Preliminary studies using in vivo model of MNU-induced rat mammary cancer and in vitro model of cultured cells had suggested – but not proved – that various nutritional and chemopreventive anti-cancer agents, including moderate dietary restriction, up-regulated the expression of p27 (Fig. 1a).

To prove or disprove this preliminary observation, each luciferase reporter vector containing proximal 5'-upstream region of the cyclin D1, cyclin A, p27 (p27-Kpn I) or p21 genes (Fig. 1b) was transiently transfected into promotion-sensitive (P+) JB6 mouse epidermal cells and then treated with all-trans-retinoic acid (atRA), 9-cis-retinoic acid (9cRA), 13-cis-retinoic acid (13cRA), phorbol 12-myristate 13-acetate (TPA), 1α, 25-dihydroxyvitamin D3 (calcitriol), or dexamethasone. Phorbol 12-myristate 13-acetate (TPA) is not a chemopreventive anti-cancer agent; rather it is a tumor promoter. But TPA was included here to demonstrate that it could stimulate the activity of the proximal 5'-upstream region (-1745) of cyclin D1. The TPA, three retinoic acids and dexamethasone exerted spurious effects on the backbone of the empty luciferase reporters when JB6 cells were used. Therefore, a special method, as described in the Methods and Materials section, was used to correct these effects when JB6 cells were exposed to these compounds.

The proximal 5'-upstream region (-1745) of cyclin D1 was activated only by TPA (Fig. 1c). Using a wild-type -963 cyclin D1 and a -963 cyclin D1 mutated at AP-1, it was confirmed that TPA activated the proximal 5'-upstream region (-1745) of cyclin D1 gene primarily through its TPA-response element (TRE).

The proximal 5'-upstream region of cyclin A and p21 genes were not activated by any of the compounds tested (Fig. 1d and 1f).

In contrast, the proximal 5'-upstream region (-1797) of p27 gene (p27-Kpn I) was activated by four nutritional and chemopreventive anti-cancer agents, namely all-trans-retinoic acid (atRA), 9-cis-retinoic acid (9cRA), 13-cis-retinoic acid (13cRA) and dexamethasone (Fig. 1e).

To investigate whether this specific activation of the proximal 5'-upstream region (-1797) of p27 gene (p27-Kpn I) recapitulates breast cancer preventive activity of various nutritional and chemopreventive anti-cancer agents, -1797 p27 (p27-Kpn I) was transfected into three different human breast cancer cell lines – estrogen receptor (ER)-positive MCF7 (Fig. 2a), ER-negative MDA-MB-231 (Fig. 2b), and ER-negative AU565 (Fig. 2c) – and then exposed to the following eighteen different compounds for 24 hours: 4-hydroxytamoxifen, tamoxifen, 17β-estradiol, ICI 182 780, genistein, genistin, daidzein, epigallocatechin-3-gallate, epigallocatechin, resveratrol, curcumin, taxifolin, mifepristone (RU486), all-trans-retinoic acid (atRA), 9-cis-retinoic acid (9cRA), 13-cis-retinoic acid (13cRA), 1α, 25-dihydroxyvitamin D3 (calcitriol), or dexamethasone. Preliminary studies indicated that none of these compounds exerted any spurious effects on the backbone of the empty luciferase reporter when human breast cancer cells were used.

4-Hydroxytamoxifen – but not tamoxifen – is the ultimate cancer preventive agents. Our results showed that 4-hydroxytamoxifen – but not tamoxifen – activated -1797 p27 in ER-positive MCF7 (Fig. 2a) and ER-negative MDA-MB-231 (Fig. 2b). These results indicated that (a) the activity of -1797 p27 recapitulated the differential breast cancer preventive efficacy of 4-hydroxytamoxifen and tamoxifen and that (b) the estrogen receptor (ER) was not involved in the activation of -1797 p27. In AU565 cells, both 4-hydroxytamoxifen and tamoxifen activated -1797 p27 (Fig. 2c), suggesting that either tamoxifen was converted to 4-hydroxytamoxifen in these cells or, as the results presented below suggest, the global rate of transcription was generally reduced in these cells, which, in turn, activated normally inactive tamoxifen by some unknown mechanisms.

Similarly, genistein – but not genistin – from soybeans is the ultimate cancer preventive agents. Our results showed that genistein – but not genistin – activated -1797 p27 in MCF7 (Fig. 2a) and MDA-MB-231 cells (Fig. 2b). These results again indicated that (a) the activity of -1797 p27 recapitulated the differential breast cancer preventive efficacy of genistein and genistin and that (b) the estrogen receptor (ER) was not involved in the activation of -1797 p27. In AU565 cells (Fig. 2c), however, both genistein and genistin activated -1797 p27 suggesting again that either genistin was converted to genistein in AU565 cells or, as the results presented below suggest, the global rate of transcription was reduced in these cells, which, in turn, activated normally inactive genistin by some unknown mechanisms.

Daidzein from soybeans activated -1797 p27 in all three cell lines (Fig. 2a,b,c).

Epigallocatechin – but not epigallocatechin-3-gallate – from green tea activated -1797 p27 in MCF7 cells (Fig. 2a), but neither epigallocatechin nor epigallocatechin-3-gallate activated -1797 p27 in MDA-MB-231 cells (Fig. 2b). In AU565 cells (Fig. 2c), both epigallocatechin and epigallocatechin-3-gallate activated -1797 p27. Resveratrol from grape skin did not activate -1797 p27 in MCF7 cells, but it did in MDA-MB-231 and AU565 cells. Curcumin from curry spice and taxifolin from citrus activated -1797 p27 in MCF7 and AU565 cells, but neither curcumin nor taxifolin activated -1797 p27 in MDA-MB-231 cells.

Of the three different forms of retinoic acid tested, 9-cis-retinoic acid (9cRA) most strongly activated -1797 p27, followed by all-trans-retinoic acid (atRA) and 13-cis-retinoic acid (13cRA) in all three human breast cancer cell lines. In JB6 mouse epidermal cells, these retinoic acids almost equally activated -1797 p27 (Fig. 1e). These results are compatible with those obtained using in vivo experimental animal models of breast cancer.

Dexamethasone activated -1797 p27 in all three human breast cancer cell lines (Fig. 2a,b,c).

Mifepristone (RU486) and 1α, 25-dihydroxyvitamin D3 (calcitriol) did not activate -1797 p27 in all three human breast cancer cell lines (Fig. 2a,b,c).

The estrogen receptor (ER)-negative AU565 cells were unusual in that sixteen of the eighteen compounds tested activated -1797 p27 (Fig. 2c); only mifepristone (RU486) and 1α, 25-dihydroxyvitamin D3 (calcitriol) did not activate it. This unusually high cancer preventive activity of nutritional and chemopreventive anti-cancer agents in AU565 cells might be due to its potentially reduced rate of transcription.

In summary, with the possible exception of AU565 cells, (a) activity of -1797 p27 in either estrogen receptor (ER)-positive or -negative human breast cancer cells fairly faithfully recapitulated the breast cancer preventive activity in vivo of the various nutritional and chemopreventive anti-cancer agents and (b) the effective doses for the activation of -1797 p27 by these agents were within the range that had been found effective for in vivo rat models of breast cancer.

To determine the core element for the activation of the proximal 5'-upstream region (-1797) of p27 gene, (P+) JB6 mouse epidermal cells or estrogen receptor (ER)-negative MDA-MB-231 human breast cancer cells were transfected with the following deletion mutants of -1797 p27 (Fig. 3a) and then treated with various nutritional and chemopreventive anti-cancer agents: -1797 p27 (p27-Kpn I) 4, -774 p27 (p27-Apa I) 4, -575 p27 (p27-5'UTR) 56, -435 p27 (p27-MB) 4, and -417 p27 (p27-IRES) 56. Preliminary studies indicated that TPA, three retinoic acids and dexamethasone exerted spurious effects on the backbone of the empty luciferase reporters when JB6 cells were used. Therefore, special method, as described in the Methods and Materials section, was used to normalize these effects in these cases. When human breast cancer cells were used, none of the agents tested exerted any spurious effects on the backbone of the empty luciferase reporters.

The results of deletion analysis suggested that, in all cases without exception (Fig. 3b to 3e), including those not shown here, the various nutritional and chemopreventive anti-cancer agents activated proximal 5'-upstream region (-1797) of p27 gene at least through -575 p27 (5'-untranslated region (5'UTR) of p27 gene). When the regions shorter than -575 p27 (p27-5'UTR) were tested, the activities tended to either stay constant or be reduced; the activities never increased above that of -575 p27 (p27-5'UTR).

To investigate if -575 p27 (p27-5'UTR) contains any cryptic transcription factor binding sites, the activity of the region was stimulated with 4-hydroxytamoxifen in the presence of an antibiotic actinomycin D. Figure 4a shows the schematic drawing of the pGL3 control-p27-5'UTR-luciferase reporter plasmid. This plasmid – pGL3 control – contains SV40 promoter in its backbone. Preliminary tests using pGL3 control without p27-5'UTR insert had demonstrated that 4-hydroxytamoxifen, tamoxifen, or vehicle (DMSO) did not exert any spurious effects on the SV40 promoter when human breast cancer cells were used. The outline of the protocol for actinomycin D experiment is shown in Figure 4b. Actinomycin D was added one hour before the addition of 4-hydroxytamoxifen, tamoxifen, or vehicle (DMSO) and then the cells were kept exposed to both compounds for another 24 hours.

The results (Fig. 4c) indicated that in the absence of actinomycin D, only 4-hydroxytamoxifen significantly up-regulated the activity of -575 p27 (p27-5'UTR) above that of vehicle (DMSO) in MDA-MB-231 cells; tamoxifen did not up-regulate it. The addition of actinomycin D in the presence of vehicle (DMSO) alone decreased the baseline activity of -575 p27 (p27-5'UTR) by about 53% relative to the activity observed in the absence of actinomycin D. Despite this decrease in the baseline activity in the presence of actinomycin D, 4-hydroxytamoxifen still significantly up-regulated the activity of -575 p27 (p27-5'UTR) above that of vehicle (DMSO). These findings suggested that transcriptional mechanisms were not involved in the up-regulation of the activity of -575 p27 (p27-5'UTR) by 4-hydroxytamoxifen, precluding the involvement of any cryptic transcription 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, significantly up-regulated the activity of -575 p27 (p27-5'UTR) in the presence of actinomycin D, suggesting that the overall level of global transcriptional rate might somehow modulate the effects of tamoxifen on the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells.

The results presented above suggested that the estrogen receptor was not involved in the activation of -1797 p27 (p27-Kpn I) by 4-hydroxytamoxifen, genistein, daidzein and probably other nutritional and chemopreventive anti-cancer agents. Tamoxifen and genistein have been known to exhibit anti-estrogenic activity, but in addition, they have been reported to inhibit receptor protein tyrosine kinase (RPTK) activity at a slightly higher concentrations 910. Epigallocatechin-3-gallate has also been reported to block activation of RPTKs such as epidermal growth factor receptor (EGFR) and HER-2/neu receptor, which are generally overexpressed or constitutively active in many human malignancies 1112.

Although multiple RPTKs are known to be expressed in human breast cancer cells, synthetic inhibitors of four RPTKs – epidermal growth factor receptor (EGFR), HER/ErbB, platelet-derived growth factor receptor (PDGFR) and insulin receptors (IR and IGF-1R) – were used to investigate whether they up-regulate the activity of -575 p27 (p27-5'UTR). Preliminary studies had demonstrated again that none of them exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in all four types of cells used in this experiment.

The following four synthetic inhibitors were used to inhibit EGFR: AG9 (inactive negative control for inhibition of EGFR), AG18 (broad spectrum protein tyrosine kinase inhibitor, inhibits EGFR autophosphorylation), AG1478 (specific inhibitor of EGFR), and PD153035 (specific inhibitor of tyrosine kinase activity of EGFR). Of the four inhibitors, two of them – AG18 and AG1478 – up-regulated the activity of -575 p27 (p27-5'UTR) in MCF7 cells (Fig. 5b). However, none of them up-regulated the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 (Fig. 5a), AU565 (Fig. 5c), and JB6 cells (Fig. 5d).

AG30, a specific inhibitor of c-ErbB, did not up-regulate the activity of -575 p27 (p27-5'UTR) in any of the cells tested (Fig. 5a to 5d).

In contrast, AG1295, a specific inhibitor of PDGFR, up-regulated the activity of -575 p27 (p27-5'UTR) in all four types of cells (Fig. 5a to 5d).

Three inhibitors of insulin receptors were investigated using MDA-MB-231 cells (Fig. 5a). Of the three inhibitors, two of them – IGF-IR inhibitor PPP and AGL2263 (IR and IGF-1R inhibitor) – up-regulated the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells, but AG1024 (IGF-1 inhibitor) failed to up-regulate it.

Taken together, these results suggested that inhibition of certain RPTKs on the cell surface could up-regulate the activity of -575 p27 (p27-5'UTR).

When the cell-surface RPTKs are inhibited, the inhibitory signal could be transmitted to the interior of the cells in two ways. The inhibition of mitogen-activated protein kinases (MAPKs) is one of them. Therefore, inhibitors of MAPKs were used to investigate whether any one or more of them could up-regulate the activity of -575 p27 (p27-5'UTR). Again, none of them exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in all four types of cells used in this experiment.

PD98059, an inhibitor of MEK, up-regulated the activity of -575 p27 (p27-5'UTR) in all four types of cells tested (Fig. 6a). The effects of other inhibitors of MAPKs were investigated using MDA-MB-231 cells (Fig. 6b). Of the two inhibitors of ERK tested, ERK activation inhibitor peptide I up-regulated the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells, but ERK activation inhibitor peptide II did not up-regulate it. And of the four p38MAPK inhibitors, only SB202190 strongly up-regulated the activity of -575 p27 (p27-5'UTR); the other three inhibitors – PD169316, SB203580 and SB202474, an inactive negative control for p38MAPK inhibitors – failed to up-regulate it. Anisomycin, which activates MAPKs and stress-activated protein kinases (SAPKs), also slightly up-regulated the activity of -575 p27 (p27-5'UTR), and the other two inhibitors – Ro-32-0432 (protein kinase C (PKC) inhibitor) and hypericin (PKC/ERK inhibitor) – did not have any significant effects on the activity of -575 p27 (p27-5'UTR).

The signal that the cell-surface RPTKs are inhibited is also transmitted to the interior of the cells by phosphoinositide 3-kinase/Akt/mammalian target of rapamycin (PI3K/Akt/mTOR) pathway. Therefore, inhibitors of these three protein kinases were used next to investigate whether they also up-regulate the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells. Again, preliminary experiments had been conducted to verify that none of them had exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in MDA-MB-231 cells.

The results indicated that LY294,002 (an inhibitor of PI3K) 5, triciribine (Akt inhibitor) and rapamycin (mTOR inhibitor) all three of them up-regulated the activity of -575 p27 (p27-5'UTR) (Fig. 7).

Also shown in Figure 7 are the effects of two cyclooxygenase inhibitors – COX-1 inhibitor FR122047 and COX-2 inhibitor II – both of which failed to up-regulate the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells.

There are two ways to suppress the global cap-dependent translation initiation of 5'-m⁷G-capped mRNAs, thereby potentially up-regulating the activity of -575 p27 (p27-5'UTR) by a set of mechanisms called cap-independent translation initiation. One way is to inhibit the methylation of 5'-m⁷G cap of mRNAs by S-(5'-adenosyl)-L-methionine (AdoMet or SAM). Another way is to stimulate the inhibitory effect of endoplasmic reticulum stress by increasing the phosphorylation of the eukaryotic translation initiation factor 2α (eIF2α). The results (Fig. 7) indicated that NSC 119889, a cell-permeable, competitive inhibitor of AdoMet (SAM) and which also acts as a global inhibitor of cap-dependent translation initiation, up-regulated the activity of -575 p27 (p27-5'UTR) in estrogen receptor (ER)-negative MDA-MB-231 cells. The salubrinal, a cell-permeable thiourea compound, that acts as a selective inhibitor of eukaryotic translation initiation factor 2α (eIF2α) dephosphorylation by phosphatase complex, failed to up-regulate the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells (Fig. 7). Again, both inhibitors had not exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in MDA-MB-231.

5'-AMP-activated protein kinase, known as AMPK, acts as a metabolic energy sensor or cellular fuel gauge playing a key role in the regulation of energy metabolism 13. A number of physiological and pathophysiological stimuli, including caloric restriction, lead to an increase in the AMP: ATP ratio within the cell, which activates AMPK by phosphorylating it at Thr172 in the activation loop. When activated, AMPK modulates tuberous sclerosis complex (TSC), down-regulates phosphorylation of mammalian target of rapamycin (mTOR), eukaryotic translation initiation 4E binding protein 1 (4EBP1) and p70 S6 kinase (S6K) 131430. Thus, caloric restriction – by way of AMPK, TSC, mTOR, and 4EBP1 – might potentially up-regulate the activity of -575 p27 (p27-5'UTR).

The Figure 8 shows the effects of various modulators of AMPK phosphorylation on the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells. Again, none of the compounds tested had exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in MDA-MB-231 cells. The results indicated that two compounds – rotenone (an inhibitor of mitochondrial electron transport) and AICA riboside (an analogue of 5'-AMP) – that are known to increase phosphorylation of AMPK up-regulated the activity of -575 p27 (p27-5'UTR) (also see a recent report 35). The results shown in Figure 9 also demonstrated that the removal of D-(+)-glucose from the cell culture medium up-regulated the activity of -575 p27 (p27-5'UTR). In contrast, the compounds that are known to decrease phosphorylation of AMPK – excess D-(+)-glucose and compound C – down-regulated the activity of -575 p27 (p27-5'UTR).

Metformin is a widely used drug for treatment of type 2 diabetes with no defined cellular mechanism of action. It has been suggested recently that metformin could increase phosphorylation of AMPK 153839. However, the results, as shown in Figure 8, indicated that metformin did not up-regulate the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells.

In addition to caloric restriction and growth factor signals, tuberous sclerosis complex (TSC) could transmit amino acid deficiency signals to regulate mTOR activity 13. We already demonstrated that rapamycin, an inhibitor of mTOR phosphorylation, up-regulated the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells (see Fig. 7). Therefore, the effects of amino acid deficiency on the activity of -575 p27 (p27-5'UTR) were investigated using MDA-MB-231 cells. Again, none of the amino acid deficiencies tested had exerted any spurious effects on the backbone of the empty luciferase reporter plasmids in MDA-MB-231 cells. As shown in Figure 9, removal of L-leucine, L-methionine, L-cysteine, or combination of L-methionine and L-cysteine, all up-regulated the activity of -575 p27 (p27-5'UTR) in MDA-MB-231 cells. The findings of L-methionine deficiency are interesting because L-methionine deficiency could up-regulate the activity of -575 p27 (p27-5'UTR) in two ways: one is to decrease methylation of 5'-m⁷G cap of mRNAs and another is to decrease phosphorylation of mTOR by TSC.

Caloric restriction could up-regulate translation initiation of p27 mRNA through its 5'-untranslated region (p27-5'UTR) by sending the signal to endoplasmic reticulum via AMPK, TSC, mTOR, and 4EBP/S6K. Amino acid deficiencies could also send the signals to endoplasmic reticulum via TSC, mTOR, and 4EBP/S6K. Deficiency of L-methionine is an interesting case because it could also up-regulate the cap-independent translation initiation of p27 mRNA by down-regulating global methylation of the 5'-m⁷G-cap of other mRNAs.

Caloric restriction has long been known to activate AMPK. The AMPK system is controlled by the balance between ATP consumption (e.g., by biosynthesis, cell growth, or muscle contraction) and ATP production via catabolism 2122. If the rate of ATP consumption exceeds its rate of production, such as during caloric restriction, ADP will tend to rise and be converted to AMP by the enzyme adenylate kinase. The rise in level of the activating ligand AMP, coupled with the fall in level of the inhibitory nucleotide ATP, activates AMPK, which then switches off ATP-consuming processes and switches on catabolism in an attempt to redress the balance.

AMPK, when activated, phosphorylates tuberous sclerosis complex 2 (TSC2), thereby inhibiting mTOR activation 13. In addition to changes in the intracellular AMP/ATP ratio, the TSC1 (hamartin)-TSC2 (tuberin) complexes might mediate amino acid signals to regulate mTOR activity 13.

In mammalian cells, mTOR generally regulates translation. Eukaryotic translation initiation factor 4E (eIF4E) binding protein 1 (4EBP1) 30 and ribosomal p70 S6 kinase (S6K), the most extensively studied substrates of mTOR, are key regulators of protein translation 13.

4EBP1 acts as a translational repressor by binding and inhibiting the eIF4E, which recognizes the 5'-end m⁷G cap of eukaryotic mRNAs. Phosphorylation of 4EBP1 by mTOR results in a dissociation of 4EBP1 from eIF4E, thereby relieving the inhibition of 4EBP1 on eIF4E-dependent (cap-dependent) translation initiation 30.

The inhibition of mTOR, therefore, results in decreased global cap-dependent translation initiation of 5'-m⁷G-capped mRNA, but it could also increase cap-independent translation initiation of p27 mRNA through its 5'UTR.

Following growth factor activation of RPTKs, phosphoinositide 3-kinase (PI3K) is recruited to the receptor and activated resulting in the production of phosphatidylinositol-3,4,5-trisphosphate (PIP₃) 23. This recruits Akt/PKB to the membrane where it is phosphorylated by phosphoinositide-dependent kinase 1 (PDK1). Akt/PKB is then released from the membrane and translocate to other subcellular compartments.

Akt/PKB is involved in mTOR activation by phosphorylating mTOR at Ser2448 1324. It is not yet settled whether Akt/PKB activates mTOR directly or indirectly, but recent biochemical studies indicated that Akt/PKB directly phosphorylates TSC2 and inhibits its function 13. TSC2 inactivation by Akt/PKB may also inhibit mTOR indirectly through inhibition of the small GTPase, Rheb 13.

The inhibition of RPTKs, therefore, leads to reduction of the phosphorylation of mTOR and 4EBP/S6K, thereby attenuating the global cap-dependent translation initiation of 5'-m⁷G-capped mRNAs, but at the same time activating the cap-independent translation initiation of p27 mRNA through its 5'UTR.

The inhibitory signals originating from RPTKs could also be transmitted to MAPK/MNK signaling pathway. It is known that the activity of eIF4E is regulated not only by interaction with 4EBP1 but also phosphorylation by mitogen-activated protein (MAP) kinase-interacting kinase (MNK) on Ser209. The phosphorylation of eIF4E via MNK is mediated by the activation of either the ERK or p38 pathway 25. The results presented above indicated that the inhibition of the MEK (an upstream MAPK of ERK), ERK or p38MAPK could also decrease the phosphorylation of eIF4E, thereby reducing the global cap-dependent translation initiation of 5'-m⁷G-capped mRNAs, but at the same time activating the cap-independent translation initiation of p27 mRNA through its 5'UTR.

Nearly all mRNAs are post-transcriptionally modified at their 5' and 3' ends, by capping and polyadenylation, respectively 262728. The m⁷G-capping at their 5' end protects the nascent pre-mRNAs against degradation and failure to cap or loss of cap leads to rapid breakdown of the mRNAs. The mRNA cap (guanine-N7) methyltransferase catalyzes methyl transfer from S-adenosylmethionine (AdoMet or SAM) to GpppRNA to form m⁷GpppRNA.

The results presented above indicated that the NSC 119889, a cell-permeable, competitive inhibitor of AdoMet (SAM), inhibits global cap-dependent translation initiation of 5'-m⁷G-capped mRNAs, but it could also increase cap-independent translation initiation of p27 mRNA through its 5'UTR. This finding suggests that the epigenetic methylation hypothesis of cancer should be based not only on DNA methylation but also on mRNA methylation.

Phosphorylation of the α subunit of eukaryotic translation initiation factor 2 (eIF2α) is a well-documented mechanism of down-regulating protein synthesis under a variety of stress conditions, but at the same time it could also up-regulate the cap-independent translation initiation of p27 mRNA through its 5'UTR. However, contrary to this expectation, the results presented above indicated that salubrinal, a cell-permeable thiourea compound, that acts as a selective inhibitor of translation initiation factor 2α (eIF2α) dephosphorylation by phosphatase complex, failed to up-regulate the activity of 5'-untranslated region of p27 gene.

Several recent papers have reported that there is a network of either synergistic or antagonistic cross-talk between 5'-AMP-activated protein kinase (AMPK) and receptor protein tyrosine kinase (RPTK) pathways, e.g. Akt 31323740, IGF-1 33, insulin signaling downstream of PKB 36, MEK 34, and c-Raf 43. Additionally, a cross-talks between AMPK and eukaryotic translation initiation factor 2α (eIF2α) 41 and between c-Raf-1 and Akt 42 were reported. These reports, if confirmed, would suggest that the molecular signaling pathways of how p27 expression might be regulated could be a "tangled web" 42 to say the least.

All-trans-retinoic acid (atRA), 9-cis-retinoic acid (9cRA), 13-cis-retinoic acid (13cRA), phorbol 12-myristate 13-acetate (TPA), 1α, 25-dihydroxyvitamin D3 (calcitriol), dexamethasone, 4-hydroxytamoxifen, tamoxifen, 17β-estradiol, genistein, genistin, daidzein, epigallocatechin-3-gallate, epigallocatechin, resveratrol, curcumin, taxifolin, mifepristone (RU486), actinomycin D, anysomycin, Ro-32-0432, hypericin, LY 294002, rapamycin, D-(+)-glucose, and rotenone were obtained from Sigma-Aldrich (St. Louis, MO, USA). ICI 182 780 was purchased from TOCRIS (Ellisville, MO, USA). Triciribine, NSC 119889, AG9, AG18, AG30, AG1024, AG1295, AG1478, AGL2263, PD98059, PD153035, PD169316, SB202190, SB202474, SB203580, IGF-1R inhibitor PPP, ERK activation inhibitor peptide I, ERK activation inhibitor peptide II, COX-1 inhibitor FR122047, COX-2 inhibitor II, salubrinal, compound C, and metformin were obtained from Calbiochem/EMD (San Diego, CA, USA). AICA riboside was purchased from Phoenix Pharmaceuticals, Inc. (Belmont, CA, USA). Dulbecco's Modified Eagle's Medium (DMEM) labeling kit was obtained from Chemicon International (Temecula, CA, USA).

Promotion-sensitive (P+, clone 41.5a) JB6 mouse epidermal cell line was kindly provided by Dr. N. H. Colburn (National Cancer Institute, Frederick, MD, USA). JB6 cells were grown in Eagle's Minimum Essential Medium (MEM) supplemented with 5% heat-inactivated fetal bovine serum (FBS), 2% L-glutamine, and antibiotics. Human breast cancer cell lines, MCF7, MDA-MB-231, and AU565, were obtained 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% heat-inactivated FBS, 100 mg/L recombinant human insulin, 2% L-glutamine, and antibiotics. MDA-MB-231 and AU565 cells were grown in the same culture medium without insulin. The incubation was carried out at 37°C in a 5% CO₂ humidified chamber. All cells were subcultured after trypsinization with 0.05% trypsin-0.02% EDTA solution. The cultures were always maintained below. The cells were checked periodically for mycoplasmal infection by DNA fluorochrome staining.

Luciferase reporter plasmids containing one of the following proximal 5'-upstream region of the genes were used to transfect the cells: -1745 cyclin D1 2, -963 cyclin D1 2, -963 AP-1mut cyclin D1 2, -8100 cyclin A 3, -1797 p27 (p27-Kpn I) 4, -774 p27 (p27-Apa I) 4, -575 p27 (p27-5'UTR) 56, -435 p27 (p27-MB) 4, -417 p27 (p27-IRES) 56, and -2320 p21 7. The control luciferase reporter plasmids that do 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. We found that none of the nutritional and chemopreventive anti-cancer agents tested did not exert spurious effects on the backbone in the human breast cancer cells, but TPA, the three retinoic acids and dexamethasone did stimulate the empty luciferase reporter plasmids in mouse epidermal JB6 cells. Therefore, a proper transfection protocols as described below at the end of the section of Transfection and Luciferase Assay were devised when JB6 cells were exposed to TPA, the three retinoic acids and dexamethasone.

Transfections were performed according to a published protocol 8 using FuGENE 6 from Roche Applied Science (Indianapolis, IN, USA). In brief, 24 hours before reporter transfection, the cells were seeded into a 60-mm tissue culture dish at a density of 1.5 × 10⁵ cells/dish and incubated at 37°C in a 5% CO₂ humidified chamber. Reporter transfection was then 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 MEM or 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 either MEM with 5% FBS or DMEM with 10% FBS for recovery. The transfected cells were then starved in either MEM or DMEM with 0.2% FBS for 24 hours. The resulting cells were treated with various compounds 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 a substrate.

Each luciferase activity driven by a specific proximal 5'-upstream region of the genes was normalized to β-galactosidase activity, a control for transfection efficiency. Since, depending on the cell types (e.g., JB6), certain nutritional and chemopreventive agents (e.g., the three retinoic acids and dexamethasone) and a tumor promoter (e.g., TPA) stimulated the normalized luciferase activity of empty luciferase reporter vector that do not contain an insert of the specific proximal 5'-upstream region of the genes, the following formula was used in these exceptional cases to correct for this false increase in the relative luciferase activity:

Relative luciferase activity (%) = (Experimental luciferase activity/Control luciferase activity) × 100,

where,

Experimental luciferase activity = {Test compound/None} {Luciferase reporter vector containing a specific promoter insert},

Control luciferase activity = {Test compound/None} {Luciferase reporter vector

NOT

Test compound/None = [Luc(Test)/βGal(Test)]/[Luc(None)/βGal(None)].

An experimental value with statistical significance of P ≤ .05 as compared to control (vehicle alone) by t test is indicated as * on top of the vertical bar.