Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

LPA and S1P are bioactive lysolipids that exert many of their effects and signaling activities through G protein-coupled receptors (GPCRs) (Goetzl and An, 1998; Moolenaar, 1999; Spiegel, 1999). We have reported previously that both LPA and S1P are important signaling molecules in ovarian cancer, regulating both growth and metastatic potentials of ovarian cancer cells (Xu et al., 1995a,b, 1998, 2001; Hong et al., 1999; Schwartz et al., 2001). We have detected both of these lysolipids in ascitic fluids in patients with ovarian cancer (Xiao et al., 2000, 2001). Moreover, we have reported that LPA is elevated in the plasma of patients with ovarian cancer but not in that of patients with breast cancer or leukemia, indicating its potential as a marker for ovarian cancer (Xu et al., 1998). LPA has been reported to have a cytoprotective effect in HEY ovarian cancer cells exposed to cis-diamminedichloroplatinum (Frankel and Mills, 1996). Furthermore, under certain conditions in vitro, ovarian cancer cells produce LPA (Shen et al., 1998; Eder et al., 2000), suggesting that LPA, and potentially S1P, function as autocrine growth factors in ovarian cancer.

LPA and/or S1P have been shown to activate extracellular signal regulated kinase (ERK) and PI3-K and/or Akt (PKB) via a PTX-sensitive pathway in a number of cell types (Marte and Downward, 1997; Weiner and Chun, 1999; Fang et al., 2000; Lee et al., 2000; Xu et al., 2001). A G_i-dependent ERK activation is essential for the mitogenic activity of LPA in fibroblasts (Van Corven et al., 1993; Fang et al., 2000). PI3-K/Akt signaling is involved in cell survival in many cellular systems and cancers (Marte and Downward, 1997; Liu et al, 1998; Yuan et al., 2000). The activation of Akt, an antiapoptotic protooncogene, is mediated by PI3-K, after receptor stimulation. PI3-K has also been shown to be the upstream activator of ERK and p38 MAPK (Krump et al., 1997; Lopez-Ilasaca et al., 1997). However, the potential interactions between ERK/p38 and Akt activation have just begun to be revealed.

Two different kinases, PDK1 and PDK2, are responsible for the phosphorylation and activation of Akt at T308 and S473, respectively. PDK1 has been cloned and sequenced (Alessi et al., 1997). The mechanism by which S473 undergoes phosphorylation remains obscure. It has been proposed that S473 can be both autophosphorylated and phosphorylated by other kinases, such as PDK1 and integrin-linked kinase-1 (ILK1), which may be promoted by interactions between PDK1 and other kinases associated with Akt (reviewed in Chan and Tsichlis, 2001). In vitro phosphorylation of Akt at S473 by MAPK activated protein kinase-2 (MK2), a downstream target of p38, has been reported previously, although it is not involved in the in vivo S473 phosphorylation induced by insulin (Alessi et al., 1996). Recently, Rane et al. (2001) have shown that in neutrophils, p38, but not ERK, activation is required for Akt S473 phosphorylation induced by fMLP, Fc-R cross-linking, or phosphatidylinositol-3,4,5-trisphosphate (PIP₃), and MK2 functions as PDK2 to phosphorylate Akt at S473 in vivo (Rane et al., 2001). Apparently, more than one molecular identity may function as PDK2 to phosphorylate Akt at S473, and this may be cell type- or stimulus-dependent.

We describe herein a series of studies examining the signaling mechanisms of LPA- and S1P-induced activation of the PI3-K/Akt pathway in HEY ovarian cancer cells and a panel of other cell lines. In this study, we have focused on the signaling mechanisms of LPA/S1P-induced S473 phosphorylation of Akt in HEY cells, which is essential for the full activation of Akt. We demonstrate here that p38 activation is required for most stimuli (LPA, S1P, PDGF, EGF, thrombin, and Et-1, but not insulin) to induce S473 phosphorylation of Akt in HEY cells. In addition, p38 is required for LPA/S1P-induced S473 phosphorylation of Akt in all 11 responsive cell lines tested. On the other hand, of the stimuli tested, MEK is required for Akt S473 phosphorylation induced only by LPA, S1P, and PDGF, and also occurs in a cell line-specific manner. MEK-dependent Akt phosphorylation occurs in all six ovarian cancer cell lines tested, as well as HeLa cells, and T-47D and MDA-MB-231 breast cancer cells, but not in PC-3 prostate cancer or GI-101A breast cancer cells. Moreover, Akt is phosphorylated in a Rho-dependent manner by LPA but not S1P or PDGF, and Rho acts upstream of PI3-K. Our results show that LPA and S1P decrease paclitaxel-induced caspase-3 activity in HEY cells, which is mediated by the PI3-K/MEK/p38 pathway, suggesting that LPA/S1P-induced Akt activation is potentially involved in survival activities of these cells. Because LPA and S1P probably activate Akt through their Edg receptors, the expression patterns of these receptors in all cell lines used in this study have been examined.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Oleoyl-LPA and S1P were purchased from Avanti Polar Lipids (Birmingham, AL) or Toronto Research Chemicals (Toronto, ON, Canada). LPA was dissolved in phosphate-buffered saline (PBS), and S1P was dissolved in Tris-saline (50 mM Tris, pH 9.5, 145 mM NaCl) to 4 and 2 mM stock solutions, respectively. LY294002, PD98059, and SB203580 were obtained from Biomol (Plymouth Meeting, PA). Wortmannin and paclitaxel were obtained from Sigma-Aldrich (St. Louis, MO). PTX was purchased from Invitrogen (Rockville, MD). PDGF-BB was a kind gift from the lab of Dr. Paul DiCorleto (Cleveland Clinic Foundation, Cleveland, OH) or was purchased from R & D Systems (Minneapolis, MN). Thrombin and EGF were obtained from Calbiochem (La Jolla, CA) and Et-1 was from Peninsula Laboratories, Inc. (San Carlos, CA). Anti-phospho-S473-Akt, anti-phospho-T308-Akt, anti-Akt, anti-phospho-ERK, anti-ERK, and anti-phospho-p38 antibodies were obtained from Cell Signaling Technology (Beverly, MA). Anti-MEK2, anti-MKK6, and anti-p38 antibodies were from StressGen (Victoria, BC, Canada).

Cell Culture and Transfection. HEY, Ovca420, Ovca429, Ovca432, and Ovca433 ovarian cancer cells were from Dr. G. Mills or Dr. R. Bast, MD Anderson Cancer Center (Houston, TX). MDA-MB-231 and T-47D breast cancer cells were from American Type Culture Collection (Manassas, VA). GI-101A cells were from the Goodwin Institute for Cancer Research, Inc. (Plantation, FL). PC-3 cells were from Dr. Warren Heston (Cleveland Clinic Foundation). All of the above cell lines were maintained in RPMI 1640 medium containing 10% fetal bovine serum (FBS) at 37^oC with 5% CO₂. A2780 cells (also from Dr. G. Mills) were maintained in DMEM/Ham's F12 medium (1:1) supplemented with 10% FBS. HeLa cells (from American Type Culture Collection) were maintained in DMEM medium containing 10% FBS. MCF10A immortalized breast cells were obtained from the Karmanos Cancer Institute (Detroit, MI) and cultured as recommended by the provider. All cells were cultured in serum-free media for 24 to 48 h before lipid treatment. For transient transfections, cells were plated into 35-mm dishes and transfected with DNA using LipofectAMINE (Invitrogen) and Transfection Booster Reagents (Gene Therapy Systems, San Diego, CA) according to the manufacturers' instructions. Transfected cells were used within 48 h after transfection. Transfection efficiency was detected by lacZ transfection and -galactosidase staining. Dominant negative and constitutively active MEK were from Dr. D. Templeton (Case Western Reserve University, Cleveland, OH). Kinase inactive p38 and constitutively active MKK6 were from Dr. Bryan R.G. Williams (Cleveland Clinic Foundation). Dominant-negative and constitutively active Rho were from Dr. Wouter Moolenaar, (Netherlands Cancer Institute, Amsterdam, The Netherlands). The C3-exoenzyme construct and constitutively active PI3-K (p110- isoform) were provided by Dr. Alan Wolfman (Cleveland Clinic Foundation).

Nonradioactive Immunoprecipitation Akt Kinase Assay. The Akt kinase assay was performed with the Nonradioactive Akt Kinase Assay Kit (Cell Signaling Technology) according to the manufacturer's instructions. All reagents were provided with the kit. Briefly, cells were treated with LPA or S1P, rinsed with ice-cold PBS, and then lysed in cell lysis buffer. Immunoprecipitation was carried out using immobilized Akt 1G1 monoclonal antibody. The immunoprecipitate was then incubated with GSK-3 fusion protein and ATP in kinase buffer. Western analyses were used to determine the extent of GSK-3 phosphorylation by active Akt using a phospho-GSK-3/ (Ser21/9) antibody.

Western Blot Analysis. After treatment with LPA, S1P, or other stimuli, cells were rinsed with ice-cold PBS, and then lysed in SDS sample buffer. Samples were electrophoresed through 10 to 12% SDS polyacrylamide gels and then transferred to PVDF membranes (Bio-Rad, Hercules, CA). Immunoblot analyses were carried out using the appropriate antibodies. Specific proteins were detected with the enhanced chemiluminescence system (Amersham Biosciences, Piscataway, NJ).

Quantitative RT-PCR of LPA/S1P Receptor Expression. Total RNA was extracted from cells using the SV Total RNA Isolation System (Promega, Madison, WI). Total RNA (1-5 µg) was reverse transcribed using Superscript II RT (Invitrogen). Derived cDNA (8 ng) was used as a template for real-time quantitative SYBR Green I PCR. Primer sequences for S1P₁ (Edg-1), S1P₂ (Edg-5), S1P₃ (Edg-3), LPA₁ (Edg-2), and LPA₂ (Edg-4) were kindly provided by Dr. Ed Goetzl (UCSF) and are as follows: S1P₁, 5'GCAGCAGCAAGATGCGAAG and 5'CGATGAGTGATCCAGGCTTTT; S1P₂, 5'GCGCCATTGTGGTGGAA and 5'GAGCCAGAGAGCAAGGTATTGG; S1P₃, 5'CTGGTGACCATCGTGATCCTC and 5'ACGCTCACCACAATCACCAC; LPA₁, 5'GCTGGTGATGGGACTTGGAAT and 5'CAACCCAGCAAAGAAGTCTGC; and LPA₂, 5'ACGCTCAGCCTGGTCAAGAC and 5'AACCATCCAGGAGCAGTACCAC. Primer sequences for S1P₅ (Edg-8) and LPA₃ (Edg-7) were developed in our lab and are: S1P₅, 5'CGCCTTCATCGTGCTAGAGA and 5'AGATCCGACAACGTGAGGCT; and LPA₃, 5'TCCAACCTCATGGCCTTCC and 5'GACCCACTTGTATGCGGAGAC. GAPDH was amplified in a separate tube as a housekeeping gene with primers 5'GAAGGTGAAGGTCGGAGT and 5'GAAGATGGTGATGGGATTTC. All SYBR Green I core reagents, including AmpliTaq Gold polymerase, were from Applied Biosystems (Foster City, CA). The thermal cycling conditions were 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C, 1 min. PCR reactions and product detection were carried out in an ABI Prism 7700 Sequence Detection System (Applied Biosystems). Amplified product was detected by measurement of the fluorescent dye, SYBR Green I, which was added to the initial reaction mixture and binds proportionally to double-stranded DNA. After completion of the PCR, a fixed threshold (DNA amount reflected by bound fluorescent dye) was selected based on the manufacturer's suggestion, and the number of cycles (the threshold cycle, or C_T) required to amplify the target to reach this threshold was used for calculations. The comparative C_T method (User Bulletin #2; Applied Biosystems) was used to determine relative amounts of each receptor. The comparative C_T method is similar to the standard curve method, except that it uses arithmetic formulas derived by Applied Biosystems to achieve the same results for relative quantitation. For this method to be valid, the efficiencies of the target (i.e., LPA/S1P receptor) and reference (i.e., GAPDH) must be approximately equal. We have validated the amplification efficiencies of each of our targets to meet this requirement.

Measurement of Caspase-3 Activity. Cells were seeded into 96-well plates, grown to 80% confluence, and then cultured overnight in serum-free media. The following day, cells were pretreated with or without various reagents, followed by exposure to paclitaxel for the indicated periods of time. Cells were then washed with PBS and lysed in caspase-3 assay kit cell lysis buffer (Calbiochem). Caspase-3 activity was measured by cleavage of the fluorogenic substrate N-acetyl-Asp-Glu-Val-Asp-amino-4-methylcoumarin with the Caspase-3 Assay Kit (Calbiochem).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

LPA and S1P Induced Akt Activation in HEY Ovarian Cancer Cells in a PI3-K-, MEK-, and p38-Dependent Manner. LPA and S1P are present in ovarian cancer ascites and are likely to be involved in proliferation and survival of ovarian tumor cells. We have shown that LPA and S1P stimulate ovarian cancer cell proliferation (Xu et al., 1995a; Hong et al., 1999). However, the effects and mechanisms of LPA/S1P-induced PI3-K/Akt activation in ovarian cancer cells have not previously been reported. To investigate the effects of LPA and S1P on Akt activation in HEY ovarian cancer cells, we treated these cells with physiological concentrations of LPA and S1P, and then measured the activity of Akt with an Akt kinase assay. LPA (10 µM) and S1P (1 µM) induced activation of Akt compared with the untreated control (Fig. 1). To determine the mechanism of LPA/S1P-induced Akt activation, and in particular, to test whether there is an interaction between a major cell proliferation signaling pathway (MEK/ERK) and a major cell survival pathway (PI3-K/Akt), we examined the sensitivity of Akt activation induced by LPA and S1P to three specific inhibitors of PI3-K, MEK, and p38 MAPK: LY294002, PD98059, and SB203580, respectively. Pretreatment of HEY cells with all three of these inhibitors abolished activation of Akt by LPA and S1P (Fig. 1), suggesting a dependence on PI3-K, MEK, and p38 for LPA/S1P-induced Akt activation.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   LPA and S1P activate Akt in a PI3-K-, MEK-, and p38-dependent manner. Nonradioactive immunoprecipitation Akt kinase assay of HEY cells pretreated with or without 10 µM LY294002 (LY), 30 µM PD98059 (PD), or 10 µM SB203580 (SB) for 30 min, and then treated with LPA (10 µM) or S1P (1 µM) for 20 min.

Time- and Concentration-Dependent Akt S473 Phosphorylation Induced by LPA and S1P. Akt activation is mediated through phosphorylation of S473 and T308. Western blot analyses with Akt-S473-phospho-specific antibodies were used to measure LPA/S1P-induced S473 phosphorylation of Akt. LPA and S1P induced a time-dependent Akt S473 phosphorylation in HEY cells (Fig. 2, A and B). HEY cells displayed a low basal level of Akt S473 phosphorylation, which was not increased over the time course (Fig. 2A). Both LPA and S1P induced a time- and concentration-dependent S473 phosphorylation of Akt, occurring as early as 5 min, with maximal stimulations of Akt S473 phosphorylation occurring at 20 min (Fig. 2B) and at 10 µM for LPA and 1 µM for S1P (Fig. 2C). Our results demonstrate that at concentrations of LPA greater than 10 µM, the Akt and ERK phosphorylation levels are significantly decreased compared with 10 µM LPA (Fig. 2C and data not shown). Furthermore, there is a correlation between the fold-change decrease in phosphorylation of Akt and ERK (with LPA concentrations greater than 10 µM). LPA and S1P had relatively narrow optimal concentration ranges for Akt activation. We observed a similar phenomenon in LPA-induced thymidine incorporation (Xu et al., 1995b). The relationship between these effects is currently under investigation.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Time- and concentration-dependent S473 phosphorylation of Akt by LPA and S1P. Western analyses of HEY cell lysates probed with phospho-specific Akt S473 (p-S473-Akt) or Akt (total Akt) antibodies. A, cells were treated without lipid (Control) or with LPA (10 µM) or S1P (1 µM) for 5, 10, and 20 min. The Western analysis shown is a representative example of at least three independent experiments. B, cells were treated with LPA (10 µM) or S1P (1 µM) for various time points as indicated, up to 60 min. Results were plotted as the mean ± S.D. of three independent experiments. Fold-change in Akt phosphorylation is normalized to the levels of total Akt. C, cells were treated with LPA and S1P for 20 min at the indicated dosages. Results were plotted as the mean ± S.D. of three independent experiments. Fold change in Akt phosphorylation is normalized to the levels of total Akt. ***, p <0.001; **, p <0.01 (Student's t-test).

PI3-K and G_i-Dependent T308 and S473 Akt Phosphorylation by LPA and S1P. Because phosphorylation of both T308 and S473 are necessary for the complete activation of Akt, we examined the ability of LPA and S1P to stimulate Akt T308 phosphorylation. Both LPA and S1P were able to induce an approximately 4- to 6-fold increase in Akt phosphorylation at T308 in HEY cells (Fig. 3). To determine whether S473 and T308 phosphorylation of Akt by LPA and S1P were dependent on PI3-K activity, we examined the effect of the specific PI3-K inhibitor, LY294002, on Akt phosphorylation. LPA- and S1P-induced S473 and T308 phosphorylation of Akt were completely abolished by pretreatment of cells with LY294002 (10 µM) (Fig. 3). The dependence on PI3-K was further confirmed by pretreatment with wortmannin (150 nM), a second, structurally different, specific inhibitor of PI3-K (Fig. 3).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   PI3-K- and Gi-dependent Akt S473 and T308 phosphorylation by LPA and S1P. HEY cells were pretreated with 10 µM LY294002 (LY) or 150 nM wortmannin (Wort) for 30 min, or 100 ng/ml PTX for 16 h before treatment with LPA (10 µM) or S1P (1 µM) for 20 min. Cell lysates were collected and analyzed for S473 and T308 phosphorylation of Akt, as well as total Akt, by Western blotting. The Western analysis shown is a representative example of at least three independent experiments.

LPA- and S1P-Induced Akt S473 and T308 Phosphorylation Are Dependent on MEK, but Only S473 Phosphorylation Is p38-Dependent, and MEK Is Upstream of p38. Phosphorylation of both S473 and T308 is essential for the full activation of Akt. However, Akt S473 and T308 may be phosphorylated through different mechanisms. Because LPA/S1P-induced Akt enzymatic activation was MEK- and p38-dependent (Fig. 1), we determined whether phosphorylation of Akt S473 and T308 by LPA/S1P also were MEK- and p38-dependent. We used PD98059 and SB203580 as specific inhibitors of MEK (the upstream kinase of ERK) and p38, respectively. To use the optimal concentration of these inhibitors, we performed titration analyses and observed that PD98059 at 3, 10, and 30 µM had an inhibitory effect on LPA/S1P-induced ERK and Akt phosphorylation of approximately 60, 90, and 100%, respectively. Similarly, SB203580 at 1, 3, and 10 µM had an inhibitory effect on LPA/S1P-induced p38 and Akt phosphorylation of approximately 60, 90, and 100%, respectively. PD98059 (30 µM) completely inhibited Akt S473 and T308 phosphorylation induced by LPA and S1P (Fig. 4A, 1 and 2), suggesting that MEK, and potentially its downstream target, ERK, was involved in phosphorylation of Akt at both S473 and T308. Interestingly, we found that although SB203580 (10 µM) completely abolished Akt S473 phosphorylation induced by both LPA and S1P (Fig. 4A, 1), T308 phosphorylation was not altered by pretreatment with up to 30 µM SB203580 (Fig. 4A, 2). These results suggest that p38 activation is required for LPA/S1P-induced Akt S473, but not T308, phosphorylation.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   MEK- and p38 MAPK-dependent Akt S473 phosphorylation and MEK-dependent Akt T308 phosphorylation by LPA and S1P. Western analyses of HEY cell lysates probed with p-S473-Akt, p-T308-Akt, phospho-specific ERK (p-ERK), phospho-specific p38 (p-p38), total Akt, or total ERK antibodies. A, HEY cells were pretreated with 30 µM PD98059 (PD) or 10 µM SB203580 (SB) for 30 min followed by treatment with LPA (10 µM) or S1P (1 µM) for 10 min (p-ERK, p-p38) or 20 min (p-Akt). B, HEY cells were transiently transfected with control vector, dominant-negative MEK (MEK/2A), or kinase inactive p38 MAPK (p38/AGF), and then treated with LPA (10 µM) or S1P (1 µM) for 10 min (p-ERK, p-p38) or 20 min (p-Akt). C, HEY cells were transiently transfected with control vector, constitutively active MEK (MEK/2E), or constitutively active MKK6 (MKK6/2E) and then treated with LPA (10 µM) or S1P (1 µM) for 10 min (p-ERK, p-p38) or 20 min (p-Akt). D, HEY cells were pretreated with 100 ng/ml PTX or 10 µM LY294002 (LY) for 16 h or 30 min, respectively, followed by treatment with LPA (10 µM) or S1P (1 µM) for 10 min. E, HEY cells were transiently transfected with constitutively active PI3-K (CA PI3-K). F, HEY cells overexpressing CA PI3-K or vector control were treated with 30 µM PD98059 (PD) or 10 µM SB203580 (SB) for 30 min. Each Western analysis shown is a representative example of at least three independent experiments.

MEK-Dependent Akt S473 Phosphorylation Is Specific to LPA, S1P, and PDGF, but Not Thrombin, EGF, Et-1, or Insulin. The kinase activity of Akt was 90% inhibited by SB203580 (Fig. 1) and the phosphorylation of Akt S473, but not T308, was sensitive to both SB203580 and transfection with MEK/2A, suggesting that S473 phosphorylation was essential for the majority of the Akt kinase activity in this system. Therefore, we focused the rest of our studies on the mechanisms of Akt S473 phosphorylation induced by LPA and S1P. Akt can be activated by a variety of growth factors through their receptors, as well as via a number of GPCR ligands. We sought to determine whether MEK-dependent S473 phosphorylation of Akt in HEY cells was specific to LPA and S1P. Treatment of HEY cells with PDGF (10 ng/ml), thrombin (1 U/ml), EGF (10 ng/ml), Et-1 (100 nM), or insulin (100 nM) for 5 min induced S473 phosphorylation of Akt (Fig. 5A). The phosphorylation of Akt by all five of these stimuli was PI3-K-dependent as evidenced by pretreatment with LY294002 (data not shown). Pretreatment of cells with SB203580 (3-10 µM) resulted in inhibition of Akt S473 phosphorylation induced by PDGF, EGF, thrombin, and Et-1, but not insulin, indicating that p38 activation is required by most stimuli tested (Fig. 5A). In contrast, of these factors tested, only PDGF required activation of MEK for S473 phosphorylation of Akt as evidenced by the sensitivity of this activation to pretreatment with PD98059 (Fig. 5A), suggesting that MEK-dependent Akt S473 phosphorylation is stimulus-specific. This was further supported by transient transfection with MEK/2A, which resulted in decreased Akt S473 phosphorylation by PDGF but not Et-1, even though the latter also activated ERK, which was sensitive to the dominant inhibitory effect of MEK/2A (Fig. 5B).

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Stimulus-specific S473 phosphorylation of Akt. Western analyses of HEY cell lysates probed with p-S473-Akt, and then reprobed with p-ERK or total Akt antibodies. A, cells were pretreated with 30 µM PD98059 (PD) or 10 µM SB203580 (SB) for 30 min followed by treatment with thrombin (1 U/ml), EGF (10 ng/ml), insulin (100 nM), Et-1 (100 nM), or PDGF-BB (10 ng/ml) for 5 min. B, cells transiently transfected with MEK/2A were treated with Et-1 (100 nM) or PDGF-BB (10 ng/ml) for 5 min.

LPA, but Not S1P or PDGF, Requires Rho for S473 Phosphorylation of Akt, and Rho Is Upstream of PI3-K. To test the potential involvement of Rho in Akt S473 phosphorylation induced by LPA, S1P, and PDGF, we transiently transfected HEY cells with dominant negative Rho (Rho/N19) or C3-exoenzyme. Interestingly, overexpression of Rho/N19 or C3-exoenzyme resulted in decreased Akt S473 phosphorylation by LPA, but not S1P (Fig. 6A, 1). Furthermore, LPA, but not S1P, required Rho for phosphorylation of ERK and p38 (Fig. 6A, 3 and 4). Similar to S1P, overexpression of Rho/N19 did not inhibit PDGF-induced Akt S473 phosphorylation (Fig. 6B). Transfection with constitutively active Rho (Rho/V14) demonstrated that Rho was sufficient for S473 phosphorylation of Akt in HEY cells (Fig. 6C, 3rd column).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Rho-dependent S473 phosphorylation of Akt by LPA but not S1P or PDGF. Western analyses of HEY cell lysates probed with p-S473-Akt, p-ERK, p-p38, total Akt, or total ERK antibodies. Cells were transiently transfected with control vector, dominant-negative Rho (Rho/N19), or C3-exoenzyme (C3) (A, B), constitutively active Rho (Rho/V14) (C), or Rho/N19 and or constitutively active PI3-K (CA PI3-K) (D). A, transfected cells were treated with LPA (10 µM) or S1P (1 µM) for 10 min (p-ERK, p-p38) or 20 min (p-Akt). B, transfected cells were treated with PDGF (10 ng/ml) for 10 min. C, transfected cells were treated with LY294002 (10 µM) for 30 min after transfection.

MEK/ERK-Dependent Akt S473 Phosphorylation Is Cell Line-Specific. In contrast to what we have observed in HEY cells with LPA and S1P, the S473 phosphorylation of Akt induced by fMLP, Fc-R cross-linking, or PIP₃ in human neutrophils was insensitive to pretreatment with PD98059 (Rane et al., 2001). To determine whether this MEK-dependent S473 phosphorylation of Akt by LPA and S1P was cell-type-specific, we tested five other ovarian cancer cell lines (Ovca420, Ovca429, Ovca432, Ovca433, A2780), three breast cancer cell lines (MDA-MB-231, T-47D, and GI-101A), and HeLa (cervical cancer), PC-3 (prostate cancer), and MCF10A (immortalized breast epithelial) cells. In all cell lines tested, except the immortalized breast MCF10A cell line (Fig. 7C), LPA and S1P induced a 1.7- to 10.1-fold increase in Akt S473 phosphorylation (Table 1). Whereas MCF10A cells were the only cells in this study that did not respond to LPA/S1P for Akt phosphorylation, we have observed that LPA and S1P can induce Akt activity in other noncancerous cell lines (specifically, Swiss 3T3 cells or mouse embryonic fibroblasts; unpublished observations). Thus, a simple, generalized conclusion pertaining to the effects of LPA and S1P on Akt induction in cancerous versus noncancerous cells cannot be drawn through our limited studies.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   The effect of LPA/S1P on Akt S473 phosphorylation in Ovca420, GI-101A, and MCF10A cells. Western analyses of cell lysates probed with p-S473-Akt, and then reprobed with p-ERK or total Akt antibodies. Cells were pretreated with 30 µM PD98059 (PD) or 10 µM SB203580 (SB) for 30 min followed by treatment with LPA (10 µM) or S1P (1 µM) for 20 min. A, Ovca420 cells; B, GI-101A cells; C, MCF10A cells.

The LPA/S1P Receptors Potentially Involved in Mediating the S473 Phosphorylation of Akt. Three [LPA₁ (Edg-2), LPA₂ (Edg-4), and LPA₃ (Edg-7)] and five [S1P₁ (Edg-1), S1P₂ (Edg-5), S1P₃ (Edg-3), S1P₄ (Edg-6), and S1P₅ (Edg-8)] GPCRs have been identified as receptors for LPA and S1P, respectively. To determine which of these receptors might be associated with the MEK-dependent Akt activation in different cell lines, we examined the expression of S1P_1-3, S1P₅, and LPA_1-3 in all cell lines used in this study with quantitative real time RT-PCR (Table 2). Because S1P₄ is predominantly expressed in lymphocytes (Graler et al., 1998; Van Brocklyn et al., 2000), and all of our cell lines are of epithelial origin, the expression of this receptor was not determined in our studies. The comparative threshold cycle (C_T) method (Experimental Procedures) was used to calculate the relative expression of each receptor in different cell lines. We arbitrarily chose the expression level of LPA₂ in HEY cells (relative to GAPDH in these cells) as 1-fold. The expression levels of all other receptors in HEY and all other cell lines (relative to GAPDH in the corresponding cell lines) are expressed as fold-change relative to this 1-fold expression of HEY LPA₂ (Table 2). The LPA/S1P receptor expression levels in HEY cells obtained through our studies are consistent in principle with the levels reported by Fischer et al. (2001), using a semiquantitative RT-PCR method (Fischer et al., 2001). In our studies, we considered 1-fold expression to be low, because LPA₂ was previously detected at a very low level in HEY cells using a semiquantitative RT-PCR method (Fischer et al., 2001). Thus, when the fold expression of the receptor was below 1.0 in our studies, we considered it to be very low or not expressed.

Caspase-3 Activity Induced by Paclitaxel in Hey Cells Was Inhibited by LPA and S1P. Akt has been described as a mediator of survival signals in many cell types (Marte and Downward, 1997), including ovarian cancer cells (Liu et al, 1998; Yuan et al., 2000). Furthermore, LPA has been shown to prevent HEY cell death induced by cis-diamminedichloroplatinum (Frankel and Mills, 1996). Because paclitaxel is a potent apoptotic inducer in many ovarian cancer cell lines, we investigated the potential for LPA and S1P to prevent paclitaxel-induced apoptosis in HEY cells and whether the effect was related to the PI3-K signaling pathway. We used a caspase-3 activity assay as a sensitive measurement of the effect of paclitaxel on HEY cells. Caspase-3 activity was measured in HEY cells treated for various times and concentrations of paclitaxel, and it was determined that optimal caspase-3 activity occurred after 24 h treatment with 1 µM paclitaxel (data not shown). Pretreatment of cells for 20 min with LPA (10 µM) and S1P (1 µM) inhibited (45%) caspase-3 activity induced by paclitaxel (Fig. 8). Pretreatment with LY294002, PD98059, and SB203580, followed by treatment with LPA and S1P, and then paclitaxel, reinstated caspase-3 activity (Fig. 8). These results suggest that the PI3-K/MEK/p38 signaling pathway mediates LPA/S1P-induced inhibition of caspase-3 activity, which is the same signaling pathway leading to LPA/S1P-stimulated Akt S473 phosphorylation in HEY cells. Therefore, Akt may mediate the LPA/S1P-induced caspase-3 inhibition in HEY ovarian cancer cells. A schematic of the pathways leading to Akt activation by LPA and S1P in HEY cells is shown in Figure 9.

View larger version (28K): [in this window] [in a new window]   Fig. 8.   LPA and S1P reduced paclitaxel-induced caspase-3 activity. HEY cells were pretreated with 30 µM PD98059 (PD), 10 µM SB203580 (SB), or 10 µM LY294002 (LY) for 30 min, then treated with LPA (10 µM) or S1P (1 µM) for 20 min, followed by exposure to paclitaxel (1 µM) for 24 h. Caspase-3 activity was determined by cleavage of the fluorogenic substrate, N-acetyl-Asp-Glu-Val-Asp-amino-4-methylcoumarin. ***, p <0.001 (Student's t-test).

View larger version (12K): [in this window] [in a new window]   Fig. 9.   Schematic of LPA/S1P-induced Akt activation in HEY ovarian cancer cells. Solid arrows indicate activation of signaling molecules but do not exclude mediation of this activation by additional effectors. A dashed arrow indicates that the relationship between the effector and activated molecule was not directly tested in this study. MK2 may function as PDK2 in our system, although this was not directly tested. *, dominant-negative and/or constitutively active alleles were used to study the involvement of that signaling molecule in this study. C3, C3-exoenzyme; LY, LY294002; PD, PD98059; PTX, pertussis toxin; SB, SB203580; Wort, wortmannin.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Although the signaling pathways of p38, MEK/ERK, and Akt activation induced by various stimuli, including LPA and S1P, have been studied in many cellular systems, a number of novel and important signaling mechanisms have been revealed through the current study. First, our work indicates for the first time that p38 is a relatively general requirement for the S473 phosphorylation of Akt, with the exception of insulin-induced Akt activation. The signaling mechanisms leading to p38 activation, however, seem to be highly cell line- and stimulus-specific. Second, we have observed LPA/S1P-induced cross communication between the two major kinase cascades (MAPK and PI3-K/Akt) involved in cell proliferation and cell survival, respectively, in ovarian cancer cells. Finally, our study reveals a cell line- and stimulus-specific MEK-dependent Akt activation, and we have explored the potential role of LPA and S1P receptors that confer a MEK-dependent Akt activation.

p38 Is a Relatively General Requirement for the S473 Phosphorylation of Akt. Full activation of Akt requires phosphorylation at both S473 and T308, which is usually regulated by different signaling mechanisms (Alessi et al., 1997; Pullen et al., 1998). T308 is phosphorylated by PDK1, which has been cloned and identified (Alessi et al., 1997). Our current understanding of the mechanism of S473 phosphorylation and the identification of PDK2 is still elusive and controversial. At least four different kinases have been suggested to be potential candidates of PDK2: Akt itself, PDK1, ILK1, and MK2 (Chan and Tsichlis, 2001). Because S473 can be phosphorylated when Akt is inactive, Akt autophosphorylation clearly cannot account for all of the S473 phosphorylation induced under different conditions. Although it has been suggested that PDK1 can acquire PDK2 activity, PDK1 is not necessary for S473 phosphorylation, which can occur even in PDK1-knockout ES cells. In cells stimulated with insulin, ILK is activated and enhances S473 phosphorylation through a PI3-K-dependent mechanism. However, it is likely that ILK may contribute indirectly to S473 phosphorylation of Akt by providing an adaptor function. The major obstacle for recognizing MK2 as one of the PDK2 candidates arises from the fact that MK2 is not involved in S473 phosphorylation induced by insulin in HEK 293 cells (Alessi et al., 1996). However, the recent work by Rane et al. (2001) illustrates the possibility that MK2 functions as PDK2 under certain conditions. Our data seem to support works from both Alessi et al. (1996) and Rane et al. (2001) suggesting that at least two types of signaling pathways are involved in S473 phosphorylation. One type of signaling pathway, which does not require MK2 and/or p38 for S473 phosphorylation, may be represented by cells that respond to insulin, insulin-like growth factor, heat shock, or hydrogen peroxide (Shaw et al., 1998). Another signaling pathway, in which MK2 functions as PDK2, may be represented by cells, such as neutrophils, stimulated by fMLP, Fc-R cross-linking, PIP₃ (Rane et al., 2001), and other potential stimuli.

The Signaling Pathways That Regulate p38 Activity Seem to Be Highly Specific. For example, although PI3-K is required for Akt activation in both keratinocytes (Zhang et al., 2001) and HEY cells (Fig. 1), we show that PI3-K is both necessary and sufficient to induce ERK and p38 activation in HEY cells (Fig. 4), whereas in keratinocytes, p38 is activated by a PI3-K-independent pathway (Zhang et al., 2001). Thus, p38 activation can be PI3-K-dependent or -independent, depending on the system. Furthermore, our results show that MEK is an upstream activator of p38 in MEK-dependent (and hence, p38-dependent) cellular systems. In MEK-independent cellular systems, p38 is still required for Akt activation. Thus, our results suggest that p38 can be activated by MEK-dependent or -independent pathways. The mechanisms regulating p38 activity in different systems require further investigation.

Cross Communication between the MAPK and PI3-K/Akt Cascades. Although simultaneous stimulation of the ERK and PI3-K/Akt pathways has been reported previously, the requirement of MEK/ERK for Akt activation was either not examined or MEK/ERK and Akt were shown to be unrelated to each other. One of the novel and important findings of our work presented here is that MEK is a necessary and sufficient activator of Akt, and MEK functions upstream of p38 in G_i/PI3-K/Akt signaling in a cell- and stimulus-specific manner. In particular, all six ovarian cancer cell lines tested demonstrate MEK/ERK-dependent Akt activation. These studies have revealed an integration of two important signaling pathways (MAPK and PI3-K/Akt) that govern two important tumorigenic processes (cell proliferation and survival, respectively). This is of potential therapeutic importance for ovarian cancer because both LPA and S1P 1) have been detected in ovarian cancer plasma and ascites (Xiao et al., 2000; Xiao et al., 2001), 2) protect ovarian cancer cells from paclitaxel-induced apoptosis in a manner dependent on the MAPK and PI3-K pathways (Fig. 8), 3) regulate pro-angiogenic factors in ovarian cancer (Hu et al., 2001; Schwartz et al., 2001) and 4) affect ovarian cancer cell proliferation, migration, and/or survival (Xu et al., 1995a, 2001; Frankel and Mills, 1996; Hong et al., 1999; Lu et al., 2002).

MEK-Dependent Akt Phosphorylation Is Cell Line- and Stimulus-Specific. Our study reveals a cell line- and stimulus-specific MEK-dependent Akt activation. Interestingly, while this article was in preparation, a MEK-dependent Akt activation by ultraviolet B irradiation was reported by Nomura et al. (2001). Thus, our results are consistent with these recent findings. However, although Nomura et al., reported a MEK- and p38-dependent T308 phosphorylation of Akt by ultraviolet B radiation in mouse epidermal cells (Nomura et al., 2001), our results show for the first time that Akt T308 phosphorylation is dependent on MEK, but not p38, in HEY cells. Thus, similar to MEK-induced Akt S473 phosphorylation, the requirement of p38 activity for Akt T308 phosphorylation may be cell line- and/or stimulus-specific and remains to be further investigated.