Akt-mediated Liver Growth Promotes Induction of Cyclin E through a Novel Translational Mechanism and a p21-mediated Cell Cycle Arrest*

The control of hepatocyte growth is relevant to the processes of liver regeneration, development, metabolic homeostasis, and cancer. A key component of growth control is the protein kinase Akt, which acts downstream of mitogens and nutrients to affect protein translation and cell cycle progression. In this study, we found that transient transfection of activated Akt triggered a 3-4-fold increase in liver size within days but only minimal hepatocyte proliferation. Akt-induced liver growth was associated with marked up-regulation of cyclin E but not cyclin D1. Analysis of liver polyribosomes demonstrated that the post-transcriptional induction of cyclin E was associated with increased translational efficiency of this mRNA, suggesting that cell growth promotes expression of this protein through a translational mechanism that is distinct from the cyclin D-E2F pathway. Treatment of Akt-transfected mice with rapamycin only partially inhibited liver growth and did not prevent the induction of cyclin E protein, indicating that target of rapamycin activity is not necessary for this response. In the enlarged livers, cyclin E-Cdk2 complexes were present in high abundance but were inactive due to increased binding of p21 to these complexes. Akt transfection of p21-/- mice promoted liver growth, activation of Cdk2, and enhanced hepatocyte proliferation. In conclusion, growth promotes cyclin E expression through a novel translational mechanism in the liver, suggesting a new link between cell growth and the cell cycle machinery. Furthermore, p21 suppresses proliferation in the overgrown livers and may play a role in preventing cell cycle progression in response to organ size homeostatic mechanisms.

Although the terms are often used interchangeably, cell growth and proliferation reflect distinct processes. Cell growth refers to an increase in cell mass resulting from enhanced synthesis of proteins and other macromolecules, whereas proliferation refers to cell cycle progression resulting in increased cell number (1)(2)(3)(4)(5)(6). In many cases, extracellular stimuli, such as mitogens, promote both growth and proliferation, so that subsequent generations of cells maintain normal size. However, cell growth can be induced in the absence of proliferation, depending on the stimulus and target cell. The intracellular pathways that regulate growth are incompletely characterized, but several key mediators have been identified (1)(2)(3)(4)(5)(6). In particular, signaling through the phosphoinositide 3-kinase pathway appears to play an important role in growth regulation.
A major downstream target of phosphoinositide 3-kinase is the serine/threonine protein kinase Akt (also known as protein kinase B), which regulates diverse processes, including glucose homeostasis, transcription, apoptosis, cell motility, angiogenesis, proliferation, and growth (7)(8)(9)(10). Akt has been identified as an oncogene, and this kinase is frequently activated in malignant cells (11). Numerous studies have implicated Akt signaling as an important mediator of cell and organ growth. For example, interruption of Akt signaling impairs growth in Drosophila, whereas constitutive activation of Akt leads to enlarged organ size (12,13). In transgenic mice, overexpression of active forms of Akt in cardiomyocytes, islet cells, and skeletal muscle leads to cardiomegaly, increased pancreatic ␤ cell mass, and muscle hypertrophy, respectively (14 -16). Studies in cell culture systems indicate that Akt promotes growth by regulating several components of the translational apparatus, including the target of rapamycin (TOR) 2 pathway (17)(18)(19). However, there is relatively little published information about the targets of Akt signaling in in vivo models.
In addition to regulating growth, Akt promotes proliferation in some systems (reviewed in Refs. 9 -11). Progression through the cell cycle is controlled by the activity of protein kinase complexes consisting of cyclins and cyclin-dependent kinases (CDKs) and associated regulatory proteins (20). Studies in a number of models have demonstrated that Akt regulates cyclin-CDK activity and cell cycle progression through several mechanisms. For example, Akt has been shown to induce expression of cyclin D1, a key G 1 phase regulatory protein, through enhanced translation and impaired protein degradation (21). Akt can also inactivate the CDK inhibitor proteins p21 and p27, thereby promoting CDK activity and cell cycle progression (9).
Although Akt is generally thought to play a proproliferative role, other reports have shown that this kinase can also lead to cell cycle inhibition. In endothelial cells, constitutive Akt activation leads to a senescence-like cell cycle arrest through a p53and p21-dependent pathway (22). Transgenic mice with constitutively active Akt in thymocytes show cellular enlargement but no change in cell cycle distribution in vivo (23). Interestingly, thymocytes with constitutively active Akt demonstrated enhanced cell cycle progression in culture, suggesting that organ size control mechanisms prevented cell cycle progression in vivo but not in culture (23). The available data suggest that the effects of Akt signaling on cell cycle progression are complex and are probably dependent upon both the cell type and extracellular cues.
In the current work, we examined the regulation of hepatocyte growth and cell cycle progression by constitutively active Akt expression in vivo. As expected from other studies (14,15,24), Akt activation led to marked liver growth in the absence of substantial proliferation. Despite the lack of proliferation, liver growth was associated with increased expression of the cell cycle proteins, particularly cyclin E. The induction of cyclin E involved a translational mechanism, establishing a new link between cell growth and expression of this protein. Our data also suggest the existence of a p21-dependent cell cycle checkpoint that may be related to organ size homeostatic mechanisms.

EXPERIMENTAL PROCEDURES
Animals-All animal studies were completed following IACUC-approved techniques and National Institutes of Health guidelines. Eight-week-old male BALB/c (Harlan Sprague-Dawley), or p21 Ϫ/Ϫ or p21 ϩ/ϩ (The Jackson Laboratories) mice were injected with 6 ϫ 10 9 plaque-forming units via tail vein injection of E1-deleted recombinant adenoviruses encoding Akt, cyclin D1, E2F2, or a ␤-galactosidase control vector followed by bromodeoxyuridine (BrdUrd) injection, liver harvest, and processing as previously described (25,26). Four to seven animals were used per condition for each time point, and representative specimens are shown for Western blot and immunoprecipitation studies. Isolation of hepatocytes and liver nuclei for forward angle light scattering and flow cytometry analysis after propidium iodide staining (respectively) were performed as previously described (26,27). The adenovirus expressing a constitutively activated form of Akt1 (myr-Akt) was provided by Dr. K. Walsh (28), and the E2F2 adenovirus was provided by Dr. J. Nevins (29). Rapamycin (1.5 mg/kg/day) or vehicle (Me 2 SO) was given by intraperitoneal injection beginning 2 h before adenovirus transfection and daily thereafter as previously described (26).
Quantification of mRNA by Real Time RT-PCR-Total RNA from each liver was isolated and quantified as previously described (26). RNA (5 g) was treated with DNase I (DNAfree TM ; Ambion) according to the manufacturer's instructions. cDNA was synthesized from 5 g of each RNA sample with a Taqman reverse transcriptase reagent kit (Applied Biosystems) primed with oligo(dT). The cyclin E primers were purchased from Integrated DNA Technologies (upper, 5Ј-CACCACT-GAGTGCTCCAGAA-3Ј; lower, 5Ј-CCACATTTGCTCA-CAACCAC-3Ј). Real time PCR was performed using the Light Cycler DNA Master SYBR Green I kit (Roche Applied Sciences). Primers were used at a concentration of 0.5 M, and MgCl 2 was used at 2.4 mM. Samples were denatured for 10 min at 95°C and then 40 cycles of 95°C for 20 s, 55°C for 20 s, and 72°C for 20 s. Samples were normalized to ␤-actin, and relative amounts were calculated as recommended by the manufacturer. Quantification of cyclin E mRNA by real time RT-PCR in polysome fractions was carried out as described above and expressed as a percentage of total RNA from each liver sample.
RNA Isolation and Polyribosome Preparations-Mouse livers were removed and flash frozen in liquid nitrogen. 300 mg of frozen liver tissue was placed in a Dounce homogenizer with 375 l of low salt buffer (10 mM NaCl, 20 mM Tris, pH 7.5, 3 mM MgCl 2 ), supplemented with 1 mM dithiothreitol, 200 units of RNase inhibitor, and 1 mg/ml cyclohexamide and crushed with three slow strokes of the tight pestle. The homogenizer was incubated on ice for 5 min, and 125 l of lysis buffer (0.2 M sucrose, 1.2% Triton X-100 in low salt buffer) was added. The tissue was homogenized with 20 strokes of the tight pestle. The homogenate was removed and microcentrifuged at 16,000 ϫ g for 30 s to pellet insoluble material. The supernatant was removed to a new tube containing 50 l of heparin (1000 USP units/ml) and 15 l of 5 M NaCl. Equivalent A 260 units of the supernatants were layered onto a 5-ml, 0.5-1.5 M linear sucrose gradient (in low salt buffer) and centrifuged at 45,000 rpm in a Beckman SW55 rotor for 80 min at 4°C. The polyribosome stratified RNA was fractionated using an ISCO density gradient fractionator as previously described (32,33). Ten fractions of 0.5 ml were collected into tubes containing 50 ml of 10% SDS. RNA in the fractions was purified using Tri-reagent (Sigma) and precipitated overnight with 0.5 volumes of isopropyl alcohol.

RESULTS
Akt Induces Massive Liver Growth-Transfection of an activated form of Akt promoted liver growth beginning at 1 day after transfection (Fig. 1A) (24). At 6 days, the livers had nearly quadrupled in size. Microscopy revealed cytoplasmic changes compatible with increased hepatocyte fat and glycogen stores (data not shown) as recently reported following Akt transfection in the liver (24). We have previously shown that cyclin D1 promotes liver growth (25,26), although less dramatically than Akt. Cyclin D1 also induces robust hepatocyte proliferation (DNA synthesis and mitosis) in the absence of other mitogenic stimuli, particularly at time points from 1 to 3 days (25). In contrast, Akt induced little DNA synthesis as measured by BrdUrd incorporation into hepatocyte nuclei (Fig. 1B), and very few mitoses were seen in these livers (data not shown). Forward angle light scattering analysis of isolated hepatocytes demonstrated that Akt induced a substantial increase in hepatocyte size (Fig. 1C). Fluorescence-activated cell sorting analysis revealed an increased number of polyploid nuclei in the Akttransfected mice (Fig. 1D), which are common in normal liver and frequently observed under conditions of hepatocyte cell cycle arrest (34). Thus, Akt induces marked liver and hepatocyte growth without triggering a proportional response in hepatocyte proliferation.

Regulation of Cell Cycle Proteins by Akt-Because
Akt is known to regulate the cell cycle in some systems (9 -11), we examined the expression of cell cycle proteins in our model. As expected, transfection of Akt was associated with phosphorylation of Akt at Ser-473, indicating activation of Akt (Fig. 2). At 1 day after transfection, no induction of cell cycle proteins was observed, in contrast to cyclin D1-transfected livers, which show robust proliferation at this time point (25). This suggests that Akt signaling does not directly regulate cell cycle protein expression prior to the induction of growth. On the other hand, in livers that had undergone substantial growth at 6 days after Akt transfection, we were surprised to find significant up-regulation of proteins associated with the G 1 and S phases (such as proliferating cell nuclear antigen, cyclin A, cyclin E, and Cdk1) despite the minimal proliferative response (Figs. 1 and 2). In particular, cyclin E expression was higher at the 6 day time point following Akt transfection than in livers transfected with cyclin D1 for 1 day. Notably, Akt did not induce detectable cyclin D1 expression at any time point. Total levels of cyclin D3, Cdk2, and Cdk4 were modestly increased by Akt. Phosphorylation of Cdk2 at its activating site (Thr-160) was increased to levels greater than that observed in the cyclin D1-transfected mice, and phosphorylation at its inhibitory site was similar in the two models. We did not observe marked changes in the expression of the p27 CDK inhibitor, but p21 was substantially induced after 6 days. These results demonstrate that Akt-mediated liver growth promotes both proliferative and anti-proliferative changes in cell cycle protein expression and phosphorylation.
Regulation of Cyclin E Expression by Akt and Growth-The marked induction of cyclin E protein at 6 days (but not 1 day) after Akt transfection was puzzling given the lack of cyclin D1 expression and cell cycle activity. Prevailing models of cell cycle control suggest that cyclin E is induced by E2F-mediated transcription following partial inactivation of the retinoblastoma (Rb) and related pocket proteins by cyclin D1-Cdk4 (20). To examine this further, we measured the expression of cyclin E mRNA by RT-PCR (Fig. 3A). At 1 day after cyclin D1 transfection, which is associated with activation of cyclin D1-Cdk4 kinase activity (25), cyclin E mRNA was induced ϳ140-fold, presumably due to E2F-mediated transcription. Similarly, transfection with E2F2 for 1 day, which promotes robust hepatocyte DNA synthesis, 3 also markedly induced cyclin E mRNA.   At 1 day after Akt transfection, cyclin E mRNA was increased ϳ4-fold, indicating that Akt signaling does promote modestly increased expression of this mRNA in the liver despite the lack of protein expression. At the 6 day time point, cyclin E mRNA was up-regulated to a similar extent as at 1 day (p not significant), but the expression of cyclin E protein was markedly induced to levels even higher than at 1 day after cyclin D1 transfection (Fig. 2).
In a previous study (25), we found that at 6 days after cyclin D1 transfection, when the liver had undergone marked growth, we saw more induction of cyclin E protein than at 1 day, although cyclin D1-Cdk4 activity and BrdUrd uptake were diminished at the later time point. At the 6 day time point after cyclin D1 transfection, there is enlargement of hepatocytes (26). This suggested that cyclin E overexpression in these enlarged livers may be regulated by a common mechanism following either Akt or cyclin D1 expression. Indeed, we now show that the expression of cyclin E mRNA was only modestly induced at 6 days after cyclin D1 transfection (Fig. 3A), in contrast to the situation at 1 day, whereas cyclin E protein is markedly induced (Fig. 3B) (25). Thus, the expression of cyclin E protein does not parallel the expression of its mRNA in these models, and this protein appears to be disproportionately up-regulated relative to its mRNA following hepatocyte growth induced by either Akt or cyclin D1.
Previous review articles have suggested the possibility that cyclin E translation may be enhanced in states of cell growth, although direct evidence is lacking (5,6). To examine this further, we performed sucrose density centrifugation of liver extracts to isolate polysomes. As shown in Fig. 3B, cyclin E mRNA expression was shifted into heavier polysome fractions in the enlarged livers. These data suggest that the marked induction of cyclin E by Akt-mediated liver growth is due, at least in part, to increased translational efficiency of its mRNA. This supports the concept that cyclin E may act as a "sensor" of cell growth through a translational mechanism and thus may link growth to the cell cycle machinery (5,6).
Involvement of TOR Signaling in Akt-mediated Cell Growth and Cyclin E Expression-The mechanisms by which Akt regulates cell growth and proliferation are incompletely characterized, but signaling through TOR and downstream effectors represents one important pathway (1)(2)(3)(4)(5). To examine the involvement of TOR in the hepatic effects of Akt activation, we treated animals with the specific inhibitor rapamycin, which inhibits hepatocyte growth and proliferation in response to 70% partial hepatectomy and mitogens (26,(35)(36)(37)(38). Compared with vehicle-treated control animals, rapamycin diminished Aktmediated liver growth by 24% at the 6 day time point, but the liver size was still markedly greater than normal (Fig. 4A). As expected, rapamycin completely inhibited phosphorylation of S6 ribosomal protein, which is a downstream target of TOR signaling (Fig. 4B) (26,37). On the other hand, the expression of cyclin E was only modestly diminished by rapamycin and was still markedly elevated compared with base line. These results suggest that both growth and the resulting induction of cyclin E can occur independently of TOR signaling in this model.
Akt-mediated Liver Growth Promotes Formation of Inactivate Cdk2 Complexes-Activation of Cdk2 is thought to lead to phosphorylation of Rb and other targets that promote progression through S phase (20). The Western blot studies in Fig. 2 suggest that Akt-mediated liver growth might lead to activation of cyclin-Cdk2 complexes, given the induction of cyclins E and A and the Thr-160 phosphorylation of Cdk2, although relatively little evidence of DNA synthesis or mitosis was observed ( Fig. 1). To examine this further, we performed immunoprecipitation studies to characterize Cdk2-containing complexes. In the Akttransfected livers, no activation of kinase activity associated with cyclin E or Cdk2 toward the Rb protein or histone H1 was observed (Fig. 5A), indicating that these complexes were inactive. This contrasts with livers harvested 1 day after cyclin D1 transfection, which trigger CDK activation and cell cycle progression (25). At 6 days after Akt transfection, Western blot analysis of cyclin E immunoprecipitates indicated that there was abundant Cdk2 in the complex, including Cdk2 phosphorylated at the activating Thr-160 site (Fig. 5B). When comparing the complexes in the Akt-transfected mice with the active complexes seen at 1 day after cyclin D1 transfection, potential inhibitory events include increased phosphorylation of Cdk2 at Tyr 15 and binding of p21 and p27. This suggests that several mechanisms could lead to diminished kinase activity in response to liver overgrowth.

Akt-mediated Liver Growth Promotes Cdk2 Activation and Enhanced Cell Cycle Progression in the Absence of p21-Previ-
ous studies have documented that p21 negatively regulates progression of hepatocytes through the G 1 phase of the cell cycle in several models and may play a role in inhibiting proliferation in human liver diseases (31, 39 -43). Analysis of cyclin E-Cdk2 in Fig. 5 suggested that increased p21 binding to the complexes may inhibit the kinase activity. To address this directly, we repeated the studies in p21 Ϫ/Ϫ and matched wildtype (WT) mice. In the absence of p21, Akt led to marked liver growth similar to that observed in the WT mice at the 6 day time point (Fig. 6A). As compared with the WT mice, p21 Ϫ/Ϫ mice demonstrated a significant increase in hepatocyte BrdUrd uptake (Fig. 6B), and mitoses were readily observable on microscopy (Fig. 6C). Western blot studies showed that Aktmediated liver growth led to similar changes in the expression of cell cycle proteins in WT and p21 Ϫ/Ϫ mice, except that cyclin A and Cdk1 were induced to even higher levels (Fig. 6D). Interestingly, in the absence of p21, the inhibitory phosphorylation at Tyr-15 was increased, perhaps reflecting activation of additional antiproliferative pathways in this model.
In Fig. 6, E and F, we performed immunoprecipitation-kinase studies to examine CDK activity. At 6 days after Akt transfection in the p21 Ϫ/Ϫ mice, we found induction of kinase activity associated with cyclin E, cyclin A, and Cdk2. Western blot analysis of immunoprecipitated proteins indicated that absence of p21 in this model increased the Tyr-15 phosphorylation of Cdk2 that would tend to down-regulate kinase activity, although the net effect was activation of kinase complex. This suggests that liver overgrowth induced in this model may activate additional antiproliferative pathways that are further activated in the absence of p21. The cumulative results indicate  . Akt promotes formation of cyclin E-Cdk2 complexes that are inactivated. Liver lysates were subjected to immunoprecipitation with cyclin E or Cdk2 followed by kinase assay or Western blot analysis. A, kinase assays for cyclin E (top) or Cdk2 using Rb or histone H1 as a substrate. B, Western blot of cyclin E immunoprecipitates. that p21 suppresses Cdk2 activity and cell cycle progression in this model.

DISCUSSION
Cell growth is a critical component of normal cell division, but the molecular mechanisms that link growth and proliferation control have not been fully characterized. In this study, we used a model of transient Akt overexpression to examine the relationship between growth and proliferation in vivo. Our results suggest that cell growth can promote cyclin E expression via a translational mechanism, supporting prior studies that suggest that this protein may be a sensor of cell growth independent of the cyclin D-Rb-E2F pathway (5,6). Furthermore, we find that liver growth beyond that which is appropriate for animal size is associated with a p21-dependent cell cycle arrest. Our results provide further insight into the relationship between growth, the cell cycle machinery, and organ size.
It is well established that mitogens promote cell cycle progression by activating cyclin-CDK complexes that regulate progression through different phases of the cell cycle (20). In most systems, one or more D-type cyclins (often cyclin D1) is induced by signaling from extracellular stimuli and leads to activation of cyclin D-Cdk4 -6 kinase activity, which phosphorylates Rb and allows transcription of cell cycle genes by proproliferative E2F transcription factors. A key target of these transcription factors during late G 1 phase is cyclin E, which forms active kinase complexes with Cdk2 (or Cdk1) that phosphorylate Rb at different sites, leading to further activation of E2F and transcription of genes acting downstream in the cell cycle. In addition to Rb and related pocket proteins, cyclin-CDK complexes phosphorylate a diverse set of proteins involved in cell cycle progression. Furthermore, cyclins regulate a number of targets through Cdkindependent mechanisms (20,44).
The interplay between growth control and the cyclin-Cdk system is complex and incompletely understood. In general, adequate cell growth is a prerequisite for normal cell cycle progression, whereas cell growth can occur without proliferation, particularly in postmitotic cells. In many cell culture systems, signaling cascades triggered by growth factors regulate growth (e.g. via phosphoinositide 3-kinase) and cell cycle (e.g. via cyclin D1) in a coordinated manner. Inhibition of key growth control proteins, such as TOR, can override the mitogenic effect of growth factor signaling. In the case of TOR inhibition, studies in several models (including hepatocytes) have shown that mitogens stimulate expression of cyclin D1 mRNA, but translation of this protein is inhibited, leading to a cell cycle arrest (26,45). This supports the concept that cell cycle proteins act downstream of growth control proteins. However, a handful of studies have suggested a more complicated arrangement, since proteins typically associated with the cell cycle (e.g. cyclin D1 and Myc) can also trigger cell and organ growth (1-3, 25, 26, 46).
In the current study, we found that transient transfection of activated Akt led to massive hepatocyte and liver growth without evidence of substantial cell cycle progression. This is similar to the effect of Akt overexpression in postmitotic organs in Drosophila and mice, which leads to hypertrophy (12)(13)(14)(15)(16). In our model, Akt had no effect on the expression of cell cycle proteins at 1 day after transfection, suggesting that this kinase had minimal direct effects on these proteins. However, after liver growth had occurred at 6 days after transfection, we found marked up-regulation of cell cycle proteins, particularly cyclin E. Notably, although cyclin E mRNA expression was modestly increased at both the 1 and 6 day time points after Akt transfection, cyclin E protein was induced only after 6 days, when liver growth had occurred. Our data suggest that Akt-mediated cell growth, rather than Akt signaling per se, was responsible for the induction of cyclin E and other cell cycle proteins in this model.
Several lines of evidence suggest that the induction of cyclin E in Akt-mediated liver growth is regulated through a mechanism different from that which occurs during typical cell cycle progression. First, we did not observe expression of cyclin D1, which plays a pivotal role in hepatocyte proliferation and liver regeneration (25,26,(47)(48)(49)(50)(51). Second, cyclin E mRNA was only modestly induced by Akt (4 -6-fold), whereas this mRNA was induced Ͼ120-fold at 1 day after cyclin D1 or E2F2 transfection (when robust cell cycle progression occurs). Third, there was a marked disparity between cyclin E mRNA and protein expression; the protein was up-regulated to a greater degree at 6 days after Akt transfection than at 1 day after cyclin D1 expression, in contrast to the expression of the mRNA.
The finding that cyclin E protein was up-regulated by Aktmediated growth to a much greater degree than its mRNA suggests that this protein may be regulated at the level of translation. Indeed, we found that the cyclin E mRNA was shifted to heavier (translating) polysomes following Akt-induced liver growth, indicating that a translational mechanism plays a role in the increased expression of this protein. Similar to the current findings, previous studies in Drosophila have suggested that cyclin E expression may be regulated by cell growth independently of Rb and E2F and may thus link growth (or translational capacity) to the cell cycle machinery (6). A recent review has also speculated that cyclin E may be regulated by a translational mechanism following cell growth induced by signaling downstream of phosphoinositide 3-kinase in mammalian cells, although direct evidence has been lacking (5).
Additional evidence has also suggested that cyclin E expression may be linked to growth in the liver. In a previous study (25), we found greater induction of cyclin E protein at 6 days after cyclin D1 transfection (when the liver had undergone marked growth) than at 1 day, although cyclin D1-Cdk4 activity and DNA synthesis were diminished at the later time point. At the 6 day time point after cyclin D1 transfection, there is enlargement of hepatocytes (26). We now show that the expression of cyclin E mRNA was induced to a much greater degree at 1 day than at 6 days after cyclin D1 transfection (Fig. 3A), whereas the reverse is true for cyclin E protein (Fig. 3B) (25). These combined studies suggest that the marked induction of cyclin E in these enlarged livers may be regulated by a common mechanism following either Akt or cyclin D1 expression. Another prior study showed that conditional deletion of the S6 ribosomal protein led to diminished cyclin E expression in the liver following 70% partial hepatectomy despite normal expression of cyclin D1 and activation of cyclin D1-Cdk4 kinase activity, suggesting that cyclin E expression is regulated by ribosomal biogenesis, a key component of growth (52). Similarly, a study in hepatocellular carcinoma cells demonstrated that cyclin E expression is linked to the accumulation of ribosomal RNA during proliferation (53). The cumulative data strongly suggest that cyclin E is induced in states of cell growth in a manner distinct from the canonical cyclin D-Rb-E2F pathway.
We considered the possibility that the translational induction of cyclin E following liver growth was mediated by TOR, which integrates growth factor and nutrient signaling pathways to regulate both growth and proliferation (45). Indeed, inhibition of TOR with rapamycin has been shown to prevent cyclin D1 expression at the level of protein but not mRNA in a number of models (including hepatocytes) by decreasing the translation of this protein (26,45,54). Similar to previous results in transgenic mice (15), treatment with rapamycin only partially diminished Akt-mediated growth. Furthermore, rapamycin did not prevent the marked induction of cyclin E in our model, indicating that TOR has distinct effects on the regulation of cyclins D1 and E. Additional studies will be required to unravel the potential TOR-independent pathways that link cell growth to cyclin E expression.
Liver growth triggered by Akt led to changes in the expression of cell cycle mediators that would be expected to promote proliferation, but marked induction of Cdk2 activity, DNA synthesis, or mitosis was not noted. This is similar to the situation previously observed after cyclin D1 transfection in this model (25); after the robust proliferative response at 1-3 days, we found diminished cyclin-CDK activity and cell cycle progression at the 6 day time point. Thus, in both models, marked liver enlargement is associated with high abundance of cyclin-CDK complexes that are inactivated, with a consequent inhibition in hepatocyte proliferation. Several mechanisms could account for the inhibition of CDK activity in these models, including markedly increased p21 expression.
Because p21 plays an important role in regulating hepatocyte proliferation in vivo (31, 39 -43), we examined the effect of Akt transfection in p21 Ϫ/Ϫ mice. This led to a similar induction of liver growth as in wild-type mice. In livers of the knock-out animals, Cdk2 kinase activity was markedly up-regulated, and hepatocyte proliferation (as measured by DNA synthesis and the appearance of mitoses) was significantly increased, indicating that p21 plays an important role in governing cell cycle progression in this setting. Interestingly, Akt-mediated liver overgrowth in the p21 Ϫ/Ϫ mice was associated with increased phosphorylation of Cdk2 at Tyr-15 that would be expected to diminish kinase activation, although the net effect was increased activity. These data suggest that liver growth beyond the normal size may trigger several different potential antiproliferative pathways that impact cyclin-Cdk activity and cell cycle progression.
We and others have shown that p21 regulates the rate of progression hepatocytes through G 1 phase (i.e. hepatocytes enter S phase more quickly during liver regeneration in p21 Ϫ/Ϫ mice) (31,(41)(42)(43). In addition, p21 may play a role in inhibiting hepatocyte proliferation in liver diseases (39,(55)(56)(57)(58). In the current study, liver overgrowth was associated with marked induction of p21 that led to a relative cell cycle arrest; this is similar to our prior studies at 6 days after cyclin D1 transfection, which promotes liver growth and a comparable increase in p21 expression (25). These data suggest that induction of p21 opposes hepatocyte proliferation in the enlarged livers of both models. The induction of p21 could be due to a number of different mechanisms (59) but was probably not a direct result of Akt expression, since we did not observe expression of this protein at early time points after transfection (Fig. 2). One possible mechanism may involve liver size homoestatic responses, which prevent hepatomegaly through unknown cell cycle-inhibitory and proapoptotic pathways (51,60,61). However, further study is required to delineate the mechanism(s) responsible for the induction of p21 and other potential antiproliferative pathways in this model.
Despite the lack of proliferation or activation of Cdk2 by Akt-mediated growth in normal mice, we did see an increase in polyploid cells (Fig. 1D). Previous studies have indicated that polypoid hepatocytes are normal and increase in abundance during aging; furthermore, these cells are more common in the liver under conditions of hepatocyte cell cycle arrest (34). Emerging data suggest that cyclin E may play an important role in regulation endoreduplication in certain polyploid cell types and that this function is independent of its ability to activate CDKs (62). Thus, the overexpressed cyclin E seen in the livers at 6 days after transfection may be leading to endoreduplication without CDK activation or cell cycle progression; the rate of DNA synthesis in this process may be lower than that which can be detected by immunohistochemistry after a 2-h pulse of BrdUrd. The control of hepatocyte ploidy is complex and poorly understood, and further studies will be required to clarify the relationship between Akt, growth, cyclin E, and polyploidy in the liver.
In summary, the data presented here suggest that liver growth induced by Akt is associated with marked up-regulation of key cell cycle genes in a manner distinct from that triggered by the cyclin D-Rb-E2F pathway. In particular, cyclin E was markedly induced by growth through a translational mechanism, and this may provide a link between cell growth pathways and the cell cycle machinery, an issue of relevance to both normal and malignant cells (1)(2)(3)(4)(5)(6). In addition, our data suggest that liver overgrowth is associated with an antiproliferative response involving p21. Further study of the mechanisms underlying these findings could provide insight into development, organ size homeostasis, regeneration, metabolic function, and cancer in the liver.