Transcriptional Induction of Mitogen-activated Protein Kinase Phosphatase 1 by Retinoids

All-trans-retinoic acid (t-RA) inhibits hydrogen peroxide (H2O2)-induced apoptosis by inhibiting the c-Jun N-terminal kinase (JNK)-activator protein 1 (AP-1) pathway. In this report, we examined the involvement of mitogen-activated protein kinase phosphatase 1 (MKP-1) in suppression of JNK and the antiapoptotic effect of t-RA and the roles of nuclear receptors in the regulation of MKP-1 by t-RA. We found that not only t-RA, but also a selective agonist of retinoic acid receptor (RAR), a selective agonist of retinoid X receptor (RXR), and a pan-agonist of RAR and RXR all induced MKP-1 at the transcriptional level. Activation of RAR was required for all of these triggering effects, but activation of RXR was required only for the RXR agonist-induced MKP-1 expression. Among the three RAR subtypes, RARα and RARγ, but not RARβ, mediated the t-RA-induced MKP-1 expression. The antiapoptotic effect of t-RA on H2O2-induced apoptosis in several cell types was correlated with the inducibility of MKP-1 by t-RA. Inhibition of MKP-1 by vanadate enhanced JNK phosphorylation and attenuated the antiapoptotic effect of t-RA. Furthermore, overexpression of MKP-1 inhibited H2O2-induced JNK phosphorylation and apoptosis. To our knowledge, this is the first to demonstrate that 1) MKP-1 is inducible by retinoids at the transcriptional level, 2) RXR and individual RAR subtypes have different roles in this process, and 3) the induced MKP-1 plays a significant role in mediating both JNK inhibition and the antiapoptotic effect of t-RA in oxidative stress.

Retinoids, the biologically active derivatives of vitamin A (retinol), have profound effects on embryogenesis, neoplasia, and maintenance of normal tissues (1). The crucial roles of retinoids in controlling cell function have been extensively investigated especially using retinoic acid (RA). 1 For example, RA is known to be essential for the development of kidneys (2) and is also effective in the treatment of glomerular diseases (3). The wide spectrum of physiological and pharmacological effects of RA is attributed to both receptor-dependent (1) and receptorindependent mechanisms (4 -6). Transcriptional regulation of target genes by RA is generally mediated by nuclear receptors: retinoic acid receptors (RAR␣, -␤, and -␥) and retinoid X receptors (RXR␣, -␤, and -␥) (7). These receptors have different ligand specificity; e.g. RARs are activated by both all-transretinoic acid (t-RA) and 9-cis-RA, whereas RXRs are activated only by 9-cis-RA (7). After ligand binding, these receptors form homodimers or heterodimers and function as transcriptional regulators. For example, t-RA binds to RARs and activates RAR-RXR heterodimers, and the complex exerts its biological effects via binding to a particular cis element, retinoic acid response element (RARE) (8).
We previously reported that t-RA inhibits hydrogen peroxide (H 2 O 2 )-induced apoptosis of renal mesangial cells and suggested that t-RA might have a therapeutic effect on glomerulonephritis by preventing oxidant stress-induced mesangial cell injury (9 -11). Mesangial cells exposed to H 2 O 2 undergo apoptosis via activator protein 1 (AP-1)-dependent pathways (e.g. the c-Jun N-terminal kinase (JNK)-c-Jun/AP-1 pathway and the extracellular signal-regulated kinase (ERK)-c-Fos/AP-1 pathway) (12). t-RA inhibits H 2 O 2 -induced apoptosis by suppression of the AP-1 pathway, at least in part, by inhibition of c-fos/c-jun expression and inactivation of JNK (10). However, it is still unclear how t-RA affects the JNK pathway and whether or not t-RA modulates activation of other major mitogen-activated protein (MAP) kinases such as ERK and p38 MAP kinase.
MAP kinases, including ERK, JNK, and p38 MAP kinase, are activated by upstream kinases including MEK1 and -2, and MKK3, -4, and -6 (13). However, once activated, MAP kinases are rapidly inactivated by the family of protein phosphatases. In particular, dual specificity protein phosphatases play crucial roles in the dephosphorylation and inactivation of MAP kinases (14). MAP kinase phosphatase 1 (MKP-1), also termed CL100, 3CH134, and ERP, is a prototypic member of the family of inducible dual specificity phosphatases (15). It selectively dephosphorylates tyrosine and threonine residues on MAP kinases and inactivates them. Although all three MAP kinases are potential targets of MKP-1 (16 -18), it has been reported that JNK and p38 MAP kinase were preferentially inactivated by MKP-1 (16).
Recently, Lee et al. (19) reported that serum-induced phosphorylation of JNK was inhibited by t-RA in tumor cells. This was associated with a post-translational increase in the level of MKP-1 protein (19). We hypothesized that induction of MKP-1 might be involved in the antiapoptotic effect of t-RA on H 2 O 2exposed mesangial cells, especially via inhibition of JNK phosphorylation. The aim of this investigation was to test this hypothesis. Our results show that MKP-1 was induced by t-RA, unexpectedly, at the transcriptional level. RXR and individual RAR subtypes had different roles in this process (i.e. activation of RAR, but not RXR, was required for the triggering effect; among the three RAR subtypes, RAR␣ and RAR␥, but not RAR␤, mediated the t-RA-induced MKP-1 expression). Furthermore, we show that the induced MKP-1 plays a significant role in mediating both JNK inhibition and the antiapoptotic effect of t-RA.

MATERIALS AND METHODS
Cells-Mesangial cells (SM43) were established from isolated glomeruli of a male Sprague-Dawley rat and identified as being of the mesangial cell phenotype as described previously (20). The rat fibroblast cell line NRK49F, the canine epithelial cell line MDCK, and the human endothelial cell line ECV304 were purchased from the American Type Culture Collection (Manassas, VA). All cells were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 (Invitrogen) supplemented with 100 units/ml of penicillin G, 100 g/ml of streptomycin, 0.25 g/ml of amphotericin B, and 10% fetal calf serum (FCS). Medium containing 1% FCS was generally used for experiments.
Pharmacological Manipulation-Confluent cells were preincubated in 1% FCS for 24 h, treated with t-RA (0.01 nM to 10 M; Sigma) or 9-cis-RA (1 nM to 10 M; Sigma) for 0.5-6 h, and subjected to Northern blot analysis. Incubation with t-RA for 1 h was generally used for induction of MKP-1 expression. In some experiments, cells were costimulated with t-RA (0.5 -5 M) and H 2 O 2 (150 M). To examine roles of RAR and RXR in the regulation of MKP-1 by retinoids, the following agonists and antagonists were used: RAR agonist TTNPB (0.  (31,32), and RAR␥ antagonist MM11253 (also known as SR11253; 0.1 M) (33,34). Cells were pretreated with or without 10 -50 times higher concentrations of antagonists for 10 min and then stimulated with agonists for 1 h.
Western Blot Analysis-To examine effects of t-RA on the basal and H 2 O 2 -induced activation of ERK, JNK, and p38 MAP kinase, confluent mesangial cells were incubated in 1% FCS for 24 h, pretreated with t-RA for 1 h and exposed to 100 -200 M H 2 O 2 for 15 min to 1 h. Phosphorylated forms of ERKs and p38 MAP kinase were detected by Western blot analysis as described before (10,12). Analyses were per-formed using the PhosphoPlus MAP kinase antibody kit and Phospho-Plus p38 MAP kinase antibody kit (New England Biolabs, Hert, UK) following protocols provided by the manufacturer. Activity of JNK was evaluated by phosphorylation of c-Jun using the stress-activated protein kinase/JNK assay kit (New England Biolabs) or by detecting phosphorylation of JNK per se using PhosphoPlus stress-activated protein kinase/JNK (Thr 183 /Tyr 185 ) antibody kit (New England Biolabs), as described previously (9,10,12).
To examine the effect of t-RA on the protein level of MKP-1, confluent cells were cultured in 0.5-1% FCS for 24 -48 h and stimulated with 5 M t-RA for up to 24 h. Western blot analysis was performed using MKP-1 polyclonal antibody (V-15, sc-1199; Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Assessment of mRNA Stability-Effects of t-RA on the stability of MKP-1 mRNA were assessed using the RNA synthesis inhibitor actinomycin D (39). In brief, mesangial cells were treated with or without t-RA (1 M) for 1 h in the presence of actinomycin D (5 g/ml; Serva, Heidelberg, Germany) for the last 0 -60 min. Northern blot analysis was performed to examine the level of MKP-1 mRNA, as described above.
Assessment of Apoptosis-Cells were pretreated with or without t-RA (5 M) for 1-2 h and stimulated by H 2 O 2 (200 -500 M) for 16 -24 h. To examine roles of phosphatases in the antiapoptotic effect of t-RA, cells were pretreated with or without the protein-tyrosine phosphatase inhibitor sodium orthovanadate (vanadate; 100 M, Sigma) for 1 h, treated with t-RA for 1 h, and stimulated by H 2 O 2 for 6 -12 h. Apoptosis was assessed quantitatively, as described before (10,40). In brief, cells were fixed with 4% formaldehyde for 10 min, stained by Hoechst 33258 (10 g/ml; Sigma) for 1 h, and subjected to fluorescence microscopy. Apoptosis was identified using morphological criteria (i.e. nuclear condensation and/or fragmentation). Both attached cells and detached cells were used for evaluation.
To further confirm the role of MKP-1 in apoptosis, another set of experiments were performed using pSG5-MKP1 encoding a wild-type MKP-1 and pSG5-MKP-1CS encoding a catalytically inactive MKP-1 (18).
Statistical Analysis-Data were expressed as means Ϯ S.E. Statistical analysis was performed using the nonparametric Mann-Whitney U test to compare data in different groups. p value of Ͻ0.05 was used to indicate a statistically significant difference.

Regulation of MAP Kinases by t-RA-We previously reported
that H 2 O 2 -induced apoptosis in mesangial cells was mediated by JNK and ERK but not p38 MAP kinase (10,12). We also found that this apoptotic process was suppressed by t-RA via suppression of JNK (10). Based on these previous data, we first examined whether or not t-RA affects other MAP kinases including ERK and p38 MAP kinases. Mesangial cells were pretreated with or without 5 M t-RA, stimulated by H 2 O 2 for up to 30 min, and subjected to kinase assays. As we have previously shown (10), activation of JNK by H 2 O 2 was markedly suppressed by the treatment with t-RA (Fig. 1A). Similarly, activation of p38 MAP kinase by H 2 O 2 was also attenuated by t-RA (Fig. 1B). In contrast, t-RA triggered basal activity of ERK and enhanced H 2 O 2 -induced ERK activation (Fig. 1C).
Induction of MKP-1 by t-RA-MKP-1 is a prototypic member of the family of inducible dual specificity phosphatases (15) and preferentially inactivates JNK and p38 MAP kinase (16,43). A previous report showed that, in some cancer cells, MKP-1 protein was increased by t-RA through a post-translational mechanism (19). To examine whether the suppression of MAP kinases by t-RA is mediated by MKP-1, we tested the level of MKP-1 protein in t-RA-exposed mesangial cells. Mesangial cells were treated with t-RA for up to 24 h, and Western blot analysis was performed. As shown in Fig. 2A, t-RA transiently increased the level of MKP-1 protein with a peak at 3-6 h. This induction was closely correlated with the preceding, transient increase of MKP-1 mRNA (i.e. the level of MKP-1 mRNA was significantly increased within 30 min, peaked to maximum at 1 h, and returned to the prestimulation level after 6 h (Fig. 2B)). These data suggested that the induction of MKP-1 by t-RA occurs at the transcriptional level in mesangial cells. We next examined the dose-dependent effect of t-RA on the level of MKP-1 mRNA. As shown in Fig. 3A, the stimulatory effect of t-RA was observed with a wide range of concentrations at 1 nM to 10 M. The maximum effect was observed at concentrations between 100 nM and 1 M.
MKP-1 mRNA is known to be induced by H 2 O 2 (15). The stimulatory effect of t-RA on MKP-1 was further tested in the presence of H 2 O 2 . Cells were treated with 0 -5 M t-RA for 2 h and stimulated by H 2 O 2 for an additional 2 h. Northern analysis showed that MKP-1 was modestly induced by H 2 O 2 , and the induction was further enhanced by the treatment with t-RA (Fig. 3B).
The up-regulation of the MKP-1 mRNA level by t-RA may be caused by transcriptional induction or increased stability of the transcript. To test the latter, the effect of t-RA on the stability of MKP-1 mRNA was examined using the RNA synthesis inhibitor actinomycin D. Mesangial cells were treated with or without t-RA for 1 h in the presence of actinomycin D for the last 0 -60 min. Northern blot analysis showed that MKP-1 mRNA was rapidly degraded in the presence of actinomycin D, and treatment with t-RA did not affect the half-life of MKP-1 mRNA (Fig. 4A). The half-lives calculated were as follows: 34 min in untreated cells and 36 min in t-RA-treated cells (Fig.  4B). Induction of MKP-1 mRNA by t-RA was not mediated by de novo synthesis of proteins, because blockade of translation using cycloheximide did not prevent but rather enhanced t-RAinduced MKP-1 expression (Fig. 4C). The increase in the level of MKP-1 mRNA by cycloheximide is similar to induction of other immediate early genes, c-fos and c-myc (44), by this agent.
Different Contribution of RAR and RXR-We found that rat mesangial cells constitutively expressed all RAR and RXR subtypes. Expression of RAR␥ and RXR␣ was found to be most abundant, and the level of RAR␣ was moderate (Fig. 5A). RAR␤, RXR␤, and RXR␥ (data not shown) were expressed only weakly. The expression level of RXR␣ was not significantly affected by t-RA. In contrast, expression of RAR␤ and RAR␥ was substantially induced by t-RA with a peak at 3 and 6 h, respectively (Fig. 5B). Expression of RAR␣ was also slightly induced within 6 h after the stimulation with t-RA.
The roles of RAR and RXR in the regulation of MKP-1 by t-RA were examined using RAR agonist TTNPB and RXR agonist AGN194204. As shown in Fig. 6, A and B, both agonists significantly increased the expression of MKP-1 dose-dependently. The minimum effective dosages were 1 and 10 nM, respectively. TTNPB binds to RAR selectively at concentrations up to 10 M (24), and AGN194204 selectively binds to RXR at concentrations up to 30 M (25). Based on these, activation of either RAR or RXR can induce MKP-1 expression.
The roles of RAR and RXR in the t-RA-induced expression of MKP-1 were further tested using RAR antagonist AGN193109 and RXR antagonist HX531. Cells were pretreated with or without these antagonists for 10 min and then stimulated by t-RA, 9-cis-RA, and AGN194204, respectively. Northern blot analysis showed that the induction of MKP-1 by t-RA, 9-cis-RA, and AGN194204 was abrogated by the treatment with RAR antagonist AGN193109. In contrast, RXR antagonist HX531 did not affect the MKP-1 induction by t-RA and 9-cis-RA but completely prevented AGN194204-induced MKP-1 expression (Fig. 6, C-E). The effect of AGN193109 observed here is due to specific suppression of RAR, because both basal and H 2 O 2 -/ serum-induced expression of MKP-1 mRNA was not obviously affected by AGN193109 (Fig. 6F).
Contrastive Roles of Three RAR Subtypes-Since RAR is required for the retinoid-induced MKP-1 expression, we further examined roles of RAR subtypes in the regulation of MKP-1. Am580 and CD666 are known to be specific RAR␣ and RAR␥ agonists at concentrations of Յ1 nM (45,46) and Յ10 nM (26), respectively. As shown in Fig. 7, A and C, both Am580 and CD666 induced expression of MKP-1 dose-dependently. Of note, the induction of MKP-1 by these agents was observed at low concentrations that allow the specificity of Am580 to RAR␣ and CD666 to RAR␥. In contrast, RAR␤-selective agonist CD2314 did not induce MKP-1 mRNA even at high concentrations (100 -1000 nM) (Fig. 7B). These data suggested the significant roles of RAR␣ and RAR␥, but not RAR␤, in mediating retinoid-induced MKP-1 expression.
Correlation between Induction of MKP-1 and Antiapoptotic Effect of t-RA-To examine whether the stimulatory effect of t-RA on MKP-1 expression is a general phenomenon in mammalian cells, rat fibroblast cell line NRK49F, canine epithelial cell line MDCK, and human endothelial cell line ECV304 were tested. These cells were treated with or without t-RA, and expression of MKP-1 was examined by Northern blot analysis. As shown in Fig. 9A, dramatic induction of MKP-1 was observed in NRK49F cells. In contrast, induction of MKP-1 was not evident in MDCK cells and ECV304 cells.
Correlation between the induction of MKP-1 by t-RA and the antiapoptotic effect of t-RA was examined using these cells exposed to H 2 O 2 . NRK49F cells, MDCK cells, and ECV304 cells were pretreated with t-RA and stimulated by H 2 O 2 . Hoechst staining showed that, like mesangial cells, pretreatment with t-RA substantially inhibited H 2 O 2 -induced apoptosis in NRK49F cells (Fig. 9B, right). The percentage of apoptotic cells was significantly reduced from 35.0 Ϯ 5.9% (H 2 O 2 alone) to 15.0 Ϯ 1.7% (t-RA ϩ H 2 O 2 ) (means Ϯ S.E., p Ͻ 0.05). In contrast, t-RA did not attenuate H 2 O 2 -induced apoptosis in MDCK cells and ECV304 cells (Fig. 9B, left and middle), in which MKP-1 was not obviously induced by t-RA (Fig. 9A) Hoechst staining revealed that the antiapoptotic effect of t-RA was significantly attenuated by the treatment with vanadate (Fig. 10A). Quantitative analysis showed that the percentage of apoptotic cells was dramatically reduced from 43.8 Ϯ 4.8 to 7.3 Ϯ 0.6% (p Ͻ 0.05) by the treatment with t-RA, and pretreatment with vanadate significantly increased the percentage of apoptosis from 7.3 Ϯ 0.6% (vanadate (Ϫ)) to 27.3 Ϯ 5.2% (vanadate (ϩ)) (Fig. 10B). Furthermore, the effect of vanadate was closely correlated with the enhanced phosphorylation of JNK in H 2 O 2 -stimulated cells (Fig. 10C). Treatment with 10 M vanadate alone for 8 h did not induce any apoptotic changes of mesangial cells (data not shown). Similarly, vanadate alone did not induce JNK phosphorylation (Fig. 10C).  (Fig. 11A).

Inhibition of H 2 O 2 -induced JNK Phosphorylation and Apoptosis by Overexpression of MKP-1-To
The effectiveness of MKP-1 in inhibiting H 2 O 2 -induced apoptosis was further confirmed using different MKP-1 expression plasmids (Fig. 11B). Consistent with the results described above, transfection with MKP-1 significantly inhibited H 2 O 2induced apoptosis. This suppressive effect was not observed when the cells were transfected with a catalytically inactive mutant of MKP-1, MKP-1CS (Fig. 11B).
To further confirm that t-RA-induced expression of MKP-1 inhibits H 2 O 2 -induced apoptosis by inhibiting JNK phosphorylation, stable transfectants of mesangial cells that conditionally express wild-type MKP-1 were created by transfection of the cells with pMEP4-MKP1 that introduces the MKP-1 gene under the control of the human metallothionein IIa promoter. After the stimulation with 5 M CdSO 4 , the established MKP-1/SM (1) cells and MKP-1/SM (14) cells, but not Control/SM cells, expressed exogenous MKP-1 with a peak after 6 h (Fig.  12A). Using these established clones, activation of JNK in response to H 2 O 2 was examined. Control/SM cells and MKP-1/SM cells were pretreated with CdSO 4 to induce MKP-1 expression. Then the cells were exposed to H 2 O 2 and subjected to the kinase assay. As shown in Fig. 12B, obvious induction of JNK phosphorylation was observed in Control/SM cells stimulated by H 2 O 2 . In contrast, the JNK phosphorylation by H 2 O 2 was abrogated in both MKP-1/SM cells that overexpress exogenous MKP-1. DISCUSSION t-RA inhibits H 2 O 2 -induced apoptosis of mesangial cells via intervention in the JNK-AP-1 pathway (9,10). This antiapoptotic effect of t-RA was attenuated by the treatment with cycloheximide, 2 suggesting that de novo synthesis of some protein is required. In this report, we describe a role of MKP-1 that is induced by t-RA and participates in the antiapoptotic action of t-RA in mesangial cells. Our results showed that MKP-1 was rapidly and dose-dependently induced by t-RA at the transcriptional level, leading to accumulation of intracellular MKP-1 protein. Very low concentrations (Ͻ1 nM) were found to be effective, indicating that induction of MKP-1 may occur under physiologic situations. Similar stimulatory effects were observed with other retinoids including 9-cis-RA and other cell types including NRK49F fibroblasts. Furthermore, t-RA-induced expression of MKP-1 was also observed in H 2 O 2 -stimulated cells. These data suggested that MKP-1 is an RA-responsive gene and may contribute to the RA-induced suppression of MAP kinases that are essential for H 2 O 2induced apoptosis.
The induction of MKP-1 by t-RA is somewhat controversial. A recent report showed a lack of increase in the level of MKP-1 in t-RA-treated, rat aortic smooth muscle cells (47). In contrast, Lee et al. (19) reported that MKP-1 protein was increased by t-RA using a human lung cancer cell line. They demonstrated that the protein level of MKP-1 was increased by the treatment with t-RA but its mRNA level was not affected. They concluded that the increase in MKP-1 protein was ascribed to its increased stability. In contrast to these previous reports, our present data using rat mesangial cells showed that MKP-1 was up-regulated by t-RA at the transcriptional level. Currently, the reason for this discrepancy from report to report is unknown. Regulation of MKP-1 by t-RA may be different depending on cell types; e.g. cell type-specific expression of RAR subtypes (48) or co-regulators (49) may cause different responses to t-RA.
A previous report showed that RA may affect mRNA levels of some genes through post-transcriptional mechanisms (50). However, it is not the case in the up-regulation of MKP-1 mRNA by t-RA. As shown in this report, stability of MKP-1 mRNA was not altered in mesangial cells treated with t-RA. The mechanisms involved in the transcriptional induction of MKP-1 by t-RA are currently unclear. Our data showed that RAR was essential in this process. Because the induction of MKP-1 mRNA by t-RA was rapid (within 30 min; Fig. 2B) and no protein synthesis was required (Fig. 4C), activated RAR/ RXR may directly induce MKP-1 expression via binding to RARE of the MKP-1 gene. Previous reports analyzed the promoter/enhancer regions of human, mouse, and rat MKP-1 genes, but RARE has not been reported (51)(52)(53). One possible explanation could be that the promoter/enhancer region of the rat MKP-1 gene might exclusively contain RARE, since the promoter/enhancer sequence of the rat MKP-1 gene has been analyzed only partially (52). The fact that MKP-1 mRNA was induced by t-RA in rat mesangial cells and rat NRK49F cells but not in human ECV304 cells and canine MDCK cells may support this possibility (Fig. 9). Another possible explanation is that t-RA-bound RAR/RXR may induce MKP-1 expression via binding to a RARE-like sequence that has not been identified. In addition, we cannot exclude the possibility that the induction of MKP-1 by t-RA is mediated by RARE-independent 2 Q. Xu, unpublished data. mechanisms (e.g. by removing elongation block on the MKP-1 gene (52)). Further investigation will be necessary to clarify these issues.
In general, transcriptional induction of RA-responsive genes is mediated by RAR and RXR. To investigate roles of RAR and RXR in t-RA-induced MKP-1 expression, we first examined expression of these nuclear receptors in rat mesangial cells. Our data showed that mesangial cells constitutively expressed all RAR and RXR subtypes. Expression of RAR␤ and RAR␥ was substantially induced by the treatment with t-RA. These findings are consistent with previous reports that showed the inducible property of RAR␤ and RAR␥ in rat tissues (54,55). Experiments using receptor agonists revealed that activation of either RAR or RXR was sufficient to induce MKP-1 expression in our system. RAR antagonist AGN193109 markedly suppressed induction of MKP-1 by RAR agonist t-RA, RXR agonist AGN194204, or RAR/RXR pan-agonist 9-cis-RA. This result indicates crucial requirement of RAR in retinoid-induced MKP-1 expression.
Although previous gene knockout studies did not show any functional difference among RAR subtypes in development (7), a few reports demonstrated different roles of RAR subtypes in regulating gene expression (45,56). We therefore examined roles of individual RAR subtypes in the induction of MKP-1 by t-RA. Experiments using selective receptor agonists and antagonists revealed involvement of RAR␣ and RAR␥, but not RAR␤, in mediating t-RA-induced MKP-1 expression. This finding was further confirmed using the cells transfected with each receptor. The distinct role of RAR␤ is consistent with a previous report showing that RAR␤ has different function from RAR␣ and RAR␥ in mediating some gene expressions in F9 embryonic carcinoma cells (45).
Previous reports showed the importance of MKP-1 in the regulation of apoptosis in various cells. For example, expression of MKP-1 in cancer cells is correlated with their resistance to apoptosis (57,58). Franklin et al. (59) showed that exogenous, conditional expression of MKP-1 conferred resistance of leukemia cells against UV-induced apoptosis. Induction of endogenous MKP-1 also played an important role in the cytoprotection by insulin (60). Although there is some controversy (61), induction of MKP-1 has been generally regarded as cytoprotective in mammalian cells. In the present report, we provided additional evidence for the cytoprotective role of MKP-1. It was based on the following findings. 1) In mesangial cells and NRK49F fibroblasts, MKP-1 was induced by t-RA at the transcriptional level. 2) The induction of MKP-1 was associated with inhibition of H 2 O 2 -induced apoptosis by t-RA in these cells. 3) Lack of MKP-1 induction in MDCK cells and ECV304 cells was associated with the lack of cytoprotection by t-RA. 4) The protein-tyrosine phosphatase inhibitor vanadate significantly attenuated the antiapoptotic effect of t-RA. 5) Transfection with MKP-1 abrogated the H 2 O 2 -induced apoptosis of mesangial cells. These data suggest an important role of MKP-1 in the regulation of cell survival in mesangial cells.
In summary, our data shed light on the mechanism involved in the antiapoptotic action of RA against oxidative stressinduced apoptosis. To our knowledge, this is the first to demonstrate that 1) MKP-1 is inducible by retinoids at the transcriptional level, 2) RXR and individual RAR subtypes have different roles in this process, and 3) the induced MKP-1 plays a significant role in mediating both JNK inhibition and the antiapoptotic effect of t-RA in oxidative stress.