Cell Cycle Progression and Proliferation Despite 4BP-1 Dephosphorylation

Cell culture.

BC3H1 and RR cells were grown in Dulbecco’s modified essential medium (DMEM) with the addition of 20% fetal bovine serum (FBS) and 1% penicillin-streptomycin. The medium was changed every 48 to 72 h. Rapamycin was added directly to the medium. The cells were cultured for multiple passages in the presence of high concentrations of rapamycin (0.1 to 1 μM) followed by dilutional cloning. Two independent clones were utilized; RR-1 clones were grown in 1 μM rapamycin, and RR-3 clones were grown in 0.1 μM rapamycin. In all studies, rapamycin was removed 1 week prior to the experiment. MELC were grown in minimal essential medium-α (MEM-α) supplemented with 10% FBS (inactivated) and 1% penicillin-streptomycin (38). CHO cells (obtained from the American Type Culture Collection) were grown in Ham’s F-12 medium supplemented with 10% FBS and 1% penicillin-streptomycin. Deprivation and restoration of amino acids were performed as previously described (48).

Cell proliferation assays.

BC3H1, RR-1, and RR-3 cells (25 × 10⁴) were grown in DMEM plus 20% FBS (in triplicate). Rapamycin (100 nM) was added directly to the medium. After 48 h, the cell number was determined with a Coulter Counter. MELC (10 × 10⁴) were grown in MEM plus 10% FBS; rapamycin (0.2 to 1 μM) or a vehicle was added directly to the medium. After 4 days, the cells were counted with a Coulter Counter.

[³H]leucine incorporation.

BC3H1 and RR-1 cells (10⁴) were plated in triplicate in 12-well dishes. After 48 h, rapamycin (1 μM) was added to the appropriate wells, and cells were pulsed with 1 μCi of [³H]leucine per well. [³H]leucine incorporation into protein was determined by precipitation with trichloroacetic acid. Experiments were repeated three times.

Fluorescence-activated cell sorter analysis.

BC3H1 and RR-1 cells were placed in DMEM plus 1% FBS for 24 h. The cells were then stimulated with DMEM plus 20% FBS and treated with either rapamycin (100 nM) or the vehicle. The cells were washed and harvested after 24 h and labeled with propidium iodide solution-RNase for 1 h. The cells were analyzed on a fluorescence-activated cell sorter, with a minimum of 15,000 cells counted.

Western blot analyses.

To detect 4BP-1, eIF-4E, and eIF-4G, lysates were prepared as previously described (23). Parental BC3H1 cells and RR cells, MELC, and CHO cells were grown in DMEM plus 20% FBS, MEM plus 10% FBS (inactivated), and Ham’s F-12 medium plus 10% FBS, respectively. Asynchronously dividing cells were treated with either the vehicle or rapamycin for 30 min, and cellular lysates were prepared. Protein was measured with the Bradford reagent, size fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose. Membranes were incubated as previously described (23) and visualized by enhanced chemiluminescence. For 4BP-1–eIF-4E complex experiments, cell extracts (130 μg) were incubated with m⁷-GTP–Sepharose in duplicate for 1 h at 4°C. The resin was washed and samples were size fractionated on SDS–8% (eIF-4G) or –13% (eIF-4E and 4BP-1) polyacrylamide gels, transferred to nitrocellulose, and blotted with eIF-4G (29), eIF-4E (Transduction Laboratories), or 4BP-1 (PHAS-I) (23) antibodies. For immunoprecipitation, cell extracts (100 μg) were incubated with 2 μl of anti-eIF-4G antibody (49) in 500 μl of lysis buffer, captured with protein A-Sepharose, washed and eluted with Laemmli buffer, and size fractionated on SDS–11.5% polyacrylamide gels.

Proliferation and cell cycle progression in RR cells.

RR and BC3H1 cells were selected on the basis of normal growth following multiple passages in the presence of high concentrations of rapamycin (either 100 or 1,000 nM). Proliferating RR and parental cells exhibited similar morphologies (Fig. 1A). Two independent clonal RR cell lines (RR-1 and RR-3) exhibited no growth arrest (Fig. 1B) and no decrease in viability (as assessed by trypan blue exclusion) in the presence of rapamycin. Rapamycin induced a delay in transition from G₁ to S phase in the parental cells but not in the RR cells (Fig. 1C).

Rapamycin inhibits activation of cyclin-dependent kinases (CDK) by preventing mitogen-induced down-regulation of the CDK inhibitor p27^kip1 (31). p27^kip1 protein regulation has been shown to be controlled at the translational level (18) and via ubiquitin-dependent degradation (32). Rapamycin resistance in RR cells results from a deficiency in p27^kip1 due to increased degradation of the protein via a ubiquitin-independent pathway (25). RR cells have constitutively low p27^kip1 levels, and unlike the case with parental cells, p27^kip1 levels are not regulated in response to mitogens or rapamycin (25). In response to serum withdrawal, RR cells undergo apoptosis because of their inability to extinguish hyperphosphorylation of retinoblastoma protein (pRb) (25).

Inhibition of 4BP-1 phosphorylation.

The ability of RR cells to proliferate in the presence of high concentrations of rapamycin provided the opportunity to determine whether inactivation of proteins thought to be required for translation interferes with cell cycle progression. Rapamycin has been previously demonstrated to minimally inhibit protein synthesis, if at all (11, 36). In both proliferating BC3H1 and RR-1 cells, rapamycin (1 μM) demonstrated no significant effect on [³H]leucine incorporation (Fig. 2A).

Rapamycin inhibited phosphorylation of both p70^s6k (data not shown and reference 25) and 4BP-1 (Fig. 2B) in proliferating RR cells. The inhibition of both p70^s6k and 4BP-1 phosphorylation in wild-type and RR cells was observed more than 48 h after removal of rapamycin from the culture medium (data not shown). Rapamycin is a potent inhibitor of cell growth; therefore, it has been difficult to determine whether the drug specifically inhibits translation or, conversely, whether the primary effect of rapamycin is to inhibit cell cycle progression (e.g., by up-regulating p27^kip1 [25, 31]). However, the ability of rapamycin to inhibit 4BP-1 (PHAS-1) phosphorylation and eIF-4E function has been interpreted as indicating that inhibition of translation is the mechanism underlying the growth-inhibitory properties of rapamycin (7). The rapamycin-induced inhibition of 4BP-1 in proliferating RR cells uniquely demonstrates that proliferation can proceed despite dephosphorylation of 4BP-1.

4BP-1–eIF-4E interaction.

Overexpression of eIF-4E increases the translation of mRNAs containing extensive secondary structures, including cyclin D1 (40, 41), ornithine decarboxylase (42), and c-myc (9), and causes transformation of fibroblasts (10, 22, 39). Immunoblot analyses showed that eIF-4E protein levels were equivalent in the parental and RR cells (Fig. 3A), indicating that the stoichiometry between 4BP-1 (Fig. 2B) and eIF-4E was unchanged in the RR cells. eIF-4E is constitutively bound to the 5′ cap, and its activity is regulated by the binding of the inhibitory protein 4BP-1 (23, 33). To determine whether 4BP-1 was bound to eIF-4E in RR cells, an affinity resin containing the 5′ cap homolog m⁷-GTP was used as previously described (23). In these experiments dephosphorylation of 4BP-1 caused 4BP-1 to bind to eIF-4E, which in turn was bound to the m⁷-GTP resin. RR cells cultured with rapamycin were able to proliferate despite persistent inactivation of eIF-4E by 4BP-1 (Fig. 3B). No significant differences in the amount of 4BP-1 binding to eIF-4E were seen in BC3H1 and RR cells treated with rapamycin.

Like RR cells, MELC exhibited no growth arrest after 4 days of rapamycin treatment (Fig. 4A). Rapamycin also inhibits p70^s6k in MELC, as previously reported (8). In proliferating MELC, rapamycin (0.2 μM) treatment for 1 h inhibited 4BP-1 phosphorylation (Fig. 4B) and increased the complex formation between 4BP-1 and eIF-4E as assessed by m⁷-GTP binding (Fig. 4C). Rapamycin had no effect on eIF-4E protein levels in MELC (data not shown).

Rapamycin does not affect eIF-4E–eIF-4G complex formation.

The binding of eIF-4G or 4BP-1 to eIF-4E is believed to be mutually exclusive (14, 36). Therefore, the increased association of eIF-4E and 4BP-1 would theoretically reduce the association of eIF-4E and eIF-4G (7, 14, 24, 47). Several groups have shown that increased eIF-4E–eIF-4G association was correlated with decreased 4BP-1–eIF-4E complex formation (12, 21). Moreover, rapamycin prevented amino acid-induced eIF-4E–eIF-4G complex formation in amino acid-depleted CHO cells (48). However, conflicting data have also been reported with regard to the effects of rapamycin on eIF-4E–eIF-4G interactions (20, 29). Morley and McKendrick demonstrated with NIH 3T3 cells that rapamycin increased the association of 4BP-1 and eIF-4E on m⁷-GTP resin but that rapamycin did not prevent serum-induced eIF-4E–eIF-4G interaction as assessed by immunoprecipitation assays (29). Moreover, they demonstrated that rapamycin does not inhibit the recruitment of eIF-4E to the ribosome and that prolonged incubation (for 20 h) causes only a 30 to 40% reduction in the amount of eIF-4E associated with eIF-4G (29). Furthermore, Beretta et al. (1) have shown a lack of temporal correlation between 4BP-1 dephosphorylation and inhibition of in vitro cap-dependent translation. Morley and McKendrick (29) and others (1) have suggested that these data may reflect a low rate of release of 4BP-1 from eIF-4E, such that eIF-4E is able to associate with 4BP-1 only after release from eIF-4G.

In CHO cells, rapamycin did not prevent the association of eIF-4E and eIF-4G as assessed by immunoprecipitation and binding to m⁷-GTP resin (Fig. 5A). However, as previously described (48), amino acid deprivation for 1 h did significantly reduce the association of eIF-4E and eIF-4G. Rapamycin (1 μM) did not prevent serum-induced association of eIF-4E and eIF-4G. Morley and McKendrick (29) suggested that there might be a population of eIF-4E that is inaccessible to dephosphorylated 4BP-1 and thus dissociation from eIF-4G. However, our findings that amino acid withdrawal rapidly induced dissociation of the eIF-4E–eIF-4G complex in cells in which serum withdrawal did not induce the same dissociation suggest that compartmentalization may not explain the lack of effect of rapamycin on this complex in both CHO cells and BC3H1 or RR cells.

In BC3H1 and RR cells, rapamycin (1 μM for 24 h) caused increased association of 4BP-1 with the m⁷-GTP resin but failed to inhibit the serum-induced association of eIF-4E and eIF-4G as assessed by binding to the m⁷-GTP (Fig. 5B) and by coimmunoprecipitation (Fig. 6). In RR cells exposed to rapamycin for more than 1 week, eIF-4G–eIF-4E association persisted despite increased 4BP-1 association with eIF-4E (data not shown), indicating that even prolonged exposure to rapamycin and 4BP-1 dephosphorylation does not induce eIF-4E–eIF-4G dissociation. Amino acid withdrawal in BC3H1 and RR cells caused increased association of 4BP-1 and eIF-4E on m⁷-GTP resin and markedly reduced eIF-4E–eIF-4G complex formation (Fig. 6). Rapamycin inhibited the amino acid-induced eIF-4E–eIF-4G complex formation in nutrient-deprived BC3H1 cells; however, rapamycin did not inhibit the serum-induced eIF-4E–eIF-4G complex formation in BC3H1 and RR cells (Fig. 6). Therefore, although growth factor (serum) withdrawal can lead to dephosphorylation of 4BP-1 and increased 4BP-1–eIF-4E association, in the absence of amino acid withdrawal, it is insufficient to cause eIF-4E–eIF-4G complex dissociation.

Conclusions.

The mechanism(s) by which rapamycin inhibits proliferation and induces G₁/S arrest have not been fully elucidated. Rapamycin induces a G₁-to-S-phase transition arrest and inhibits a mitogen-activated signaling pathway that involves mTOR, p70^s6k, p27^kip1, 4BP-1, eIF-4E, cell cycle kinases, and retinoblastoma protein (5, 7, 11, 23, 28, 33). However, the immediate downstream targets of mTOR and the significance of rapamycin’s inhibition of p70^s6k, 4BP-1, and eIF-4E remain uncertain. The observations by several groups that the inhibition of mTOR, p70^s6k, and 4BP-1 phosphorylation by rapamycin were coupled to growth arrest and to inactivation of eIF-4E led to the hypothesis that rapamycin’s antiproliferative properties are mediated via translational control (7, 13, 19, 37). In the present study, we used two RR muscle cell lines (RR-1 and RR-3) and a third hematopoietic cell line (MELC) that is also resistant to rapamycin and confirmed our findings with rapamycin-sensitive CHO cells. This strategy allowed us to address the question of whether cell cycle progression and translation can proceed despite dephosphorylation of 4BP-1 and inactivation of p70^s6k by rapamycin. We showed that rapamycin-mediated dephosphorylation and increased binding of eIF-4E to 4BP-1 are not required for rapamycin’s antiproliferative effects. Rapamycin does not cause the dissociation of the eIF-4E–eIF-4G complex in serum-stimulated cells; therefore, inhibition of protein translation does not appear to be the mechanism through which rapamycin exerts its antiproliferative effects. Rapamycin caused no significant reduction in protein synthesis (Fig. 2A) (11, 36), in contrast to amino acid withdrawal, which has a more profound effect on generalized protein synthesis (16). Although amino acid withdrawal and rapamycin cause similar degrees of dephosphorylation of p70^s6k and 4BP-1, presumably through inhibition of mTOR activity, they do not have similar effects on the association of eIF-4E and eIF-4G. This suggests that the mTOR pathway, which is regulated by amino acids and rapamycin, may play a greater role in p70^s6k inhibition and 4BP-1 phosphorylation than in eIF-4E–eIF-4G complex regulation.

The ability of rapamycin to inhibit cellular proliferation may have important applications in the treatment of disorders such as accelerated arteriopathy that occurs in transplanted hearts and restenosis following the placement of coronary stents (27). The present study indicates that the antiproliferative properties of rapamycin are probably not mediated by inhibition of protein translation via inactivation of p70^s6k or eIF-4E. While it appears that rapamycin does not inhibit cell growth via translational control, the CDK inhibitors, in particular p27^kip1, remain attractive candidates for mediators of the drug’s important antiproliferative properties.