balkwill, 2004). Our studies have explored the role for autocrine chemokine鈥揷hemokine receptor interactions specifically during metastatic invasion, where CXCR4 is known to be of critical importance (Muller et al., 2001). Clearly, we show that mammary carcinoma cells re-expressing CXCL12 have diminished invasive potential, but once established, metastasized tumors grow out as efficiently as cells lacking CXCL12. These latter data support prior reports suggesting that constitutive activation of CXCR4 is mitogenic in mammary carcinoma results (Orimo et al., 2005) and are consistent with results from our mammary fat-pad engraftment approach showing primary tumors expressing CXCL12 grow faster and form overall larger tumors compared to CXCL12 null cells. This increase in proliferation is in stark contrast to our previous report where colorectal carcinoma cells underwent increased apoptosis upon CXCL12 re-expression (Wendt et al., 2006). Taken together, these data suggest that, in mammary carcinoma, intratumoral stimulation of CXCR4 can be pro-mitogenic, increasing proliferation of primary tumor cells where CXCL12 levels are relatively high. Cells within the primary tumor in which CXCL12 expression has been silenced may be at a selective advantage to receive endocrine CXCL12 signals promoting their exit and driving more active metastasis to ectopic sources of the CXCR4 ligand. Support for that model was illustrated herein by our in vitro calcium flux and chemotaxis data showing that CXCL12 re-expressing cells respond poorly to exogenous CXCL12.CXCL12 was robustly expressed by normal mammary epithelial cells in vivo. This consistent expression was noticeably absent in primary mammary carcinoma and in vitro breast cancer cell lines. Further, we showed that absence of CXCL12 expression reflects DNA promoter hypermethylation, where 40% of primary tumors examined exhibited hypermethylation of the CXCL12 promoter. This percentage is slightly lower than the 62% frequency we have previously reported for colorectal carcinoma, possibly reflecting an alternate mechanism for modifying expression of CXCL12. Support for this interpretation is evidenced in the MCF-7 cells where little if any methylation was observed yet a majority of those cells expressed little to no detectable CXCL12 protein. Lack of CXCL12 hypermethylation in MCF7 cells parallels a recent study showing lack of gene silencing by DNA methylation in those cells as compared to MDA-MB-231 cells (McGarvey et al., 2006). Therefore MCF7 cells seem to downregulate CXCL12 expression by means other than DNA methylation, stressing the importance of this process to the cancer cell phenotype. Moreover, a recent in vivo comparative analysis of a subset of MCF7 cells expressing CXCL12 against MDA-MB-231 cells completely lacking expression of the gene indicate that the former failed to metastasize, mirroring the phenotype of our MDA-MB-231 cells stably re-expressing CXCL12 (Dewan et al., 2006). These data suggest that complete silencing of CXCL12 through DNA methylation establishes a leukocyte phenotype promoting more effective metastasis of CXCR4-expressing cells to ectopic sources of CXCL12.Several clinical studies have recently examined increased expression of CXCR4 as a possible predictor of mammary carcinoma metastasis (Salvucci et al., 2006). Detection of an increase in CXCR4 requires a large tissue sample for immunohistochemical or mRNA expression analysis and need for CXCR4 to be increased in a large percentage of the tumor. Data shown here utilize small biopsy samples for DNA isolation and MSP to identify even a very few number of cells that are of high metastatic potential due to epigenetic silencing of CXCL12. Moreover, our data in normal breast epithelium, together with our previous data in 20 normal colonic epithelial specimens, strongly support CXCL12 methylation as a pathologic event during neoplastic transformation (Wendt et al., 2006). Thus, detection of CXCL12 methylation appears to be more sensitive and reliably diagnostic of tumorigenesis than assessment of increased CXCR4 expression.In summary, the homeostatic expression of CXCL12, but not CXCR4, is a target for gene silencing in mammary carcinoma. Our data support the notion that chemokine silencing aids in carcinoma disease progression as reestablishing normal CXCL12 expression reduced in vitro chemotaxis and in vivo metastasis by mammary carcinoma cells. These findings are consistent with and expand upon previous data concerning the role of CXCR4 signaling in carcinoma cell metastasis. Our results, together with recent findings emphasizing the importance of CXCR4 signaling in cancer cell migration and invasion, constitute a unique observation that loss of autocrine CXCL12 and the concordant imbalance in CXCR4 signaling plays a role in the increased metastasis of mammary and colorectal cancer cells.Materials and methodsHuman mammary carcinoma cell linesMDA-MB-231, MCF-7, MCF-10A, MCF-12A, MDA-MB-134-IV, MDA-MB-453, MDA-MB-435s and BT549 breast cancer cell lines were purchased from ATCC and maintained using recommended culture conditions.RT鈥揚CR analysisTotal RNA was isolated from cultured cells and converted to cDNA as previously described (Wendt et al., 2006). Primer sequences and amplification conditions are listed in Supplementary Material.Immuno analysesProtein detection by immunoblot and immunohistochemistry was conducted as previously described (Smith et al., 2005; Wendt et al., 2006). Cell surface CXCR4 was detected using fluorescein isothiocyanate-conjugated antibodies specific for human CXCR4 (R D Systems, Minneapolis, MN, USA) according to the manufacturer\'s instructions and quantified by flow cytometry.Methylation-specific PCR and bisulfite sequencingIsolated genomic DNA was bisulfite-treated and subjected to MSP and bisulfite sequencing PCR as described in Supplementary Material.Construction of stable cell linesMDA-MB-231 cells were transfected with pcDNA3.1 (Invitrogen, Carlsbad, CA, USA) encoding a firefly luciferase, and stable plasmid integration was selected using G418 sulfate (EMD Biosciences, La Jolla, CA, USA) yielding the 231-luc cell line. Those cells were then transfected with the pTre plasmid (Stratagene, La Jolla, CA, USA) encoding either eGFP or human CXCL12 and stable plasmid integration was selected using hygromycin to yield the 231-luc-eGFP and 231-luc-CXCL12 cell lines.ChemotaxisU937 monocytes (24-h serum-starved) were pretreated with AMD3100 (Sigma, St Louis, MO, USA) or remained untreated, and loaded with Calcein-AM (Molecular Probes, Eugene, OR, USA). Cells (5 脳 105) were plated to the upper well of a chemotaxis chamber (5鈥壩糾 pore size, Corning Costar, Corning, NY, USA), with conditioned media from the varying cell lines added to the bottom chamber. Cells were incubated 90鈥塵in and the cell migration quantified by fluorescence spectroscopy (Victor Wallac, Perkin Elmer, Turku, Finland). Mammary carcinoma chemotaxis was assayed with recombinant CXCL12 (50鈥塶g ml鈭?) added to the bottom chamber, incubated 18鈥塰 and migrated cells quantified by luminescence.SCID mouse bioluminescence analysesUsing protocols approved by the Medical College of Wisconsin Institutional Animal Care and Use Committee, cells suspended in a 100鈥壩糽 volume of sterile phosphate-buffered saline were injected into the lateral tail vein (5 脳 105 cells) or the mammary fat-pad (2 脳 106 cells) of 8-week-old female SCID mice (cr-PrdkcScid, Charles Rivers, Wilmington, MA, USA). Mice were monitored by bioluminescence imaging for tumor development using the Lumina IVIS-100 In Vivo Imaging System (Xenogen Corp, Alameda, CA, USA).Cell cycle and proliferation analysisEthanol-fixed cells were stained in 50鈥壩糶 ml鈭? propidium iodide (EMD Biosciences) and 10鈥壩糶 ml鈭? RNAse A (Promega, Madison, WI, USA) and analysed by flow cytometry. DNA synthesis was measured by the addition of 3H thymidine (1鈥壩糃i ml鈭?) with fresh media and incorporated thymidine quantified by scintillation counting (Beckman Coulter, Fullerton, CA, USA).Apoptosis assayProgrammed cell death in mammary carcinoma cells was assayed using a Caspase 3/7 kit (Promega) as we defined previously (Wendt et al., 2006).Intracellular calcium fluxSubconfluent cells were loaded with cell permeant Fluo-4 AM (Molecular Probes). CXCL12 was added at indicated concentrations and intracellular calcium flux was measured by fluorescence spectroscopy (Victor Wallac). ReferencesAgace WW, Amara A, Roberts AI, Pablos JL, Thelen S, Uguccioni M et al. (2000). Constitutive expression of stromal derived factor-1 by mucosal epithelia and its role in HIV transmission and propagation. Curr Biol 10: 325鈥?28.CAS聽 Article聽Google Scholar聽 Amara A, Gall SL, Schwartz O, Salamero J, Montes M, Loetscher P et al. (1997). HIV coreceptor downregulation as antiviral principle: SDF-1alpha-dependent internalization of the chemokine receptor CXCR4 contributes to inhibition of HIV replication. J Exp Med 186: 139鈥?46.CAS聽 Article聽Google Scholar聽 Balkwill F . (2004). The significance of cancer cell expression of the chemokine receptor CXCR4. Semin Cancer Biol 14: 171鈥?79.CAS聽 Article聽Google Scholar聽 Baylin SB, Chen WY . (2005). Aberrant gene silencing in tumor progression: implications for control of cancer. Cold Spring Harb Symp Quant Biol 70: 427鈥?33.427鈥?33.CAS聽 Article聽Google Scholar聽 Brauweiler A, Merrell K, Gauld SB, Cambier JC . (2007). Cutting Edge: acute and chronic exposure of immature B cells to antigen leads to impaired homing and SHIP1-dependent reduction in stromal cell-derived factor-1 responsiveness. J Immunol 178: 3353鈥?357.CAS聽 Article聽Google Scholar聽 Clark SJ, Harrison J, Molloy PL . (1997). Sp1 binding is inhibited by (m)Cp(m)CpG methylation. Gene 195: 67鈥?1.CAS聽 Article聽Google Scholar聽 Dewan MZ, Ahmed S, Iwasaki Y, Ohba K, Toi M, Yamamoto N . (2006). Stromal cell-derived factor-1 and CXCR4 receptor interaction in tumor growth and metastasis of breast cancer. Biomed Pharmacother 60: 273鈥?76.CAS聽 Article聽Google Scholar聽 Ellison G, Klinowska T, Westwood RF, Docter E, French T, Fox JC . (2002). Further evidence to support the melanocytic origin of MDA-MB-435. Mol Pathol 55: 294鈥?99.CAS聽 Article聽Google Scholar聽 Garcia-Moruja C, Alonso-Lobo JM, Rueda P, Torres C, Gonzalez N, Bermejo M et al. (2005). Functional characterization of SDF-1 proximal promoter. J Mol Biol 348: 43鈥?2.CAS聽 Article聽Google Scholar聽 Gazitt Y . (2004). Homing and mobilization of hematopoietic stem cells and hematopoietic cancer cells are mirror image processes, utilizing similar signaling pathways and occurring concurrently: circulating cancer cells constitute an ideal target for concurrent treatment with chemotherapy and antilineage-specific antibodies. Leukemia 18: 1鈥?0.CAS聽 Article聽Google Scholar聽 Heidemann J, Ogawa H, Dwinell MB, Rafiee P, Maaser C, Gockel HR et al. (2003). Angiogenic effects of interleukin 8 (CXCL8) in human intestinal microvascular endothelial cells are mediated by CXCR2. J Biol Chem 278: 8508鈥?515.CAS聽 Article聽Google Scholar聽 Herman JG, Baylin SB . (2003). Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med 349: 2042鈥?054.CAS聽 Article聽Google Scholar聽 Hesselgesser J, Liang M, Hoxie J, Greenberg M, Brass LF, Orsini MJ et al. (1998). Identification and characterization of the CXCR4 chemokine receptor in human T cell lines: ligand binding, biological activity, and HIV-1 infectivity. J Immunol 160: 877鈥?83.CAS聽Google Scholar聽 Jenkins DE, Hornig YS, Oei Y, Dusich J, Purchio T . (2005). Bioluminescent human breast cancer cell lines that permit rapid and sensitive in vivo detection of mammary tumors and multiple metastases in immune deficient mice. Breast Cancer Res 7: R444鈥揜454.CAS聽 Article聽Google Scholar聽 Kimura R, Nishioka T, Ishida T . (2003). The SDF1-G801A polymorphism is not associated with SDF1 gene expression in Epstein-Barr virus-transformed lymphoblastoid cells. Genes Immun 4: 356鈥?61.CAS聽 Article聽Google Scholar聽 McGarvey KM, Fahrner JA, Greene E, Martens J, Jenuwein T, Baylin SB . (2006). Silenced tumor suppressor genes reactivated by DNA demethylation do not return to a fully euchromatic chromatin state. Cancer Res 66: 3541鈥?549.CAS聽 Article聽Google Scholar聽 Moyer RA, Wendt MK, Johanesen PA, Turner JR, Dwinell MB . (2007). Rho activation regulates CXCL12 chemokine stimulated actin rearrangement and restitution in model intestinal epithelia. Lab Invest 87: 807鈥?17.CAS聽 Article聽Google Scholar聽 Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME et al. (2001). Involvement of chemokine receptors in breast cancer metastasis. Nature 410: 50鈥?6.CAS聽 Article聽Google Scholar聽 Nagasawa T, Hirota S, Tachibana K, Takakura N, Nishikawa S, Kitamura Y et al. (1996a). Defects of B-cell lymphopoiesis and bone-marrow myelopoiesis in mice lacking the CXC chemokine PBSF/SDF-1. Nature 382: 635鈥?38.CAS聽 Article聽Google Scholar聽 Nagasawa T, Nakajima T, Tachibana K, Iizasa H, Bleul CC, Yoshie O et al. (1996b). Molecular cloning and characterization of a murine pre-B-cell growth-stimulating factor/stromal cell-derived factor 1 receptor, a murine homolog of the human immunodeficiency virus 1 entry coreceptor fusin. Proc Natl Acad Sci USA 93: 14726鈥?4729.CAS聽 Article聽Google Scholar聽 Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delaunay T, Naeem R et al. (2005). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121: 335鈥?48.CAS聽 Article聽Google Scholar聽 Phillips RJ, Burdick MD, Lutz M, Belperio JA, Keane MP, Strieter RM . (2003). The stromal derived factor-1/CXCL12-CXC chemokine receptor 4 biological axis in non-small cell lung cancer metastases. Am J Respir Crit Care Med 167: 1676鈥?686.Article聽Google Scholar聽 Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, Schuebel KE et al. (2002). DNMT1 and DNMT3b cooperate to silence genes in human cancer cells. Nature 416: 552鈥?56.CAS聽 Article聽Google Scholar聽 Rot A, von Andrian UH . (2004). Chemokines in innate and adaptive host defense: basic chemokinese grammar for immune cells. Annu Rev Immunol 22: 891鈥?28.CAS聽 Article聽Google Scholar聽 Salvucci O, Bouchard A, Baccarelli A, Deschenes J, Sauter G, Simon R et al. (2006). The role of CXCR4 receptor expression in breast cancer: a large tissue microarray study. Breast Cancer Res Treat 97: 275鈥?83.CAS聽 Article聽Google Scholar聽 Schrader AJ, Lechner O, Templin M, Dittmar KE, Machtens S, Mengel M et al. (2002). CXCR4/CXCL12 expression and signalling in kidney cancer. Br J Cancer 86: 1250鈥?256.CAS聽 Article聽Google Scholar聽 Shirozu M, Nakano T, Inazawa J, Tashiro K, Tada H, Shinohara T et al. (1995). Structure and chromosomal localization of the human stromal cell-derived factor 1 (SDF1) gene. Genomics 28: 495鈥?00.CAS聽 Article聽Google Scholar聽 Smith JM, Johanesen PA, Wendt MK, Binion DG, Dwinell MB . (2005). CXCL12 activation of CXCR4 regulates mucosal host defense through stimulation of epithelial cell migration and promotion of intestinal barrier integrity. Am J Physiol Gastrointest Liver Physiol 288: G316鈥揋326.CAS聽 Article聽Google Scholar聽 Smith MC, Luker KE, Garbow JR, Prior JL, Jackson E, Piwnica-Worms D et al. (2004). CXCR4 regulates growth of both primary and metastatic breast cancer. Cancer Res 64: 8604鈥?612.CAS聽 Article聽Google Scholar聽 Tachibana K, Hirota S, Iizasa H, Yoshida H, Kawabata K, Kataoka Y et al. (1998). The chemokine receptor CXCR4 is essential for vascularization of the gastrointestinal tract. Nature 393: 591鈥?94.CAS聽 Article聽Google Scholar聽 Ueda Y, Neel NF, Schutyser E, Raman D, Richmond A . (2006). Deletion of the COOH-terminal domain of CXC chemokine receptor 4 leads to the down-regulation of cell-to-cell contact, enhanced motility and proliferation in breast carcinoma cells. Cancer Res 66: 5665鈥?675.CAS聽 Article聽Google Scholar聽 Wendt MK, Johanesen PA, Kang-Decker N, Binion DG, Shah V, Dwinell MB . (2006). Silencing of epithelial CXCL12 expression by DNA hypermethylation promotes colonic carcinoma metastasis. Oncogene 25: 4986鈥?997.CAS聽 Article聽Google Scholar聽 Zeelenberg IS, Ruuls-Van Stalle L, Roos E . (2003). The chemokine receptor CXCR4 is required for outgrowth of colon carcinoma micrometastases. Cancer Res 63: 3833鈥?839.CAS聽Google Scholar聽 Zou YR, Kottmann AH, Kuroda M, Taniuchi I, Littman DR . (1998). Function of the chemokine receptor CXCR4 in haematopoiesis and in cerebellar development. Nature 393: 595鈥?99.CAS聽 Article聽Google Scholar聽 Download referencesAcknowledgementsWe thank Dr Richard Komorowski (Department of Pathology, Medical College of Wisconsin) and Dr Eric Luedke for assistance in obtaining and processing primary human mammary carcinoma tissue and Dr Robert Truitt (Cancer Center of the Medical College of Wisconsin and Director of the Biophotonic Imaging Core) for assistance in these studies. These studies were supported in part by grants from the Cancer Center of the Medical College of Wisconsin and the American Cancer Society (IRG-84-004).Author informationAffiliationsDepartment of Microbiology and Molecular Genetics, Medical College of Wisconsin, Milwaukee, WI, USAM K Wendt,聽A N Cooper聽 聽M B DwinellAuthorsM K WendtView author publicationsYou can also search for this author in PubMed聽Google ScholarA N CooperView author publicationsYou can also search for this author in PubMed聽Google ScholarM B DwinellView author publicationsYou can also search for this author in PubMed聽Google ScholarCorresponding authorCorrespondence to M B Dwinell.Additional informationSupplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).Supplementary information Supplementary Methods (DOC 24 kb)Rights and permissionsReprints and PermissionsAbout this articleCite this articleWendt, M., Cooper, A. Dwinell, M. Epigenetic silencing of CXCL12 increases the metastatic potential of mammary carcinoma cells. Oncogene 27, 1461鈥?471 (2008). https://doi.org/10.1038/sj.onc.1210751Download citationReceived: 09 May 2007Revised: 25 July 2007Accepted: 01 August 2007Published: 03 September 2007Issue Date: 28 February 2008DOI: https://doi.org/10.1038/sj.onc.1210751KeywordschemokineDNA hypermethylationcancermetastasisCXCR4 Cl谩udia S. Marques, Ana Rita Santos, Andreia Gameiro, Jorge Correia Fernando Ferreira BMC Cancer (2018) Sylvain Lecomte, Frederic Chalmel, Fran莽ois Ferriere, Frederic Percevault, Nicolas Plu, Christian Saligaut, Claire Surel, Marie Lelong, Theo Efstathiou Farzad Pakdel Cell Communication and Signaling (2017) Harsh Samarendra, Keaton Jones, Tatjana Petrinic, Michael A Silva, Srikanth Reddy, Zahir Soonawalla Alex Gordon-Weeks British Journal of Cancer (2017) Ishan Roy, Kathleen A Boyle, Emily P Vonderhaar, Noah P Zimmerman, Egal Gorse, A Craig Mackinnon, Rosa F Hwang, Janusz Franco-Barraza, Edna Cukierman, Susan Tsai, Douglas B Evans Michael B Dwinell Laboratory Investigation (2017) Alexander Semaan, Dimo Dietrich, Dominik Bergheim, J枚rn Dietrich, J枚rg C. Kalff, Vittorio Branchi, Hanno Matthaei, Glen Kristiansen, Hans-Peter Fischer Diane Goltz Virchows Archiv (2017)