Age-dependent Decline in Mitogenic Stimulation of Hepatocytes

The proliferative potential of the liver has been well documented to decline with age. However, the molecular mechanism of this phenomenon is not well understood. Cellular proliferation is the result of growth factor-receptor binding and activation of cellular signaling pathways to regulate specific gene transcription. To determine the mechanism of the age-related difference in proliferation, we evaluated extracellular signal-regulated kinase-mitogen-activated protein kinase activation and events upstream in the signaling pathway in epidermal growth factor (EGF)-stimulated hepatocytes isolated from young and old rats. We confirm the age-associated decrease in extracellular signal-regulated kinase-mitogen-activated protein kinase activation in response to EGF that has been previously reported. We also find that the activity of the upstream kinase, Raf kinase, is decreased in hepatocytes from old compared with young rats. An early age-related difference in the EGF-stimulated pathway is shown to be the decreased ability of the adapter protein, Shc, to associate with the EGF receptor through the Shc phosphotyrosine binding domain. To address the mechanism of decreased Shc/EGF receptor interaction, we examined the phosphorylation of the EGF receptor at tyrosine 1173, a site recognized by the Shc phosphotyrosine binding domain. Tyrosine 1173 of the EGF receptor is underphosphorylated in the hepatocytes from old animals compared with young in a Western blot analysis using a phosphospecific antibody that recognizes phosphotyrosine 1173 of the EGF receptor. These data suggest that a molecular mechanism underlying the age-associated decrease in hepatocyte proliferation involves an age-dependent regulation of site-specific tyrosine residue phosphorylation on the EGF receptor.

The proliferative potential of several tissues, including the liver, declines with age (1,2). The ability to proliferate or regenerate hepatocytes is particularly important in the elderly, who, due to drug metabolism and other environmental exposures, need to replace cells that are destroyed due to toxic reactions. Thus, an age-related decline in the ability to proliferate may limit the ability of the elderly to recover from toxic exposures. The molecular mechanism underlying the age-related decrease in proliferative response is not well understood. Cellular proliferation is activated by a cascade of signals initiated at the cell membrane in response to growth factor binding to its receptor. In response to the mitogen, epidermal growth factor (EGF), 1 hepatocytes from old rats compared with young rats have diminished induction of DNA synthesis (2). However, hepatocyte EGF receptor number and binding to the receptor are similar with age (3). Such data suggest that age-related differences in proliferation may involve differences in the signal transduction pathways stimulated by growth factors.
The binding of EGF to its receptor initiates a series of signaling events to induce DNA synthesis and cell division. In response to ligand binding, the EGF receptor is phosphorylated, which leads to membrane recruitment of adapter and exchange proteins for activation of Ras and subsequent Raf kinase activation (4 -6). Downstream of Raf kinase, activation of a series of kinases proceeds to activate primarily ERK-MAP kinase in response to EGF (7,8). The MAP kinase family of proteins are activated by dual phosphorylation on Thr-X-Tyr residues (reviewed in Refs. 9 and 10). Other members of the MAP kinase family include JNK and p38 MAP kinases, which are primarily activated by proinflammatory cytokines and cellular stress (reviewed in Ref. 11). Upstream activators of these kinases can function both specifically and with cross-specificity, resulting in activation of more than one MAP kinase. For example, in response to EGF, ERK-MAP kinase is activated as much as 20-fold, while JNK-MAP kinase is activated up to 6-fold (12)(13)(14). Activation of ERK-MAP kinase results in phosphorylation of a number of target proteins including Elk-1 of the ternary complex factor. Elk-1 mediates transcriptional activation of many genes, including c-fos and, consequently, AP-1 target genes (15)(16)(17)(18)(19).
We have used a model of primary hepatocytes obtained from young (4 -6-month) or old (Ͼ32-month) rats to investigate the mechanism for the age-related differences in signal transduction pathways that are activated in response to EGF. ERK-MAP kinase activity was previously reported to be substantially reduced in hepatocytes from old compared with young rats (20). In addition to decreased ERK-MAP kinase activation by EGF, we found that activity of Raf kinase, an upstream signal transducer, was also substantially decreased without a change in Raf protein level in hepatocytes from old rats. The tyrosine phosphorylation of the EGF receptor was similar in hepatocytes from young and old after EGF stimulation, suggesting that the age-related difference is after receptor activation and before Raf kinase activation. Therefore, we examined the ability of the adapter protein, Shc, to associate with the EGF receptor. Using both direct immunoprecipitation and GST fusion protein capture assays, we observed a decrease in the association of the activated EGF receptor with Shc in hepatocytes from old animals compared with young. Furthermore, we show by Western blot analysis that site-specific phosphorylation at tyrosine 1173 of the EGF receptor, a Shc interaction site, is reduced in hepatocytes from old compared with the young. These results indicate that the reduced ability of the EGF receptor to associate with Shc is an early step in the EGF signal transduction pathway that is altered with age.

EXPERIMENTAL PROCEDURES
Hepatocyte Isolation-Hepatocytes were isolated from young (4 -6month) or old (Ͼ30-month) male F1 F344 X BN rats by perfusion of the liver with collagenase, type 2 (Worthington) according to the procedure of Seglen (21) with the modification of using a HEPES (10 mM)-based buffer (22). Cell viability (Ͼ85%) was determined by trypan blue exclusion. The cells (5.5 ϫ 10 6 /100-mm dish) were plated in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin (100 g/ml), and streptomycin (100 g/ml) on tissue culture dishes precoated with Matrigel (1.2 mg/100-mm dish; Collaborative Research, Lexington, MA). Matrigel supports the differentiated state of hepatocytes similar to that in the native liver (22). Two hours after plating, the medium was removed, cells were washed gently two times with Dulbecco's modified Eagle's medium alone, and treatment medium was added (Dulbecco's modified Eagle's medium supplemented with 0.5% fetal bovine serum plus glucose, 5 mM; insulin, 0.14 M; hydrocortisone, 5 M; sodium selenite, 0.2 g/ml; and transferrin, 1 g/ml).
Immune Complex Kinase Assay-ERK-MAP kinase, JNK-MAP kinase, and Raf kinase activities were measured using an immune complex kinase assay (23). Briefly, whole cell extract (50 g of protein) was immunoprecipitated using specific antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) bound to protein A-Sepharose beads (Repligen Corp., Needham, MA). Myelin basic protein (Sigma) was used as the substrate for the p44 ERK kinase assay and the N-terminal c-Jun obtained from a GST fusion protein purified on glutathione-agarose beads (Amersham Pharmacia Biotech, Uppsala, Sweden) was used as substrate for p46 JNK kinase. Purified MEK protein (Santa Cruz Biotechnology) was used as substrate for Raf kinase. Activity was visualized by autoradiography of the dried 10% SDS-polyacrylamide gel, and quantitation was assessed by PhosphorImager analysis using Molecular Dynamics phosphor imaging equipment.
Immunoprecipitation and Western Blot Analysis-Specific antibody was bound to Protein A-Sepharose beads (Repligen Corp., Needham, MA). Human Shc antibody and phosphotyrosine antibody, PY20, were obtained from Transduction Laboratories (Lexington, KY). The phosphospecific antibody, pY 1173, of the EGF receptor was obtained from Upstate Biotechnology, Inc. (Lake Placid, NY). The specificity of the antibody was confirmed by Drs. H. Hoschuetzky and P. Schuessler, nanoTools Antikorpertechnik, Denzlingen, Germany. 2 All other antibodies were purchased from Santa Cruz Biotechnology. Whole cell extracts (100 -400 g of protein) were immunoprecipitated with the bead-antibody complex. The beads were washed with Triton lysis buffer (0.1% Triton X-100, 20 mM Tris, pH 7.4, 137 mM NaCl, 2 mM EDTA, 25 mM ␤-glycerophosphate, 1 mM sodium orthovanadate, 2 mM sodium pyrophosphate, 10% glycerol, 0.4 mM phenylmethylsulfonyl fluoride, leupeptin (1 g/ml), and pepstatin (1 g/ml)), and the protein was eluted into 4ϫ Laemmli's buffer, boiled, and electrophoresed by SDSpolyacrylamide gel electrophoresis. Western blot analysis was done by standard procedures after protein transfer to nitrocellulose membrane. Proteins were detected by a chemiluminescent procedure (NEN Life Science Products).
Shc Domain GST "Capture"-Bacterially expressed GST-Shc domain fusion proteins were generated from pGEX-Shc-PTB and pGEX-Shc SH2 domain constructs obtained from K. Ravichandran (University of Virginia, Charlottesville, VA). The expressed proteins were bound to glutathione Sepharose beads. After washing three times with buffer (20 mM Tris, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 0.4 mM phenylmethylsulfonyl fluoride, leupeptin (1 g/ml), and pepstatin (1 g/ml)), whole cell extracts (120 g of protein) were mixed with the beads/fusion protein and rocked for 2-4 h at 4°C. After washing, the "captured" proteins were eluted into 4ϫ Laemmli's buffer, boiled, and electrophoresed by SDS-polyacrylamide gel electrophoresis. The "captured" EGF receptor was identified by Western blot analysis using the EGF receptor antibody and detected with chemiluminescent reagents (NEN Life Science Products).

Age-related Differences in MAP Kinase
Activation-In the signal transduction cascade to activate transcription of genes involved in proliferation, ERK-MAP kinase is the final regulatory step that directly interacts with transcriptional regulators. Therefore, we first measured activation of ERK-MAP kinase in response to EGF in hepatocytes isolated from young and old rats. In hepatocytes isolated from young rats, ERK-MAP kinase was maximally activated 15-fold (15 min, Fig. 1). Activation of ERK-MAP kinase was only 5-fold in hepatocytes from old animals (Fig. 1). The activity returned to near base-FIG. 1. ERK-MAP kinase activation in response to EGF in hepatocytes from young and old rats. Top panels, representative activity (ACT.) of ERK-MAP kinase after 15, 60, or 120 min of EGF treatment (50 ng/ml) of hepatocytes isolated from young and old rats. The activity was determined by an immune complex kinase assay using myelin basic protein as a substrate. Middle panels, Effect of EGF on ERK-MAP kinase protein levels in primary hepatocytes as detected by Western blot analysis using an ERK-1 antibody. Bottom panels, graphic results of ERK-MAP kinase activity in response to EGF treatment of hepatocytes from three pairs of animals. Results are expressed as the mean Ϯ S.E. and normalized to the average activity of the young controls.
line levels by 2 h. after the addition of EGF to the medium. (Note that the Fig. 1, top panels, show a higher than base line response of one pair of animals at 2 h. The average of all animals at this time was near base line.) JNK-MAP kinase was modestly activated (5-fold) by 60 min after the addition of EGF in cells from young animals (Fig. 2). Cells from old animals showed little EGF-stimulated JNK-MAP kinase activity compared with cells from young animals. The amount of protein determined by Western blot analysis for either JNK or ERK-MAP kinases stimulated with EGF was not different in the hepatocytes from young compared with old ( Figs. 1 and 2). These results are similar to the data reported for ERK-and JNK-MAP kinases previously (20) and confirm that there is a significant decline in EGF-dependent signaling in hepatocytes from old animals.
MKP-1 Protein Levels in Young and Old Animals-MKP-1 has been identified as a dual specificity phosphatase that inhibits ERK-MAP kinase activity (24,25). An age-dependent increase in basal MKP-1 RNA expression was previously suggested to be involved in the mechanism for the decreased ERK-MAP kinase activation in response to EGF (13). We measured the protein level of MKP-1 in whole cell extracts from hepatocytes of young and old rats as well as the level in the native liver. Fig. 3 shows that there is actually a decline in the amount of MKP-1 protein in both the liver (4.5-fold) and cultured hepatocytes (3-fold) with age. This suggests that the mechanism for the age-related difference in ERK-MAP kinase activation does not involve MKP-1.
EGF-stimulated Raf Kinase Activity-To determine if the age-dependent decline in ERK-MAP kinase activity was due to differences in upstream activation of the kinase pathway, we measured the activity of Raf kinase. Raf kinase is activated by Ras GTPase (6). Ras is induced into the active GTP-bound form following the coupling of Ras to the EGF receptor via the SOS-GRB2-SHC complex (5). Subsequent to Raf kinase activation, a series of kinases including MEK1 and -2 are activated by phosphorylation leading to activation of ERK-MAP kinase. Raf kinase activity was measured using an immune complex kinase assay with purified MEK as substrate. Raf kinase activity was significantly reduced at 15 min in EGF-stimulated hepatocytes from old animals compared with those from young animals (Fig. 4). Importantly, Raf kinase activity was greater in the young than the old without a change in the protein level, indicating that specific activity of Raf kinase declined with age. These data suggest that age-related changes in the activation of the ERK-MAP kinase pathway may be upstream of Raf kinase.
Total Tyrosine Phosphorylation of the EGF Receptor-Upon activation of the EGF receptor, autophosphorylation of specific receptor tyrosines initiate the signaling cascade by creating phosphotyrosine sites for binding of SH2 domain-containing proteins that are critical for the recruitment of SHC, GRB2, and SOS proteins to the membrane. To establish if age-related differences in Raf kinase activity were due to differences in signaling initiated at the membrane, we determined the amount of tyrosine-phosphorylated EGF receptor by Western blot analysis. As shown in Fig. 5, no age-related differences in the amount of EGF receptor protein nor in the phosphorylation of the receptor upon EGF stimulation were found. Although no age-related changes in total phosphorylation of the EGF receptor were detected, we could not conclude that membrane signaling events are not different with age, since phosphorylation levels at specific sites were not determined by these methods. Therefore, as a functional measure of membrane membranecoupled signaling, we examined the adapter protein Shc, in response to EGF.
Tyrosine Phosphorylation of Shc and Association with the EGF Receptor-The adapter protein Shc associates with phosphotyrosine residues at positions 1148 and 1173 of the EGF receptor (26,27) and, itself, is phosphorylated on tyrosine residues upon activation of the EGF receptor (28,29). These Shc phosphotyrosine residues create binding sites for the SH2 domain of the adapter protein, Grb2 (29,30). To determine differences in the age-associated ability to phosphorylate Shc, extracts from hepatocytes isolated from young and old rats were immunoprecipitated with Shc antibody and immunoblotted with a phosphotyrosine antibody. The age-dependent ability to phosphorylate Shc in response to EGF was not different (Fig. 6A), nor was there any age-associated difference in the amount of Shc protein (data not shown). However, the association of a 170-kDa protein with Shc was markedly reduced in extracts from hepatocytes of old animals compared with those from young animals (Fig. 6A). In order to determine if this 170-kDa protein was the EGF receptor, Shc immunoprecipitates were analyzed by Western blot analysis using an EGF receptor antibody. As shown in Fig. 6B, the Shc-associated EGF receptor protein was decreased in hepatocytes from old animals as compared with the young.
To confirm the age-related decrease in the association between the EGF receptor and Shc, a GST capture experiment using bacterial fusion proteins of GST-Shc domains was performed. Three domains (shown in Fig. 7A) have been identified in the Shc protein. The N-terminal phosphotyrosine binding (PTB) domain and the C-terminal SH2 domain recognize phosphotyrosine residues of the activated EGF receptor (27,28). The internal collagen homology domain contains tyrosine residues that are phosphorylated upon activation and create the binding sites for the SH2 domain of Grb2 (29,30). As shown in Fig. 7B, the amount of EGF receptor "captured" from the extracts of the hepatocytes from old animals is decreased in comparison with the young when using the bacterially expressed PTB domain of the Shc protein. The amount of decreased association between the proteins ranged from 2-to 3-fold in three independent experiments. Tyrosine 1148 has been identified to be a recognition site for binding of the Shc PTB domain (27,32), and tyrosine 1173 has been reported to be recognized by the Shc-SH2 domain (26,32) as well as the PTB domain (32). Therefore, it is possible that the EGF receptor may be less phosphorylated specifically at tyrosine 1148 and/or 1173 with age. A difference in the phosphorylation state at these positions could exist and would not have been detectable by Western blot analysis for total phosphotyrosine in the EGF receptor-immunoprecipitated extracts (Fig. 5) due to the abundance of tyrosine phosphorylation on multiple tyrosines on the EGF receptor. Furthermore, due to the low levels of EGF receptor in primary hepatocytes, we could not directly map the amount of phosphotyrosine residues at these specific sites.
Phosphorylation Level of Tyrosine 1173 on the EGF Receptor-To determine whether the age-related difference in the ability of Shc to associate with the EGF receptor is due to a difference in level of phosphorylation at a specific site on the EGF receptor, we determined the amount of EGF receptor phosphorylation at tyrosine 1173 by Western blot analysis. Whole cell extracts were immunoprecipitated with the EGF receptor antibody and immunoblotted with an antibody specific for the epitope containing phosphotyrosine 1173 of the EGF receptor. We found decreased levels of EGF receptor phosphotyrosine 1173 in response to EGF in hepatocytes isolated from old rats compared with young (Fig. 8).
The results with both direct Shc immunoprecipitation and the GST-Shc capture assay clearly suggest that a major cause of the age-related decrease in ERK-MAP kinase signaling pathway is due to an alteration in the EGF receptor coupling to the membrane signaling machinery. Furthermore, the results of the GST capture experiment (Fig. 7), using the PTB domain of Shc, indicate that the age-related difference is in the activated EGF receptor protein. Western blotting for phosphotyrosine 1173 (Fig. 8) confirms that the age-related difference in the ability of the EGF receptor to associate with Shc is due to a site-specific difference in tyrosine phosphorylation of the EGF receptor.

DISCUSSION
The molecular basis of the age-related decrease in the capacity of some tissues to proliferate is not well understood. Cellular proliferation is initiated by growth factor binding to specific receptors to initiate signal transduction. Two models of agerelated responses to proliferative stimuli have shown differences in either receptor number and/or receptor phosphorylation (33)(34)(35). However, such differences are not found in the hepatocyte model for the liver (3). In a human skin fibroblast model of aging, the EGF receptor number and phosphorylation kinetics were reduced in old compared with young (33,34). However, the comparison was done in cells derived from different locations i.e. foreskin (young) and forearm (old). This may reflect two very different sources of cells that are influenced by differences in environmental exposures as well as aging. Agerelated differences were also reported in tyrosine phosphorylation of the CD3 chain of the T-cell receptor (35). Additionally, these age-related differences suggest cell type differences in the mechanism of responses to proliferative stimuli with age. Our results agree with those of Ishigami (3) and show that EGF receptor protein levels are similar, and in addition that total tyrosine phosphorylation of the receptor is similar in hepatocytes from young or old rats. Thus, the age-related differences in EGF-stimulated hepatocyte proliferation do not seem likely to be caused by gross overall differences in receptor number or phosphorylation. These data suggested that the mechanistic age-related differences in the response to proliferative stimuli in hepatocytes may be downstream in the signal transduction pathway.
MAP kinases are key regulators of transcription, since these are the catalysts for phosphorylation of transcription factors regulating proliferation. Blocking the ERK-MAP kinase pathway has been shown to down-regulate the mitogenic response (36). Reduced activity of ERK-MAP kinase in response to EGF was previously identified in hepatocytes isolated from old rats compared with young. The mechanism of this decreased activation was suggested to be due to an increase in the basal RNA expression of the dual specificity phosphatase, MKP-1 (13). MKP-1 functions as a dual site phosphatase to inactivate the dually phosphorylated ERK-MAP kinase (24,25) and inhibits cell division (37). Our data also show ERK-MAP kinase activity to be diminished in hepatocytes from old animals in response to EGF (Fig. 1). However, rather than measuring MKP-1 RNA expression, we determined the amount of MKP-1 protein and found the levels to be decreased in hepatocytes as well as in the native livers from old rats (Fig. 3), suggesting that MKP-1 is not regulating the age-related decrease in ERK-MAP kinase activity. Further support for the lack of a role for MKP-1 in regulating the decrease in ERK-MAP kinase comes from the work of Shapiro and Ahn (38). They reported that MKP-1 has a positive effect on Raf kinase activity independent of the basal inhibition of ERK-MAP kinase activity. These results are consistent with our own, where we observed a decrease in Raf kinase activation in hepatocytes from old compared with young ( Fig. 4), not an increase as would be expected if MKP-1 were increased.
Our results indicate that the age-related difference in the proliferative response to EGF is upstream from ERK-MAP kinase activation. Since Raf kinase activation also declined with age (Fig. 4), we examined the coupling of the EGF receptor to the ERK-MAP kinase signaling pathway. Age-related differences in the membrane complexes involved in signaling have previously been reported in T-lymphocytes (35,39,40) and B-cells (41). We have identified an early step in the EGF receptor signal transduction pathway to be a decrease in the association of the adapter protein, Shc, with the activated EGF receptor in hepatocytes from old compared with young (Fig. 6). Furthermore, the results of the GST capture experiment (Fig.  7B), which employed bacterially expressed GST fusion protein containing the PTB domain of Shc, indicate that the age-related difference in hepatocytes is in the activated EGF receptor protein. The PTB domain recognizes the tyrosine phosphoryl- ation at position 1148 of the EGF receptor (27,32,57) and position 1173 (32,57). We also observed a similar age-related decrease in the association of the EGF receptor with the SH2 domain of the Shc protein, which recognizes tyrosine 1173, although the magnitude of the association was not as great as with the Shc-PTB domain (data not shown). The idea that reduced association of Shc with the EGF receptor due to defects in autophosphorylation sites with the effect of reducing mitogenic activity is supported by the data of Gotoh et al. (42). A cell line containing a deletion mutant of the EGF receptor that lacked the autophosphorylation sites downstream of residue 1011 was used to show that EGF treatment could still result in stimulation of Shc phosphorylation but retained only approximately 20% of its mitogenic activity when measured as cell growth. Others have shown that EGF-stimulated mutant EGF receptors containing a deleted C-terminal portion of the EGF receptor or site-specific mutations of C-terminal autophosphorylation sites also stimulate Shc phosphorylation in NIH 3T3 cells (43,44). Shc phosphorylation in the absence of binding to the EGF receptor (Figs. 6 and 7; Refs. 42-44) may be in response to other non-autophosphorylation-dependent pathways initiated by EGF transphosphorylations, such as through the ErbB3 receptor (45,46). The transforming ability of the Cterminal deletion mutant of the EGF receptor (42) is not inconsistent with our data, since it can be mediated in some cell types by Eps8, a substrate for EGF receptor tyrosine kinase activity (47,48). Clearly, the complexities of the pathways that regulate the mitogenic response initiated by EGF ligand binding are not completely understood.
The mechanism of the age-related decline in the ability of the liver to proliferate is identified from our data to probably involve factors that regulate the phosphorylation of the EGF receptor at the specific tyrosine residues that function as sites for binding of the adapter protein Shc. Our data identify tyrosine 1173 of the EGF receptor as one of the residues that is differentially phosphorylated with age in response to EGF (Fig.  8). Candidates of regulatory enzymes for the tyrosine phosphorylation include both kinases and phosphatases. Phosphatases may function as negative or positive regulators of signaling by directly acting on a phosphorylated tyrosine. For example, the phosphatase SHP-2 has been suggested to have a positive effect on EGF signaling (49 -51). We found no age-related difference in the level of SHP-2 protein expression or in the association of SHP-2 in the complex that was immunoprecipitated by Shc antibody following EGF stimulation (data not shown). Another phosphatase, SHP-1, has recently been reported to bind the activated EGF receptor at tyrosine 1173 (52). Binding at this site resulted in decreased ERK-MAP kinase activity in response to EGF. Clearly, the balance between kinase and phosphatase activity is important in regulating the EGF signaling pathway. An alternative mechanism for regulation of signal transduction is suggested by the ability of proteins to mask specific SH2 domains of EGF receptor-coupled proteins (53). However, our data, particularly the GST capture experiments, do not suggest such a mechanism in the age-dependent decline in EGF signal transduction.
Understanding the molecular mechanisms for the age-related differences in the proliferative response to EGF will require more information about the subtle mechanisms that modulate the EGF receptor-ERK signaling pathway. There are certain to be newly discovered regulators that will play a role. For example, it has only recently been shown through genetic and biochemical studies that the kinase suppressor of Ras is an important regulator of Raf activity (54, 55); however, this regulation has also been reported to occur in a kinase-independent manner (56). Thus, the complexity of the molecular mecha-nisms of age-related differences in response to proliferative stimuli in different tissues will require a more in depth knowledge of the EGF signaling pathway.