Rapamycin-sensitive induction of eukaryotic initiation factor 4F in regenerating mouse liver



Following acute injuries that diminish functional liver mass, the remaining hepatocytes substantially increase overall protein synthesis to meet increased metabolic demands and to allow for compensatory liver growth. Previous studies have not clearly defined the mechanisms that promote protein synthesis in the regenerating liver. In the current study, we examined the regulation of key proteins involved in translation initiation following 70% partial hepatectomy (PH) in mice. PH promoted the assembly of eukaryotic initiation factor (eIF) 4F complexes consisting of eIF4E, eIF4G, eIF4A1, and poly-A binding protein. eIF4F complex formation after PH occurred without detectable changes in eIF4E-binding protein 1 (4E-BP1) phosphorylation or its binding eIF4E. The amount of serine 1108-phosphorylated eIF4G (but not Ser209-phosphorylated eIF4E) was induced following PH. These effects were antagonized by treatment with rapamycin, indicating that target of rapamycin (TOR) activity is required for eIF4F assembly in the regenerating liver. Rapamycin inhibited the induction of cyclin D1, a known eIF4F-sensitive gene, at the level of protein expression but not messenger RNA (mRNA) expression. In conclusion, increased translation initiation mediated by the mRNA cap-binding complex eIF4F contributes to the induction of protein synthesis during compensatory liver growth. Further study of factors that regulate translation initiation may provide insight into mechanisms that govern metabolic homeostasis and regeneration in response to liver injury. (HEPATOLOGY 2004;40:537–544.)

In response to acute injuries that reduce functional hepatic mass, the remaining liver tissue undergoes rapid growth that eventually restores organ size and function. The best-studied model of liver regeneration is that of 70% partial hepatectomy (PH) in rodents, which triggers most of the remaining hepatocytes to enter the cell cycle in a relatively synchronous fashion.1, 2 Another important component of the response to PH is the marked induction of protein synthesis, which is required for normal regeneration.3, 4 The mechanisms that regulate protein synthesis in the regenerating liver have not been well characterized.

In general, the induction of protein synthesis after growth stimulation involves augmentation of both translational capacity and efficiency, which depend on increased numbers of ribosomes and enhanced translation initiation, respectively.5–8 Previous studies have documented enhanced ribosome biogenesis in the regenerating liver.4, 9–11 However, there is little published data regarding changes in translation initiation in this model. Translation initiation is a complex process that requires collaboration between multiple eukaryotic initiation factors (eIFs) and signaling pathways.6–8, 12, 13 A major determinant of translation initiation is binding of the eIF4F protein complex to the m7GpppN 5′ cap present on most cellular messenger RNA (mRNA) transcripts. The translation of mRNAs possessing a high degree of secondary structure in the 5′ untranslated region is particularly sensitive to changes in eIF4F. The eIF4F complex consists of the eIF4E cap-binding protein and the large eIF4G scaffolding protein, which binds the eIF4A RNA helicase, poly-A binding protein (PABP), and other regulatory components.

Assembly of eIF4F is thought to be a rate-limited step in the formation of translationally competent ribosomes at the 5′ AUG start sites of mRNA transcripts.6–8, 12, 13 A key regulator of eIF4E · eIF4G complex formation is the eIF4E-binding protein 1 (4E-BP1). Binding of 4E-BP1 to eIF4E prevents the recruitment of eIF4G to the cap-binding complex and thereby inhibits translation initiation. Phosphorylation of 4E-BP1 leads to its dissociation from eIF4E, which can then bind eIF4G and promote translation. 4E-BP1 phosphorylation is mediated by the target of rapamycin (TOR) kinase, which integrates signals from growth factors, hormones, and nutrients and regulates diverse processes including growth and proliferation.7, 14 In many systems, inhibition of TOR with the immunosuppressive drug rapamycin (sirolimus) leads to diminished translation initiation and protein synthesis. In addition to phosphorylation of 4E-BP1, other TOR-dependent and TOR-independent mechanisms are believed to regulate eIF4F, including the phosphorylation of eIF4E and eIF4G.6–8, 12

The phosphorylation of translation initiation proteins and the regulation of eIF4F complex assembly during cell proliferation have been primarily studied in tissue culture models. Although a number of previous papers have documented the regulation of eIF4F in the liver in response to dietary and hormonal stimuli, prior studies have not examined this complex in models of liver regeneration. A previous report by Jiang et al. examined the effect of rapamycin treatment on 4E-BP1 phosphorylation and eIF4E · 4E-BP1 complex formation after PH in the rat.15 Surprisingly, they found that rapamycin increased 4E-BP1 phosphorylation and decreased its binding to eIF4E, in apparent contradiction to results from other systems. In the current study, we examined the regulation of eIF4F components after PH in mice; this evokes a similar growth and proliferation response to that in the rat. We find that PH induced little change in 4E-BP1 phosphorylation but promoted eIF4E · eIF4G complex assembly, phosphorylation of eIF4G, and association of PABP and eIF4A with the eIF4F complex. Each of these effects was markedly inhibited by rapamycin. These results provide further insight into the regulation of translation initiation in the regenerating liver and suggest that this process may be distinctly regulated in different species.


PH, 70% partial hepatectomy; eIF, eukaryotic initiation factor; mRNA, messenger RNA; PABP, poly-A binding protein; 4E-BP1, eIF4E-binding protein 1; TOR, target of rapamycin; SDS, sodium dodecyl sulfate; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; AP, alkaline phosphatase; Ig, immunoglobulin; RT-PCR, real-time polymerase chain reaction.

Materials and Methods


Male Balb/c mice (Harlan Sprague-Dawley) were housed in an Association for Assessment and Accreditation of Laboratory Animal Care-approved facility and provided laboratory chow and water ad libitum. At 8 weeks, PH or sham surgery was performed as previously described,16 between the hours of 3 and 5 PM. Rapamycin or vehicle (DMSO) was administered at a dose of 1.5 mg/kg/d by intraperitoneal injection as previously described,17 beginning 2 hours prior to PH. Nonhepatectomized mice (0 hours) received DMSO or rapamycin on the same schedule as the mice that underwent PH and harvest at 42 hours. Two hours prior to harvest, the animals were injected with bromodeoxyuridine (50 mg/kg) by intraperitoneal injection. For each time point studied, 3 to 5 animals were harvested. All animal studies were performed in accordance with Institutional Animal Care and Use Committees approval and National Institutes of Health guidelines.

Liver Harvest and Homogenization.

Immediately after harvest, 500 mg of liver tissue was homogenized in 7 volumes of eIF4E buffer (20 mmol HEPES, pH 7.4; 2 mmol EGTA; 50 mmol NaF; 100 mmol KCl; 0.2 mmol ethylenediaminetetraacetic acid [EDTA]; 50 mmol β-glycerophosphate) containing protease and phosphatase inhibitors (1 mmol DTT, 1 mmol benzamidine, 1 mmol Na3VO4 , 87 μg/mL phenylmethylsulfonyl fluoride, 1 μg/mL aprotinin, 0.1 μg/mL leupeptin, 1 μg/mL pepstatin). The tissue was homogenized using a VirtiShear homogenizer (Virtis Co., Gardiner, NY). The homogenates were cleared by centrifugation. The protein concentration of the supernatants was measured using BioRad DC protein reagent (BioRad, Hercules, CA) following the manufacturer's protocol. Fresh supernatants were used for eIF4E immunoprecipitation, and the remainder were stored as aliquots at −80°C.

Fresh Lysate Immunoprecipitation.

eIF4E immunoprecipitation from fresh liver lysates was performed as previously described.18, 19 In brief, 500 μL of each lysate (containing 7 mg of protein in eIF4E buffer) was incubated with 75 μL eIF4E monoclonal antibody, 175 μL phosphate buffered saline, and 12.5 μL Triton-X 100 in a 1.5 mL microcentrifuge tube at 4°C overnight. BioMag Goat Anti-Mouse IgG magnetic beads (310004; Qiagen, Valencia, CA) were prepared by washing 1 mL BioMag beads per sample. The beads were washed 3 times with 5 mL low salt buffer (20 mmol Tris-HCl, pH 7.4; 150 mmol NaCl; 5 mmol EDTA; 0.5% Triton-X100; 0.1% β-mercaptoethanol) by placing them in a 50 mL tube in a magnetic rack. After washing, the beads were resuspended in low salt buffer plus 0.1% nonfat dried milk, and 500 μL of the resuspended beads were added to the sample tubes from the night before. The tubes were allowed to rotate for 1 hour at 4°C.

After the 1 hour incubation time, beads were collected using a magnetic tube rack. The supernatant was removed and the beads were washed, twice with low salt buffer and once with high salt buffer (50 mmol Tris-HCl, pH 7.4; 500 mmol NaCl; 5 mmol EDTA; 1% Triton-X100; 0.1% sodium dodecyl sulfate [SDS]; 0.04% β-mercaptoethanol). The beads were washed by adding 500 μL of buffer to the pellet and rotating the tube 180°. The beads were resuspended in 100 μL SDS sample buffer, boiled for 5 minutes, centrifuged briefly, and removed from the supernatant. The supernatant was placed in a new tube and stored at −80°C.

m7GTP Binding Assays.

Lysates (1 mg in 300 μL of buffer) were precleared with 30 μL bed volume of Protein A Sepharose 4 Fast Flow beads (Amersham, Piscataway, NJ) for 1 hour at 4°C. After removal of the protein A beads, 20 μL bed volume of 7-methyl GTP Sepharose 4B beads (Amersham) were added, and the samples were rocked at 4°C overnight. The beads were washed 3 times, resuspended in SDS sample buffer, and stored at −20°C until used for Western blots.

Western Blot Analysis.

Lysates (50 μg) and IP samples (20 μL) were loaded on 8% or 12% SDS-polyacrylamide gel electrophoresis (PAGE) gels for Western blot analysis as previously described.16, 17 The following additional antibodies were used: mouse monoclonal anti-eIF4E (610269; BD Biosciences, Palo Alto, CA), rabbit polyclonal anti-4EBP1 (9452; Cell Signaling, Beverly, MA), rabbit polyclonal anti-PABP and anti-eIF4A1,20, 21 rabbit polyclonal anti-eIF4G,18, 19 rabbit polyclonal anti–phospho-eIF4E (Ser209; 9741, Cell Signaling), and anti–phospho-eIF4G (Ser1108; 2441, Cell Signaling). Two alkaline phosphatase (AP)-conjugated secondary antibodies from Applied Biosystems (Bedford, MA) were used: AP-conjugated goat anti-rabbit immunoglobulin G (IgG; AC31RL), and AP-conjugated goat anti-mouse IgG + IgM (AC32ML). The chemiluminescent CDP-Star kit (S250R, Applied Biosystems) was used to illuminate the Western blots, and the signals were captured using an Image Station 2000R (Kodak, Rochester, NY). Densitometry was performed using the Kodak 1D Image Analysis Software according to the manufacturer's instructions. Three to 5 samples were used for densitometry determinations, and the expression of each immunoprecipitated protein was normalized to the eIF4E content in the precipitated sample, as previously described.18, 19 Densitometry values are shown as the mean ± SD expression relative to DMSO-treated liver (Figs. 2, 3, and 5).

Figure 2.

Effect of PH and rapamycin treatment on the binding of 4E-BP1 to eIF4E in the liver. Liver extracts obtained at the indicated time points were subjected to eIF4E immunoprecipitation followed by Western blot analysis and densitometry as described in Materials and Methods. Values represent the mean (± SE) of 4E-BP1 protein present in the immunoprecipates, normalized for the recovery of eIF4E . A representative western blot of the immunoprecipitates is shown (inset). *P < .04. **P < .03 as compared to vehicle-treated mice.

Figure 3.

Regulation of hepatic eIF4E · eIF4G complex formation by PH and rapamycin. Liver extracts were subjected to eIF4E immunoprecipitation followed by Western blot analysis and densitometry as described in Fig. 2. (A) Total eIF4G expression in the complexes (normalized to eIF4E). *P < .04 as compared to all other groups. (B) Ser1108-phosphorylated eIF4G in the precipitates. *P < .01.

Figure 5.

Effect of PH and rapamycin on the association of PABP and eIF4A1 with eIF4E in the liver. eIF4E immunoprecipitates were subjected to Western blot analysis as in Figs. 2 and 3. (A) PABP content in the complexes *P < .03. (B) eIF4A1 content in the complexes *P < .05.

Quantification of Cyclin D1 mRNA by Real-Time Polymerase Chain Reaction (RT-PCR).

Total RNA from liver specimens was extracted and quantified as previously described.16 RNA samples were electrophoresed on a 0.7% agarose 2.2 mol formaldehyde gel and visualized by ethidium bromide staining to ensure that there was not overt RNA degradation. An aliquot of each RNA (5 μg) was treated with DNAse I (DNA-free; Ambion, Austin, TX) according to the manufacturer's instructions. Oligo-dT primed complementary DNA was generated from 4 μg of each RNA with TaqMan Reverse Transcriptase Reagent kit (Applied Biosystems). Mouse cyclin-D1 DNA sequences for upper (5′-GCGTACCCTGACACCAATCT) and lower (3′-ATCTCCTTCTGCACGCACTT) primers and mouse β-actin sequences for upper (5′-AACCCTAAGGCCAACCGTGAAAAG) and lower (3′-ACCGCTCGTTGCCAATAGTGATGA) primers were selected using the PrimerSelect program (DNASTAR, Inc., Madison, WI), and the resulting sequences were synthesized in the University of Minnesota microchemical facility and purified by HPLC. RT-PCR was completed using a LightCycler FastStart DNA Master SYBR Green I Kit (Roche Molecular Biochemicals, Indianapolis, IN). Samples were denatured for 10 minutes at 95°C, then 40 cycles of 95°C for 20 seconds, 60°C for 20 seconds, and 68 °C for 20 seconds. Optimization of MgCl2 and primer concentrations were completed as recommended by the manufacturer with 3 mmol for MgCl2, and 0.5 μmol and 0.1 μmol for cyclin D1 and β-actin, respectively. For each mRNA, quantification was completed by comparison (linear interpolation) of the cycles to saturation in each sample. Cyclin-D1 mRNA was normalized to β-actin mRNA in each sample, and the relative amounts were quantified as recommended by the manufacturer.


The response to 70% PH has been well characterized in rodent species and involves the relatively synchronous progression of the remnant hepatocytes into the cell cycle.1, 2 Previous studies have shown that in the mouse, hepatocytes enter S phase at 1.5 to 2 days after PH.1, 2, 16 The proliferative response after PH is accompanied by up-regulation of cell cycle control proteins as well as the induction of global protein synthesis.1–4 Previous studies have shown that treatment of rats or mice with the specific TOR inhibitor rapamycin inhibits protein synthesis, cell cycle protein expression, hepatocyte proliferation, and liver growth after PH.15, 17, 22, 23 To further explore the regulation of translation initiation in liver regeneration, we examined eIF4F complex assembly after PH and its modulation by rapamycin. Mice underwent PH with tissue harvest at time points corresponding to G1 phase (24 hours) and S phase (42 hours). We compared mice treated with vehicle alone, which did not induce any changes compared to untreated animals (data not shown), to mice treated with rapamycin.

Prior studies have documented that the expression and phosphorylation of the S6 protein are dependent on TOR activity,15, 17, 23, 24 and as expected, Western blot analysis demonstrated that rapamycin prevented the induction of S6 expression and phosphorylation after PH (Fig. 1). As previously shown,17 PH induced the expression of cyclin D1, cyclin A and PCNA in a rapamycin-sensitive manner (Fig. 1). We then examined the expression and phosphorylation of proteins involved in eIF4F complex assembly. When 4E-BP1 is subjected to SDS-PAGE, it separates into 3 distinct phosphorylated forms19, 25, 26: the slowest migrating band (γ) corresponds to the most highly phosphorylated form that does not bind eIF4E, and the faster-migrating bands (β and α) represent hypophosphorylated forms that can bind eIF4E. After PH, a modest increase in the amount of the hyperphosphorylated (γ) form was noted. As previously shown in regenerating rat liver, the abundance of this form was not diminished by rapamycin treatment.15 However, rapamycin markedly increased expression of hypophosphorylated 4E-BP1 in liver extracts, particularly the α form. These findings suggest that TOR inhibition increases the abundance of 4E-BP1 forms capable of binding eIF4E in the regenerating liver.

Figure 1.

Expression of cell cycle and translation initiation proteins in regenerating mouse liver. BALB/c mice were treated with vehicle (DMSO) or rapamycin (1.5 mg/kg/d) and underwent 70% partial hepatectomy (PH) as described in Materials and Methods (A) Liver extracts (50 μg) were isolated at the indicated time points and representative samples were subjected to Western blot analysis using the indicated antibodies. The 3 distinct phosphorylated forms of 4E-BP1 are indicated. (B) The expression of Ser1108-phosphorylated eIF4G (normalized to total eIF4G) was quantified by densitometry. *P < .003.

As shown in other systems,7, 8, 15 eIF4E expression was invariant after PH and was not influenced by rapamycin treatment (Fig. 1). In proliferating cells, eIF4E can be phosphorylated at ser209 in an ERK-dependent manner.27, 28 In the regenerating mouse liver, the abundance of Ser209-phosphorylated eIF4E did not vary after PH but was modestly induced by rapamycin at all time points. The expression of eIF4G was modestly induced after PH but not substantially decreased in the rapamycin-treated mice. In contrast, the Ser1108-phosphorylated form of eIF4G was markedly induced after PH, and this response was inhibited by rapamycin treatment. In animals that underwent sham surgery, no change in the expression of cell cycle proteins or eIF4F components (including phosphorylated eIF4G) was observed (data not shown). The expression of eIF4A1 and PABP did not substantially vary after PH or rapamycin treatment. Thus, by Western blot analysis of eIF4F components, eIF4G phosphorylation was the most striking change induced by PH. The most apparent effects of rapamycin treatment were inhibition of eIF4G phosphorylation and increased expression of the β and α forms of 4E-BP1.

To examine the regulation of eIF4F complex assembly in the regenerating liver, we performed immunoprecipitation of eIF4E from liver lysates using established techniques.18, 19 The precipitated protein complexes were then subjected to Western blot and densitometry analysis to determine changes in complex formation. As predicted, the amount of eIF4E recovered in precipitates did not vary between the conditions examined in this study (data not shown). We first examined the association of 4E-BP1 and eIF4E (Fig. 2). Compared to resting liver, PH did not induce any detectable change in eIF4E · 4E-BP1 complex formation in vehicle-treated mice. Rapamycin treatment led to an approximately 2-fold increase in eIF4E · 4E-BP1 complex formation in resting liver, and an approximately 3-fold increase after PH. The rapamycin-induced increase in these complexes was primarily due to enhanced binding of the least-phosphorylated (α) form.

The best-characterized mechanism that promotes attachment of the mRNA to the 40S ribosomal subunit involves the reversible binding of eIF4E to eIF4G.7, 12, 13, 19 In Fig. 3A, we observed a 2.3-fold increase in eIF4E · eIF4G complex formation after PH; this was inhibited by rapamycin. The Ser1108-phosphorylated form of eIF4G, which may facilitate binding to eIF4E,7 was markedly induced in the complexes (Fig. 3B). In sham-operated mice, no change in the binding of total or phosphorylated eIF4G to eIF4E was observed (data not shown). These results demonstrate for the first time that eIF4F abundance is increased in the regenerating liver.

The results differ substantially from a prior report on regenerating rat liver, in which rapamycin paradoxically led to diminished eIF4E · 4E-BP1 complex formation.15 In that report, complexes were precipitated with sepharose beads coupled to m7GTP, which mimics the 5′ cap on mRNA. In Fig. 4, we used this technique to precipitate eIF4E from liver extracts and again found that PH did not substantially alter eIF4E · 4E-BP1 complex formation, while rapamycin increased the abundance of these complexes. Furthermore, PH led to increased binding of eIF4G to the m7GTP beads, indicating an induction of functional eIF4F complexes capable of binding capped mRNA, and this response was inhibited by rapamycin. Thus, as predicted by studies in numerous other systems,7, 8, 14 inhibition of TOR led to the accumulation of hypophosphorylated 4E-BP1 and increased binding of this protein to eIF4E in regenerating mouse liver.

Figure 4.

m7GTP-sepharose chromatography. Liver extracts were subjected to m7GTP binding assays as described in Materials and Methods, and the precipitated proteins were visualized by Western blot analysis as shown.

To further study the regulation of eIF4F in the regenerating liver, we examined the binding of additional subunits to the complex. In Fig. 5, we found that the abundance of PABP and eIF4A1 in eIF4E immunoprecipitates increased after PH, in a manner similar to eIF4G. This suggests that the increase in eIF4E · eIF4G complex formation after PH is functionally significant because it results in recruitment of additional translation initiation factors into the complex.

To determine whether the inhibition of eIF4F complex formation by rapamycin affected the expression of a known target gene in the regenerating liver, we examined cyclin D1. Previous studies in cultured rat hepatocytes,17 as well as in other cell culture systems,29 have indicated that rapamycin inhibits cyclin D1 at the level of protein but not mRNA expression. This is presumably due to a reduction in functional eIF4F complex formation because the cyclin-D1 mRNA contains a complex 5′ untranslated region that is particularly dependent on eIF4F.30, 31 As previously shown,17 rapamycin inhibited the induction of cyclin-D1 protein by more than 90%, whereas the abundance of cyclin-D1 mRNA after PH was not altered by this treatment (Fig. 6). These data suggest that the translation of cyclin D1 (and possibly other mRNAs that are highly dependent on eIF4F) is inhibited by rapamycin in the regenerating mouse liver.

Figure 6.

Effect of rapamycin on cyclin D1. (A) Cyclin D1 protein expression in liver lysates was evaluated by Western blot analysis and quantified by densitometry *P < .003 as compared to all other samples. (B) Cyclin-D1 mRNA expression was quantified by RT-PCR as described in Materials and Methods and normalized to the expression of β-actin mRNA in each sample *P < .001 as compared to 0-hour animals.


After PH or other injuries that diminish parenchymal liver mass, the remaining hepatocytes must substantially up-regulate the rate of protein synthesis to meet increased metabolic demands and to accommodate rapid liver growth.1–4 The mechanisms that govern enhanced hepatic protein synthesis in liver regeneration have not been well characterized. Older studies have documented an increased rate of ribosome biogenesis after PH; this augments the capacity for protein translation.4, 9–11 The current results suggest a second mechanism by which protein synthesis is enhanced in the regenerating liver, through increased assembly of the eIF4F translation initiation complex. This complex may be an important target of stimuli that govern liver regeneration under physiological and pathological conditions.

Numerous agents that have been shown to regulate translation initiation in the liver (e.g., diet, hormones, endotoxin, alcohol) also modulate liver regeneration.1, 2, 12, 13, 19, 32 It is tempting to speculate that regulation of the protein synthetic apparatus by these agents plays a role in altering the regenerative response to PH and other liver injuries. There are abundant data from genetic studies to implicate translation initiation factors in the control of cell and tissue growth.7, 8, 31 Furthermore, key components of the cell cycle apparatus are regulated at the level of translation and may act as “sensors” of translational capacity that permit proliferation only under conditions that allow sufficient protein synthesis.31, 33 This is most evident for the G1 cyclins, which play an important role in the cellular “decision” to proceed through replication.34 In yeast systems, the G1 cyclin CLN3 is significantly regulated at the level of translation, and down-regulation of this protein plays an important role in inhibiting proliferation under conditions that impair protein synthesis.35–37 Previous studies have shown that cyclin-D1 expression is regulated by eIF4E in several types of cells, indicating that this cell cycle control protein is responsive to alterations in the translational apparatus.30, 31 In the regenerating liver, conditional knockout of the ribosomal protein S6 gene markedly inhibits cyclin-E expression and hepatocyte proliferation, suggesting that cyclin E is a sensor of ribosome function.11 We have recently found that down-regulation of cyclin D1 plays an apparently important role in the cell cycle arrest produced by amino acid deficiency or rapamycin treatment in models of hepatocyte proliferation.17, 38 These studies suggest that perturbations of the ribosome or translation initiation apparatus significantly regulate cell cycle progression in the liver and that further investigation of the protein synthetic machinery may provide insight into mechanisms of altered liver regeneration.

The current data indicate that eIF4E · eIF4G complex assembly was induced in the regenerating mouse liver, but this was not due to comparable changes in 4E-BP1 phosphorylation or binding. Assembly of eIF4E · 4E-BP1 complexes is the most clearly characterized mechanism of eIF4F regulation.6–8, 12 However, numerous studies in other systems indicate that eIF4E · eIF4G complexes can be regulated independently of 4E-BP1.8 For example, eIF4G phosphorylation, which is rapamycin-sensitive, may promote its association with eIF4E.7 Furthermore, additional eIF4E binding proteins have been identified, but their role has not been clearly defined.7, 12 Finally, other TOR-independent mechanisms appear to regulate eIF4F.6–8, 12, 39 Further study of the factors that regulate eIF4F (which will depend in part on the development of new reagents) will likely provide insight into the mechanisms that govern translation initiation in the liver.

Our results suggest that TOR activity is necessary for the induction of eIF4F in the regenerating liver. However, we do not yet have a clear understanding of the regulation of TOR activity during liver regeneration. Assays to determine the activity of endogenous TOR are not widely available (to our knowledge), and therefore indirect assessments of TOR activity have been frequently reported in the translation control literature.7, 14, 15, 23 S6 and 4E-BP1 phosphorylation have been used as surrogate markers of TOR signaling because both events are regulated in part by the activity of this kinase. In the current study, the dramatic increase in S6 phosphorylation was markedly greater than the change in 4E-BP1 phosphorylation after PH, suggesting that the 2 events are not regulated in parallel. Activation of S6 kinases is thought to require TOR activity, but other mitogen-stimulated pathways also regulate these kinases.5, 7 Although 4E-BP1 phosphorylation can be mediated by TOR, other kinases are likely to regulate phosphorylation of this protein, possibly including cyclin-dependent kinase 1.7, 8, 40 In the current study, rapamycin substantially increased the appearance of the least phosphorylated (α) species of 4E-BP1, confirming that TOR activity plays a role. However, rapamycin did not diminish the appearance of the hyperphosphorylated (γ) form, indicating that TOR-independent pathways also regulate phosphorylation of 4E-BP1. In addition to its effect on 4E-BP1, rapamycin potently inhibited the appearance of Ser1108-phosphorylated eIF4G in the regenerating liver; this may have affected eIF4F assembly.7 The control of TOR and downstream mediators is complex and incompletely understood,5, 7, 14 and further studies are required to delineate the mechanisms that regulate this kinase during liver regeneration.

These results differ in several regards from a prior paper by Jiang et al.15 In that study, 4E-BP1 phosphorylation was assessed in regenerating rat liver after 70% PH in the presence or absence of rapamycin. Binding of 4E-BP1 to eIF4E was determined after precipitating the complexes with m7GTP-sepharose. In contrast to our results, the abundance of the hyperphosphorylated (γ) form increased substantially after PH, and eIF4E · 4EBP1 complex assembly was dramatically reduced. Surprisingly, rapamycin paradoxically increased the abundance of the hyperphosphorylated (γ) form of 4E-BP1 in the rat PH model. These results were confirmed using antibodies that recognized specific phosphorylated forms of 4E-BP1 (Threonine 36/45, serine 64, and Threonine 69). Consistent with these findings, rapamycin decreased eIF4E · 4EBP1 complex formation in the regenerating rat liver.

A clear explanation for the differences between the current results and the study by Jiang et al.15 is not readily apparent. In the present study, we focused on a time point corresponding to peak DNA synthesis (42 hours) following PH in the mouse; the corresponding time point (24 hours) was used in the rat PH studies. We have also examined a time point corresponding to G1 phase (24 hours) in mice and found similar patterns of eIF4E · 4EBP1 and eIF4F complex formation in both control and rapamycin-treated livers (data not shown), suggesting a consistent effect in the mouse model. The differences are not easily explained by technical factors because we obtained similar results using eIF4E immunoprecipitation or m7GTP-sepharose chromatography. In the current study, we used a higher dose of rapamycin (1.5 vs. 0.4 mg/kg/d); this may have resulted in more complete inhibition of TOR effects. Another possibility is that rapamycin pharmacokinetics and pathways of 4E-BP1 phosphorylation may differ between species. Our data are consistent with studies in other systems showing that rapamycin significantly diminishes functional 4E-BP1 phosphorylation, promotes eIF4E · 4E-BP1 binding, and inhibits eIF4F complex assembly.6–8, 12, 41

In summary, these results demonstrate that eIF4F complex formation is enhanced in the regenerating mouse liver in a TOR-dependent fashion, suggesting an additional mechanism by which global protein synthesis is enhanced during liver regeneration. Accumulating data indicate that translation initiation regulates diverse biological processes including differentiation, growth, cell proliferation, glucose homeostasis, viral infection, apoptosis, and carcinogenesis.5–8, 12, 13, 30, 31 Efforts are currently underway to develop therapeutic agents targeting components of the translational apparatus for diseases such as diabetes and cancer.12, 29, 41 A better understanding of the mechanisms that regulate translation initiation in the liver may facilitate the use of novel agents to promote hepatic regeneration, inhibit carcinogenesis, or ameliorate liver injury.


The authors thank Dr. Richard Z. Lin for helpful discussions about these experiments.