Abstract
Endothelial differentiation gene (edg)-encoded G protein-coupled receptors (Edg Rs)-1, -3, and -5 bind sphingosine 1-phosphate (S1P), and Edg-2 and -4 bind lysophosphatidic acid (LPA). Edg Rs transduce signals from LPA and S1P that stimulate ras- and rho-dependent cellular proliferation, enhance cellular survival, and suppress apoptosis. That high levels of LPA in plasma and ascitic fluid of patients with ovarian cancer correlate with widespread invasion suggested the importance of investigating expression and functions of Edg Rs in ovarian cancer cells (OCCs) as compared with nonmalignant ovarian surface epithelial cells (OSEs). Analyses of Edg Rs by semiquantitative reverse transcription-PCR, a radioactively quantified variant of PCR, and Western blots developed with monoclonal antibodies showed prominent expression of Edg-4 R in primary cultures and established lines of OCCs but none in OSEs. In contrast, levels of Edg-2, -3, and -5 were higher in OSEs than OCCs. LPA stimulated proliferation and signaled a serum response element-luciferase reporter of immediate-early gene activation in OCCs but not OSEs, whereas S1P evoked similar responses in both OSEs and OCCs. Pharmacological inhibitors of Edg R signaling suppressed OCC responses to LPA. A combination of monoclonal anti-Edg-4 R antibody and phorbol myristate acetate, which were inactive separately, evoked proliferative and serum response element-luciferase responses of OCCs but not OSEs. Thus the Edg-4 R may represent a distinctive marker of OCC that transduces growth-promoting signals from the high local concentrations of LPA characteristic of aggressive ovarian cancer.
INTRODUCTION
The lysolipid phosphoric acid mediators LPA 3 and S1P are generated enzymatically from membrane lipid precursors in many different types of normal and malignant cells (1, 2, 3) . Extracellular LPA and S1P both stimulate cellular proliferation and differentiation, enhance cellular survival, and evoke specific cellular functional responses such as migration and adhesion (3, 4, 5) . A subfamily of at least five G protein-coupled Edg Rs, which are encoded by edgs, bind and transduce signals from LPA or S1P (3 , 6, 7, 8, 9, 10) . Two homology clusters, defined by highest amino-acid sequence identity and specificity for the same ligand, consist of the Edg-1, -3, and -5 set of S1P Rs and the Edg-2 and -4 LPA receptors. LPA and S1P stimulate cellular proliferation directly by eliciting the SRF and TCF, which together bind to and transcriptionally activate the SRE in promoters of many immediate-early genes (11) . LPA and S1P also stimulate cellular proliferation by increasing secretion and effectiveness of autocrine polypeptide growth factors (12 , 13) .
The potential importance of LPA in ovarian cancer growth and tissue invasion has been suggested by findings of high concentrations in the ascitic fluid of many patients with local metastases, in which it is a major growth factor, and in the plasma of patients with widespread ovarian cancer (14 , 15) . However, neither the patterns of expression of Edg Rs on OCCs nor the types of distinctive Edg R-mediated signals to OCCs have been elucidated to define further the roles of LPA in ovarian tumor growth, local spread, and metastasis. Thus, parallel studies were begun to examine the role of Edg Rs in LPA signaling of the proliferation of OCCs (16) and of the induction of expression of vascular endothelial growth factors by OCCs (17) . The present data show different levels of expression of Edg-1, -2, -3, and -5 Rs on human OCCs in primary cultures, established lines of human OCCs, normal OSEs, and IOSEs, but expression of Edg-4 Rs only by the OCCs. It is further shown that LPA stimulates proliferation, biochemical signaling, and secretion of IGF-II in OCCs but not in IOSEs, and that Edg-4 Rs transduce these responses in OCCs but not in IOSEs.
MATERIALS AND METHODS
Chemical Reagents and Antibodies.
The sources of chemicals were: (for S1P) Biomol, Plymouth Meeting, PA; (for 1-oleoyl-LPA, PMA, and faf-BSA) Sigma Chemical Co., St. Louis, MO; and (for human IGF-II) Peprotech, Inc., Rocky Hill, NJ. Ovarian cells were treated with: (a) PTX (Calbiochem, Inc., La Jolla, CA); (b) recombinant Clostridium botulinum C3 ADP-ribotransferase (C3 exoenzyme, List Biological Laboratories, Inc., Campbell, CA), which ADP-ribosylates rho specifically; and (c) the MEK inhibitor 2′-amino-3′-methoxy-flavone (PD98059, Calbiochem) as described previously (18 , 19) . Mouse monoclonal antibodies specific for substituent peptides of human Edg-3 (amino acids 1-21), Edg-4 (9-27), and Edg-5 (303-322) were generated, purified, and used to develop Western blots at 0.1- 0.3 μg/ml (18 , 19) . The cross-reactivity of each antibody with heterologous Edg proteins was less than 1%, as determined by Western blots of 0.1-100 μg of membrane proteins isolated from HTC4 rat hepatoma cells stably transfected with human Edg-2, -3, -4, or -5 (18 , 19) . A mouse monoclonal IgG1 that specifically neutralizes activity of human/rat IGF-II but not IGF-I was purchased (Upstate Biotechnology, Inc., Lake Placid, NY). A rabbit polyclonal antiserum to rodent and human Edg-2 was kindly provided by Dr. Jerold Chun (University of California, San Diego, CA).
Cell Culture and Quantification of Cellular Proliferation.
OV202 primary cultures of human OCCs (Ref. 20 ; cell culture passage 13), the IOSE 29 and IOSE 80 lines of SV40 T antigen-immortalized nonmalignant human ovarian surface epithelial cells, and OSE normal human ovarian surface epithelial cells (from Dr. Nelly Auersperg, University of British Columbia, Vancouver, Canada; Ref. 21 ) and the OCC lines OVCAR3 (from American Type Culture Collection, Rockville, MD), DOV13 (from Dr. Robert Bast, M. D. Anderson Cancer Center, Houston, TX), and A2780 (from Dr. Thomas C. Hamilton, Fox Chase Cancer Center, Philadelphia, PA) were cultured in α-MEM, containing Earle’s salts and nucleosides, with 20% FBS, 4 mm l-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. RNA was extracted from OSE by Y-L. H. Serum-deprived conditions were attained by culture in α-MEM with 2% FBS for 12 h followed by α-MEM: Waymouth’s medium (1:1, v:v) without serum for 12 h, prior to the introduction of stimuli in the serum-free latter medium with 0.1 mg/ml faf-BSA (MEM-faf-BSA). To assess proliferation, replicate layers of 0.25 × 104 OV202 or IOSE 29 cells per well of 96-well flat-bottomed plates were serum-deprived and cultured in 100 μl of MEM-faf-BSA. Some cells were treated with C3 exoenzyme for 30 h or MEK inhibitor for 2 h before the addition of PMA, a lysophospholipid, and/or an anti-Edg receptor antibody, followed by incubation for 24, 48, 72, or 96 h. In the longer incubation intervals, LPA and S1P were added every 24 h. Cells received 1 μCi of [3H]thymidine (specific activity = 81 Ci/mmol; NEN-DuPont, Inc., Boston, MA) 24 h before harvesting for β -scintillation counting. Ovarian cell proliferative responses to stimulation are expressed as the fold-increase in cpm above the control value obtained for medium alone or as a percentage of the control. In some studies, cells were harvested, fixed in 5% buffered formalin, and stained with H&E before quantification by microscopic counting of 10 1-mm3 fields in a hemocytometer.
RT-PCR Analysis of Edg Rs.
Total cellular RNA was extracted by the TRIzol method (Life Technologies, Inc., Grand Island, NY) from suspensions of OSEs, IOSEs, OCCs, and four lines of stably transfected rat HTC4 hepatoma cells. Each HTC4 cell line had very low background expression of native Edg Rs but overexpressed one recombinant human Edg R. RT-PCR was performed as described previously (18 , 19) . Two μCi of [α-32P]dCTP were added to some sets of reaction mixtures to allow radioactive quantification of the mRNA encoding each Edg R relative to that of the standard G3PDH (18 , 19) . The sequences of oligonucleotide primer pairs have been provided previously (18 , 19) . PCR products were resolved by electrophoresis in a 2-g/100-ml agarose gel with ethidium bromide staining. G3PDH and Edg R cDNA bands were cut from gels and solubilized for β -scintillation counting in 0.5 ml of sodium perchlorate solution at 55°C for 1 h (Elu-Quick, Schleicher and Schuell, Keene, NH). Relative quantities of cDNA encoding each Edg R were calculated by the ratio of radioactivity to that in the corresponding G3PDH band (18 , 19) .
Western Blots.
Replicate suspensions of 1 × 107 ovarian cells that had been incubated without or with LPA or S1P for 16 h were washed three times with 10 ml of ice-cold Ca2+- and Mg2+-free PBS, resuspended in 0.3 ml of cold 10 mm Tris-HCl (pH 7.4) containing a protease inhibitor cocktail (Sigma Chemical Co.), 0.12 m sucrose, and 5% glycerol (v:v), and disrupted and extracted for proteins as described previously (18 , 19) . Aliquots of supernatant containing 1-100 μg of protein were mixed with 4× Laemmli’s solution, heated to 100°C for 3 min, and electrophoresed in a SDS-12% polyacrylamide gel for 20 min at 100 v and 1.5 h at 140 v, along with a rainbow-prestained set of molecular weight markers (NEN-DuPont, or Amersham, Inc., Arlington Heights, IL). Proteins in each gel were transferred electrophoretically to a nitrocellulose membrane (Hybond, Amersham) for sequential incubation with:
(a) 5 g/100 ml reconstituted nonfat milk powder to block unspecific sites;
(b) dilutions of mouse monoclonal anti-Edg R antibody or rabbit anti-Edg-2 R antibody; and
(c) horseradish peroxidase-labeled goat antimouse or antirabbit IgG before development with a standard ECL kit (Amersham).
RIA Quantification of IGF-II.
RIAs were conducted according to the instructions of Research and Diagnostic Antibodies, Inc. (Berkeley, CA), after removal of some IGF-binding proteins by Sep-Pak chromatography (Millipore Corp., Milford, MA), as directed (13) . Replicate layers of 1.5-1.8 × 106 serum-deprived OV202 or IOSE 29 cells in 1 ml of MEM-BSA were cultured in six-well plates for 24 h with each stimulus before harvesting and before processing the supernatant fluid.
Transfections and SRE Reporter Assay.
Replicate suspensions of 0.3-1 × 105 ovarian cells in 1 ml of α-MEM-2% FBS were layered in 12-well plates at 40-50% confluency and incubated for 4 h. The monolayers were washed twice, covered with 1 ml of serum-free MEM-faf-BSA, and lipotransfected with 100 ng/well SRE firefly luciferase reporter plasmid (7) and 5 ng/well pRL-CMV Renilla luciferase vector (Promega, Madison, WI) using FuGENE 6 (Boehringer-Mannheim Corp., Indianapolis, IN). After 20 h of incubation, the medium was replaced with fresh MEM-faf-BSA, and PMA, antibodies, and/or lysophospholipids were added in MEM with 0.1 mg/ml faf-BSA, as for the proliferation assay. Similarly, some wells were pretreated with PTX for 6 h, C3 exoenzyme for 30 h or MEK inhibitor for 2 h. After 4 h of incubation at 37°C, the luciferases were extracted in Reporter Lysis Buffer (Promega), and their activities were quantified sequentially by luminometry using Luciferase Assay and Stop & Glo reagents (Promega), with integration of light emitted during the 15 s after the addition of each reagent (EG & G Berthold microplate luminometer, model LB96V). Each firefly luciferase value was corrected for any difference in apparent cell number or transfection efficiency by division with the Renilla luciferase signal from the same well.
RESULTS
Semiquantitative RT-PCR analyses of mRNA derived from OCCs, OSEs, and IOSE 29 showed that Edg-4 and Edg-2 LPA Rs were the predominant members of this subfamily of G protein-coupled receptors (Fig. 1 ⇓ ; Table 1 ⇓ ). The most striking finding was that Edg-4 R mRNA is expressed at high levels in OV202 primary cultures of OCCs (Fig. 1) ⇓ and several established lines of OCCs (Table 1) ⇓ but was not detected in IOSE 29 (Fig. 1) ⇓ or in OSEs (Table 1) ⇓ . In contrast, levels of mRNA encoding Edg-2 R in OV202 primary cultures of OCCs and different lines of OCCs encompassed a broad range, but they often were even higher in OSE and IOSE lines (Table 1) ⇓ . For further contrast, levels of mRNA encoding Edg-3 and Edg-5 S1P Rs were higher in OSE and IOSE lines than in the OCCs, whereas levels of mRNA encoding the Edg-1 S1P R in OCCs were similar to those in OSEs and IOSEs (Table 1) ⇓ . The highest levels of Edg-4 in OCCs and of Edg-2 in OSEs and IOSEs were similar to those found in transfectants selectively expressing each recombinant R.
RT-PCR semiquantification of mRNA encoding Edg Rs in human ovarian cells. The volume of cDNA mixture generated by RT of RNA from IOSE 29 cells (A) and OV202 cells (B) was selected to equalize the level of amplified G3PDH cDNA product. Lanes 1-5 in each frame show the cDNA of Edg-1 to Edg-5 receptors, respectively. Number below each lane, the ratio of 32P in cDNA for an Edg R to that for G3PDH.
Relative levels of mRNA encoding Edg Rs in ovarian cells
The results of Western blot analyses of Edg R proteins supported the findings of a much higher level of Edg-4 LPA Rs in OV202 primary OCCs than in IOSE 29 cells and of higher levels of Edg-3 and Edg-5 S1P Rs and Edg-2 LPA Rs in IOSE 29 than in OV202 cells (Fig. 2) ⇓ . This latter difference was not predicted by the findings of only slightly higher levels of Edg-2 R mRNA in IOSEs than these same lines of OCCs (Fig. 1 ⇓ ; Table 1 ⇓ ). Northern and Western blots also revealed high levels of Edg-4 R mRNA and protein in OVCAR-3 and DOV-13 OCCs, as contrasted with barely detectable levels in IOSE 29 and IOSE 80 cells (17) .
Western blot analysis of the expression of Edg-2, -3, -4, and -5 Rs by OV202 and IOSE 29 cells. The three samples analyzed for content of each Edg R are: T, 3 μg of protein from HTC4 rat liver cells that were stably transfected with the respective Edg Rs; OV202, 10 μg of protein from OV202 OCCs; and IOSE, 10 μg of protein from IOSE 29-transformed human ovarian epithelial cells. No Edg-3, -4, or -5 R proteins were detected in untransfected HTC4 cells with monoclonal antibodies, but a low level of Edg-2 R was revealed by the rabbit anti-Edg-2 R antibodies. Marginal lines, positions of the Mr 45,000 and 66,000 protein molecular weight standards. This result is representative of a total of three such studies.
The capacity of the Edg-2 and -4 R ligand LPA and of the Edg-1, -3, and -5 R ligand S1P to stimulate ovarian cell growth and the secretion of cancer-related proteins was examined in three independent assay systems: (a) DNA synthesis; (b) SRE-driven reporter gene activity; and (c) production of IGF-II, which is a potent mitogen for ovarian cells. DNA synthesis, assessed by uptake of [3H]thymidine, was stimulated by LPA in OV202 primary OCCs but not in IOSE 29 cells (Fig. 3) ⇓ . OV202 cell uptake of [3H]thymidine was increased significantly by 10−6 m but not by 10−8 m LPA after 1 day of incubation but was increased significantly by both concentrations of LPA after 3 and 5 days, respectively. The increase achieved by 10−6 m LPA after 5 days was similar in magnitude to that attained by FBS. IOSE 29 cell uptake of [3H]thymidine decreased on days 3 and 5 in the absence of LPA to respective mean levels of 78 and 19% of that seen on day 1 (Fig. 3) ⇓ . IOSE 29 cell uptake of [3H]thymidine was increased marginally on day 5 by both concentrations of LPA, but did not respond on days 1 or 3. Only FBS increased uptake of [3H]thymidine by IOSE 29 cells significantly on day 5 by over 20-fold. Cell counts changed in parallel with uptake of [3H]thymidine in these experiments. On days 3 and 5, respectively, OV202 mean cell counts were 2.2-fold and 5.1-fold higher than on day 1 in the absence of LPA and were increased by respective means of 63 and 102% by 10−6 M LPA on days 3 and 5, and by 132% by FBS on day 5. Under the same conditions, IOSE 29 mean cell counts on days 3 and 5 were only 84 and 45%, respectively, of that on day 1 with no significant effect of LPA and a mean increase by FBS on day 5 of 8.4-fold.
Stimulation of proliferation of OV202 and IOSE 29 cells by LPA. Each column, the mean of the results of two studies carried out in duplicate; bar, the range. The mean control uptake of [3H]thymidine by OV202 cells was 1,348 and 2,338 cpm at day 1; 5,217 and 9,581 cpm at day 3; and 17, 212 and 31,473 cpm at day 5. For IOSE 29 cells, the mean control uptake of [3H]thymidine was 2,288 and 4,339 cpm at day 1; 1,621 and 3,666 cpm at day 3; and 436 and 814 cpm at day 5. The significance of LPA-induced increases relative to medium alone after each incubation interval: +, P < 0.05; *, P < 0.01.
SRE-luciferase activity of OV202 cells, which represents one proliferation-related index of immediate-early gene responses to Edg R signaling, was stimulated significantly by 10−9 m to 10−6 m LPA, whereas that of IOSE 29 cells did not respond to LPA (Fig. 4) ⇓ . The SRE-luciferase responses of OVCAR-3 OCCs resembled those of OV202 cells, with significant increases to 10−7 and 10−6 m LPA up to a mean maximal level of over 300% of control (17) . In contrast, SRE-luciferase activity of IOSE 29 cells was enhanced significantly and to a greater extent than that of OV202 cells by 10−9 m to 10−7 m S1P (Fig. 4 ⇓ ; 17). IOSE 80 cell SRE-luciferase activity also responded to S1P, but not to LPA, (n = 1, data not shown). IGF-I production by IGFR-1-bearing OCC establishes an autocrine growth loop under serum-free conditions (22) . IGF-II also binds to IGFR-1 with resultant mitogenesis. OV202 cells (19) , but not IOSE 29 cells, produce IGF-II in culture in amounts sufficient to attain ng/ml concentrations. IGF-II generation by OV202 cells was increased 10-fold or more by 10− 8 m and 10−7 m LPA, and 9-fold or more by 10−8 m and 10−7 m S1P (Fig. 5) ⇓ .
SRE-luciferase reporter assay of nuclear signaling of OV202 and IOSE 29 cells by LPA and S1P. Each column, the mean of the results of three studies; bar, ±SD. The medium-alone control values were 3,626, 12,770, and 7,250 luminometer units for OV202 cells and 2,906, 862, and 4,026 for IOSE 29 cells. The levels of significance of increases above medium controls were determined by a paired Student’s t test: +, P < 0.05; *, P < 0.01.
LPA and S1P enhancement of the secretion of IGF-II by OV202 cells. Each column, the mean of the results of two studies performed in triplicate; bar, ± the range. The mean concentration of radioimmunoreactive IGF-II in media from OV202 cells incubated for 24 h without stimulation in the two studies was 4.6 ng/ml and 9.7 ng/ml. Mean increases in IGF-II were significant (P < 0.05) by a paired t test at 10−8 m, 10−7 m, and 10−6 m LPA and at 10− 8 m and 10−7 m S1P.
Biochemical signal transduction required for coupling Edg-2 and/or -4 LPA Rs and Edg-1, -3, and/or -5 S1P Rs to enhancement of SRE-luciferase activity next were examined with pharmacological inhibitors selective for distinct components of the signaling pathways (Table 2) ⇓ . Suppression of the Gi-ras-MAP kinase cascade by either PTX or the MEK inhibitor, each of which prevents increases in functional TCF, and suppression of rho by C3 exoenzyme, which reduces activation of SRF—all diminished LPA- and S1P-induced enhancement of SRE-luciferase activity in OV202 cells and OVAR-3 OCCs (17) . This is the pattern of susceptibility expected for Edg R signaling through the SRF-TCF-SRE ternary complex, which initiates transcription of numerous immediate-early genes.
Pharmacological inhibition of LPA and S1P signaling to the SRE-luciferase reporter in OV202 human primary OCCs
Although highly selective agonists and antagonists for Edg Rs are not available presently, a mouse monoclonal anti-Edg-4 R antibody to a substituent peptide of the extracellular NH2-terminal domain of Edg-4 initiates signals to the SRE-luciferase reporter in OCCs primed by concurrent exposure to PMA. Neither PMA nor the anti-Edg-4 R antibody separately elicited a consistent SRE-luciferase response. Together, however, 0.1-3 ng/ml of PMA and 0.1 μg/ml or 1 μg/ml of anti-Edg-4 R antibody increased SRE-luciferase activity significantly in OV202 cells to maximal responses at 0.1 μg/ml anti-Edg 4 R antibody with 1 ng/ml PMA and 1 μg/ml anti-Edg-4 R antibody with 1 ng/ml or 3 ng/ml PMA (Fig. 6) ⇓ . A2780 OCCs responded significantly and similarly to OV202 cells to anti-Edg 4 R antibody plus PMA (n = 1, data not shown). In contrast, IOSE 29-cell SRE-luciferase activity did not respond either to any combination of PMA and anti-Edg-4 antibody or substantially to LPA (Fig. 6) ⇓ . SRE-luciferase of IOSE 80 cells similarly failed to respond to PMA and anti-Edg-4 antibody. However, IOSE 29-cell SRE-luciferase activity did respond significantly to polyclonal anti-Edg-2 R antibody with PMA, confirming expression of Edg-2 Rs and intact signaling pathways that apparently are not accessible to or responsive to extracellular LPA (Fig. 6) ⇓ . In contrast, OV202 SRE-luciferase showed no increase after the addition of 1 and 3 μg/ml anti-Edg 2 R antibody with 1 and 3 ng/ml of PMA in the same protocol, possibly because of their low level of Edg-2 Rs (Fig. 2) ⇓ . With 3 ng/ml PMA, 3 μg/ml anti- Edg 2 R antibody suppressed the SRE-luciferase activity of OV202 cells to a mean of 62% of the control level (P < 0.05). Thus, immunostimulation of SRE-luciferase in these cell lines was characterized by OCC responses through Edg-4 R and by IOSE responses through Edg-2 R.
Edg-4 R-mediated specific immuno-stimulation of SRE-luciferase activity in OV202 cells. Each column, the mean of the results of three separate studies; bar, ± SD. The significance of each increase in signaling observed was calculated relative to the medium-alone control for each type of cell (100%), as described for Fig. 4 ⇓ , and is shown by the same symbols except for #, P = 0.05.
Proliferation of OV202 cells, but not of IOSE 29 cells, also was enhanced by anti-Edg-4 antibody and PMA in combination (Fig. 7) ⇓ . As was demonstrated, 10− 8 M to 10−6 M LPA evoked [3H] thymidine uptake by OV202 cells but not by IOSE 29 cells (Fig. 3) ⇓ . Neither concentration of anti-Edg-4 antibody alone altered [3H]thymidine uptake, whereas PMA alone led to slight suppression. In combination with 1 and 3 ng/ml PMA, both concentrations of anti-Edg-4 antibody elicited substantial proliferative responses of OV202 cells, but not of IOSE 29 cells, after 48 h (Fig. 7) ⇓ . In contrast, S1P evoked greater uptake of [3H]thymidine by IOSE 29 cells than by OV202 cells.
Edg-4 R-mediated immunostimulation of the proliferation of OV202 cells. Each value is the mean ± the range of results of two studies carried out in triplicate and expressed as a percentage of the medium control. Mean control unstimulated values in medium alone were 1916 and 2784 dpm for OV202 cells and 1494 and 1688 dpm for IOSE 29 cells after 48 h. The concentrations of antibody are shown as μg/ml.
DISCUSSION
The patterns of Edg R expression by human ovarian cells from diverse sources suggest that the Edg-4 LPA R appears in association with malignant transformation (16 , 17) . Edg-4 LPA R was the quantitatively predominant Edg R in OV202 primary cultures of OCCs and in several different established lines of OCC, as assessed in terms of both mRNA and protein antigen, whereas no Edg-4 was detected in either IOSE 29 SV40 T antigen-immortalized nonmalignant OSE or primary cultures of normal OSE (Figs. 1 ⇓ and 2 ⇓ ; Table 1) ⇓ . In contrast, IOSE 29 and OSEs had higher levels of Edg-2 LPA Rs and of Edg-3 and Edg-5 S1P Rs than any of the OCCs. The most striking functional correlate of the different patterns of Edg R expression was that OCCs, but not IOSEs, responded to LPA.
The difference in responsiveness to LPA was observed in three distinct systems, using primary cultures of OV202 OCCs (16 , 20) , established lines of OCCs, and IOSE 29 cells (17 , 21) , that grow more rapidly and reliably than normal OSEs. LPA stimulated the proliferation of OV202 OCCs, but not of IOSE 29 cells, as assessed by the uptake of [3H]thymidine (Fig. 3) ⇓ and cell counts. In two other established OCC lines, namely DOV13 and A2780, LPA also enhanced proliferation (data not shown). Similarly, LPA increased SRE-luciferase activity in OV202 OCCs but not in IOSE 29 cells, whereas S1P evoked greater responses in IOSE 29 cells than in OV202 OCCs (Fig. 4) ⇓ . LPA also stimulated SRE-luciferase activity in the same established lines that responded proliferatively (data not shown). LPA augmented the secretion of IGF-II by OV202 OCCs, as did S1P (Fig. 5) ⇓ , but IOSE 29 cells did not secrete detectable IGF-II before or after stimulation (not shown). Thus in each functional assay, LPA effects were restricted to OCCs, whereas S1P had effects on IOSE 29 cells that were similar to or greater than the effects on OCCs (Fig. 3 ⇓ 4 ⇓ 5) ⇓ . Expression of Edg-4 Rs exclusively by OCCs led to the hypothesis that Edg-4 Rs, but not Edg-2 Rs, transduced effects of LPA on ovarian cells. The Edg-4 R, but not Edg-2 R, also mediated the induction of vascular endothelial growth factor by OCCs and not by OSEs (17) .
The susceptibility of LPA- and S1P-evoked SRE-luciferase responses of OV202 OCCs to pharmacological inhibitors acting on known components of Edg R signaling pathways is consistent with mediation by Edg Rs (Table 2) ⇓ . However, confirmation of the assumed central role of the Edg-4 R in the transduction of signals from LPA to OCCs required further investigation. The incubation of OV202 cells with concentrations of a non-cross-reactive anti-Edg-4 antibody that lacks effect alone, together with inactive or only marginally active levels of PMA, elicited significant increases in SRE-luciferase reporter responses up to levels equal to or exceeding those attained by 10−7 M LPA (Fig. 6) ⇓ . OV202 SRE-luciferase activity did not respond to polyclonal anti-Edg-2 antibody, possibly because of the low level of expression of Edg-2 Rs (Fig. 2) ⇓ . Proliferative responses of OV202 cells, quantified by the increased uptake of [3H] thymidine, similarly were evoked by a combination of inactive concentrations of monoclonal anti-Edg-4 antibody and PMA (Fig. 7) ⇓ . In contrast, IOSE 29 cells failed to respond to the same combination of anti-Edg-4 R antibodies and PMA but responded significantly when challenged with a combination of polyclonal anti-Edg-2 antibodies and PMA (Fig. 6) ⇓ . The patterns of response of OCCs and IOSEs to anti-Edg R antibodies are consistent with the concept that functional Edg-4 R is present only in OCCs, but not OSEs. This is the first demonstration of selective expression of an Edg R in relation to a malignant phenotype.
The reason for the lack of LPA recognition and/or transduction of signals from LPA by the Edg-2 Rs of IOSE 29 cells has not been elucidated. Several possible explanations for the lack of Edg-2 R-mediation of LPA responses in OSE cells are suggested by the results of ongoing studies of Edg R coupling to G proteins and consequent patterns of signaling. Recombinant Edg-2 couples principally to Gi, whereas Edg-4 associates with Gq and Gi, and both may be required for an optimal response (23) . Perhaps because of this difference in pairing with G proteins, Edg-2 R-mediated mobilization of intracellular Ca2+ is nearly completely inhibited by PTX, whereas Edg-4 R-transduced Ca2+ responses are incompletely inhibited (24) . In contrast, rho inactivation by C3 exoenzyme more completely reduces Edg-4 R-dependent responses than those transduced by Edg-2 Rs. Other recent results of studies of overexpression of Edg Rs in OCCs have shown that a high level of Edg-2 R suppresses proliferation and promotes apoptosis, whether the Edg-2 Rs are native or recombinant and introduced by transfection (25) . This suppressive effect of high levels of Edg-2 Rs may be independent of LPA and expressed in the presence of Edg-4 Rs. Perhaps the high level of Edg-2 Rs in OSEs, unopposed by Edg-4 Rs, may blunt any positive response to LPA mediated by the Edg-2 Rs alone. That anti-Edg-2 R antibodies but not LPA evoke a positive nuclear response and proliferation in OSEs (Figs. 6 ⇓ and 7) ⇓ may reflect the bivalent presentation and bridging of Edg-2 Rs by the antibodies or the engagement of a site of interaction different from that characteristic of LPA binding. Further definition of these and other mechanisms will require studies of sets of OSEs and OCCs expressing different ratios of Edg-2:Edg-4 Rs and the application of monovalent Fab as well as polyvalent presentations of LPA.
That the Edg-4 LPA R expression by OCCs in primary and established cultures sharply distinguishes these malignantly transformed cells from normal and nonmalignant IOSEs suggests that as clear a contrast will be found in the equivalent cells in tumors. If ongoing immunohistological analyses of biopsies and surgically resected specimens of ovarian cancers and benign ovarian lesions confirm the basic observation in situ, then the Edg-4 LPA R may be considered a candidate marker for malignant ovarian tumors that is worthy of additional clinical studies.
The finding of Edg-4 Rs on ovarian cancer cells already asserts two points of clinical significance. First, the expression of active Edg-4 LPA Rs creates an autocrine growth factor system capable of promoting ovarian cancer growth because many OCCs produce LPA in amounts that result in functionally relevant local concentrations. The stimulation of ovarian cancer cell proliferation by LPA involves both the direct nuclear signaling of transcription of immediate-early growth-related genes through one or more copies of SRE in their promoters and the enhancement of production of autocrine polypeptide growth factors such as IGF-II (Fig. 5) ⇓ . Epidemiological studies will be important to determine whether there is any prognostic value in finding a low or high level of expression of Edg-4 in ovarian cancers. Second, Edg-4 LPA Rs may have therapeutic as well as diagnostic potential. Classical pharmacological antagonists of Edg-4 Rs may both block LPA direct growth effects and reduce the growth-promoting effects of LPA mediated by increased levels of one or more polypeptide growth factors. Monoclonal antibodies to Edg-4 Rs may permit selective local delivery of conjugates with covalent antagonists of Edg-4 Rs or with chemotherapeutic agents too toxic for systemic administration.
Acknowledgments
We thank Bethann Easterly for expert preparation of graphics.
Footnotes
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
↵1 Supported by Grant HL31809 from NIH and CRP 1 PF0265 from the California Department of Health Services (to E. J. G.).
-
↵2 To whom requests for reprints should be addressed, at Immunology and Allergy Research, UB8B, Box 0711, University of California Medical Center, 533 Parnassus, San Francisco, CA 94143-0711. Phone: (415) 476-5339; Fax: (415) 476-6915; E-mail: egoetzlitsa.ucsf.edu.
-
↵3 The abbreviations used are: LPA, lysophosphatidic acid; S1P, sphingosine 1-phosphate; SRF, serum response factor; TCF, ternary complex factor; SRE, serum response element; OCC, ovarian cancer cell; OSE, ovarian surface epithelial cell; IOSE, immortalized OSE; IGF-I, type I insulin-like growth factor; IGF-II, type II insulin-like growth factor; IGFR-1, type 1 receptor for IGF-I and IGF-II; edg, endothelial differentiation gene; Edg R, edg-encoded receptor; FBS, fetal bovine serum; PMA, phorbol myristate acetate; faf-BSA, fatty acid-free BSA; PTX, pertussis toxin; MAP, mitogen-activated protein; MEK, MAP kinase kinase; RT, reverse transcription; α-MEM, α-minimum essential medium; G3PDH, glyceraldehyde 3-phosphate dehydrogenase.
- Received April 9, 1999.
- Accepted August 19, 1999.
- ©1999 American Association for Cancer Research.
References
Volume 59, Issue 20
