Ether-linked diglycerides inhibit vascular smooth muscle cell growth via decreased MAPK and PI3K/Akt signaling

Kristy L. Houck, Todd E. Fox, Lakshman Sandirasegarane, Mark Kester

Abstract

Diglycerides (DGs) are phospholipid-derived second messengers that regulate PKC-dependent signaling pathways. Distinct species of DGs are generated from inflammatory cytokines and growth factors. Growth factors increase diacyl- but not ether-linked DG species, whereas inflammatory cytokines predominately generate alkyl, acyl- and alkenyl, acyl-linked DG species in rat mesenchymal cells. These DG species have been shown to differentially regulate protein kinase C (PKC) isotypes. Ester-linked diacylglycerols activate PKC-ε and cellular proliferation in contrast to ether-linked DGs, which lead to growth arrest through the inactivation of PKC-ε. It is now hypothesized that ether-linked DGs inhibit mitogenesis through the inactivation of ERK and/or Akt signaling cascades. We demonstrate that cell-permeable ether-linked DGs reduce vascular smooth muscle cell growth by inhibiting platelet-derived growth factor-stimulated ERK in a PKC-ε-dependent manner. This inhibition is specific to the ERK pathway, since ether-linked DGs do not affect growth factor-induced activation of other family members of the MAPKs, including p38 MAPK and c-Jun NH2-terminal kinases. We also demonstrate that ether-linked DGs reduce prosurvival phosphatidylinositol 3-kinase (PI3K)/Akt signaling, independent of PKC-ε, by diminishing an interaction between the subunits of PI3K and not by affecting protein phosphatase 2A or lipid (phosphatase and tensin homologue deleted in chromosome 10) phosphatases. Taken together, our studies identify ether-linked DGs as potential adjuvant therapies to limit vascular smooth muscle migration and mitogenesis in atherosclerotic and restenotic models.

  • cell migration
  • proliferation
  • bioactive lipids
  • mitogen-activated protein kinase
  • phosphatidylinositol 3-kinase

diglycerides (DGs) are bioactive phospholipid-derived second messengers. Our laboratory and others have shown that there are distinct DG species produced endogenously in mammalian cells (20, 23, 27, 32). It has been demonstrated that growth factors, such as platelet-derived growth factor, lead to the formation of ester-linked diacylglycerols (DAGs), whereas cytokines, such as interleukin-1, generate ether-linked DGs, which contain either an alkyl- or alkenyl-chain, via hydrolysis of ether-linked phosphatidylethanolamine (32). DAGs are linked to vascular smooth muscle (VSM) cell mitogenesis (27). In contrast, we have demonstrated that ether-linked DGs inhibit mesenchymal cell mitogenesis (27). In fact, these novel ether-linked, phospholipid-derived, second messengers mimic the effect of interleukin-1 to inhibit growth factor-induced cellular proliferation (27).

DAG acts as an allosteric activator of protein kinase C (PKC), which is the most prominent DAG target. PKC is a family of homologous serine/threonine, lipid-regulated protein kinases linked to a myriad of biological functions, including cell cycle progression, differentiation, and cell survival (33). PKCs are subdivided into three classes: conventional or calcium dependent (α, β, and γ), novel or calcium independent (δ, ε, η, and θ), and atypical or DG and calcium independent (ζ and λ/ι). We have chosen to predominantly investigate the DAG-activatable PKC-ε, since this is the most strongly activated PKC isotype in rat VSM cells (1). During VSM activation, DAG causes the translocation of cytosolic PKC-ε to the plasma membrane, where it becomes fully activated and hence is able to activate promitogenic signaling cascades, such as MEK/ERK (35).

Ether-linked DGs have been shown to inhibit mitogenesis via an inability to activate PKCs. We have shown that ether-linked DGs competitively inhibit ester-linked DAG- and growth factor-stimulated PKC-δ and -ε but not -ζ (27). In previous studies, ether-linked DG inhibition of PKC-ε was more potent than inhibition of PKC-δ (27). Additional studies support this signaling mechanism. For example, ether-linked DGs fail to activate total PKC activity or only activate PKC in the presence of pharmacological concentrations of calcium (7, 12, 16, 21). Ether DGs competitively bind to, but do not activate, PKC (31) since they lack the carboxyl group of the sn-l ester bond, as well as the corresponding bond angle essential for PKC activation (25, 34). The ester bond in the sn-1 position is a major determinant for DAG-mediated PKC activation, and hence an ether, or a vinyl ether, bond at the sn-1 position leads to the loss of PKC activity (21). However, both ester-linked and ether-linked DGs most likely bind at the same site on PKC and hence interact competitively (27). The inactivation of PKC-ε by ether-linked DGs is consistent with a growth-arrested phenotype (42). Whereas a downregulation of PKC-α or -ε inhibits Gl/S transition in vascular smooth muscle cells (36), an overexpression of PKC-ε induces tumorigenicity in fibroblasts (8, 29). Additionally, ether-linked phospholipids and alkyl, but not acyl, lysophosphatidylcholine have been identified as antineoplastic, antiproliferative, and apoptosis-inducing agents (2, 15, 43).

We now investigate putative downstream targets that regulate ether-linked DG-induced growth arrest in VSM cells. We specifically investigated the ability of DG species to regulate ERK and Akt promitogenic signaling cascades, downstream targets of activated PKC-ε and phosphatidylinositol 3-kinase (PI3K), respectively. It was hypothesized that ether-linked DGs inhibit mitogenesis through inactivation of ERK and Akt signaling cascades. Understanding the mechanisms by which ether-linked DGs limit VSM cell growth or migration has direct clinical significance to inflammatory VSM pathologies, including atherosclerosis or restenosis after angioplasty/stenting.

MATERIALS AND METHODS

Materials and cell culture.

Rat aorta smooth muscle cells (A7r5) were obtained from American Type Culture Collection (Rockville, MD). Human coronary arterial smooth muscle cells were obtained from Cascade Biologics (Portland, OR). Since physiological DG cannot penetrate the plasma membrane of intact cells, we used cell-permeable, sn-2 short-chain, DG analogs: oleoyl,acetylglycerol (1-oleoyl-2-acetyl-sn-glycerol) (OAG), an ester-linked DAG mimetic, and palmityl,acetylglycerol (1-palmityl-2-acetyl-sn-glycerol) (PAG), an ether-linked DG mimetic (purchased from Doosan Serdary Research Laboratories; Toronto, Canada). Both OAG and PAG were reconstituted in ethanol, which served as the vehicle control. Since PAG can be metabolized into platelet-activating factor (1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine), we have previously used a platelet-activating factor receptor blocker (BN-52021) to show that this autacoid does not mediate the inhibitory actions of PAG upon physiological events, such as migration and proliferation (27). We have previously shown that the differential actions of OAG and PAG are due to the differences in the sn-1 bond position (ester vs. ether) and are not due to the length of the fatty acid constituents (27, 32).

Recombinant human PDGF-BB was purchased from R&D Systems (Minneapolis, MN). Anti-PKC-ε, pp38, phospho- (p)MEK, pERK, and ERK antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Akt, pAkt, pp70, pJNK, JNK, p-cRaf, Raf-1, p85 PI3K, p110 PI3K, phospho-myristoylated alanine-rich PKC substrate (pMARCKS), and phospho-phosphatase and tensin homologue deleted on chromosome ten (pPTEN) antibodies were purchased from Cell Signaling Technology (Beverly, MA).

A7r5 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimicotic solution (Invitrogen, Carlsbad, CA). The cells were incubated at 37°C in 5% CO2-95% filtered air.

Western blot analysis.

Western blot analysis was performed as described previously (4) with some modifications. Briefly, after the selected treatment, the cells were washed once with cold PBS followed by the addition of cold lysis buffer containing 1% Triton X-100, 20 mM Tris, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM Na4P2O7, 1 mM β-glycerolphosphate, 1 mM Na3VO4, protease inhibitor cocktail, and 1 mM phenylmethylsulfonylfluoride in ddH2O (pH 7.5) on ice. Cells were lysed for 15 min on ice, and the cell lysate was harvested and centrifuged at 14,000 μg for 10 min at 4°C. The cell lysates were loaded in 4–12% precasted SDS-polyacrylamide gel electrophoresis gradient gels from Invitrogen, and the resolved proteins were transferred to nitrocellulose membranes (Hybond C, GE Healthcare, Piscataway, NJ). The membranes were blocked in 5% nonfat milk in Tris-buffered saline containing 0.1% Tween 20 (TBST) for 1 h and then incubated with the appropriate primary antibody overnight at 4°C. After incubation, the membranes were washed with TBST (3 × 10 min). The membranes were then incubated with the appropriate secondary antibody for 3 h at room temperature. After three more washes with TBST, the protein bands were detected by the enhanced chemiluminescence method from GE Healthcare and quantitated by laser densitometry using the Bio-Rad GS800-calibrated densitometer with Quantity One software.

Transfections.

A7r5 cells were transiently transfected with either Fugene 6 from Roche (Indianapolis, IN) or an electroporation kit from Amaxa Biosystems (Cologne, Germany). For Fugene 6, A7r5 cells were seeded to 50–60% confluency and transfected with a 3:1 ratio of Fugene 6 to DNA. A7r5 cells were transiently transfected with a vector control, wild-type, or kinase-dead PKC-ε mutant construct. The wild-type construct is a full-length PKC-ε. The kinase-dead mutant construct is a full-length PKC-ε with a point mutation in the catalytic domain at the ATP-binding site that renders the enzyme inactive and functions as a dominant-negative mutant. Transfection efficiency was consistently 50–60%, as determined by green fluorescent protein transfections, and was similar for both Fugene and Amaxa transfection strategies. Controls for these transfection experiments included Western blot quantification of HA-tag constructs, total downstream kinase expression, and vector-only and transfection agent-only conditions.

RT-PCR.

A7r5 cells were seeded onto 60-mm tissue culture dishes, grown to 80–90% confluency, and then serum starved in basal media for 24 h. Cells were treated with 10 μM OAG or PAG, followed by PDGF (10 ng/ml) for 4 h before RNA extraction. RNA was extracted using the Qiagen RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. First-strand complementary DNA (cDNA) was transcribed from 1 μg RNA using Invitrogen's Superscript III Reverse Transcriptase according to manufacturer's protocol. The relative expression of p85 regulatory subunit of PI3K (Qiagen, Valencia, CA) was determined by quantitative real-time polymerase chain reaction assay using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA), maintained at the Pennsylvania State College of Medicine Functional Genomics Core Facilities. Relative quantities were calculated using ABI SDS 2.0 RQ software and the 2−ΔΔ{C↓t} analysis method (26) with β-actin as the endogenous control. Final results are given as relative expression.

PI3K assay.

PI3K activity was measured as described previously (41). Lysates from A7r5 cells that were treated with or without 10 ng/ml PDGF were incubated with PI3K p85 antibody for 3 h, followed by incubation with agarose-conjugated protein A. The immunoprecipitated PI3K was treated with 10 μM OAG or PAG. Kinase buffer, [32P]ATP, and phophoinositide were incubated with the immunocomplex for 10 min. The reaction was stopped with the addition of HCl and chloroform-methanol. The lower phase was spotted on a TLC plate precoated with 1% potassium oxalate solution in water-methanol (1:1) and dried. The TLC was developed in a TLC tank containing 100 ml of chloroform-methanol-water-ammonium hydroxide (60:47:11.3:2, vol/vol/vol/vol) solvent system followed by detection of 3′-phosphatidylinositides by autoradiography.

Coimmunoprecipitation of p85 subunit with p110 subunit of PI3K.

Lysates from A7r5 cells were incubated overnight with the immunoprecipitating antibody (p85 subunit of PI3K) conjugated with agarose beads overnight. The lysates were then centrifuged briefly and the supernatants removed. The pellet was washed three times with cold PBS. Sample buffer with reducing agent was added to each tube, and the standard NuPage Western blot analysis protocol was performed. Blotting was performed with the PI3K p110 α-subunit antibody. The blots were stripped and reprobed with p85 for normalization.

Scratch wound assay.

A7r5 cells were plated in 100-mm tissue culture dishes and grown to 90% confluency. The cells were serum starved for 24 h. After the medium was aspirated, the cells were wounded using a 1-ml micropipette tip. The plates were then rinsed with PBS, and serum-free media was added. Cells were treated with either 10 μM OAG or PAG for 30 min followed by treatment with 10 ng/ml PDGF or vehicle. Wound closure was monitored and photographed at 0, 24, and 48 h.

[3H]thymidine proliferation assay.

A7r5 cells were seeded at 3 × 104 cells/well in 24-well plates and grown overnight before 24 h of serum deprivation. Following serum starvation, cells were treated with either 10 μM OAG or PAG for 30 min followed by treatment with 10 ng/ml PDGF or vehicle for an additional 16 h. Cellular proliferation was assayed with the addition of 0.5 mCi/ml [3H]thymidine from MP Biomedicals (Costa Mesa, CA) for the final 4 h of treatment. Cells were washed once with cold 10% trichloroacetic acid (TCA) and then incubated with cold trichloroacetic acid for 20 min. Cells were solubilized with 0.4 N NaOH for 30 min, and [3H]thymidine incorporation into acid-insoluble DNA was assessed with a Beckman LS 6500 scintillation counter.

Caspase-3/7 assay.

A7r5 cells were seeded at 1,000–2,000 cells/well in 96-well plates and grown for 24–48 h in DMEM growth media (Cell Applications). At 70–90% confluency, the media was removed and replaced with basal media containing 0.1% FBS for all treatment groups and 0% FBS for the positive control (100% serum deprived). After 24 h of serum deprivation, cells were treated with OAG or PAG. After 24 h of treatment, apoptosis was measured using the Apo-ONE Homogenous Caspase-3/7 Assay kit (Promega, Madison, WI) according to manufacture's protocol. Briefly, 100 μl of assay substrate/buffer mixture were added to each well, and the plate was subsequently mixed on a plate shaker at 100 rpm for 90 min. Caspase-3/7 activity was determined by measuring the relative fluorescence with a Molecular Devices Gemini XS spectrophotometer at an excitation wavelength of 485 nm and an emission wavelength of 530 nm.

Apoptosis detection.

Apoptosis was assessed and quantified by flow cytometry analysis of annexin V-stained cells using the Vybrant apoptosis assay kit according to the instructions of the manufacturer. Briefly, A7r5 cells were seeded at 1.0 × 106 cells/plate in 100-mm tissue culture dishes and grown for 48 h in culture media containing 10% FBS. At 70–80% confluency, the media was removed and replaced with basal media containing 0.1% FBS. After serum deprivation for 24 h, the cells were treated with OAG or PAG for 24 h. Media deprivation, which was the addition of PBS without any media for 4 h, was used as a positive control. Following treatment, cells were collected and washed once with cold PBS. Pelleted cells were then stained with FITC-labeled annexin V and propidium iodide (PI) according to the instructions of the manufacturer. Labeled cells were immediately analyzed using the Becton Dickinson FACScans for flow cytometry. Viable cells were double negative, early apoptotic cells were positive for annexin V staining and negative for PI staining, and late apoptotic cells were double positive.

Statistical analysis.

One-way analysis of variance with Bonferroni multiple comparison post hoc test and t-tests analyses were performed using GraphPad Prism 4.0 software. A statistically significant difference was reported if P < 0.05. Data are reported as means ± SE from at least n = 3 separate experiments.

RESULTS

Ether-linked DGs inhibit PDGF-stimulated ERK-1 and -2 activation.

We have previously shown that the cell-permeable ether-linked DG mimetic PAG, but not the cell-permeable DAG mimetic OAG, inhibits mesengial cell proliferation (32). We have now investigated the role of ether-linked DG to inhibit promitogenic ERK and Akt signaling cascades, downstream of PKC, in VSM cells. We initially pretreated A7r5 cells for 0.5 h with 10 μM U-0126 (MEK inhibitor) or with either 50 μM LY-294002 or 10 μM wortmannin (PI3K inhibitors), followed by addition of PDGF. We then assessed the phosphorylation/activation of ERK. As shown in Fig. 1, the pretreatment with U-0126 significantly decreased the PDGF activation of ERK, whereas the PI3K inhibitors did not have any effect upon PDGF activation. To validate the PI3K inhibitors, we probed for pAkt and demonstrated a significant decrease in Akt activation when the cells were pretreated with the PI3K inhibitors but not the MEK inhibitor. In data not shown, we demonstrate that both U-0126 and LY-294002 inhibit PDGF-induced cellular proliferation, implicating that MEK/ERK and PI3K/Akt are involved in PDGF-stimulated VSM growth. In supportive data, we have shown that the inactivation of Akt, through Ser34 phosphorylation, also leads to VSM growth arrest (14).

Fig. 1.

Ether-linked diglycerides (DGs) inhibit PDGF-stimulated ERK activation. A7r5 rat aorta smooth muscle cells were pretreated with either the MEK inhibitor (U-0126) or phosphatidylinositol 3-kinase (PI3K) inhibitors LY-294002 (LY) or wortmannin (Wort) for 30 min, followed by the addition of PDGF. Western blot analysis revealed that U-0126 pretreatment prevented PDGF-induced ERK phosphorylation, whereas the PI3K inhibitors had no effect on PDGF-induced ERK phosphorylation. The PI3K inhibitor LY-294002 did, however, elevate basal phospho- (p)ERK levels. Western blot analysis on Akt phosphorylation was performed to ensure that the PI3K inhibitors were efficacious. The PI3K inhibitors blocked PDGF-induced Akt phosphorylation. Data are reported as means ± SE from n = 4 separate experiments. Con, control. *P < 0.05.

As the ERK-MAPK pathway is critical to cell proliferation, we examined the effects of PAG and OAG on PDGF-stimulated ERK activation in A7r5 cells. We have previously shown that PDGF (10 ng/ml) treatment of A7r5 cells leads to maximal phosphorylation of ERK activity within 5–10 min, which correlates to increases in ERK activity (5). Hence, in the present study, we used a time point of 5 min to demonstrate ERK phosphorylation in response to PDGF. We now demonstrate that PAG (Fig. 2A), but not OAG (Fig. 2B), induces significant inhibition of PDGF-induced ERK phosphorylation in a dose-dependent manner. In contrast, there was a slight trend for OAG to increase basal and growth factor-stimulated ERK activity. To confirm these experiments, we ran another set of studies directly comparing 10 μM OAG or PAG in the same experimental design (Fig. 2C). Again, we demonstrate that PAG, but not OAG, significantly reduced ERK phosphorylation. In all subsequent experiments, we used 10 μM PAG or OAG as a therapeutic in vitro dose. Since A7r5 cells are a clonal cell line derived from fetal rat aorta, we also examined the effect of OAG and PAG in low-passage human coronary arterial smooth muscle cells. We demonstrate that PAG, but not OAG, significantly reduced ERK phosphorylation (Fig. 2D), confirming that PAG has similar effects in arterial cells derived from both rat and human origin.

Fig. 2.

Ether-linked DGs inhibit PDGF-stimulated ERK activation. A and B: A7r5 cells were treated with either 1-oleoyl-2-acetyl-sn-glycerol (OAG) or 1-palmityl-2-acetyl-sn-glycerol (PAG) at varying concentrations for 30 min, followed by PDGF stimulation for 5 min. PAG, but not OAG, inhibited PDGF-induced ERK phosphorylation in a dose-dependent manner. C: in a separate experimental design, A7r5 cells were treated with 10 μM OAG or PAG for 30 min and then stimulated with PDGF (10 ng/ml). PAG, but not OAG, inhibited PDGF-induced ERK phosphorylation. Data are reported as means ± SE from n = 3 to 4 separate experiments. D: human coronary arterial smooth muscle cells (HCASMCs) were treated with 10 μM OAG or PAG for 30 min and then stimulated with PDGF (10 ng/ml). PAG, but not OAG, inhibited PDGF-induced ERK phosphorylation. Data are reported as means ± SE from n = 4 separate experiments. Veh, vehicle. *P < 0.05.

Our studies demonstrated that PAG pretreatment leads to the inhibition of PDGF-induced ERK activation in A7r5 VSM cells. Upstream protein kinases for ERK, including MEK-1/2 and c-Raf (Raf-1) were next examined. Raf-1 and MEK were activated in response to PDGF (Fig. 3A). Neither DG species had an effect on PDGF-induced phosphorylation of Raf-1. However, PAG was able to inhibit the PDGF-induced activation of MEK-1/2, mimicking the effects on ERK. Surprisingly, OAG, to some extent, also inhibited PDGF-induced activation of MEK. The inhibition of MEK, but not Raf-1, may be, in part, due to the translocation of MEK to the plasma membrane upon PKC-ε translocation and activation in VSM cells (28, 35). We also examined other family members of the MAPKs including p-38 MAPK (Fig. 3A) and JNK (Fig. 3B). PDGF significantly activated p38 MAPK. We found that neither OAG nor PAG had a significant effect on PDGF-activated p38. The JNK antibody used detects endogenous levels of both p46 and p54 SAPK/JNK dually phosphorylated at Thr183 and Tyr185 (JNK-1 and JNK-2). PDGF induced phosphorylation of JNK-2 but did not change JNK-1 levels. We found that neither DG species had a significant effect on PDGF-induced JNK phosphorylation. We also found that neither OAG nor PAG had a significant effect on PDGF-activated p70 S6 kinase (Fig. 3B), suggesting that the action of PAG on cell signaling pathways is relatively MEK/ERK specific.

Fig. 3.

Ether-linked DGs reduce PDGF-induced MEK, but not Raf, phosphorylation. A7r5 cells were treated with 10 μM OAG or PAG, followed by 10 ng/ml PDGF. Western blot analysis revealed that PAG and OAG reduced PDGF-induced MEK (A), but not Raf (A), p38 (A), p70 (B), JNK (B) phosphorylation. As loading controls, PDGF had no effect on total Raf-1 (A), JNK (B), and PKC-ε (B) expression. Representative blots of 3 separate experiments are shown. GF, PDGF.

PKC-ε is a necessary component for PAG inhibition of PDGF-stimulated A7r5 cell growth and ERK activation but not Akt activation.

Kester's laboratory (27) has previously shown that ether-linked DGs inhibit growth in both G protein-linked receptor- and tyrosine kinase receptor-stimulated mesangial cells by attenuating PKC activity (27). We thus examined whether ether-linked DG mimetics were functioning in a PKC-dependent fashion in A7r5 VSM. As shown in Fig. 4A, PDGF caused an increase in the phosphorylation of MARCKS, a major PKC substrate. When the cells were pretreated with PAG, PDGF activation of pMARCKS was reduced, suggesting that PAG signaling events are PKC dependent. To confirm the requirement for PKC in the PDGF activation of ERK, we pretreated A7r5 cells with 5 μM bisindolylmalemide I, the classical and novel PKC inhibitor. Upon PKC inhibition, PDGF was no longer able to activate ERK (data not shown), suggesting that PKC is necessary for PDGF-induced ERK activation.

Fig. 4.

PKC-ε is a necessary component for PAG-inhibition of PDGF-stimulated A7r5 cell growth and ERK activation but not Akt activation. A: A7r5 cells were treated with 10 μM PAG, followed by PDGF stimulation. PAG inhibited PDGF-induced myristoylated alanine-rich PKC substrate (MARCKS) phosphorylation. Since the PDGF group is set to 100%, the raw values for the groups are as follows: (in arbitrary units) control, 473 ± 6.8; PDGF, 4,839 ± 197.4; PAG: 693 ± 17.5; PAG + PDGF: 1,696 ± 585.2. B: A7r5 cells were transfected with either a wild-type (WT) or kinase-dead PKC-ε vector. A7r5 cells were pretreated with or without OAG (10 μM) or PAG (10 μM) for 30 min and then stimulated with PDGF (10 ng/ml). [3H]thymidine incorporation revealed that PDGF increased cellular proliferation, and PAG inhibited PDGF-stimulated growth in the WT transfected cells. The same results were seen with the vector control (VC) transfected cells (data not shown). After kinase-dead transfections, PDGF was no longer able to increase thymidine incorporation, and PAG was unable to further reduce PDGF-induced thymidine incorporation. C: A7r5 cells were transfected with either a vector control or kinase-dead PKC-ε vector, followed by DG treatment. Western blot analysis revealed that PDGF significantly increased the phosphorylation of ERK, whereas PAG, but not OAG, inhibited PDGF-induced phosphorylation in the WT transfected cells. In the kinase-dead transfected cells, PDGF was no longer able to increase the phosphorylation of ERK, and PAG was unable to further reduce PDGF-activated ERK. Kinase-dead PKC-ε transfected cells had no affect on the activation of Akt. PAG inhibited the PDGF-induced activation of Akt in both the WT and kinase-dead transfected cells. Controls included total ERK, as well as HA-tagged PKC-ε. Data are reported as means ± SE from n = 3–6 separate experiments. DN, dominant negative. *P < 0.05.

To further examine this PKC dependence, we determined whether PAG exerted its cell growth inhibitory actions through the inactivation of PKC-ε. We transfected A7r5 cells with either wild-type or kinase-dead/dominant-negative PKC-ε and assessed cellular proliferation by means of [3H]thymidine incorporation (Fig. 4B) and proliferating cell nuclear antigen levels (data not shown). In the wild-type PKC-ε transfections, PAG, but not OAG, reduced PDGF-induced cellular proliferation as expected. After dominant-negative PKC-ε transfections, PDGF no longer increases thymidine incorporation and PAG is unable to further reduce PDGF-induced thymidine incorporation. the inhibition of endogenous PKC-ε through the overexpression of a kinase-dead mutant PKC-ε thus mimics the effects of PAG.

To definitively determine whether PDGF-activated ERK requires PKC-ε in A7r5 cells, we transfected cells with either a vector control or kinase-dead PKC-ε followed by Western blot analysis of pERK (Fig. 4C). In vector control-transfected cells, PDGF significantly increased the phosphorylation of ERK, whereas PAG, but not OAG, inhibited PDGF-induced phosphorylation, consistent with Fig. 2C. In the kinase-dead transfected cells, PDGF was no longer able to increase the phosphorylation of ERK and PAG was unable to further reduce PDGF-activated ERK. In data not shown, transfection with wild-type versus vector control did not change PDGF-stimulated ERK. In Fig. 4C, we also determined whether PKC-ε has an effect on Akt phosphorylation. We observed that kinase-dead PKC-ε-transfected cells had no affect on the PDGF-induced activation of Akt. PAG was still able to inhibit the PDGF-induced activation of Akt in both the wild-type and dominant-negative transfected cells. Taken together, we demonstrate that PAG reduces ERK activity in a PKC-ε-dependent manner, whereas PAG reduces Akt activity in a PKC-ε-independent manner. We thus next investigated a PKC-ε-independent mechanism responsible for the ability of PAG to reduce Akt activity.

Ether-linked DGs inhibit Akt signaling via inhibition of p85/p110 PI3K interactions.

We initially determined whether PAG reduced PDGF-induced Akt Ser473 phosphorylation, a surrogate to Akt activity. As shown in Fig. 5A, PDGF increased Akt phosphorylation, whereas PAG significantly inhibited PDGF-induced activation. The ester-linked DG, OAG, had no significant effect on PDGF-induced phosphorylation of Akt.

Fig. 5.

Ether-linked DGs inhibit PDGF-stimulated PI3K/Akt activation. A: A7r5 cells were treated with 10 μM OAG or PAG, followed by 10 ng/ml PDGF. Western blot analysis revealed that PAG, but not OAG, inhibited PDGF-induced Akt phosphorylation. B: A7r5 cells were pretreated with 100 nM okadaic acid for 30 min, followed by DG treatment. Western blot analysis revealed that okadaic acid had no effect on the ability of PAG to inhibit PDGF-induced Akt phosphorylation. C: A7r5 cells were treated with OAG or PAG. Western blot analysis revealed that phosphorylation of phosphatase and tensin homologue deleted on chromosome 10 (PTEN) did not change upon treatment. D: PI3K was immunoprecipitated from PDGF-BB treated or untreated A7r5 vascular smooth muscle (VSM) cells with an antibody to the p85 subunit. An in vitro reconstitution assay using exogenous phosphatidylinositols and [32P]ATP was performed. Cell-free immunoprecipitates were treated with PAG. PAG markedly decreased PDGF-stimulated PI3K activity by reducing the radioactivity associated with the band corresponding to 3′-phosphatidylinositide species. Data are reported as means ± SE from n = 3–5 separate experiments. Ptdins-3-P, phosphatidylinositol 3-phosphate; Ptdins-3,4-P2, phosphatidylinositol 3,4-bisphosphate; Ptdins-3,4,5-P3, phosphatidylinositol 3,4,5-trisphosphate. *P < 0.05.

As it is well established that protein phosphatase type 2A (PP2A) has the ability to inactivate Akt (22, 37, 38), we investigated whether PAG diminished Akt phosphorylation in the presence or absence of a PP2A inhibitor, okadaic acid. We treated A7r5 cells with the okadaic acid for 30 min, followed by the addition of OAG or PAG in the presence or absence of PDGF. As shown in Fig. 5B, the PP2A inhibitor had no effect on the ability of PAG to inhibit PDGF-induced Akt phosphorylation, suggesting that PAG is not affecting Akt phosphorylation via the activation of a protein phosphatase. Akt activity can also be regulated by reduction of PI3 lipid messengers as a result of lipid-dependent phosphatases. It is known that the lipid phosphatase PTEN is a major negative regulator of the PI3K/Akt signaling pathway (3, 18). We therefore investigated whether PAG was inducing the dephosphorylation of PI3 lipids via activating this lipid phosphatase (Fig. 5C). PTEN phosphorylation, which correlates with activity, did not change upon treatment, suggesting that PAG is decreasing PI3K/Akt activity independent of lipid phosphatases (Fig. 5C), as well as protein phosphatases (Fig. 5B).

We next investigated the role of DGs to directly regulate PI3K, an upstream mediator of Akt. Here we performed a PI3K assay on immunoprecipitated PI3K, which was then directly treated with PAG. PI3K was immunoprecipitated from untreated or PDGF-treated A7r5 VSM cells with an antibody to the p85 subunit. An in vitro reconstitution assay using exogenous phosphatidylinositols and [32P]ATP was performed. This assay is linear with time and with the concentration of the exogenous substrate (data not shown). PAG markedly decreased PDGF-stimulated PI3K activity as evidenced by a reduction of radioactivity associated with the bands corresponding to 3′-phosphatidylinositides (Fig. 5D). The PI3K inhibitor, LY-294002, was used as a positive control, and lysis buffer alone (no immunoprecipitate) was used as a negative control.

To further examine the role of PAG-induced inhibition of PI3K, we investigated PI3K protein and RNA expression. As shown in Fig. 6A, Western blot analysis revealed no significant changes of p110 or p85 protein expression upon treatment. To further evaluate the PI3K subunits, RT-PCR analysis revealed no significant changes in RNA expression upon treatment (Fig. 6B). Although there was no change in the protein or RNA expression of either PI3K subunit, there was a significant change in the interaction between the two subunits as shown through a coimmunoprecipitation assay (Fig. 6C). PDGF significantly increased the interaction between p85 and p110, whereas PAG inhibited this interaction. This suggests that PAG inhibited PI3K/Akt activation by diminishing the interaction of the two subunits (p85 and p110) and not through altered gene or protein expression.

Fig. 6.

Ether-linked DG inhibits Akt signaling via inhibition of p85/p110 PI3K interactions. A: A7r5 cells were treated with 10 μM OAG or PAG for 30 min, followed by PDGF (10 ng/ml) stimulation for an additional 5 min. Western blot analysis revealed no significant changes of p110 or p85 protein expression upon treatment. B: A7r5 cells were pretreated with either OAG or PAG for 30 min in the presence or absence of PDGF for an additional 4 h, followed by RT-PCR analysis. RT-PCR analysis revealed no significant changes in RNA expression upon treatment. C: A7r5 cells were treated with PAG for 30 min, followed by PDGF or vehicle stimulation for an additional 5 min. The lysates were incubated with PI3K p85 antibody conjugated with agarose beads. Western blot analysis was performed with the PI3K p110 α-antibody. PDGF increased the interaction between p85 and p110, whereas PAG inhibited this interaction. Controls for these coimmunoprecipitations included using cell lysates that are not immunoprecipitated (IP) with p85. In addition, the immunoprecipitated p85 was immunoblotted (IB) with p85 antibody to ensure equal immunoprecipitation and loading between conditions (data not shown). Data are reported as means ± SE from n = 3 separate experiments. *P < 0.05.

Ether-linked DGs inhibit VSM migration.

To validate the physiological consequences of PAG-inhibited MEK/ERK and PI3K/Akt signaling cascades, we investigated the actions of PAG on VSM migration and proliferation, phenotypes exhibited in restenotic or atherosclerotic arterial models (10). To assess the physiological effect of DGs on rat aorta VSM cell migration, we used a scratch wound assay. A7r5 cells were pretreated with either OAG or PAG, followed by treatment with PDGF. We again used a dose of 10 μM OAG and PAG, consistent with the dose-response data in previous work (27), as well as the dose-response experiments shown in Fig. 2. As shown in Fig. 7, PDGF significantly increased migration compared with vehicle-treated A7r5 cells. OAG was able to augment the migration ability of PDGF, even though OAG did not significantly induce migration by itself. PAG, unlike OAG, limited PDGF-induced VSM migration.

Fig. 7.

Ether-linked DGs inhibit VSM migration. Cellular migration was assessed by means of a scratch wound assay. Cells were pretreated with 10 μM OAG or PAG and then stimulated with PDGF (10 ng/ml). After 48 h, PDGF increased VSM migration, which was augmented with the addition of OAG. On the other hand, PAG inhibited PDGF-stimulated migration. There were no significant changes in wound closure after 24 h of treatment (data not shown). Six individual fields were photographed from each treatment, and the distance between cell scratch layers was measured in the ImageJ program after calibration using the known distance of the hemacytometer squares. Calculated distances for time 0 was set to 100% and compared with distances for 24 (data not shown) and 48 h. The inverse of the percent wound closure was plotted as percent migration. Data are reported as means ± SE from n = 6 representative fields. *P < 0.05; **P < 0.01.

Ether-linked DGs induce VSM growth arrest without apoptosis.

The consequence of PAG inhibition of VSM migration could be consistent with diminished VSM mitogenesis. To investigate this, we used [3H]thymidine incorporation into acid-insoluble DNA as a measure of cell proliferation. When A7r5 cells were pretreated with PAG, but not OAG, there was a reduction in PDGF-stimulated [3H]thymidine incorporation (Fig. 8A). OAG at the higher dose (10 μM) increased basal thymidine incorporation. PAG and the lower-dose OAG (1 μM) did not significantly effect basal VSM cell proliferation.

Fig. 8.

Ether-linked DGs induce VSM growth arrest without apoptosis. A: cellular proliferation was assessed by means of incorporation of [3H]thymidine in acid-insoluble DNA in A7r5 cells. Cells were pretreated with OAG (1 and 10 μM) or PAG (1 and 10 μM), and then stimulated with PDGF (10 ng/ml). PDGF-increased cellular proliferation and PAG, but not OAG, inhibited PDGF-stimulated growth. RFU, relative fluorescent unit. B: apoptosis was measured by caspase-3/7 assay in A7r5 cells in 0.1% reduced serum conditions. Neither OAG (10 μM) nor PAG (10 μM) induced caspase-3/7 activation after 24 h. The positive control was the presence of total serum deprivation (Serum Dep; 0% FBS). C and D: apoptosis was assessed by fluorescent-activated cell sorting analysis of annexin V in 0.1% reduced serum conditions. After 24 h of treatment, most of the cells were annexin V-FITC and propidium iodide (PI) negative and hence were viable and not undergoing apoptosis. Media deprivation (Media Dep), the addition of PBS without any media, significantly increased cells undergoing apoptosis (annexin V positive and PI negative), and necrosis (annexin V positive and PI positive) (C). Neither 10 μM OAG- or PAG-induced apoptosis or necrosis compared with control cells (D). Data are reported as means ± SE from n = 3–6 separate experiments. *P < 0.05.

Since growth-arrested cells often undergo apoptosis, we assessed the effects of ether-linked DGs upon apoptosis using a caspase-3/7 assay. As shown in Fig. 8B, neither OAG nor PAG at 10 μM induced caspase-3/7 activation compared with vehicle-treated cells. To demonstrate the capability to induce apoptosis, we used short-term media deprivation where PBS, without any serum, was added to the wells as a positive control. To further confirm these results, we examined apoptosis and necrosis through 4,6-diamidino-2-phenylindole staining (data not shown) and through fluorescent-activated cell sorting analysis of annexin V and PI (Fig. 8, C and D). In these assays, neither OAG nor PAG induced apoptosis or necrosis at 10 μM concentrations after 24 h, compared with control cells. Taken together, these data suggest that PAG, but not OAG, limits VSM migration and proliferation without inducing apoptosis.

Ether-linked DGs inhibit VSM growth and ERK phosphorylation in a therapeutic model.

Our observations that PAG inhibits VSM growth and ERK phosphorylation have been in an in vitro system where OAG and PAG were used as a pretreatment followed by the addition of PDGF. However, in vivo atherosclerotic and restenotic systems already have mitogenic responses underway. We therefore used a therapeutic in vitro model to simulate a natural effect. A7r5 cells were pretreated with PDGF, followed by the addition of OAG or PAG. As shown in Fig. 9, the addition of PAG, but not OAG, inhibited both ERK phosphorylation and VSM proliferation as observed through the inhibition of proliferating cell nuclear antigen.

Fig. 9.

Ether-linked DGs inhibit VSM growth and ERK phosphorylation in a therapeutic model. A7r5 cells were pretreated with 10 ng/ml PDGF for 15 min, followed by the addition of 10 μM OAG or PAG for 24 h. PAG, but not OAG, inhibited PDGF-induced ERK phosphorylation and proliferation through inhibition of proliferating cell nuclear antigen (PCNA). Data are reported as means ± SE from n = 4 separate experiments. *P < 0.05.

DISCUSSION

Our data indicate that the ether-linked DG (PAG), but not the ester-linked DG (OAG), inhibited PDGF-induced VSM migration (Fig. 7) and/or mitogenesis via cell-cycle growth arrest (Fig. 8A) and not apoptosis (Fig. 8, B–D). We further demonstrated that both pretreatment and posttreatment of PAG, but not OAG, consistently reduced promitogenic and prosurvival ERK (Figs. 2 and 9) in both rat and human models. Additionally, PAG inhibited Akt (Fig. 5)-signaling cascades associated with stenotic vascular injury (10). We provided evidence for both ERK and Akt signaling mechanisms responsible for PAG-induced growth arrest in VSM. We demonstrated that PAG-induced inhibition of ERK, but not Akt, signaling is PKC-ε dependent. Furthermore, PAG inhibited PI3K/Akt signaling through a novel mechanism involving disruption of the interaction between p110 and p85 PI3K subunits.

We examined putative PKC-ε-dependent, downstream signaling kinases responsible for PAG-induced VSM cell growth arrest. We have shown that ether-linked DGs preferentially inhibited MEK and ERK, but not cRaf, JNK, p38, or p70. These results suggest that ether-linked DGs inhibited PKC-ε activation of the ERK-1/2 cascade at the level of MEK. This is consistent with our previous studies demonstrating that ceramide inhibited MEK activity through the inhibition of PKC-ε (5). Even though OAG had, for the most part, no effect upon PDGF signaling, the modest effects of OAG on VSM PDGF-induced proliferation may reflect the dichotomous actions of OAG upon MEK (Fig. 3) and ERK (Fig. 2).

Evidence for PKC isotypes being upstream of PI3K has not been firmly established. However, phorbol esters have been shown to activate PI3K activity (39). We have shown that ether-linked DG inhibition of PI3K/Akt activity is not dependent on PKC-ε (Fig. 4C) but is dependent on the interaction of PI3K p110 and p85 subunits (Fig. 6C). The protein/protein interactions that regulate the interactions between the p110 catalytic and the p85 adapter subunits of PI3K have been characterized (13). The multiple SH2/SH3 and Bcl-2 homology domains that regulate p85 interactions with multiple signaling complexes that coordinate the assembly and activity of PI3K have also been studied (9, 19, 24). Yet, even though these protein/protein interactions are defined, the putative binding site for PI lipids on the p110 catalytic subunit has not been rigorously proven. Moreover, the regulation of the interaction between PI lipids and the p110 catalytic subunit is poorly understood. It is suggested that a phosphatidylinositol kinase (PIK) domain (also known as HR2) could be involved in substrate recognition and presentation to the catalytic subunit (11). In addition, a substrate binding domain (SBR) has been proposed to be a phosphatidylinositol ring-binding region in the HR1 domain of class 1 p110 (44). Thus the PIK and the SBR domains need to be further investigated to determine whether they are possible sites at which ether-linked DGs regulate PI3K.

Ether-linked glycerolipids are widely distributed in nature and are required for normal physiological processes (17). In fact, in diseases where ether-linked phospholipids are not generated, such as cerebro-hepato-renal (Zellweger) syndrome and rhizomelic chondrodysplasia punctata, inflammation and/or mitogenesis is observed (6, 30, 40). On the flip side, exogenous ether-linked phospholipids and alkyl-lysophosphatidlycholine can induce apoptosis in cancer models (2, 43). Our studies now indicate that exogenous ether-linked DGs can induce growth arrest, but not apoptosis, in vascular/aortic cell lines. It can be further inferred that endogenous ether-linked DGs could be a compensating, yet failing, mechanism to limit PDGF-induced proliferation and/or migration in proinflammatory vascular diseases. Thus ether-linked DG inhibition of ERK and Akt signaling cascades may, in part, induce a growth-arrested VSM phenotype, despite an inflammatory milieu. Our studies also implicate ether-linked DGs as a possible growth-arresting therapeutic for inflamed and proliferating VSM as evidenced in atherosclerosis and restenosis. In an analogous fashion to ceramide-coated balloon embolectomy catheters (10), cell-permeable ether-linked DGs have the potential to be delivered from vascular devices.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant RO1-HL-066371 and RO1-HL-076789 (to M. Kester).

Acknowledgments

We acknowledge the support of the Cell Science/Flow Cytometry Core Facility of the Section of Research Resources, Pennsylvania State College of Medicine.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract