Insulin regulates heregulin binding and ErbB3 expression in rat hepatocytes.

The heregulin-ErbB system of ligands and receptors are newly described epidermal growth factor (EGF) and EGF receptor-related proteins that regulate growth, differentiation, and gene expression in numerous cell types. This study describes a receptor for heregulin β-1 (HRGβ1) on cultured rat hepatocytes and an inhibitory influence of insulin on HRGβ1 binding. HRGβ1 (30 nM) stimulated DNA synthesis 2-fold and was not augmented by insulin as is the case with EGF receptor ligands. A labeled peptide corresponding to the EGF domain of HRGβ1 bound to a single population of 19,600 ± 1,800 binding sites/cell with a Kd of 360 ± 22 pM. Cross-linking experiments showed binding of HRGβ1 to ErbB3 but not ErbB2 or ErbB4. HRGβ1 induced phosphorylation of ErbB3 and decreased ErbB3 protein levels, suggesting that HRGβ1 activates signaling through the ErbB3 receptor and influences receptor trafficking. Following plating, [125I]HRGβ1 binding and ErbB3 protein levels increased 8- and 3-fold, respectively, over the first 12 h in culture. These increases required de novo protein synthesis and were inhibited with 50 nM insulin resulting in 3500 binding sites with a Kd of 265 pM. These data suggest that the heregulin-ErbB system can regulate liver functions and may be linked to the metabolic and nutritional status of the animal.

The heregulin-ErbB system of ligands and receptors are newly described epidermal growth factor (EGF) and EGF receptor-related proteins that regulate growth, differentiation, and gene expression in numerous cell types. This study describes a receptor for heregulin ␤-1 (HRG␤1) on cultured rat hepatocytes and an inhibitory influence of insulin on HRG␤1 binding. HRG␤1 (30 nM) stimulated DNA synthesis 2-fold and was not augmented by insulin as is the case with EGF receptor ligands. A labeled peptide corresponding to the EGF domain of HRG␤1 bound to a single population of 19,600 ؎ 1,800 binding sites/cell with a K d of 360 ؎ 22 pM. Cross-linking experiments showed binding of HRG␤1 to ErbB3 but not ErbB2 or ErbB4. HRG␤1 induced phosphorylation of ErbB3 and decreased ErbB3 protein levels, suggesting that HRG␤1 activates signaling through the ErbB3 receptor and influences receptor trafficking. Following plating, [ 125 I]HRG␤1 binding and ErbB3 protein levels increased 8-and 3-fold, respectively, over the first 12 h in culture. These increases required de novo protein synthesis and were inhibited with 50 nM insulin resulting in 3500 binding sites with a K d of 265 pM. These data suggest that the heregulin-ErbB system can regulate liver functions and may be linked to the metabolic and nutritional status of the animal.
Heregulins signal through a family of EGFr-like tyrosine kinase receptors that include ErbB2 (neu, HER2), ErbB3 (HER3), and ErbB4 (HER4) (15). Several studies have suggested that heregulins are ligands for ErbB3 and ErbB4 (16) and that heterodimeric complexes of ErbB proteins form a high affinity heregulin binding site (17) and designate the signal transduction pathway(s) activated following binding (18,19). Although heregulins do not bind the EGFr and EGFr ligands do not bind ErbB2, ErbB3, or ErbB4, interactions between these receptors occur through inter-receptor transphosphorylations (15,20). Ligands of the EGFr, especially TGF-␣, have been strongly implicated in the burst of compensatory hyperplasia following partial hepatectomy (21)(22)(23)(24)(25), a process that is also strongly influenced by other hormones, growth factors, and nutritional signals (26 -30). Liver regeneration is a striking example of the synchronized proliferation of multiple cell types within a tissue that simultaneously maintains complex metabolic and other differentiated functions. The purpose of our study was to explore the possibility that the heregulins and their receptors are involved in the control of growth or differentiated functions in the liver. We demonstrate that a recombinant peptide corresponding to the EGF domain of the ␤-1 isoform of heregulin (HRG␤1) bound to rat hepatocytes via the ErbB3 receptor, induced receptor phosphorylation, and stimulated DNA synthesis. HRG␤1 treatment led to decreased levels of ErbB3 in freshly plated hepatocytes in a manner consistent with ligandmediated down-regulation. Insulin inhibited the up-regulation of ErbB3 in vitro, suggesting a mechanism whereby metabolic controls could modulate proliferative and differentiative signals within liver.

EXPERIMENTAL PROCEDURES
Peptides, Reagents, and Radiochemicals-Human recombinant HRG␤1 (amino acids 177-244) was prepared as described (4). HRG␤1 was iodinated to a specific activity of 250 -300 Ci/g by the lactoperoxidase method (4). This protein corresponds to the EGF domain of the mature secreted form of HRG␤1 amino acids 177-244. Insulin was from Eli Lily & Co. (Indianapolis, IN). Synthetic rat TGF-␣ was purchased from Peninsula Laboratories Inc. (Belmont, CA). [ 125 -I]TGF-␣ was prepared using the chloramine T method as described (22,31 (32). The livers of ether-anesthetized rats were perfused through the portal vein with a calcium-free solution consisting of 150 mM NaCl, 2.8 mM KCl, 5.5 mM glucose, and 25 mM HEPES (pH 7.6) for 10 min followed by the same solution containing 3.8 mM CaCl 2 , 10 g/ml soybean trypsin inhibitor, and 0.5 mg/ml collagenase type I. The cells were dispersed in medium supplemented with 10% calf serum and filtered through 61-m nylon mesh. The hepatocytes were then purified by a 5-min sedimentation at 1 ϫ g in serum-containing medium followed by centrifugation at 50 ϫ g in isotonic percoll (specific gravity ϭ 1.06) to reduce contamination by nonparenchymal cells (33). Percoll was removed by two washes in serum-containing medium, and the hepatocytes were assessed for viability by trypan blue exclusion (Ͼ 95% viable). Cells (375,000 cells/well) were plated in type-1 collagen-coated 35-mm wells. After a 30-min attachment period, the serum-containing medium was replaced with 1.5 ml of serum-free medium containing growth factors as indicated.
DNA Synthesis Assays-Except where indicated otherwise, after 48 h of stimulation, the cells were refed with medium containing specified growth factors and 1 Ci/ml of [ 3 H-methyl] thymidine (0.36 Ci/mmol). At 72 h the experiment was terminated, and the incorporation of [ 3 Hmethyl]thymidine was determined. Cells were fixed and washed free of unincorporated label with 5% trichloroacetic acid at 4°C. The cells were then lysed in 1 N NaOH, and the DNA was precipitated by adding perchloric acid to a concentration of 1.5 N. The DNA was hydrolyzed at 90°C for 15 min, and aliquots were taken for scintillation counting and determination of DNA content using Richards' modification of the Burton diphenylamine assay (34). The results of assays in triplicate are expressed as the specific activity of the DNA (dpm/g DNA Ϯ S.D.).
In autoradiography studies, cells incorporated [ 3 H-methyl]thymidine from 12 to 72 h in culture. After incubation, ascorbic acid ([final] ϭ 230 mM) was added for 20 min at room temperature. Then the cells were washed twice in methanol, rehydrated in deionized water, and overlaid with photographic emulsion for 72 h before developing and counterstaining with Giemsa. The results are represented as the percentage of labeled nuclei from a total of at least 300 nuclei.
HRG␤1 Binding Studies-[ 125 I]HRG␤1 binding to cultured hepatocytes was assessed using a modification of our previous method for [ 125 I]EGF binding (32,35). After the indicated times and conditions of incubation, hepatocytes were placed at 4°C for 10 min. Medium was then aspirated and replaced with ice-cold binding medium (Williams' medium E containing 2 mg/ml bovine serum albumin, 20 mM HEPES, pH 7.35) containing the indicated concentrations of [ 125 I]HRG␤1 with or without unlabeled HRG␤1. After a 4-h binding period at 4°C, the monolayers were washed twice with 2.0 ml of ice-cold Hanks' balanced salt solution containing 2 mg/ml bovine serum albumin, 20 mM HEPES (pH 7.35) and twice with cold phosphate-buffered saline. The cells were then solubilized in 1 ml of NaOH (1 N) and counted (Micromedic Systems; Horsham, PA). The data are expressed as total cpm bound/dish or as the amount of bound HRG␤1 (means of duplicate dishes), corrected for nonspecific binding in the presence of 50 -200-fold excess cold HRG␤1. The data were subjected to a Scatchard-type analysis (36) using the program Ligand (37).
Immunoprecipitation and Western Blotting-Whole rat liver or cultured hepatocytes were lysed in TGH buffer (20 mM HEPES, 1% Triton X-100, 10% glycerol, 50 mM NaCl) with 1 mM phenylmethylsulfonyl fluoride, 100 M sodium orthovanadate, 10 g/ml aprotinin, and 1 g/ml leupeptin. The indicated amounts of protein lysate were incubated overnight at 4°C with antibodies directed toward ErbB2, ErbB3, or ErbB4 receptors or EGF receptor. Complexes were precipitated for 2 h at 4°C with protein G-Sepharose (Pierce). Pellets were washed four times with TGH at room temperature then dissolved in 20 l 1 ϫ SDS gel sample buffer and heated to 95°C for 5 min. Supernatants were loaded in a 6% SDS-polyacrylamide gel and electrophoresed. Resolved proteins were then electrotransferred onto nitrocellulose membranes and Western blotted with antibodies against EGFr, ErbB2, ErbB3, or ErbB4 receptors. Immunoreactive species were resolved using the ECL method (Amersham Corp.) and exposed to x-ray film for radiography. Films were scanned on a IS-1000 digital densitometer for quantification (Alpha Innotech Corp, SanLeandro, CA).
Chemical Cross-linking-Following perfusion and plating, hepato-cytes were incubated at 37°C for 12 h. Cells were then chilled to 4°C and incubated 4 h in binding buffer containing 10 ng/ml [ 125 I]HRG␤1 or [ 125 I]TGF-␣ with or without a 50-fold molar excess of unlabeled HRG␤1 or EGF, respectively. Bis(sulfosuccinimidyl) suberate was then added to a final concentration of 2 nM, and the cells were incubated for an additional 40 min at room temperature. The reaction was stopped by the addition of Tris-HCl to a final concentration of 50 mM. The cells were then washed two times with phosphate-buffered saline and lysed in TGH lysis buffer (with protease inhibitors). As a control, MCF-7 cells were subjected to the same reaction conditions and analysis. Total protein lysate (100 g/lane) was electrophoresed in a 6% acrylamide gel, dried, and exposed to x-ray film. 400 g of hepatocyte lysate was incubated in separate reactions with the above anti-ErbB antibodies. Immune complexes were precipitated with protein G, electrophoresed, dried, and exposed to x-ray film. Phosphorylation of ErbB3 by HRG␤1-Hepatocytes were cultured for 12 h and then stimulated for 5 min at 37°C with 300 nM HRG␤1 and lysed in TGH with protease inhibitors (listed above) and phosphatase inhibitors (10 mM sodium molybdate, 10 mM ␤-glycerol phosphate). ErbB3 proteins were immunoprecipitated, transferred to membranes, and blotted with antiphosphotyrosine antibodies (RC20H, Transduction Laboratories, Lexington, KY). Identical membranes were blotted with anti-ErbB3 antibodies to normalize level of phosphorylation to amount of receptor precipitated.

HRG␤1
Stimulates Hepatocyte DNA Synthesis-The combination of EGF and insulin can stimulate 60 -80% of rat hepatocytes cultured in serum-free medium to undergo at least one round of replicative DNA synthesis with a smaller number actually undergoing mitosis (27,38). Fig. 1 shows the effect of increasing concentrations of TGF-␣ or HRG␤1 on [ 3 H-methyl] thymidine incorporation both alone and in the presence of 150 nM insulin. In contrast to ligands for the EGFr, the effects of HRG␤1 were not appreciably amplified by the presence of 150 nM insulin. In a nuclear labeling assay, maximal stimulatory doses of HRG␤1 (30 nM) stimulated 8.7 Ϯ 0.84% of nuclei to enter S phase by 72 h in culture as compared with 2.6 Ϯ 0.83% of control nuclei (data not shown). The stimulatory effects of HRG␤1 alone and in the presence of 150 nM insulin were completely inhibited by TGF␤1 (data not shown).
HRG␤1-Hepatocyte Binding Studies-Initial studies established that binding equilibrium of [ 125 I]HRG␤1 was reached by 4 h at 4°C (data not shown). The specificity of the HRG␤1 binding site in hepatocytes is shown in Fig. 2. Whereas the binding of [ 125 I]HRG␤1 was displaced by unlabeled HRG␤1, 500-fold molar excess concentrations of EGF, TGF-␣, or amphiregulin, ligands for the EGFr, were ineffective. As reported by Holmes et al. (4), a 500-fold excess of unlabeled HRG␤1 was ineffective in displacing the binding of [ 125 I]TGF-␣ from the EGFr (data not shown). These results suggest a specific bind- ing site for HRG␤1 on hepatocytes that is distinct from the EGFr.
We next sought to determine the influence of hepatocyte isolation and purification on HRG␤1 binding. Previous studies showed that the number of EGF receptors on rat hepatocytes varies dramatically over time in culture (32,39), increasing in freshly isolated cells to 230,000 sites/cell 12 h after plating and then declining rapidly. The estimated EGFr receptor density after 12 h in culture corresponds well with independent estimates made on membranes from whole liver (40), suggesting that cell isolation results in a loss of EGFr that is transiently reversed over the first 12 h in culture. Fig. 3A shows a similar pattern of recovery and decline in [ 125 I]HRG␤1 binding over time in culture. All subsequent estimates of HRG receptor number and affinity were performed on hepatocytes 12 h after plating. To determine if the increase in [ 125 I]HRG␤1 binding requires de novo protein synthesis or is regulated by physiologic signals, binding assays were performed on hepatocytes cultured for 12 h with or without 10 g/ml cycloheximide or 150 nM insulin. Fig. 3B shows that the increase in [ 125 I]HRG␤1 binding was blocked by cycloheximide and by insulin. Concentrations of insulin as low as 50 nM were able to completely inhibit the increase in HRG␤1 binding (data not shown).
Hepatocyte HRG␤1 binding sites could be saturated at 4°C. Fig. 4 shows the binding of increasing amounts of [ 125 I]HRG␤1 and the data resulting from Scatchard-type analysis using the Ligand program (37). A single population of 19,600 Ϯ 1,800 binding sites per cell (mean Ϯ S.D.) with a K d of 360 Ϯ 22 pM was identified in three separate hepatocyte preparations. Scatchard analysis of HRG␤1 binding to hepatocytes cultured in the presence of insulin revealed that the diminished HRG␤1 binding resulted from fewer binding sites of comparable affinity as control culture (K d ϭ 263 pM, sites/cell ϭ 3532) (Fig. 4B).
Immunodetection of ErbB Proteins in Hepatocytes-To determine if the known components of the HRG␤1 receptor complex are present in rat liver, lysates of cultured hepatocytes and whole liver were immunoprecipitated with antibodies specific for rat EGFr, ErbB2, ErbB3, and ErbB4 receptors. Fig. 5 shows that of the proteins analyzed, only the EGFr and ErbB3 were immunodetectable in rat liver at 170 and 180 kDa, respectively. The sensitivity of the ErbB2 antibodies to rat protein was confirmed by the detection of the receptor in the rat hepatoma line H4IIe (data not shown).
HRG␤1-Hepatocyte Cross-linking and Immunoprecipitation-Following incubation to equilibrium with [ 125 I]HRG␤1 or [ 125 I]TGF-␣ as described above, ligand-receptor complexes were cross-linked with the cell-membrane impermeable coupling agent bis(sulfosuccinimidyl) suberate, lysed, and electrophoresed for autoradiography. As shown in Fig. 6A, [ 125 I]HRG␤1 was associated with a single major molecular complex of approximately 180 kDa, and this association was completely inhibited in the presence of unlabeled HRG␤1. As a control, [ 125 I]TGF-␣ was cross-linked to a single major species of 160 kDa and was competed with unlabeled EGF. Anti-ErbB2, anti-ErbB3, anti-ErbB4, and anti-EGFr antibodies were then used to identify the components of the [ 125 I]HRG␤1 binding site in hepatocytes. As shown in Fig. 6B, only anti-ErbB3 antibodies precipitated a radioactive 180-kDa species from a lysate of hepatocytes cross-linked to [ 125 I]HRG␤1. As a control, anti-EGFr antibody was used to immunoprecipitate [ 125 I]TGF-␣ cross-linked to cultured rat hepatocytes, yielding a radioactive 170-kDa species. Taken together, these results suggest that ErbB3 is the primary binding protein in rat hepatocyte HRG receptors.
Phosphorylation of ErbB3 in Hepatocytes by HRG␤1-After 12 h of incubation, hepatocytes were stimulated for 5 min with 300 nM HRG␤1 at 37°C. Western blotting with anti-phosphotyrosine antibodies showed more than a 2-fold increase in the phosphotyrosine content of ErbB3 stimulated with HRG␤1 as compared with controls or insulin-treated ErbB3. (Fig. 7) Effect of Growth Factors on ErbB3 Expression in Cultured Hepatocytes-Hepatocytes were incubated for 8 h with HRG␤1, insulin, or EGF. Figs. 8 and 9 show that within 8 -12 h under control conditions, ErbB3 protein expression increased 2-3fold, consonant with the observed increase in HRG␤1 binding. Both HRG␤1 (Fig. 8) and insulin (Fig. 9A) and to a lesser extent EGF (Fig. 8) inhibited the increase in ErbB3 protein levels. Neither insulin (Fig. 9B) nor HRG␤1 (data not shown) affected EGFr expression. In three separate hepatocyte preparations, HRG␤1 or insulin treatment resulted in ErbB3 expression that was 25 Ϯ 2 or 32 Ϯ 10% of controls, respectively. These results are consistent with a down-regulation of ErbB3 by HRG␤1 in a ligand-dependent manner and suggests that the depressed HRG␤1 binding induced by insulin results from a decrease in the expression of ErbB3 protein. DISCUSSION Heregulins elicit proliferation and/or differentiation in cells expressing ErbB receptors (20,41). We questioned whether the heregulin system of ligands and receptors is important in the control of liver growth and function. HRG␤1 at 30 nM stimulated DNA synthesis about 2-fold and was not augmented by insulin. This is in contrast to TGF-␣, which at 10 nM stimulated a 30-fold increase in [ 3 H]thymidine incorporation and was augmented at lower concentrations by insulin. Relative to TGF-␣, which stimulates labeling of 80 -85% of nuclei in cultured hepatocytes, HRG␤1 stimulated only 6% of the nuclei, suggesting possible compartmentalization of HRG-sensitive hepatocytes (42) and an underestimation of the number of binding sites per cell. Alternately, heregulins may not be primary mitogens in rat liver but could regulate differentiation during FIG. 5. Immunoreactivity of ErbB receptors in liver and hepatocyte lysates. Receptors were precipitated from the indicated amount of liver or hepatocyte (HC) protein, transferred to membranes, and blotted. As a control, ErbB2, ErbB3, and ErbB4 antibodies were neutralized with immunizing peptide prior to precipitation (ϩNP). Nonspecific bands are indicated in reactions without protein.
FIG. 6. Cross-linking of HRG␤1 to cultured cells. 12 h after plating, hepatocytes or MCF-7 cells were incubated for 4 h at 4°C in 1.5 nM [ 125 I]HRG␤1 with or without 200-fold excess cold competitor. Cells were cross-linked with bis(sulfosuccinimidyl) suberate for 40 min at room temperature, washed, and lysed. A, 200 g of protein was electrophoresed, dried, and exposed to film. As a control, labeled TGF-␣ was cross-linked to hepatocytes. B, lysates from [ 125 I]HRG␤1 cross-linked hepatocytes were immunoprecipitated with the indicated ErbB antibody. As a positive control, noncompeted lysates from TGF-␣ crosslinked hepatocytes were immunoprecipitated with EGFr antibody. Immunoprecipitates were electrophoresed in a 5% polyacrylamide gel, dried, and exposed to x-ray film.
FIG. 7. Phosphorylation of ErbB3 by HRG␤1. After 12 h in culture, hepatocytes were stimulated with 300 nM HRG␤1, 1.5 M insulin (Ins), or control medium (Con) for 5 min at 37°C and then lysed. ErbB3 protein was immunoprecipitated, electrophoresed, and blotted with anti-phosphotyrosine or anti-receptor antibodies. Films were scanned and quantified by densitometry.
FIG. 8. Effect of HRG␤1 and EGF on hepatocyte ErbB3 expression. A, freshly isolated hepatocytes were incubated at 37°C for 8 h with (f) or without (Ⅺ) 30 nM HRG␤1 or 10 nM EGF (å). B, identical cultures of hepatocytes 12 h after isolation were incubated for 8 h with (f) or without (Ⅺ) 30 nM HRG␤1 or 10 nM EGF (å). At the indicated times, cells were lysed, immunoprecipitated and Western blotted with anti-ErbB3 antibodies. Films were scanned and quantified by densitometry. con, control. development, maintenance of differentiated functions during regeneration, or metabolism in response to nutritional status.
HRG␤1 binds a specific receptor on rat hepatocytes that is distinct from the EGFr, as described previously in other cell types (4,43). In a wide variety of cells, heregulin appears to bind ErbB3 or ErbB4, but no ligands have been identified that directly bind ErbB2 (16). Dissociation constants of 1.9 (17), 11 (44), and 0.9 nM (45) have been reported in ErbB3 transfected COS-7 and insect cells, whereas a K d of 1.5 nM was measured in ErbB4-transfected COS-7 cells (44). In contrast, COS-7 cells co-transfected with ErbB2 and ErbB3 showed a K d of 0.13 nM (17). In our studies, hepatocytes were shown to have approximately 20,000 HRG binding sites per cell with a K d of 0.36Ϯ.02 nM, a dissociation constant intermediate between those shown in cells transfected with ErbB3 or ErbB4 alone and those co-transfected with ErbB2 and ErbB3. Of these three receptors, only ErbB3 was immunodetectable in rat hepatocytes and appears to be at least one component of the hepatic HRG␤1 receptor. Subsequent reverse transcriptase-polymerase chain reaction analysis failed to detect mRNA corresponding to ErbB2 or ErbB4 in adult rat liver (data not shown). Labeled HRG␤1 chemically cross-linked to hepatocyte plasma membranes was immunoprecipitated in a 180-kDa complex only with antibodies to ErbB3, consonant with a primary role for ErbB3 in hepatocyte HRG signaling. Stimulation with HRG␤1 led to phosphorylation of ErbB3 receptors in hepatocytes within 5 min. Hepatocytes cultured as briefly as 2 h in the presence of HRG␤1 expressed lower levels of ErbB3 compared with control or EGF-treated cells consistent with ligand-mediated down-regulation. However, further experimentation is required to fully elucidate the mechanics of receptor-ligand trafficking after HRG binding in hepatocytes. The origins of the up-regulation of ErbB3 protein following hepatocyte isolation are unclear but may be associated with the induction of immediate early genes that is observed in hepatocyte cultures (46).
Our studies do not yet define the entire hepatocyte heregulin receptor complex, which is presumed to include at least one other ErbB species. Several lines of evidence suggest that ErbB3 requires the association of other ErbB receptors or the EGFr to form an active signaling complex in hepatocytes: (a) the previously cited data indicating that heterodimers of ErbB2 and ErbB3 constitute high affinity HRG binding sites in transfected cells, whereas monomers of ErbB3 exhibit low affinity binding (17); (b) the demonstration that despite its homology to other tyrosine kinase receptors, ErbB3 has weak or nonexistent kinase activity due to an Asp 3 Asn substitution in the catalytic site of the otherwise highly conserved kinase domain (47); (c) the demonstration that ErbB3 is known to be a substrate of ErbB2, ErbB4, and the EGFr and can designate the signal transduction pathway(s) activated following ligand binding (15,18,19). For example, activation of inositol 1,4,5trisphosphate kinase by EGFr has been recently shown to be dependent upon the intermediacy of ErbB3, which contains seven repeats of the binding motif specific for the p85 subunit of inositol 1,4,5-trisphosphate kinase (48). In Ba/F3 cells transfected with EGFr and ErbB3, heregulin treatment resulted in phosphorylation of both receptors, suggesting that ErbB3 can be a substrate for the EGFr kinase in a heregulin-induced ErbB3/EGFr heterodimer (49). No data exist to speculate on the possible influence of the EGFr on ErbB3 affinity for HRG in heteromeric complexes. We were unable to demonstrate any high molecular weight species consistent with ErbB multimers from cross-linked hepatocyte lysates, presumably due to relatively low levels of receptor expression as compared with transformed or transfected cell lines. However, the data in Fig. 8 suggest that EGF may also influence ErbB3 expression in a fashion similar to HRG␤1, perhaps due to internalization of ErbB3 and EGFr heterodimers. Heterodimerization with EGFr may explain the mechanism of ErbB3 phosphorylation and high affinity binding in our model. Current studies aim to identify the other proteins involved in HRG binding and signaling in hepatocytes and other liver cells.
Insulin regulation of hepatic HRG binding and ErbB3 expression is of potential physiological significance. Insulin continuously bathes the liver through the portal circulation and is a mitogen for hepatocytes in vitro, acting synergistically with glucagon, EGF, and other peptide growth factors to stimulate DNA synthesis in cultured hepatocytes and regenerating liver (27,38,50). Although HRG␤1-mediated DNA synthesis was not significantly augmented by insulin, the inhibitory effect of insulin on ErbB3 protein expression and HRG␤1 binding in cultured hepatocytes was striking. Insulin could affect ErbB3 expression directly or down-regulate the receptor by inducing the secretion of a ligand. Insulin may prove to regulate the expression of ErbB3 protein and HRG binding in vivo because the concentrations of insulin effective in inhibiting ErbB3 expression were within physiological ranges detected in the portal vein (51,52). We speculate that insulin, perhaps with other hormones and growth factors, may permit complex nutritional or metabolic controls to influence the intensity or duration of growth and differentiative signals generated by activated ErbB3 receptors.
Our purpose in this study was to describe the interaction of a recently described ErbB ligand, heregulin, with hepatocytes. This ligand has been purified from human breast cancer cells and ras-transformed rat fibroblasts, but transcripts have been detected in normal human liver as well (4). Here we show that the HRG␤1 isoform of heregulin can stimulate DNA synthesis in rat hepatocytes and binds to ErbB3, but the current studies do not define the role played by this ligand-receptor system in liver physiology. As has been reported in the case of astrocytes in culture, heregulin may act as a survival and/or differentiation factor for hepatocytes (53). The signaling pathways and cellular responses that result from activation of ErbB3 are currently under study in this laboratory. These studies set the FIG. 9. Inhibition of ErbB3 protein expression by insulin in cultured hepatocytes. Freshly isolated hepatocytes were incubated at 37°C for 12 h with (f) or without (Ⅺ) 150 nM insulin (Ins). A, at the indicated time points, cells were lysed, immunoprecipitated, and Western blotted with anti-ErbB3 antibodies. B, the same lysates were immunoprecipitated and Western blotted with anti-EGFr antibodies. Films were scanned and quantified by densitometry. con, control. groundwork for the clarification of the role of the heregulins and their receptors in the normal growth and function of the liver, and their possible involvement in the proliferative response to injury and partial hepatectomy, in the acute phase response, and in the development of hepatic neoplasia.