Induction of Low Density Lipoprotein Receptor (LDLR) Transcription by Oncostatin M Is Mediated by the Extracellular Signal-regulated Kinase Signaling Pathway and the Repeat 3 Element of the LDLR Promoter*

Oncostatin M (OM) activates the transcription of the human low density lipoprotein receptor (LDLR) in HepG2 cells through a sterol-independent mechanism. Our previous studies showed that mutations within the repeat 3 sequence of the LDLR promoter significantly decreased OM activity on LDLR promoter luciferase reporter constructs that contain the sterol responsive element-1 (repeat 2) and Sp1 binding sites (repeats 1 and 3). In this study, we investigated the signal transduction pathways that are involved in OM-induced LDLR transcription. In HepG2 cells, OM induced a rapid increase in LDLR mRNA expression, with increases detected at 30 min and maximal induction at 1 h. This OM effect was not blocked by protein synthesis inhibitors, inhibitors of p38 kinase, phosphatidylinositol 3-kinase, or c-Jun N-terminal kinase, but OM activity was completely abolished by pretreating cells with inhibitors of the extracellular signal-regulated kinase (ERK) kinase (mitogen/ERK kinase (MEK)). To investigate whether the repeat 3 sequence of the LDLR promoter is the OM-responsive element that converts ERK activation at the promoter level, three luciferase reporters, pLDLR-TATA containing only the TATA-like elements of the promoter, pLDLR-R3 containing repeat 3 and the TATA-like elements, and pLDLR-234 containing repeats 1, 2, 3 and the TATA-like elements were constructed and transiently transfected into HepG2 cells. OM had no effect on the basal promoter construct pLDLR-TATA; however, including a single copy of repeat 3 sequence in the TATA vector (pLDLR-R3) resulted in a full OM response. The activity of OM on pLDLR-R3 was identical to that of pLDLR-234. Importantly, the ability of OM to increase luciferase activities in both pLDLR-R3- and pLDLR-234-transfected cells was blocked in a dose-dependent manner by inhibition of MEK. These results demonstrate that the mitogen-activated protein kinase MEK/ERK cascade is the essential signaling pathway by which OM activates LDLR gene transcription and provide the first evidence that the repeat 3 element is a new downstream target of ERK activation.


From the Department of Veterans Affairs Palo Alto Health Care System, Palo Alto, California 94304
Oncostatin M (OM) activates the transcription of the human low density lipoprotein receptor (LDLR) in HepG2 cells through a sterol-independent mechanism. Our previous studies showed that mutations within the repeat 3 sequence of the LDLR promoter significantly decreased OM activity on LDLR promoter luciferase reporter constructs that contain the sterol responsive element-1 (repeat 2) and Sp1 binding sites (

repeats 1 and 3). In this study, we investigated the signal transduction pathways that are involved in OM-induced LDLR transcription. In HepG2 cells, OM induced a rapid increase in LDLR mRNA expression, with increases detected at 30 min and maximal induction at 1 h. This OM effect was not blocked by protein synthesis inhibitors, inhibitors of p38 kinase, phosphatidylinositol 3-kinase, or c-Jun Nterminal kinase, but OM activity was completely abolished by pretreating cells with inhibitors of the extracellular signal-regulated kinase (ERK) kinase (mitogen/ ERK kinase (MEK)).
To investigate whether the repeat 3 sequence of the LDLR promoter is the OM-responsive element that converts ERK activation at the promoter level, three luciferase reporters, pLDLR-TATA containing only the TATA-like elements of the promoter, pLDLR-R3 containing repeat 3 and the TATA-like elements, and pLDLR-234 containing repeats 1, 2, 3 and the TATA-like elements were constructed and transiently transfected into HepG2 cells. OM had no effect on the basal promoter construct pLDLR-TATA; however, including a single copy of repeat 3 sequence in the TATA vector (pLDLR-R3) resulted in a full OM response. The activity of OM on pLDLR-R3 was identical to that of pLDLR-234. Importantly, the ability of OM to increase luciferase activities in both pLDLR-R3-and pLDLR-234transfected cells was blocked in a dose-dependent manner by inhibition of MEK. These results demonstrate that the mitogen-activated protein kinase MEK/ERK cascade is the essential signaling pathway by which OM activates LDLR gene transcription and provide the first evidence that the repeat 3 element is a new downstream target of ERK activation.
The transcription of the human low density lipoprotein re-ceptor (LDLR) 1 gene is under the control of intracellular cholesterol and other non-sterol mediators. Cholesterol regulates LDLR transcription through a negative feedback mechanism (1,2). When cellular cholesterol levels rise, LDLR transcription is reduced. When cellular cholesterol storage is depleted, LDLR transcription is activated. The promoter region and the cisacting elements that are responsible for the basal and the cholesterol-controlled transcription of LDLR have been localized to three GC-rich imperfect 16-bp direct repeats (3)(4)(5). The three repeats lie within 100 bp upstream of the transcriptional start site. Repeats 1 and 3 contain Sp1 binding sites that support the basal transcriptional activity of LDLR. Interference of Sp1 binding to either repeat severely decreases basal transcription. Cholesterol regulation is mediated through a 10-bp sequence (5Ј-ATCACCCCAC-3Ј) within repeat 2 designated sterol responsive element-1 (SRE-1) (6,7). Under low intracellular cholesterol conditions, the SRE-1-binding proteins SREBP1 and SREBP2 bind to the SRE-1 sequence and interact synergistically with Sp1 in repeat 3, leading to the activation of LDLR transcription (8 -11). In addition to the SRE-1 site, another cis-acting element, designated FP1, was recently described that is located upstream of the repeat 1 sequence (-145 to -126) (12). The FP1 site participates in LDLR transcription only under conditions of sterol depletion.
In addition to cholesterol and its derivatives, LDLR transcription is also subject to regulation by physiological non-sterol mediators, including growth hormone (13,14), hepatocyte growth factor (15), cytokines (16), and some pharmaceuticals (17). The increased transcription of the LDLR gene induced by these agents has been ascribed to activation of protein kinase C, protein kinase A, and an increase in intracellular Ca 2ϩ (18,19). However, the cis-acting elements in the LDLR promoter and the transcription factors that are responsible for regulation by cholesterol-independent pathways have not been elucidated.
Our laboratory has been involved in characterizing the mechanism by which the cytokine oncostatin M (OM) activates LDLR transcription. OM, a member of the interleukin-6 family of cytokines, is a strong inducer of LDLR transcription in HepG2 cells and in mouse liver (20 -23). The mechanism of OM induction is clearly sterol-independent, because we have demonstrated that mutations within the SRE-1 site that totally abolished cholesterol regulation had no effect on OM inducibility of LDLR transcription. Instead, mutations within the re-peat 3 sequence severely impaired OM activity (22).
In order to identify and characterize the trans-acting factors that mediate OM-induced LDLR transcription by interacting with the repeat 3 sequence, in the present studies, we examined the OM-elicited signal transduction pathways and specific kinases that are involved in LDLR transcription.
Two major signaling pathways can be activated by OM and its related cytokines interleukin-6 (IL-6) and leukemia inhibitory factor (LIF): the JAK/signal transducer and activator of transcription (STAT) pathway (24) and the Ras-mitogen-activated protein (MAP) kinase pathway (25)(26)(27). Binding of OM to its cell surface receptors activates JAK kinases (JAK 1, JAK 2, and tyk), which, in turn, convert latent cytoplasmic transcription factors, known as STATs, into activated forms by tyrosine phosphorylation. The tyrosine-phosphorylated STATs (STAT 1, 3, and 5b) form homo-or heterodimers and translocate into the nucleus, where they bind to their recognition sequences. This signaling pathway is responsible for OM-induced transcription of the cyclin-dependent kinase inhibitor p21 (28), aromatase P450 (29), and genes coding for acute phase proteins (30,31). However, the STAT pathway does not appear to be directly involved in OM-mediated LDLR transcription because no STAT binding sites exist in the OM-responsive region of the LDLR promoter (Ϫ142 to ϩ35). Furthermore, cotransfection of STAT 3 or STAT 5b expression vectors with LDLR promoter reporter constructs did not alter basal or OM-inducible promoter activity (22). Therefore, in the present studies, we focused our investigation on the MAP kinase pathway.
Our findings in this study demonstrate that LDLR transcription is also regulated by the MAP kinase cascade. By using specific inhibitors that block each signaling pathway, we demonstrate that inhibition of ERK activation by inactivating its upstream kinase, MEK, totally prevented OM-induced LDLR mRNA expression. We also demonstrate that the activity of OM on the LDLR promoter is mediated through the repeat 3 element. More importantly, we show that the inhibition of ERK activation totally abolished OM activity on the LDLR promoter. These data for the first time localize an ERK responsive cisacting element in the LDLR promoter and provide the first evidence that the repeat 3 sequence is a novel downstream target of MAP kinase activation.

MATERIALS AND METHODS
Cells and Reagents-The human hepatoma cell line HepG2 was obtained from American Type Culture Collection (Manassas, VA) and was cultured in Eagle's minimum essential medium supplemented with 10% fetal bovine serum (Summit Biotechnology, Fort Collins, CO). Purified human recombinant OM was provided by Bristol-Myers Squibb (Princeton, NJ). Cholesterol and 25-hydroxycholesterol, anisomycin, cycloheximide, puromycin, emetine, and the PI 3-kinase inhibitor wortmannin were purchased from Sigma. The MEK inhibitor U0126 was obtained from DuPont Merck Pharmaceutical Co. The MEK inhibitor PD98059 was purchased from New England Biolabs (Beverly, MA). The p38 kinase inhibitor SB-203580 was obtained from Smith-Kline Beecham Pharmaceuticals (King of Prussia, PA). The rabbit polyclonal antibodies to detect inactive ERK2 (C-14, catalog no. sc-154) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The rabbit polyclonal antibodies (V803A) against activated ERK1 and ERK2 were obtained from Promega, Inc. (Madison, WI).
Plasmid Vectors and Oligonucleotides-The plasmid pLDLR-234 was constructed by subcloning a 177-bp fragment of the LDLR promoter obtained by HindIII digestion of pLDLR-CAT 234 (3) into HindIII-digested pGL3-basic vector (Promega). To construct the pLDLR-TATA vector, a double-stranded oligonucleotide corresponding to LDLR promoter sequence ϩ13 to -36 (TCGAGtagaaacctcacattgaaatgctgtaaatgacgtgggccccgagtgcaatA) was synthesized with XhoI and HindIII sites (italic) at the 5Ј and 3Ј ends, respectively, and was cloned between the XhoI and HindIII sites of pGL3basic. The vector pLDLR-R3 was constructed by cloning of a doublestranded oligonucleotide containing the LDLR promoter sequence from ϩ13 to -52 with SacI and HindIII sites (italic) at the 5Ј and 3Ј ends (Caaactcctc-cccctgctagaaacctcacattgaaatgctgtaaatgacgtgggccccgagtgcaatA) into pGL3-basic vector. The repeat 3 sequence is underlined.
RNA Isolation and Northern Blot Analysis-Cells were lysed in Ultraspec RNA lysis solution (Biotecxs Laboratory, Houston, TX), and total cellular RNA was isolated according to the vendor's protocol. Approximately 15 g of each total RNA sample was used to analyze the LDLR mRNA as described previously. The RNA blots were first hybridized to a 0.84-kilobase 32 P-labeled human LDLR probe (22) and then stripped and reprobed with a human GAPDH probe to ensure that equivalent amounts of RNA were being loaded. Differences in hybridization signals of Northern blots were quantitated by the Bio-Rad Fluro-S MultiImager system. Densitometric analysis of autoradiographs in these studies included various exposure times to ensure linearity of signals.
Western Blot Analysis-HepG2 cells in 60-mm culture dishes were scraped into 0.2 ml of cold lysis buffer (20 mM HEPES, pH 7.4, 30 mM p-nitrophenyl phosphate, 10 mM NaF, 10 mM MgCl 2 , 2 mM EDTA, 5 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 0.1 mM Na 2 MoO 4 , 10 mM sodium ␤-glycerophosphate, 10 nM okadaic acid, 10 nM cypermethrin, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 1 g/ml leupeptin, and 1.25 g/ml pepstatin) and disrupted by repeated passages through a 26 gauge needle. Extracts were centrifuged for 30 min at 10,000 ϫ g and stored at Ϫ80°C. Samples of 30 g of soluble protein per lane were electrophoresed through a 12% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The membranes used to probe the inactive ERK1 and ERK2 were first blocked in 5% powered milk and 0.05% Tween 20 in phosphate-buffered saline for 2 h, and then incubated with the same blocking buffer containing anti-ERK antibody at the concentration of 1 g/ml overnight at 4°C. The antibodies against inactive ERK primarily recognize ERK2. The membranes used to probe activated ERK1 and ERK2 were blocked in TBST buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween 20) containing 1% bovine serum albumin for 2 h and then incubated in TBST buffer containing 0.1% bovine serum albumin and antibody at a 1:5000 dilution overnight at 4°C. Detection was performed using a horseradish peroxidase-conjugated secondary anti-rabbit antibody and enhanced chemiluminescence (ECL kit, Amersham Pharmacia Biotech).
Transient Transfection Assays-HepG2 cells were transiently transfected with plasmid DNA by the method of calcium phosphate coprecipitation (22). To demonstrate sterol-independent regulation, cells transfected with pLDLR-234 vector were cultured in media containing 10% fetal bovine serum and cholesterol (10 g/ml, plus 1 g/ml 25hydroxycholesterol). For the studies of kinase inhibitors, 24 h after termination of transfection, inhibitors at the indicated concentrations were added 1 h before 50 ng/ml OM was added. After 4 h of OM treatment, cells were washed twice with phosphate-buffered saline and lysed with 150 l of reporter lysis buffer (Promega). The luciferase assay (70 l of lysate/sample) was conducted with Promega luciferase assay reagents, and the luciferase activity was measured in a luminometer (Promega, model TD-20/20). ␤-Galactosidase activity (50 l of lysate) was measured according to standard methods. Absolute luciferase activity was normalized against ␤-galactosidase activity to correct for transfection efficiency. Triplicate wells were assayed for each transfection condition, and at least three independent transfection assays were performed for each reporter construct.

RESULTS
The kinetics of OM-induced LDLR gene expression in HepG2 cells was examined by Northern blot analysis. Total RNA was isolated from cells that were treated with 50 ng/ml OM for different periods from 30 min to 48 h. OM induced a rapid increase in LDLR mRNA level. This increase was detected as early as 30 min after OM addition to the cells, reached a maximal level (3.4-fold of control) at 1 h, and slowly declined to baseline at 48 h (Fig. 1). Results from five separate experiments, after normalization with GAPDH, showed that stimulation of cells with OM for 1 h increased LDLR mRNA to an average of 4.2-fold (4.2 Ϯ 1.2, S.D.), as compared with unstimulated cells.
The rapid action of OM on LDLR mRNA expression suggests that OM may activate LDLR transcription in the absence of new protein synthesis. To investigate this issue, HepG2 cells were first incubated with the protein synthesis inhibitors, anisomycin (12.5 g/ml), cycloheximide (10 g/ml), puromycin (150 g/ml), and emetine (10 g/ml) for 1 h prior to addition of OM to the cells. Analysis of protein synthesis with labeled [ 3 H]leucine indicated that more than 99% of protein synthesis occurring in HepG2 cells was blocked by these inhibitors at the doses used. Fig. 2 shows that the protein synthesis inhibitors had different effects on the basal LDLR mRNA levels. The normalized LDLR mRNA level was slightly decreased in cycloheximide treated cells (71% of control), but it was significantly increased in anisomycin-treated (258% of control) and emetinetreated (183% of control) cells. However, the OM-induced upregulation of LDLR mRNA expression was not inhibited by any of the inhibitors; instead, LDLR mRNA was further increased. This result confirmed our speculation that activation of LDLR transcription by OM does not require new protein synthesis.
Three kinase pathways have been shown to be activated by OM and its related cytokines: JAK/STAT pathway (24,30), MAP kinase pathway (25)(26)(27), and PI 3-kinase (32). Because our previous studies already excluded the role of activated STAT in LDLR transcription (22), we focused the present investigation on PI 3-kinase and the MAP kinases. In mammalian cells, extracellular signals are transduced through three separate MAP kinase cascades, including MAP kinase (also known as extracellular signal-regulated kinase (ERK) (33)), the c-Jun N-terminal kinase (JNK) (34), and P38 kinase (35).
To determine which kinase pathway is involved in OM-induced LDLR mRNA expression, HepG2 cells were stimulated with 50 ng/ml OM for 1 h in the presence or the absence of MEK inhibitors U0126 and PD98059, p38 kinase inhibitor SB-203580, and PI 3-kinase inhibitor wortmannin. Total RNA was isolated and the level of LDLR mRNA was analyzed by Northern blot. The result showed that the MEK inhibitors, U0126 and PD90586, were extremely effective in preventing OM-induced LDLR mRNA expression. The majority of the OM activity was inhibited by U0126 at 1 M and PD98059 at 10 M. In contrast, wortmannin at concentrations up to 100 nM, which has been shown to completely inhibit PI 3-kinase activity (32), had no effect. The p38 kinase inhibitor SB-203580 did not affect OM activity at its effective concentration (IC 50 ϭ 0.6 M) (36) but moderately lowered OM activity (30%) at 25 M (Fig. 3).
The possible role of the third MAP kinase cascade, JNK, in OM-induced LDLR transcription was examined by using curcumin. Curcumin has been shown to inhibit the activations of JNK and ERK at different concentrations (37). The JNK pathway is more sensitive to curcumin with an IC 50 of 5-10 M. At higher concentrations, activation of ERK by MEK is inhibited with an IC 50 of 20 M. Treatment of cells with up to 25 M curcumin lowered the OM-induced increase in LDLR mRNA only 35% (Fig. 4). This small effect noted at high concentrations of curcumin suggests that the inhibitory activity of curcumin was more likely mediated through its effect on ERK activation rather than JNK activation.
The compound U0126 is a newly developed inhibitor that specifically inhibits MEK (38). It strongly inhibits the activity of MEK1 and MEK2 (IC 50 ϭ 1 M) without detectable effects on other kinases, such as PKC, Raf, ERK, and JNK. Because our data showed a higher potency of U0126 than PD98059 (39) on OM-stimulated LDLR mRNA expression, the dose-dependent inhibitory effect of U0126 on OM activity was further examined. Fig. 5 shows a representative Northern blot. U0126 inhibited OM activity with an IC 50 of 0.5-1 M, with complete inhibition achieved at 5 M and higher. This dose range is in excellent agreement with its reported effect on MEK. The basal level of LDLR mRNA was slightly decreased by the compound. These results together demonstrate that inhibition of MEK activity totally prevented OM induction of LDLR transcription, whereas inhibitors of p38 kinase and JNK kinase only partially inhibited OM activity at doses far exceeding their IC 50 s. Based upon these data, we conclude that the MAP kinase MEK/ERK cascade is the major signaling pathway through which OM activates LDLR transcription.
MAP kinases ERK1 and ERK2 are the only known substrates of MEK1 and MEK2 (40). MEK activates ERK by inducing phosphorylation on both threonine and tyrosine residues of ERK. Inhibition of MEK kinase activity will directly block ERK activation. Western blot analysis using antibody that detects inactive ERK and antibody that only recognizes the double-phosphorylated, activated ERK was conducted to examine OM-induced activation of ERK. As shown in Fig. 6, neither OM nor U0126 treatments altered the level of expression of inactive ERK1 and ERK2. In contrast, the detectable level of activated ERK1 and ERK2 was rapidly increased by OM stimulation. A 3.7-fold increase was detected at 5 min, reached a maximal level (4.2-fold of control) at 20 min, and slowly declined to baseline after 60 min. Complete inhibition of OM-induced ERK activation was demonstrated by using 10 M U0126 (Fig. 6, ϩ). This result documents that OM activates both ERK1 and ERK2 and that this activation immediately precedes OM-induced LDLR transcription.
Our next goal was to determine the promoter region and the cis-acting element that serves as the functional downstream target of the OM-activated MEK/ERK cascade. Previously, by deletion analysis, we localized the OM response to the LDLR promoter region from -142 to ϩ35 (pLDLR-234), which contains repeats 1, 2, and 3 and the TATA-like elements. Mutagenesis studies of each repeat showed that OM activity was only diminished when repeat 3 was mutated (22). To prove that repeat 3 is the response element that mediates OM effects, in the present studies, we constructed two luciferase reporter vectors that contain LDLR promoter TATA-like sequences with or without repeat 3. PLDLR-TATA was made by insertion of FIG. 1. Kinetics of LDLR mRNA expression in HepG2 cells treated with OM. HepG2 cells cultured in Eagle's minimum essential medium containing 10% fetal bovine serum were incubated with 50 ng/ml OM for different times as indicated. Total RNA was isolated, and 15 g per sample was analyzed for LDLR mRNA by Northern blot. The membrane was stripped and hybridized to a human GAPDH probe. The fold induction of LDLR mRNA level with OM treatment normalized to GAPDH mRNA over untreated cells was as follows: 30   the LDLR promoter sequence from -36 to ϩ13 relative to the transcription start site into the pGL3-basic luciferase vector. PLDLR-R3 contains the LDLR promoter sequence from -52 to ϩ13. The effect of OM on pLDLR-TATA and pLDLR-R3 was then compared with its effect on pLDLR-234. These reporters, along with the ␤-galactosidase expression vector pRSV-␤-galactosidase, were transiently transfected into HepG2 cells and luciferase activity and ␤-galactosidase activity were measured. Fig. 7A illustrates a representative experiment that shows the basal promoter activities and the response to OM of the three reporter vectors. PLDLR-TATA produced a low level of luciferase activity and showed no response to OM. Including a single copy of repeat 3 with the TATA sequence (pLDLR-R3) increased the basal level of luciferase activity 3-5-fold and, importantly, created a promoter that was responsive to OM. OM increased luciferase activity in pLDLR-R3-transfected cells to an extent similar to that observed in pLDLR-234-transfected cells. Fig. 7B summarizes the results from over 10 independent transfections and shows that OM increased pLDLR-234 and pLDLR-R3 promoter activities 3.3-fold and 3.4-fold, respectively, without affecting pLDLR-TATA. These data provide strong evidence that the OM effect on the LDLR promoter is mediated through the repeat 3 sequence.
We next examined whether blocking ERK activation by inhibition of MEK could affect the ability of OM to increase the activity of the LDLR promoter. HepG2 cells were transfected with pLDLR-234 or pLDLR-R3. Twenty hours after transfection, U0126 was added to the cells at different doses for 1 h, followed by OM. U0126 produced a dose-dependent inhibition of OM stimulation of both the full promoter (pLDLR-234) and the minimum promoter containing a single copy of repeat 3 (pLDLR-R3) (Fig. 8). The dose of U0126 that effectively blocked OM-induced LDLR promoter activity is in a range similar to that of inhibition of OM-induced LDLR mRNA expression.
These data suggest that an ERK-responsive transcription activator is involved in LDLR transcription through interaction with the repeat 3 sequence.
Finally, to investigate whether the repeat 3 element in the LDLR promoter only mediates OM signaling or whether it is used by other members of the IL-6 family of cytokines, the activities of IL-6 and LIF on pLDLR-234 and pLDLR-R3 were examined and compared with OM in the presence or the absence of the MEK inhibitor U0126. All the cytokines at their maximally effective concentrations increased LDLR promoter activity, and their activities were inhibited by U0126 (Fig. 9). The abilities of these cytokines to increase LDLR promoter activity was in the order OM 3 IL-6 3 LIF, which is consistent with their abilities to activate MAP kinase in HepG2 cells (25).

DISCUSSION
Cytokines bind to their cell surface receptors and initiate intracellular signaling by activating members of the Janus kinase family of tyrosine kinase. Activated JAK can initially activate two separate pathways, the STAT pathway and the Ras pathway. Activated Ras has an essential role in the acti-vation of Raf kinase, which directly phosphorylates and activates the ERK upstream kinases MEK1 and MEK2. Activated MEK, in turn, activates ERK1 and ERK2 by inducing phosphorylation of threonine and tyrosine residues in ERK. Because it has been previously reported that OM activates Ras and Raf (27,(41)(42)(43), we were interested in examining the downstream effector of this signaling cascade that might play a role in the stimulation of LDLR transcription. By using inhibitors that block different signal pathways, we show that the MEK inhibitors U0126 and PD98059 are effective in blocking OM-induced LDLR transcription, whereas inhibitors of PI 3-kinase, JNK, and p38 kinase had little or no effect. The observed small inhibitory effects of p38 kinase inhibitor SB-203580 at 25 M on OM-induced LDLR mRNA expression was due to its nonspecific effects on MEK, as the OM-induced ERK activation detected by Western blot was also inhibited by SB-203580 at 25 M (data not shown). The IC 50 (0.5-1 M) of U0126 on OM activity is very similar to that reported for its inhibition of MEK enzyme activity. These data provide convincing evidence that the MEK/ERK cascade is the essential signaling pathway by which OM activates LDLR transcription.
The involvement of ERK activation in OM-induced LDLR transcription is also supported by the tight association of these two biological events. Activation of ERK by OM was detected within 5 min, reaching a maximum at 12-20 min and declining to baseline by 60 min. This process was immediately followed by an increase in LDLR mRNA expression. LDLR mRNA levels started to rise after 30 min of exposure to OM, reached a maximal level at 1 h, and slowly decreased afterward. The closely linked kinetics of ERK activation and LDLR transcription suggest that these two events elicited by OM are mechanistically related.
Establishment of the essential role of ERK activation in OM-induced LDLR transcription raises a critical question: which cis-acting element in the LDLR promoter is the downstream target of ERK activation? Previously, we have localized the OM response to the LDLR promoter region from -142 to ϩ35 (pLDLR-234), which contains repeats 1, 2, and 3 and the TATA-like elements. Mutations within the repeat 3 sequence of pLDLR-234 significantly lowered OM activity (22). In this study, we further demonstrate that OM activity is mediated through the repeat 3 sequence. OM had no effect on the basal LDLR promoter (pLDLR-TATA, Ϫ36 to ϩ13) that contains the TATA-like elements. Placing the 16-bp repeat 3 sequence 5Ј to the basal promoter (pLDLR-R3, Ϫ52 to ϩ13), as is found in the native promoter configuration, generated the same OM response as that seen with the longer promoter fragment (pLDLR-234). Importantly, the OM-induced responses of pLDLR-234 and pLDLR-R3 were both prevented by the MEK inhibitor U0126. These results identify for the first time a region in the LDLR promoter that responds to MAP kinase activation and demonstrate that the repeat 3 element is a new downstream target of the MEK/ERK cascade.
Several cis-acting elements that mediate the effects of the IL-6 type cytokines on gene transcription have been described. These include nuclear factor-IL-6 (44), STAT binding site (24), and Ap1 (41,45). In this study, we demonstrate that OM, IL-6, and LIF are all capable of inducing LDLR transcription through the repeat 3 sequence, although their relative potencies differ. These data suggest that the repeat 3 sequence contains a novel responsive element that mediates the effects of OM and its related cytokines on transcriptions of a subset genes.
Upon activation, ERK translocates to the nucleus, where it phosphorylates a number of transcription factors, including members of the Ets family, ELK, c-Jun, and c-Myc (46 -48). Phosphorylation of ELK by ERK potentiates the DNA binding The LDLR promoter luciferase reporter constructs pLDLR-TATA, pLDLR-R3, and pLDLR-234 were transiently transfected into HepG2 cells along with a ␤-galactosidase expression vector, pRSV-␤galactosidase. Twenty hours after transfection, cells were stimulated with 50 ng/ml OM for 4 h. Luciferase expression was normalized to ␤-galactosidase activity to correct for variations in transfection efficiency. A, basal and OM-induced luciferase activity of the three promoter constructs. B, effects of OM on the three promoter constructs. The normalized luciferase activity of transfected cells that were untreated is expressed as 100%. The data (mean Ϯ S.E.) shown were derived from 10 -12 independent transfection assays. Solid columns, without OM; checked columns, with OM. of ELK and the formation of a ternary complex at the serum response element of the c-fos promoter, resulting in an immediate activation of transcription of the c-fos gene (49). Interestingly, comparison of the published DNA recognition sequences of the known ERK substrates, such as SRF or c-Myc, with the repeat 3 sequence (AAACTCCTCCCCCTGC) does not reveal any significant sequence homology.
Currently, Sp1 and the Sp1-related protein, Sp3, are the only transcription factors that have been shown to bind to the repeat 3 sequence. Sp1 was shown to be phosphorylated by a DNA-dependent protein kinase, but this phosphorylation did not affect the extent and specificity of DNA binding or Sp1-dependent transcription (50). On the other hand, a recent study showed that Sp1 in HL-60 leukemia cells was phosphorylated by protein kinase A, and the Sp1 DNA binding and transactivating activities were also stimulated by protein kinase A (51). We have considered the possibility that Sp1 is phosphorylated by ERK or an ERK substrate upon OM stimulation, resulting in a increased transcriptional activity of LDLR gene. However, we could not detect any OM-induced changes in Sp1 binding by gel shift assay or in the ratio of hyperphosphorylated versus the hypophosphorylated Sp1 by Western blot analysis (22). Furthermore, the MEK inhibitor U0126 at concentrations that totally blocked ERK activation had no effect on Sp1 DNA binding activity (data not shown). At present, it is uncertain whether Sp1 is phosphorylated by ERK or downstream kinases at specific residues without changing Sp1 DNA binding activity. This phosphorylation might somehow alter the affinity of Sp1 to interact with other transcription factors or cofactors. Alternatively, an unknown transactivator might be a direct or indirect substrate of ERK. Upon activation by OM, this factor could either associate with Sp1 as a cofactor or directly bind to the repeat 3 sequence, thereby stimulating LDLR transcription. These important issues are currently under investigation.
Recently, Kumar et al. (52) reported that IL-1␤-induced LDLR transcription was mediated through ERK activation. However, comparison of the characteristics of IL-1␤ with OM on LDLR transcription demonstrates that these two cytokines regulate LDLR transcription through different mechanisms. First, the kinetics of OM in inducing LDLR mRNA expression is much faster than that of IL-1␤. OM activity reached a maximum at 1 h and slowly declined afterward, whereas a significant increase in LDLR mRNA level induced by IL-1␤ was not seen until 4 h and levels continued to increase after that. Second, cholesterol has no effect on OM induction of LDLR mRNA expression or on LDLR promoter activity, but cholesterol totally abolished the IL-1␤-induced increase in LDLR mRNA expression. Third, unlike IL-1␤, the activity of which on LDLR mRNA was potentiated by SB-203580 at concentrations of 1-5 M, the OM-induced LDLR mRNA expression was not significantly affected by SB-203580 at 1-10 M and was slightly inhibited at 25 M, suggesting that p38 kinase may be involved in IL-1␤ signaling, but it does not play a role in OM-induced LDLR transcription. Furthermore, the effect of IL-1␤ on the LDLR promoter was examined within a reporter construct containing a 1563-bp fragment of the LDLR promoter. Therefore, it is not known what region of the LDLR promoter was responsible for mediating the IL-1␤ effect and, critically, whether the IL-1␤-responsive element is the same cis-acting element that responds to ERK activation. These striking differences, as well as the involvement of ERK activation in both OM-induced and IL-1␤-induced LDLR transcription, suggest that ERK activation might be involved both in cholesterol-dependent and cholesterol-independent regulation of LDLR transcription.
In conclusion, we have demonstrated that the MAP kinase MEK/ERK cascade is the essential signaling pathway by which OM activates LDLR gene transcription. The repeat 3 sequence of the LDLR promoter not only functions as an Sp1 binding site that controls basal transcription but also is an OM-responsive element that converts ERK activation at the LDLR promoter level. These results provide critical clues for further characterizing the transcriptional complex that controls LDLR transcription independent of the SRE-1/SREBP pathway.