The prostaglandin F2α (PGF2α) receptor (FP) is elevated in endometrial adenocarcinoma. This study found that PGF2α signaling via FP regulates expression of chemokine (C-X-C motif) ligand 1 (CXCL1) in endometrial adenocarcinoma cells. Expression of CXCL1 and its receptor, CXCR2, are elevated in cancer tissue compared with normal endometrium and localized to glandular epithelium, endothelium, and stroma. Treatment of Ishikawa cells stably transfected with the FP receptor (FPS cells) with 100 nmol/L PGF2α increased CXCL1 promoter activity, mRNA, and protein expression, and these effects were abolished by cotreatment of cells with FP antagonist or chemical inhibitors of Gq, epidermal growth factor receptor, and extracellular signal-regulated kinase. Similarly, CXCL1 was elevated in response to 100 nmol/L PGF2α in endometrial adenocarcinoma explant tissue. CXCL1 is a potent neutrophil chemoattractant. The expression of CXCR2 colocalized to neutrophils in endometrial adenocarcinoma and increased neutrophils were present in endometrial adenocarcinoma compared with normal endometrium. Conditioned media from PGF2α-treated FPS cells stimulated neutrophil chemotaxis, which could be abolished by CXCL1 protein immunoneutralization of the conditioned media or antagonism of CXCR2. Finally, xenograft tumors in nude mice arising from inoculation with FPS cells showed increased neutrophil infiltration compared with tumors arising from wild-type cells or following treatment of mice bearing FPS tumors with CXCL1-neutralizing antibody. In conclusion, our results show a novel PGF2α-FP pathway that may regulate the inflammatory microenvironment in endometrial adenocarcinoma via neutrophil chemotaxis. [Cancer Res 2009;69(14):5726–33]
- Prostaglandin F2α
- endometrial cancer
Endometrial adenocarcinoma is the most common gynecological malignancy in Western countries, affecting mainly postmenopausal women with a frequency of 15 to 20 per 100,000 women per year ( 1). Overexpression of the cyclooxygenase (COX) enzymes and prostaglandins has been shown in endometrial adenocarcinoma as well as a number of other cancer types and gynecological pathologies ( 2, 3).
In the reproductive tract, the most commonly synthesized prostaglandins are the E- and F-series prostanoids ( 4). These are synthesized from arachidonic acid by COX enzymes and prostaglandin synthases, and are then transported out of the cell by a prostaglandin transporter ( 5) to act in an autocrine/paracrine manner on G-protein–coupled receptors. The G-protein–coupled receptor for PGF2α (FP) is a Gq coupled receptor, which upon activation leads to release of inositol-1,4,5-trisphosphate and diacylglycerol ( 6). Recently, we have shown a role for FP in endometrial adenocarcinoma, with evidence for elevated PGF2α-FP signaling up-regulating angiogenic and tumorigenic genes including COX-2 ( 7), FGF2 ( 8), and vascular endothelial growth factor (VEGF; ref. 9), and increasing proliferation and migration of neoplastic epithelial cells ( 10– 12).
Chemokine (C-X-C motif) ligand 1 (CXCL1, also known as growth-regulated oncogene α) has angiogenic, chemoattractant, and inflammatory activities ( 13). A link between prostaglandins and CXCL1 has been shown, as prostaglandin E2 signaling induces CXCL1 expression in colorectal cancer cell lines ( 14). CXCL1 is up-regulated in melanoma ( 15, 16), colorectal ( 17, 18), and prostate cancer ( 19) and binds to the CXCR2 receptor ( 20) to promote the recruitment of neutrophils to sites of inflammation ( 21). Although the role of neutrophils in cancer is unclear, recent evidence suggests that they may promote tissue remodeling by production of proteases including matrix metalloproteinase (MMP)-9 ( 22) and angiogenic factors such as VEGF ( 23), in addition to their classic role as the first line of defense against invading pathogens ( 24).
In this study, we used a chemokine antibody array to identify CXCL1 as a target gene of PGF2α. We investigated the expression, localization, and regulation of CXCL1 expression mediated via PGF2α-FP signaling in endometrial adenocarcinoma and its downstream regulation of neutrophil influx into endometrial tumors in vitro and in endometrial tumor xenografts in vivo.
Materials and Methods
Reagents. Indomethacin, PBS, bovine serum albumin (BSA), AL8810, Tri-reagent, and PGF2α were purchased from Sigma Co. PD98059, AG1478, Cyclosporin A, and 4C3MQ were purchased from Calbiochem. CXCR2 and Gr-1 (a murine neutrophil marker) antibodies were purchased from R&D systems, and CXCL1 and neutrophil elastase antibodies from Santa Cruz Biotechnology and DAKO, respectively. FITC-CD11b, PE-GR-1, and Cy5-CD11c antibodies were obtained from eBioscience. The chemokine antibody array was purchased from RayBiotech.
Patients and tissue collection. Endometrial adenocarcinoma tissue and normal tissue was obtained as detailed in our prior studies ( 9, 10). Cancer patients were prediagnosed to have adenocarcinoma of the uterus, and diagnosis was confirmed histologically in all cases. Normal endometrial tissue was collected from women undergoing surgery for minor gynecologic procedures with no underlying endometrial pathology. Ethical approval was obtained from Lothian Research Ethics Committee, and written informed consent was obtained from all subjects before tissue collection.
Cell lines, culture, and treatments. Wild-type (WT) Ishikawa cells and Ishikawa cells engineered to stably express the full-length human FP receptor to the levels observed in endometrial adenocarcinomas, called FPS cells, were cultured as described previously ( 9). Transient transfections were performed using Superfect (Qiagen) as per the manufacturer's protocol. The optimal concentrations of all chemical inhibitors and antibodies were determined empirically by titration using the manufacturer's guidelines as described in our previous studies ( 25). Cell viability was determined for each inhibitor using the CellTitre 96 AQueous One Solution assay (Promega) as described in our previous studies ( 10, 25). Within the time frame of incubation in this study, the inhibitors had no adverse effect on cell viability, at the concentration used. Cells were treated with vehicle, inhibitor alone, or 100 nmol/L PGF2α alone or in the presence of YM254890 (1 μmol/L), AL8810 (50 μmol/L), 43CMQ (1 μmol/L), AG1478 (200 nmol/L), or Cyclosporine A (CsA; 1 μmol/L) for the time indicated.
Chemokine antibody array. Conditioned medium (CM) was prepared as described ( 8). Briefly, FPS cells were stimulated with vehicle or 100 nmol/L PGF2α for 24 h. The CM (V-CM or P-CM, respectively) was analyzed for cytokine expression using the RayBio Human Cytokine Antibody Array 3 kit, following manufacturer's protocol.
CXCL1 luciferase reporter assays. The CXCL1 reporter plasmid consisting of the CXCL1 promoter fused to the firefly luciferase reporter (as described in ref. 26) was kindly supplied by Professor Ann Richmond (Department of Cell Biology, Vanderbilt University School of Medicine, Nashville, Tennessee). The CXCL1 promoter firefly luciferase reporter was transfected into FPS cells with pRL-TK (containing the renilla luciferase coding sequence; Promega) as an internal control. FPS cells were cotransfected with control vector (pcDNA3.0) or vector encoding a dominant negative (DN) isoform of EGFR, Ras, MAP/ERK kinase (MEK), or NFAT (kindly supplied by Professor Zvi Naor, Tel Aviv University, Israel). The DN constructs have been previously characterized and described ( 27, 28). Cells were stimulated with vehicle or 100 nmol/L PGF2α for the time indicated in the figure legend. The activity of firefly and renilla was determined using the dual luciferase assay kit (Promega) and total luciferase activity determined relative to the internal renilla control. Data are expressed as fold increase in luciferase activity compared with vehicle-treated cells and are presented as mean ± SE from at least three independent experiments.
Taqman quantitative reverse transcription-PCR. CXCL1 and CXCR2 expression in endometrial tissues and FPS cells was measured by quantitative reverse transcription-PCR analysis as described previously ( 1, 2). FPS cells were treated with vehicle, 100 nmol/L PGF2α alone or in the presence of inhibitor, or vehicle and inhibitor alone for 8 h. RNA samples were then extracted using Tri-reagent following manufacturer's guidelines. RNA samples were reverse transcribed and RT-PCR performed as described previously ( 1, 2) using sequence specific primers and probes: CXCL1 forward, 5′-GTT TTC AAA TGT TCT CCA GTC ATT ATG-3′; reverse, 5′-CCG CCA GCC TCT ATC ACA GT-3′; probe, 5′-TTC TGA GGA GCC TGC AAC ATG CCA-3′; CXCR2 forward, 5′-TGC TCT TCT GGA GGT GTC CTA CA-3′; reverse, 5′-AGA TCT TCA CCT TTC CAG AAA TCT T-3′; probe, 5′-CCC AGC GAC CCA GTC AGG ATT TAA-3′. Primers and data were analyzed and processed using Sequence Detector v1.6.3 (Applied Biosystems). Expression of analyzed genes was normalized to RNA loading for each sample using the 18S rRNA as an internal standard. Results are expressed as fold increase above cells treated with vehicle and inhibitor. Data are presented as mean ± SE from at least three independent experiments.
CXCL1 ELISA. CXCL1 protein secretion by FPS cells in the culture media was measured by the Human CXCL1 Quantikine ELISA kit (R&D systems). FPS cells were treated as described above for mRNA, for time indicated in the figure legend. The ELISA was then carried out according to manufacturer's instructions. Absorbance of wells was determined by spectrophotometry at 450 nm. Data are presented as mean ± SE from at least three independent experiments.
Immunohistochemical analysis. Expression of CXCL1, neutrophil elastase, CXCR2, and Gr-1 was localized in endometrial tissue and xenografts by immunohistochemistry using standard techniques as described previously ( 1, 2). Briefly, following antigen retrieval, sections were blocked in 5% normal rabbit serum (CXCL1 and Gr-1) or normal goat serum (CXCR2 and neutrophil elastase) diluted in TBS with 5% BSA. Subsequently, tissue sections were incubated with goat anti-human CXCL1 polyclonal antibody (2 μg/mL), mouse anti-human CXCR2 (5 μg/mL), mouse anti-human neutrophil elastase monoclonal antibody (2 μg/mL), or rat anti-mouse Gr-1 (5 μg/mL) overnight at 4°C. Control sections included the following: no primary antibody, nonimmune goat, mouse and rat IgG, or CXCL1 antibody preabsorbed with blocking peptide (20 μg/mL; Santa Cruz Biotechnology). Subsequently, sections were incubated with rabbit anti-goat/rat biotinylated or goat anti-mouse biotinylated antibodies (DAKO), followed by streptavidin-horse radish peroxidase complex (DAKO). Color reaction was developed with 3′3 diaminobenzidine (DAKO). Sections were counterstained in hemotoxylin. Images were obtained on a PROVIS microscope at ×200 or ×400 magnification (Olympus Optical) using Canon EOS image capture software (Canon). The number of neutrophils was quantified using neutrophil elastase staining and standard stereological techniques. Briefly, each section was examined using ×40 plan apo objective from a BH2 microscope (Olympus) fitted with an automatic stage (Prior Scientific Instruments Ltd.) using a video camera (HV-C20; Hitachi) and analyzed with Image-ProPlus 4.5.1 software with a Stereology 5.0 plug-in (Media Cybernetics). A total of 40 randomized fields of view were examined and counted (n = 7 normal endometrium; n = 30 carcinoma), and data are expressed as mean number of cells per mm2 of tumor examined.
Immunofluorescence microscopy. CXCR2 expression was colocalized with neutrophil elastase by immunofluorescence microscopy as described previously ( 8, 29). Briefly, sections were blocked in 5% normal goat serum diluted in PBS with 5% BSA before incubation with mouse anti-neutrophil elastase (1 μg/mL). Following overnight incubation at 4°C, sections were incubated with goat-anti mouse biotinylated Fab, then tyramide signal amplification kit (TSA Fluorescein System; 1:50 dilution; Perkin-Elmer). Sections were then microwaved in 0.01 M citrate buffer for 30 min and endogenous peroxidase blocked using hydrogen peroxide. Nonspecific binding was blocked with 5% normal goat serum and sections were incubated with mouse anti-CXCR2 antibody (1 μg/mL) at 4°C overnight. Sections were again incubated with goat-anti-mouse biotinylated Fab and tyramide signal amplification kit. Nuclei were stained using ToPro (Molecular Probes). Fluorescent images were visualized and photographed using a Carl Zeiss laser scanning microscope LSM510 (×400 objective; Jena).
Neutrophil chemotaxis assay. Neutrophil chemotaxis was analyzed using transwell inserts (5-μm pore size; Corning Costar). Neutrophils were purified as previously described ( 30) and resuspended in serum free media. Cells (750,000) were added to the top chamber of the transwell insert and 600 μL of V-CM or P-CM was added to the bottom chamber. Serum-free media alone or with 50,000 pg/mL CXCL1 were added as negative and positive controls, respectively. Cells were incubated at 37°C in a 5% CO2 atmosphere for 1 h and the plate was gently tapped to dislodge cells adhered to the underside of the membrane. Cells in the bottom chamber were collected and counted at least six times using a haemocytometer. Data are expressed as mean ± SE from at least three independent experiments.
Xenograft tumor model. A suspension of 500,000 Ishikawa WT or FPS cells in a total volume of 0.2 mL DMEM was injected s.c. into each dorsal flank of CD1-Foxn1nu mice (Charles River). The mice (n = 30) were divided into two groups of equal tumor size after engraftment (1 wk). The mice were injected twice weekly with 100 μg IgG (WT and FPS) or CXCL1 neutralizing antibody (FPS) via i.p. injection for 4 wk. One tumor from each mouse was placed in PBS for flow cytometry analysis and RNA extracted from the second tumor from each mouse. The animals were maintained under sterile conditions in individually vented cages.
Flow cytometry analysis. Xenografts from nude mice were assessed for immune cell infiltrate using flow cytometry (n = 15). Briefly, tumors were digested by collagenase treatment at 37°C for 45 min. Tissue was then mechanically disrupted into a single-cell solution using a syringe and 40 μm mesh and resuspended in fluorescence-activated cell sorting (FACS) wash (PBS + 1%BSA + 2% formalin). Cells were incubated at 4°C for 30 min in FACS wash containing the following monoclonal antibodies and appropriate isotype controls: FITC-CD11b, PE-Gr-1, and Cy5-CD11c. RBC were lysed using BD FACS lysing solution according to manufacturer's instructions (BD Biosciences). Samples were analyzed using a FACScalibur cytometer (BD biosystems) using BD CellQuest software. Neutrophils were defined by expression of Gr-1 and CD11b epitope, absence of CD11c, and scatter profile.
Statistical analysis. Where appropriate, data were subjected to statistical analysis with ANOVA and Student's t test (GraphPad Prism).
CXCL1 expression in FPS cells. Changes in cytokine expression in FPS cells in response to PGF2α treatment were examined by cytokine antibody array ( Fig. 1A ). A combined up-regulation of CXCL1, CXCL2, and CXCL3 as well as CXCL1 alone was observed following 100 nmol/L PGF2α treatment of FPS cells for 24 hours compared with vehicle-treated cells. To verify this finding, the promoter activity ( Fig. 1B), mRNA ( Fig. 1C), and protein ( Fig. 1D) expression of CXCL1 in response to PGF2α treatment was examined. All were significantly increased (P < 0.01) in response to PGF2α treatment in a time-dependent manner compared with vehicle-treated cells.
Involvement of epidermal growth factor receptor and MEK signaling in CXCL1 production. To determine signaling pathways mediating CXCL1 production in FPS cells, we treated cells with vehicle, 100 nmol/L PGF2α alone or with a panel of chemical inhibitors of cell signaling, or inhibitor alone ( Fig. 2 ). Treatment of FPS cells with PGF2α for 8 and 24 hours induced a 91.5- ± 8.4-fold and 22.3- ± 4.7-fold increase in CXCL1 mRNA and protein expression, respectively, compared with vehicle treated cells ( Fig. 2A and B). This increase was abolished by treatment of cells with a selective inhibitor of Gq (YM254890; P < 0.01) and significantly inhibited with the FP receptor antagonist AL8810 (P < 0.05) and inhibitors of epidermal growth factor receptor (EGFR; AG1478; P < 0.05) and MEK (PD98059; P < 0.01). Inhibitors of calcineurin (CsA) and protein kinase A (43CMQ) did not significantly affect CXCL1 mRNA and protein production.
We confirmed a role for EGFR and extracellular signal-regulated kinase (ERK) in PGF2α-mediated CXCL1 production by cotransfecting FPS cells with the CXCL1 promoter and either an empty vector (pcDNA3.0) or DN EGFR, DNRas, DNMEK, or DN-Nuclear factor of activated T-cells (NFAT; Fig. 2C). Treatment of control-vector–transfected cells with 100 nmol/L PGF2α showed an elevation of CXCL1 promoter activity of 18.9- ± 3.6-fold, which was significantly reduced by cotransfection of cells with DN-EGFR (P < 0.05), DN-ras (P < 0.01), and DN-MEK (P < 0.01), but no significant difference was shown when cells were transfected with DN-NFAT.
CXCL1 and CXCR2 expression in endometrial adenocarcinoma and normal endometrium. Because we had ascertained a role for the FP receptor in regulating CXCL1 in an endometrial adenocarcinoma cell line, we next investigated the expression and regulation of CXCL1 in endometrial adenocarcinoma explants by PGF2α. Quantitative RT-PCR analysis showed an increase in the expression of CXCL1 and its receptor CXCR2 mRNA expression in human endometrial adenocarcinoma tissue (n = 58) compared with normal endometrium (n = 45; 5.9- and 4.2-fold, respectively; P < 0.001; Fig. 3A and B ). We investigated whether CXCL1 expression in endometrial adenocarcinoma explants was regulated via FP and MEK signaling pathways. Carcinoma tissue was treated with PGF2α in the absence/presence of AL8810 and PD98059 for 24 hours. CXCL1 mRNA was found to be elevated 5.3- ± 0.8-fold in response to PGF2α (P < 0.05). Cotreatment of tissue with AL8810 or PD988059 significantly reduced this increase in CXCL1 expression (P < 0.05).
Localization of CXCL1 and CXCR2 in endometrial adenocarcinoma. The site of expression of CXCL1 and CXCR2 protein in carcinoma tissue was then determined by immunohistochemistry ( Fig. 4A ). CXCL1 and CXCR2 immunoreactivity was localized to glandular epithelium, vascular endothelial cells, and stroma in all well, moderately, and poorly differentiated carcinoma sections studied (n = 4 each group).
Using serial sectioning, CXCL1 expression could be localized to the same glandular epithelial and vascular endothelial cells as the FP receptor ( Fig. 4B, arrowheads). CXCL1 has been previously described as a potent neutrophil chemoattractant. We next colocalized CXCR2 expression throughout the stroma with expression of neutrophil elastase, a neutrophil marker, in endometrial adenocarcinoma using dual immunofluorescence immunohistochemistry ( Fig. 4C). No immunoreactivity was observed in sections incubated with nonimmune IgG. The number of neutrophils present in endometrial tissue was then quantified using immunohistochemistry for neutrophil elastase ( Fig. 4D) and was found to be 13.9- ± 2.3-fold higher in cancer compared with sections of normal endometrium (P < 0.01).
PGF2α-stimulated CXCL1 induces neutrophil chemotaxis in vitro and in vivo. We next determined whether the CXCL1 expressed in FPS cells via PGF2α-FP receptor interaction could induce neutrophil chemotaxis. Human neutrophils were purified from peripheral blood and used in a chemotaxis assay. We found a significant increase in neutrophil chemotaxis in response to conditioned media from FPS cells treated with 100 nmol/L PGF2α (P-CM) compared with vehicle-treated cells (V-CM; Fig. 5A ). This effect was significantly inhibited with immunoneutralization of CXCL1 before incubation with neutrophils or with the addition of the CXCR2 antagonist SB225002 to P-CM (P < 0.001).
To explore whether FP receptor signaling could promote neutrophil migration in vivo, we injected WT or FPS cells s.c. in nude mice. Mice were then regularly injected with control IgG (WT and FPS xenografts) or CXCL1 antibody (FPS xenografts). Tumors formed from FPS cells expressed significantly higher CXCL1 mRNA compared with WT tumors ( Fig. 5B) and, when analyzed by flow cytometry, had increased neutrophil infiltration (P < 0.001; Fig. 5C). This infiltration was significantly decreased in FPS xenografts injected with CXCL1 neutralizing antibody compared with those treated with nonimmune IgG (P < 0.001). This analysis was confirmed further by immunohistochemistry ( Fig. 5D), where increased neutrophils were seen distributed throughout FPS xenografts compared with WT or CXCL immunoneutralized FPS xenografts.
The link between inflammation and tumor progression has been shown in a range of studies. For example, elevated expression of inflammatory COX-2 and prostaglandins has been correlated with tumor growth and angiogenesis in prostate, pancreatic, and colon cancer ( 31– 33), and the risk of long-term inflammation has been shown by studies showing that continued use of specific COX-2 inhibitors nonsteroidal anti-inflammatory drugs can significantly reduce cancer occurrence in patients at high risk ( 34). In the present study, we show that PGF2α-FP signaling can regulate expression of the inflammatory chemokine CXCL1 in endometrial adenocarcinoma cells to modulate neutrophil influx in tumors. To our knowledge, this is the first study to provide a link between inflammatory prostanoids, specifically PGF2α, and neutrophil recruitment in endometrial cancers.
Prostaglandins have been shown to regulate chemokine expression in vitro ( 35, 36). Prostaglandin E2 is overexpressed in many cancer types and has been shown to induce CXCL1 production in colon cancer cells, which can then promote tube formation and migration of endothelial cells ( 14). We have previously ascertained a role for the FP receptor and PGF2α signaling in regulating endometrial adenocarcinoma ( 8– 11). In the present study, we investigated a role for the FP receptor in modulating the expression of chemokines using an in vitro model system of Ishikawa cells stably expressing the human FP receptor (FPS cells; ref. 9) and a human cytokine antibody array. The array identified CXCL1 as a key cytokine induced by PGF2α-FP signaling. Using FPS cells, which we have previously shown to reproduce the ex vivo effects of PGF2α on endometrial adenocarcinoma tissue explants ( 9), we elaborated the signaling pathways mediating the role of FP on CXCL1 expression using chemical inhibitors and DN mutants of cell signaling pathways. A key effector pathway that has been previously shown to regulate tumorigenic signaling molecules in response to G-protein–coupled receptor signaling is the mitogen-activated protein kinase (MAPK) pathway. The signaling components of this pathway in FPS cells have been identified in our laboratory, where the phosphorylation of the downstream component of the MAPK pathway, ERK1/2, was shown to be mediated by EGFR trans-activation and c-src phosphorylation ( 9). We found that chemical inhibitors of EGFR and MEK could inhibit CXCL1 production, as did cotransfection of DN EGFR, Ras, and MEK. However, NFAT, a common regulator of cytokine expression ( 37), was not involved in PGF2α-mediated CXCL1 production in this cell type. These data are supported by previously published evidence in colorectal adenocarcinoma cell lines where the ERK pathway was also shown to be crucial in the regulation of CXCL1 expression after stimulation with prostaglandin E2 ( 14).
Overexpression of CXCL1 has previously been shown in a variety of tumor types, including colorectal ( 18) and melanoma ( 15), and promotes a variety of cellular functions including cell proliferation in esophageal cancer ( 38) and cell invasion in bladder cancer ( 39). Here, we showed elevated expression of CXCL1 and its receptor CXCR2 in endometrial adenocarcinoma compared with normal endometrium. Expression of both was localized to glandular epithelium, stroma, and vascular endothelial cells. In addition, treatment of endometrial adenocarcinoma explants with PGF2α caused an increase in CXCL1 expression via FP receptor and ERK1/2 signaling pathways confirming the importance of this signaling cascade in regulating CXCL1 expression ex vivo. CXCR2 localization in neutrophils in endometrial adenocarcinoma suggested that CXCL1 via CXCR2 could play a role in immune cell function. A role for CXCL1 in neutrophil influx has been previously shown in an angiogenic sponge model in the mouse, as endogenous CXCL1 expression increased immediately preceding a neutrophil influx ( 40).
Here, we show by immunohistochemistry that neutrophils are elevated in human endometrial adenocarcinomas. We have also confirmed that CXCL1 is strongly chemotactic to neutrophils as conditioned media from PGF2α-stimulated FPS cells induced chemotaxis of peripheral neutrophils. This chemotaxis was significantly reduced by CXCL1 immunoneutralization and CXCR2 inhibition using a specific antagonist. To determine a role for CXCL1 induced by PGF2α-FP interaction in vivo, we inoculated nude mice with FPS and WT cells. The increased neutrophils in the resulting FPS tumors compared with WT were significantly reduced by injection of CXCL1 neutralizing antibodies, demonstrating that PGF2α signaling via CXCL1 is influencing neutrophil cell infiltrate in endometrial adenocarcinomas. Neutrophil infiltration into tumors has also been shown to be dependent on CXC chemokine-CXCR2 signaling in a model of melanoma in a CXCR2 null nude mouse ( 41).
A chemokine-mediated influx of neutrophils is seen in the late secretory phase of the normal endometrium ( 42). Their role may be dependent on the activating agents and cytokines present, but they are thought to be involved in the breakdown and repair at menstruation by degranulation and the release of proteases that degrade the extracellular matrix ( 43). They may also be capable of remodeling vasculature, as neutrophils found close to or associated with endothelial microvessels express VEGF during or coincident with angiogenesis in the normal menstrual cycle ( 44).
The role of neutrophils in endometrial adenocarcinoma is unclear, and in our study, similar to other reports ( 45, 46), neutrophil influx in our xenograft model did not effect on tumor size. However, considering their profound tissue-remodeling capabilities, which have been shown in a number of animal models of other cancer types, it is possible that they play a similar role in endometrial cancer. For example, neutrophils have been shown to uniquely produce a tissue inhibitor of metalloproteinase–free MMP-9, a key protease involved in extracellular matrix degradation, which may affect the tumor microenvironment by tissue remodeling ( 24). In addition, the depletion of neutrophils in a mouse model was shown to prevent metastasis of fibrosarcoma cells from the primary tumor ( 45), suggesting a role for neutrophils in the switch to a metastatic phenotype. In a nude mouse model of breast cancer, overexpression of interleukin-8, a chemokine related to CXCL1, caused an infiltration of neutrophils, which increased invasiveness of the tumor, likely due to an increase in protease production ( 47). Similarly, the decrease in neutrophil infiltration caused by an inhibition of CXCL1 expression in a nude mouse model of colon cancer significantly decreased metastasis in these animals ( 48). Furthermore, neutrophils may also promote tumorigenesis through means other than tissue remodeling. In an in vitro model of colon cancer, neutrophils promote cellular stress by inducing transient errors in DNA replication in epithelial cells ( 49), which could ultimately lead to carcinogenesis, whereas neutrophils from ovarian cancer patients released higher levels of reactive oxygen species, which could potentially lead to cellular changes that support tumor progression ( 50).
In conclusion, we provide evidence for a novel PGF2α-FP pathway that can regulate the inflammatory microenvironment in endometrial adenocarcinoma via CXCL1-induced neutrophil chemotaxis.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Medical Research Council core funding (H.N. Jabbour; U.1276.00.004.00002.01).
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.
We thank Anne Saunderson for patient recruitment and tissue collection, and Professor S. Howie, Dr. S. Battersby, S. Wright, V. Grant, and M. Keightley for helpful discussion and technical assistance.
- Received February 2, 2009.
- Revision received May 1, 2009.
- Accepted May 8, 2009.
- ©2009 American Association for Cancer Research.