Lysophosphatidic acid stimulates cell migration, invasion, and colony formation as well as tumorigenesis/metastasis of mouse ovarian cancer in immunocompetent mice

Epithelial ovarian cancer (EOC) is the most deadly gynecologic disease mainly due to the lack of highly sensitive and specific methods for early detection and effective treatments for late-stage diseases. Since our first report of elevated lysophosphatidic acid (LPA) levels in human ovarian cancer ascites and the identification of the stimulatory role of LPA in EOC more than 13 years ago (1), extensive information has accumulated showing that LPA plays an important role in EOC development. LPA regulates almost every aspect of EOC cell biology, including cell proliferation, apoptosis, morphology, drug resistance, adhesion, migration, and invasion (2–4). Most of the studies on the mechanisms of LPA in ovarian cancer have been conducted in vitro. Only recently, animal studies showing a role for LPA and/or its receptors in the development of cancers in vivo have been published (5–9). We have recently provided the first direct evidence that LPA stimulates tumor metastasis in vivo, which is inhibited by LY294002, a phosphatidylinositol 3-kinase inhibitor, and by 17-dimethylaminoethylamino-17-demethoxygeldanamycin, an inhibitor of the heat shock protein 90, which stabilizes hypoxia-inducible factor-1α (10, 11). However, these experiments were conducted using nude or severe combined immunodeficient mice, which are immunocompromised. It will be important to test the same concepts in immunocompetent mice.

The two most common and critical processes in metastasis are cell migration and invasion. We and others have shown that LPA stimulates the migration and invasion of ovarian cancer cells (10–22). Although most, if not all, of these studies were conducted in established ovarian cancer cell lines, we have recently conducted LPA- and S1P-induced cell migration and invasion experiments in primary EOC and nonmalignant human ovarian surface epithelial cells (11, 23).

All three LPA receptors have been implicated in the migration of different cell types and under different assay conditions (24, 25). Whereas LPA₁ plays a very important role in migration of breast, pancreatic, and prostate cancer cells (25, 26), our data suggest that LPA₂ and LPA₃ are more important in cell adhesion, migration, and invasion of EOC cells, which are consistent with reports showing that LPA₁ is involved in the negative growth regulation in ovarian cancer cells and that LPA₂ and LPA₃, but not LPA₁, expression is up-regulated in last-stage EOC (27). In addition, we have shown that either inhibitors and/or dominant-negative forms of G_i protein, phosphatidylinositol 3-kinase, or cytosolic phospholipase A₂ (cPLA₂) completely or nearly completely block LPA-induced cell migration in human EOC cells, suggesting that these three molecules play a pivotal role in this process (2, 10, 11, 24). The PLA₂ family of enzymes catalyzes the hydrolysis of the sn-2 position of phospholipids to generate free fatty acids (arachidonic acid in particular) and lysophosholipids. We have recently shown that LPA regulates both cPLA₂ expression and activation (assessed by its phosphorylation) in human EOC cells (11, 24). The role of cPLA₂ in cell migration is likely related to its ability to produce arachidonic acid (15, 17, 24). Arachidonic acid is converted to a group of bioactive compounds called eicosanoids (including prostagladins and leukotrienes) by cyclooxygenase (COX) and by lipoxygenase (LOX) enzymes. These enzymes and their lipid products play an important role in inflammation and cancer (28). The role of COX and LOX isoforms, particularly COX-2, in inflammation and cancer has been well established (28). However, COX-1, but not COX-2, is overexpressed in human EOC tissues, which has been confirmed by several independent studies (29, 30). In addition, COX-1, but not COX-2, is also overexpressed in all of the major current EOC mouse models, including the TVA mouse models and the models with deletions of both p53 and Rb, with induction of the transforming region of SV40 under the control of Mullerian inhibitory substance type II receptor, or with local activation of K-Ras in the absence of Pten in the ovarian surface epithelium (29). Thus, COX-1 has been proposed as a potential marker and a possible target for the prevention and/or treatment of EOC (29). There are three major groups of LOX enzymes: 5-LOX, 12-LOX, and 15-LOX. Expression of LOX pathway receptors have been found in human EOC tissues (31). 15-LOX-2, but not 15-LOX-1 or COX-2, is overexpressed in human EOC and EOC metastases (32).

Roby et al. have established mouse EOC cell lines using mouse ovarian surface epithelial cells isolated from normal mouse ovaries through repetitive passaging and spontaneous transformation in vitro (33). Ten clonal cell lines were established with distinguishable chromosomal changes and tumorigenic potential. Among these cell lines, ID8 cells form tumors in syngeneic C57/BL6 mice with the highest incidence and shortest lag time when i.p. injected (33). Mouse ovarian surface epithelial cells can be used in immunocompetent mice.

In this work, we characterized the LPA responsiveness in ID8 cells and examined the signaling pathways by which LPA induced cell migration. The in vivo effects of LPA on tumorigenesis, metastasis, and mouse survival were determined in syngeneic mice. Our results suggest that the ID8 cells represent a valuable system to develop and evaluate therapeutic reagents targeting LPA in immunocompetent mice.

18:1 LPA was from Avanti Polar Lipids. Human fibronectin and other extracellular matrix proteins were purchased from Chemicon International. AACOCF3, Akt inhibitor, LY294002, and U0126 were from Calbiochem. Ki16425, thiazolyl blue tetrazolium bromide, AA-861, PD146176, baicalein, and poly(2-hydroxyethyl methacrylate) were from Sigma. SC-560 was from Cayman Chemical. Nordihydroguaiaretic acid and NS398 were from Biomol. Antibodies against phosphorylated cPLA₂ and total cPLA₂ were from Cell Signaling Technology. We have used nontoxic doses of the inhibitors. The effects on cell survival were checked for all of the inhibitors used.

ID8 cells were obtained from Dr. Paul F. Terranova (University of Kansas Medical Center). Cells were cultured in DMEM supplemented with 4% fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomycin, 5 μg/mL insulin, 5 μg/mL transferrin, and 5 ng/mL sodium selenite. To establish ID8 cells that stably expressing luciferase, cells were transfected with LUCIE (a kind gift from Dr. Mark Kelley, Indiana University School of Medicine) and selection for resistance to zeocin (100 μg/mL) was initiated 48 h after transfection. Cells were transfected with a dominant-negative cPLA₂^514-733 (a kind gift from Dr. Sabine Adam, Christian-Albrechts-Universität Kiel) using the Nucleofector Solution V and program T-020 (Amaxa). 15-LOX or COX-1 small interfering RNA (siRNA; Ambion; s62269 and s72391) were transfected into cells using the same method as above and harvested for experiments 36 h post-transfection.

Cells (10⁴) in 100 μL serum-free DMEM were added in 96-well plates and incubated at 37°C overnight. LPA (in 100 μL serum-free DMEM) at a 2-fold concentration was then added to the wells and cells were incubated for an additional 24 or 48 h. Cells were trypsinized and counted in the presence of trypan blue using a hemocytometer. For colony assays, CytoSelect 96-well cell transformation assay (Soft Agar Colony Formation; Cell Biolabs) was done according to the company's manual with some modifications. In brief, the cell agar layer was prepared by mixing equal volumes of the 1.2% agar in 2× DMEM and cell suspension (4 × 10⁵ cells/mL in DMEM with various stimulators). DMEM (100 μL) containing the stimulus was layered on the top of the cell agar layer. The assay was done for 8 days. For anoikis assays, 96-well plates were coated twice with poly(2-hydroxyethyl methacrylate) (10 mg/mL in 95% ethanol) and rinsed thoroughly with PBS. ID8 cells in 100 μL serum-free medium with various concentrations of LPA were plated on the coated plates. At the indicated times, the cells were subjected to a brief trypsin treatment to dissociate spheroids and counted in a hemocytometer.

Cell migrations were done in 24-well modified Boyden chambers (Corning Life Sciences) with 8 μm pore size polycarbonate polyvinylpyrrolidone-free Nucleopore filters. Briefly, the bottom phase of the top chamber was coated with 10 μL laminin or other extracellular matrix at a concentration of 10 μg/mL. ID8 cells were serum starved overnight and 5 × 10⁴ cells in 300 μL DMEM were added to the top chamber. DMEM (300 μL) containing the indicated concentrations of LPA was added to the bottom chambers as the chemoattractant. The chambers were then incubated at 37°C for 4 h. After incubation, the top chambers were rinsed with PBS and swabbed with a cotton swab to remove nonmigrated cells. The migrated cells were then fixed in methanol for 30 min and stained with crystal violet (Fisher Scientific) for 30 min followed by a wash with water. After the chambers were dried, pictures were taken with a Nikon SMZ 1000 microscope and migrated cells were counted in three different fields. Invasion assays were done using modified Boyden's chamber (Becton Dickinson Labware). The Matrigel in the chamber was allowed to swell for 2 h in PBS before the experiment. Cells were starved for 24 h and 5 × 10⁴ cells were resuspended in 300 μL DMEM. The cells were added to the top chamber and 300 μL serum-free medium with or without LPA was introduced to the bottom chamber. The invasion assays were conducted for 16 h. The same method for fixing and counting cells described above for cell migration was used.

The expression levels of LPA_1-3 receptors were evaluated by reverse transcription-PCR. In brief, total RNA was extracted from cells using RNeasy mini kit (Qiagen) and reverse transcribed by MMLV (Invitrogen). Derived cDNAs were amplified using PCR master mix (Promega). Primer sequences for LPA_1-3 are as follows: LPA₁, 5′-AATCGAGAGGCACATTACGG-3′ and 5′-GTTGAAAATGGCCCAGAAGA-3′; LPA₂, 5′-TTGTCTTCCTGCTCATGGTG-3′ and 5′-TCAGCATCTCGGCAAGAGTA-3′; and LPA₃, 5′-TGCTCATTTTGCTTGTCTGG-3′ and 5′-GCCATACATGTCCTCGTCCT-3′. β-Actin was amplified as a housekeeping gene with primers 5′-ACCGCTCGTTGCCATTAGTGATGA-3′ and 5′-AAGGCCAACCGTGAAAAGATGACC-3′. PCR was also done using 15-LOX primers: 5′-TACCTGTGGTTGATCGGACA-3′ and 5′-AGTTCCTCCTCCCTGTGGTT-3′ and COX-1 primers: 5′-CATTGCACATCCATCCACTC-3′ and 5′-CCCAGAGCCAGTATCCATGT-3′.

Cells in 6-well plate were rinsed with PBS and lysed in 100 μL Laemmli sample buffer (Bio-Rad). Proteins were separated by 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Bio-Rad). Membranes were blocked in 5% skim milk for 2 h and then with the indicated primary antibodies overnight followed by incubation with appropriate secondary antibodies for 1 h. Specific proteins were detected using Enhanced Chemiluminescence Plus Western blotting detection system (GE Healthcare). Protein loading was verified by stripping and reprobing blots with antibodies against β-actin or total cPLA₂.

Four- to 6-week-old female C57/BL6 mice were purchased from Charles River Labs and housed at the Laboratory Animal Resource Center at the Indiana University School of Medicine. All surgeries were done under approved animal protocols and conducted as described previously (10, 11). LPA was delivered using Alzet microosmotic pumps (model 1002) as described previously (10, 11). Briefly, the pumps were first siliconized with Sigmacote (Sigma) to prevent LPA from adhering to the wall of the pumps. Pumps containing 100 μL concentrated LPA (4 mmol/L in PBS) or PBS were implanted into the peritoneal cavities. Following surgery, ID8-Luc cells (5 × 10⁶) in 200 μL PBS were i.p. injected into the peritoneal cavity. Pumps were replaced after 14 days. In a separate model, LPA was also delivered by daily i.p. injection. Mice were i.p. injected with 5 × 10⁶ ID8 cells in 1 mL PBS and then administered daily with LPA (100 μmol/L) in 200 μL PBS. The LPA injection lasted for 4 to 8 weeks. In the orthotopic model, 5 × 10⁵ ID8 cells in 5 μL PBS were injected into the bursa of a mouse ovary, and LPA was administered by daily i.p. injection for 4 weeks. Animal's health was monitored daily. Animals were euthanized when they developed large volumes of ascites (evidenced by abdominal detention). Tumorigenesis was recorded by counting the numbers and sizes of tumor foci on each organ.

The results were presented as mean ± SE of at least three individual experiments unless otherwise stated. Statistical significance was assessed by Kaplan-Meier for animal average survival and Student's t test for the other experiments. P < 0.05 was considered to be significant.

We conducted a series of experiments to characterize ID8 mouse ovarian cancer cells for their responsiveness to LPA to test the effect of LPA in vivo in immunocompetent mice. The effect of LPA on cell proliferation was first tested in two-dimensional tissue culture dishes in the presence or absence of LPA (0.1-4 μmol/L) and no significant effect was seen after 24 or 48 h, suggesting that LPA did not affect survival or proliferation of ID8 cells. In comparison, LPA (4 μmol/L) significantly (but rather weakly) stimulated proliferation of human ovarian cancer SKOV3 cells at 48 h (Fig. 1A). Using a 96-well soft agar colony formation assay, we found that although it was not as potent as fetal bovine serum (10%) or human ovarian cancer ascites (10%), LPA (4 μmol/L) significantly stimulated colony formation of ID8 cells (Fig. 1B), which is similar to the effect of LPA in human ovarian cancer cells (34). We showed further that the promoting role of LPA in colony formation was not likely through protecting cells from anoikis, because LPA did not increase cell survival when ID8 cell attachment was prevented by poly(2-hydroxyethyl methacrylate) (Fig. 1C).

Similar to human EOC cells (24, 35–40), LPA (1 μmol/L) induced migration of ID8 cells toward four different extracellular matrix proteins: collagen I, fibronectin, laminin, and vitronectin (Supplementary Fig. S1A).³

We tested whether cPLA₂ and/or iPLA₂ were similarly involved in LPA-induced migration of ID8 cells as those human EOC cells (2, 10, 11, 13, 24). A general inhibitor (AACOCF3) for both cPLA₂ and iPLA₂ almost completely blocked LPA (1 μmol/L)-induced ID8 cell migration, suggesting that one or both of these enzymes are involved in the process (Fig. 2A). To distinguish the effect between iPLA₂ and cPLA₂, we tested a more specific inhibitor for iPLA₂, HELSS. As shown in Fig. 2B, HELSS significantly reduced basal levels of cell migration as well as migration induced by LPA. To investigate the effect of cPLA₂, ID8 cells were transiently transfected with a dominant-negative form of cPLA₂ (42). We found that the LPA-induced cell migration were significantly (∼70%) inhibited (Fig. 2C), suggesting that the cPLA₂ signaling pathway is also involved in the effect of LPA on mouse ovarian cancer cell migration.

Activation of cPLA₂ involves phosphorylation by either extracellular signal-regulated kinase (ERK) or p38 mitogen-activated protein kinase (43). LPA (1 μmol/L) time-dependently induced phosphorylation of cPLA₂ in ID8 cells, which was inhibited by the MEK-ERK inhibitor U0126 (10 μmol/L; Supplementary Fig. S2A³; Fig. 3A, a), suggesting that ERK is the upstream kinase involved in cPLA₂ phosphorylation induced by LPA in ID8 cells. When tested in cell migration, the MEK-ERK inhibitor U0126 (10 μmol/L), but not an inhibitor for Akt (10 μmol/L), inhibited LPA-induced cell migration, suggesting that the MEK-ERK-cPLA₂ signaling pathway is involved in LPA-induced ID8 cell migration. Similar to human cells (2, 10, 11, 24), LY294002, a phosphatidylinositol 3-kinase inhibitor, also inhibited LPA-induced migration in ID8 cells (Fig. 3A, b).

To explore downstream effectors of cPLA₂, we first used some inhibitors. The general LOX inhibitor nordihydroguaiaretic acid (5 μmol/L), but not the COX-2 inhibitor NS398 (5 μmol/L), significantly inhibited LPA-induced migration of ID8 cells, suggesting that one or more LOX enzymes are involved in LPA-induced cell migration. Interestingly, COX-1, but not COX-2, is overexpressed in human ovarian cancer tissues and all major mouse ovarian cancer models (29–32). Therefore, we tested whether COX-1 was likely to be involved in LPA-induced cell migration. The COX-1 inhibitor SC-560 (2 μmol/L) partially inhibited LPA-induced cell migration, suggesting that COX-1 is involved in this activity. Next, specific LOX enzyme inhibitors were used to determine which LOX enzymes are involved in LPA-induced cell migration, and 2 μmol/L PD146176, AA-861, and baicalein were used to inhibit 15-LOX, 5-LOX, and 12-LOX, respectively. Our results suggested that 15-LOX, but not 5-LOX or 12-LOX, was involved in the effect of LPA. A combination of PD146176 and COX-1 specific inhibitor SC-560 (2 μmol/L for each inhibitor) almost completely blocked both basal level and LPA-induced cell migration (Fig. 3B), suggesting that each of the signaling pathways contribute partially to the activity and together they mediate the major effects of LPA on cell migration. To confirm these results, ID8 cells were transfected with siRNAs for COX-1 or 15-LOX. The reverse transcription-PCR results showed that the mRNA level of 15-LOX was diminished and the level for COX-1 was only partially reduced (Supplementary Fig. S2B).³ However, each of these siRNAs resulted in reduced cell migration (Fig. 3C). Combined siRNAs did not further reduce LPA-induced cells migration, which was more likely to be related to the transfection and down-regulation efficiency of these siRNAs. To examine whether human EOC cells also use the same or similar signaling pathways, the same set of inhibitors were used to pretreat SKOV3 cells before LPA-induced cell migration assays were conducted. As shown in Supplementary Fig. S2C,³ similar sensitivities to the inhibitors tested were also observed in SKOV3 cells, suggesting that ID8 is a good system to recapitulate human EOC cells in terms of LPA responsiveness.

LPA dose-dependently induced cell invasion of ID8 cells, which peaked at 1 to 4 μmol/L (Fig. 4A). Similar to LPA-induced cell migration, AACOCF3 also inhibited cell invasion induced by LPA (Fig. 4B). Similar to human EOC cells, ID8 cells expressed the three major LPA receptors, LPA_1-3 (Fig. 4C). A LPA_1/3-specific inhibitor, Ki16425, almost complete inhibited LPA-induced cell invasion in ID8 cells (Fig. 4D), suggesting that either LPA₁, LPA_3, or both of them are involved in LPA-induced invasion in these cells.

ID8 cells i.p. injected resulted in the formation of numerous tumors as reported previously (33). However, these tumors were small in size (<0.05 mm³). Daily injections of LPA (100 μmol/L × 200 μL PBS/mouse) for 8 weeks enhanced tumorigenesis of ID8 cells in C57/BL6 syngeneic and immunocompetent mice as evidenced by (a) early onset of initial tumors detected, (b) early onset of ascites formation (60-70 days in the LPA group versus 80-100 days in the PBS group) with larger ascites volumes, (c) formation of tumors that were >1 mm³, and (d) reduced mouse survival times (Fig. 5A-C; Table 1A). In the LPA group, 80% of the mice formed ascites at 77 ± 7 days compared with only 20% of mice formed ascites in the PBS group at the same time. Also, the volumes of ascites formed in the PBS group were less (10 ± 5 mL in the LPA group versus 6 ± 2 mL in the PBS group). However, at later days, most of the mice in the PBS group also developed ascites (n = 21 and 19 in the LPA and the PBS groups, respectively; Fig. 5A). Both groups presented tumors on the peritoneal wall, diaphragm, omentum, and mesentery. Tumors were also found on the surface of the spleen, liver, kidney, and small intestine in 10% mice from both groups of mice. Thus, LPA did not significantly affect the location of tumor formation. However, LPA did significantly stimulate tumor growth, with tumors >0.05 mm³ found only in the LPA group. Approximately 30% of tumors in the LPA-treated group were 3 ± 2 mm³, mainly located on the peritoneal wall and the diaphragm (Fig. 5B). The representative H&E staining of the tumors from the PBS and LPA groups are shown in Fig. 5C. In the PBS group, tumors were small and not highly invasive, although the mesothelial layer was disorganized (the normal tissue layers are shown by arrows and the tumor areas are labeled by “tumor”). In contrast, LPA-treated tumors were bigger and the tumor covered the entire slide.

To track the tumor development in living mice, we generated a luciferase-expressing ID8 cell line and used them in the i.p. injection mouse model. The mice were imaged every 2 weeks. At week 4 post-injection, tumor development was detected in the LPA group but not in the PBS group (Supplementary Figs. S3A and B).³ We found an excellent correlation between the intensity of the bioluminescent signal and the tumor load. Even very small tumors could be detected by the imaging. For the ID8 model, tumors mostly grew on the peritoneal wall, including the ventral (top) and both sides of the body, which were very well imaged. These results showed the promoting roles of LPA in tumorigenesis and/or metastasis. Enhanced tumorigenesis/metastasis likely results in early mouse death. Therefore, we determined mouse survival in LPA- and PBS-treated mice. Fifty percent of the mice receiving LPA died in 75 ± 3 versus 95 ± 5 days in the PBS group, suggesting LPA accelerated cancer-associated mouse deaths (Fig. 5D).

Although the i.p. injection model has been considered as an experimental metastasis model of EOC, it is hard or impossible to distinguish primary tumor and metastatic tumor growth in this model and the initial steps of EOC metastasis have been bypassed. Thus, we used an orthotopic mouse model to further test the effect of LPA. ID8 cells (5 × 10⁵) were orthotopically injected into the bursa of a mouse ovary and LPA was administered by daily i.p. injection for 4 weeks (Table 1B). We found that LPA treatment for a shorter time (4 versus 8 weeks) in the early stage of cancer development was sufficient to induce differences in tumorigenesis and/or metastasis. The orthotopic model results showed that LPA significantly affected tumor growth at the primary site and enhanced tumor metastasis to the distant organ, such as diaphragm, and ascites formation (n = 7 in the PBS group and n = 8 in the LPA group). Although tumor development in the orthotopic model was slower than that in the i.p. injection model, LPA still accelerated mortality of mice compared with the PBS group (50% of mouse death at 128 versus 160 days in the control group; Table 1B).

All cell types in the immune system, including T cells, B cells, natural killer cells, macrophages, and dendritic cells, participate in tumorigenesis and metastasis. Therefore, it is necessary to include these cells and to examine their effects in a preclinical cancer model. Because human xenografts cannot be done in immunocompetent mice, identification of mouse cells that closely resemble the properties and signaling pathways of their human counterparts is critical. In this work, we have compared the cellular functional responses to LPA in mouse EOC cells and human EOC cells. Both human and mouse EOC cells similarly responded to LPA in colony formation, cell migration, and cell invasion assays. In addition, the signaling pathways involved in LPA-induced cell migration are similar in ID8 and SKOV3 cells. In particular, we have found that both human SKOV3 EOC and mouse ID8 EOC cell lines require COX-1, but not COX-2, in LPA-induced cell migration. Although COX-2 has been shown to be involved in one human EOC cell line, Dov13 cells (15), the majority of human ovarian cancer tissues and mouse EOC models have elevated COX-1, but not COX-2, expression (29–32). These data suggest that the ID8 cell line is a suitable model for testing the effects of LPA on ovarian cancer cells and evaluating LPA-targeted therapeutics in immunocompetent mouse models.

We have used two models (i.p. and orthotopic injection) to observe the effects of LPA in vivo. Consistent with our xenographic models using human EOC cells (10, 11), our results suggest that LPA affects both tumorigenesis and metastasis in vivo. These results are also highly consistent with the effects of LPA on colony formation, cell migration, and invasion in vitro (44).

There are two major strategies to target LPA for EOC treatment. The first is to target LPA signaling pathways including its receptors and downstream signaling molecules and the second is to target LPA production and/or degradation. We have shown that, similar to human EOC cells, ID8 cells express at least three major LPA receptors, LPA_1-3, with one or more of these receptors likely to be involved in cell invasion induced by LPA (Fig. 4). To date, at least six LPA receptors have been identified (27). These receptors may have overlapping or even opposing effects in vitro and in vivo. Although many LPA receptor-specific antagonists have been developed or are under development, specifically targeting of these receptors in vivo is still challenging. Alternatively, downstream signaling pathways/molecules can be identified and tested for blocking the effects of LPA in vivo. For example, phosphatidylinositol 3-kinase has been considered to be an effective cancer target for LPA-induced and other factor-induced tumor development (45–47). We show here that phosphatidylinositol 3-kinase is also likely to be involved in LPA-induced migration of mouse EOC cells. Moreover, to our knowledge, our work is the first to show that COX-1 and 15-LOX are involved in LPA-induced migration of both human and mouse EOC cells. The potential complete inhibition of LPA-induced cell migration when both of these pathways are inhibited provides promising results that warrant further testing in vivo using the mouse model developed here. Targeting downstream signaling molecules may result in more specific inhibition of LPA effects related to tumor metastasis, whereas upstream targets, such as LPA receptors, may be involved in a broader cellular effects, some of which should not be inhibited in normal physiology. Interestingly, PLA₂s are not only involved in LPA signaling but also involved in LPA production (48). Thus, targeting these enzymes may present special advantages.

Ascites formation is a major burden for patients with ovarian cancer. It represents a tumor microenvironment and contains large number of tumor and other cells as well as promoting growth factors, cytokines, and lipids factors. We present here the first evidence that LPA has a promoting activity in ascites formation in vivo. This activity may be related to its ability to stimulate VEGF secretion, because VEGF has been shown to be an important factor for the regulation of blood vessel permeability and ascites formation (49, 50). Therefore, the mouse model established here may be used to study the mechanisms of ascites formation in vivo.

In summary, we have established a mouse model using ID8 to test the in vivo effect of LPA on EOC. Both our in vitro and in vivo results suggest that ID8 cells closely recapitulate responses to LPA. In addition, the similarity in the effects of LPA on tumorigenesis and metastasis observed in immunocompromised and immunocompetent mice suggest that the promoting functions of LPA are not diminished by the presence of endogenous T or B cells. Thus, this is a valuable system to identify and test therapeutics that target LPA in EOC and to study the mechanisms by which ascites forms in EOC.

No potential conflicts of interest were disclosed.

The imaging work was supported in part by the Indiana Genomics Initiative. The Indiana Genomics Initiative is supported in part by the Lilly Endowment.