Upstream molecular signaling pathways of p27(Kip1) expression: Effects of 4-hydroxytamoxifen, dexamethasone, and retinoic acids

Background p27(Kip1) is a cyclin-dependent kinase inhibitor that inhibits G1-to-S phase transition of the cell cycle. It is known that a relatively large number of nutritional and chemopreventive anti-cancer agents specifically up-regulate expression of p27 without directly affecting the expression of other G1-to-S phase cell cycle regulatory proteins including p21(Cip1Waf1). However, the upstream molecular signaling pathways of how these agents up-regulate the expression of p27 have not been well characterized. The objective of this study was to identify such pathways in human breast cancer cells in vitro using 4-hydroxytamoxifen, dexamethasone, and various retinoic acids as examples of such anti-cancer agents. Results Experimental evidence presented in the first half of this report was obtained by transfecting human breast cancer cells in vitro with proximal upstream region of p27 gene-luciferase reporter plasmids. 1) The evidence indicated that 4-hydroxytamoxifen, dexamethasone, and various retinoic acids up-regulated expression of p27 in both estrogen receptor-positive and negative human breast cancer cells in vitro. 2) The degree of up-regulation of p27 expression by these anti-cancer agents in human breast cancer cells in vitro linearly correlated with the degree of inhibition of methylnitrosourea (MNU)-induced rat mammary adenocarcinoma in vivo. 3) Lastly, up-regulation of the expression of p27 was likely due to the activation of translation initiation rather than transcription of p27 gene. The experimental evidence presented in the second half of this report was obtained by a combination of Western immunoblot analysis and transfection analysis. It indicated that 4-hydroxytamoxifen and dexamethasone up-regulated expression of p27 by down-regulating phosphorylation of eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (4E-BP1) at Ser65 and this phosphorylation was likely to be mediated by upstream receptor tyrosine kinases/phosphoinositide-3-kinase/Akt/5'-AMP-activated protein kinase/mammalian target of rapamycin (RTKs/PI3K/Akt/AMPK/mTOR) protein kinase signaling pathways. Retinoic acids up-regulated expression of p27 without using either 4E-BP1 or RTKs/PI3K/Akt/AMPK/mTOR protein kinase signaling pathways. Conclusions 4-Hydroxytamoxifen and dexamethasone up-regulated translation initiation of p27 by down-regulating 4E-BP1 phosphorylated at Ser65 and this down-regulation seemed to be mediated by upstream RTKs/PI3K/Akt/AMPK/mTOR protein kinase signaling pathways. Retinoic acids also up-regulated translation initiation of p27, but without using any of these pathways.


Background
Cyclin-dependent kinases (CDKs), together with cyclins, their regulatory subunits, govern cell cycle progression in eukaryotic cells. p27(Kip1) is a member of a family of CDK inhibitors (CDIs) that bind to cyclin/CDK complexes and arrest cell cycle progression from G1 to S phase.
Previously, we identified four different upstream molecular signaling pathways of p27 expression using p27-luciferase reporter plasmids and numerous specific inhibitors and stimulators of p27 expression [10]. (We will call these four pathways as pathway #1, #2, #3 and #4.). This approach was very efficient and sensitive in identifying upstream molecular signaling pathways of p27 expression, but it had a major drawback; namely, it could not tell which specific anti-cancer agent uses which specific pathway to up-regulate p27 expression. To address this question, Western immunoblot analysis, although cumbersome and not as sensitive as p27-luciferase reporter assays, must have been performed. The objective of the present study, therefore, was to perform Western immunoblot analysis using 4-hydroxytamoxifen, dexamethasone, and retinoic acids as examples of anti-cancer agents to identify which specific upstream molecular signaling pathway each one of these anti-cancer agents uses to up-regulate the expression of p27 in human breast cancer cells in vitro.

4-Hydroxytamoxifen
, dexamethasone, all-trans-retinoic acid and 9-cis-retinoic acid up-regulated expression of p27 in both estrogen receptor-positive and -negative human breast cancer cells in vitro The diagram in Figure 1a shows the outline of how various anti-cancer agents specifically up-regulate expression of p27 and arrest cell cycle progression from G1 to S phase. The upstream molecular signaling pathways of how these anti-cancer agents up-regulate the expression of p27 was investigated using a p27-luciferase reporter plasmid containing proximal upstream region (-1797) of p27 gene (p27-Kpn I) (Figure 1b) [32]. This plasmid was transfected into the estrogen receptor (ER) -positive as well as negative human breast cancer cells in vitro and then the transfected cells were exposed to 1 μM each of the following five different anti-cancer agents, namely In all experiments, the cells were exposed to 1 μM each of tamoxifen, 4-hydroxytamoxifen, dexamethasone, all-trans-retinoic acid or 9-cis-retinoic acid for 24 hours. All assays were performed in triplicates and repeated three times. tamoxifen, 4-hydroxytamoxifen, dexamethasone, alltrans -retinoic acid (atRA), and 9-cis-retinoic acid (9cRA) for 24 hours. The results (Figure 1c and 1d) indicated first that tamoxifen did not up-regulate the expression of p27 in both MDA-MB-231 and MCF7 cells, but other four anti-cancer agents up-regulated the expression of p27 in both ER-positive (also LKB1-positive) and ER-negative (also LKB1-negative) human breast cancer cells in vitro. Next, expression of p27 protein in ER-negative MDA-MB-231 cells was examined by Western immunoblot analysis. The results ( Figure  1e) indicated that tamoxifen and all-trans-retinoic acid (atRA) did not up-regulate the expression of p27 protein, but 4-hydroxitamoxifen, dexamethasone and 9-cisretinoic acid (9cRA) did. It should be noted that, although all-trans-retinoic acid (atRA) did not up-regulate the expression of p27 protein in a statistically significant manner, average expression of p27 protein tended to be higher in the presence of all-trans-retinoic acid (atRA) than in the absence of all-trans-retinoic acid (atRA).
In summary, these results suggested that 4-hydroxytamoxifen (but not tamoxifen), dexamethasone, 9-cis-retinoic acid (9cRA) and probably all-trans-retinoic acid (atRA) up-regulated the expression of p27 in both ERpositive and negative human breast cancer cells in vitro (Figures 1c, 1d and 1e).
The degree of up-regulation of p27 in human breast cancer cells in vitro linearly correlates with the degree of inhibition of methylnitrosourea (MNU)-induced rat mammary adenocarcinoma in vivo In the next experiment, we used various chemically synthesized retinoic acids to investigate whether the degree of up-regulation of the -1797 p27-luciferase reporter activity (p27-Kpn I) in human breast cancer cells in vitro correlates with the degree of inhibition of methylnitrourea (MNU)-induced rat mammary adenocarcinoma in vivo. The results presented in the Figures  2a and 2b indicated that the up-regulation of the in vitro p27-luciferase reporter activity by various retinoic acids indeed correlated with the in vivo activity of the inhibition of MNU-induced rat mammary cancer by the same retinoic acids [33]. The Figure 2c graphically represents the results in Figure 2a; it shows that the in vitro and in vivo parameters of the inhibition of breast cancer linearly correlated with each other and the correlation is statistically significant. One note of caution about this linear correlation: if a particular anti-cancer agent (e.g., tamoxifen) needs to be metabolized into an ultimately active anti-cancer agent (e.g., 4-hydroxytamoxifen) in vivo, then the in vitro and in vivo activities of this particular anti-cancer agent (e.g., tamoxifen) do not follow this linear correlation.
The results shown in the left half of the Figure 4b [33]. The cells were exposed to 1 μM each of the retinoic acids for 24 hours. In vitro transfection assays were performed in triplicates and repeated three times. (b) Chemical structure of the retinoic acids used in this experiment [33]. (c) Graphical representation of the results in Figure 2a above.

Figure 4
The 5'-untranslated region (5'-UTR) (-575) of p27 gene is unlikely to contain cryptic transcription factor binding sites. (a) Schematic drawing (adapted from the references [11,34]) of the pGL3-control-p27-5'-UTR-luciferase reporter plasmid. (b) The ER-negative MDA-MB 231 cells were transfected with -575 p27 (p27-5'-UTR) -luciferase reporter plasmid and then treated with either vehicle or actinomycin D(0.5 μg/ml) [55]. One hour after the addition of either vehicle or actinomycin D, the cells were exposed to vehicle, tamoxifen (1 μM) or 4hydroxytamoxifen (1 μM) for another 24 hours. (c) Same as in Figure 4b above, except that the cells were exposed to vehicle, all-trans-retinoic acid (atRA) (1 μM) or 9-cis-retinoic acid (9cRA) (1 μM) for 24 hours. (d) Same as in Figure 4b above, except that the cells were exposed to vehicle, 4-methyl-UAB30 (4meUAB30) (1 μM) or UAB30 (1 μM) for 24 hours. (e) Same as in Figure 4b, except that the cells were exposed to vehicle or dexamethasone (1 μM). All assays were performed in triplicates and the transfection experiments were repeated three times. addition of actinomycin D in the presence of vehicle (DMSO) alone decreased the baseline p27-luciferase activity of -575 p27 (p27-5'-UTR) by about 50% compared to the baseline luciferase activity observed in the absence of actinomycin D. Despite this decrease in the baseline p27-luciferase activity in the presence of actinomycin D, 4-hydroxytamoxifen significantly up-regulated the p27-luciferase activity of -575 p27 (p27-5'-UTR) above that of the vehicle (DMSO) in the presence of actinomycin D. These results suggested that the transcriptional mechanisms were not involved in a significant manner in the up-regulation of the luciferase activity of -575 p27 (p27-5'-UTR) by 4-hydroxytamoxifen, precluding the involvement of any cryptic transcription factor binding sites in this region. What was more surprising was the finding that tamoxifen, which had previously been inactive in the absence of actinomycin D, now significantly up-regulated the p27-luciferase activity of -575 p27 (p27-5'-UTR) in the presence of actinomycin D, suggesting that the overall rate of global transcription might somehow exerted effects on the p27-luciferase activity of -575 p27 (p27-5'-UTR) in MDA-MB-231 cells.
4-Hydroxytamoxifen and dexamethasone up-regulated the expression of p27 by down-regulating 4E-BP1 phosphorylated at Ser65 and this down-regulation was likely to be mediated by upstream RTKs/Akt/AMPK/mTOR protein kinase signaling pathways. Retinoic acids also upregulated the expression of p27 but they did so without using any of these pathways Previous study identified four potential upstream molecular signaling pathways that might be involved in the up-regulation of the expression of p27 by these anticancer agents in the ER-negative MDA-MB-231 human breast cancer cells in vitro [10]. These four potential upstream molecular signaling pathways of p27 were pathway #1 (Figures 5a and 5b), pathway #2 (Figures 5a and 5b), pathway #3 ( Figure 5b) and pathway #4 ( Figure  5b). To investigate which one of these upstream molecular signaling pathways was used by 4-hydroxitamoxifen, dexamethasone, all-trans-retinoic acid (atRA) and 9-cis-retinoic acid (9cRA) to up-regulate the expression of p27, Western immunoblot analysis was performed using the ER-negative MDA-MB-231 human breast cancer cells in vitro (Figures 5 and 6). We investigated only the pathways #1, #2 and #3 in this Western immunoblot study; the pathway #4 was not investigated.
Most notable result of this Western immunoblot study was the expression of eukaryotic translation initiation repressor protein 4E-BP1 (eukaryotic translation initiation factor 4E-binding protein 1) phosphorylated at Ser65. As the results in Figure 5c indicate, expression of total 4E-BP1 was neither up nor down-regulated by any of the anti-cancer agents tested (Figure 5c). However, the 4E-BP1 phosphorylated at Ser65 was significantly down-regulated by two of the anti-cancer agents tested, namely 4-hydroxytamoxifen and dexamethasone ( Figure  5d). The 4E-BP1 phosphorylated at Ser65 was neither up nor down-regulated by tamoxifen, all-trans-retinoic acid (atRA) or 9-cis-retinoic acid (9cRA) (Figure 5d). These results suggested that 4-hydroxytamoxifen (but not tamoxifen) and dexamethasone used either the upstream molecular signaling pathway #1 or #2 or both to up-regulate the expression of p27. They also suggested that the two retinoic acids tested did not use pathways#1 and #2 to up-regulate the expression of p27.
The second most notable result of this study was the expression of the following two proteins which were significantly up or down-regulated by one or more of these anti-cancer agents tested: (a) one was AMPKα (5'-AMPactivated protein kinase α) phosphorylated at Thr172 (Figure 5f) and (b) another was Akt/PKB phosphorylated at Thr308 (Figure 6b).
(a) In the case of AMPKα, expression of total AMPKα was neither up nor down-regulated by any of the anticancer agents tested (Figure 5e), but the expression of AMPKα phsophorylated at Thr172 was up-regulated by dexamethasone ( Figure 5f). Therefore, it is reasonable to assume that dexamethasone up-regulated the expression of p27 by using upstream molecular signaling pathway #2 (Figures 5a, 5b and 7).
(b) In the case of Akt/PKB, expression of total Akt/ PKB was neither up nor down-regulated by any of the anti-cancer agents tested (Figure 6a), but the expression of Akt/PKB phosphorylated at Thr308 was down-regulated by 4-hydroxytamoxifen and dexamethasone ( Figure  6b). Since 4-hydroxytamoxifen did not up-regulate the expression of AMPKα phosphorylated at Thr172 (Figure  5f), it is likely that 4-hydroxytamoxifen used the upstream molecular signaling pathway #1 exclusively to up-regulate the expression of p27 (Figures 5a, 5b and 7). As for dexamethasone, expression of p27 could have been up-regulated by dexamethasone using either one or both of the following two pathways: namely either (a) dexamethasone used both pathways #1 and #2, or (b) Figure 5 4-Hydroxytamoxifen and dexamethasone up-regulate the expression of p27 by down-regulating phosphorylation of 4E-BP1 and this down-regulation is likely to be mediated by upstream Akt/AMPK/mTOR protein kinase signaling pathways. Retinoic acids also up-regulate the expression of p27 but they do so without using any of these pathways. (a) and (b): Schematic drawings of the four upstream molecular signaling pathways of p27 expression identified in our previous study [10]. These pathways are: pathway #1 (Figures 5a and  5b), pathway #2 (Figures 5a and 5b), pathway #3 ( Figure 5b) and pathway #4 (Figure 5b). The specific inhibitors and activators used previously to identify these four pathways are indicated next to each of the four pathways. From (c) to (f): Estrogen receptor (ER) -negative MDA-MB-231 human breast cancer cells in vitro were exposed to vehicle, tamoxifen (1 μM), 4-hydroxytamoxifen (1 μM), dexamethasone (1 μM), all-transretinoic acid (atRA) (1 μM) or 9-cis-retinoic acid (9cRA) (1 μM) for 24 hours. Western immunoblot assays of the cells exposed to these anti-cancer agents were performed using antibodies against (c) total 4E-BP1, (d) 4E-BP1 phosphorylated at Ser65, (e) total AMPK, and (f) AMPK phosphorylated at Thr172. All assays were performed in triplicates and repeated three times. Abbreviations: RTK, receptor tyrosine kinase; PI3K, phosphoinositide 3-kinase; PKB, protein kinase B; AMPK, 5'-AMP-activated protein kinase; TSC, tuberous sclerosis complex; mTOR, mammalian target of rapamycin; eIF4E, eukaryotic translation initiation factor 4E; 4E-BP1, eIF4E-binding protein 1; MAPK, mitogen-activated protein kinase; Raf, MAP kinase kinase kinase; MEK, MAP kinase kinase; MKK, MAP kinase kinase; MNK, MAP kinase-interacting kinase; eIF2α, eukaryotic translation initiation factor 2α. Figure 6 Further results (continuation of Figure 5) of the Western immunoblot analysis. Same as in Figure 5, except that Western immunoblot assays of the cells were performed using antibodies against (a) total Akt/PKB, (b) Akt/PKB phosphorylated at Thr308, (c) total IRS-1 (insulin receptor substrate 1), IRS-1 phosphorylated at Ser636/639, total PTEN (phosphatase and tensin homolog), PTEN phosphorylated at Ser380, total eIF4E (eukaryotic translation initiation factor 4E), eIF4E phosphorylated at Ser209, total eIF2α (eukaryotic translation initiation factor 2α), and eIF2α phosphorylated at Ser52, (d) PDGFRβ (platelet-derived growth factor receptor b) phosphorylated at Tyr751, and (e) p44/42 MAPK or ERK1/2 phosphorylated at Thr202Tyr204. All assays were performed in triplicates and repeated three times.  Figure 5a are pathways #1 and #2. The pathway #1 consisted of receptor tyrosine kinases/phosphoinositide 3-kinase/Akt/tuberous sclerosis complex/mammalian target of rapamycin/eukaryotic translation initiation factor 4E (eIF4E) -binding protein 1 (RTKs/PI3K/Akt/TSC/mTOR/4E-BP1). The pathway #2 consisted of 5'-AMP-activated protein kinase (metabolic energy sensor or cellular fuel gauge)/tuberous sclerosis complex/mammalian target of rapamycin/eIF4E-binding protein 1 (AMPK/TSC/mTOR/4E-BP1). (b) In addition to these two pathways, two more upstream molecular signaling pathways of p27 expression were previously identified. They were pathways #3 and #4. The pathway #3 consisted of receptor tyrosine kinases/MAPKs/eIF4E (RTKs/MAPKs/eIF4E). The pathway #4 consisted of global hypomethylation of the 5'-7-methylguanosine (m 7 G) cap of mRNAs. The specific inhibitors and activators used previously to identify these four pathways are indicated next to each of the four pathways. The results of this study suggested that 4-hydroxytamoxifen used pathway #1 and dexamethasone primarily used pathway #2 to up-regulate the expression of p27. Dexamethasone could also use a portion of pathway #1 secondarily. We also believe, but could not conclude, that 4-hydroxytamoxifen up-regulated the expression of p27 using MAP kinase pathways (Pathway #3 in Figures 5b and 7b). Retinoic acids up-regulated p27 expression without using pathways #1, #2 and #3. We propose a hypothesis that retinoic acids are likely to have used pathway #4 to up-regulate the expression of p27. Abbreviations: see the legend of Figure 5. dexamethasone primarily up-regulated AMPKα phosphorylated at Thr172 (pathway #2 in Figures 5a, 5b and  7), the up-regulation of which could have in turn secondarily down-regulated the Akt/PKB phosphorylated at Thr308 (pathway #1 in Figures 5a, 5b and 7).

Discussion
The cell cycle repressor protein p27 exhibits a set of unique characteristics that are not seen in other G1-to-S phase cell cycle regulatory proteins including p21 [10]. First, a relatively large number of nutritional and chemopreventive anti-cancer agents specifically up-regulate the expression of p27 without directly affecting expression of other G1-to-S phase cell cycle regulatory proteins. Secondly, the degree of up-regulation of the expression of p27 by these anti-cancer agents in human breast cancer cell lines in vitro linearly and positively correlates with the degree of inhibition of methylnitrosourea (MNU)-induced rat mammary adenocarcinoma by the same anti-cancer agents. If a particular anti-cancer agent must be converted to an active metabolite in vivo to up-regulate the expression of p27, the degree of up-regulation of p27 in vitro and the degree of inhibition of MNU-induced rat mammary adenocarcinoma in vivo by the same anti-cancer agent do not follow this linear relationship. An example of such anti-cancer agent is tamoxifen which must be converted to 4-hydroxytamoxifen in vivo to up-regulate the expression of p27. Lastly, unlike other G1-to-S phase cell cycle regulatory proteins, expression of p27 is regulated primarily at the level of translation, not at the level of transcription.
In the 1980s and 1990s, it was observed that, during the progression of cell cycle, the level of p27 protein expression oscillated cyclically, but the level of p27 mRNA remained constant. This observation led investigators to suggest that, during the cell cycle, expression of p27 is regulated primarily at the level of translation, not at the level of transcription [11,[17][18][19]. The expression of p27 during the cell cycle could also be regulated by various post-translational mechanisms including ubiquitin-proteasome-induced degradation [20][21][22][23], complex formation [24], subcellular localization [25][26][27][28][29][30] and phosphorylation [12,30,31]. Based on the results of our present and previous studies [10], we believe that a relatively large number of nutritional and chemopreventive anti-cancer agents up-regulate the expression of p27primarily by activating the rate of translation.

4-Hydroxytamoxifen (but not tamoxifen) up-regulates p27 expression by down-regulating eukaryotic translation initiation repressor protein 4E-BP1 phosphorylated at
Ser65 and this down-regulation is likely to be mediated by upstream receptor tyrosine kinases/phosphoinositide-3-kinase/Akt/tuberous sclerosis complex/mammalian target of rapamycin (RTKs/PI3K/Akt/TSC/mTOR) protein kinase signaling pathway (pathway #1) 4-Hydroxytamoxifen (but not tamoxifen) up-regulated expression of p27 in estrogen receptor (ER) -positive as well as negative breast cancer cells in vitro [also see the reference [10], suggesting that 4-hydroxytamoxifen upregulates the expression of p27 regardless of the status of estrogen receptor in the breast cancer cells.
The results also indicated that 4-hydroxytamoxifen (but not tamoxifen) down-regulates eukaryotic translation initiation repressor protein 4E-BP1 phosphorylated at Ser65. It was reported in 2001 that co-expression of the mutant 4E-BP1, which was altered at five different amino acid positions that are normally the targets for phosphorylation, up-regulated the expression of p27 through 5'-untranslated region (5'-UTR) in the proximal upstream region of p27 gene in D6P2T Schwannoma cells [35]. Based on this observation and our results taken as a whole, we conclude that down-regulation of 4E-BP1 phosphorylated at Ser65 constitutes an essential component of the upstream molecular signaling pathways of the up-regulation of p27 expression induced by 4-hydroxytamoxifen.
It is worth noting in this respect that decreased phosphorylation of 4E-BP1 normally leads to decreased translation initiation of mRNAs in general, but for p27 the effect is opposite; it leads, instead, to increased translation initiation of p27 mRNA. This opposite effect of phosphorylated 4E-BP1 on p27 translation initiation is likely to be achieved through its unusually long 5'untranslated region (5'-UTR) (-575) in the p27 gene  [11,34]. (b) The most stable secondary stem and loop structure of the wild-type IRES region of Hengst (-417 to -1) [11,[34][35][36] at several lowest Gibbs free energy values. This structure was generated using Zucker's RNA mfold software version 2.3 [56,57]. (c) Same as in Figure 8b, except that UU at -57 and -56 in the polypyrimidine tract (-66 to -41) were mutated to GG. This mutation completely destroyed the overall secondary stem and loop structure of IRES. Various other mutations introduced into the polypyrimidine tract also significantly modified the overall stem and loop structure of IRES. [11,[34][35][36], which contains two unusual nucleotide motifs, namely uORF (upstream open reading frame) and IRES (internal ribosome entry site) (Figure 8a). Combination of these two elements makes it possible for p27 mRNA to achieve the reverse, cap-independent translation initiation mechanisms as opposed to the normal, cap-dependent translation initiation mechanisms of mRNAs in general.
The critical nucleotide sequence within the IRES motif in the 5'-untranslated region (UTR) of the p27 gene resides in the polypyrimidine tract located between -66 and -41 relative to the translation initiation start site (Figures 8a and 8b) [11,[34][35][36]. If this polypyrimidine tract is disrupted by mutations, expression of p27 significantly decreases due to the failure of 40S ribosomal subunit to recognize and bind to the IRES motif (see, for example, Figure 8c). In 2005, an article was published in which the authors induced two mutations in what was called "FOXO response element" located at around -57 relative to the translation initiation start site of p27 gene. They observed that these mutations significantly reduced the p27 promoter activity and stated that the disruption of this putative FOXO responsive element decreased the "transactivation of the FOXO response element that was present in the p27 promoter" [37]. Unfortunately, they did not perform gel shift assay to investigate if any transcription factor binds to this element. We believe, along with Koff, Miskimins, Hengst and other investigators [11,[34][35][36], that the transcription of human p27 gene starts significantly upstream of -51in fact it is -575 in human p27 gene -rather than somewhere between -51 and -1 from the translation initiation start site. The so-called FOXO responsive element located at around -51 seems to represent the polypyrimidine tract within the IRES motif which is located between -66 and -41 relative to the translation initiation site.
As for the issue of the upstream molecular signaling pathways of how 4-hydroxytamoxifen up-regulates the expression of p27, 4-hydroxytamoxifen seems to downregulate phosphorylated 4E-BP1 using upstream receptor tyrosine kinase/phosphoinositide-3-kinase/Akt/ tuberous sclerosis complex/mammalian target of rapamycin (RTKs/PI3K/Akt/TSC/mTOR) protein kinase signaling pathway (Pathway #1 in Figure 5a, 5b and 7). This is based on the observation that 4-hydroxytamoxifen down-regulated Akt/PKB phosphorylated at Thr308 without up-regulating AMPKα phosphorylated at Thr172. The results of our previous study also indicated that inhibitors of several receptor tyrosine kinases [see also, for example, reference [38], LY294,002 (inhibitor of PI3K), triciribine (inhibitor of Akt/PKB), and rapamycin (inhibitor of mTOR) upregulated the p27-luciferase reporter activity in estrogen receptor (ER) -negative MDA-MB-231 human breast cancer cells in vitro [10]. We also believe, but could not conclude, that 4-hydroxytamoxifen up-regulated the expression of p27 via MAP kinase pathways (Pathway #3 in Figures 5b and 7b).
Dexamethasone up-regulates the expression of p27 by down-regulating phosphorylated eukaryotic translation initiation repressor protein 4E-BP1 and this downregulation is likely to be mediated -primarily -by upstream 5'-AMP-activated protein kinase/tuberous sclerosis complex/mammalian target of rapamycin (AMPK/TSC/mTOR) protein kinase signaling pathway (pathway #2) Similar to 4-hydroxytamoxifen, dexamethasone also upregulated the expression of p27 in estrogen receptor (ER) -positive as well as negative breast cancer cells in vitro [also see the reference [10]. In addition, dexamethasone down-regulated eukaryotic translation initiation repressor protein 4E-BP1 phosphorylated at Ser65 through 5-untranslated region (5'UTR) (-575) in the proximal upstream region of p27 gene.
The effect of dexamethasone on p27 expression appears to be somewhat different from the effect of 4hydroxytamoxifen in terms of the molecular signaling pathway upstream of 4E-BP1 (Pathways #1 and #2 in Figures 5a, 5b and 7). Unlike 4-hydroxytamoxifen, dexamethasone up-regulated AMPKα phosphorylated at Thr172. The results of previous studies published by other investigators seem to agree with our observation (for example, [39][40][41][42][43][44]). Since MDA-MB-231 human breast cancer cells are negative not only in estrogen receptor (ER), but also in LKB1 (Drosophila par-4 homologue gene), we believe that dexamethasone upregulated AMPKα phosphorylated at Thr172 without activating LKB1. The up-regulation of AMPKα phosphorylated at Thr172 by dexamethasone probably led to the up-regulation of p27 expression by way of tuberous sclerosis complex (TSC) proteins, mammalian target of rapamycin (mTOR) and 4E-BP1.
It should be noted that the up-regulation of AMPKα phosphyrylated at Thr172 by dexamethasone could indirectly down-regulate Akt/PKB phosphorylated at Thr308.
In summary, we believe that dexamethasone up-regulated the expression of p27 by down-regulating phosphorylated 4E-BP1 and this down-regulation was mediated primarily by 5'-AMP-activated protein kinase α/tuberous sclerosis complex/mammalian target of rapamycin (AMPKα/TSC/mTOR) protein kinase signaling pathway (Pathway #2 in Figures 5a, 5b and 7). The results of our previous study [10] also indicated that AMPK is involved in both up and down-regulation of p27 expression, namely (a) AICAR (aminoimidazole carboxamide riboside; activator of AMPK), (b) rotenone (inhibitor of Complex I in mitochondrial oxidative phosphorylation), and (c) rapamycin (inhibitor of mTOR) up-regulated the expression of p27. In contrast, Compound C (inhibitor of AMPK) down-regulated the expression of p27 in estrogen receptor (ER) -negative MDA-MB-231 human breast cancer cells in vitro [10].
Finally, we do not believe that dexamethasone upregulated expression of p27 using upstream MAP kinase pathways (Pathway #3 in Figures 5b and 7b).
Of the four upstream molecular signaling pathways of p27 expression identified previously, we investigated only three pathways (#1, #2 and #3) in the present study. The pathway #4 was not investigated. Potential involvement of the pathway #4 in the expression of p27 by retinoic acids was suggested by the results of our previous study where NSC 119889, an inhibitor of the global methylation of 5'-m 7 G-cap of mRNAs, up-regulated the p27-luciferase reporter activity of the 5'untranslated region (5'UTR) (-575) within the proximal upstream region of the p27 gene (Figures 5b and 7b) [10].
It is known that nearly all mRNAs are post-transcriptionally modified at their 5' and 3' ends by capping and polyadenylation, respectively [45][46][47]. The m 7 G-capping at the 5' end protects the nascent pre-mRNAs against degradation. Therefore, failure to cap or loss of cap leads to rapid breakdown of mRNAs. The enzyme 5'-mRNA cap (guanine-N 7 ) methyltransferase catalyzes transfer of methyl group from S-adenosylmethionine (AdoMet or SAM) to GpppRNA to form m 7 GpppRNA. We observed in our previous study [10] that NSC 119889, a cell-permeable, competitive inhibitor of Ado-Met (SAM), inhibited global cap-dependent translation initiation of 5'-m 7 G-capped mRNAs in general, but it increased cap-independent translation initiation of p27 mRNA through its 5'-UTR in estrogen receptor (ER) -negative MDA-MB-231 human breast cancer cells in vitro.
Schalinske and other investigators have been reporting for almost two decades that retinoic acids decrease the ratio of S-adenosylmethionine (AdoMet or SAM) to Sadenosylhomocysteine (AdoHcy or SAH) presumably by inducing glycine N-methyl transferase [48][49][50][51][52][53]. This observation suggests that retinoic acids decrease the ratio of SAM/SAH thereby inducing global hypomethylation of 5'-m 7 G-cap of mRNAs, which in turn up-regulates the expression of p27 by increasing reverse, capindependent translation initiation of p27 mRNA through its 5'-UTR.
Based on these considerations, we propose that retinoic acids up-regulate the expression of p27 by reducing the methylation of the 5'-m 7 G -cap of mRNAs in general, and at the same time, increasing the reverse, capindependent translation initiation of p27 mRNA through its 5'-UTR (pathway #4 in Figures 5b and 7b).

Conclusions
Based on the results presented above, we conclude that: (a) 4-Hydroxytamoxifen (but not tamoxifen) up-regulates the expression of p27 in both estrogen receptorpositive and negative human breast cancer cells in vitro by down-regulating phosphorylation of 4E-BP1 and this down-regulation is mediated by upstream receptor tyrosine kinases/phosphoinositide 3-kinase/ Akt/tuberous sclerosis complex proteins/mammalian target of rapamycin (RTKs/PI3K/Akt/TSC/mTOR) protein kinase signaling pathway (pathway #1 in Figures 5a, 5b and 7). We also believe, but could not conclude, that 4-hydroxytamoxifen up-regulates the expression of p27 using MAP kinase pathways (Pathway #3 in Figures 5b and 7b).
(b) Dexamethasone up-regulates the expression of p27 in both estrogen receptor-positive and negative human breast cancer cells in vitro by down-regulating phosphorylation of 4E-BP1 and this down-regulation is mediated primarily by upstream 5'-AMP-activated kinase/tuberous sclerosis complex proteins/mammalian target of rapamycin (AMPK/TSC/mTOR) protein kinase signaling pathway (pathway #2 in Figures 5a, 5b and 7). We do not believe that dexamethasone up-regulated expression of p27 using MAP kinase pathways (Pathway #3 in Figures 5b and 7b) (c) Retinoic acids also up-regulate the expression of p27 in both estrogen receptor-positive and negative human breast cancer cells in vitro, but they do so without using any of the pathways described above for 4hydroxytamoxifen and dexamethasone. We propose that retinoic acids up-regulate the expression of p27 by decreasing the ratio of SAM/SAH thereby inducing hypomethylation of the 5'-m 7 G-cap of p27 mRNA. Hypomethylation of the 5'-cap of p27 mRNA in turn activates the reverse, cap-independent translation initiation of p27 mRNA through its 5'-untranslated region (5'-UTR), which contains upstream open reading frame (uORF) and internal ribosome entry site (IRES) (pathway #4 in Figures 5b and 7b).
The following antibodies were purchased from Cell Signaling Technology, Inc.

Cell Cultures
Human MCF7 (estrogen receptor-positive and LKB1positive) and MDA-MB-231 (estrogen receptor-negative and LKB1-negative) breast cancer cell lines were purchased from the American Type Culture Collection (Rockville, MD, USA). MCF7 cells were grown in Dulbecco's Modified Eagle's Medium (DMEM) containing 4.5 g/L of D-(+)-glucose, supplemented with 10% heatinactivated fetal bovine serum (FBS), 100 mg/L recombinant human insulin, 2% L-glutamine, and antibiotic/ antimycotic solution. MDA-MB-231 cells were grown in the same culture medium without insulin. The incubation was carried out at 37°C in a 5% CO 2 humidified chamber. All cells were subcultured after trypsinization with 0.05% trypsin-0.02% EDTA solution. The cell cultures were always maintained below confluency. The cells were checked periodically for mycoplasmal infection by DNA fluorochrome staining.

Plasmids
Luciferase reporter plasmids containing one of the following proximal 5'-upstream region of the p27 gene were used to transfect the human breast cancer cells: -1797 p27 (p27-Kpn I) [32], -774 p27 (p27-Apa I) [32], -575 p27 (p27-5'-UTR) [11,34], -435 p27 (p27-MB) [32], and -417 p27 (p27-IRES) [11,34] (see Figures 1b and  3a). The control luciferase reporter plasmids that did not contain these inserts were also prepared to test if nutritional and chemopreventive anti-cancer agents were exerting any spurious effects on the backbone rather than the insert of the luciferase reporter plasmids. All of the nutritional and chemopreventive anti-cancer agents tested did not exert spurious effects on the backbone of the luciferase reporter plasmid in the human breast cancer cells in vitro.

Transfection and Luciferase Assay
Transfections were performed according to the published protocol [54] using FuGENE 6 purchased from Roche Applied Science (Indianapolis, IN, USA). Briefly, 24 hours before the transfection of luciferase-reporter plasmid, cells were seeded into a 60-mm tissue culture dish containing 3 mL of Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% heatinactivated fetal bovine serum (FBS), 2% L-glutamine, and antibiotic/antimycotic solution at a density of 1.5 × 10 5 cells/dish and incubated at 37°C in a 5% CO 2 humidified chamber. Transfection was carried out with 1 μg of luciferase reporter plasmid and 0.2 μg of pSV-β-galactosidase internal control plasmid (Promega, Madison, WI, USA) mixed with 3 μL of FuGENE 6 solution in 3 mL of FBS-free DMEM supplemented with only 2% L-glutamine. A minimum of 5-hour incubation at 37°C was needed for transient transfection, followed by 18-hour incubation in DMEM with 10% FBS for recovery. The transfected cells were then partially synchronized in DMEM with 0.2% FBS for 24 hours. The resulting cells were then treated with various anti-cancer agents in the same culture medium as described in the figure legends. After 24 hours, the treated cells were collected and lysed using Reporter Lysis Buffer (Promega, Madison, WI). The resulting cell lysates were assayed for luciferase activity using Luciferase Assay Kit (Promega, Madison, WI, USA) and TD-20/20 Luminometer (Turner Designs, Sunnyvale, CA, USA). β-Galactosidase activity was measured using chlorophenol red-β-D-galactopyranoside (CPRG) (Sigma-Aldrich, St. Louis, MO, USA) as substrate.
Each luciferase activity driven by a specific proximal 5'-upstream region of the p27 gene was normalized to β-galactosidase activity, a control for transfection efficiency. Since certain nutritional and chemopreventive anti-cancer agents could sometimes stimulate the normalized luciferase activity of empty luciferase reporter that do not contain any insert of the proximal 5'upstream region of the p27 gene, a special formula [9,10] was used in these exceptional cases to correct for this false increase in the relative luciferase activity. With human breast cancer cell lines used in this study, we have not encountered any such exceptional cases.

Western Immunoblot Analysis
Western immunoblot analysis of the upstream molecular signaling pathways of p27 expression was performed using estrogen receptor (ER) -negative MDA-MB-231 human breast cancer cells in vitro. The analysis was performed without either transfecting the cells with proximal 5'-upstream region of p27 gene-luciferase reporter plasmid or adding growth factors to stimulate the proliferation of the cells.
The cells were first seeded at a density of 5.5 × 10 6 cells/dish into a 100-mm tissue culture dish containing 10 mL of DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2% L-glutamine, and antibiotic/antimycotic solution and incubated at 37°C in a 5% CO 2 humidified chamber for 24 hours. After 24 hours, the cells were partially synchronized for another 24 hours in DMEM containing 0.2% FBS. Then, the cells in the 0.2% FBS-DMEM culture medium were treated with vehicle (DMSO), tamoxifen, 4-hydroxytamoxifen, dexamethasone, all-trans-retinoic acid (atRA), or 9cis-retinoic acid (9cRA) for another 24 hours. After 24 hours, the cells were washed twice with cold 1× PBS and scraped in 1× RIPA Lysis Buffer (Santa Cruz Biotechnology, Santa Cruz, CA, USA) containing phenylmethylsulphonyl fluoride (PMSF), protease inhibitor cocktail and sodium orthovanadate, and supplemented with 50 mM NaF. The cells were then sonicated and the supernatant was collected by centrifugation and stored at -80°C.
The supernatants (50 μg protein/lane) were applied to the SDS-PAGE and, after fractionation, proteins were transferred to nitrocellulose membrane, which was then blocked and incubated in a solution containing first primary antibody. After shaking overnight at 4°C, the target proteins bound to the first primary antibody were further treated with alkaline phosphatase (AP)-conjugated secondary anti-immunoglobulin antibody and detected by chemiluminescence using TROPIX Western-Star Kit (Applied Biosystems, Foster City, CA, U.S. A.). After exposure to X-ray film, the blots were stripped using Western Re-Probe solution (G-Biosciences, St. Louis, MO, U.S.A.), checked for removal of the chemiluminescence and then re-probed with second primary antibody.
Densitometric measurement of the intensity of the bands on the X-ray film was performed using UN-SCAN-IT Gel & Graph Digitizing Software Version 6.1 (Silk Scientific Corporation, Orem, UT, U.S.A.). Background corrections were done by four corner interpolation and optical density calculations were performed using linear standard reflective scan method.

Statistical Analysis
All of the significant P values were between 0.01 and 0.05. So, the results with P values less than 0.05 are simply indicated as asterisk on top of the vertical bars. The statistical significance information for the regression analysis, however, was provided in more detail in the panel Figure 2c.