Introduction
Human esophageal carcinoma, one of the most common causes of cancer death worldwide, occurs at a very high frequency in China [
1,
2]. Esophageal carcinomas often have poor prognosis due to early lymph node metastasis and invasion of neighboring organs such as the aorta, trachea, bronchus, pericardium and lung [
2]. Therefore, disrupting the aggressive metastatic phenotype is essential for developing an effective treatment for esophageal cancer. Although several molecules have been reported to contribute to the ability of esophageal carcinoma cells to metastasize and invade normal tissue, such as N-cadherin [
3], TSLC1 [
4] and MTA1 [
5], the underlying mechanism remains obscure. Considering the complexity of tumor invasion and metastasis, various experimental approaches have been developed to systematically identify genes that are involved in the process. By directly comparing the differentially expressed genes between liver metastatic and primary tumor tissues,
POSTN, encoding the periostin protein, was identified as a gene associated with colon cancer and liver metastasis [
6]. Cancer metastasis is thought to originate from a small proportion of cancerous cells in primary tumors. Therefore, screening for a subpopulation of cells with high metastatic potential from a parent tumor cell line in experimental models is a well-defined method for discovering genes that play roles in metastasis, especially that which preferentially occurs in specific organs. For example, microarray analysis of sublines of the MDA-MB-231 cell line with high lung or bone metastatic selection in nude mice led to identification of a set of genes that mark or mediates breast cancer metastasis in these tissues [
7‐
9].
In order to derive a subpopulation of cells with high metastatic potential from tumor cell lines, we have established a model system to inspect genes involved in different steps of metastasis including invasion, survival and arrest. We screened for and selected an esophageal tumor cell subline with high invasive potential and analyzed genes which may correspond to this phenotype by gene microarray. Through this analysis, we identified sphingosine kinase 1 (SPHK1) as one such gene that participates in esophageal carcinoma invasion and metastasis.
SPHK1 is a conserved lipid kinase that catalyzes formation of important regulators of inter- and intracellular signaling. It is a ubiquitously expressed, evolutionary conserved enzyme that catalyzes phosphorylation of sphingosine (Sph) and dihydrosphingosine (dhSph) to sphingosine 1-phosphate (S1P) and dhS1P, respectively. SPHK1 is transiently activated in response to a large variety of agonists and has been shown to contribute to signaling cascades elicited by TNF-α [
10], VEGF and 17β-estradiol [
11]. Accordingly, SPHK1 can mediate biological effects of TNF-α such as induction of COX2 and generation of PGE2 in fibroblastic cells and MCP-1 in endothelial cells [
12].
Of interest is that
SPHK1 mRNA is frequently overexpressed in a variety of solid tumors, suggesting an important role for its encoded enzyme during tumorigenesis. Overexpression of
SPHK1 is thought to be oncogenic and renders transfected cells chemoresistant. Bonhoure
et al. [
13] reported that targeting
SPHK1 overcomes the multi-drug- resistant gene (MDR)-associated chemoresistance of HL60 cells. The oncogenic properties of
SPHK1 have been demonstrated both
in vitro and
in vivo in experimental models based on its overexpression in NIH3T3 cells [
14]. Overexpression of
SPHK1 facilitates anchorage-independent growth
in vitro and leads to tumor formation in SCID mice. Furthermore, more recent studies have demonstrated that phosphorylation of SPHK1 and subsequent membrane translocation are required for its pro-oncogenic function. Overexpression of
SPHK1 up-regulates
MMP1 mRNA and promoter activity as well as its protein levels, and this action by
SPHK1 requires activation of the ERK1/2-Ets1 and NF-kB pathways [
12]. In addition, high levels of the SPHK1 enzyme and its downstream product, S1P, correlate with poor survival of glioblastoma patients [
15,
16], and this unusual enhancement of neural cell invasion requires upregulation of the matricellular protein CCN1/Cyr61. These data suggest that enhanced SPHK1 activity in glioblastoma may drive the invasive potential of these cells.
Few studies by other investigators have specifically analyzed the effect of SPHK1 on invasiveness of cancer cells. Our report is the first to implicate the involvement of SPHK1 in the metastasis of esophageal cancer. We observed the overexpression of SPHK1 transcripts in esophageal carcinoma and identified downstream mediators that may mediate enhanced malignant behavior in these tumor cells. As several of these mediators may be useful as therapeutic targets of esophageal carcinoma, our findings will have implications for cancer drug development.
Materials and methods
Cell culture conditions
The esophageal carcinoma cell (ESCC) line EC9706 (a gift from Dr. Minrong Wang, Chinese Academy of Medical Sciences Cancer Institute Hospital, Peking University Medical School, Beijing, China) was cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS). Other ESCC lines (KYSE30, KYSE150, NEC, KYSE510, and KYSE2) were generously provided by Dr.Y. Shimada, Kyoto University.
Selection of invasive sublines from EC9706 cells
To select a highly invasive subpopulation, EC9706 cells were seeded on a Matrigel (Becton Dickinson, Franklin Lakes, NJ) coated, 8 μm-pore transwell (Costar, Cambridge, MA). 24 hours later, cells that had invaded to the other side of the transwell membrane were collected, expanded and then re-seeded into another Matrigel coated transwell. Such selection rounds for highly invasive cells were repeated four times, resulting in a subline from each generation designated as EC9706-P1, EC9706-P2, EC9706-P3 and EC9706-P4.
Quantitative RT-PCR
QRT-PCR was performed as my previous manuscript described [
17‐
20], briefly cells were harvested in Trizol
® reagent (Invitrogen, Carlsbad, CA), and total RNA was isolated according to the manufacturer's instructions. Single-stranded cDNA was synthesized from 4 μg total RNA using M-MLV reverse transcriptase (Invitrogen), with an oligo(dT) 18-mer as the primer, in a final reaction volume of 25 μl. Amplifications of specific transcripts from the cDNA were performed using the following primers:
SPHK1 sense 5'-ACAGTGGGCACCTTCTTTC-3', antisense 5'-CTTCTGCACCAGTGTAGAGGC -3';
SPHK2 sense
5'-GCACGGCGAGTTTGGTTC-3', antisense
5'-GAGACCTCATCCAGAGAGACTAG-3'; TM4SF3 sense 5'-CCAAACCCAGTATCTCA-3', antisense 5'- GACAAGCCTGTAACGAA-3';
Integrin α6 sense 5'- GCC TTG CAC GAT GAT ATG GAG-3', antisense 5'- GAT GAG CTG TCT GGAGAA-3';
Integrin β4 sense 5'- GCATCGTGGTCATGG AGAGCAG-3', antisense 5'- AATGTCCCTCGTGCACACAGC-3';
MET sense 5'- CATGCCGACAAGTGCAGTA-3', antisense 5'- TCTTGCCATCATTGTCCAAC-3';
GAPDH sense 5'- TGGTATCGTGGAAGGACTCATGAC-3', antisense 5'- ATGCCAGTGAGCTTCCCGTTCAGC-3';
ADAM12m sense 5'- GCACCTCCCTTCTGTGACAAGTTT-3', antisense 5'- CTTGGTGTGGATATTGTGGAGCAG-3';
CLDN1 sense 5'- GCGCGATATTTCTTCTTGCAGG-3', antisense 5'- TTCGTACCTGGCATTGACTGG-3';
IGFBP1 sense 5'- ATCTGATGGCCCCTTCTGAA-3', antisense 5'- AGCCTTCGAGCCATCATAGGTA-3';
A2M sense 5'- ACCAGGACGATGAAGACTGCAT-3', antisense 5'- CCACGAATCACAGAGTAAGGCA-3';
Integrin α3 sense 5'- AAGCCAAGTCTGAGACT-3', antisense 5'- GTAGTATTGGTCCCGAGTCT-3';
TFPI2 sense 5'- CGCCGGATCCTTTCTCGGAC-3', antisense 5'- GAATGTCTCGAGTTGCTTCTTCCGA-3';
LOXL2 sense 5'- GGCCGCCACGCGTG GATC-3', antisense 5'- CCCAAGGGTCAGGTAGCAGCCCC-3';
CYR61 sense 5'- AGCCCAACTGTAAACATCAG-3', antisense 5'- CATCCAGCGTAAGTAAACCT-3';
TIMP3 sense 5'- TTCGGTTACCCTGGCTACCA-3', antisense 5'- CTGCAG TAGCCG CCTTCT-3'.
Tissue samples and IHC
Samples were obtained from patients who attended Chinese Academy of Medical Sciences Cancer Hospital from January 2001 to December 2005. Ethical approval was provided by the Chinese Academy of Medical Sciences Cancer Hospital Ethics Committee. None of the patients received any neoadjuvant therapy prior to surgery. Prior patient consent and approval from the Institute Research Ethics Committee were obtained before we used these clinical materials for research purposes. Paired samples of fresh normal esophageal tissues and esophageal carcinoma tissues from the same patient were collected by the Department of Pathology in the Chinese Academy of Medical Sciences Cancer Hospital (Beijing, China). Each patient had received a pathologically and clinically confirmed diagnosis of esophageal squamous cell carcinoma. Primary tumor regions and the corresponding histologically normal esophageal mucosa were separated by experienced pathologists and immediately stored at -70°C until use. None of the patients received treatment before surgery.
The expression of SPHK1 in the esophageal tumors was determined by assessing its staining using tissue microarrays from 154 clinical cases, of which 124 of the esophageal cancer specimens had clinical follow-up records, these patients were followed up 8 years and 117 patients died at the end-points. In addition, 30 of these specimens had paired normal epithelia. For immunostaining of SPHK1, a DAKO CSA kit was used. The anti-SPHK1 antibody (ab16491, Abcam, 10 μg/ml) was incubated with the sectioned tissues for 1 h in citrate buffer. After staining, the slides were evaluated by two pathologists. SPHK1 expression was determined by scoring intensity and percentage of staining. Tissues with no staining were rated as 0; those with faint staining or moderate to strong staining in < 25% of cells as 1; with moderate staining or strong staining in 25-50% of cells as 2; and with strong staining in > 50% of cells as 3.
Immunofluorecence
Cells were fixed in 4% paraformaldehyde and then permeablized with 0.2% Triton X-100. Coverslips were blocked with 10% non-immune sheep serum. Primary antibodies were diluted in 1% bovine serum albumin/PBS and incubated overnight at 4°C. The antibody against SPHK1 (ABIN552700, poly rabbit antibody, 1:50) was obtained from
http://antibodies-online.com. The secondary antibody was Cy5-conjugated goat anti-rabbit IgG (Jackson Immunoresearch Laboratories Inc., West Grove, PA). Coverslips were mounted on glass sides with 0.5 μg/ml DAPI/fluorescence protection agent. Images were acquired with an Olympus confocal microscope.
Generation of SPHK1 overexpressing cells
The human
SPHK1 cDNA was cloned into the pcDNA4.0 expression vector (Invitrogen) containing a c-myc epitope at the C-terminus to give the pcDNA4.0-SPHK1-c-myc vector. The SPHK1 expression vector was transfected into EC9706 cells using Lipofectamine 2000 (Invitrogen), according to the manufacturer's recommendations. Cells carrying the recombinant SPHK1 or empty (control) vector were selected by culturing in the presence of 200 ng/ml Zeocin (Invitrogen) for more than 4 weeks. Two
SPHK1 stably transfected EC9706 cell clones (C17, C24) and one
SPHK1 negative clone (C7) were chosen for subsequent experiments. Transfection of duplex siRNAs was performed in the same manner. The
SPHK1 siRNA sequences as used before [
21] were 5'-GGGCAAGGCCUUGCAGCUCd(TT)-3' and 5'-GAGCUGCAAGGCCUUGCCCd(TT)-3', control siRNA 5'-UUCUCCGAACGUGUCACGUd(TT)-3'.
Invasion assays
Invasion assays were carried out similarly to the procedure for selecting invasive cell lines described above [
17]. Briefly, 1 × 10
4 cells were placed in a 24-well transwell unit on polycarbonate filter with 8 μm pores coated with Matrigel. After a 24-h incubation period, the cells that had passed through the filter into the lower wells were stained, counted and photographed. All experiments were performed in triplicate and repeated twice.
Western blot analysis
The proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA). Blots were blocked and then probed with antibodies against the tag of SPHK1 protein c-myc (9E10 1:3000, Sigma, St. Louis, MO) or directlly against the SPHK1 protein (1 μg/ml, ab16491, Abcam), β-actin (1:5000, Sigma, St. Louis, MO), phospho-EGFR (Y1068,1:1000, Invitrogen), EGFR(1:1000, Invitrogen). After washing, the blots were incubated with horseradish peroxidase-conjugated secondary antibodies and visualized using an enhanced chemiluminescence kit (Pierce, Rockford, IL).
Xenograft assays in nude mice
Female nu/nu mice, obtained from the Jackson Laboratory (Vitalriver, China), were kept in a specific pathogen-free facility at the Experimental Center of the Chinese Academy of Medical Sciences, Beijing, China. We are ensure animals used for scientific purposes are treated and cared for ethically and humanely. Female mice, aged 4-6 weeks, were used in these experiments. Parental EC9706, SPHK1 transfected EC9706 clones C17 and C24, or vector transfected cells were each subcutaneously injected into eight nude mice. Three months after the injection, the mice were sacrificed and examined for subcutaneous tumor growth and metastasis development. The tumor volumes were calculated according to the following formula: volume = length × width2/2. After the last treatment, the mice were sacrificed, and the tumor volumes and weights were measured. Xenografts were detected using the anti-SPHK1 antibody (10 μg/ml, ab16491, Abcam) with each section examined.
Statistical analysis
All data are presented as mean ± standard deviation (SD). Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) (SPSS Inc., Chicago, IL). Student's two-tailed t-test was used to compare the groups. P ≤ 0.05 was considered significant.
Discussion
Metastasis is a multi-factorial process, including tumor cells capable of escaping their normal microenvironment, traversing into and out of lymphatic or blood vessels and proliferating in new ''soil'' [
36]. Implicit in these stages, invasion is the critical ability for tumor cells to metastasize [
36]. During invasion, malignant cells reside on or within two major types of extracellular matrices, the basement membrane and the stromal matrix [
37]. The basement membrane is one of the most important barriers against cancer cell invasion [
37]. Therefore, for this study, we used Matrigel, a solubilized basement membrane preparation from the Engelbreth-Holm-Swarm (EHS) mouse sarcoma, as a model basement matrix to mimic esophageal carcinoma invasion
in vivo. Although EC9706, an esophageal squamous carcinoma cell line, can invade and form spontaneous lung metastasis nodules in
nu/nu mice, its metastatic potential is relatively low [
38]. The metastatic ability of EC9706 may arise from a few subclones with high metastatic potential among the parental cells. By screening with our
in vitro model, the subline EC9706-P4 with high invasion potential was established. This subline also exhibited high spontaneous metastatic potential
in vivo. Microarray analysis was used to determine which genes may be involved in invasion and metastasis. However, the microarray analysis of esophageal cancer tissues demonstrated that
SPHK1 was significantly overexpressed in these tumor tissues, and that this expression significantly correlated with tumor invasion, lymph node metastasis and clinical stage, indicating that SPHK1 is involved in esophageal carcinoma invasion and metastasis.
SPHK1 is up-regulated in many types of cancers and has been suggested as a potentially new therapeutic target. However, it is not yet known what signals the cancer cells use to apparently constitutively up-regulate expression of this enzyme, nor is it clear why it has such a profound role in tumorigenesis. A recent study examined the role of
SPHK1 in intestinal tumorigenesis in the Min mouse in which intestinal adenomas develop spontaneously [
39]. Deletion of the
SPHK1 gene in these mice resulted in reduction of adenoma size. Concomitantly, epithelial cell proliferation in the polyps was attenuated, suggesting that
SPHK1 regulates adenoma progression [
39].
Exogenous expression of
SPHK1 in vitro and
in vivo further showed that it is a key factor in esophageal carcinoma cell invasion. In the transwell invasion assay, upregulation of
SPHK1 expression significantly increased the invasion of EC9706 cells. Furthermore, upregulation of
SPHK1 expression significantly increased the proliferation of EC9706 cells
in vitro as well as increased EC9706 cell growth and spontaneous metastasis in nude mice. These studies support the view that
SPHK1 expression is vital for the maintenance of invasive and metastatic potential of esophageal carcinoma cells. Interestingly, neutralizing S1P, the product of SPHK1 enzymatic activity, with a specific monoclonal antibody was remarkably effective in slowing progression of cancers, such as lung [
40], colon [
41], breast [
42,
43], melanoma [
44] and ovarian cancers [
21,
45] in murine xenograft and allograft models [
46]. A critical question raised by these observations is how neutralization of this simple lysophospholipid can have such dramatic effects on tumor progression.
To glean mechanistic insight into the role of SPHK1 in invasion and metastasis, we surveyed potential links between SPHK1 and key molecules related to EGFR. Western blot analysis indicated that the expression of SPHK1 was significantly correlated with the phosphorylation of EGFR. In clustering the upregulated genes in the SPHK1 overexpression clones compared with control cells, SPHK1 expression was significantly correlated with that of IL6, ITGA2, IL8, EREG, MMP1, ITGA5, MMP3 and AREG. Western blotting indicated that the ligands of EGFR, EGF, EREG and AREG induced the expression of SPHK1 protein. These findings provide evidence for cross-talk between SPHK1 and the EGFR pathway and reveal a key role for SPHK1 in integrating events downstream of EGF receptors. An intriguing possibility is that many growth and angiogenic factors such as EGF and AREG involved in tumorigenesis may act through SPHK1 activation [
47]. Because both EGF and AREG have been implicated in progression of esophageal cancer, it was of interest to examine the involvement of SPHK1. There are numerous reports of rapid and transient activation of SPHK1 by growth and angiogenic factors [
48,
49]) that stimulate its phosphorylation at Ser225 [
50] and subsequent translocation to the plasma membrane [
51] where its substrate sphingosine resides, resulting in local formation of S1P [
52]. Some cross-talk models between EGFR and SPHK1/S1P have been proposed previously,
Estrada-Bernal, A et al[
53] have reported that treatment of glioma cell lines with EGF led to increased expression and activity of SphK1. Expression of EGFRvIII in glioma cells also activated and induced SphK1. In addition, siRNA to SphK1 partially inhibited EGFRvIII-induced growth and survival of glioma cells as well as ERK MAP kinase activation. SphK1 activity is necessary for survival of GBM-derived neurosphere cells, and EGFRvIII partially utilizes SphK1 to further enhance cell proliferation.
Shida, D. et al[
54]reported that LPA markedly enhanced SphK1 mRNA and protein in gastric cancer MKN1 cells, DLD1 colon cancer cells and MDA-MB-231 breast cancer cells. LPA transactivated the epidermal growth factor receptor (EGFR) in these cells, and the EGFR inhibitor AG1478 attenuated the increased SphK1 and S1P(3) expression induced by LPA. Their research finally showed that SphK1 is a convergence point of multiple cell surface receptors for three different ligands, LPA, EGF, and S1P, which have all been implicated in regulation of motility and invasiveness of cancer cells. In breast cancer,
Sukocheva, O et al[
55]demonstrated that E2-induced EGFR transactivation in human breast cancer cells is driven via a novel signaling system controlled by the lipid kinase sphingosine kinase-1 (SphK1). E2 stimulates SphK1 activation and the release of sphingosine 1-phosphate (S1P), by which E2 is capable of activating the S1P receptor Edg-3, resulting in the EGFR transactivation in a matrix metalloprotease-dependent manner. These findings reveal a key role for SphK1 in the coupling of the signals between three membrane-spanning events induced by E2, S1P, and EGF. However, it is still difficult to understand how such short-lived activation can be responsible for the profound involvement of SPHK1 in tumorigenicity or how this relates to its up-regulation in cancer. Our results imply that SPHK1 may be the central controller of amplification loops of EGF, AREG and EREG-EGFR interactions that can contribute to cancer progression.
Pan Jian, Ph.D. Immunology. Graduated from State Key Laboratory of Molecular Oncology, Cancer Institute (Hospital), Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing, PR China. Now is an associate professor of Department of Hematology and Oncology, Children's Hospital of Soochow University, Suzhou, China, and a gust professor of Translational Research Center, Second Hospital, The Second Clinical School, Nanjing Medical University, Nanjing, China.
Acknowledgements
We thank Professor Zhihua Yang (Cancer Institute/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China) for her kind help. We thank Dr. Mingrong Wang (Cancer Institute/Cancer Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China) for kindly providing the EC9706 cell line and Dr. Shimada Y (Kyoto University Graduate School of Medicine, Japan) for the KYSE esophageal cancer cells.
This work was supported by grants from the Natural Science Foundation for the Youth (No.81100371), National Key Basic Research Program (NKBRP) (973 program) (No.2010CB933902) and the National Natural Science Foundation (30570818 and 30600279).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PJ and NJ designed the study and wrote the manuscript, FX and ZWL participated in data analysis, PJ, WJ, ZYL and TYF finished the most experiment, HSY and WSY performed flow cytometry analysis. CBR and LZ collected the samples and made great contribution in making the tissue microarray. All authors read and approved the final manuscript.