Growth hormone-induced phosphorylation of epidermal growth factor (EGF) receptor in 3T3-F442A cells. Modulation of EGF-induced trafficking and signaling.

Growth hormone (GH) promotes signaling by causing activation of the non-receptor tyrosine kinase, JAK2, which associates with the GH receptor. GH causes phosphorylation of epidermal growth factor receptor (EGFR; ErbB-1) and its family member, ErbB-2. For EGFR, JAK2-mediated GH-induced tyrosine phosphorylation may allow EGFR to serve as a scaffold for GH signaling. For ErbB-2, GH induces serine/threonine phosphorylation that dampens basal and EGF-induced ErbB-2 kinase activation. We now further explore GH-induced EGFR phosphorylation in 3T3-F442A, a preadipocytic fibroblast cell line that expresses endogenous GH receptor, EGFR, and ErbB-2. Using a monoclonal antibody that recognizes ERK consensus site phosphorylation (PTP101), we found that GH caused PTP101-reactive phosphorylation of EGFR. This GH-induced EGFR phosphorylation was prevented by MEK1 inhibitors but not by a protein kinase C inhibitor. Although GH did not discernibly affect EGF-induced EGFR tyrosine phosphorylation, we observed by immunoblotting a substantial decrease of EGF-induced EGFR degradation in the presence of GH. Fluorescence microscopy studies indicated that EGF-induced intracellular redistribution of an EGFR-cyan fluorescent protein chimera was markedly reduced by GH cotreatment, in support of the immunoblotting results. Notably, protection from EGF-induced degradation and inhibition of EGF-induced intracellular redistribution afforded by GH were both prevented by a MEK1 inhibitor, suggesting a role for GH-induced ERK activation in regulating the trafficking itinerary of the EGF-stimulated EGFR. Finally, we observed augmentation of early aspects of EGF signaling (EGF-induced ERK2 activation and EGF-induced Cbl tyrosine phosphorylation) by GH cotreatment; the GH effect on EGF-induced Cbl tyrosine phosphorylation was also prevented by MEK1 inhibition. These data indicate that GH, by activating ERKs, can modulate EGF-induced EGFR trafficking and signaling and expand our understanding of mechanisms of cross-talk between the GH and EGF signaling systems.

Growth hormone (GH) 1 is a four-helix bundle protein that shares structural similarity with a large class of hormones and cytokines, including prolactin and various interleukins and colony stimulating factors. GH exerts its profound somatogenic and metabolic regulatory effects by interacting with the GH receptor (GHR), a cell surface glycoprotein member of the cytokine receptor superfamily (1)(2)(3). Like other cytokine receptors, the GHR initiates its signal transduction by physical and functional association with JAK2, a cytoplasmic tyrosine kinase member of the JAK family (4,5). GH activates several intracellular signaling pathways. GH-induced STAT5b activation has been particularly intensively studied and is strongly implicated in the regulation of certain GH-responsive hepatic genes (6). Other signaling pathways induced by GH include the mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways, both also targets of a variety of other growth factors and cytokines.
GH induces activation of ERK1 and ERK2 in several model systems, including the murine 3T3-F442A preadipocyte fibroblast (7)(8)(9)(10). In contrast to STAT5 activation, which requires the presence of the receptor cytoplasmic domain in addition to activation of JAK2, the ability of GH to activate ERKs appears to require none of the distal GHR tail and instead corresponds to whether it can promote JAK2 activation (9,(11)(12)(13). However, ERK is not activated in all cells in which GH activates JAK2 (14). Though it remains uncertain what factors determine whether a GHR-expressing cell or tissue will respond to GH with ERK activation, evidence exists among responsive cells for the potential involvement of a number of molecules. These include the Shc-Grb2-Sos-Ras-Raf pathway (15,16), the phosphatidylinositol 3-kinase (or a related enzyme) pathway (17)(18)(19), IRS-1 (19) and the IRS-like Gab-1 adapter molecule (20), and c-src (21).
Recent studies of the mechanisms and consequences of GHinduced ERK activation have revealed interesting relationships between the GH and epidermal growth factor (EGF). These studies suggest previously unrecognized cross-talk between the GHR and members of the EGFR family, examples of seemingly disparate types of signaling receptors (a cytokine receptor and a family of tyrosine kinase receptors, respective-ly). The EGFR family of structurally related transmembrane glycoproteins includes the EGFR itself (ErbB-1), ErbB-2 (cneu), ErbB-3, and ErbB-4 (22,23). Except for ErbB-3, each has intrinsic tyrosine kinase activity in its cytoplasmic domain. Ligands such as EGF, transforming growth factor-␣, and neuregulins induce signaling through these receptors by binding to specific EGFR family members in such a way as to promote particular homo-or heterodimers among the family members. ErbB-2 has no known ligand, but it is the preferred heterodimer partner for other family members when they engage their ligands. EGF, for example, promotes formation of EGFR homodimers or EGFR/ErbB-2 heterodimers and thus causes activation of both of these tyrosine kinases in cells that express them. EGF signaling through EGFR and ErbB-2 has a number of biologically relevant signaling outcomes in both normal and neoplastic cells.
Yamauchi et al. (24) first demonstrated that GH caused tyrosine phosphorylation of the EGFR, both in vivo in the livers of mice and in cell culture. This GH-induced EGFR tyrosine phosphorylation was shown to require JAK2, but not EGFR, kinase activity. Partial mapping by mutagenesis suggested that EGFR Tyr-1068, which when phosphorylated resides in a consensus Grb-2 association motif, was a site of GH-induced phosphorylation and that GH caused enhanced EGFR-Grb-2 association. Further, GH-induced EGFR tyrosine phosphorylation was shown in cell culture to likely contribute to GHinduced ERK activation. Thus, this study suggested that EGFR may be a docking molecule involved in GH-induced, JAK2-dependent ERK activation. Our previous study (25) in 3T3-F442A cells confirmed that GH caused EGFR kinase-independent EGFR tyrosine phosphorylation but also suggested GH effects on ErbB-2. GH caused a decrease in basal and EGF-induced ErbB-2 tyrosine kinase activation and tyrosine phosphorylation that was associated with retardation of the electrophoretic migration of ErbB-2. This retarded migration was shown to be because of serine or threonine, rather than tyrosine, phosphorylation of ErbB-2, and both the GH-induced change in migration and inhibited tyrosine kinase activation were prevented by blocking GH-induced ERK activation. These findings suggested that GH caused an ERK pathway-dependent phosphorylation of ErbB2 that rendered it desensitized to activation in response to EGF.
We now explore further the mechanisms and consequences of cross-talk between GH and EGF signaling. In particular, we examine GH-induced phosphorylation of EGFR and ErbB-2. Our findings suggest that GH causes an ERK pathway-dependent threonine phosphorylation of both ErbB-2 and EGFR that is recognized by a state-specific antibody reactive with ERK consensus phosphorylation sites. In contrast to our findings for ErbB-2, we observe that this GH-induced EGFR phosphorylation does not significantly alter the intrinsic tyrosine kinase activation of EGFR, but instead slows the rate of EGF-induced EGFR intracellular redistribution and degradation, thereby potentiating EGF-induced EGFR signaling. These results suggest that GH may affect EGF signaling by multiple mechanisms, including modulation of EGF-induced EGFR trafficking.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human GH was kindly provided by Lilly. Recombinant human EGF was purchased from Upstate Biotechnology (Lake Placid, NY), and recombinant human PDGF-BB was from Intergen (Purchase, NY). Murine 125 I-EGF (150 -200 Ci/g) was purchased from PerkinElmer Life Sciences. PMA was obtained from Sigma. The PKC inhibitor, GF109203X (Calbiochem), and the MEK1 inhibitors, PD98059 (New England Biolabs, Beverly, MA) and U0126 (Promega, Madison, WI), were purchased commercially.
To generate stable 3T3-L1 transfectants expressing the CFP-tagged EGFR (EGFR-CFP), cells were seeded in 60-mm dishes and used at 50 -80% confluency. The plasmid pCFP/EGFR, kindly provided by Dr. L. Samelson, Laboratory of Cellular and Molecular Biology, NCI, National Institutes of Health, Bethesda, MD, encodes the human EGFR with the CFP fused to its C terminus (27). pCFP/EGFR was transfected into 3T3-L1 cells using GenePORTER transfection reagent (Gene Therapy Systems, Inc., San Diego, CA) as described previously (28). 3T3-L1 cells expressing EGFR-CFP were selected in 1 mg/ml of G418 (Invitrogen) and cloned. Transfectants were maintained in culture medium containing 200 g/ml of G418.
Immunoprecipitation and Immunoblotting-For immunoprecipitation, cell extracts (500 -1000 g) were mixed with 5 l of polyclonal anti-ErbB-2, -EGFR, -PDGFR, or -Cbl-b antibody (1 g) and incubated at 4°C for 2 h with continuous agitation. Protein A-Sepharose (20 l) (Amersham Biosciences) was added and incubated at 4°C for an additional hour. The beads were washed four times with lysis buffer. Laemmli sample buffer eluates were resolved by SDS-PAGE and immunoblotted as indicated below.
Proteins resolved by SDS-PAGE were transferred to Hybond ECL nitrocellulose membranes (Amersham Biosciences). The membranes were blocked with TBST buffer (20 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 0.1% (v/v) Tween 20) containing 2% (w/v) bovine serum albumin and incubated with primary antibodies (0.5-1 g/ml) as specified in each experiment. After three washes with TBST, the membranes were incubated with appropriate secondary antibodies (1:10,000 dilution) and washed. The bound antibodies were detected with SuperSignal chemiluminescent substrate (Pierce). Membrane stripping was performed according to the manufacturer's suggestions (Amersham Biosciences).
Densitometric Analysis-Densitometry of immunoblots was per-formed using a solid state video camera (Sony-77; Sony Corp.) and a 28-mm MicroNikkor lens over a light box of variable intensity (Northern Light Precision 890; Imaging Research, Inc., Toronto, Canada). Quantification was performed using a Macintosh II-based image analysis program (Image 1.49, developed by W. S. Rasband; Research Services Branch, NIMH, Bethesda, MD). Pooled data from several experiments are displayed as mean Ϯ S.E. The significance (p value) of differences of pooled results was estimated using unpaired t tests.
To verify the fidelity of this densitometric method, we performed a control experiment in which total protein concentration of 3T3-F442A cell extract was determined using BCA protein assay reagents (Pierce), and serially diluted total protein aliquots (2.5-100 g) were resolved by SDS-PAGE and immunoblotted with anti-EGFR. Densitometry of the EGFR intensities plotted against the known loaded protein amounts yielded a straight line with a correlation coefficient (R) of 0.97 (not shown). This suggested a high degree of reliability for this densitometric analysis.
Fluorescence Microscopy-3T3-L1 transfectants expressing EGFR-CFP were grown on Corning glass coverslips precoated with gelatin (Sigma) in 6-well plates for 48 h in culture medium until they reached ϳ50% confluency. The cells were starved, pretreated with PD98059 (100 M) or vehicle for 1 h, and stimulated with GH (500 ng/ml), EGF (1 nM), or vehicle as described above and specified in the experiment. The cells were rinsed with PBS and fixed with 4% formaldehyde solution in PBS for 15 min at room temperature. After rinsing with PBS, the coverslips were mounted on microscope slides (Fisher) in Vectorshield mounting medium for fluorescence (Vector Laboratories Inc., Burlingame, CA). Fluorescence patterns were visualized with an Olympus fluorescence microscope at the University of Alabama Cell Biology Imaging Core Facility. Images were collected and analyzed using IPLab Spectrum software (Scanalytics Inc., Fairfax, VA).

125
I-EGF Internalization Experiments-3T3-F442A cells were grown in six-well plates in culture medium until they formed monolayers. The cells were starved and pretreated with GH (500 ng/ml) or vehicle in binding medium (Dulbecco's modified Eagle's medium containing 4.5 g/l glucose, supplemented with 20 mM HEPES and 0.5% (w/v) bovine serum albumin) at 37°C for 10 min. To measure the internalization of 125 I-EGF, the cells were incubated with 125 I-EGF (1 ng/ml) in binding medium in the continued presence or absence of GH at 37°C for 0 -10 min in duplicate. At the end of incubation, the medium was aspirated, and the monolayers were rapidly washed twice with ice-cold binding medium to remove unbound ligand. The cells were then incubated with 0.2 M sodium acetate (pH 4.5) containing 0.5 M NaCl at 4°C for 5 min. The acid wash was combined with another short rinse in the same buffer and used to determine the amount of surface-bound 125 I-EGF. The cells were finally lysed in 100 mM NaOH containing 0.1% (w/v) SDS to determine the amount of internalized 125 I-EGF. Radioactivity was counted with a ␥-counter. The acid-inaccessible internalized 125 I-EGF was presented as a fraction of total cell-associated radioactivity (the sum of cell surface-bound and internalized 125 I-EGF) at each time point. Nonspecific binding was measured in the presence of 100-fold molar excess of unlabeled EGF and was not more than 5% of the total counts.

GH Acutely Promotes ERK-dependent Phosphorylation of EGFR at Sites Recognized by an ERK Phosphorylation Site
Antibody-The murine 3T3-F442A fibroblast expresses EGFR and ErbB-2, as well as the receptor for GH (25). We and others (4,10,18,24,25,29,30) have detected biochemical responses to both GH and EGF in this cell line, making it an appealing target for evaluation of potential cross-talk between the GH and EGF signaling pathways. Our previous studies (25) indicated that in 3T3-F442A cells GH treatment causes ErbB-2 to undergo serine/threonine phosphorylation that results in a decrease in its basal activation state and its desensitization to EGF-induced activation. These effects of GH were inhibited by drugs that block activation of the MEK1/ERK pathway.
The data in Fig. 1 indicated that both GH and EGF caused phosphorylation of ErbB-2 and EGFR at potential ERK consensus sites, suggesting that this is mediated by activation of the ERK pathway by these stimuli. To further test this propo- sition, we employed inhibitors of MEK1, the upstream activator of ERK1 and ERK2 (Fig. 2). Two separate inhibitors, PD98059 and the more potent U0126 (34 -36), were utilized. As already observed, GH and EGF induced the PTP101 reactivity of both ErbB-2 ( Fig Previous studies suggested that serine/threonine phosphorylation of EGFR family members induced by the phorbol ester PMA and PDGF are attributable to the activation of PKC (37-41). We tested whether the GH-induced ErbB-2 and EGFR phosphorylations detected by PTP101 were related to activation of PKC by using the PKC inhibitor, GF109203X (42) ( . As expected, the PMA-induced ERK activation and modifications of ErbB-2 and EGFR were prevented by preincubation of the cells with GF109203X or with the MEK1 inhibitor U0126 (Fig. 3, A-C, upper panels, lanes 8 and 9 versus 7). In contrast, neither GH-induced ERK activation nor ErbB-2 and EGFR PTP101 reactivity were affected by the PKC inhibitor, though U0126 was inhibitory, as expected (Fig. 3, A-C, upper panels, lanes 5 and 6 versus 4). These data strongly suggest that PTP101 is detecting ERK-dependent, rather than PKC-dependent, phosphorylation of EGFR family members that occurs in response to GH and that the PMA-induced ErbB-2 and EGFR PTP101reactive phosphorylations are also mediated by ERK activation; however, the mechanism of PMA-induced (in distinction to the GH-induced) ERK activation in these cells apparently involved the activation of PKC forms that are sensitive to GF109203X. Further, PMA caused retardation of SDS-PAGE migration of both ErbB-2 and EGFR, and this PMA-induced retarded migration was prevented by both the PKC and MEK1 inhibitors (Fig. 3, A and B, lower panels, lanes 7-9 versus 1-3), bolstering the conclusion that the PTP101 reactivity reflects ERK-, rather than PKC-, mediated phosphorylation.
We next investigated the degree to which the propensity of GH to cause ErbB-2 and EGFR phosphorylation reflected a specific preference for cross-talk with EGFR family members. Like the EGFR and ErbB-2, the PDGFR is a transmembrane glycoprotein receptor that harbors intrinsic tyrosine kinase activity in its cytoplasmic domain. 3T3-F442A cells express the PDGFR, and previous studies (43,44) indicate that PDGF stimulation can functionally impact upon GH responsiveness in these cells. Thus, we tested the effect of GH on PDGFR phosphorylation (Fig. 4). As expected, PDGF treatment acutely promoted PDGFR tyrosine phosphorylation, as assessed by anti-PDGFR immunoprecipitation followed by anti-pTyr blotting (Fig. 4A, upper panel, lane 2 versus 3), verifying the responsiveness of these cells to PDGF. In contrast, GH failed to promote either tyrosine phosphorylation (Fig. 4A, upper panel, lane 1) or PTP101-reactive phosphorylation of PDGFR (Fig. 4A, middle panel, lanes 1 and 2 versus 3), despite the rough equivalence of PDGFR immunoprecipitation in GH, PDGF, and control stimulations (Fig. 4A, lower panel, lanes 1-3). As a positive control, the GH-induced PTP101-reactive phosphorylation of ErbB2 in the same experiment was monitored (Fig. 4B, upper panel, lane 1 versus 2), confirming the ability of these reagents and methods to detect such a modification. We conclude that GH-induced activation of ERKs results in ERK kinase consensus sequence phosphorylation of the EGFR family members, ErbB-2 and EGFR, but not of the PDGFR.
GH Affects EGF-induced EGFR Trafficking in 3T3-F442A Cells-We demonstrated previously (25) that acute treatment of 3T3-F442A cells with the combination of GH and EGF resulted in substantially less activation of ErbB-2 than did exposure to EGF alone, suggesting a dampening effect of GH signaling on the ErbB-2 activation mechanism. As detailed above, GH promoted both tyrosine-and PTP101-reactive phosphorylation of the EGFR. Thus, we examined whether cotreatment with GH appreciably affected EGF-induced EGFR tyrosine phosphorylation. For these experiments, we employed a panel of state-specific antibodies, each specific for phosphorylation of particular EGFR tyrosine residues, Tyr-845, Tyr-992, Tyr-1045, Tyr-1068, or Tyr-1173, in addition to anti-pTyr antibody (Fig. 5). Consistent with our prior treatment protocol (25), cells were treated with vehicle (Ϫ), GH, or EGF alone for 15 min or EGF for 15 min in the presence of GH (added 10 min prior to EGF). As expected, GH and EGF each alone caused EGFR tyrosine phosphorylation (Fig. 5, row F (anti-pTyr blot), lane 2 and 3 versus 1), with the effect of EGF being more potent. Yamauchi et al. (24) used EGFR mutagenesis to suggest that Tyr-1068 was a preferred site for GH-induced EGFR tyrosine phosphorylation. Blotting of EGFR precipitates with the vari-ous phosphotyrosine state-specific antibodies (Fig. 5, rows A-E) confirmed that this site (row D, lane 2 versus 1) is phosphorylated in response to GH. However, we also detected substantial GH-induced phosphorylation of Tyr-845 (row A, lane 2 versus 1), a site not known previously to be targeted. Additionally, we observed less substantial phosphorylation of Tyr-992 and Tyr-1173 but no phosphorylation of Tyr-1045. Notably, we detected no change in overall or site-specific tyrosine phosphorylation of the EGFR in cells treated with EGF plus GH (added 10 min prior to EGF) compared with EGF alone (rows A-F, lanes 4  versus 3). In other experiments (not shown), we likewise saw no difference in the level of EGFR tyrosine phosphorylation in EGF-treated versus EGF plus GH-treated samples in which the GH cotreatment was simultaneous to the EGF treatment, rather than being added 10 min prior. Thus, unlike our findings for ErbB-2, GH-induced activation had no appreciable effect on the level of acute EGF-induced EGFR tyrosine phosphorylation, a reflection of the earliest steps in EGF-induced EGFR triggering.
We next investigated the effects of GH on the EGF-induced fate of the EGFR and ErbB-2. It is well documented that EGF induces down-regulation of EGFR, but not ErbB-2, mass (reviewed in Ref. 45). We used immunoblotting of total cell lysates (extracted in the presence of 1% SDS) to estimate the abundance of the EGFR and ErbB-2 in 3T3-F442A cells treated with EGF and GH over a 4 h period (Fig. 6). As expected, EGF treatment caused a time-dependent loss of EGFR, detectable after 30 min and progressive over the remainder of the period (Fig. 6A, upper panel, lanes 2-6 versus 1). In contrast, no loss of EGFR mass was observed in response to treatment with control vehicle alone (not shown) or with GH alone (Fig. 6A, upper panel, lanes 12-16 versus 1), likely consistent with the lack of EGFR tyrosine kinase activation by GH (24,25). Interestingly, however, treatment with EGF in the presence of GH (added 10 min prior to EGF) resulted in attenuation of the loss of EGFR mass in comparison with exposure to EGF alone (Fig. 6A, upper  panel, lanes 7-11 versus 2-6). These parameters were estimated quantitatively by densitometric analysis of the results of five such experiments (Fig. 6B). Though GH itself did not affect EGFR mass, GH cotreatment significantly inhibited the EGFR loss induced by EGF treatment of 60, 120, or 240 min duration (p values less than 0.05, 0.05, and 0.01, respectively). This analysis indicated that the EGFR loss promoted by 240 min of EGF treatment was reduced by roughly 50% in samples also exposed to GH. Thus, GH partially, but substantially, protects against the EGF-induced down-regulation of the EGFR in this system. Notably, ErbB-2 abundance was not affected by treatment with EGF alone, GH alone, or EGF plus GH (Fig. 6A,  lower panel, lanes 2-6, 12-16, and 7-11, respectively, versus 1), consistent with previous observations of the lack of ErbB-2 down-regulation in response to EGF.
Because GH did not affect acute EGF-induced EGFR tyrosine phosphorylation (Fig. 5) but did cause EGFR PTP101reactive phosphorylation (Figs. 1-4) and attenuated EGF-in-duced EGFR loss (Fig. 6), we considered whether the GHinduced ERK-mediated phosphorylation might relate to the effects of GH on EGFR trafficking. To address this issue, cells were treated with GH, EGF, or EGF plus GH for 2 h in the absence or presence of the MEK1 inhibitor, PD98059. As expected, in the absence of PD98059, EGF caused substantial loss of EGFR (Fig. 7A, lane 3 versus 1

), GH alone did not affect EGFR abundance (lane 2 versus 1), and the presence of GH blunted the EGF-induced EGFR loss (lane 4 versus 3).
In the presence of PD98059, EGF-induced EGFR loss was similar to that seen in the absence of the drug (compare lane 6 versus 8 to lane 3 versus 1), whereas GH caused no loss of EGFR (lane 7 versus 8). Notably, however, the EGF-induced loss of EGFR was not prevented by GH cotreatment in the presence of PD98059 (lane 5 versus 6), in contrast to the findings in the absence of the MEK1 inhibitor (lane 3 versus 4). For ease of visualization, the immunoblot in Fig. 7A was scanned densitometrically, and the relative intensity of the EGFR band (relative EGFR mass) is plotted for each lane in Fig. 7B. Pooled data from four independent experiments were densitometrically evaluated and are shown in Fig. 7C. In this display, the loss of EGFR induced by a 2-h incubation with EGF is considered as 100%. Cotreatment with GH in the absence of PD98059 blunted the EGF-induced loss, resulting in a 64 Ϯ 8% loss of EGFR (p Ͻ 0.05 versus EGF alone). In contrast, in the presence of PD98059, GH plus EGF resulted in a 102.9 Ϯ 1.7% loss of EGFR, not different from the 100% loss induced by EGF alone. Thus, blockade of the ERK activation pathway reversed the effect of GH on EGF-induced EGFR loss. This suggests that the GH-induced activation of ERKs and possibly ERK-dependent EGFR PTP101-reactive phosphorylation substantially affect the ligand-activated trafficking of the EGFR, slowing its degradation and down-regulation.
These biochemical studies suggested that GH may impinge on early aspects of EGF-induced EGFR trafficking, because the protective effect of GH on EGFR loss could be discerned as early as 30 min into EGF treatment, albeit not statistically significantly so until 1 h (Fig. 6B). We thus sought to examine the effects of GH on acute EGF-induced EGFR trafficking. As we have shown previously (25,44), 3T3-L1 fibroblasts, like 3T3-F442A cells, endogenously express GHRs and biochemically respond similarly to GH. Notably, the abundance of EGFRs in 3T3-L1 cells was found to be substantially less than that seen in 3T3-F442A cells (Fig. 8A, lane 2 versus 1). Further, we have observed that 3T3-L1 is generally more transfectable than is 3T3-F442A. 2 Thus, to pursue fluorescence microscopy studies, we employed 3T3-L1 cells as the recipient for stable transfection of a plasmid encoding the human EGFR fused at its C terminus to the CFP. This EGFR-CFP fusion has been shown recently (27) in COS-7 cells to bind EGF and become activated and rapidly redistribute intracellularly in response to EGF.
Stably transfected 3T3-L1 cell clones expressing EGFR-CFP were negatively selected in G418 and screened for fusion protein expression by anti-EGFR and anti-GFP immunoblotting. A clone, designated L1/EGFR-CFP (Fig. 8B), was chosen for further analysis. As anticipated, the EGFR-CFP chimera in these cells responded to EGF by becoming tyrosine-phosphorylated (Fig. 8C, lane 2 versus 1) and underwent PTP101-reactive phosphorylation in response to GH (Fig. 8C, lane 4 versus 3), suggesting that the chimera expressed in 3T3-L1 cells behaved similarly to the endogenous EGFR in 3T3-F442A cells. Fur-thermore, GH treatment resulted in ERK activation in the L1/EGFR-CFP, as expected (data not shown).
We next performed fluorescence microscopy of 3T3-L1/ EGFR-CFP cells to track the subcellular distribution of the EGFR chimera (Fig. 8D, upper panel). As expected, untreated serum-starved cells (control) exhibited prominent steady-state cell surface EGFR-CFP expression, as has been noted previously (27) for this chimera and for endogenous EGFRs (reviewed in Ref. 45). Treatment with GH for 25 min had no discernable effect on the distribution of EGFR-CFP. In contrast, EGF treatment for 15 min caused a substantial redistribution of EGFR-CFP away from the cell surface to perinuclear and peripheral intracellular aggregates, similar to those observed by others (27,41) in the course of EGFR endocytosis. Notably, treatment with EGF in the presence of GH (GH/EGF, in which GH was added 10 min prior to EGF) markedly reduced this translocation, resulting in substantial remaining EGFR-CFP on the cell surface and diminished accumulation of intracellular chimera. We further tested the impact of inhibition of the ERK pathway on these EGF-and GH-induced effects on EGFR trafficking by performing the same experiments in the presence of the MEK1 inhibitor, PD98059 (Fig. 8D, lower panel). ERK inhibition had no discernable effect on the distribution of EGFR-CFP in control, GH-treated, or EGF-treated cells, consistent with the biochemical data on EGFR abundance in Fig. 7A. Remarkably, however, PD98059 treatment precluded the inhibitory effect of GH on EGF-induced EGFR-CFP intracellular redistribution (Fig. 8D, GH/EGF treatments, upper versus lower panel). Together with the data in Fig. 7, these fluorescence microscopy experiments suggest that GH lessens EGF-induced EGFR loss in an ERK pathway-dependent fashion, likely by inhibiting an early step(s) in EGFR trafficking.
To further probe the effect of GH, we assessed radiolabeled EGF internalization in these cells in the presence or absence of GH (Fig. 9). 3T3-F442A cells were treated at 37°C for 10 min with GH prior to the addition of murine 125 I-EGF (1 ng/ml). At intervals over 10 min after radiolabeled EGF addition, internalized 125 I-EGF was measured, as described under "Experimental Procedures." Radiolabeled EGF internalization, expressed as a fraction of total specific EGF binding, is plotted as a function of the duration of incubation in Fig. 9. Notably, GH pretreatment did not appreciably change the rate of 125 I-EGF internalization. We thus conclude that the changes in EGFinduced EGFR fate promoted by GH are not likely exerted at the earliest steps corresponding to EGF internalization.
GH Affects Acute EGF-induced EGFR Signaling-The findings above suggested that GH affects EGF-induced EGFR trafficking in such a way as to delay its degradation and downregulation. We explored whether this effect of GH might have impact upon early EGF signaling events. We first examined the effects of GH on EGF-induced ERK activation (Fig. 10). 3T3-F442A cells were exposed to GH for 0 -60 min, and detergent extracts were evaluated with anti-active ERK immunoblotting (Fig. 10A, upper panel). Consistent with previous reports (7,8,10), treatment with GH (500 ng/ml) caused ERK activation as early as 5 min with peak activation occurring between 5 and 15 min. Thereafter, ERK activation markedly declined. We next compared the level of ERK activation induced by EGF alone (15 or 30 min) with that induced by EGF in the presence of GH (Fig. 10B). As in the experiments above, in the cotreated samples, GH was added 10 min prior to EGF; thus, comparison treatments with GH alone were for 25 or 40 min. The level of ERK activation achieved by the combination of EGF plus GH was greater than EGF alone at both 15 and 30 min of EGF stimulation (Fig. 10B, upper panel, lane 4 versus 3 and lane 7  versus 6, respectively).  1 and 8), 500 ng/ml GH (lanes 2 and 7), 1 nM EGF (lanes 3 and 6), or with GH plus EGF (GH was added 10 min prior to EGF) (lanes 4 and 5) for 2 h. Total cell lysates (extracted in the presence of 1% SDS) (50 g) were resolved by SDS-PAGE and immunoblotted with an anti-EGFR. The experiment shown is representative of four such experiments. B, quantitative analysis by densitometry of the EGFR mass in the experiment shown in A. C, data shown in A, along with those obtained from three other experiments, were subjected to densitometric analysis. In each experiment, the EGFinduced loss of EGFR is considered as 100%. The data are plotted as a percentage of loss of EGFR mass relative to EGF-induced loss and are mean Ϯ S.E. of four experiments. The p value less than 0.05 is indicated.
To determine whether this augmentation reflected simply the summation of GH-induced plus EGF-induced ERK activation or instead reflected synergistic ERK activation (more than the summation) by EGF in the presence of GH, we performed densitometric evaluation of three such experiments. The results for ERK2 activation are displayed in Fig. 10C. In this figure, the open bars indicate the sum of the relative activation of ERK2 elicited by 15 min (left side) or 30 min (right side) of EGF alone (corresponding to lanes 3 and 6, respectively of Fig.  10B, upper panel) plus the ERK2 activation caused by GH alone (25 min, left side; 40 min, right side, corresponding to lanes 2 and 5, respectively, in Fig. 10B, upper panel). That summation is compared with the closed bars, which reflect the activation elicited by the cotreatment with EGF plus GH (15/25 min, left side; 30/40 min, right side, corresponding to lanes 4 and 7, respectively, in Fig. 10B, upper panel). Notably, no synergy was detected in comparing EGF only for 15 min with EGF for 15 min in the presence of GH (no difference in the height of the open and closed bars on the left of Fig. 10C), but significant synergistic augmentation was observed in comparing EGF treatment for 30 min in the presence of GH with EGF treatment alone for 30 min (p Ͻ 0.05 for the roughly 60% increase in the height of the closed bar versus the open bar on the right of Fig. 10C). Thus, GH cotreatment appeared to sensitize the EGF-induced ERK2 activation signals that continue after 30 min (but not 15 min) of EGF exposure. Because GH-induced ERK2 activation is minimal at this time (40 min of GH exposure), this suggests that changes in either the EGFR activation state or the itinerary of the EGF-activated EGFR brought about acutely by GH might affect the duration or strength of the EGF-induced ERK2 activation signal.
EGF-induced tyrosine phosphorylation of Cbl is a well documented early event in EGFR signaling in a range of cell types (46 -50). To further investigate the effects of GH on EGFRmediated signaling, we examined the tyrosine phosphorylation of Cbl (Fig. 11). We first tested whether EGF also induced tyrosine phosphorylation of Cbl in 3T3-F442A cells (Fig. 11A). Serum-starved cells were treated with EGF for 15, 30, or 90  1 and 3) and EGFR-CFP transfectants (lanes 2 and 4) (50 g) were resolved by SDS-PAGE and immunoblotted with anti-EGFR (lanes 1 and 2) or anti-GFP (lanes 3 and 4). Anti-GFP also recognizes CFP. C, GH-and EGF-induced tyrosine phosphorylation and PTP101 reactivity of EGFR-CFP in 3T3-L1/EGFR-CFP cells. Serum-starved 3T3-L1/EGFR-CFP cells were stimulated with GH (500 ng/ml), EGF (1 nM), or vehicle for 15 min. Detergent extracts (500 g) were immunoprecipitated with anti-EGFR. Eluted proteins were analyzed by immunoblotting with anti-pTyr (upper panel, lanes 1 and 2), PTP101 (upper  panel, lanes 3 and 4), and anti-EGFR (lower panels, lanes 1-4), respectively. D, effect of GH on EGF-induced EGFR-CFP redistribution. Serum-starved 3T3-L1/ EGFR-CFP cells were preincubated without (upper panel) or with 100 M PD98059 (lower panel) and then treated with vehicle, 1 nM EGF alone, or 1 nM EGF in the presence of 500 ng/ml GH (GH was added 10 min prior to EGF) for 15 min or 500 ng/ml GH for 25 min. The cells were fixed and visualized by fluorescence microscopy (ϫ100). The cell surface localization of EGFR-CFP is indicated by arrows. The experiments shown are representative of three such experiments. min, after which anti-Cbl-b antibodies were used for immunoprecipitation of detergent cell extracts, and resolved proteins were immunoblotted sequentially with anti-pTyr (upper panel) and anti-Cbl-b (lower panel). As anticipated, EGF stimulation induced Cbl tyrosine phosphorylation, which was detectable at 15 and 30 min but no longer after 90 min of EGF exposure (Fig.  11A, upper panel, lanes 3, 6, and 9 versus 1). In other experiments (not shown), we also detected the tyrosine-phosphorylated EGFR and SHC in anti-Cbl-b precipitates in response to EGF, as has been observed in other cell types. In contrast to EGF treatment, GH stimulation for 15, 30, or 90 min did not result in detectable tyrosine phosphorylation of Cbl (Fig. 11A,  upper panel, lanes 2, 5, and 8 versus 1). Notably, however, cotreatment with GH plus EGF at both 15 and 30 min of EGF exposure resulted in substantially increased Cbl tyrosine phosphorylation when compared with treatment with EGF alone (Fig. 11A, upper panel, lanes 4 versus 3 and lanes 7 versus 6). These findings strongly suggested that GH enhanced the ability of EGF to cause Cbl tyrosine phosphorylation.
To examine the basis for this augmentation of EGF-induced Cbl tyrosine phosphorylation by GH, we tested the effect of inhibition of the ERK pathway (Fig. 11B). Pretreatment with the MEK1 inhibitor, PD98059, had no effect on EGF-induced Cbl tyrosine phosphorylation (Fig. 11B, upper panel, lane 6  versus 3), nor did the drug reverse the inability of GH to cause Cbl tyrosine phosphorylation (Fig. 11B, lane 5 versus 2). PD98059 did, however, prevent the augmentation of EGF-induced Cbl tyrosine phosphorylation by GH (Fig. 11B, lane 7  versus 6 compared with lane 4 versus 3). Densitometric evaluation of three such experiments (Fig. 11C) indicated that GH cotreatment in the absence of PD98059 augmented EGF-induced Cbl tyrosine phosphorylation (15 min of treatment with EGF) by 70 Ϯ 13% (p Ͻ 0.05) but that cotreatment in the presence of PD98059 eliminated the GH-induced augmentation. DISCUSSION Emerging data suggest that the EGFR family signaling system has interrelationships with various other signaling systems. Interactions between EGFR itself or EGFR family members and the GHR have been recognized in the past several years. Yamauchi et al. (24) demonstrated that GH promotes EGFR tyrosine phosphorylation that is dependent on JAK2 activation but independent of an intact EGFR kinase domain. The implication of that study was that the EGFR serves as a docking molecule for the GHR-JAK2 system, allowing GH activation to trigger loading of the tyrosine phosphorylated EGFR with Grb-2, a key activator of the ERK pathway. Our studies in 3T3-F442A cells have confirmed that GH-induced EGFR tyrosine phosphorylation is not dependent on EGFR kinase activation (25). In the current study, we extend these findings to indicate that several EGFR tyrosine residues become phosphorylated in response to GH, most notably Tyr-845 and the previously implicated Tyr-1068. Our inability to discern more abundant overall EGFR tyrosine phosphorylation in the presence both GH and EGF is likely related to the overall stoichiometry for tyrosine phosphorylation induced by each peptide, with that induced by EGF far outweighing that in-  1 and 8). Total cell lysates (15 g) were resolved by SDS-PAGE and immunoblotted with anti-active ERK1/2 (upper panel) and anti-ERK1/2 (lower panel). C, augmentation of EGF-induced ERK2 activation by cotreatment with GH. Data shown in B, along with those obtained from two other experiments, were subjected to densitometric analysis. Data are normalized to the level of GH-induced ERK2 activation (the GH-induced ERK2 activation is considered as 100%) and are mean Ϯ S.E. of three experiments. Comparison of the level of ERK2 activation induced by cotreatment of GH plus EGF where GH was added 10 min prior to EGF (referred to as GH/EGF; closed bars) with the sum of those induced by EGF alone plus GH alone (referred to as GH alone ϩ EGF alone; open bars) at 15 and 30 min, respectively, is shown. For ERK2 activation by GH/EGF versus GH alone ϩ EGF alone at 30 min, p Ͻ 0.05, as indicated. duced by GH. It is difficult to determine without more detailed mutagenesis whether the GH-induced tyrosine phosphorylation of each EGFR site has particular implications for signaling in these cells.
Interestingly, engagement of various G protein-coupled receptors has also been shown to induce EGFR tyrosine phosphorylation, but in those instances this results from autophosphorylation associated with activation of the EGFR kinase domain. This so-called EGFR "transactivation" occurs with stimulation of receptors for angiotensin II, thrombin, carbachol, bombesin, lysophosphatidic acid, endothelin, and glucagon-like peptide-1, among others, and in those instances is mediated by metalloprotease-catalyzed release of EGF family ligands from their cell-surface precursors (51)(52)(53)(54)(55)(56)(57)(58)(59). The ligands then can engage EGFRs on the same or neighboring cells, causing their activation. Activation of the intrinsic kinase activity of ErbB-2 by interleukin (IL)-6 has also been demonstrated (60). In that study in prostate cancer cells, the ErbB-2 transactivation was unaccompanied by generation of EGF or EGF-family ligands but was instead envisioned to occur by oligomerization of ErbB-2 molecules associated with the gp130 subunit of the IL-6R complex. Because GH-induced activation of the EGFR kinase does not underlie GH-induced EGFR tyrosine phosphorylation, we view the cross-talk of GH with EGFR as mechanistically and probably functionally distinct from these examples of EGFR family member transactivation by G proteincoupled receptor activators and IL-6. Our findings that GH alone does not promote EGFR internalization and degradation are consistent with the lack of induction of EGFR kinase activation by GH.
Although we cannot completely rule out the possibility that GH can cause phosphorylation of particular ErbB-2 tyrosine residues, our data in 3T3-F442A and 3T3-L1 cells strongly suggest that, unlike its effects on EGFR, GH induces a lessening of both basal and EGF-induced overall tyrosine phosphorylation of ErbB-2 and also decreases ErbB-2 tyrosine kinase activity (25). This finding is also in contrast to the effects of IL-6 on ErbB-2 in cancer cells noted above (60). We note that both GH and IL-6 couple to signaling pathways that involve JAK kinases. The dissimilarities of the effects of the two ligands on ErbB-2 kinase activity in these two different systems remain as yet unexplained, but strongly suggest that the milieu in which the JAKs are activated and/or the other downstream signaling pathways activated likely dictate the net effect of non-EGF family ligands on EGFR family responses.
The current study in 3T3-F442A, a preadipocytic fibroblast that endogenously expresses GHR, ErbB-2, and EGFR, highlights the finding that both ErbB-2 and EGFR are also targets for GH-induced phosphorylation at sites other than tyrosine residues. In particular, use of the PTP101 monoclonal antibody for immunoblotting allowed us to detect prominent GH-induced phosphorylation of both molecules at what are most likely ERK consensus phosphorylation sites. The GH-induced PTP101-reactive phosphorylations of ErbB-2 and EGFR were strongly inhibited by two different MEK1 inhibitors and not inhibited by a PKC inhibitor, furthering the conclusion that GH-induced ERK activation is involved. The EGFR family members are at least relatively specifically targeted in that, by contrast, GHinduced ERK activation does not lead to PTP101-reactive phosphorylation of the PDGFR in the same cells. Notably, although GH induces ERK consensus phosphorylation of both ErbB-2 and EGFR, the consequences of that phosphorylation may differ substantially. As noted previously, ErbB-2 becomes relatively desensitized for basal and EGF-induced kinase activation in a manner correlated to the GH-induced phosphorylation. In contrast, EGFR kinase activation and trafficking appear unaffected by GH treatment alone. Rather, EGFinduced EGFR trafficking is substantially influenced by GH in a pattern corresponding to the ability of GH to activate the ERK signaling pathway. Our data indicate that both EGFinduced EGFR degradation and early EGF-induced EGFR redistribution from the cell surface to internal pools are substantially lessened by GH and that these GH effects are blocked by MEK1 inhibition.
Previous studies have implicated serine/threonine kinases as important in regulation of EGFR function. In particular, two major threonine phosphorylation sites in the juxtamembrane cytoplasmic domain have been identified, Thr-654 and Thr-669 (37,(61)(62)(63). PMA can lead to the phosphorylation of both, but protein kinase C is believed to directly mediate Thr-654 phosphorylation, whereas Thr-669 phosphorylation is thought to be phosphorylated by ERKs (64, 65). Indeed, Thr-669 lies in the only ERK phosphorylation consensus in the EGFR cytoplasmic domain and is a known target for EGF-induced EGFR phos-  1 and 8), 500 ng/ml GH (lanes 2 and 5), 1 nM EGF (lanes 3 and 6), or GH plus EGF (GH was added 10 min prior to EGF) (lanes 4 and 7) for 15 min. In A and B, detergent extracts (500 g) were immunoprecipitated with anti-Cbl-b. Eluted proteins were resolved by SDS-PAGE and immunoblotted with anti-pTyr (A and B, upper panels) and anti-Cbl-b (A and B, lower panels), respectively. C, data shown in B, along with those obtained from two other experiments, were subjected to densitometric analysis. In each experiment, the EGF-induced Cbl-b tyrosine phosphorylation is considered as 100%. Data are mean Ϯ S.E. of three experiments. For Cbl-b tyrosine phosphorylation by GH/EGF versus EGF alone in the absence of PD98059, p Ͻ 0.05, as indicated.
phorylation. The exact role of these phosphorylations, if any, in EGFR signaling is as yet uncertain. For example, although PKC activation leads to decreased EGF-induced EGFR kinase activation, direct phosphorylation of EGFR by either PKC or ERKs is not believed to be sufficient to mediate this effect (66).
In contrast to the MEK1-dependent dampening effects of PKC activation on EGFR kinase activity reported previously (66), GH treatment of 3T3-F442A cells in our experiments does not cause detectable decrease in EGF-induced EGFR tyrosine phosphorylation. Instead, the effect we observe is a GH-induced, MEK1-dependent alteration of the EGF-induced fate of the EGFR. Notably, Bao et al. (41) reported a PMA-induced change in EGF-induced EGFR trafficking and down-regulation that was PKC-dependent and attributable to Thr-654 phosphorylation. Our findings that GH protects the EGFR from EGFinduced intracellular redistribution, degradation, and downregulation, in each case in a fashion reversible by PD98059, suggest, to our knowledge, a heretofore unappreciated role of the MEK1-ERK pathway in regulation of EGF-induced EGFR trafficking. We have yet to map the EGFR site that becomes phosphorylated in a MEK1-dependent fashion in response to GH, but Thr-669, a known ERK phosphorylation site, is a leading candidate, particularly in light of our findings that GF109203X, a PKC inhibitor, had no effect on GH-induced EGFR PTP101 reactivity. Likewise, we do not yet know whether threonine phosphorylation of EGFR per se (for example at Thr-669) is required for GH to promote these effects on EGFR trafficking. It will be important to pursue both these lines of investigation in discerning the mechanism(s) at play to explain our novel observations. The determinants of EGF-induced EGFR down-regulation are incompletely understood but appear to include activation of the tyrosine kinase activity of the receptor and its association with the docking protein Cbl, which harbors ubiquitin ligase activity that may target the EGFR for degradation (67)(68)(69). Whether GH is affecting the EGF-induced ubiquitination status of EGFR or other proteins involved in its intracellular trafficking and degradation is also a topic worthy of study, in light of our current findings. In this context, it is informative that the earliest measurable stages of EGF internalization were unaltered by GH in this system.
Our findings may also be relevant in the context of understanding the relationship between EGFR trafficking and EGFR signaling. Early observations suggested that EGF-induced EGFR trafficking was critical in down-regulating signals emanating from the EGF-engaged activated EGFR and that receptor mutants that were activable, but non-internalizing, exhibited augmented signaling (70). More recent studies have highlighted the findings that signaling may in some measure emanate from EGFRs that are internalized, making it difficult to predict the exact signaling consequences of perturbations in the post-ligand engagement trafficking of the activated EGFR (71)(72)(73). Interestingly, in 3T3-F442A cells, the attenuation of EGF-induced EGFR intracellular redistribution and loss by GH was accompanied by augmentation of early aspects of EGFinduced signaling (i.e. EGF-induced ERK activation and EGFinduced Cbl tyrosine phosphorylation), and PD98059 blocked this GH effect (at least for Cbl tyrosine phosphorylation). This suggests that the inhibition of EGFR down-regulation by GH may enable the activated EGFR to transmit signal(s) more effectively, either because it is not degraded and/or because it remains in a favorable proximity for signal generation because of its altered itinerary. Although changes in EGFR signaling attributable to manipulations in its trafficking have been described, our observations are notable in that they suggest that GH (and possibly other ERK activating cytokines) may be non-pharmacologic modulators of these processes. We do not yet know, nor may it necessarily be predictable in all instances, what effects such modulators may have on biological responses to EGF. We expect that these effects may vary, depending on the cellular context (e.g. neoplastic versus non-neoplastic), EGFR expression level, and the degree to which the modulator can activate ERKs.
GH is a major regulator of growth, differentiation, and metabolism in vertebrates (74). A rich body of knowledge indicates that GH may exert these effects in various ways. Some GH actions are indirect but mediated via the GH-induced systemic and/or local elicitation of IGF-1 production; other actions of GH may be independent of IGF-1 (reviewed in Ref. 75). The role of STAT5 in GH actions, including the expression of IGF-1 and other genes, is becoming increasingly appreciated (6,76). The work described herein indicates a role for GH-induced activation of ERKs in modulating the trafficking itinerary of the EGF-activated EGFR. In this regard, we suggest that GH, in addition to its known IGF-1-generating and direct effects, may also exert its actions by regulatory cross-talk with an important tyrosine kinase growth factor receptor system.