Retinoic acid-mediated growth arrest requires ubiquitylation and degradation of the F-box protein Skp2.

The mechanism by which all-trans retinoic acid (ATRA) leads to a G(1) arrest of the cell cycle remains unclear. We show here that the decrease in D-type cyclin levels observed following ATRA treatment correlates with an increase in the rate of cyclin D1 ubiquitylation in both T-47D and MCF-7 breast cancer cell lines. However, MCF-7 cells are more resistant to ATRA than T-47D cells indicating that cyclin D1 degradation is not sufficient for ATRA-mediated arrest. We found a striking difference between these cells in that while ATRA induces an elevation in the cdk inhibitor p27 in T-47D cells, this is not observed in the ATRA-resistant MCF-7 cells. Furthermore, we demonstrate that ATRA promotes the ubiquitylation of Skp2, an F-box protein that targets p27 for degradation. Moreover, overexpression of Skp2 in T-47D cells prevents accumulation of p27 and promotes resistance to ATRA. In addition, overexpression of cyclin D1 in T-47D cells also promotes ATRA resistance. We found that the mechanism of ATRA-induced ubiquitylation of cyclin D1 and Skp2 is independent of CUL-1 expression and that ATRA can rescue cyclin D1 degradation in the uterine cell line SK-UT-1, where D-type cyclins are stabilized due to a specific defect in proteolysis. These data suggest that ATRA induces a novel pathway of ubiquitylation and that the degradation of the F-box protein Skp2 is the mechanism underlying p27 accumulation and cyclin E-cdk2 inactivation following ATRA treatment.

It has been known for more than 50 years that retinoids are potent agents for the control of cellular differentiation and proliferation. Several studies have shown that retinoids can suppress the process of tumorigenesis both in vitro and in vivo (1)(2)(3) and that cells exposed to all-trans retinoic acid (ATRA) 1 arrest in the G 1 phase of the cell cycle. However, the molecular mechanism underlying this arrest remains unclear.
Cyclin-dependent kinases (cdk) are key cell cycle regulators. Their activities are regulated at several levels including phosphorylation, binding to their regulatory subunits, the cyclins, and binding to small inhibitory proteins called the cyclin-dependent kinase inhibitors (CKI). Among the cyclin subunits, D-type cyclins associate with CDK 4 and 6 to phosphorylate the retinoblastoma (Rb) protein. Hyperphosphorylation of Rb promotes the release of the E2F family of transcription factors that then promote entry into S phase through activation of key target genes. Cyclin E-CDK2 complexes act downstream of cyclin D-CDK4/6 to maintain Rb phosphorylation and events leading to the premature activation of cyclin E-CDK2 complex trigger inappropriate entry into S phase. However, the activity of cyclin E-CDK2 complexes are restricted by binding to the CDK inhibitors p27 and p21. The importance of these proteins is best illustrated by the fact that mutations leading to either accumulation of cyclin D1, or elevated cyclin E-associated kinase activity due to loss of p27 expression, are both frequent events contributing to tumorigenesis (4 -7). Therefore, agents able to promote cyclin D1 destruction and/or p27 stabilization are of major therapeutic interest. D-type cyclins and p27 are degraded by the ubiquitin-dependent pathway of proteolysis (8 -11). The assembly of ubiquitin molecules onto specific proteins to form multiubiquitylated chains results in recognition and subsequent degradation by the 26 S proteasome. Linkage of ubiquitin to a protein requires the activity of the ubiquitin activating enzyme or E1, an ubiquitin conjugating enzyme or E2 and an ubiquitin ligase or E3 (12,13). The SCF complex provides the E3 ligase activity in the ubiquitylation of several proteins involved in cell cycle progression at the G 1 /S transition and is composed of four proteins: Skp1, a Cullin, Rbx 1, and an F-box protein (for a review, see Ref. 14). Yeast and mammalian cells encode several different cullins and F-box proteins that offer the potential for differential assembly of multiple SCF complexes where the F-box protein is thought to provide substrate specificity to the ligase complex (15)(16)(17). In mammals, strong experimental evidence has recently been obtained that demonstrates the involvement of the SCF SKP2 complex in the ubiquitylation of p27 (18 -20).
We and others have previously shown that human cyclin D1 and cyclin D3 can associate with the cullin CUL-1 (11) and that expression of antisense against the F-box protein Skp2 promotes the accumulation of cyclin D1 (21). In addition, Skp2 Ϫ/Ϫ mouse embryonic fibroblasts show a modest elevation in cyclin D1 levels (22). Collectively, these data indicate that the SCF SKP2 complex affects the ubiquitylation of D-type cyclins, but unlike p27, whether this is due to a direct effect on ubiquitylation or indirect effect due to p27 elevation has not been demonstrated.
Recently, ATRA has been shown to accelerate the ubiquitylation of cyclin D1 (23), suggesting that the elimination of cyclin D1 may contribute to the G 1 arrest observed following ATRA treatment. We show here that although ATRA increases the degradation of cyclin D1 in MCF-7 and T-47D, two independently derived breast cancer cell lines, this effect is not sufficient to result in cell cycle arrest following ATRA treatment. We found that despite the degradation of cyclin D1 in both T-47D and MCF-7 cells, these cells show marked difference in their sensitivity to ATRA. Furthermore, we show that in T-47D cells but not in MCF-7 cells, ATRA induces cell cycle arrest and an elevation in p27 protein levels. We also show that the F-box protein targeting p27 for ubiquitylation, Skp2, is itself ubiquitylated, that ATRA stimulates its ubiquitylation and that Skp2 overexpression leads to resistance to ATRA. Furthermore, we present data suggesting that ATRA-inducible degradation is independent of an SCF complex containing CUL-1. Our results indicate that the stabilization of p27 results from the elimination of the F-box protein Skp2 by an ATRA-inducible ubiquitylation pathway. These molecular markers of ATRA-induced cell cycle arrest may prove invaluable in predicting a therapeutic response to retinoids.

MATERIALS AND METHODS
Cell Culture, Transfection, and Reagents-MCF-7, T-47D, T-47D ⌬MT-1, and T-47D D1-3 cells were cultured in RPMI medium with Hepes buffer and 10% fetal cals serum (CSL Biosciences) at 37°C in 5% CO 2 . T-47D ⌬MTcycD1-3 cells represent a clone of T-47D cells that stably express cyclin D1 under the control of the zinc-inducible metallothionein promoter. T-47D ⌬MT-1 cells are a clone of T-47D cells stably transfected with the control plasmid and do not express cyclin D1.
ATRA (Sigma) was diluted to 100 mM in Me 2 SO and then diluted to 2 mM in ethanol. Untreated control incubations (CTL) were performed by adding equal dilutions of ethanol. For induction of cyclin D1 from the inducible promoter, cells were treated with 50 M zinc sulfate. Transfections of SKP2 and Myc-ubiquitin were performed using Fugene reagent as described by the manufacturer.
For immunoprecipitation, cells were lysed in lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% Nonidet P-40, 0.1 mM sodium fluoride, 0.1 mM sodium orthovanadate) and the lysates clarified by centrifugation at 13,000 rpm for 20 min at 4°C. The supernatant was transferred to a new tube and affinity purified primary antibodies were added to lysates at a concentration of 0.8 g/l for anti-cyclin D1 and 0.4 g/l for Skp2. After incubation on ice for 60 min with the appropriate antibodies, 50 l of protein A-Sepharose was added and the incubation was continued for another 60 min. The peptide aldehyde leucinyl-leucinyl-norleucinyl (LLnL) was solubilized in Me 2 SO at a concentration of 20 mM. Cells were treated for 14 h with LLnL diluted to 100 M or with Me 2 SO alone. For detection of ubiquitinated Skp2, T-47D cells co-transfected with pCDNA3-ubiquitin MYC and pCDNA3-SKP2 were treated for 4 h with LLnL before protein extraction. 30 g of protein was diluted in SDS-PAGE sample buffer, heated for 5 min at 95°C, loaded on a 12% polyacrylamide gel, transferred onto a nitrocellulose membrane, subjected to Western analysis, and developed by ECL (Amersham Pharmacia Biotech).
Pulse-Chase Labeling-Cells were washed twice in phosphate-buffered saline at 37°C and then incubated for 30 min in methionine and cysteine-free medium. Cells were labeled for 30 min with 200 Ci/ml Trans 35 S (PerkinElmer Life Science), washed three times with phosphate-buffered saline, and refed with Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 5 M ATRA. At various times after 35 S-pulse labeling, aliquots were taken and cells lysed as described above for immunoprecipitations.
RNA Extraction and RT-PCR-RNA was extracted from cells using Trizol reagent (Life Technologies, Inc.) followed by gentle sonication (2-3 pulses) on ice. Following this homogenization, RNA extraction was carried out as described by the manufacturer (Life Technologies, Inc.). 0.4 g of total RNA was used per RT-PCR reaction using the Titan one tube RT-PCR system (Roche Molecular Biochemicals).
Ubiquitylation Assays-Crude cellular lysates were prepared from cell pellets washed twice in ice-cold phosphate-buffered saline, resus-pended in 1-2 volumes of 10 mM Tris, pH 8.0, incubated on ice for 10 min, and sonicated briefly on ice. Cellular debris was removed by centrifugation at 14,000 rpm for 20 min at 4°C and protein concentration determined. Substrates for the ubiquitylation assay were prepared using Promega rabbit reticulocyte lysate TNT kit and either pcDNA3-cyclin D1-HA or pcDNA3-SKP2 plasmids.
Equal amounts (1-10 l) of 35 S-substrate was added to 200 g of total crude lysate and the volume adjusted to 125 l, respectively, with reaction buffer (10 mM Tris, pH 7.5, 5 mM CaCl 2 , 5 mM MgCl 2 ). Reactions were incubated at 30°C and 20-l aliquots were taken following the indicated incubation period. Reactions were stopped by the addition of 5 l of 5 ϫ SDS sample buffer, proteins separated by SDS-PAGE, the gel fixed in 20% methanol, 10% acetic acid for 30 min, dried, and exposed to Bio-Max film (Kodak) overnight at Ϫ70°C.

T-47D and MCF-7 Cells Show Distinct Sensitivity to ATRA
Treatment-ATRA has been reported to lead to a decrease in D-type cyclin levels (24) and to promote cyclin D1 ubiquitylation (23,25). In agreement with these observations, we also found a dose-dependent decrease in the level of cyclin D1 in T-47D and MCF-7 cells after treatment with ATRA (Fig. 1A). Using an in vitro ubiquitylation assay recently described (26), we tested whether ATRA increased the ubiquitylation rate of cyclin D1 in these breast cancer cell lines. T-47D cells were treated in the presence or absence of 5 M ATRA for 3 days and proteins extracted. The resulting extracts were incubated in the presence of 1 l of in vitro translated 35  presence of ubiquitylated 35 S-cyclin D1 detected by autoradiography. Accumulation of high molecular weight laddering corresponding to ubiquitylated species of 35 S-cyclin D1 was observed following incubation with the T-47D extract in the absence of ATRA (Fig. 1B). However, the intensity of the laddering was increased in the presence of T-47D extract treated with ATRA (Fig. 1B). Similar results were observed in MCF-7 cells (data not shown). These data confirm that ATRA promotes the ubiquitylation of cyclin D1.
However, despite the ability of ATRA to decrease the level of cyclin D1 in both cell lines, when the sensitivity to ATRA was tested in MCF-7 and T-47D cells, we found a drastic difference between the two cell lines. MCF-7 and T-47D cells were treated with varying doses of ATRA for 3 days. After 3 days, cells were harvested and counted. Treatment of T-47D cells with 1 M ATRA for 3 days resulted in 60% reduction in proliferation relative to the untreated control T-47D cells (Fig. 1C). In contrast, treatment of MCF-7 cells with 1 M ATRA for 3 days resulted in only a 18% reduction compared with untreated MCF-7 cells (Fig. 1C). The difference in sensitivity to ATRA between the two cell lines was also detected at higher and lower doses of ATRA (Fig. 1C). This observation suggested that accelerated degradation of D-type cyclins is not sufficient to confer sensitivity to ATRA and that additional events are required for response to ATRA.
ATRA Promotes Skp2 Ubiquitylation-The levels of the cdk inhibitor p27 are elevated following ATRA treatment (27)(28)(29), suggesting that p27 may be an important mediator for response to ATRA. To determine whether the accumulation of p27 correlates with the sensitivity to ATRA observed in MCF-7 and T-47D cells, the level of p27 protein was analyzed by Western blotting. We found that ATRA treatment resulted in an increase of p27 in the ATRA-sensitive T-47D cells ( Fig. 2A) but that p27 levels were unaffected by ATRA in MCF-7 cells ( Fig.  2A). This result therefore raises the possibility that the lack of p27 accumulation in MCF-7 cells contributes to the ATRA resistance observed in this cell line.
p27 levels are tightly regulated by the ubiquitin-mediated pathway of proteolysis and recently the SCF Skp2 ubiquitin ligase complex has been shown to be required for its ubiquitylation (18 -20). We therefore tested whether the elevation in p27 levels induced by ATRA correlates with a loss in Skp2 expression. The level of Skp2 was determined by Western analysis in both T-47D and MCF-7 cells following ATRA treatment. We found that ATRA treatment reduced Skp2 levels below detection in the ATRA-sensitive T-47D cells (Fig. 2B). Conversly, ATRA treatment led only to a partial reduction in Skp2 levels in MCF-7 cells, and residual levels of Skp2 protein were clearly detectable in this cell line (Fig. 2B). This result is consistent with the hypothesis that the lack of p27 accumulation observed in MCF-7 cells may result from the residual presence of Skp2 and hence SCF Skp2 activity against p27, following ATRA treatment.
To address the mechanism by which ATRA promotes a decrease in Skp2 levels, T-47D cells were transfected either with a plasmid expressing Skp2 constitutively from the cytomegalovirus promoter, then treated with increasing doses of ATRA for 3 days and Skp2 levels determined by Western analysis. Consistent with our previous observation (Fig. 2B), endogenous levels of Skp2 in vector-only transfected cells decreased with increasing doses of ATRA (Fig. 3A, top panel). In addition, ATRA also led to a decrease in Skp2 levels in cytomegalovirus-Skp2 transfected cells (Fig. 3A, lower panel). As the cytomegalovirus promoter does not contain RARE sequences (30), this observation suggests that the ATRA-induced decrease in Skp2 protein involves a post-translational mechanism. However, a slight reduction in endogenous Skp2 mRNA was observed in control cells following ATRA treatment (Fig. 3B) indicating that a decrease in transcription may also contribute to the decrease in Skp2 levels.
F-box proteins have been proposed to be themselves regulated by ubiquitylation, a mechanism that is thought to ensure that binding of a specific F-box protein to the core of the ubiquitin ligase would not block access of other F-box proteins to the ubiquitin ligase complex (31). We therefore tested whether Skp2 is itself a substrate for ubiquitylation. T-47D cells were transfected with plasmids expressing Myc-tagged ubiquitin and Skp2, and the presence of ubiquitylated forms of Skp2 was analyzed by Western blotting. In the absence of the proteasome inhibitor LLnL, a modest laddering of ubiquitylated forms of Skp2 was detected (Fig. 3C). However, upon inhibition of the proteasome, ubiquitylated intermediates accumulated drastically (Fig. 3C), indicating that Skp2 is itself a substrate for ubiquitylation. This result raised the possibility that, in addition to cyclin D1, Skp2 may also be a target of ATRA-mediated ubiquitylation. We tested this possibility using three separate approaches. First, T-47D cells overexpressing Skp2 were treated with increasing doses of ATRA, the proteasome inhibited and Skp2 levels analyzed by Western blotting. In contrast to what was observed in the absence of LLnL (Fig. 3, A, lower panel, and E), inhibition of the proteasome by addition of LLnL reduced the ability of ATRA to promote Skp2 proteolysis by ϳ50% (Fig. 3, D and E). Second, T-47D cells were treated with ATRA and the presence of ubiquitylated forms of SKP2 were detected by Western blotting. We found that despite the low levels of Skp2 in cells treated with ATRA (Fig. 3F, lower panel), increased levels of ubiquitylated forms of Skp2 were evident (Fig. 3F, top panel). Third, using our in vitro ubiquitylation assay, we found that extracts from ATRA-treated T-47D cells had a higher ubiquitylation activity against 35 S-labeled in vitro translated Skp2 as a substrate when compare with extracts from untreated cells (Fig. 3G). Taken together, these data indicate that, as observed for cyclin D1, ATRA also increases the ubiquitylation of the F-box protein Skp2.
The residual levels of Skp2 detected in MCF-7 cells following ATRA (Fig. 2B) suggests that either the ATRA-mediated degradation of Skp2 or an ATRA-mediated decrease in Skp2 mRNA may be deficient in this cell line and allow cells to continue to proliferate. To distinguish between these two possibilities, the in vitro ubiquitylation activity of MCF-7 cell extracts toward 35 S-labeled Skp2 was analyzed. We found that extracts of MCF-7 treated with ATRA contained similar ubiquitylation activity against 35 S-labeled Skp2 (Fig. 4A) as extracts of T-47D treated with ATRA (Fig. 3G). However, treat- ment with ATRA did not significantly affect the level of Skp2 mRNA (Fig. 4B) indicating that even if the effect of ATRA on the level of Skp2 mRNA is modest (Fig. 3B), the absence of this effect is sufficient to maintain the residual level of Skp2 observed in MCF-7 cells.
Overexpression of Either Skp2 or Cyclin D1 Inhibits the Antiproliferative Effect of ATRA on T-47D Cells-Our observation that Skp2 is a target of ATRA-mediated degradation and that the resistance to ATRA observed in MCF-7 cells correlates with detectable levels of Skp2 and failure to accumulate p27, sug-gest that maintaining Skp2 expression impairs the anti-proliferative effect of ATRA. Therefore, enforced expression of Skp2 should result in ATRA resistance. To test this hypothesis, T-47D cells were transfected with a plasmid expressing Skp2 and the level of Skp2 detected by Western analysis. As expected, the level of Skp2 was elevated following Skp2 transfection compared with mock transfected cells (Fig. 5A). We then treated T-47D cells overexpressing Skp2 with ATRA and determined the effect of Skp2 overexpression on p27 levels. We found that overexpression of Skp2 abolished the accumulation of p27 normally observed following ATRA treatment (Fig. 5B) indicating that Skp2 levels are an important regulator of p27 in response to ATRA. Furthermore, while in the absence of Skp2 overexpression, treatment of T-47D cells with 1 M ATRA resulted in ϳ50% reduction of the proliferation relative to untreated cells (Fig. 5C), when Skp2 was overexpressed, treatment with 1 M ATRA resulted in only a 10% reduction in proliferation (Fig. 5C). Therefore, Skp2 expression led to an increased resistance in ϳ40% of the cells, which is in agreement with the efficiency of transient transfection in T-47D cells (30 -40%).
To determine whether overexpression of cyclin D1 can also promote resistance to ATRA we used a stable clone of T-47D cells (T-47D D1-3) expressing cyclin D1 under the metallothionein promoter. Upon addition of 50 M zinc sulfate to the culture media, expression of cyclin D1 was induced to elevated levels in T-47D D1-3 cells compared with T-47D ⌬MT-1 cells, a stable clone of T-47D transfected with a control plasmid that does not contain the cyclin D1 gene (Fig. 5D).
Similarly, to what we observed with Skp2 overexpression, we found that overexpression of cyclin D1 also led to ATRA resistance (Fig. 5E). These data indicate that overexpression of either cyclin D1 or Skp2 is sufficient to abolish the anti-proliferative effect of ATRA.
ATRA-mediated Ubiquitylation Is Independent of CUL-1 Expression-We next addressed whether ATRA stimulates Skp2 and cyclin D1 ubiquitylation by activating the expression of a specific component of the ubiquitin ligase that normally promotes their ubiquitylation. Skp2 ubiquitylation is dependent on a complex containing the cullin, CUL-1 (32). Furthermore, we have previously reported that cyclin D1 can co-immunoprecipitate with CUL-1 (11) raising the possibility that ATRA may act on Skp2 and cyclin D1 by leading to an elevation in either the expression or activity of CUL-1-associated ubiquitin ligase. To test this possibility, we measured the level of CUL-1 expression following ATRA treatment in T-47D cells by Western analysis. We found that CUL-1 levels were not increased, but rather drastically reduced following ATRA treatment (Fig. 6A). This result suggests that either CUL-1 expression is not required for ATRA-mediated ubiquitylation, or that CUL-1 is itself a target of this pathway. To distinguish between these two possibilities, we first tested whether CUL-1 is a substrate for ubiquitylation by inhibiting the proteasome using LLnL. As shown in Fig. 6B, CUL-1 levels were unaffected by LLnL treatment. Furthermore, we found that inhibiting the proteasome did not affect the ability of ATRA to reduce the CUL-1 level (Fig. 6C). These data indicate that the decrease in CUL-1 level following ATRA is not due to an increase in CUL-1 degradation but is rather consistent with the recent finding that CUL-1 is a transcriptional target of Myc (33). As c-Myc is itself downregulated by ATRA (34), an observation we also confirmed (data not shown), the decrease in CUL-1 is likely a consequence of ATRA mediated inhibition of c-Myc. Therefore, our data suggest that ATRA-mediated ubiquitylation is independent of a CUL-1 containing ubiquitin ligase and raises the possibility of induction of an alternative proteolytic pathway for the degradation of both cyclin D1 and Skp2 following ATRA treatment.
To test the possibility of an alternative pathway of degradation induced by ATRA, we took advantage of the uterine cell line SK-UT1, which is defective in the normal pathway of degradation of D-type cyclins (35) and tested whether treatment with ATRA affected cyclin D1 turnover. SK-UT 1 cells were treated for 3 days with 5 M ATRA and the level of cyclin D1 determined by Western analysis. As observed in T-47D and MCF-7 cells, the levels of cyclin D1 were reduced following ATRA treatment in SK-UT 1 cells (Fig. 6D). We next determined the half-life of cyclin D1 in this cell line in the presence and absence of ATRA. We found that cyclin D1 is very stable in SK-UT 1 cells, with a half-life longer than 4 h, but that in the presence of ATRA the half-life of cyclin D1 was reduced to 1 h (Fig. 6E). These results indicate that ATRA by-passes the ubiquitylation defect observed in SK-UT1 cells and induces an alternative pathway of ubiquitylation.

DISCUSSION
The anti-proliferative effect of ATRA has been associated with its ability to induce a G 1 arrest. Our results indicate that in addition to an increase in the degradation of D-type cyclins, the anti-proliferative effect of ATRA correlates with the stabilization of the CDK inhibitor p27 and that this is due to the combined effect of ATRA on Skp2 protein turnover and Skp2 mRNA expression. In agreement with this conclusion, we showed that in ATRA-resistant MCF-7 cells, the lack of p27 accumulation correlated with the presence of residual Skp2 levels. Moreover, despite an intact ATRA-inducible ubiquitylation activity against Skp2 in MCF-7 cells, maintenance of Skp2 mRNA following ATRA was sufficient to promote resistance. Conversely, we showed that enforcing Skp2 expression in T-47D cells resulted in an increased resistance to ATRA and a lack of p27 accumulation.
In addition to Skp2, we also found that constitutive expression of cyclin D1 in T-47D led to ATRA resistance. Considering the ability of D-type cyclins to sequester p27 away from cyclin E-cdk2 complexes (36), it is likely that overexpression of either cyclin D1 or Skp2 led to resistance by both lowering the level of p27 associated with cyclin E-cdk2.
Several lines of evidence suggest that loss of Skp2 affects cyclin D1 levels: 1) we found that in SK-UT1 cells, lack of Skp2 expression contributes to cyclin D1 stabilization (37); 2) reduction of Skp2 mRNA by antisense Skp2 leads to elevated level of cyclin D1 (21); and 3) loss of Skp2 in mice leads to a modest increase in cyclin D1 levels (22). Therefore, loss of Skp2 following ATRA treatment would have been expected to also lead to a modest stabilization of cyclin D1. In contrast, ATRA treatment led to the destabilization of cyclin D1 suggesting that ATRA-mediated degradation of Skp2 and cyclin D1 defines a distinct pathway. In agreement with this possibility, our data indicated that the mechanism by which ATRA promotes the degradation of D-type cyclins and Skp2 does not require the expression of CUL-1. Therefore, the ATRA-mediated ubiquitylation pathway is independent of an SCF complex containing CUL-1. This result is important because Skp2 is degraded by a CUL-1 containing complex in G 1 . As ATRA leads to an arrest in G 1 , the increased ubiquitylation of Skp2 we observed could have simply reflected a consequence of the G 1 arrest. However, this possibility was ruled out by the fact that CUL-1 is not expressed in ATRA-arrested cells. Furthermore, the observation that ATRA-inducible in vitro ubiquitylation activity against cyclin D1 and Skp2 is similar in both T-47D and MCF-7 cells indicates that this activity is independent of the phase of the cell cycle as MCF-7 cells fail to arrest following ATRA treatment.
Additional support for the existence of an alternative pathway of degradation came from our observation that ATRA was able to rescue the degradation of cyclin D1 in the uterine cell line SK-UT1, which is deficient in the degradation of D-type cyclins (35). However, the rescue was partial in SK-UT1 cells as cyclin D1 was detectable after treatment with ATRA (Fig. 6D), while similar treatment completely abolished cyclin D1 detection in T-47D and MCF-7 cells (Fig. 1A). This observation suggests that ATRA may promote the degradation of only a subset of cyclin D1. Furthermore, experiments would be required to determine this possibility.
The identification of Skp2 as a downstream target of ATRA represents an important step in our understanding of the antiproliferative action of this drug. The best known substrate of ATRA-mediated ubiquitylation and degradation is the PML-RAR␣ fusion protein, which results from a characteristic chromosomal translocation involving the RAR-␣ gene on chromosome 17 and the PML gene on chromosome 15. The resulting fusion protein acts as a dominant negative over the normal cellular function of PML and RAR-␣ and causes acute promyelocytic leukemia. The ability of ATRA to promote the destruction of this dominant negative form of PML allows the restoration of PML nuclear localization and function and result in complete recovery from the disease (for a review, see Ref. 38). However, how ATRA stimulates the ubiquitylation of PML-RAR␣ remains unknown. The identification of cyclin D and Skp2 as additional targets of ATRA-mediated ubiquitylation uncovers a broader effect of ATRA on ubiquitylation of several proteins and raises the possibility that they may be degraded by the same mechanism. Retinoic acid resistance is a serious problem for patients with acute promyelocytic leukemia who have received ATRA treatment. Interestingly, acute promyelocytic leukemia cell lines with features of resistance to ATRA show a defect in the accumulation of p27 following ATRA treatment (39). Results shown in this study raise the possibility that these cells may acquire mutations leading either to a defect in the ATRA-mediated ubiquitylation of Skp2 or to the amplification of the Skp2 gene. Furthermore, our results also raise the possibility that in some patients, ATRA resistance may arise from cyclin D1 overexpression. Skp2 overexpression is observed in several cancers and has been shown to correlate with loss of p27 expression (40). Considering that overexpression of cyclin D1, loss of p27, and overexpression of Skp2 are all frequent events associated with cancers, understanding how ATRA affect the degradation of these proteins and how we can overcome the resistance associated with their overexpression represent important areas for future research.