RAPAMYCIN INHIBITS LIVER GROWTH DURING REFEEDING IN THE RAT VIA CONTROL OF RIBOSOMAL PROTEIN TRANSLATION BUT NOT CAP-DEPENDENT TRANSLATION INITIATION

Abstract

We have examined the role of the mammalian target of rapamycin (mTOR) in hepatic cell growth. In order to dissociate cell growth from cell proliferation, we employed an in vivo model of non-proliferative liver growth in the rat, refeeding following 48 h of food deprivation. Starvation resulted in a decrease in liver mass, liver protein and cell size, all of which were largely restored after 24 h of refeeding. Administration of the mTOR inhibitor, rapamycin, before the refeeding period partially inhibited the restoration of liver protein content. Refeeding was also associated with an increase in ribosomal protein S6 phosphorylation and phosphorylation of the eukaryotic initiation factor (eIF) 4E binding protein 1 (4E-BP1). 4E-BP1 phosphorylation was accompanied by a decrease in the abundance of the complex containing 4E-BP1 with eIF4E. These changes were prevented by rapamycin administration. However, association of eIF4E and eIF4G and eIF2α phosphorylation, both of which are stimulated by refeeding, were insensitive to rapamycin. The functional significance of these observations was confirmed by polysome fractionation, which showed that translation initiation of 5′ oligopyrimidine tract-containing mRNAs, which encode ribosomal proteins, was inhibited by rapamycin while translation of STAT1, a cap-dependent mRNA, was unaffected. The abundance of ribosomal proteins paralleled total protein content during refeeding in both control and rapamycin-injected rats. We conclude that accretion of liver protein during refeeding is dependent on mTOR-mediated activation of the translation of ribosomal proteins but not dependent on mTOR-mediated activation of cap-dependent translation initiation.

INTRODUCTION

Signal transduction via the mammalian target of rapamycin (mTOR)³ has been the subject of intense investigation in recent years. Rapamycin, an inhibitor of the mTOR kinase, has been used to great effect in dissecting this pathway (1-3). mTOR, which has been referred to as the “central controller of cell growth” (4), is responsive to nutrients and to activation by receptor tyrosine kinases, including those for insulin, insulin-like growth factor I (IGF-I) and a panoply of peptide growth factors (5). mTOR signaling to various components of the translational apparatus is thought to regulate the initiation of mRNA translation, the rate limiting step in protein synthesis (6).

Targets of mTOR signaling include several key regulators of translation. Eukaryotic initiation factor (eIF) 2α and the eIF4E/eIF4G-containing mRNA cap-binding complex (termed eIF4F) play important roles in global control of translation initiation (7). Phosphorylation of ribosomal protein S6 is thought to specifically play a role in translation initiation of ribosomal proteins, which are encoded by mRNAs that have 5′ terminal oligopyrimidine tracts (5′ TOP mRNAs) (8,9). However, recent work (10-12), including our own (13,14), has challenged this long-standing regulatory paradigm. Mechanism of action aside, numerous studies are consistent in reporting a decrease in polysomal recruitment of 5′ TOP mRNA with acute exposure to rapamycin both in vitro and in vivo. Rapamycin-induced inhibition of 5′ TOP mRNA translation is thought to inhibit cell growth, resulting in decreased cell size, and cell proliferation. Both effects are thought to be a consequence of limited abundance of ribosomal proteins and, as a direct consequence, a reduction in the number of ribosomes per cell.

mTOR can activate the key translation initiation factor eIF4E via phosphorylation of the eIF4E inhibitory binding protein, 4E-BP1 (7). Phosphorylation of 4E-BP1 results in its dissociation from eIF4E, which permits eIF4E to interact with the scaffold protein eIF4G. eIF4G, in turn, recruits multiple protein factors to assemble the eIF4G complex, which links the mRNA to the 40S subunit of the ribosome (15). mTOR signaling has also been shown to be involved in controlling the phosphorylation of eIF2α by its cognate kinase, GCN2. GCN2-dependent phosphorylation of eIF2α, which inhibits the ability of this initiation factor to participate in the activation of translation (7), is rapamycin-sensitive in yeast, suggesting a role for mTOR in this phosphorylation event (16,17).

It has been reported that rapamycin-sensitive mTOR signaling is critical to the regulation of cell size (5,18). This observation is based largely on experiments that have been conducted on cells in culture. In order to determine if this association holds true in vivo, we studied the effect of rapamycin in a model of hepatic cell growth in the rat that involves refeeding after 48 h of food deprivation (19). We chose this model, in part, because of the absence of an effect of refeeding on hepatic cell proliferation, thus allowing us to dissociate effects on cell cycle activation and progression from those on ribosomal biogenesis and protein translation. We performed analyses to examine the effect of rapamycin on the aforementioned translational control events. In addition, we extended our studies to examine not only the acute effects of rapamycin but also its effects during the latter stages of the period that is required to restore liver mass in this model.

MATERIALS AND METHODS

Materials. S6K substrate peptide (RRRLSSLRA) was purchased from Upstate USA, Inc. (Lake Placid, NY). Antibodies to S6, phosphorylated S6 (Ser235/236 and Ser240/244), phosphorylated eIF2α (Ser51) and eIF4E were purchased from Cell Signaling Technology, Inc. (Beverly, MA). Antibodies to 4E-BP1, eIF4G, eIF2α, S6K1, S6K2, L28 and L11, as well as protein kinase A and protein kinase C inhibitors, were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Protein A-Sepharose CL-4B, protein G-Sepharose 4FF and 7-methyl-GTP (7mGTP)-Sepharose 4B were purchased from Amersham Pharmacia Biotech (Piscataway, NJ). [γ-³²P]ATP (3000 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA). All primers for Realtime-polymerase chain reaction were purchased from Invitrogen Corp. (Carlsbad, CA).

Animal Studies. Male Sprague-Dawley rats age 4 to 5 weeks (Charles River Laboratories, Wilmington, MA) were used for all studies. Control rats were given ad libitum access to water and food (Purina rat chow, 3300 kcal/kg with 23.4% as protein, 4.5% as fat, and 72.1% as carbohydrate). Food-deprived rats had their laboratory chow withheld for 48 h. During this time, water was provided ad libitum. Prior to refeeding, rats were administered vehicle (DMSO) or rapamycin (50 mg/kg body weight) by intraperitoneal injection. Fifteen minutes after injection, the refeeding period was initiated by replacing the laboratory chow in their cage covers. Ad libitum feeding was allowed to proceed for 1, 4, 8 or 24 h. All rats were sacrificed by exsanguination under pentobarbital sodium anesthesia (50 mg/kg by intraperitoneal injection). Each experimental group had an n of 3. Three sets of animals were required to complete all of the experiments and analyses that were performed. All studies complied with guidelines adhered to by the Rhode Island Hospital Institutional Animal Care and Use Committee.

Biochemical Analyses. Liver tissue was frozen in liquid nitrogen before storage at -70°C. Frozen sections were stained using standard procedures with 4′,6-diamidino-2-phenylindole (DAPI) to detect nuclei and Oil Red O to detect fat droplets. Liver protein content was calculated based on total liver weight and the determination of protein content in samples prepared as described previously for ribosomal protein S6 kinase (S6K) assays (13). S6K1 and S6K2 activities were measured by immunoprecipitation kinase assay (13). Sample preparation and analyses for 7mGTP cap-binding studies have also been previously described (14). The association of 4E-BP1 and eIF4G with eIF4E was evaluated as their ability to bind to 7mGTP-Sepharose beads via the interaction with eIF4E. Sample preparation, 7mGTP-Sepharose affinity purification and Western immunoblotting were carried out as described previously (20). Western immunoblotting was accomplished using standard electrophoresis and transfer methods with chemiluminescent detection (20). Quantification of Western blots was performed using the ChemiDoc-It Imaging System with Labworks Version 4.5 software (UVP, Inc., Upland CA). In cases where immunoblotting was used to quantify both protein phosphorylation and total content, analyses were performed sequentially on the same Western blot. Protein determinations were made using the bicinchoninic acid method (BCA, Pierce Chemical Company, Rockford, IL) with bovine serum albumin as the standard.

Polysome preparation and fractionation was carried out as described previously (14). Fractions were collected and pooled to obtain three samples containing long polysomes, intermediate polysomes and short polysomes/monosomes. RNA was extracted as described previously (14) and quantified by spectral absorption.

In order to quantify the abundance of specific RNAs in the polysome pools, cDNA was generated using 1 μg of heparinase-treated RNA and Taqman reverse transcription reagents (Applied Biosystems, Foster City, CA) by following the manufacturer's instructions. Ribosomal protein S6 (S6) and GAPDH primer sequences were provided by Dr. Scot Kimball (Penn State College of Medicine, Hershey, PA). Rat-derived sequence data were used for L28, Signal Transducer and Activator of Transcription 1 (STAT1) and β-Actin primer design. PCR primers were designed using the MIT Primer3 design program (21). Quantitative real-time PCR was performed in an ABI 7700 instrument using the SYBR green system. Results are expressed as abundance of target mRNA relative to GAPDH mRNA. The following primer sequences were used: S6, 5′-ACTGGCTGTCAGAAACTCAT-3′ and 5′-CCACATAACCCTTCCACTCT-3′ (product size 120 base pairs [bp]); L28, 5′-TACGTGAGGACCACCATCAA-3′ and 5′-CAGATCAGGGCGGTACTTGT-3′ (90 bp); STAT1, 5′-AGGTCCGTCAGCAGCTTAAA-3′ and 5′-CGATCGGATAACAACTGCTT-3′ (97 bp); GAPDH, 5′-GGGCTGCCTTCTCTTGTGA-3′ and 5′-TGAACTTGCCGTGGGTAGA-3′ (84 bp); β-Actin, 5′-GTAGCCATCCAGGCTGTGTT-3′ and 5′-CCCTCATAGATGGGCACAGT-3′ (104 bp). The annealing temperature for all was 60°C.

Statistical Analyses. All data are shown as mean and standard deviation. In the case of a direct comparison between two groups, an unpaired t-test was used. Comparisons between multiple groups utilized one-way analysis of variance with a Tukey post-hoc test to identify specific differences. Analyses were done using Prism 2.01 (GraphPad Software, Inc., San Diego, CA). Level of significance was set at P<0.05.

RESULTS

Our initial experiment was aimed at examining whether rapamycin has an effect on reaccumulation of liver mass and liver protein on refeeding. Starvation for 48 h induced a 25% reduction in liver weight:body weight ratio (Table 1). This ratio increased to levels greater than that in control rats after 24 h of refeeding. Unexpectedly, there was no difference between the vehicle and rapamycin-injected refed groups. Total liver protein (Table 1) fell by more than a third with starvation. Nearly complete recovery of liver protein content was attained within 24 h in vehicle-injected refed rats. This recovery was attenuated in the rapamycin-injected refed group (83% recovery in vehicle-injected versus 45% recovery in rapamycin-injected rats).

Table 1.

The effects of fasting, refeeding and rapamycin administration on liver mass and composition. All data are shown as the mean ± 1 SD.

			Refed, 24 h
	Control	Starved	Vehicle	Rapamycin
Liver:Body Weight Ratio	0.047±0.005	0.036±0.002*	0.058±0.002*	0.059±0.001
Total Liver Protein (g)	0.841±0.008	0.512±0.074*	0.783±0.008	0.659±0.032*†
Cell Density (Nuclei/mm2)	695±107	1033±65*	539±75	622±61
Oil Red O Staining (% of surface area)	947±576	32±1	594±245	1781±618†

^*P<0.05 versus Control Group.

^†P<0.05 versus Vehicle-Injected Refed Group.

Given the incomplete inhibition of liver growth seen in the rapamycin group, we sought to confirm that the single intraperitoneal dose of rapamycin was effective for a full 24 h. Liver samples were homogenized and analyzed by Western immunoblotting with antibodies towards S6 and phosphorylated S6 (Fig. 1). In agreement with our previous observations (14), we saw an increase in S6 phosphorylation after 48 h of food deprivation. After vehicle administration and 24 h of refeeding, the level of S6 phosphorylation approximated that seen in the starved group. Rapamycin induced a profound and persistent inhibition of S6 phosphorylation at the 24 h time point.

Fig. 1.

Hepatic ribosomal protein S6 content and phosphorylation in rats that were fed ad libitum (control, C), starved for 48 h (F), or starved for 48 h then refed for 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). Hepatic S6 phosphorylation was analyzed by Western immunoblotting with antibodies directed towards phosphorylated S6 (P-S6; Ser 235/236) and total S6. The immunoblots were analyzed by densitometry, the results of which are shown in the graph as the ratio of P-S6 to total S6. Results are shown as means + 1 SD, n = 3 rats per condition. *, P < 0.05 versus the control group. †, P < 0.01 versus the vehicle-injected (V) refed group.

In order to obtain an indirect measure of cell size, liver sections were stained with DAPI and analyzed for cell density by determining the number of DAPI positive cells per unit area (Table 1). Cell density was highest in the starved group, consistent with a reduction in cell size. We did not observe a difference in hepatocyte density between vehicle and rapamycin-injected refed rats.

Given that rapamycin-injected, refed rats showed full recovery of liver mass, partial recovery of liver protein and no difference in cell density, we hypothesized that accumulation of non-protein cellular constituents in rapamycin-injected rats might account for the apparent absence of a rapamycin effect on liver mass and cell size. To test the hypothesis that fat deposition could account for the observed discrepancy, liver sections were stained for intracellular fat using Oil Red O. Results (Table 1) demonstrated that starvation resulted in a complete disappearance of fat droplets from the liver. The fat content of livers from rapamycin-injected refed rats was 3-fold higher than in the vehicle-injected group.

Direct measurement of S6K1 and S6K2 activities by immunoprecipitation kinase assay (data not shown) demonstrated a rapid activation of both kinases in response to refeeding. This activation was potently inhibited by rapamycin. Taken together with the S6 phosphorylation results (Fig. 1), these findings predicted parallel inhibition of mTOR-mediated activation of eIF4F complex formation. This was assessed by analysis using 7mGTP affinity-purification of liver homogenates followed by Western immunoblotting for 4E-BP1 and eIF4E (Fig. 2A). The ratio of 4E-BP1:eIF4E in the affinity-purified complexes confirmed our previous result (14) that starvation produces a counter-intuitive loss of the complex resulting from the association between these two proteins. Refeeding was associated with an even greater decrease in this ratio, which would generally be considered consistent with further activation of cap-dependent translation. In rats that were refed for 1 h and 8 h, administration of rapamycin induced an increase in the 4E-BP1:eIF4E ratio compared to corresponding controls. This effect was lost at 24 h, apparently as a result of an increase in the vehicle-injected rats. Direct Western blot analysis of the homogenates (Fig. 2B) was performed to assess 4E-BP1 phosphorylation state. The ratio of the least phosphorylated α form of 4E-BP1 to total 4E-BP1 (Fig. 2B) showed that starvation was not associated with a change in the 4E-BP1 phosphorylation state. However, refeeding induced a rapid and marked increase in 4E-BP1 phosphorylation (a decrease in the proportion of 4E-BP1 in the α form) that was sustained for 24 h. This increase was potently inhibited by rapamycin throughout the 24 h refeeding period. eIF4E content (Fig. 2B) was not affected by starvation or refeeding.

Fig. 2.

Regulation of the interaction of 4E-BP1 with eIF4E in the cap-binding complex in rats that were fed ad libitum (control, C), starved for 48 h (F), or starved for 48 h then refed for 1, 8 or 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). Panel A: eIF4E-containing complexes were affinity purified from liver homogenates using 7mGTP-Sepharose beads. The affinity-purified proteins were analyzed by Western immunoblotting with antibodies directed towards 4E-BP1 and eIF4E. The positions of the α and β 4E-BP1 bands are shown to the left. The Western blot results were analyzed by densitometry to obtain the ratio of 4E-BP1 to eIF4E for each condition. Bars represent the mean of the duplicate results obtained for each condition. Panel B: The same liver homogenates used to obtain the results shown in Panel A were analyzed by direct Western immunoblotting for total 4E-BP1 and eIF4E. The positions of the α, β and γ 4E-BP1 bands are shown to the left. The Western blot results were analyzed by densitometry to obtain the ratio of the intensity of the α form to the total 4E-BP1 content (combined α, β and γ). Bars represent the mean of the duplicate results obtained for each condition.

The eIF4E/4E-BP1 results predicted a marked inhibition of protein synthesis during refeeding by rapamycin. The substantial recovery of liver protein in rapamycin-injected animals seemed inconsistent with this result. We therefore performed further studies to examine the regulation of eIF4F formation. Analysis of hepatic eIF4G content by direct immunoblotting (Fig. 3A) showed that hepatic eIF4G content fell during starvation. Control levels were attained upon refeeding only after 24 h. However, analysis of eIF4E-associated proteins obtained by 7mGTP affinity-purification gave an unexpected result (Fig. 3B). While starvation was associated with lower levels of eIF4G in the affinity-purified preparations, refeeding resulted in a rapid increase that was insensitive to rapamycin.

Fig. 3.

Regulation of the interaction of eIF4E with eIF4G in the cap-binding complex in control rats that were fed ad libitum (Cont), starved for 48 h (Starv), or starved for 48 h then refed for 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). Panel A: Liver homogenates were analyzed by direct Western immunoblotting for total eIF4G. Panel B: These same homogenates were subjected to affinity purification using 7mGTP-Sepharose beads in order to obtain eIF4E-containing cap-binding complexes. The affinity-purified proteins were analyzed by Western immunoblotting with antibodies directed towards eIF4G.

To examine whether rapamycin treatment affects translation via inhibition of eIF2α through phosphorylation at Ser-51, we performed Western immunoblotting for phosphorylated and total eIF2α. Results (Fig. 4A) did not show an effect of starvation on eIF2α phosphorylation. However, refeeding was associated with a decline in eIF2α phosphorylation at 1 h and 8 h. By 24 h, there was an increase in eIF2α phosphorylation that was, contrary to the expected result, attenuated in the rapamycin-injected rats. Because this result was counterintuitive, we repeated the analysis using samples from triplicate vehicle and rapamycin-injected 24 h refed rats (Fig. 4B). The results confirmed that the phosphorylation of eIF2α was indeed lower in the rapamycin-injected samples at the 24 h time point.

Fig. 4.

Hepatic eIF2α phosphorylation in control rats that were fed ad libitum (Cont), starved for 48 h (Starv), or starved for 48 h then refed for 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). Panel A: Liver homogenates were analyzed by Western immunoblotting with antibodies directed towards phosphorylated eIF2α (Ser 51; P-eIF2α) and total eIF2α. Panel B: Liver homogenates from triplicate vehicle- and rapamycin-injected 24 h refed rats were analyzed by Western immunoblotting for total and phosphorylated eIF2α. The graph shows the ratio of phosphorylated to total eIF2α as the mean + 1 SD. *, P < 0.02 versus the control group.

In order to assess the effect of rapamycin on translation of specific mRNAs, homogenates from control, 48 h starved and vehicle- or rapamycin-injected, refed rats were fractionated by sucrose gradient centrifugation. RNA extracted from long polysome, intermediate polysome and short polysome/monosome pools were analyzed by quantitative RT-PCR. We studied the abundance of two 5′ TOP mRNAs, S6 and L28, a mRNA with a highly structured 5′ untranslated region (UTR), STAT1, and a housekeeping mRNA, β-Actin. The abundance of these mRNAs in each pool was expressed relative to GAPDH abundance in the same pool in order to correct for variable efficiency of RNA extraction. Replicate analyses were carried out for two sets of samples, one of which is shown (Fig. 5). The polysomal distribution of β-Actin was not affected by rapamycin. In contrast, translation of the two 5′ TOP mRNAs was sensitive to rapamycin. That is, a rapamycin-associated increase in the proportion in the short polysome/monosome fractions was consistently seen at 1, 4 and 8 h for both 5′ TOP mRNAs in both sets of samples. The distribution of STAT1 mRNA showed no reproducible effects of rapamycin administration at any time point.

Fig. 5.

Polysomal distribution of specific mRNAs in control rats that were fed ad libitum (C), starved for 48 h (S), or starved for 48 h then refed for 1, 4, 8 or 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). RNA was extracted from polysome-fractionated pools and analyzed by quantitative RT-PCR for β-Actin, L28, S6, STAT1 and GAPDH mRNA content. Results are expressed as abundance of the mRNA indicated relative to GAPDH abundance in long polysomes (filled bars), intermediate polysomes (hatched bars) and short polysomes/monosomes (open bars).

Given the rapamycin-induced inhibition of 5′ TOP mRNA translation, we predicted a decrease in the cellular abundance of all ribosomal proteins in livers from refed rats that were given rapamycin. We analyzed steady state levels of the 60S ribosomal proteins L28 and L11 and the 40S ribosomal protein S6 by direct Western immunoblotting of liver homogenates (Fig. 6). There were subtle decreases in the abundance of these proteins in livers from rapamycin-exposed refed rats, but the timing of the rapamycin effect differed from protein to protein. The abundance of these ribosomal proteins relative to total liver protein showed no consistent changes in response to rapamycin. Immunoblotting for phospho-S6 confirmed the efficacy of rapamycin in samples from all rapamycin-injected rats.

Fig. 6.

Ribosomal protein content in control rats that were fed ad libitum (Control or C), starved for 48 h (Starv or S), or starved for 48 h then refed for 1, 4, 8 or 24 h. The latter were injected at the start of the refeeding period with vehicle (V) or rapamycin (R). Liver homogenates from 2 rats per condition were analyzed by Western immunoblotting with antibodies directed towards L28, L11, phosphorylated S6 (P-S6; Ser 235/236) and total S6.

DISCUSSION

Recent studies have established the role of mTOR as a positive regulator of cell growth. Assignment of this role to mTOR is in part accounted for by the ability of its inhibitor, rapamycin, to inhibit cell growth and decrease cell size (18,22). These studies have employed cell culture systems. In the model we employed, food deprivation for 48 h is accompanied by a profound decrease in liver size, liver protein content and hepatocyte protein synthetic rate (23-26). Refeeding for 24 h was shown previously to be sufficient for recovery of liver mass (19). We chose to use this model of hepatic cell growth, rather than that of liver regeneration, to examine the role of mTOR in the regulation of cell size and translational control under circumstances in which hepatocyte proliferation is not a factor.

The present studies were aimed at further characterization of rapamycin as an inhibitor of cell growth in vivo. Our initial results were unexpected. Rapamycin administration prior to a 24 h refeeding period did not show impaired recovery of liver mass compared to vehicle-injected controls. However, the reaccumulation of liver protein in these rats was partially inhibited. Histochemical staining and morphometric analysis showed that cell density was, as expected, highest in the starved group. There was no difference in cell density and, by extension, cell size comparing control and rapamycin-exposed groups. Oil Red O staining revealed that rapamycin-injected refed rats had higher hepatic fat content compared to control rats. This latter observation may be secondary to the known effect of rapamycin to induce hyperlipidemia (27,28). We concluded that intracellular fat accumulation contributed to the apparent restoration of liver mass and hepatic cell density seen in rapamycin-exposed rats.

5′ TOP mRNAs encode all ribosomal proteins, thus giving them a central role in the regulation of translation and ribosomal biogenesis (29). It is well established that rapamycin selectively inhibits recruitment of 5′ TOP mRNA into translating polysomes (8,9). Based on this observation, it is generally accepted that one mechanism by which rapamycin inhibits cell growth is through a reduction in the cellular abundance of ribosomal proteins. The previous demonstration that essential amino acid restriction during refeeding in the rat blocks the activation of hepatic 5′ TOP mRNA translation (30) is an indication that mTOR signaling to ribosomal protein translation is activated in the in vivo model used in the present studies. However, these experiments were confined to the effects of a 1 h period of refeeding, as were more recent findings (31) demonstrating that leucine administration after starvation induced polysomal redistribution of ribosomal protein mRNAs. The present studies, by extending the period of observation during refeeding, indicate that there is a component of hepatic protein accretion and synthesis of ribosomal proteins that is independent of S6 phosphorylation, insensitive to rapamycin and, by extension, not dependent on signaling via mTOR.

As noted above, we (13,14) and others (10-12) have been able to dissociate S6 phosphorylation from 5′ TOP translation. Other mechanisms that may contribute to inhibition of ribosomal biogenesis by rapamycin could involve mTOR signaling to rDNA transcription. In recent years, it has been shown that RNA polymerase I binding and transcription of rDNA is mTOR-mediated and rapamycin-sensitive (32-36). In addition, a recent report (37) identified the protein SKAR (S6K1 Aly/REF-like target) as a novel and specific binding partner and substrate of S6K1. SKAR is a nuclear protein and its abundance correlates directly with cell size.

Rapamycin has been shown to inhibit global translation rate and, more specifically, the translation of mRNAs with highly structured 5′ UTRs through inhibition of the formation of the eIF4F complex (2,38). Refeeding following starvation in neonatal pigs has been shown to result in rapamycin-sensitive stimulation of the formation of eIF4F (39). Rapamycin was able to block the refeeding-associated activation of protein synthesis. However, the observations in this study also were limited to an acute (1.5 h) period of refeeding. Our results indicate rapamycin-resistant formation of the cap-binding complex as indicated by the association of eIF4E and eIF4G. The apparent discrepancy between this result and previous observations presumably relates to the reliance of prior studies on examination of 4E-BP1 phosphorylation and eIF4E/4E-BP1 association. eIF4G was not examined. We have previously published a paradoxical increase in the formation of the hepatic eIF4F complex during starvation, which we attributed to the starvation-associated increase in circulating branched-chain amino acids (14). In the present studies, we went on to demonstrate that refeeding was associated with a marked decrease in the association of 4E-BP1 and eIF4E, consistent with decreased inhibition of eIF4E. Rapamycin prevented the dissociation of 4E-BP1 and eIF4E and increased the proportion of 4E-BP1 in the hypophosphorylated state (detected as 4E-BP1α). We therefore predicted a corresponding decrease in the association of eIF4G and eIF4E in rapamycin-treated samples. However, the amount of eIF4G associated with 7mGTP beads was not different between samples obtained from vehicle- and rapamycin-injected rats. Based on this result, we hypothesized that rapamycin would not affect the translation of STAT1, a mRNA with a high secondary structure energy-containing 5′ UTR (14). This was indeed the case. Taken together with our prior observations in the starved rat, these results indicate a mode of upregulation of global protein synthesis during refeeding that is independent from regulation of cap-dependent translation.

Recent studies in yeast have demonstrated that another mechanism of rapamycin-induced translational inhibition may involve modulation of GCN2-mediated phosphorylation and inhibition of eIF2α (16,17). eIF2α phosphorylation inhibits nucleotide exchange in the eIF2 complex, an event important for the delivery of methionyl-tRNA for translation initiation (7). We studied eIF2α phosphorylation in response to rapamycin in our liver growth model. There was no change with rapamycin treatment at 1 h and 8 h after refeeding. We observed a paradoxical decrease in eIF2α phosphorylation at 24 h. We speculate that at this time point livers from vehicle-injected refed rats have attained their pre-starvation protein content and that there is an intracellular mechanism in place that senses this, thus signaling to the cell that protein synthesis should be slowed by increasing eIF2α phosphorylation. In contrast, rapamycin-injected rats have a hepatic protein deficit even at 24 h after refeeding, and low eIF2α phosphorylation levels may be maintained so that translation may continue until such time that there is complete reaccumulation of hepatic protein.

Consistent with numerous other studies (8,9,40), quantitative PCR performed on RNA from fractionated polysomes showed translational inhibition of both 5′ TOP mRNAs examined, S6 and L28. Taken together with the apparent lack of a rapamycin effect on the cap-binding complex, these results suggest that the inhibitory effect of rapamycin on recovery of liver protein after starvation may be fully accounted for by inhibition of ribosomal biogenesis. This is consistent with our observation that at all times examined, and following both vehicle and rapamycin administration, ribosomal protein abundance paralleled total protein content. Our data do not allow us to assign a functional role to the subtle differences in the abundance of the different ribosomal proteins during refeeding. Volarevic et al. (19) showed that reduced abundance of a single ribosomal protein, S6, was sufficient to impede ribosomal biogenesis. It is possible that decreased abundance of any ribosomal protein at a given time point may reflect an inhibition of ribosomal biogenesis, possibly accounting for the partial inhibition of liver protein accretion that was seen with rapamycin administration.

In summary, our studies show that rapamycin partially inhibits reaccumulation of cellular protein in a model of liver growth. This inhibition is incomplete even though the inhibition of S6 phosphorylation is profound and persistent. We interpret these findings as indicating parallel rapamycin-sensitive and -insensitive mechanisms for the activation of global protein synthesis during refeeding. Neither of two well-characterized mechanisms of translation initiation, formation of the eIF4F cap-binding complex or modulation of eIF2α phosphorylation, could account for the partial inhibition of protein reaccumulation over 24 h by rapamycin. However, rapamycin clearly inhibited recruitment of 5′ TOP mRNA into polysomes. In keeping with a primary role for ribosomal protein abundance in controlling global protein synthesis, L28, L11 and S6 steady-state levels were constant relative to hepatic protein content in both control and rapamycin-injected rats. Thus, we conclude that rapamycin is indeed an inhibitor of cell growth in a non-proliferative, in vivo model, and that this inhibition is a function of the translational control of ribosomal proteins and not mechanisms that control cap-dependent translation initiation.