One-way Cross-talk between p38 MAPK and p42/44 MAPK INHIBITION OF p38 MAPK INDUCES LOW DENSITY LIPOPROTEIN RECEPTOR EXPRESSION THROUGH ACTIVATION OF THE p42/44 MAPK CASCADE*

In this paper, we report that SB202190 alone, a specific inhibitor of p38 MAPK , induces low density lipoprotein (LDL) receptor expression (6–8-fold) in a sterol-sensitive manner in HepG2 cells. Consistent with this finding, selective activation of the p38 MAPK signaling pathway by expression of MKK6b(E), a constitutive activator of p38 MAPK , significantly reduced LDL receptor promoter activity. Expression of the p38 MAPK a -isoform had a similar effect, whereas expression of the p38 MAPK b II-isoform had no significant effect on LDL receptor promoter activity. SB202190-dependent increase in LDL receptor expression was accompanied by induction of p42/44 MAPK , and inhibition of this pathway completely prevented SB202190-induced LDL receptor expression, suggesting that p38 MAPK negatively regulates the p42/ 44 MAPK cascade and the responses mediated by this kinase. Cross-talk between these kinases appears to be one-way because modulation of p42/44 MAPK activity did not affect p38 MAPK activation by a variety of stress in-ducers. Taken together, these findings reveal a hitherto unrecognized one-way communication that exists between p38 MAPK and p42/44 MAPK and provide the first evidence that through the p42/44 MAPK signaling cascade, the p38 MAPK a -isoform negatively regulates LDL receptor expression, thus representing a novel mechanism of fine tuning cellular levels of cholesterol in response to a diverse

In this paper, we report that SB202190 alone, a specific inhibitor of p38 MAPK , induces low density lipoprotein (LDL) receptor expression (6 -8-fold) in a sterolsensitive manner in HepG2 cells. Consistent with this finding, selective activation of the p38 MAPK signaling pathway by expression of MKK6b(E), a constitutive activator of p38 MAPK , significantly reduced LDL receptor promoter activity. Expression of the p38 MAPK ␣-isoform had a similar effect, whereas expression of the p38 MAPK ␤II-isoform had no significant effect on LDL receptor promoter activity. SB202190-dependent increase in LDL receptor expression was accompanied by induction of p42/44 MAPK , and inhibition of this pathway completely prevented SB202190-induced LDL receptor expression, suggesting that p38 MAPK negatively regulates the p42/ 44 MAPK cascade and the responses mediated by this kinase. Cross-talk between these kinases appears to be one-way because modulation of p42/44 MAPK activity did not affect p38 MAPK activation by a variety of stress inducers. Taken together, these findings reveal a hitherto unrecognized one-way communication that exists between p38 MAPK and p42/44 MAPK and provide the first evidence that through the p42/44 MAPK signaling cascade, the p38 MAPK ␣-isoform negatively regulates LDL receptor expression, thus representing a novel mechanism of fine tuning cellular levels of cholesterol in response to a diverse set of environmental cues.
Analysis of the signal transduction pathways using the above inhibitors revealed a critical role of p42/44 MAPK activation during induction of low density lipoprotein (LDL) receptor gene expression by a variety of extracellular stimuli (43,44). 2 In this paper, we have directly addressed the physiological role of p38 MAPK in the regulation of LDL receptor expression by using highly specific pharmacological and molecular tools. Results presented demonstrate that simple inhibition of the p38 MAPK basal activity is sufficient to induce LDL receptor expression in a p42/44 MAPK -dependent manner. Co-transfection studies established that SB202190-induced LDL receptor expression is mediated by the activation of p42/44 MAPK resulting from the inhibition of p38 MAPK ␣-isoform. Therefore, in intact cells, p38 MAPK ␣-isoform negatively regulates p42/ 44 MAPK and the responses mediated by this kinase. We speculate that cross-talk between these MAPKs, which mediate the effects of numerous extracellular stimuli, could be crucial for controlling a wide array of biological processes.  MN). TRIzol and all tissue culture supplies were from Life Technologies, Inc. Zeta probe blotting membrane and the protein assay reagent were purchased from Bio-Rad. [␣-32 P]dCTP (3000 Ci/mmol) was obtained from DuPont, and the enhanced chemiluminescence (ECL) detection kit was obtained from Amersham Pharmacia Biotech. A light chemiluminescent reporter gene assay system for the detection of luciferase activity was purchased from TROPIX, Inc.

Materials
Cell Culture-Human hepatoma cell line HepG2 and its derivative HepG2-⌬Raf-1:ER cell line that stably expresses the ⌬Raf-1:ER chimera were maintained as monolayer cultures in a humidified 5% CO 2 atmosphere at 37°C in Eagle's minimum essential medium (BioWhittaker) supplemented with 10% fetal bovine serum (Life Technologies), 2 mM L-glutamine, 20 units/ml penicillin, and 20 g/ml streptomycin sulfate.
Northern Analysis-Total RNA was isolated using TRIzol, and Northern blotting was done essentially as described earlier (43,44). Briefly, 20 g of total cellular RNA was fractionated on 1% formaldehyde agarose gel and transferred to Zeta Probe membrane by capillary blotting. RNA blots were hybridized with LDL receptor and squalene synthase-specific single-stranded M13 probes labeled with [␣-32 P]dCTP. In most cases, the same blot was rehybridized with 32 Plabeled single-stranded M13 probe specific for ␤-actin. Hybridized filters were washed and exposed to Kodak x-ray film. The relative intensities of specific bands were determined densitometrically within the linear range of the film on a model 300A laser densitometer (Molecular Dynamics, Inc., Sunnyvale, CA) with ImageQuant software. LDL receptor mRNA was normalized to either squalene synthase or ␤-actin mRNA level, and data for each point were plotted as the percentage of LDL receptor mRNA as compared with controls.
Expression Vectors and Reporter Constructs-Expression vectors of MKK6b (wild type), MKK6b(E) (a constitutive mutant of MKK6b in which serine 207 and threonine 211 are replaced with glutamic acid), MKK6b(A) (a dominant negative mutant of MKK6b in which lysine 82 is replaced with alanine), p38 MAPK ␣, p38 MAPK ␣(AF) (a double mutant of p38 MAPK ␣-isoform in which threonine 188 and tyrosine 190 are substituted with phenylalanine), and p38 MAPK ␤II have been described elsewhere (19,26). Reporter gene of LDL receptor promoter-luciferase (plasmid A), in which the 5Ј-flanking region of human LDL receptor promoter was fused to firefly luciferase gene, has also been described previously (46). 2 Transient Transfections and Luciferase Assay-HepG2 cells were transfected by the Lipofectamine method. For transfection experiments, HepG2 cells were seeded at a density of 2.5 ϫ 10 5 cells of six-well plate 1 day in advance. Transfections were performed in duplicate with 0.75 g of DNA for plasmid A and the indicated amounts of the relevant expression vector or the corresponding empty vector (46). 2 After 5 h, the cells were washed with basic medium and refed with Eagle's minimum essential medium containing 10% fetal calf serum. Approximately 12-16 h later, transfected cells were switched to medium containing 0.5% fetal calf serum, and the cells were incubated for an additional 8 h. Finally, dishes were washed with phosphate-buffered saline and lysed with luciferase lysis buffer (100 mM potassium phosphate, pH 7.8, 0.2% Triton X-100, and 0.5 mM dithiothreitol). Luciferase activity was measured according to the TROPIX protocol. Data are representative for at least four independent experiments performed in duplicate and are expressed as -fold increase in luciferase activity, which was calculated relative to the basal level of LDL receptor promoter reporter activity (set to 1 unit) and corrected for empty vector effects for each expression vector.

RESULTS
The Specific p38 MAPK Inhibitor SB202190 by Itself Induces LDL Receptor Expression-To test whether the p38 MAPK is involved in the regulation of LDL receptor expression, we examined first the effects of inhibition of this kinase by using a specific inhibitor. The activity of p38 MAPK was inhibited by using SB202190, a highly selective inhibitor of p38 MAPK that does not affect the activity of other relevant kinases even at high concentrations. HepG2 cells were treated with SB202190 for various times, and the effect on LDL receptor expression was measured by Northern blotting. SB202190 induces LDL receptor expression in a time-dependent manner without significantly affecting expression of a housekeeping actin gene or another sterol-sensitive gene of the cholesterol biosynthetic pathway, squalene synthase (SS, Fig. 1). A significant increase in LDL receptor expression was apparent at 4 h, and a plateau was reached after an 8-h treatment of SB202190. Another specific p38 MAPK inhibitor SB203580 also induced LDL receptor expression, whereas the inactive derivative SB202474 had no effect on LDL receptor expression (data not shown). The most effective concentrations were 2.5 M for SB202190 and 10 M for SB203580.
Because LDL receptor is negatively regulated by sterols, we then tested whether SB202190 can induce LDL receptor expression in the presence of sterols. As shown in Fig. 2, the presence of sterols suppressed SB202190-induced LDL receptor expression. However, a slight increase in the suppressed levels of LDL receptor was observed in a dose-dependent manner upon treatment with SB202190.
Because pyridinyl imidazoles that are closely related to SB203580 have cyclooxygenase-inhibitory activity (47,48), ex-periments were conducted to determine whether blockade of this activity with indomethacin (49) can alter LDL receptor expression. Importantly, unlike SB202190, up to 20 M indomethacin did not significantly alter LDL receptor expression (results not shown). From the above data, it seemed probable that the effect of SB202190 on LDL receptor expression is related to its inhibition of p38 MAPK .
The p38 MAPK Pathway Negatively Regulates LDL Receptor Expression-To investigate the role of the p38 MAPK signaling pathway in the regulation of LDL receptor expression more directly, we used previously characterized expression constructs to modulate either positively or negatively the endogenous p38 MAPK activity. LDL receptor transcription was monitored by transfecting HepG2 cells with a previously cloned fragment of the human LDL receptor promoter fused to the luciferase reporter gene (46), together with the relevant constructs. In several cell systems, p38 MAPK activity can be stimulated by coexpression of p38 MAPK with MKK6b or by expression of constitutively activated MKK6b(E), in which the activating phosphorylation residues Ser 207 and Thr 211 were replaced by glutamic acids (19,50). Expression of MKK6b(E) reduced luciferase gene expression at least 3-fold, when compared with the luciferase expression in control cells transfected with the empty vector (Fig. 3). Conversely, expression of the dominant negative regulatory MKK6b(A) construct slightly increased LDL receptor-luciferase expression, correlating with a partial inhibition of endogenous MKK6b activity in cells expressing this mutated form of MKK6b. To rule out the possibility that MKK6b(E) may suppress LDL receptor promoter activity in a p38 MAPK -independent manner, we tested whether MKK6b(E)-inhibited LDL receptor expression is due to activation of p38 MAPK . HepG2 cells were transfected with the human LDL receptor promoter-luciferase reporter gene along with the expression vectors encoding p38 MAPK ␣-isoform, p38 MAPK ␤IIisoform, MKK6b, or empty vector. It is interesting to note that expression of either the p38 MAPK ␣-isoform or MKK6b was able to suppress reporter gene expression mildly. However, coexpression of the p38 MAPK ␣-isoform and MKK6b together strongly reduced luciferase activity 6-fold (Fig. 3). The effect of MKK6b is dependent on p38 MAPK ␣-isoform activation, since MKK6b failed to stimulate reporter gene activity when it was co-transfected with an inactive p38 MAPK ␣(AF) mutant, in which one of the activating phosphorylation residues, Tyr 182 , was replaced by phenylalanine. At the same time, expression of the p38 MAPK ␤II-isoform either alone or together with MKK6b had no significant effect on LDL receptor promoter activity. These results demonstrate that activation of the p38 MAPK by

p38 MAPK Negatively Regulates LDL Receptor Expression
itself is sufficient to suppress LDL receptor gene expression. In addition, we conclude that it is the p38 MAPK ␣-isoform, and not the p38 MAPK ␤II-isoform, that mediates SB202190-induced LDL receptor expression.
SB202190-induced LDL Receptor Expression Is Mediated by the p42/44 MAPK Signaling Cascade-Since p42/44 MAPK have been found to play an important role during induction of LDL receptor gene expression (43, 44), 2 we initially determined whether the inhibition of p38 MAPK by SB202190 leads to the activation of other MAPKs. HepG2 cells were incubated with SB202190, and the activities of the p42/44 MAPK and p46/54 JNK were determined by immunoblot analysis with antibodies that recognized the activated phosphorylated forms of the these kinases in cell extracts obtained at different times. In contrast to the rapid and transient growth factor-induced activation of p42/44 MAPK , 2 SB202190 treatment caused a delayed and prolonged activation with no effect on p42/44 MAPK or MKP-1 protein levels in HepG2 cells (Fig. 4A). Weak activation was observed at 4 h, and maximal induction was obtained approximately 8 h following stimulation. The activated p42/ 44 MAPK remained elevated at least until 24 h, the maximum time point measured. Furthermore, changes in p42/44 MAPK phosphorylation took place with little or no effect on the phosphorylation (Fig. 4A) or expression levels of p46/54 JNK (data not shown). Kinetics of MEK-1/2 phosphorylation paralleled an increase in p42/44 MAPK phosphorylation, suggesting that MEK-1/2 is also activated upon SB202190 treatment. Because MEK-1/2 directly phosphorylates and activates p42/44 MAPK , PD98059 was used to determine the role of this pathway in SB202190-induced LDL receptor expression. Inhibition of SB202190-induced LDL receptor expression by PD98059 (Fig.   4B) suggests that the induction of LDL receptor expression is mediated by the p42/44 MAPK signaling pathway. The lack of signficant effect of sterols on SB202190-induced p42/44 MAPK phosphorylation ruled out involvement of a sterol-sensitive step in p42/44 MAPK activation (Fig. 4C). Overall, the above results suggested that p38 MAPK negatively regulates p42/ 44 MAPK phosphorylation and that SB202190-dependent increase in LDL receptor expression is mediated by the p42/ 44 MAPK signaling cascade.
We then tested whether the inhibition of p38 MAPK by SB202190 induces LDL receptor expression in other cell types. In HeLa cells, SB202190 treatments induced p42/44 MAPK activity and caused an increase in LDL receptor expression (Fig.  5). One interesting feature is the rapid and transient kinetics of p42/44 MAPK phosphorylation in nonhepatic HeLa cells as compared with delayed and sustained kinetics of p42/44 MAPK phosphorylation by SB202190 in HepG2 cells (Figs. 4 and 5). The differences in the kinetics and extents of stimulation of p42/ 44 MAPK activity and LDL receptor induction may reflect the underlying differences in signal transduction between these cell types. Notably, despite these differences, PD98059 blocked SB202190-induced LDL receptor expression.
Modulation of Endogenous p42/44 MAPK Activity Does Not Affect p38 MAPK Activation-We then attempted to determine whether inhibition or activation of p42/44 MAPK regulates p38 MAPK phosphorylation. To test this possibility, we generated a HepG2-derived cell line (HepG2-⌬Raf:ER) expressing an estradiol-dependent human Raf-1 protein kinase. In this cell line, the ⌬Raf-1:ER chimera is activated in response to estradiol or anti-estrogen ICI182780, thereby activating MEK-1/2 and then p42/44 MAPK (51,52). The addition of ICI182780 to these cells stimulated p42/44 MAPK within minutes (Fig. 6). The p42/44 MAPK activity increased for up to 1 h and remained elevated in the presence of ICI182780. 3 We used this cell system to directly measure the effect of ICI182780 on p38 MAPK activation by a variety of stress inducers. As shown in Fig. 6, we observed no significant effects on p38 MAPK activation by anisomycin or IL-1␤ in the absence or presence of ICI182780. Likewise, the inhibition of p42/44 MAPK with PD98059 had no effect on activity of other MAPKs (data not shown). This observation is consistent with our earlier demonstration that PD98059 failed to block p38 MAPK activation in response to IL-1␤ or TNF at concentrations that completely blocked p42/ 44 MAPK activation by MEK-1/2 (43). Similar results have also been obtained by Frasch et al. (53) in human neutrophils. Collectively, the above results suggest that the p42/44 MAPK does not regulate p38 MAPK activation by a variety of transcriptional modulators.

SB202190-induced LDL Receptor Expression Is Independent of New Protein Synthesis-
To determine whether the induction of LDL receptor expression by SB202190 may require new protein synthesis, HepG2 cells were incubated with SB202190 in the absence or presence of different concentrations of cycloheximide. Data shown in Fig. 7 demonstrate that inhibition of translation has no effect on LDL receptor expression, suggesting that additional protein synthesis is not required to stimulate LDL receptor expression by SB202190. DISCUSSION Although p38 MAPK and p42/44 MAPK are members of different MAPK subfamilies, we provide evidence for cross-talk between these MAPKs and the role of this communication in controlling LDL receptor expression. The conclusions that SB202190-induced LDL receptor expression is due to inhibition of p38 MAPK and that this pathway exerts its effect on LDL receptor expres-3 C. Golden and K. D. Mehta, manuscript in preparation.

p38 MAPK Negatively Regulates LDL Receptor Expression
sion through p42/44 MAPK activation are based on the following observations. First, under experimental conditions where PD98059 completely inhibited phosphorylation of p42/44 MAPK , this MEK-1/2 inhibitor completely blocked SB202190-induced LDL receptor expression. The selectivity of PD98059 has been described in several studies, and the possibility that SB202190 inhibits another p38 MAPK is unlikely, because this inhibitor is without effect even at higher concentrations (100 M) on either p46/54 JNK or p42/44 MAPK or multiple other related protein kinases, including the homologues p38 MAPK ␥ (stress-activated protein kinase 3) and stress-activated protein kinase 4, which share 60% identity with p38 MAPK ␣and ␤-isoforms. We have also demonstrated that both PD980569 and SB202190 failed to block other MAPKs in HepG2 cells (43). Furthermore, to rule out the possibility that SB202190 may inhibit other unknown targets to fully execute its effect on p42/44 MAPK and LDL receptor expression, we show that expression of MKK6b(E), the constitutive activator of p38 MAPK , significantly suppressed LDL receptor promoter activity (Fig. 3). The effect of MKK6b(E) was dependent on activation of p38 MAPK , since coexpression of MKK6b with p38 MAPK , but not the inactive p38 MAPK (AF) mutant, inhibited LDL receptor expression. Finally, consistent with the role of p42/44 MAPK in SB202190induced LDL receptor expression, inhibition of p38 MAPK resulted in induction of its activity in a time-dependent manner. In fact, induction of p42/44 MAPK phosphorylation and LDL receptor expression with similar kinetics suggests the requirement of an activated p42/44 MAPK to continuously increase expression of the LDL receptor gene. According to Fig. 4, p42/ 44 MAPK phosphorylation started at 4 h, with a maximal FIG. 4. SB202190-induced LDL receptor expression is mediated by the p42/44 MAPK signaling cascade. A, kinetics of p42/44 MAPK and MEK-1/2 activation by SB202190. 2 ϫ 10 5 HepG2 cells were grown and treated as described in Fig. 1. After the indicated times, equal amounts of cell lysates were blotted with anti-phospho-p42/44 MAPK antibody, phosphorylation-independent p42/44 MAPK antibody, anti-phospho-p46/54 JNK antibody, anti-phospho-MEK-1/2, or phosphorylation-independent MAPK phosphatase-1. B, HepG2 cells were grown as described in Fig. 1 and were either untreated or treated with SB202190 (2.5 M) for 8 h either in the absence or presence of indicated concentrations of PD98059 that was added 30 min prior to the SB202190 (2.5 M) addition. Total RNA was subjected to Northern blot analysis, and the filter was hybridized with a 32 P-labeled LDL receptor probe. Values shown are the averages of two different experiments. Ethidium bromide staining of RNA gel before blotting onto nitrocellulose is shown to demonstrate equal loading of RNA in all lanes. Values obtained from the control cells grown in the absence of SB202190 or PD98059 were arbitrarily set at 1. C, cells were grown as described in the legend to Fig. 2. On day 4, medium was changed to either 0.5% LPDS alone or 0.5% LPDS supplemented with three different concentrations of 25-hydroxycholesterol (2, 5, or 10 g/ml) together with cholesterol (10 g/ml). After 2 h, cells were treated with 2.5 M SB202190 for an additional 8 h. Cell extracts were prepared, and equal amounts were subjected to SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-phospho-p42/44 MAPK or phosphorylation-independent anti-p42/44 MAPK (data not shown). Results shown are indicative of three separate experiments.

p38 MAPK Negatively Regulates LDL Receptor Expression
activation observed after 8 h, and remained persistent for up to 24 h following SB202190 treatment. This delay in SB202190induced p42/44 MAPK is most likely due to the late appearance of an immediate or downstream substrate of the p38 MAPK pathway through de novo synthesis, post-translational modification, or localization. Our finding that SB202190-induced LDL receptor induction is not inhibited by cycloheximide (Fig. 7), a protein synthesis inhibitor, rules out de novo synthesis. It is more likely that induction of p42/44 MAPK by SB202190 requires post-translational modification of the already existing substrate(s). Moreover, the lack of effect of sterols on SB202190induced p42/44 MAPK activation suggests that sterols block SB202190-induced LDL receptor expression at a point downstream of p42/44 MAPK . Alternatively, the sterol-sensitive step is part of a parallel signaling pathway.
This study represents the first demonstration of a functional difference between p38 MAPK isoforms in regulating gene expression. Recent evidence indicates the existence of at least four distinct isoforms of p38 MAPK : p38 MAPK ␣, also named CSBP2 or stress-activated protein kinase 2a; p38 MAPK ␤, also named stress-activated protein kinase 2b (functionally inactive in vivo) and its splice isoform ␤II (functionally active in vivo) (26); p38 MAPK ␥, also termed stress-activated protein kinase 3 or extracellular signal-regulated kinase 6; and p38 MAPK ␦, otherwise known as stress-activated protein kinase 4. The four isoforms are similar in size, show about 60 -75% sequence homology, and are all activated by TNF, IL-1, ultraviolet radiation, and hyperosmolar medium. Some isoforms show a pronounced preference in tissue expression and selective interaction with upstream kinases and downstream substrates, pointing to highly specialized functions. These isoforms also differ in susceptibility to inhibition by SB202190; both ␣and ␤II-isoforms of p38 MAPK are inhibited by SB202190, whereas the ␥and ␦-isoforms are insensitive (26). Therefore, it seems

p38 MAPK Negatively Regulates LDL Receptor Expression
less likely that either the ␥or ␦-isoform is responsible for the induction of LDL receptor expression by SB202190. Since hepatic cells contain similar levels of both ␣and ␤II-isoforms, we asked the question of which isoform is responsible for SB202190-induced LDL receptor expression in HepG2 cells. Availability of expression vectors encoding the individual p38 MAPK isoforms allowed us to directly test their involvement by co-transfection studies. We found that, unlike the p38 MAPK ␤II-isoform, expression of the p38 MAPK ␣-isoform dramatically suppressed LDL receptor promoter activity. These studies suggest that SB202190-induced LDL receptor expression is mediated by the inhibition of the ␣-isoform of p38 MAPK . This difference in the role of ␣and ␤-isoforms is consistent with a recent report that showed opposite effects of these isoforms on SB202190-induced apoptosis (54). At this point, it is not clear why the ␣-isoform, and not the ␤-isoform, negatively controls LDL receptor expression via p42/44 MAPK . One possibility is that these two p38 MAPK isoforms may differ in their substrate specificity, and such differences could allow coupling of the p38 MAPK ␣-isoform to the MEK-1/2/p42/44 MAPK signaling cascade. At this point, the mechanism underlying the cross-talk is not clear, but it is safe to predict that cross-talk between p42/44 MAPK and p38 MAPK pathways clearly requires signaling via a protein kinase that lies upstream of MEK-1/2. Because the number of substrates for p38 MAPK that have been characterized in any system is few and those that might contribute to activation of p42/44 MAPK pathways have not been explored, this remains a question for future endeavors.
The suppression of p42/44 MAPK -mediated responses by p38 MAPK activation provides a critical link in the signaling events preceding apoptosis. The ability of stressful signals to stimulate p38 MAPK activity has led to the suggestion that this pathway may function to communicate growth-inhibitory and apoptotic signals within the cell. p38 MAPK is involved in the regulation of apoptosis, since overexpression of kinases that can activate p38 MAPK resulted in the induction of apoptosis (55,56), and inhibition of p38 MAPK activity has been shown to suppress apoptosis (57). However, simple transient activation of the stress kinase cascades is not always sufficient to induce apoptosis. For example, TNF promotes a significant induction of p46/54 JNK and p38 MAPK but does not invariably induce apoptosis through induction of caspases (58). In this regard, it was suggested that concomitant inactivation of survival signals may be a prerequisite for p46/54 JNK and p38 MAPK to induce cell death (59). Interestingly, deprivation of neurotrophic factors in PC-12 cells or ultraviolet irradiation of NIH-3T3 cells not only activates the stress kinase cascades but also leads to a dramatic inhibition of the p42/44 MAPK pathway (55,60). In fact, overexpression of p42/44 MAPK in NIH-3T3 cells impaired a large part of the ultraviolet-induced apoptotic response (60). Furthermore, inhibition of the p42/44 MAPK alone has been shown to induce apoptosis through activation of caspases (45). The negative regulation of p42/44 MAPK by p38 MAPK is consistent with the above results and provides a critical link between the p38 MAPK activation and the concomitant inhibition of p42/ 44 MAPK signaling cascade. It is likely that the induction of apoptosis due to p38 MAPK activation could be partly due to inhibition of p42/44 MAPK . The contradiction regarding the role of p38 MAPK in apoptosis may be due to differences in the degree and extent of cross-talk between the "death" signal mediated by p38 MAPK and the survival signal generated by activation of p42/44 MAPK , and the differential responses may lead to an outcome in a cell-and stimuli-specific manner. Such a mechanism may reconcile the contradictory roles that have been suggested for p38 MAPK in apoptosis.
In conclusion, we have unraveled a one-way inhibitory crosstalk between p42/44 MAPK and p38 MAPK signaling pathways that plays an important role in controlling LDL receptor expression. A signal transduction cascade involving p42/44 MAPK and p38 MAPK has been proposed to account for the SB202190induced LDL receptor induction by a staurosporine-sensitive protein kinase (Fig. 8). This pathway can also account for our earlier observation that shows that p38 MAPK antagonizes IL-1␤-induced LDL receptor expression through p42/44 MAPK (43). The signaling events leading to MEK-1/2/p42/44 MAPK activation in response to p38 MAPK inhibition are still not known and might involve a complex combinatorial, spatial cross-talk of already synthesized messenger molecules. The interplay between these signaling cascades should be an important process, and dynamic balance may be critical for determining the outcome of a wide array of biological processes. Elucidation of the signaling components involved in this communication will significantly advance our ability to design novel strategies for the treatment of hypercholesterolemia and for understanding other important pathophysiological processes.