Amino acids regulate hepatocyte proliferation through modulation of cyclin D1 expression.

The mechanisms by which amino acids regulate the cell cycle are not well characterized. In this study, we examined the control of hepatocyte proliferation by amino acids and protein intake. In short-term culture, hepatocytes demonstrated normal entry into S phase and cell cycle protein expression in the absence of essential amino acids. However, deprivation of a set of nonessential amino acids (NEAA) potently inhibited cell cycle progression and selectively down-regulated the expression of proliferation-control proteins. Notably, NEAA withdrawal after the mitogen restriction point still inhibited entry into S phase, suggesting that these amino acids regulate a distinct checkpoint. Cyclin D1, an important mediator of hepatocyte proliferation, was markedly inhibited at the transcriptional level by NEAA deprivation, and transfection with cyclin D1 (but not cyclin E) overcame the cell cycle arrest. As previously shown, protein-deprived mice demonstrated impaired hepatocyte proliferation in vivo after 70% partial hepatectomy. The expression of cyclin D1 and downstream cell cycle proteins after partial hepatectomy was inhibited in these mice. Transfection with cyclin D1 in vivo triggered hepatocyte DNA synthesis and the expression of S phase proteins in the absence of dietary protein. Cyclin D1 also induced global protein synthesis in NEAA-deprived hepatocytes and promoted liver growth in vivo in the setting of protein deprivation. These results indicate that cyclin D1 is a key target of amino acid signaling in hepatocytes.

In addition to providing the substrates for protein synthesis, amino acids serve as regulatory molecules that modulate numerous cellular functions (1). It has long been recognized that the availability of amino acids controls growth and proliferation in cell culture systems, but the mechanisms have not been well defined (2,3). A substantial body of literature indicates that amino acids regulate overall protein synthesis, an essential component of growth and proliferation, through several pathways that control the translational apparatus (4,5). However, the downstream elements that link amino acids to the cell cycle remain to be determined. In yeast systems, nutrient insufficiency inhibits the expression of G 1 cyclins, which are required for normal entry into S phase (6 -9). However, a similar mechanism has not been clearly established in mammalian cells.
In the normal adult liver, hepatocytes rarely replicate, but these cells rapidly enter the cell cycle following injuries that reduce functional liver mass (10). Hepatocyte proliferation is an important component of the adaptive response to liver diseases. In the best-studied model of hepatocyte proliferation in vivo, that of 70% partial hepatectomy (PH) 1 in rodents, most of the remaining hepatocytes enter the cell cycle in a relatively synchronous manner, and liver mass is restored within 1-2 weeks. In addition, primary hepatocytes in short-term culture proliferate readily in response to appropriate mitogens. Thus the hepatocyte culture and PH models offer excellent systems to study proliferation of normal parenchymal cells. Older studies (11,12) have shown that amino acid withdrawal inhibits hepatocyte proliferation in culture, and protein deprivation impairs liver regeneration after PH. However, these previous studies did not identify potential intracellular mediators of this antiproliferative response.
Cell cycle progression is controlled by protein kinase complexes consisting of cyclins, cyclin-dependent kinases (cdks), and associated regulatory proteins (reviewed in Refs. 2 and 3). During G 1 phase, mitogens up-regulate expression of the D-type cyclins, which bind cdk4 and cdk6 to form active kinases that phosphorylate the retinoblastoma protein (Rb) and the related proteins p107 and p130. This is followed by activation of cyclin E/cdk2 in late G 1 phase, which phosphorylates Rb at different sites. The combined phosphorylation of Rb leads to derepression of E2F transcription factors, which promote entry into S phase. Phosphorylation of Rb may represent the biochemical basis of the mitogen restriction point in late G 1 phase, when the cell no longer requires mitogens to complete the cell cycle. Antiproliferative signals impact on cyclin/cdk activity through several mechanisms, including decreased cyclin expression, changes in cdk phosphorylation, and induction of cdk inhibitory proteins.
Previous studies have suggested that cyclin D1 plays an important role in hepatocyte proliferation (10,(13)(14)(15)(16)(17). In the current study, we examined whether this protein might also be a target of amino acid-dependent signaling. Our results suggest that cyclin D1 is significantly regulated by selective amino acid withdrawal in culture and dietary protein deprivation in vivo. Furthermore, cyclin D1 expression induced growth and proliferation in the absence of normal amino acids or dietary protein.
The results suggest that in mammalian cells, cyclin D1 regulates cell cycle progression in response to nutrients, similar to the previously characterized response in yeast.
Western and Northern Blot Analysis-Liver tissue and hepatocyte harvest, Western blots, and Northern blot analysis were performed as previously described (14,15,20). The antibody to CHOP was obtained from Santa Cruz Biotechnology.
Cyclin D1 Promoter-luciferase Assays-The Ϫ1745 CD1LUC plasmid containing the human cyclin D1 promoter was previously described (21). At 48 h after plating, hepatocytes on 35 mm-dishes were transfected in triplicate with 1 g/ml of the Ϫ1745 CD1LUC (firefly) plasmid and the 1 g/ml pRL-null (Renilla control) plasmid (Promega) using the FuGENE 6 reagent (Roche Applied Science) as recommended by the manufacturer. At 72 h after plating, cells were harvested following the instructions included in the dual-luciferase reporter assay system (Promega). Luminescence assays were performed using an EG&G Berthold 9507 luminometer. Firefly luciferase was normalized to Renilla luciferase activity.
Animals-Male BALB/c mice were purchased from Harlan Sprague-Dawley. At 8 weeks of age, PH or adenovirus-mediated transfection was performed as previously outlined (16,19). Rapamycin (or Me 2 SO vehicle alone) was administered at a dose of 1.5 mg/kg/day beginning 2 h prior to PH as previously described (17). Liver harvest, BrdUrd immunohistochemistry, and tissue homogenization were performed as described elsewhere (19).
In Figs. 6 and 8, mice were either provided normal laboratory chow and water ad libitum or provided only 10% dextrose in the drinking water beginning 24 h prior to PH or adenovirus injection (22). In Fig. 7, the protein-deprived mice were provided a protein-free chow (Teklad) beginning 24 h before PH. The daily amount of food intake was recorded every 24 h. The protein-fed mice received an isocaloric chow containing 18% casein (Teklad). For each 24 h period, the protein-fed mice received an amount of chow equal to the average intake of the protein-deprived group, divided in three equal doses.

Selective Deprivation of Nonessential Amino Acids Inhibits
Cyclin D1 and Cell Cycle Progression in Cultured Hepatocytes-To explore potential mechanisms by which amino acids regulate the hepatocyte cell cycle, we prepared culture media in which amino acids had been selectively omitted by using commercially available kits. One mixture contained 12 amino acids, including those that are essential for human adults (EAA). In addition, we used a mixture of seven other nonessential amino acids (NEAA). These kits do not include glutamine, which was added to all experiments. We examined mitogenstimulated primary rat hepatocytes in culture, which is a well established model of hepatocyte proliferation (10). When cultured in the presence of EGF and insulin, these cells demonstrate a relatively long G 1 phase of about 48 h and undergo peak DNA synthesis (S phase) at 72 h (13,14).
Somewhat surprisingly, in these short-term experiments omission of EAA had little effect on hepatocyte proliferation as measured by DNA synthesis at 72 h or on the expression of cell cycle proteins (Fig. 1). On the other hand, omission of NEAA markedly inhibited DNA synthesis and selectively down-regulated the expression of cell cycle proteins, including cyclin D1. Previous studies have found that cyclin E, proliferating cell nuclear antigen, and p21 are induced downstream of cyclin D1 in hepatocytes (14,16), and these proteins were down-regulated in response to NEAA deprivation. CHOP, a stress response protein which is up-regulated in response to amino acid deprivation (4), was induced in the absence of NEAA. Addition of twice the normal concentration of EAA did not overcome the cell cycle arrest induced by NEAA deprivation (data not shown), suggesting that a lack of calories or overall amino acid load was not responsible for the effect. These studies indicate that selective amino acid deprivation inhibits cyclin D1 expression and cell cycle progression in hepatocytes. We then used the NEAA deprivation conditions to further examine the role and regulation of cyclin D1 in this response.
NEAA Regulate a Distinct Cell Cycle Checkpoint-Studies in other systems (23,24) suggest that amino acid availability regulates the activity of the target of rapamycin (TOR) protein, which modulates cell cycle progression and protein synthesis in many types of cells, including hepatocytes. We have recently found that cyclin D1 is a key rapamycin-dependent mediator of proliferation in hepatocytes (17). To examine whether inhibition of TOR with rapamycin and NEAA withdrawal had similar effects, we evaluated phosphorylation of the S6 protein, which is regulated by TOR (24). As previously shown, rapamycin treatment inhibited S6 phosphorylation (Fig. 1D) (17,25). As noted in other systems (26), rapamycin also diminished the expression of total S6 protein. NEAA deprivation also inhibited S6 phosphorylation, although not as completely as rapamycin treatment. This suggests that NEAA may regulate hepatocyte proliferation in part through TOR. However, as shown below, NEAA withdrawal and rapamycin treatment have distinct effects on cyclin D1 expression and hepatocyte cell cycle progression.
Previous studies have shown that mitogen-stimulated rat hepatocytes progress through the G 1 restriction point at ϳ40 -44 h after plating, which corresponds with induction of cyclin D1 (13,14). In Fig. 2A, we confirmed that withdrawal of EGF/insulin after the restriction point (i.e. at 48 h) did not significantly alter DNA synthesis or cyclin D1 expression at 72 h. However, withdrawal of NEAA at 48 h potently inhibited DNA synthesis and cyclin D1 expression at 72 h. We have previously shown that addition of rapamycin did not inhibit cell cycle progression if added after the restriction point (17); as is shown in Fig. 2B, addition of rapamycin at 48 h did not substantially inhibit either DNA synthesis or the expression of cyclin D1 at 72 h. These data suggest that NEAA regulate a cell cycle checkpoint that is distinct from the mitogen restriction point and that rapamycin treatment and NEAA deprivation do not have equivalent effects on cell cycle progression.
To further characterize the regulation of cyclin D1 by NEAA, we performed Northern blot analysis of cells cultured in the presence or absence of EGF/insulin or NEAA (Fig. 3A). Cyclin D1 mRNA was markedly down-regulated in the absence of mitogens or NEAA. In addition, the activity of a cyclin D1 promoter-luciferase reporter gene was inhibited 59% by NEAA deprivation (Fig. 3B, p Ͻ 0.007). These data suggest that NEAA regulate cyclin D1 expression, at least in part, at the level of transcription. In addition, because rapamycin inhibits hepatocyte cyclin D1 expression at the level of protein but not mRNA (17,23), these data suggest that NEAA deprivation and rapamycin treatment regulate cyclin D1 through distinct mechanisms.
Cyclin D1 Promotes Cell Cycle Progression and Protein Synthesis in NEAA-deprived Hepatocytes-To further address the question of whether cyclin D1 is the key target of NEAAmediated signaling in hepatocytes, we transfected cells with an adenovirus encoding cyclin D1. Previous studies have demonstrated that recombinant adenoviruses readily transfect hepatocytes in culture and in vivo (14 -16, 27). We found that cyclin D1 expression was sufficient to induce DNA synthesis and the expression of downstream cell cycle proteins in the absence of NEAA (Fig. 4A). Transfection with cyclin E, on the other hand, did not promote cell cycle progression (Fig. 4B). These results suggest that cyclin D1 is the key NEAA-dependent mediator of cell cycle progression in hepatocytes.
In addition to promoting proliferation, recent studies have suggested that cyclin D1 can induce cell growth, and by inference, cellular protein synthesis (2,8,9). Based on previous studies, we anticipated that selective amino acid withdrawal would diminish global protein synthesis (4,5). As expected, NEAA deprivation inhibited the mitogen-stimulated induction of hepatocyte protein synthesis (as measured by [ 3 H]leucine incorporation into cellular proteins) by 60% (Fig. 5, p Ͻ 0.003). Transfection with cyclin D1 restored protein synthesis to a level similar to that seen in mitogen-stimulated cells in the presence of complete amino acids. These results suggest that cyclin D1 acts downstream of nutrients to induce protein synthesis under mitogenic conditions. Because enhanced protein synthesis is a requirement for cell growth (8,9,24), this further suggests that cyclin D1 plays a role in growth as well as cell cycle regulation.
Protein Deprivation Inhibits Cyclin D1 Expression and Liver Regeneration in Vivo-A number of previous studies (11,12) have demonstrated that normal liver regeneration depends on adequate intake of dietary protein and amino acids, but the molecular events linking nutritional status and hepatocyte cell cycle progression have not been determined. To examine this further, we performed standard 70% PH in mice under conditions of protein deprivation. In normal mice after PH, a large population of hepatocytes progresses through the cell cycle in a relatively synchronous fashion, with peak DNA synthesis occurring at 36 -42 h after the procedure (10,20). Previous studies have demonstrated that cyclin D1 is induced in G 1 phase in  (20). In Fig. 6, we examined mice that were fed normal diets or provided only 10% dextrose in the drinking water (to prevent hypoglycemia), adapting a previously described model (22). As expected (11,22), hepatocyte proliferation was markedly inhibited in the nutrient-deprived mice. At 42 h after PH, hepatocyte DNA synthesis (as measured by BrdUrd immunohistochemistry) was 26.7% in the fed mice as compared with 1.2% in the mice provided only dextrose (p Ͻ 0.001). The expression of cyclin D1 was markedly down-regulated, as was the induction of proteins acting downstream in the cell cycle. The inhibition of protein expression was selective, because other proteins (e.g. p27, S6, eIF-4E, and actin) were not affected. Interestingly, nutrient deprivation did not prevent the phosphorylation of S6 after PH, whereas rapamycin was inhibitory (as previously reported, Refs. 17,25). The induction of S6 phosphorylation after PH suggests either that nutrient intake does not regulate hepatocyte proliferation in vivo via TOR inhibition or that other pathways lead to phosphorylation of S6 during the process of liver regeneration. In either case, the results further suggest that nutrient deprivation and rapamycin treatment produce distinct effects in hepatocytes.
In the mice provided only 10% dextrose, we cannot rule out the possibility that overall calorie restriction or the absence of nutrients other than protein was responsible for the inhibitory effects. To clarify this, we performed a second experiment in which mice were fed chow containing no protein (Fig. 7). These mice were compared with mice fed an otherwise matched protein-containing chow and an equivalent caloric intake. Compared with the protein-fed mice, the protein-deprived mice showed diminished hepatocyte proliferation and cyclin D1 expression similar to that seen in the mice fed only dextrose. The diminution of hepatocyte DNA synthesis was slightly more pronounced in the mice provided 10% dextrose as compared with the mice fed protein-free chow, suggesting that calorie restriction may also have contributed somewhat to inhibitory effect in Fig. 6. However, even in the setting of isocaloric diets, cyclin D1 expression and cell cycle progression in the regenerating liver were inhibited by protein deprivation, consistent with prior studies demonstrating that protein intake significantly regulates liver regeneration (11,12). These data further support the concept that cyclin D1 is a target of amino acidregulated signaling in hepatocytes in vivo.
Cyclin D1 Promotes Hepatocyte Cell Cycle Progression and Growth in the Setting of Protein Deprivation in Vivo-Our results in primary hepatocytes suggest that expression of cyclin D1 is sufficient to promote both cell cycle progression and protein synthesis in the absence of NEAA. To examine whether a similar effect occurs in vivo, mice were injected with the cyclin D1 adenovirus. Prior studies have shown that intravenously injected adenoviruses transfect hepatocytes in vivo with high efficiency, and numerous investigators have used this system to study the effect of single-gene expression in the liver (27)(28)(29). Previous studies found that this method of transfection leads to expression of cyclin D1, activation of cyclin D1/ cdk4, hepatocyte proliferation, and liver growth in normal mice (16). In mice provided only 10% dextrose (Fig. 8) or fed proteinfree chow (data not shown), cyclin D1 induced the expression of S phase-specific genes and promoted hepatocyte proliferation. Furthermore, in the absence of dietary protein, cyclin D1 triggered growth of the liver by over 40% (p Ͻ 0.008). On the other hand, transfection with cyclin E or the control virus did not induce hepatocyte proliferation or liver growth. Thus, cyclin D1 expression was down-regulated by protein deprivation in the regenerating liver, and expression of cyclin D1 promoted hepatocyte proliferation and growth under these conditions. DISCUSSION Amino acids and other nutrients are known to regulate proliferation, but the link between nutrient sufficiency and the cell cycle machinery has not been clearly defined in a mammalian system. In the current study, we found that selective amino acid deprivation in culture and protein deprivation in vivo impaired hepatocyte cyclin D1 expression and that transfection of cyclin D1 promoted cell cycle progression under these conditions. In addition, cyclin D1 induced global protein synthesis in culture and liver growth in vivo in the absence of these nutrients. In contrast to studies in other systems, we found that amino acid deprivation regulated the hepatocyte cell cycle through mechanisms that were distinct from TOR inhibition with rapamycin. These results suggest that cyclin D1 is a pivotal regulator of cell growth and proliferation downstream of amino acids after mitogen stimulation, and they identify an apparently novel cell cycle checkpoint in hepatocytes.
To our knowledge, this is the first study to show that amino acids regulate cell cycle progression through modulation of cyclin D1. Previous studies have shown that the supply of nutrients (including amino acids) regulates expression of the G 1 cyclin CLN3 in the budding yeast Saccharomyces cerevisiae (6,7). CLN3 and cyclin D1 have similar biochemical functions and appear to play similar roles in integrating extracellular signals into the cell cycle (2,3,30). However, expression of CLN3 under poor growth conditions leads to cell division without comparable cell growth, resulting in smaller cell sizes (31). On the other hand, transgenic models of cyclin D overexpression in several species suggest that this protein induces both growth and proliferation, depending on the context (8,9). In the current study, cyclin D1 induced proliferation as well as global protein synthesis in NEAA-deprived hepatocytes and promoted liver growth in protein-deprived mice. These data suggest that cyclin D1 is a "sensor" of nutrient sufficiency that regulates cell cycle progression in hepatocytes, similar to the function ascribed to CLN3 (8). However, our studies also suggest that even under conditions of nutrient deprivation, constitutively expressed cyclin D1 is capable of "driving" the protein synthetic apparatus and cell growth (2,8).
Our results suggest that amino acids regulate cyclin D1 expression at the level of transcription, but further studies are required to fully characterize this response. Amino acid deprivation is thought to regulate at least two pathways, modulated by the TOR and GCN2 kinases, respectively (4,5). NEAA deprivation diminished S6 phosphorylation in cultured hepatocytes in a manner similar to rapamycin treatment, suggesting the possibility of TOR inhibition. However, addition of rapamycin after the mitogen restriction point had little effect on cyclin D1 expression or S phase entry, whereas withdrawal of NEAA was inhibitory, suggesting a distinct checkpoint. Furthermore, rapamycin down-regulates the expression of cyclin D1 protein but not its mRNA (17,23), whereas NEAA deprivation led to diminished expression of both cyclin D1 mRNA and protein. Thus, TOR inhibition does not seem sufficient to explain the effect of NEAA withdrawal. GCN2 is activated in response to amino acid deprivation in yeast and mammalian cells. A major target of GCN2 is the translation initiation factor eIF2␣, which is phosphorylated and inactivated in response to several types of cellular stress. Previous studies indicate that stimulation of another eIF2␣ kinase, PERK, by tunicamycin inhibits cyclin D1 expression primarily through diminished translation (32,33). Similarly, activation of protein kinase R, a distinct eIF2␣ kinase, also appears to inhibit cyclin D1 translation (34). In our studies, NEAA deprivation down-regulated cyclin D1 mRNA and transcription from a cyclin D1 promoterreporter gene, suggesting that impaired translation was not the primary mechanism. Another downstream effect of amino acid deprivation is induction of the CHOP protein, which inhibits proliferation through unknown mechanisms (4). NEAA deprivation induced the expression of CHOP in hepatocytes, suggesting activation of this pathway. Nonessential amino acids are also utilized for gluconeogenesis, and it is therefore conceivable that NEAA withdrawal affects cellular energy balance. However, in our experimental system, the concentration of glucose in the medium was 4.5-fold greater than the total concentration of the NEAA, and therefore it seems unlikely that diminished availability of glucose metabolites was a ratelimiting factor. Furthermore, addition of twice the normal concentration of EAA did not overcome the cell cycle arrest induced by NEAA withdrawal. Ongoing studies are attempting to clarify the mechanisms that regulate cyclin D1 expression in response to amino acids, which may involve several pathways.
This study did not comprehensively evaluate the role of amino acids and protein intake on hepatocyte proliferation and liver regeneration. It will be of interest to determine whether withdrawal of single amino acids, or small combinations of amino acids, similarly inhibits cyclin D1 expression and cell cycle progression. Older studies, which have been somewhat conflicting, have suggested that withdrawal of proline, glutamine, or arginine inhibits hepatocyte cell cycle progression in culture (35)(36)(37). Despite the fact that leucine and other essential amino acids play a significant role in regulating the protein synthetic apparatus (5), our results do not suggest that this amino acid is necessary for hepatocyte cell cycle progression in short-term experiments. Additional older studies have shown that protein deprivation inhibits liver regeneration (11,12,22), although the effects of this treatment on cell cycle protein expression have not been examined. Individual amino acids may regulate hepatocyte proliferation through distinct pathways, and the effects may be cell-type dependent.
The two models of impaired hepatocyte cell cycle progression used in these studies, NEAA withdrawal in culture and protein deprivation in vivo, are clearly not equivalent. For example, NEAA withdrawal in culture impaired S6 phosphorylation, but S6 phosphorylation was not diminished following PH in the absence of dietary protein. It is also important to point out that animals have the capacity to synthesize nonessential amino acids and that protein-deprived animals will maintain detectable (albeit diminished) levels of plasma amino acids (11). In the protein-deprived animals, we speculate that impaired hepatocyte proliferation after PH may be due in part to diminished availability of one or more key amino acids that regulate cyclin D1 expression. However, other factors, such as altered plasma insulin levels, may contribute to the mitoinhibitory effects of diminished dietary protein intake (10 -12). In our cell culture studies, NEAA deprivation down-regulated cyclin D1 expression and cell cycle progression even though the concentration of insulin was held constant. More comprehensive studies will be necessary to clarify the roles of plasma amino acids, insulin, and other factors in the models of protein deprivation in vivo. Despite the differences between the cell culture and in vivo systems reported here, our findings suggest that cyclin D1 plays an important role in both models.
In summary, these studies identify cyclin D1 as a key mediator of hepatocyte proliferation in response to amino acids and protein intake. Further study of the individual amino acids that regulate this response, and the intracellular pathways involved, should provide insight into the relationship between nutrients and cell proliferation. A better understanding of these pathways could eventually lead to the development of therapies for liver diseases and other conditions. For example, if combinations of specific amino acids maximize cyclin D1 expression and hepatocyte proliferation, these might promote adaptive liver regeneration in patients with rapidly progressing liver disease. On the other hand, enhanced cyclin D1 expression (through several mechanisms) and increased hepatocyte proliferation in the cirrhotic liver are associated with hepatocellular carcinoma (38,39). The current studies suggest that selective amino acid deficiency may override mitogenic signaling in hepatocytes. Thus, reduced intake of dietary protein or key amino acids may possibly forestall the development of cancer in patients with stable cirrhosis or other premalignant conditions. Protein intake plays an important role in the development of kidney diseases (40), and possibly other conditions (41), and could conceivably affect the natural history of end-stage liver disease and hepatocellular carcinoma.