Role of tumor-associated lymphatic endothelial cells in metastasis: a study of epithelial ovarian tumor in vitro.

(Cancer Sci 2010; 101: 679–685) Tumor‐associated lymphatic endothelial cells (TLEC) could play a key role in the process of tumor metastasis. The aim of this study was to investigate the effect of TLECs that were isolated from human epithelial ovarian tumor (EOT) on ovarian cancer cell line CAOV‐3 in vitro. First, TLECs in EOT were detected by immunochemistry and flow cytometry, then marked by lymphatic endothelial cell (LEC) marker LYVE‐1, isolated by magnetic beads, and cultured in vitro. The cells were identified by immunostaining of LEC markers LYVE‐1, Prox‐1, Podoplanin, VEGFR‐3, and pan‐endothelial cell marker CD31. TLECs from EOT can be detected, cultured, and identified in vitro successfully. The effects of TLECs on invasion and migration of CAOV‐3 cells were investigated by 12‐well Boyden chamber; the proliferation effect was studied by counting the Trypan blue exclusion cell number. Furthermore, changes in MMP‐2/9 secreted by CAOV‐3 cells treated with TLEC were shown using real‐time PCR and zymography, and TIMP‐1/2 was detected by real‐time PCR. In vitro, TLECs can enhance invasion and migration of CAOV‐3 cells, but have no significant effect on proliferation. It was clear that the expression of MMP‐9 increased and TIMP‐2 decreased in CAOV‐3 cells treated by TLECs, and the increasing of MMP‐9 was confirmed by zymography. TLECs from EOT can enhance migration and invasion of human ovarian carcinoma cell line in vitro, and the possible mechanism was through activation of MMP‐9/TIMP‐2. Tumor‐associated lymphatic endothelial cells may play a key role in the process of tumor metastasis. With the discovery of several markers of LECs, such as LYVE‐1,( 1 ) Prox‐1,( 2 ) Podoplanin (D2‐40),( 3 , 4 ) and VEGFR‐3,( 5 ) the detection of LECs became feasible. Using magnetic microbeads technology or flow cytometry, researchers successfully isolated and cultured LECs in vitro. Kriehuber et al. isolated Podoplanin(+) LECs from newborn foreskin cell suspensions by multicolor flow cytometry in 2001.( 6 ) Podgrabinska et al. isolated primary lymphatic and blood microvascular endothelial cells from human foreskin by immunoselection with the LEC marker LYVE‐1 in 2002.( 7 ) Garrafa et al. successfully isolated LECs from human lymph node, spleen, thymus, tonsil, and iliac lymphatic vessels by immunoselection with D2‐40 marked magnetic beads.( 8 , 9 ) Highly lymph node‐metastatic prostate cancer cell lines conditioned medium can enhance prostate LEC tube formation and migration.( 10 )Lymphangiogenesis has been reported in many solid tumors, not only in peritumoral but also intratumoral.( 11 , 12 ) However, little is known about the direct effects of TLEC on tumor cells, including EOT. Lymphatic spread is significant in aiding metastases in ovarian cancer,( 13 ) and the most important lymphangiogenic factor, VEGF‐C, is closely related to invasive phenotype and affected patients’ survival in ovarian carcinomas.( 14 ) However, lymphangiogenesis in EOC had not been seen before.( 15 ) There were many questions: was TLEC the same as NLEC (normal LEC from normal tissues such as newborn foreskin and other organs with abundant lymphatics)? If lymphangiogenesis was doubtable as the former study,( 15 ) was LEC absent in the metastatic spread of ovarian cancer? In order to help answer these questions, TLECs from EOT were isolated and cultured in vitro, and their effects on ovarian carcinoma cell line CAOV‐3 were detected in this study.

Patients. Fresh tumor samples were obtained from patients with advanced EOT who underwent primary surgical treatment at the Department of Obstetrics and Gynecology, Union Hospital (Wuhan, China) between May 2006 and June 2007. The collection followed national ethical standards. This work was directed by a colleague from the Department of Pathology of Union Hospital. It focused on tumors not containing any peritumoral tissues, so as not to interfere with the pathological diagnosis. Immediately before the experiment, frozen section microscopy was done to select EOT samples only. Forty‐one pathologically verified samples were studied, staged from IIIA to IIIC according to the classification of the International Federation of Gynecology and Obstetrics. None of the patients was subjected to chemotherapy or radiotherapy before the operation. The median age was 41 years (range 16–62 years). Twenty‐five cases were diagnosed as malignant ovarian tumor and the other 16 cases were borderline ovarian tumor. Five benign ovarian tumors and five normal ovary tissues from women with adenomyosis who underwent laparoscopic hysterectomy and unilateral ovariotomy were collected as control. Major agents. The MACS mini‐magnetic beads system, anti‐FITC magnetic beads, and basic beads were from Miltenyi Biotech (Bergisch Gladbach, Germany). Rabbit antihuman LYVE‐1 Ab was from Upstate Biotechnology (Charlottesville, VA, USA), goat antihuman Prox‐1 Ab and human recombination VEGF‐C were from R&D Systems (Minneapolis, MN, USA), and mouse antihuman VEGFR‐3 Ab was from Chemicon (Temecula, CA, USA). Mouse antihuman CD31 Ab and SABC FITC kit were supplied by Boster (Wuhan, China), We used mouse antihuman D2‐40 Ab and streptavidin‐perosidase kit from Zhongshan (Beijing, China), a double immunostaining kit from Maxin (Fuzhou, China), collagenase from Sigma (St. Louis, MO, USA), dispase from Invitrogen (Carlsbad, CA, USA), DMEM/F12 from Gibco (Grand Island, NY, USA), FBS from Gibco (Melbourne, Australia), and TRIzol from the Molecular Research Center (Cincinnati, OH, USA). Avian myeloblastosis virus and Taq polymerase were from Takara (Shiga, Japan). SYBR Green Real‐time PCR Master Mix was supplied by Toyobo (Osaka, Japan), 12‐well Transwell flasks were supplied by Costar (Cambridge, MA, USA), and Matrigel by Becton Dickinson (Bedford, MA, USA). Human dermal LEC was provided by ScienCell Research Laboratories (San Diego, CA, USA). Immunohistochemistry analysis. Samples (10 malignant, 5 borderline, 5 benign and 5 normal tissues) were obtained, 4 μm thin slices were cut, drawn upon 3‐aminopropyltrietoxysilane‐coated slides, and dried overnight at 37°C. Dewaxed tissue sections were pretreated with 10 mm citrate buffer (pH 6.0) at 121°C for 15 min. After rinsing in 1/15 m PBS, the sections were immersed in 0.3% H2O2 in methanol for 20 min to block the activity of endogenous peroxidase. They were incubated in 4% Block Ace (AbD, Oxford, UK) solution and then mouse antihuman D2‐40 Ab (diluted 1:400) at 4°C overnight. After washing with PBS, they were treated with peroxidase‐conjugated antimouse IgG for 1 h at room temperature and developed with diaminobenzidine. Following a rinse in PBS, the cells were then treated with a biotinylated antimouse rabbit IgG for 30 min, followed by alkaline phosphatase‐conjugated streptavidin for 30 min at room temperature. Double immunostaining was carried out according to the manufacturer’s instructions (Maxin). Monoclonal antihuman D2‐40 and Ki‐67 were sequent used. Flow cytometry analysis. Twenty‐five samples were also obtained from patients who were investigated in immunohistochemistry analysis. The samples were immediately transferred from the operating room to the laboratory in cold temperature preservation barrels. They were cut into very small pieces (approximately 1 mm3) and washed with cold PBS more than three times. Malignant and borderline samples were digested by preheated collagenase (200 U/mL) at 37°C for 2 h; benign and normal samples were digested by dispase at 4°C overnight and collagenase (200 U/mL) at 37°C for 2 h. Finally, single cell suspensions were obtained by filtration through sterile nylon meshes with 7.4 μm aperture. Antibody to human LYVE‐1 was used to mark single cell suspension for flow cytometry detection. Isolation and culture of TLECs from EOT. EOT single cell suspension was obtained from 15 malignant and 11 borderline tissues, resuspended in precooled MACS buffer (2 mm EDTA/0.5% BSA dissolved in PBS), adjusted to 106–107 cells/mL and exposed simultaneously to rabbit antihuman LYVE‐1 (final concentration 1:500) at 4°C for 30 min, then incubated with goat antirabbit F(ab’)2 FITC (10 μg/mL for 30 min at 4°C). Finally, anti‐FITC Ab with magnetic microbeads were used to mark the cell (final concentration 1:50 at 4°C for 15 min). The staining procedure did not interfere with cell viability, as determined by Trypan blue exclusion. Cells were washed by PBS after incubation, and we did our best to keep the cells in a cold environment. The LYVE‐1(+) cell was selected with LS columns and immediately placed into 2% gelatin‐coated flasks for culture in a humidified incubator in 5% CO2 at 37°C to yield an enriched LEC population. LECs were cultured in DMEM/F12 medium which 20% FBS supplemented with VEGF‐C (50 ng/mL) and without any antibiotics. This medium was named “LEM”. Basic beads were used to discard the non‐specificity particles if necessary. A little mixed fibroblast was washed out by incomplete trypsin digestion several times. Identification of LYVE‐1(+) cells from EOT. Immunocytochemistry was carried out on LYVE‐1(+) cells derived from EOT seeded, under the culture conditions described, on glass slides coated with 0.1% collagen, fixed with either methanol at −20°C for 10 min. Cells were then washed twice with PBS before incubation with primary and secondary antibodies (30 min each at room temperature). Cells were blocked in 10% goat serum in PBS for 15 min then incubated for 90 min at 37°C with antibodies to Prox‐1 (dilution 1:150) and LYVE‐1 (dilution 1:400). After washing with PBS, the cells were treated with peroxidase‐conjugated rabbit antigoat IgG for 1 h at room temperature and developed with diaminobenzidine. For immunofluorescence, LYVE‐1(+) cells were fixed with paraformaldehyde at room temperature for 90 min, blocked with normal goat serum at room temperature for 20 min, then incubated with mouse antihuman VEGFR‐3, Podoplanin, LYVE‐1, and rabbit antihuman CD31 antibodies at 4°C overnight. The secondary antibodies were all derived from goat and incubated at room temperature for 30 min. SABC FITC was used before observation. To identify the presence of false positives, due to non‐specific binding of the secondary Ab, all cells were treated with the same procedure. DAPI was used to show nucleii. Preparation of conditioned medium. In this study, TLEC was LYVE‐1(+) cells isolated from human ovarian epithelial tumor and identified by LEC markers. NLEC was human dermal LEC (ScienCell Research Laboratories). TLEC conditioned medium (TLM) was collected from subconfluent second generation TLECs (1.0 × 106 cells per dish) cultured in serum‐free LEM for 24 h. NLEC conditioned medium (NLM) was collected from subconfluent second generation NLECs (1.0 × 106 cells per dish) cultured in serum‐free LEM for 24 h. TLM was a special medium that contained the possible factors secreted by TLECs, and NLM was a control of TLM. Migration and invasion assays. Cell invasion assay was assessed using a 12‐well Boyden chamber (Transwell) and a reconstituted extracellular matrix membrane. The transwell inserts were coated with 100 mL of 1:8 dilution of Matrigel per well, and incubated at 37°C for 30 min to allow Matrigel polymerization. CAOV‐3 cells were treated with TLM or NLM for 24 h, then harvested by trypsinization, washed twice with PBS, and resuspended in serum‐free DMEM/F12. The cells were then added to the upper chamber, 3 × 105 cells per well in 500 mL medium without serum. The cells were then incubated in a humidified incubator in 5% at 37°C. The non‐migrating cells on the upper side of the filter were scraped and washed out; the transmigrating cell on the reverse side of the filter were stained with H&E. The transmigrating cells were counted and a photomicrograph was taken under (Olympus, Tokyo, Japan) microscope. Each experiment was done in triplicate and divided into four groups: (a) control group, in which CAOV‐3 cells treated with NLM were added to the upper chamber and 500 mL NLM added to the bottom chamber; (b) CAOV‐3 cells treated with TLM were added to the upper chamber and 500 mL NLM added to the bottom chamber; (c) CAOV‐3 cells treated with NLM were added to the upper chamber and 500 mL TLM added to the bottom chamber; (d) CAOV‐3 cells treated with TLM were added to the upper chamber and 500 mL TLM added to the bottom chamber. Migration assay was similar to the invasion assay but without Matrigel added to the filter. Four treatment groups were divided in the same way as the subgroups in the invasion assay. Proliferation assay. CAOV‐3 cells were treated with LEM, NLM, or TLM for 24 h before being harvested by trypsinization, and washed twice with PBS. The cells were then stained with Trypan blue solution and counted under an inverted microscope. Quantitative real‐time RT‐PCR. CAOV‐3 cells were treated with TLM or NLM for 24 h. Cells were harvested by trypsinization and washed twice with PBS before use. The mRNA expressions of MMP‐2, MMP‐9, TIMP‐1, TIMP‐2, and CXCR4 were detected by real‐time PCR, using the LightCycler system (Roche, Indianapolis, IN, USA) in combination with SYBR Green Real‐time PCR Master Mix and sequence‐specific primers for each gene (Table 1), β‐actin was used as the internal control (IC). Total RNA was extracted from cells by TRIzol according to the manufacturer’s instructions. The content of RNA was measured by spectrophotometer. cDNA synthesis followed the manufacturer’s instructions (Promega, Tokyo, Japan). RNA and Avian Myeloblastosis Virus was mixed with supplement in a sterile Eppendorf tube and kept at 42°C for 1 h. The cDNA was preserved in an icebox at −20°C. Real‐time PCR was carried out as recommended by the manufacturer. Gene expression in cells treated by TLM relative to NLM was analyzed using the comparative threshold cycle (Ct) method (2−ΔΔCt ).( 16 ) Expression levels of all genes of interest (GI) were normalized in each sample: ΔCtGI = CtGI − CtIC. Changes in normalized gene expression in TLM treated CAOV‐3 cells were than calculated relative to the average ΔCtGI value obtained in NLM treated CAOV‐3 cells (avg. ΔCtGI,NLM), using the formula 2−ΔΔCtGI, where −ΔΔCtGI = −(ΔCtGI − avg. ΔCtGI,NLM). Gene expression of NLM treated CAOV3 cells was defined as avg. ΔCtGI,NLM ± 1‐fold SD. Gene expression alterations in TLM treated CAOV3 cells of more than ±1‐fold SD from avg. ΔCtGI,NLM were regarded as increased and reduced, respectively, gene expression levels relative to NLM treated CAOV3 cells. Table 1. | Gene | Sequence | | |---|---|---| | Forward | Reverse | | | CXCR4 | GGTGGTCTATGTTGGCGTCT | TGGAGTGTGACAGCTTGGAG | | TIMP‐1 | AATTCCGACCTCGTCATCAG | GAAAGATGGGAGTGGGAACA | | TIMP‐2 | ATCAGGGCCAAAGCGGTCAGTGAG | ATCTTGCACTCGCAGCCCATCTGG | | MMP‐9 | CCGAGCTGACTCGACGGTGATGG | GAGGTGCCGGATGCCATTCACGTC | | MMP‐2 | TACTGGATCTACTCAGCCAGCA | CTTCAGGTAATAGGCACCCTTG | | β‐actin | AAGAGAGGCATCCTCACCCT | GGAAGGAAGGCTGGAAG | Gelatin zymography. CAOV‐3 cells were treated with TLM or NLM for 24 h. Conditioned media were collected and cells were lysed with 200 mm NaCl, 20 mm Tris‐HCl (pH 8.3), 1% Triton X‐100, 1 mm EDTA, and protease inhibitor cocktail tablets. Proteins of this conditioned media (5 μg per lane) were subjected to 9% SDS‐PAGE with 0.1% gelatin under non‐reducing conditions, followed by washing in 2.5% Triton X‐100 for 1 h. Gels were then incubated overnight at 37°C in activating buffer (50 mm Tris‐HCl [pH 7.5], 5 mm CaCl2, 0.03% Brij 35, 0.02% NaN3, and 1 μm ZnCl2) and gelatinolytic activities were revealed with Coomassie blue staining. Identical gels were carried out in the presence of 30 mm EDTA, attesting to the metalloproteinase nature of detected enzymes. Statistical analysis. Analysis was carried out using SAS version 6.12 (SAS Institute, Cary, NC, USA) or SPSS version 12.0 (SPSS, Chicago, IL, USA). Normal distributed results that were tested for normality were expressed as mean ± SEM obtained from at least three separate experiments in each group. Fisher’s exact test by SAS version 6.12 was used to compare the successful ratios between different clinicopathologic variables groups. Differences between groups were assessed by one‐way ANOVA and Newman–Keuls multiple comparison test where appropriate. P < 0.05 was considered statistically significant.

Detection of LYVE‐1(+) cells in EOT samples. It was easy to find TLECs in EOT samples by immunochemistry because D2‐40 (Podoplanin) stained yellow vessels. Most lymphatic vessels located near tumor cell and lymphatic vascular invasion was a common phenomenon in EOT (Fig. 1A). Intratumor lymphatic vessels were found in BOT. By double immunochemistry of Ki‐67 and Podoplanin, the expression of Ki‐67 on LECs of EOT was mild (Fig. 1A). Some LYVE‐1(+) cells in the EOT (not including benign) single cell suspension were detected by flow cytometry, approximately 4.33 ± 2.21% (range 3.0–9.5%) in the 15 tested samples. The rate in benign and normal samples was lower than 1%. It was only possible to isolate LECs from malignant and borderline cells by magnetic beads (a sample is shown in Figure 1B). Isolation, culture, and identification of LEC. It was not easy to isolate and culture LYVE‐1(+) cells from EOT in vitro. The most important factor was the age of the patients, only five cases from 22 samples of EOT were cultured more than five passages (patients age 16–29 years, all with BOT). As the number of vigor LYVE‐1(+) cells isolated directly from EOT single suspensions was relatively low, the primary passage had cultured more than 20 days before confluence (Fig. 2, bright field). Identification of the second generation of these LYVE‐1(+) cells showed that not only pan‐endothelial cell marker CD31 and isolation marker LYVE‐1, but also other LEC markers, such as VEGFR‐3, Prox‐1, and D2‐40 (Podoplanin), were expressed at these cells (Fig. 2). The lymphatic endothelial character of these cells was clear and the purity of second generation TLECs was approximately 95%. Effects of TLECs on tumor cell invasion, migration, and proliferation. Transmigration of cells across a complex basement membrane is often regarded as an indicator of the invasive behavior of cells. The migratory and invasive abilities of four groups of CAOV‐3 cells were examined. Grouped as shown in ‘Materials and Methods’, the invasive and migratory abilities of groups b and d were significantly more powerful than groups a and c (P 0.05). The detail of these effects is shown in Figure 3(A,B). No significant proliferation effect of TLECs on CAOV‐3 cells was found compared to control (P > 0.05; Fig. 3C). Changes in expression of MMPs and TIMPs in CAOV‐3 cells treated by TLECs. The expression of MMP‐9 mRNA in CAOV‐3 cells treated by TLM for 24 h increased significantly compared to the cells treated by NLM (P < 0.05), but the expression of TIMP‐2 mRNA in CAOV‐3 cells treated by TLM was decreased compared to cells treated by NLM (P < 0.05). The expression of MMP‐2, TIMP‐1, and CXCR4 mRNA in CAOV‐3 cells treated by TLM relative to NLM did not change significantly. Data from real‐time PCR is shown in Figure 4(A). The activity of the MMP‐2 and MMP‐9 enzyme was shown by gelatin zymography, the variation was indicated by electrophoretogram and analyzed. As shown in Figure 4(B), the level of MMP‐9 clearly increased (P < 0.05).

During the last decade, markers of LEC, such as Prox‐1, VEGFR‐3, LYVE‐1, and D2‐40 (Podoplanin),( 17 ) have made it possible to investigate lymphangiogenesis regulatory mechanisms, and to analyse their involvement in tumor progression. Among the regulatory systems, growth factors VEGF‐C and ‐D, which bind and activate their common receptor VEGFR‐3,( 18 ) appear to play an important role in this process. Recombinant VEGF‐A and VEGF‐C potently promoted prostate LEC tube formation, migration, and proliferation in vitro, indicating that signaling through VEGFR‐2, rather than VEGFR‐3, is involved in these responses.( 10 ) It was reported that PDGF‐BB can induce lymphangiogenesis and promote lymphatic metastasis. ( 19 ) But whether and how TLECs influence tumor cells is still a pending problem. Blood vascular endothelial cell compartment can produce growth factor to promote tumor cell growth.( 20 ) LECs might also produce some factors that promote tumor cell growth and lymphatic metastasis. BOT can be further substratified into risk categories by the presence of nodular aggregates of BOT in lymph nodes. This is a more common feature in cases with micropapillary architecture and associated stromal reaction in the intranodal tumor.( 21 ) The presence of lymphatic vessel invasion in microinvasive BOT corroborates the view that microinvasion represents an early, but very low risk, invasive process that morphologically links BOT and low‐grade serous carcinoma.( 22 ) Lymphatic spread might be significant in aiding metastases in ovarian cancer but requires other biological factors to act in conjunction, as it does not have clear‐cut prognostic significance. Dissemination of ovarian cancer does not occur primarily through vascular or lymphatic routes but might occur through direct intraperitoneal spread of disease.( 14 ) Like most solid tumors, both peritumoral and intratumoral LECs were found in EOT tissues, and approximately 3–9.5% LYVE‐1(+) cells were detected in its single cell suspensions by flow cytometry. The fewer LECs in benign and normal ovarian tissues indicated that TLECs from EOT were rarely interfered with by NLECs. The isolation of pure populations of LECs, the investigation of lymphatic metastases in animal models, and the identification of markers that discriminate lymphatics from blood vessels at immunohistochemistry are current advances in the field of lymphangiogenesis.( 23 ) In this study, the age of the patients was the decisive factor in the successful isolation and culture of LECs from EOT. The result was better if the sample was from younger patient, but the morbidity of EOC in young women is very low,( 24 , 25 ) so it was too difficult to culture enough malignant ovarian TLECs in vitro. However, TLECs from BOT were available in this research (data shown in Table 2). Table 2. | n | Successful† (%) | P‡ | || |---|---|---|---|---| | Age of patients (years) | >40 | 13 | 0 (0.00) | 0.01 | | <40 | 13 | 6 (46.15) | || | Pathological diagnosis | Malignant (EOC) | 15 | 1 (6.67) | 0.03 | | Borderline (BOT) | 11 | 5 (45.45) | †Sample cultured more than three passages. ‡Fisher’s exact test. BOT, borderline ovarian tumors; EOC, epithelial ovarian cancer. In the past, the sialomucin CD34 was used as a marker of blood endothelial cells, CD31+/CD34− magnetic beads were successfully used to isolate LECs from newborn foreskin,( 1 ) but this rule was seen to be unsuitable for TLECs as a research reported the existence of CD34 in TLECs, by immunohistochemistry.( 26 ) LYVE‐1 was not only limited to LECs, endothelial cell of hepatic sinus and tumor‐associated macrophages were also reported.( 17 ) The identification showed the majority of LYVE‐1+ cells from EOT, cultured by LEM, were LECs. Perhaps macrophages cannot grow in LEM. Meanwhile, a published paper reported tumor cells produced multiple hemangiogenic and lymphatic angiogenic factors that stimulate the growth of LECs. Highly lymph node‐metastatic prostate cancer cell lines and their conditioned medium enhanced prostate LEC tube formation and migration, whereas poorly lymph node‐metastatic prostate cancer cells or normal prostate epithelial cells or their conditioned medium had no effect.( 10 ) This result suggested LECs treated by tumor, especially highly lymph node‐metastatic tumor, were different from others. Further study is required to analyze the different phenotypes between TLECs (not tumor cell treated LECs) and NLECs, as it was proposed that not only CD34 but also some lymphatic metastasis key molecules are located on TLECs. This study found that TLECs might secrete factors to enhance the migration and invasion of ovarian cancer CAOV‐3 cells in vitro, and this effect was not only chemotaxis, the difference between tumor‐associated and normal LECs was clear. LYVE‐1,( 27 ) secondary lymphoid chemochine,( 28 ) Mannose receptor, and common lymphatic endothelial and vascular endothelial receptor‐1( 29 ) were the possible factors secreted by TLEC to direct the binding of cancer cells to the lymph vessel endothelium. Coincidentally, a recent study found that LEC induced by oral cancer cells displayed abnormal characteristics by gene expression profile and were distinct at the molecular level.( 30 ) These factors could increase the expression of MMP‐9 and decrease TIMP‐2 in CAOV‐3 cells. The mechanism of MMP‐9 increasing and TIMP‐2 decreasing was unknown, but it might be the direct cause to promote the migration and invasion of tumor cells. Published reports suggest that LECs can secrete CXCR12‐like factors that can regulate EOC cell invasion by activating MMP‐9.( 31 , 32 ) However the results of this study showed that the expression of its ligand CXCR4 in CAOV‐3 cells treated by TLECs did not obviously increase. While CXCL12 mRNA was abesnt in CAOV‐3 cells, the level of CXCR4 mRNA in CAOV‐3 cells can increase by the stimulation of SDF‐1.( 33 ) In this study, we did not find changes in CXCR4 in TLM‐treated CAOV‐3 cells, so the involvement of SDF‐1 in this process was absent or mild. The proliferation effect of TLECs on cancer cell line CAOV‐3 cells was not seen in this study. This is also a pending problem that should be studied by MTT or other proliferation assay. In our opinion, the possible unknown key factor of lymphatic metastasis was located on tumor‐associated (especially high lymphatic metastatic tumor) LECs. Overall, our study initially explored that TLECs can enhance migration and invasion of ovarian cancer cell line CAOV‐3 in vitro. Lymphangiogenesis occurring in EOT was not investigated in this study, and TLECs were not absent in metastasis of EOT. Abbreviations - BOT borderline ovarian tumors - EOC epithelial ovarian cancer - EOT epithelial ovarian tumor - LEC lymphatic endothelial cell - LEM lymphatic endothelial cell medium - LYVE‐1 lymphatic vessel endothelial hyaluronic acid receptor‐1 - NLEC normal lymphatic microvascular endothelial cell - NLM NLEC conditioned medium - TLEC tumor‐associated microvascular lymphatic endothelial cell - TLM TLEC conditioned medium - VEGF vascular endothelial growth factor - VEGFR vascular endothelial growth factor receptor Acknowledgment This work was supported by Grant No. 30772325 from the National Natural Science Foundation of China.

- 1. Banerji S, Ni J, Wang SX et al. LYVE‐1, a new homologue of the CD44 glycoprotein, is a lymph‐specific receptor for hyaluronan. Cell Biol 1999; 144: 789–801. [DOI] [PMC free article] [PubMed] [Google Scholar] - 2. Wilting J, Papoutsi M, Christ B et al. The transcription factor Prox1 is a marker for lymphatic endothelial cells in normal and diseased human tissues. FASEB J 2002; 16: 1271–3. [DOI] [PubMed] [Google Scholar] - 3. Dumoff KL, Chu C, Xu X et al. D2‐40 immunoreactivity correlates with lymphatic invasion and nodal metastasis in early‐stage squamous cell carcinoma of the uterine cervix. Mod Pathol 2005; 18: 97–104. [DOI] [PubMed] [Google Scholar] - 4. He Y, Karpanen T, Alitalo K. Role of lymphangiogenic factors in tumor metastasis. Biochim Biophys Acta 2004; 1654: 3–12. [DOI] [PubMed] [Google Scholar] - 5. Akagi K, Ikeda Y, Miyazaki M et al. Vascular endothelial growth factor‐C (VEGF‐C) expression in human colorectal cancer tissues. Br J Cancer 2000; 83: 887–91. [DOI] [PMC free article] [PubMed] [Google Scholar] - 6. Kriehuber E, Breiteneder‐Geleff S, Groeger M et al. Isolation and characterization of dermal lymphatic and blood endothelial cells reveal stable and functionally specialized cell lineages. J Exp Med 2001; 194: 797–808. [DOI] [PMC free article] [PubMed] [Google Scholar] - 7. Podgrabinska S, Braun P, Velasco P et al. Molecular characterization of lymphatic endothelial cells. PNAS 2002; 99: 16069–74. [DOI] [PMC free article] [PubMed] [Google Scholar] - 8. Garrafa E, Trainini L, Benetti A et al. Isolation, purification, and heterogeneity of human lymphatic endothelial cells from different tissues. Lymphology 2005; 38: 159–66. [PubMed] [Google Scholar] - 9. Garrafa E, Alessandri G, Benetti A et al. Isolation and characterization of lymphatic microvascular endothelial cells from human tonsils. J Cell Physiol 2006; 207: 107–13. [DOI] [PubMed] [Google Scholar] - 10. Zeng Y, Opeskin K, Goad J et al. Tumor‐induced activation of lymphatic endothelial cells via vascular endothelial growth factor receptor‐2 is critical for prostate cancer lymphatic metastasis. Cancer Res 2006; 66: 9566–75. [DOI] [PubMed] [Google Scholar] - 11. Van der Auwera I, Van den Eynden GG, Colpaert CG et al. Tumor lymphangiogenesis in inflammatory breast carcinoma: a histomorphometric study. Clin Cancer Res 2005; 11: 7637–42. [DOI] [PubMed] [Google Scholar] - 12. Gao P, Zhou GY, Yin G et al. Lymphatic vessel density as a prognostic indicator for patients with stage I cervical carcinoma. Hum Pathol 2006; 37: 719–25. [DOI] [PubMed] [Google Scholar] - 13. Harter P, Gnauert K, Hils R et al. Pattern and clinical predictors of lymph node metastases in epithelial ovarian cancer. Int J Gynecol Cancer 2007; 17: 1238–44. [DOI] [PubMed] [Google Scholar] - 14. Ueda M, Hung YC, Terai Y et al. Vascular endothelial growth factor‐C expression and invasive phenotype in ovarian carcinomas. Clin Cancer Res 2005; 11: 3225–32. [DOI] [PubMed] [Google Scholar] - 15. Sundar SS, Zhang H, Brown P et al. Role of lymphangiogenesis in epithelial ovarian cancer. Br J Cancer 2006; 94: 1650–7. [DOI] [PMC free article] [PubMed] [Google Scholar] - 16. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real‐time quantitative PCR and the 2(‐Delta Delta C(T)) Method. Methods 2001; 25: 402–8. [DOI] [PubMed] [Google Scholar] - 17. Baluk P, McDonald DM. Markers for microscopic imaging of lymphangiogenesis and angiogenesis. Ann N Y Acad Sci 2008; 1131: 1–12. [DOI] [PubMed] [Google Scholar] - 18. Vittet D, Feige JJ. Lymphangiogenesis and tumor progression. Bull Cancer 2007; 94: 881–6. [PubMed] [Google Scholar] - 19. Vincent L, Rafii S. Vascular frontiers without borders: multifaceted roles of platelet‐derived growth factor (PDGF) in supporting postnatal angiogenesis and lymphangiogenesis. Cancer Cell 2004; 6: 307–9. [DOI] [PubMed] [Google Scholar] - 20. Cao Y. Opinion: emerging mechanisms of tumour lymphangiogenesis and lymphatic metastasis. Nat Rev Cancer 2005; 5: 735–43. [DOI] [PubMed] [Google Scholar] - 21. McKenney JK, Balzer BL, Longacre TA. Lymph node involvement in ovarian serous tumors of low malignant potential (borderline tumors): pathology, prognosis, and proposed classification. Am J Surg Pathol 2006; 30: 614–24. [DOI] [PubMed] [Google Scholar] - 22. Sangoi AR, McKenney JK, Dadras SS et al. Lymphatic vascular invasion in ovarian serous tumors of low malignant potential with stromal microinvasion: a case control study. Am J Surg Pathol 2008; 32: 261–8. [DOI] [PubMed] [Google Scholar] - 23. Sundar SS, Ganesan TS. Role of lymphangiogenesis in cancer. J Clin Oncol 2007; 25: 4298–307. [DOI] [PubMed] [Google Scholar] - 24. Chen R, Shen K, Wu M et al. Analysis of 21 cases of epithelial ovarian malignancies in women aged below 30 years. Zhonghua Fu Chan Ke Za Zhi 2005; 40: 417–20. [PubMed] [Google Scholar] - 25. Rodriguez M, Nguyen HN, Averette HE et al. National survey of ovarian carcinoma XII. Epithelial ovarian malignancies in women less than or equal to 25 years of age. Cancer 1994; 73: 1245–50. [DOI] [PubMed] [Google Scholar] - 26. Fiedler U, Christian S, Koidl S et al. The sialomucin CD34 is a marker of lymphatic endothelial cells in human tumors. Am J Pathol 2006; 168: 1045–53. [DOI] [PMC free article] [PubMed] [Google Scholar] - 27. Jackson DG, Prevo R, Clasper S et al. LYVE‐1, the lymphatic system and tumor lymphangiogenesis. Trends Immunol 2001; 22: 317–21. [DOI] [PubMed] [Google Scholar] - 28. Muller A, Homey B, Soto H et al. Involvement of chemokine receptors in breast cancer metastasis. Nature 2001; 410: 50–6. [DOI] [PubMed] [Google Scholar] - 29. Irjala H, Alanen K, Grénman R et al. Mannose receptor (MR) and common lymphatic endothelial and vascular endothelial receptor (CLEVER)‐1 direct the binding of cancer cells to the lymph vessel endothelium. Cancer Res 2003; 63: 4671–6. [PubMed] [Google Scholar] - 30. Zhuang Z, Jian P, Longjiang L et al. Identification of oral cancer cell‐induced changes in gene expression profile of lymphatic endothelial cell. Cancer Invest 2008; 26: 1002–7. [DOI] [PubMed] [Google Scholar] - 31. Yuecheng Y, Xiaoyan X. Stromal‐cell derived factor‐1 regulates epithelial ovarian cancer cell invasion by activating matrix metalloproteinase‐9 and matrix metalloproteinase‐2. Eur J Cancer Prev 2007; 16: 430–5. [DOI] [PubMed] [Google Scholar] - 32. Kajiyama H, Shibata K, Terauchi M et al. Involvement of SDF‐1alpha/CXCR4 axis in the enhanced peritoneal metastasis of epithelial ovarian carcinoma. Int J Cancer 2008; 122: 91–9. [DOI] [PubMed] [Google Scholar] - 33. Jiang YP, Wu XH, Xing HY et al. Role of CXCL12 in metastasis of human ovarian cancer. Chin Med J (Engl) 2007; 120: 1251–5. [PubMed] [Google Scholar]