Ras/Mitogen-activated Protein Kinase (MAPK) Signaling Modulates Protein Stability and Cell Surface Expression of Scavenger Receptor SR-BI

Abstract

The mitogen-activated protein kinase (MAPK) Erk1/2 has been implicated to modulate the activity of nuclear receptors, including peroxisome proliferator activator receptors (PPARs) and liver X receptor, to alter the ability of cells to export cholesterol. Here, we investigated if the Ras-Raf-Mek-Erk1/2 signaling cascade could affect reverse cholesterol transport via modulation of scavenger receptor class BI (SR-BI) levels. We demonstrate that in Chinese hamster ovary (CHO) and human embryonic kidney (HEK293) cells, Mek1/2 inhibition reduces PPARα-inducible SR-BI protein expression and activity, as judged by reduced efflux onto high density lipoprotein (HDL). Ectopic expression of constitutively active H-Ras and Mek1 increases SR-BI protein levels, which correlates with elevated PPARα Ser-21 phosphorylation and increased cholesterol efflux. In contrast, SR-BI levels are insensitive to Mek1/2 inhibitors in PPARα-depleted cells. Most strikingly, Mek1/2 inhibition promotes SR-BI degradation in SR-BI-overexpressing CHO cells and human HuH7 hepatocytes, which is associated with reduced uptake of radiolabeled and 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyane-labeled HDL. Loss of Mek1/2 kinase activity reduces SR-BI expression in the presence of bafilomycin, an inhibitor of lysosomal degradation, indicating down-regulation of SR-BI via proteasomal pathways. In conclusion, Mek1/2 inhibition enhances the PPARα-dependent degradation of SR-BI in hepatocytes.

Introduction

Anti-atherosclerotic properties of HDL and the major HDL apolipoprotein, apoA-I, are believed to include their ability to induce signaling events that promote cholesterol export from peripheral cells to the liver for disposal. However, the signal transduction pathways that contribute to stimulate reverse cholesterol transport in macrophages and hepatocytes are not fully understood (1–3). HDL binding to receptors such as SR-BI⁶ activates various cellular processes, including endothelial nitric-oxide synthase activation in endothelial cells (1–3) and proliferation in smooth muscle cells (4) but also cell surface localization of SR-BI in hepatocytes (5). Downstream targets of HDL include Src family kinases, phospholipase C and D, Ras, phosphatidylinositol 3-kinase (PI3K), Akt, the mitogen-activated protein kinase (MAPK) pathway (Mek1/2 and Erk1/2), and Rac/Rho GTPases (1–8). Other kinases implicated in HDL- or apoAI-inducible cholesterol transport include protein kinase C (PKC), protein kinase A (PKA), c-Jun N-terminal kinase (JNK), and p38 MAPK (9–14).

Alternatively, signaling pathways can modulate the activity of nuclear receptors, including PPAR and LXR, to alter the ability of cells to transport cholesterol (15–18). Indeed, post-translational phosphorylation via various kinases, including Erk1/2, can alter PPARα and PPARα co-activator activity in a ligand-dependent and -independent manner (17–25). Recent findings link Erk1/2 kinases with nuclear receptors and lipid export via ABCA1. First, enhanced Erk1/2 signaling increases ABCA1 expression and ABCA1-mediated phospholipid efflux via up-regulation of PPARα levels in lung epithelial cells (26). Second, inhibition of Erk1/2 and activation of LXR synergistically induce macrophage ABCA1 expression and cholesterol efflux (27). Third, Mek1/2 inhibition is involved in the regulation of PPARγ- and LXRβ-dependent ABCA1 protein degradation in HepG2 cells (28). These findings indicate that Erk1/2 kinases might exert opposite effects on nuclear receptor-inducible lipid transport depending on the cell type analyzed.

SR-BI is another PPARα target gene that is up-regulated by PPARα agonists in human and mouse macrophages to promote HDL-inducible cholesterol efflux (29). Although the role of SR-BI for cholesterol efflux in peripheral cells is not fully understood (2, 30, 31), SR-BI is also highly expressed in liver to regulate hepatic uptake of HDL cholesteryl esters for excretion into bile (2, 32–35). Surprisingly, in hepatocytes PPARα activation induces SR-BI protein degradation and reduces SR-BI cell surface expression (36, 37). The mechanisms that down-regulate hepatic SR-BI have yet to be fully elucidated. Recently, the PDZ domain containing protein PDZK1 was identified to interact and stabilize SR-BI cell surface expression in mouse hepatocytes (38–40). Loss of hepatic SR-BI in PDZK1 KO mice suggest that PDZK1 is essential for maintaining hepatic SR-BI levels (40). However, PPARα-inducible SR-BI degradation in mouse liver appears independent of PDZK1 (37).

We and others identified the HDL-induced, PKC- and G-protein-dependent activation of Ras, Raf-1, Mek1/2, and Erk1/2, commonly known as the Ras/MAPK pathway (6–8, 41). We also showed that HDL-induced Ras/Raf-1 signaling via SR-BI represents an alternative and PKC-independent signaling cascade for MAPK activation (6, 8). Earlier attempts did not reveal a potential role for MAPKs in apoA-I-dependent cholesterol efflux (41). However, based on the link between Erk1/2 and PPARα (17–26), we hypothesized that Erk1/2 could impact on PPARα-dependent SR-BI expression through modulating SR-BI mRNA expression or protein stability.

In this study, we show that MAPK inhibition down-regulates SR-BI in CHO and HEK293 cells treated with PPARα agonists, whereas enhanced Ras/MAPK activity increases PPARα Ser-21 phosphorylation and SR-BI expression levels, respectively. This correlates with MAPK inhibition reducing SR-BI activity, as judged by reduced cholesterol efflux onto HDL. Most relevant for the function of SR-BI in the liver, Ras/MAPK inhibition reduces SR-BI protein stability in HuH7 hepatocytes, which correlates with reduced uptake of [³H]cholesteryl ester and DiI-labeled HDL. Mek1/2 inhibition most likely promotes proteasomal degradation of SR-BI, whereas levels of the SR-BI adaptor PDZK1 remain unaffected. Thus, Erk kinases appear to target PPARα-dependent degradation pathways to regulate hepatic SR-BI protein stability. The role of the Ras/MAPK pathway to alter SR-BI protein levels and fine-tune hepatic cholesterol homeostasis is discussed.

EXPERIMENTAL PROCEDURES

Reagents and Antibodies

Nutrient mixture Ham's F-12, DMEM, RPMI 1640, geneticin, glutathione, 2-mercaptoethanol, cycloheximide, fenofibrate (FF), Wy-14643, MK886, 12-O-tetradecanoylphorbol-13-acetate, bafilomycin, lactacystin, and paraformaldehyde were from Sigma. 1,1′-Dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine (DiI) was from Molecular Probes. Endothelial basal medium was from Lonza. [³H]Cholesterol was from Amersham Biosciences. PD98059, U0126, and Mowiol were from Calbiochem. Horseradish peroxidase (HRP)-labeled antibodies and SDS-PAGE molecular weight markers were from Bio-Rad. Polyclonal anti-SR-BI was from Novus. Polyclonal anti-Ser(P)-21 PPARα and anti-PDZK1 were from Abcam. Rabbit polyclonal anti-PPARα and mouse monoclonal anti-H-Ras were from Santa Cruz Biotechnology. Mouse monoclonal anti-Pan Ras, rabbit polyclonal anti-β-actin, and anti-caveolin were from BD Transduction Laboratories. Rabbit polyclonal antibodies against activated Mek1/2 (P-Mek1/2), Erk1/2 (P-Erk1/2), Akt (P-Akt), total Mek1/2, total Erk1/2, total Akt, and Lamin A/C were purchased from Cell Signaling. Monoclonal anti-LBPA was kindly provided by Dr. J. Gruenberg (Geneva, Switzerland). Expression vectors encoding constitutively active H-Ras (H-RasG12V), K-Ras (K-RasG12V), Mek1 (Mek215-DD), dominant-negative Erk1 (DN-Erk1), and PPARα were kindly provided by John F. Hancock (Dallas, TX), Brian Gabrielli (Brisbane, Australia), Charles Hii (Adelaide, Australia), and Bart Staels (Lille, France), respectively. CHOldlA− and CHO cells overexpressing SR-BI (CHO-SRBI) (31) were kindly provided by Monty Krieger (Cambridge, MA). High density lipoproteins (HDL₃, density 1.125–1.21 g/ml) were isolated from the plasma of normolipidemic volunteers by sequential density gradient ultracentrifugation as described previously (6–8). Recombinant A85K mutant RBD protein fused to glutathione S-transferase (GST-A85K-RBD) was expressed in Escherichia coli strain BL21 pLysE and purified by glutathione-Sepharose chromatography as described previously (6–8).

Cell Culture

CHOwt, CHOldlA, and CHO-overexpressing SR-BI (CHO-SRBI) were grown in Ham's F-12, HEK293 in DMEM, HuH7 in DMEM, and F-12 (1:1), THP1 in RPMI 1640 medium, and bovine aortic endothelial cells in endothelial basal medium together with 10% fetal calf serum (FCS), l-glutamine (2 mm), penicillin (100 units/ml), and streptomycin (100 μg/ml) at 37 °C, 5% CO₂. THP1 monocytes were differentiated with 2 nm 12-O-tetradecanoylphorbol-13-acetate for 24 h before treatment with MEK inhibitors. For transient transfections with expression vectors encoding constitutively active H-Ras (HRasG12V), K-Ras (K-RasG12V), Mek1 (Mek215-DD), and dominant-negative Erk1 (DN-Erk1), 1–2 × 10⁵ cells were transfected with 1.5 μg of Qiagen-purified DNA and 6 μl of Lipofectamine 2000 (Invitrogen) as described previously (7, 8).

Real Time RT-PCR

Total RNA from HEK293 cells was extracted using the TRIzol and RNeasy system (Macherey-Nagel, Germany) according to the manufacturer's instructions. 1 μg of RNA was reverse-transcribed using the High Capacity cDNA archive kit (Applied Biosystems) as per the manufacturer's instructions. Real time RT-PCR was performed as described previously (42). Assay-on-Demand primer sets to amplify cDNA fragments encoding human SR-BI and TATA box-binding protein sequences were from Applied Biosystems. Relative SR-BI expression was calculated by normalization to the housekeeper mRNA (TATA box-binding protein) as described previously (43).

RNAi-mediated Inhibition of PPARα

PPARα knockdown studies in HEK293 cells were performed with a mixture of three SureSilencing shRNA plasmids (SABiosciences) designed to specifically knock down expression of human PPARα. 1–2 × 10⁵ cells were transfected in 2 ml of medium with 1.5 μg of shRNA plasmids targeting human PPARα at positions 552–572 (5′-ggagcattgaacatcgaatgt-3′), 954–974 (5′-atgggtttataactcgtgaat-3′), and 1273–1293 (5′-tcaggaaaggccagtaacaat-3′) and Lipofectamine 2000 as described above. Studies were conducted after 72 h when PPARα depletion was most significant. Scrambled shRNA served as negative control (5′-ggaatctcattcgatgcatac-3′).

Cholesterol Efflux and Uptake Assays

For the determination of HDL₃-induced cholesterol efflux, 2–5 × 10⁵ cells (in triplicate) were labeled overnight with [³H]cholesterol (2 × 10⁶ cpm/ml) as described previously (44). Noninternalized radioactivity was removed by extensive washing with PBS. Cells were incubated in Ham's F-12, 0.1% BSA ± 50 μg/ml HDL₃ for 4–8 h, respectively. The media were harvested; cells were lysed in 0.1 n NaOH, and the total cellular protein was determined (45). The radioactivity in the media and cell lysate was determined by scintillation counting (44, 46). The ratio of released and cell-associated radioactivity was determined and is given in %.

For uptake assays, HDL₃ containing [³H]cholesteryl esters were prepared by incubating [³H]cholesterol-HDL₃ with lecithin:cholesterol acyltransferase at 37 °C for 24 h as described previously (47). 2–5 × 10⁵ CHO-SRBI cells (in triplicate) were preincubated overnight ± 20 μm Wy-14643 and 10 μm PD98059. Cells were washed, incubated with Ham's F-12, 0.1% BSA for 30 min, followed by addition of 10 μg/ml [³H]cholesteryl ester-HDL₃ for 2 h. Cells were washed extensively and lysed in 0.1 n NaOH. The cell-associated radioactivity was quantified and normalized to total cell protein as above (44–46).

Measurement of Ras Activation

The capacity of Ras-GTP to bind RBD (Ras-binding domain of Raf-1) was used to measure Ras activity (7, 48). CHOldlA− and CHO-SRBI cells (2 × 10⁶) were incubated for 3 min ± 50 μg/ml HDL₃ and harvested in lysis buffer (20 mm Tris-HCl, pH 7.5, 2 mm EDTA, 100 mm NaCl, 5 mm MgCl₂, 1% (v/v) Triton X-100, 5 mm NaF, 10% (v/v) glycerol, 0.5% (v/v) 2-mercaptoethanol, 0.1 mm Na₃VO₄, and protease inhibitors). After centrifugation at 10,000 × g, the protein concentration of the cleared cell lysate was determined. Then 600 μg of cellular protein was incubated for 2 h at 4 °C with glutathione-Sepharose-4B beads pre-coupled to the A85K-RBD mutant (7, 48). Beads were washed four times in lysis buffer; bound proteins were solubilized in Laemmli loading buffer and electrophoresed on 12.5% SDS-PAGE. Proteins were transferred and immunoblotted using anti-H-Ras antibody.

Fluorescence Microscopy

CHO-SRBI cells were grown on coverslips, treated ± bafilomycin (50 ng/ml), lactacystin (25 ng/ml), and PD98059 (10 μm) overnight, fixed with 4% paraformaldehyde, permeabilized, and incubated with first and secondary antibodies as described previously (44, 49). Coverslips were washed extensively and mounted with Mowiol. SR-BI and LBPA were visualized using rabbit anti-SR-BI, mouse anti-LBPA, Cy5- and Cy3-donkey anti-rabbit or anti-mouse as secondary antibody, respectively (Jackson ImmunoResearch). Images were taken using a Leica DMI3000B inverted fluorescent microscope.

For DiI-HDL uptake assays, HDL₃ was labeled with DiI according to the manufacturer's instructions (Molecular Probes). CHO-SRBI and HuH7 cells were preincubated ± bafilomycin, lactacystin, and PD98059 as above and washed, and 10 μg/ml DiI-HDL was added for 5–15 min at 37 °C, respectively.

Preparation of Cellular Extracts and Western Blot Analysis

Cells were harvested in lysis buffer (20 mm Tris-HCl, pH 7.5, 2 mm EDTA, 100 mm NaCl, 5 mm MgCl₂, 1% (v/v) Triton X-100, 5 mm NaF, 10% (v/v) glycerol, 0.5% (v/v) 2-mercaptoethanol, 0.1 mm Na₃VO₄, and protease inhibitors). After centrifugation at 10,000 × g, the protein concentration of the cleared cell lysate was determined. Cell lysates were separated by 10–12.5% SDS-PAGE and transferred to Immobilon-P (Millipore). Proteins were detected using their specific primary antibodies, followed by HRP-conjugated secondary antibodies and enhanced chemiluminescence detection (ECL, Amersham Biosciences).

Subcellular Fractionations

For the isolation of plasma membrane-enriched fractions, lysates from 1 × 10⁷ cells were separated on Percoll gradients as described previously (7). Cells were washed twice in 0.25 m sucrose, 1 mm EDTA, 20 mm Tris-HCl, pH 7.8, plus protease inhibitors, collected, and centrifuged. The postnuclear supernatant was layered on top of 10 ml of 30% Percoll and centrifuged at 84,000 × g for 30 min in a Beckman 70.1 TI rotor. The plasma membrane fraction in the middle of the gradient was isolated (1 ml), concentrated, and analyzed for the amount of SR-BI and Ras.

For the isolation of nuclear fractions (50), cells were harvested in 10 mm Tris-HCl, pH 7.5, 2 mm EDTA and incubated on ice for 10 min. An equal volume of 0.5 m sucrose, 0.1 m KCl, 10 mm MgCl₂, 2 mm CaCl₂, 2 mm EDTA in 10 mm Tris-HCl, pH 7.5, was added. Lysates were passed 10 times through a 25-gauge needle and centrifuged (700 × g, 10 min, 4 °C). Pellet and supernatant were collected as nuclear and cytoplasmic fractions, separated by SDS-PAGE, and transferred to Immobilon-P.

For the isolation of late endosomes, 4–6 × 10⁷ CHO-SRBI cells were incubated overnight ± bafilomycin and PD98059 and homogenized by 10 passages through a 22-gauge needle. Endosomes were separated on sucrose gradients as described previously (44). In brief, the cell homogenate was centrifuged, and the postnuclear supernatant was brought to a final 42% sucrose (w/v) concentration. Then 35% sucrose, 25% sucrose, and homogenization buffer were poured stepwise on top of the postnuclear supernatant. The gradient was centrifuged for 90 min at 35,000 rpm, 4 °C in a Beckman SW40 rotor. After centrifugation, fractions representing late endosomes at the interphase of 25% sucrose and homogenization buffer were collected and analyzed by Western blotting.

Cell Growth

Cell proliferation in CHOwt and CHO-SRBI cells was determined using colorimetric proliferation assays (Promega; G5421 and G4000) according to the manufacturer's instructions. Cells (0.5 × 10⁵, in triplicate) were plated in 96-well plates ± 50 μg/ml HDL, 10 μm PD98059, or both and incubated with tetrazolium compound (MTS in G5421 or MTT in G4000) for 0–6 h. MTS and MTT are reduced by cells into formazan products, which was determined by absorbance at 490 or 570 nm without any further processing, respectively.

RESULTS

Inhibition of MAPK Reduces SR-BI Expression

Recent studies implicate cell type-specific and differential effects of MAPK signaling on ABCA1 transporter expression and activity (26–28). To address the potential involvement of the Ras/MAPK pathway in SR-BI expression, we first investigated if inhibition of Erk1/2 signaling would alter the expression of endogenous SR-BI in the presence or absence of the PPARα agonist Wy-14643. Therefore, we initially analyzed CHO wild-type (CHOwt) cells, which have been proven valuable for Ras/MAPK signaling studies (6–8, 32, 51) and express low but detectable levels of SR-BI and PPARα as judged by Western blot analysis using antibodies against human SR-BI and PPARα, respectively (Figs. 1, A and B, and 2A). Similar to results obtained from other cell types (29), Wy-14643 induced SR-BI expression 4.1 ± 1.7-fold (Fig. 1A, compare lane 1 and 3) (**, p < 0.01; see quantification in Fig. 1A). Upon addition of the commonly used Mek1/2 inhibitor PD98059 (52), which reduced basal P-Erk1/2 levels by 46 ± 5% in this set of experiments, the expression of SR-BI ± Wy-14643 decreased by 57.5 ± 9% (*, p < 0.05) and 73.1 ± 19.5% (*, p < 0.05), respectively) (Fig. 1A, compare lanes 1 and 2 and 3 and 4). We next determined the effect of PD98059 on SR-BI protein levels in CHO cells incubated with the PPARα antagonist MK886 (Fig. 1B). Incubation with MK886 reduced SR-BI levels by ∼75% (Fig. 1B, compare lane 1 and 3), indicating that MK886 inhibits the involvement of PPARα to establish basal SR-BI levels. To enhance detection of the low endogenous SR-BI levels in MK886-treated cells (Fig. 1B, lanes 3 and 4), overnight exposures of Western blots were analyzed. Similar to Fig. 1A, SR-BI expression was reduced by ∼30% with PD98059 (Fig. 1B, lanes 1 and 2). However, residual SR-BI expression in MK886-treated cells was not reduced any further by PD98059, supporting that Erk1/2 inhibition decreases SR-BI levels via modulating PPARα activity (Fig. 1A, compare lanes 3 and 4).

FIGURE 1.

Mek1/2 inhibition reduces SR-BI expression. A and B, CHOwt cells were incubated for 24 h ± 20 μm Wy-14643 (A) and ± 10 μm MK886 (B) with or without 10 μm PD98059 as indicated. Cell lysates were prepared and analyzed by Western blotting for expression levels of SR-BI, activated (P-Erk1/2) and total Erk1/2, and β-actin. Expression of SR-BI in each lysate was quantified and normalized to the amount of β-actin. The mean values ± S.D. of four independent experiments are given in A. * and **, p < 0.05 and p < 0.01 for Student's t test, respectively. C, HEK293 cells were incubated for 24 h ± 20 μm fenofibrate (FF), 20 μm Wy-14643 (Wy), and 10 μm PD98059 as indicated. Western blot analysis of SR-BI, activated Erk1/2 (P-Erk1/2) and total Erk1/2, and β-actin was quantified as described above. A representative Western blot is shown in supplemental Fig. 1A. D, HEK293 cells were transfected with control RNAi (Scramble) and RNAi targeting PPARα (see “Experimental Procedures”). 72 h after transfection, cells were treated for an additional 24 h ± 20 μm Wy-14643 and 10 μm PD98059 as indicated. Expression of SR-BI, PPARα, activated (P-Erk1/2) and total Erk1/2, and β-actin was analyzed and quantified as in A–C. PPARα expression levels were reduced 45 ± 15% compared with controls (n = 2). The mean values ± S.D. from two independent experiments are given in C and D.

FIGURE 2.

Enhanced Ras/MAPK activity increases SR-BI expression. A, CHOwt cells were transfected with empty vector (control), constitutively active H-Ras (HRasG12V), or PPARα. 24 h after transfection, cells were treated ± 20 μm Wy-14643 as indicated. Cell lysates were analyzed and quantified for expression levels of SR-BI, PPARα, Ras, β-actin, and caveolin (cav-1) as described in Fig. 1. The mean values ± S.D. from four independent experiments are given. **, p < 0.01 for Student's t test. B and C, HEK293 cells were transfected with empty vector (control) or constitutively active Mek1 (Mek215-DD) and treated (B) ± 20 μm Wy-14643 or (C) ± 10 μm MK886 as indicated. Cell lysates were analyzed for expression of SR-BI, total Mek1, activated Erk1/2 (P-Erk1/2), total Erk1/2, and β-actin. SR-BI expression levels were quantified (B).

To verify our findings in human cells, we examined the involvement of the Ras/MAPK pathway on the regulation of SR-BI expression in the human embryonic kidney cell line HEK293, a common model to study cholesterol transport and nuclear receptor activity (1, 14–18). Cells were incubated ± PPARα agonist Fenofibrate (FF) or Wy-14643 in the presence or absence of the Mek1/2 inhibitor PD98059 (Fig. 1C, for representative Western blot see supplemental Fig. 1A). Consistent with data from CHOwt cells, FF and Wy-14643 increased SR-BI levels 1.7–2.0-fold (Fig. 1C), whereas Mek1/2 inhibition reduced SR-BI levels in FF-treated HEK293 cells by 45 ± 7% and almost completely repressed the stimulatory effect of Wy-14643 on SR-BI expression (Fig. 1C). To further substantiate these findings, HEK293 cells were transfected with PPARα siRNA to knock down the expression of endogenous protein (Fig. 1D). Similar to Fig. 1C, in HEK293 cells transfected with control siRNA, PD98059 reduced SR-BI protein levels by ∼38 ± 10% (compare lane 1 and 2). RNAi-mediated down-regulation of PPARα by 66 ± 9% correlated with 63 ± 6% reduced SR-BI levels (Fig. 1D, compare lane 1 and 3). Most strikingly, PPARα-depleted HEK293 cells did not exhibit further down-regulation of SR-BI upon Mek1/2 inhibition (54 ± 8% compared with control), supporting a model of MAPK signaling modulating SR-BI expression via regulating PPARα activity (Fig. 1D, compare lane 3 and 4).

Enhanced Ras/MAPK Signaling Up-regulates SR-BI Expression

To address if enhanced Ras/MAPK activity increases SR-BI expression, possibly in a PPARα-dependent manner, CHOwt cells transfected with HRasG12V were monitored for SR-BI levels ± the PPARα agonist Wy-14643 (Fig. 2A). Overexpression of PPARα served as a positive control and increased SR-BI levels 3.1- and 4.1-fold ± Wy-14643, respectively, compared with the nontransfected control (Fig. 2A, compare lanes 1, 4, and 5). Similarly, ectopic expression of constitutively active H-RasG12V increased SR-BI levels in untreated (2.3 ± 0.2-fold; **, p < 0.01) and Wy-14643-stimulated cells (3.5 ± 0.5-fold; **, p < 0.01; compare lanes 1–3). To date, only H-Ras has been linked to HDL-induced and SR-BI-mediated activation of the MAPK pathway (6–8). To determine whether other Ras isoforms could also impact on SR-BI expression, we overexpressed KRasG12V in HEK293 cells treated ± FF or Wy-14643 and determined SR-BI levels (supplemental Fig. 1B). As expected, KRasG12V increased Erk1/2 activity, which correlated with strongly elevated SR-BI levels in untreated (11.4 ± 2.8) and FF- (11.2 ± 1.1)- and Wy-14643 (9.8 ± 1.2)-treated cells compared with controls. Thus, consistent with Erk1/2 kinases potentiating the activity of PPARα-dependent pathways, enhanced K-Ras activity increases SR-BI levels in the presence and absence of PPARα agonists. To further verify that elevated MAPK activity facilitates SR-BI up-regulation, HEK293 cells were transfected with constitutively active Mek1 (Mek215-DD), and SR-BI levels were monitored ± Wy-14643 (Fig. 2B). SR-BI levels were increased 1.8 ± 0.3 and 3.2 ± 0.8 with and without Wy-14643 in Mek215-DD-transfected cells, respectively, further pointing at Mek/Erk kinases stimulating SR-BI expression. Then we compared SR-BI levels in HEK293 cells transfected ± Mek215-DD in the presence or absence of PPARα antagonist MK886 (Fig. 2C). Similar to Fig. 1B, MK886 reduced SR-BI levels in HEK293 cells (compare lanes 1 and 2) and overexpression of active Mek1 increased SR-BI levels (compare lanes 1 and 3), respectively. However, ectopic expression of active Mek1 did not overcome the inhibitory effect of MK886 on SR-BI expression (Fig. 1B, lane 4), hence providing additional evidence that MAPK signaling regulates SR-BI expression via modulating PPARα activity.

Ras/MAPK Signaling Increases PPARα Phosphorylation

Erk1/2 could increase PPARα activity through transcriptional up-regulation (26), by promoting its nuclear translocation as shown for PPARγ (53), or through the phosphorylation of serine residues at positions 12 and 21 within the N-terminal region of PPARα (17, 19). To address these different modes of action, we first examined expression and cellular distribution of PPARα upon overexpression of constitutively active H-Ras (HRasG12V). Results from Fig. 1D (lanes 1 and 2) already revealed that Mek1/2 inhibition did not down-regulate endogenous PPARα expression in HEK293 cells. Conversely, as shown in Fig. 2A (compare lanes 1–3) and supplemental Fig. 2A (lanes 1 and 2), HRasG12V overexpression did not increase endogenous PPARα levels. Also, HRasG12V did not alter levels of ectopically expressed PPARα (supplemental Fig. 2A). In addition, HRasG12V did not change the predominant nuclear localization of PPARα as indicated by the constant nuclear/cytoplasmic ratio of ectopically expressed PPARα ± HRasG12V (supplemental Fig. 2B). However, phosphorylation of Ser-21 in the N-terminal region of PPARα was increased 4.0–4.7-fold upon overexpression of HRasG12V, KRasG12V, or Mek215-DD (**, p < 0.01; Fig. 3A) in CHO cells, respectively. Similarly, ectopically expressed active Mek1 increased Ser-21 phosphorylation of PPARα ∼4.0–10-fold in HEK293 cells treated with or without Wy-14643 (Fig. 3B). Hence, phosphorylation events targeting the N-terminal Ser-21 residue, and possibly Ser-12 of PPARα, might alter PPARα activity, which could contribute to the stimulatory effect of Ras/MAPK signaling on SR-BI expression.

FIGURE 3.

Ras/MAPK signaling induces PPARα phosphorylation. A and B, CHOwt and HEK293 cells were transfected with empty vector (control), constitutively active H- or K-Ras (HRasG12V and GFP-KRasG12V) or Mek1 (Mek215-DD) ± 20 μm Wy-14643 (Wy) as indicated. After 24 h, cell lysates were prepared and analyzed by Western blotting for the amount of phosphorylated PPARα (P-Ser-21 PPARα), total PPARα, activated Erk1/2 (P-Erk1/2), GFP-KRasG12V, H-Ras, Mek1/2, and β-actin. The mean values ± S.D. of phosphorylated PPARα normalized to the amount of total PPARα from three independent experiments are given. **, p < 0.01 for Student's t test.

Inhibition of Mek1/2 Reduces SR-BI Activity

We next reasoned that reduced SR-BI expression upon Mek1/2 inhibition should result in reduced SR-BI activity. Indeed, cell surface expression of SR-BI, as judged by Western blot analysis of Ras-containing plasma membrane fractions isolated from Percoll gradients, was robustly decreased in CHOwt cells treated with PD98059 (Fig. 4A). Together with the data presented above (Figs. 1–3), we hypothesized that Ras/MAPK inhibition could reduce SR-BI activity, as judged by HDL-inducible cholesterol efflux. As shown previously (6–8), HDL₃ is a potent activator of the MAPK pathway in CHOwt cells. PD98059 strongly reduced HDL-induced Mek1/2 and Erk1/2 activity (Fig. 4B, compare lanes 2 and 3). Other effectors downstream of Ras, such as PI3K/Akt, which could possibly control cholesterol transporter activity and HDL-inducible efflux (1–5), were not significantly affected by PD98059 (Fig. 4B).

FIGURE 4.

MAPK inhibition reduces HDL-inducible cholesterol efflux. A, CHOwt cells were incubated ± 10 μm PD98059 for 24 h and subjected to subcellular fractionation through Percoll gradients. Ras-containing plasma membrane fractions in the middle of the gradient were isolated and analyzed for the amount of SR-BI. B, CHOwt were starved, preincubated for 60 min ± 10 μm PD98059 (PD), and stimulated ± HDL₃ (50 μg/ml) for 3 min at 37 °C as indicated. Western blot analysis of activated Mek1/2 (P-Mek1/2), Erk1/2 (P-Erk1/2), Akt (P-Akt), total Mek1/2, Erk1/2, Akt, and β-actin from each lysate is shown. C, CHOwt were incubated with [³H]cholesterol (2 × 10⁶ cpm/ml) for 24 h, washed with PBS, and incubated with HDL₃ (50 μg/ml) for 4 h ± 10 μm PD98059. Alternatively, CHOwt cells were transfected ± empty control vector and dominant-negative Erk1 (DN-Erk1) and then incubated with [³H]cholesterol and HDL₃ as described above. The ratio of released and cell-associated radioactivity was determined and normalized to total cell protein, and the amount of efflux is given in %. The background efflux obtained from CHOwt was equivalent to 2.0–4.0 × 10⁵ cpm/mg cell protein, respectively. D, CHOwt cells were incubated ± HDL₃ (50 μg/ml) and 10 μm PD98059 (PD) (in triplicate) as described above. Cell proliferation was monitored for 6 h at 37 °C using an MTT assay. Mean values ± S.D. of a representative experiment (n = 3 in A; n = 2 in B) with triplicate samples are given. *, p < 0.05 for Student's t test.

Next CHOwt cells were labeled for 24 h with [³H]cholesterol, preincubated for 3–4 h ± PD98059, followed by an incubation with 50 μg/ml HDL₃ (Fig. 4C). Cells and media were assayed for radioactivity after 4 h, and efflux was determined as the percentage of total cholesterol in the culture. Similar to previous experiments (44), HDL stimulated cholesterol efflux 2.0–2.5-fold in CHOwt compared with controls (Fig. 4C). However, incubation of HDL in the presence of PD98059 reduced efflux by ∼15–20% (*, p < 0.05). Treatment with PD98059 alone did not affect basal cholesterol efflux levels. Comparable results were obtained when determining cholesterol efflux in CHOwt in the presence of U0126, another Mek1/2-specific inhibitor (data not shown) (54). To further validate that inhibition of Erk1/2 signaling reduces SR-BI activity, we transiently transfected CHOwt cells with dominant-negative Erk1 (DN-Erk1), which is known to inhibit Erk1/2 signaling (55). Similar to the results obtained with Mek1/2 inhibitors, overexpression of DN-Erk1 inhibited HDL-inducible cholesterol efflux by 43.7 ± 4.3% (Fig. 4C). Reduced cholesterol efflux onto HDL upon addition of MAPK inhibitors was not due to unequal internalization of [³H]cholesterol, which was comparable in cells incubated with and without HDL and treated ± PD98059 (4.1 ± 0.3 × 10⁸ cpm/mg cell protein) or transfected ± DN-Erk1 (4.4 ± 0.6 × 10⁸ cpm/mg cell protein), respectively. Furthermore, colorimetric (MTT) proliferation assays verified that 10 μm PD98059 (and U0126; data not shown) did not significantly reduce cell growth in CHOwt within the time frame of the experimental setting (Fig. 4D).

CHO cells stably expressing SR-BI (CHO-SRBI) are very suitable for HDL-dependent cholesterol transport studies and are characterized by increased HDL surface binding (32). We therefore examined if MAPK signaling affects efflux pathways in CHO-SRBI. Consistent with our previous findings, HDL-incubated CHO-SRBI displayed a strong activation of H-Ras (4.0 ± 1.0) and Mek1/2 (Fig. 5, A and B). PD98059 efficiently reduced Mek1/2 phosphorylation and, as shown for CHOwt (Fig. 4B), HDL-induced activation of other Ras effectors such as Akt was not significantly affected by PD98059 in CHO-SRBI cells (Fig. 5B). Importantly, PD98059 reduced HDL-induced cholesterol efflux in CHO-SRBI by 28–32% (*, p < 0.05) (Fig. 5C). As shown for CHOwt, cell growth of CHO-SRBI ± 10 μm PD98059 was not altered within the time frame of this efflux experiment, as determined by proliferation assays (data not shown).

FIGURE 5.

MAPK inhibition reduces HDL-induced cholesterol efflux in SR-BI-overexpressing cells. A, CHO-SRBI cells were starved overnight and incubated ± 50 μg/ml HDL₃ for 3 min as indicated. Lysates were subjected to RBD pulldowns to determine activated H-Ras (H-Ras GTP) levels. Total H-Ras in each lysate is shown. H-Ras GTP levels were quantified and normalized to the amount of total H-Ras in each lysate. B, CHO-SRBI cells were starved, preincubated for 60 min ± 10 μm PD98059 (PD), and stimulated ± HDL₃ (50 μg/ml) for 3 min at 37 °C as indicated. Lysates were prepared and analyzed for the amounts of activated Mek1/2 and Akt (P-Mek1/2 and P-Akt), total Mek1/2 and Akt, SR-BI, and caveolin (cav-1). C, CHO-SRBI cells were incubated with [³H]cholesterol (2 × 10⁶ cpm/ml) for 24 h. Efflux (%) in cells treated for 4 h ± HDL₃ (50 mg/ml) and ± 10 μm PD98059 (PD) was determined as described in Fig. 4C. The background efflux obtained from CHO-SR-BI was equivalent to 6.0–8.0 × 10⁵ cpm/mg cell protein, respectively. Mean values ± S.D. of three (A) and two (B) independent experiments with triplicate samples are given. *, p < 0.05 for Student's t test. D, CHO-SRBI cells were transfected with PPARα, preincubated ± 10 μm PD98059 (PD), and stimulated for 3 min ± HDL₃ (50 μg/ml) as described above. Lysates were prepared and analyzed for phosphorylated PPARα (P-Ser-21 PPARα) and total PPARα. The mean values ± S.D. of phosphorylated PPARα normalized to the amount of total PPARα are given (n = 2).

To link HDL-inducible efflux with MAPK signaling and PPARα phosphorylation, we examined Ser-21 phosphorylation of ectopically expressed PPARα in CHO-SRBI incubated with HDL ± PD98059 (Fig. 5D). PPARα Ser-21 phosphorylation was increased 1.9 ± 0.5-fold with HDL, whereas the addition of PD98059 reduced PPARα Ser-21 phosphorylation in HDL-incubated cells by 52.1 ± 19.2% (Fig. 5D, compare lanes 2 and 3). Similarly, Ser-21 phosphorylation of endogenous PPARα in HDL-incubated CHOwt was reduced by ∼50–60% upon addition of PD98059 (data not shown). Taken together, it is tempting to speculate that MAPK inhibition reduced cholesterol efflux in CHOwt and CHO-SRBI cells due to inhibition of HDL-mediated PPARα Ser-21 phosphorylation, followed by reduced SR-BI expression and activity.

Enhanced Ras/MAPK Signaling Stimulates Cholesterol Efflux

Next, we analyzed if overexpression of constitutively active Ras could promote cholesterol efflux (supplemental Fig. 3A). As expected, overexpression of constitutively active H-Ras (and K-Ras; data not shown) resulted in elevated HRasG12V-GTP levels and increased Mek1/2 and Erk1/2 phosphorylation (supplemental Fig. 3A, compare lane 5 and 6) without affecting endogenous H-Ras activity (supplemental Fig. 3A, compare lanes 1–4). Consistent with MAPK inhibition inhibiting cholesterol export, overexpression of HRasG12V and KRasG12V correlated with a statistically significant 23–28% stimulation (*, p < 0.05) of cholesterol efflux in HDL-incubated cells, respectively (supplemental Fig. 3B). Finally we examined if Mek1/2 inhibition would interfere with efflux in cells overexpressing HRasG12V or Mek215-DD. Similar to results presented above, HRasG12V stimulated HDL-inducible cholesterol efflux 3–4-fold compared with the negative control (supplemental Fig. 3C). PD98059 reduced HDL-inducible efflux in HRasG12V-transfected cells by 25–30%, supporting a model of Ras stimulating efflux via Mek1/2 and Erk1/2 signaling. Interestingly, 10 μm PD98059 was insufficient to inhibit efflux in cells with highly elevated levels of the active Mek1 mutant.

Mek1/2 Inhibition Reduces SR-BI Protein Stability

Endogenous SR-BI expression in CHOwt is driven by its homologous promoter, whereas ectopically expressed SR-BI in CHO-SRBI (33) is under the control of the heterologous CMV promoter. Hence, it appears unlikely that Ras/MAPK inhibition reduces SR-BI mRNA, possibly in a PPARα-dependent manner, in both cell lines. To identify if Ras/MAPK signaling modifies SR-BI mRNA or protein levels (29, 36, 37), we first measured SR-BI mRNA by real time PCR (Fig. 6A). However, treatment of HEK293 ± Wy-14643 and U0126 did not significantly alter SR-BI mRNA levels.

FIGURE 6.

Mek1/2 inhibition reduces SR-BI protein stability. A, 1 μg of RNA extracted from HEK293 cells treated ± PPARα agonists (20 μm Wy-14643) and Mek1/2 inhibitor (10 μm U0126) was reverse-transcribed, and real time RT-PCR to amplify SR-BI and TATA box-binding protein (TBP) cDNA fragments was performed as described previously (42). Relative expression from two independent experiments with duplicate samples is given and was calculated by normalization to the housekeeper mRNA (TBP) (43). B and C, CHOwt (B) and CHO-SRBI (C) were preincubated with 20 μm Wy-14643 overnight, followed by the addition of 20 ng/ml cycloheximide (CHX) ± 10 μm PD98059 (PD) for 0–8 h as indicated. Western blot analysis of SR-BI and β-actin in each lysate of a representative experiment is shown. The mean values ± S.D. of SR-BI expression levels (n = 2) are given.

Therefore, we next examined SR-BI protein stability. CHOwt cells were incubated with 20 ng/ml cycloheximide to inhibit protein synthesis for 0–8 h ± Mek1/2 inhibitor PD98059 (Fig. 6B). SR-BI levels were significantly down-regulated by 44.8 ± 2.7% after 4 h in CHOwt controls (Fig. 6B, compare lanes 1–4). In contrast, addition of PD98059 accelerated SR-BI protein degradation, and after 2 h, reduction of SR-BI protein levels by 57.4 ± 2.5% was evident (Fig. 6B, lanes 5–8). Thus inhibition of MAPK signaling appears to reduce protein stability of SR-BI.

We next reasoned that ectopically expressed SR-BI in the CHO-SRBI cell line (33) should also be sensitive toward Mek1/2 inhibitors. Indeed, treatment of CHO-SRBI with cycloheximide (Fig. 6C) reduced SR-BI protein by 41.6 ± 2.1% in controls after 8 h (compare lanes 1–4). Yet Mek1/2 inhibitors strongly increased sensitivity toward SR-BI degradation, which was already detectable after 2 h (65.1 ± 1.4% reduction; Fig. 6C, compare lanes 5–8).

Mek1/2 Inhibition Enhances PPARα-inducible SR-BI Degradation in Hepatocytes

SR-BI protein stability was then analyzed in other cell types. Like CHO cells, HEK293 exhibited enhanced SR-BI degradation in the presence of cycloheximide and PD98059 (data not shown). In contrast, Mek1/2 inhibitors did not significantly alter SR-BI protein stability in endothelial (bovine aortic endothelial cells) and monocytic (THP1) cell lines (Fig. 7A). These findings suggest that Erk inhibition reduces SR-BI protein stability in a cell-specific manner that does not include peripheral cells. Because fibrates induce SR-BI degradation in liver (36, 37), we next analyzed SR-BI protein stability in human HuH7 hepatocytes. As shown for primary hepatocytes from fibrate-fed mice (36, 37), Wy-14643 strongly down-regulated SR-BI protein levels in HuH7 cells (Fig. 7B, compare lanes 1 and 3). Furthermore, PD98059 reduced SR-BI protein expression in the absence and presence of PPARα agonists (Fig. 7B, compare lanes 1 and 2 and 3 and 4). Similar to the mRNA expression data obtained from HEK293 cells (Fig. 6A), SR-BI mRNA levels in mouse hepatocytes remained unchanged upon Mek1/2 inhibition (data not shown). Enhanced SR-BI degradation in the presence of PD98059 and cycloheximide further suggest that Mek1/2 inhibition down-regulates SR-BI protein levels in HuH7 hepatocytes (Fig. 7C, compare lanes 4 and 8).

FIGURE 7.

Mek1/2 inhibition reduces SR-BI protein stability in hepatocytes. Bovine aortic endothelial cells (BAEC) and THP1 macrophages (A) and HuH7 hepatocytes (B and C) were preincubated ± 20 μm Wy-14643 overnight, followed by the addition of 20 ng/ml cycloheximide (Cyclohex) ± 10 μm PD98059 (PD) for 0–8 h as indicated. Western blot analysis of SR-BI and β-actin in each lysate of a representative experiment is shown. Overnight exposures (ON exposure) for SR-BI are shown in B and C. D, 2–5 × 10⁵ CHO-SRBI cells were incubated with 10 μg/ml [³H]cholesteryl ester-labeled HDL₃ for 2 h ± 20 μm Wy-14643 and 10 μm PD98059 as indicated. Cells were washed and lysed, and cell-associated radioactivity was quantified and normalized to total cell protein (see “Experimental Procedures”). C and D, mean values ± S.D. from three independent experiments (with triplicates in D). *, **; p < 0.05, p < 0.01 for Student's t test. E, CHO-SRBI cells were grown on coverslips and preincubated overnight ± 10 μm PD98059. 10 μg/ml DiI-HDL was added for 5 min at 37 °C. Cells were washed and fixed, and the intensity and number of DiI-stained vesicles per cell (150 cells per treatment) in images captured using identical contrast and exposure times was quantified using Image J as described previously (44). Bar, 10 μm. **, p < 0.01 for Student's t test.

Importantly, these findings correlate with ∼40–50% down-regulation of [³H]cholesteryl ester uptake from HDL₃ (*, p < 0,05) with PD98059 in CHO-SRBI cells ± WY-14643 (Fig. 7D). Similarly, uptake of DiI-labeled HDL₃ was significantly reduced (**, p < 0.01) in CHO-SRBI incubated with PD98059, as judged by the reduced number (44.4 ± 9.1%) and fluorescence intensity (41.4 ± 18.1%) of DiI-stained vesicles/cell, respectively (Fig. 7E). Similarly, DiI-HDL uptake in PD98059-treated HuH7 hepatocytes was reduced by 59.5 ± 8.1% (DiI-containing vesicles/cells) and 39.2 ± 11.0% (staining intensity), respectively (data not shown). Taken together, Mek1/2 inhibition strongly reduces SR-BI expression and activity in fibroblasts and hepatocytes.

To gain mechanistic insights into SR-BI down-regulation, we incubated HuH7 cells with inhibitors of proteasomal (25 ng/ml lactacystin) and lysosomal (50 ng/ml bafilomycin) degradation, respectively (Fig. 8, A and B). Lactacystin increased SR-BI protein levels ± PPARα agonists, indicating that substantial amounts of SR-BI are degraded via the proteasome (Fig. 8A, compare lanes 1 and 3 and 5 and 7). Interestingly, in the presence of lactacystin, PD98059 did not reduce SR-BI levels (Fig. 8A, compare lanes 3 and 4 and 7 and 8). Bafilomycin also increased SR-BI levels in HuH7 (Fig. 8B, compare lanes 1 and 3 and 5 and 7), suggesting that SR-BI turnover can occur as well via lysosomal degradation pathways. Most strikingly, Mek1/2 inhibition decreased SR-BI levels in bafilomycin-treated cells (Fig. 8B, compare lanes 3 and 4 and 7 and 8), pointing at Mek1/2 regulating proteasomal SR-BI degradation. In addition, incubation of HuH7 cells with bafilomycin as well as Mek1/2 inhibitor did not significantly alter SR-BI adaptor PDZK1 levels in fibrate-stimulated HuH7 cells (Fig. 8C). Hence, the previously identified PPARα-inducible but PDZK1-independent SR-BI degradation in hepatocytes (36, 37) appears to be stimulated by the inhibition of Mek1/2 kinases.

FIGURE 8.

Mek1/2 inhibition promotes proteasomal degradation of SR-BI. A–C, HuH7 hepatocytes were preincubated ± 20 μm Wy-14643 (WY) overnight, followed by the addition of 25 ng/ml lactacystin (A) or 50 ng/ml bafilomycin (B) ± 10 μm PD98059 (PD) for 12 h as indicated. Western blot analysis of SR-BI (A and B), PDZK1 (C) and β-actin in each lysate is shown. Overnight exposures were required to visualize SR-BI protein expression in fibrate-induced HuH7 cells (A and B). The mean values ± S.D. of SR-BI expression levels from three independent experiments are shown in A and B. * and **, p < 0.05 and p < 0.01 for Student's t test, respectively.

Results described above (Fig. 8B) implicated the accumulation of SR-BI in late endosomes of bafilomycin-treated HuH7 cells. Therefore, we compared SR-BI localization in the CHO-SRBI ± bafilomycin and Mek1/2 inhibitor (Fig. 9A). As expected, large amounts of SR-BI are localized at the plasma membrane in control cells (see arrows in Fig. 9A, panel a). Consistent with SR-BI down-regulation in HuH7 cells by Western blot analysis (Fig. 8B), PD98059 strongly reduced SR-BI staining intensity and led to a punctate staining pattern throughout the cytoplasm (Fig. 9A, panel b). As hypothesized, bafilomycin induced an accumulation of SR-BI in perinuclear, possibly late endosomal vesicles (Fig. 9A, panel c). In agreement with the expression analysis (Fig. 8B), PD98059 reduced bafilomycin-inducible SR-BI accumulation (Fig. 9A, panel d) in perinuclear vesicles. To verify accumulation of SR-BI in late endosomes in the presence of bafilomycin, we performed co-localization studies of SR-BI with the late endosomal marker LBPA (44). Indeed, significant co-localization of SR-BI and LBPA in bafilomycin-treated cells was observed (Fig. 9B, see panel h and arrows in panel i) supporting a model of SR-BI being targeted to (pre)-lysosomes/late endosomes upon inhibition of lysosomal degradation. As control, lactacystin did not induce perinuclear accumulation of SR-BI (data not shown). In contrast, SR-BI staining is found in cellular sites not resembling late endosomes upon Mek1/2 inhibition (Fig. 9B, panels d–f). In addition, PD98059 treatment of cells incubated with bafilomycin reduced co-localization of SR-BI and LBPA (Fig. 9B, panels j–l). Western blot analysis of membrane fractions enriched with late endosomes from CHO-SRBI cells support these findings (Fig. 9, panel c). SR-BI was not detectable in late endosomal fractions from untreated CHO-SRBI cells ± PD98059 (data not shown) but accumulated in Rab7-positive late endosomes with bafilomycin. Incubation of bafilomycin together with PD98059 reduced late endosomal SR-BI levels, further indicating that Mek1/2 inhibition promotes degradation pathways that do not involve targeting of SR-BI to lysosomes.

FIGURE 9.

Mek1/2 inhibition reduces SR-BI cell surface expression. A and B, CHO-SRBI cells were grown on coverslips and incubated overnight ± 10 μm PD98059 and 20 ng/ml bafilomycin as indicated. Cells were fixed and immunolabeled with anti-SR-BI (A) or anti-SR-BI (red) and anti-LBPA (green) (B) as indicated. A, to compare the intensity of SR-BI staining in cells treated ± PD98059 (compare panel a with b and c with d), experiments were performed in parallel, and images were captured using identical contrast and exposure times. Arrows point at SR-BI at the cell surface (panel a) and perinuclear vesicles (panel c), respectively. B, merged images are shown in panels c, f, i, and l. Arrows point at SR-BI and LBPA co-localization in bafilomycin-treated cells (panel i). Bar, 10 μm. C, late endosomes from 4–6 × 10⁷ CHO-SRBI cells treated overnight with 20 ng/ml bafilomycin ± 10 μm PD98059 were isolated as described previously (44) and analyzed for SR-BI by Western blotting. The purity of isolated fractions was assessed by immunoblotting with a marker for late endosomes (Rab 7).

DISCUSSION

In this study we demonstrate that members of the Ras/MAPK signaling cascade, including H-Ras, K-Ras, Mek1/2, and Erk1/2, regulate the protein levels of SR-BI via PPARα-inducible degradation pathways in hepatocytes and fibroblasts. Modulating Ras/MAPK signaling correlates with altered SR-BI activity, as judged by cholesterol export onto HDL in fibroblasts and cholesteryl ester uptake in HuH7 hepatocytes. Thus, Ras isoforms and Mek1/2 kinases upstream of Erk1/2 are novel modulators of reverse cholesterol transport via regulating SR-BI protein stability and cell surface expression.

Earlier studies aiming to link Mek and Erk kinases with cholesterol transport showed that cholesterol loading and apoA-I-mediated cholesterol efflux did not alter MAPK activity (41), However, NIH/3T3 cells transformed with constitutively active HRas (HRasG12V) exported cholesterol much more rapidly to HDL compared with controls (56). This was mainly attributed to Ras-mediated down-regulation of caveolin-1, a cholesterol-binding protein that inhibits HDL-induced cholesterol efflux in some cell types, including NIH3T3 (56, 57). However, transient HRasG12V overexpression or Mek1/2 inhibition did not significantly alter caveolin expression in our cell model systems (see Figs. 2A and 5B). Results presented in this study suggest that Ras/MAPK signaling modulates SR-BI expression and activity via PPARα, a major regulator of cholesterol homeostasis (16, 58). Yet the role of PPARα in SR-BI expression is complex and can involve transcriptional and post-transcriptional mechanisms, depending on the cell type analyzed (2, 29, 36, 37). PPARα agonists enhance SR-BI mRNA expression in macrophages but down-regulate SR-BI protein in hepatocytes. Fibrates have therefore been proposed to possibly regulate SR-BI protein synthesis, trafficking, degradation, and interaction with SR-BI adaptors, such as PDZK1 (36, 37). Results presented here suggest that Mek1/2 inhibitors enhance PPARα-inducible proteasomal degradation pathways to down-regulate SR-BI in fibroblasts and hepatocytes, which is most relevant for the involvement of SR-BI in reverse cholesterol transport.

Activation of PPARs has been mainly considered a ligand-dependent mechanism, but more recently, ligand-independent activation of PPARs has also been reported (17–19, 26). In vitro assays as well as cellular studies implicate a strong cross-talk between kinase cascades and PPARs (17, 19). Both PPARα and PPARγ are phosphoproteins and MAPK, in particular Erk2, can modulate PPAR activity (59). Several consensus sites for PPARα phosphorylation have been identified, and treatment of cells with PD98059 blocks PPARα activity (60, 61). PPARα is phosphorylated exclusively on serine residues, and Erk1/2-mediated transactivation of PPARα is specific for serine residues at positions 12 and 21 within the N-terminal transactivation domain of PPARα (17–19, 26, 62). Depending on the cell type and stimuli, Erk1/2-mediated phosphorylation of PPARα can either lead to activation or inactivation of PPARα possibly involving recruitment or dissociation of PPARα co-repressors (17–19, 62–65). Erk1/2 up-regulates PPARα expression in lung epithelial cells, thereby increasing PPARα activity (26). This up-regulates ABCA1 during the inflammatory response triggered by bacterial infection, as Mek1 overexpression increased PPARα expression coordinately with ABCA1 levels (26).

Alternatively, oxidized LDL can induce Erk1/2 to up-regulate COX2, which in turns activates PPARα and PPARγ to increase ABCA1 expression (66). In contrast, Erk1/2 inhibition and LXR activation synergistically induce macrophage ABCA1 expression and efflux (27). Another study recently implicated Mek1/2 in the regulation of PPARγ- and LXRβ-dependent ABCA1 protein degradation in HepG2 cells (28). In addition, Erk1/2 inhibition might affect the ability of other nuclear proteins, including PPARα co-activators and co-repressors (20–25) or even retinoic acid receptors (67), to interact or heterodimerize with PPARα.

Our findings suggest that PPARα-inducible SR-BI degradation pathways identified in mouse liver (36, 37) may also exist in human hepatocytes. Because fibrates down-regulate SR-BI, but also PDZK1, which is essential for maintaining hepatic SR-BI levels (2, 36, 37, 40), it was originally speculated that decreased hepatic SR-BI levels upon PPARα activation might be secondary to decreased PDZK1. Also, an atherogenic diet induces post-translational down-regulation of both SR-BI and PDZK1 in mouse liver (68). However, PPARα agonists reduced hepatic SR-BI in PDZK1-deficient mice (37), pointing at PPARα targeting SR-BI protein stability independent of PDZK1. Similar pathways may exist in humans, as PDZK1 levels remained unaffected in HuH7 hepatocytes treated with PPARα agonists and Mek1/2 inhibitors.

Little is yet known about the degradation pathways that promote PPARα-inducible SR-BI down-regulation (37). This study provides insights into the mechanism driving SR-BI protein turnover. Inhibitors of proteasomal (lactacystin) as well as lysosomal (bafilomycin) degradation increased SR-BI protein levels in HuH7 hepatocytes, indicating that SR-BI can be targeted to the proteasome and lysosome for degradation. Importantly, Mek1/2 inhibition enhanced SR-BI degradation in the presence of bafilomycin, but not lactacystin, suggesting that MAPK are possibly involved in proteasomal degradation pathways. Subcellular fractionation and fluorescence microscopy demonstrate that SR-BI accumulates in (pre)-lysosomes/late endosomes upon inhibition of lysosomal degradation. PD98059 not only reduces SR-BI cell surface expression but also co-localization of SR-BI with the late endosomal marker LBPA in the presence of bafilomycin. Hence, SR-BI down-regulation might occur through various degradation pathways, with the Ras/MAPK pathway regulating the targeting of SR-BI from the cell surface to the proteasome.

Identifying the molecular mechanism that enables Mek1/2 kinases to decrease SR-BI protein stability may provide new clues on the physiological and pathophysiological regulation of hepatic SR-BI. Mouse models have identified SR-BI as a key molecule facilitating hepatic uptake of HDL cholesterol and secretion into bile. This study indicates that PPARα-inducible SR-BI degradation pathways are triggered by HDL or other ligands that activate the Ras/MAPK pathway. As hepatic SR-BI expression positively correlates with biliary cholesterol secretion (2, 33–40), one can speculate that HDL-induced activation of Mek1/2 kinases, followed by PPARα phosphorylation and subsequent SR-BI degradation, is a feedback loop to fine-tune cholesterol homeostasis in the liver.

It is important to note that Erk2 can also phosphorylate PPARγ, which is considered to decrease transcriptional activity of PPARγ (17–19, 69, 70). In fact, HDL-induced and MAPK-mediated phosphorylation of PPARγ inhibited expression of PPARγ-responsive genes in RAW macrophages (53). Hence Ras/MAPK-mediated phosphorylation events could be an important switch to determine the contribution of PPARα and PPARγ in peripheral and hepatic cholesterol metabolism. Depending on the PPARα and PPARγ levels in a given cell, one can envisage that enhanced Ras/MAPK signaling could have opposite effects on HDL receptor and ATP-binding cassette transporter expression.

The diverse action of Erk1/2 kinases on PPARα and PPARγ activity in peripheral and hepatic cells reflects the complexity of signaling cascades and the still poorly understood physiological relevance of PPAR phosphorylation. To date, the following three kinase families have been implicated in PPAR phosphorylation: MAPK (Erk1/2, p38, and JNK), PKA, and PKC (17–19). All of these kinases are involved in interconnected signaling cascades and provide multiple controls at the level of the receptor, ligand, cell type, cofactors, and gene promoter. Several studies suggest that Ras/MAPK overactivation contributes to atherosclerotic lesion development (71). Ras activation because of oxidative stress and H-Ras minisatellite instability in atherosclerotic plaque indicated that increased Ras activity may be involved in atherosclerosis (71). Ras inhibition using farnesyltransferase inhibitors attenuated atherosclerotic lesion formation and reduced oxidative stress in apoE-deficient mice (71, 72) and intimal thickening in the rat carotid injury model (73). Statins, which interfere with Ras prenylation and activity, reduce angiogenesis and plaque progression (74). Although those studies favor Ras/MAPK inhibition to be beneficial in the prevention of atherosclerotic lesion formation, results presented here implicate that Mek1/2 inhibition promotes PPARα-inducible SR-BI degradation in hepatocytes. Future studies to clarify how Ras/MAPK signaling affects hepatic SR-BI protein stability will add to a better understanding of HDL metabolism and reverse cholesterol transport in vivo.