Nature Cell Biology
- 8, 1369 - 1375 (2006)
Published online: 26 November 2006; | doi:10.1038/ncb1507
Tumour-mediated upregulation of chemoattractants and recruitment of myeloid cells predetermines lung metastasisSachie Hiratsuka1, Akira Watanabe2, Hiroyuki Aburatani2 & Yoshiro Maru11 Department of Pharmacology, Tokyo Women's Medical University School of Medicine, Shinjyuku-ku, Tokyo 162–8666, Japan. 2 Genome Science Division, Research Center for Advanced Science and Technology, University of Tokyo, Meguro-ku, Tokyo 153–8904, Japan.
Correspondence should be addressed to Yoshiro Maru ymaru@research.twmu.ac.jp Primary tumours influence the environment in the lungs before metastasis1,
2. However, the mechanism of metastasis is not well understood. Here, we show that the inflammatory chemoattractants S100A8 and S100A9, whose expression is induced by distant primary tumours, attract Mac 1 (macrophage antigen 1)+-myeloid cells in the premetastatic lung. In addition, tumour cells use this mechanism, through activation of the mitogen-activated protein kinase (MAPK) p38, to acquire migration activity with pseudopodia for invasion (invadopodia). The expression of S100A8 and S100A9 was eliminated in lung Mac 1+-myeloid cells and endothelial cells deprived of soluble factors, such as vascular endothelial growth factor A (VEGF-A), tumour necrosis factor  (TNF) and transforming growth factor  (TGF) both in vitro and in vivo. Neutralizing anti-S100A8 and anti-S100A9 antibodies blocked the morphological changes and migration of tumour cells and Mac 1+-myeloid cells. Thus, the S100A8 and S100A9 pathway may be common to both myeloid cell recruitment and tumour-cell invasion.
During metastasis, cancer cells shed from a primary tumour can travel through blood vessels, migrate to a secondary site, invade the tissue and form metastatic nodules3,
4. Tumour-associated cells, such as macrophages and haematopoietic bone-marrow progenitors enhance metastasis because they prepare the environment for invasion in potential premetastatic organs1,
2. Previously, we showed that matrix metalloproteinase 9 (MMP9), expressed in lung macrophages and endothelial cells in response to primary tumours, promoted the invasion of lung tissues by tumour cells and this induction, in a premetastatic phase, was dependent on VEGF-A secreted from the primary tumours2. In addition, VEGF receptor 1 (VEGFR1)-positive cells may regulate the homing of tumour cells1. However, it is unclear which molecule in the premetastatic tissues, into which tumour-associated cells infiltrate, induces the migration of tumour cells. Here, we hypothesize that the lungs are sensitized by primary tumours at an early stage and searched for molecules that are induced by primary tumours and directly regulate the migration of tumour cells, as well as cells associated with them.
To elucidate the host response to a primary tumour in the premetastatic phase, it is necessary to separate the metastatic time course into two states: premetastatic and metastatic. As many primary tumours may affect the lungs before metastasis1,
2, we focused on the lungs before metastasis actually took place, as judged by examining histological sections. We further checked for micrometastasis using two different methods: we detected fluorescence-stained tumour cells in primary tumour tissues, but not in lungs for up to 1.5 months; and no PCR products for the vector sequence inserted into tumour cells were found in the lungs (see Supplementary Information, Fig. S1a–c). Using this model, we could mimic the premetastatic lung responding to a primary tumour. Occasionally, we injected tumour cells intravenously into tumour-bearing mice, before metastasis to the lungs from primary tumours, to examine the effect of the change in the lungs caused by the primary tumour. This model represented the metastatic phase2 (Fig. 1a and see Methods).
 | | |
To search for genes related to the premetastatic phase, the gene-expression profile was examined in lungs derived from normal, benign tumour (F2 cells; mouse angioma5)-bearing and malignant tumour (Lewis lung carcinoma (LLC), B16 melanoma and highly metastic LLC (3LL))-bearing nude mice. RNAs extracted from normal or tumour-bearing lungs were used to prepare complementary RNAs (cRNAs), which were each hybridized to oligonucleotide microarrays to determine candidates for differentially expressed genes (see Supplementary Information, Fig. S1d).
We focused on S100A8 and S100A9 for two reasons: first, these genes encode secreted proteins, which act as strong chemoattractants for monocytes and neutrophils6,
7,
8. The expression of S100A8 and S100A9 in LLC cells and B16 cells was not detectable. Second, the induction of S100A8 and S100A9 expression was strong (see Supplementary Information, Fig. S1d). Both genes were clearly detectable at the protein level in the lungs of tumour-bearing mice (Fig. 1b).
To specify the cells responding to the primary tumour, endothelial cells, Mac 1+-myeloid cells and the remaining cells from the lungs of normal and tumour-bearing mice were sorted with anti-VEGFR2 and anti-Mac1 antibodies, respectively (see Methods), and the expression levels of S100A8 and S100A9 were examined. The expression of both genes was increased in the Mac 1+-myeloid cells and endothelial cells in the LLC-bearing mice compared with normal mice (Fig. 1c). Moreover, the primary tumours induced recruitment of Mac 1+-myeloid cells in the lungs (Fig. 1d). To confirm the location of the S100A8 and S100A9 proteins, lung tissues were stained with anti-S100A8 (Fig. 1e) and anti-S100A9 antibodies (data not shown). Positive signals were detected in both VE–cadherin+(an endothelial cell adhesion molecule)-endothelial cells and Mac 1+-myeloid cells.
When normal lung tissue was cultured with serum from tumour-bearing mice, expression of S100A8 and S100A9 was induced, and was then eliminated by heat inactivation of the serum (Fig. 2a). We next looked for possible mediators in the serum that upregulate S100A8 and S100A9 expression in the lungs. When tumour-bearing mouse serum was deprived of TNF , TGF or VEGF-A with a neutralizing antibody for each protein, the upregulation of both genes by the serum in vitro was partially disturbed (Fig. 2a). Similar effects were observed in vivo (Fig. 2b). Furthermore, the addition of these factors stimulated the expression of S100A8 and S100A9 in a lung organ culture (see Supplementary Information, Fig. S2a). Next, 3  g each of VEGF-A, TNF or TGF, or a combination of VEGF-A and TNF was injected once per day for two days before the intravenous injection of rhodamin-labelled LLC cells. VEGF-A or TNF alone supported the recruitment of the LLC cells. Remarkably, a combination of VEGF-A and TNF further stimulated the colonizing activity of the LLC cells (see Supplementary Information, Fig. S2b). Thus, multiple factors including TNF, VEGF-A and TGF, possibly secreted from the primary tumour, could at least partially induce the expression of S100A8 and S100A9 in premetastatic lungs.
| | |
The role of S100A8 and S100A9 in tumours has not been well established. We examined whether these proteins could induce the migration of LLC and B16 cells as both are strong chemoattractants for inflammatory cells6,
7,
8,
9. These proteins triggered the migration of alveolar macrophages (Fig. 3a) and similar results were obtained with peritoneal macrophages (data not shown). Both LLC cells and B16 cells showed maximal migration with S100A8 at 100 pg ml-1 and S100A9 at 1 ng ml-1 (Fig. 3b,c).
| | Figure 3. S100A8 and S100A9 induce migration activities of macrophages and tumour cells. | | |  | (a) S100A8 and S100A9 elicit the migration of alveolar macrophages in vitro. S100A8–GST and S100A9–GST but not GST as a negative control, acted as chemoattractants for macrophages in the chemotaxis Boyden chamber (lane 1, 0.1; lane 2, 1; lane 3, 10; lane 4, 1  102, lane 5, 1 103; lane 6, 1 104 pg ml-1). Means  s.d. are shown (n = 10, asterisk indicates P <0.01). (b, c) S100A8 and S100A9 induce migration of LLC cells (b) and B16 cells (c) in the Boyden chamber (lane 1, 1; lane 2, 10; lane 3, 1 102; lane 4, 1 103; lane 5, 1 104; lane 6, 1 105 pg ml-1). Means s.d. are shown (n = 10, asterisk indicates P <0.01). (d) Morphological changes in LLC cells cultured with conditioned medium from lungs stimulated with S100A8–GST or S100A9–GST for 1.5 h. The scale bar represents 10 m. (e) The percentage of LLC cells migrating in response to LCM stimulated by S100A8–GST, S100A9–GST or GST as a control (100%) over 1.5, 6 or 22 h (n = 5). Means s.d. are shown.
 Full Figure and legend (41K) |
| |
We next examined whether S100A8 and S100A9 cause the migration directly, or by inducing the secretion of secondary molecules. Lung tissue was cultured with S100A8–GST, S100A9–GST or GST alone for 30 min and the conditioned medium was collected. When tumour cells were incubated in the presence of S100A8- or S100A9-stimulated lung conditioned medium (S100-stimulated LCM), they changed shape from being small and round to unusually spiky with invadopodia (Fig. 3d). LCM stimulated with either of the factors for 1.5 h showed maximum induction of LLC-cell migration (Fig. 3e). The number of LLC cells that migrated in response to S100A8-stimulated LCM was fourfold larger than that with S100A8 alone (data not shown).
To examine whether Mac 1+-myeloid cells and endothelial cells secrete these migration-stimulating factors (MSFs), these cells were purified from normal or tumour-bearing mice and the S100A8-stimulated LCM was collected (Fig. 4). The combined LCM derived from endothelial cells and Mac 1+-myeloid cells of normal mice exposed to S100A8 for 30 min stimulated the migration of tumour cells (Fig. 4c, lane 10). In the same assay, the effect was more prominent in the cells derived from tumour-bearing mice (Fig. 4c, lanes 4, 8 and 12). This suggests that S100A8 and S100A9 induce the secretion of MSFs from lung Mac 1+-myeloid cells and endothelial cells.
| | |
Next, we attempted to identify the MSFs. Several molecules, such as VEGF-A, TGF and TNF have been reported as being chemoattractants for macrophages and tumour cells10,
11,
12. Chemokine macrophage inflammatory protein-2 (MIP2) transcriptionally regulates the expression of G protein-coupled receptor kinase 2 (grk2) in migrating polymorphonuclear leukocytes (PMNs)13. MIP2 was also a candidate MSF given that S100A8-stimulated LCM induced grk2 expression in tumour cells (data not shown). We examined whether any of these molecules were MSFs. When S100A8-stimulated LCM were deprived of TNF, MIP2, TGF or VEGF-A using a neutralizing antibody for each, the conditioned medium caused 47%, 29%, 19% and 5% reduction, respectively, in the migration of LLC cells toward IgG-treated S100A8-LCM (data not shown). These results suggest that S100A8-stimulated MSFs include TNF, MIP2 and TGF, at least.
It is well established that p38 is involved in inflammation, apoptosis, cardiomyocyte hypertrophy and cell differentiation. The activity of p38 is stimulated by many growth factors, cytokines and chemotactic substances, such as VEGF-A, fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), TNF, interleukins, lipopolysaccharide (LPS) and formyl-methionyl-leucyl-phenylalanine (fMLP)14. Recently, several studies have suggested that p38 is also involved in the migration of diverse cell types. SB203580 and SB202190, inhibitors of p38, inhibit chemoattractant-induced migration of various cells14. We examined whether S100A8 could stimulate the activation of p38 in peritoneal macrophages and LLC cells. Both cells showed a similar time course of phosphorylation of p38 responding to S100A8- and S100A8-LCM (Fig. 5a). SB203580 suppressed the phosphorylation of p38 that was induced by S100A8–GST in LLC cells (Fig. 5a) and blockage of the p38 signalling cascade using SB203580 suppressed the migration of these cells (Fig. 5b). Thus, the inhibition of p38 activation is a cause of the reduced migration.
| | Figure 5. Neutralizing anti-S100A8 and anti-S100A9 antibodies block the migration of macrophages and tumour cells to lungs. | | |  | (a) Phosphorylation of p38 MAP kinase stimulated by S100A8–GST, GST, S100A8–GST–LCM or GST–LCM in macrophages (M ) and LLC cells. Cells were treated for 5–30 min and analysed by western blotting. SB203580 (SB) or anti-S100A8 antibodies suppressed phosphorylation of p38 stimulated by S100A8–GST. The intensity of the bands in phospho-p38 was normalized to those in p38 and the relative values are shown under panels. (b) The inhibition of p38 signalling suppressed migration of LLC cells in response to S100A8. S100A8–GST (100 pg ml-1), GST (100 pg ml-1), SB203580 (2 M) in DMSO, or an equal volume of DMSO (D, control) was added to lower wells (n = 5). Means s.d. are shown. (c) The numbers of Mac 1+-myeloid cells in premetastatic lungs from LLC-bearing or B16-bearing mice, which were treated with control IgG, anti-S100A8 antibody alone (anti-A8), or both anti-S100A8 and anti-S100A9 (anti-A9) antibodies. Means s.d. are shown (n = 5, asterisk indicates P <0.05). (d) The numbers of rhodamine-labelled tumour cells in lungs from normal (N), LLC-bearing or B16-bearing mice, which were treated with control IgG, anti-S100A8 antibody alone, or both anti-S100A8 and anti-S100A9 antibodies before the injection of these cells. Detection was carried out at 5 h and 24 h after the injection of tumour cells into the tail vein. Means s.d. are shown (n = 10, asterisk indicates P <0.05).
Full Figure and legend (41K) |
| |
We next attempted to generate antibodies to neutralize the functions of S100A8 and S100A9 and characterized the effect of candidate antibodies using a migration assay. Of eight independent antibodies against S100A8 and S100A9, one antibody for each protein (A81 for S100A8 and A91 for S100A9) blocked the migration of LLC cells to the proteins (see Supplementary Information, Fig. S2c, d). Furthermore, the anti-S100A8 antibody suppressed the S100A8-stimulated phosphorylation of p38 (Fig. 5a), suggesting that the reduction in migration caused by the anti-S100A8 antibody is likely to be mediated by the phosphorylation of p38. In addition, both antibodies suppressed S100A8- or S100A9-induced secretion of MSFs, as the migration of tumour cells to LCMs containing both proteins was inhibited (see Supplementary Information, Fig. S2e, f).
Next, we next determined the potency of these antibodies in vivo. First, to establish a direct relationship between lung metastasis and the expression of S100A8 and S100A9, we examined macrometastasis to the lungs with highly metastatic 3LL tumours. A combination of the anti-S100A8 and anti-S100A9 antibodies blocked spontaneous metastasis by 3LL cells (see Supplementary Information, Fig. S3a). Second, to focus on the migration step, we examined whether tumour-bearing mouse lungs pretreated with these neutralizing antibodies are protected from circulating tumour cells. A81 and/or A91 antibodies were intravenously administered to tumour-bearing mice once daily for two days before injecting tumour cells into the tail vein. The number of Mac 1+-myeloid cells remarkably decreased in antibody-treated lungs derived from tumour-bearing mice in the premetastatic phase (Fig. 5c). An 82% and 93% reduction in the colonization of the lung by tumour cells was observed at 5 h and 24 h, respectively, after the injection of tumour cells into anti-S100A8 antibody-treated tumour-bearing mice (Fig. 5d). This demonstrates that the blockade of S100A8 and S100A9 expression and infiltration by Mac 1+-myeloid cells into the lung at the premetastatic stage can inhibit the migration of primary tumour cells.
The application of both antibodies had a remarkable effect on the recruitment of tumour cells responding to S100A8 and S100A9 in the lungs — which was induced by primary tumours as the expression of S100A8 and S100A9 is stronger in the lungs than other organs (such as the liver and kidney) at the protein level (see Supplementary Information, Fig. S3b, c). In addition, as the sites harbouring tumours were scattered widely, rather than concentrated in a specific area, consistent with the expression pattern of S100A8 and S100A9, the antibodies seemed to target endothelial cells and Mac 1+-myeloid cells at non-specific sites in primary tumour-stimulated lungs.
Many monocyte or macrophage-specific chemokines, including MCP-1, stromal cell-derived factor (SDF-1) and MIP1, are also expressed in response to TNF or TGF15,
16,
17 We suspect that the production of these factors may be induced by the primary tumours: in fact, production of MIP1 in premetastatic lungs was stimulated by LLC or 3LL but not by B16 in a microarray analysis (see Supplementary Information, Fig. S4a). We confirmed this result using real-time PCR analysis (see Supplementary Information, Fig. S4b). To determine whether inducing the expression of MIP1 is effective in the recruitment of circulating tumour cells to the lungs, we intravenously injected LLC cells into LLC-bearing mice treated with anti-MIP1 neutralizing antibodies (see Supplementary Information, Methods). Inhibition of MIP1 in tumour-bearing lungs blocked 20% of tumour-cell colonization (see Supplementary Information, Fig. S4c). Potential secretion of TNF, VEGF-A and TGF from primary tumours could induce S100A8 and S100A9, which in turn could stimulate the secretion of TNF and MIP2 and Mac 1+-myeloid-cell migration into the lungs. These results suggest that S100A8 and S100A9 are key molecules for the enrichment of inflammatory chemoattractants capable of causing inflammatory cell infiltration, and S100A8 and S100A9 and chemokines may synergistically regulate this cycle.
S100A8 and S100A9 are strong chemotactic agonists for inflammatory cells and we showed that both attract Mac 1+-myeloid cells through the p38 cascade in vitro. Furthermore, the lung Mac 1+-myeloid cells that increased in number produced both proteins in response to primary tumours in the premetastatic phase. Tumour cells mimic Mac 1+-myeloid cells in that abundance of S100A8 and S100A9 in the lungs induces the migration of Mac 1+-myeloid cells by signalling through p38. Moreover, LCM stimulated by S100A8 attracted tumour cells in vitro fourfold more efficiently than S100A8 alone because of the strong activation of p38 (see Fig. 5a). Thus, Mac 1+-myeloid cells and tumour cells use a common pathway for recruitment to the lungs (see Supplementary Information, Fig. S4d).
Recently, a link between chronic inflammation and cancer has been established, but the precise mechanism involved has not been uncovered18. The immune system sustains inflammation and can occasionally inhibit tumour development. Conversely, chronic inflammation positively contributes to the development of cancer19.
Metastasis is thought to be a multistep process requiring the concerted actions of multiple genes3,
4. Specific genes allow tumour cells to shed from the primary tumour and these cells travel through the blood vessels to secondary sites, attach to endothelial cells, invade through proteolysis of the extracellular matrix and finally cause macrometastases at distant organs4. Currently, a focus on primary tumour cells and the surrounding environment is a major approach to therapy. As specific primary tumours are composed largely of cells with a genetic make-up that compels them to metastasize, copious numbers of target molecules (such as the small GTPase, Rho C and the transcription factor,Twist) have been identified using gene expression profiling20,
21,
22. Anti-angiogenic drugs block the growth of new blood vessels at metastatic sites, as well as in primary tumours23,
24.
Our results suggest that stimulation of metastasis by the host in the premetastatic phase is an attractive target for the prevention of metastasis. Even if the tumour cells enter into the circulation, one can reduce the metastatic risk by blocking the expression of guide molecules, such as S100A8 and S100A9, in the target organs. Drugs targeting metastatic tumours themselves show only a limited effect in the prevention of metastasis. We propose that suppression of the host response triggered by the primary tumour at early clinical stages or, hopefully, in the premetastatic phase, can prevent the spread of tumours.
Methods
Animals.
Nude mice and C57BL/6 mice were purchased from CLEA (Tokyo, Japan). VEGFR-1TK homozygous null animals, whose genetic background is 50% that of 129 mice and 50% that of C57BL/6 mice, have been previously described11.
Antibodies.
Neutralizing antibodies for mVEGF-A, mMIP1, mMIP2, mTNF, mTGF, rat IgG2A, rabbit IgG and goat IgG were purchased from R&D (Minneapolis, MN). To confirm the neutralizing effect of the antibodies, an anti-MIP2 antibody and an anti-TNF antibody were purchased from Santa Cruz (Santa Cruz, CA) and Genzyme (Cambridge, MA), respectively.
Tumour-cell labelling and in vivo metastasis assay.
LLC and B16 cells were double-marked by infection with a pSRMSVtk-neo retroviral vector and by labelling with a PKH26 fluorescent staining kit (Zynaxis, Malvern, PA). For the metastasis assay, 1 105 labelled B16 or LLC cells were intravenously injected 10 days after subcutaneous implantation of 1 107 tumour cells into the back. At the times indicated, lungs were collected after perfusion with PBS to eliminate circulating tumour cells. PCR analysis of the retroviral vector sequence and visualization of tumour cells were performed to examine for non-spontaneous metastasis to the lungs.
Semi-quantitative RT–PCR.
Total RNA was isolated with Trizol (Invitrogen, Carlsbad, CA), and cDNA was prepared for hybridization. Primers used for RT–PCR were as follows: 5'-TGAGCAACCTCATTGATGTCTACC-3' and 5'-ATGCCACACCCACTTTTATCACC-3' for S100A8 and 5'-GAAGAAAGAGAAGAGAAATGAAGCC-3' and 5'- CTTTGCCATCAGCATCATACACTCC-3' for S100A9. All the samples were run in triplicate, and the results were averaged.
Proteins and antibodies.
GST-fused constructs comprising the mouse S100A8 and S100A9 proteins were generated in pGEX4T-1 (Amersham, Piscataway, NJ). The proteins were expressed in Eshcerichia coli BL21 and purified on a glutathione–Sepharose column. Polyclonal antibodies were raised in rabbits using GST-fused proteins, and were purified using a protein A–Sepharose column.
Isolation of endothelial cells and Mac 1+-myeloid cells.
Lung endothelial cells were collected using a modification of a previously described method25. Briefly, minced mouse lungs were digested in collagenase at 37 °C for 90 min and then filtered through a sterile 58 m nylon mesh. Washed cells were separated into a layer of endothelial and myeloid cells using Histpaque (Sigma, St Louis, MO). Collected cells were incubated with an anti-mouse VEGFR2 rat antibody (Pharmigen, San Diego, CA) or an anti-Mac 1 antibody (Pharmingen), respectively, followed by Dynabeads (Dynal, Oslo, Norway) coated with anti rat-IgG antibody.
Culture.
For organ cultures, 2 mm2 specimens of tissue were cultured in serum-free or 2% FCS–DMEM containing mouse S100A8–GST, S100A9–GST or GST alone for 30–90 min. For coculture, 1 103 lung endothelial cells were cultured on 1 103 Mac 1+-myeloid cells derived from non-tumour-bearing or tumour-bearing mice, in 2% FCS–DMEM containing S100A8–GST or GST. Alveolar macrophages were collected intratracheally with 1% BSA–PBS from anesthetized mice. All experiments were carried out in the presence of polymyxin B (10 g ml-1).
Immunohistochemistry.
Mouse lung tissue sections were immunohistochemically stained with an anti-S100A8 or anti-S100A9 antibodies, as well as antibodies against mouse VE–cadherin (Pharmingen) as an endothelial cell-specific marker or Mac1/CD11b (Serotec, Oxford, UK) as a monocyte/macrophage marker. The fluorescence-conjugated secondary antibodies were used for double staining.
In vitro migration assay.
The migration of tumour cells and macrophages was evaluated using a chemotaxis Boyden chamber (Neuroprobe, Gaithersburg, MD). The upper and lower wells were separated by a 5 m pore size polyvinylpyrrolidone-free polycarbonate filter (Nucleopore, Costar, Cambridge, MA) Chemoattractants or various tissue-conditioned media stimulated by S100A8–GST, S100A–GST or GST alone were applied to the lower wells. An aliquot (50 l) of the cell suspension (5 104 cells per well) was seeded in each of the upper wells and incubated for 3 h at 37 °C with 5% CO2.
Application of anti-S100A8 and anti-S100A9 antibodies in vivo.
The antibody (500 g) for S100A8 or S100A9 was injected intravenously into the tails of non-tumour bearing or tumour-bearing mice two days before the intravenous injection of rhodamin-labelled tumour cells. For the treatment of 3LL-bearing mice, normal IgG or antibodies (2 mg) against S100A8 and S100A9 were injected intravenously into the tail every two days for two weeks starting on the tenth day after tumour implantation.
Statistical analysis.
For statistical analysis, the data were expressed as means s.d. and compared using a Student's t-test. A P value <0.05 was considered significant.
Note: Supplementary Information is available on the Nature Cell Biology website.
Received 12 May 2006; Accepted 7 September 2006; Published online: 26 November 2006.
REFERENCES
- Kaplan, R. N. et al. VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature 438, 820–827 (2005). | Article | PubMed | ISI | ChemPort |
- Hiratsuka, S. et al. MMP9 induction by vascular endothelial growth factor receptor-1 is involved in lung-specific metastasis. Cancer Cell 2, 289–300 (2002). | Article | PubMed | ISI | ChemPort |
- Couzin, J. Medicine. Tracing the steps of metastasis, cancer's menacing ballet. Science 299, 1002–1006 (2003). | Article | PubMed | ChemPort |
- Kang, Y. & Massague, J. Epithelial–mesenchymal transitions: twist in development and metastasis. Cell 118, 277–279 (2004). | Article | PubMed | ISI | ChemPort |
- Toda, K. et al. Establishment and characterization of a tumorigenic murine vascular endothelial cell line (F-2). Cancer Res. 50, 5526–5530 (1990). | PubMed | ChemPort |
- Lackmann, M. et al. Identification of a chemotactic domain of the pro-inflammatory S100 protein CP-10. J. Immunol. 150, 2981–2991 (1993). | PubMed | ChemPort |
- Devery, J. M., King, N. J. & Geczy, C. L. Acute inflammatory activity of the S100 protein CP-10. Activation of neutrophils in vivo and in vitro. J. Immunol. 152, 1888–1897 (1994). | PubMed | ChemPort |
- Hobbs, J. A. et al. Myeloid cell function in MRP-14 (S100A9) null mice. Mol. Cell. Biol. 23, 2564–2576 (2003). | Article | PubMed | ChemPort |
- Roth, J., Vogl, T., Sunderkotter, C. & Sorg, C. Chemotactic activity of S100A8 and S100A9. J. Immunol. 171, 5651 (2003). | PubMed | ChemPort |
- Barleon, B. et al. Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor flt-1. Blood 87, 3336–3343 (1996). | PubMed | ISI | ChemPort |
- Hiratsuka, S., Minowa, O., Kuno, J., Noda, T. & Shibuya, M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice. Proc. Natl Acad. Sci. USA 95, 9349–9354 (1998). | Article | PubMed | ChemPort |
- Muraoka, R. S. et al. Blockade of TGF- inhibits mammary tumor cell viability, migration, and metastases. J. Clin. Invest. 109, 1551–1559 (2002). | Article | PubMed | ISI | ChemPort |
- Fan, J. & Malik, A. B. Toll-like receptor-4 (TLR4) signaling augments chemokine-induced neutrophil migration by modulating cell surface expression of chemokine receptors. Nature Med. 9, 315–321 (2003). | Article |
- Huang, C., Jacobson, K. & Schaller, M. D. MAP kinases and cell migration. J. Cell Sci. 117, 4619–4628 (2004). | Article | PubMed | ISI | ChemPort |
- Watanabe, T. et al. Monocyte chemotactic protein-1 regulates leukocyte recruitment during gastric ulcer recurrence induced by tumor necrosis factor-. Am. J. Physiol. Gastrointest. Liver Physiol. 287, 919–928 (2004). | Article |
- Chen, S. et al. Transforming growth factor-1 increases CXCR4 expression, stromal-derived factor-1-stimulated signalling and human immunodeficiency virus-1 entry in human monocyte-derived macrophages. Immunology 114, 565–574 (2005). | Article | PubMed | ChemPort |
- Ohira, S. et al. Possible regulation of migration of intrahepatic cholangiocarcinoma cells by interaction of CXCR4 expressed in carcinoma cells with tumor necrosis factor- and stromal-derived factor-1 released in stroma. Am. J. Pathol. 168, 1155–1168 (2006). | PubMed | ChemPort |
- Marx, J. Cancer research. Inflammation and cancer: the link grows stronger. Science 306, 966–968 (2004). | Article | PubMed | ISI | ChemPort |
- Greten, F. R. et al. IKK links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell 118, 285–296 (2004). | Article | PubMed | ISI | ChemPort |
- Clark, E. A., Golub, T. R., Lander, E. S. & Hynes, R. O. Genomic analysis of metastasis reveals an essential role for RhoC. Nature 406, 532–535 (2000). | Article | PubMed | ISI | ChemPort |
- Kang, Y. et al. A multigenic program mediating breast cancer metastasis to bone. Cancer Cell 3, 537–549 (2003). | Article | PubMed | ISI | ChemPort |
- Yang, J. et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 117, 927–939 (2004). | Article | PubMed | ISI | ChemPort |
- Folkman, J. The role of angiogenesis in tumor growth. Semin. Cancer Biol. 3, 65–71 (1992). | PubMed | ChemPort |
- Hanahan, D. & Folkman, J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86, 353–364 (1996). | Article | PubMed | ISI | ChemPort |
- Dong, Q. G. et al. A general strategy for isolation of endothelial cells from murine tissues. Characterization of two endothelial cell lines from the murine lung and subcutaneous sponge implants. Arterioscler. Thromb. Vasc. Biol. 17, 1599–1604 (1997). | PubMed | ISI | ChemPort |
Acknowledgements
We thank M. Shibuya and T. Noda for preparation of the VEGFR1–TK-/- mouse and H. Meguro and S. Yamamoto for GeneChip analysis and data processing. We are also grateful to O. N. Witte for critical reading of and comments on the manuscript. This study was partly supported by Grants-in-Aid for Scientific Research from the Japanese government (No. 12147210) from the Ministry of Education, Culture, Sports, Science and Technology and the Program for Promotion of Fundamental Studies in Health Sciences of the National Sciences of the National Insititute of Biomedical innovation to Y.M., and from the NFAT project of New Energy and Industrial Technology Development Organization to H.A.
Competing interests statement:
The authors declare that they have no competing financial interests. |
