HER4-mediated Biological and Biochemical Properties in NIH 3T3 Cells

The EGF receptor family of tyrosine kinase growth factor receptors is expressed in a variety of cell types and has been implicated in the progression of certain human adenocarcinomas. The most recent addition to this family of receptors, HER4, was expressed in NIH 3T3 cells to determine its biological and biochemical characteristics. Cells expressing HER4 were responsive to heregulin (cid:98) 2 as demonstrated by an increase in HER4 tyrosine phosphorylation and ability to form foci on a cell monolayer. HER4 exhibited in vitro kinase activity and was able to phosphorylate the regulatory subunit of phosphatidylinositol 3-kinase and SHC. Peptide compe-tition studies identified tyrosine 1056 of HER4 as the phosphatidylinositol 3-kinase binding site and tyrosines 1188 and 1242 as two potential SHC binding sites. Inter- estingly, transfection of HER4 into NIH 3T3 cells con-ferred responsiveness to EGF with respect to colony formation in soft agar. It was also found that in response to heregulin (cid:98) 2, endogenous murine HER1 or transfected human HER1 became phosphorylated when HER4 was present. This demonstrates that HER1 and HER4 can exist in a heterodimer complex and likely activate each other by transphosphorylation. The human epidermal growth factor receptor (HER) 1 family of type I receptor tyrosine kinases has been linked to the progression of certain human adenocarcinomas. The mechanism by which these receptors function with respect to tumor development is thought to involve overproduction of the gene product resulting from an increase in gene copy number or expression (1–12). Receptor overexpression amino acids. Phosphotyrosine was incorporated into the peptide either by direct coupling of N -Fmoc- O -phosphotyrosine or by coupling of N -Fmoc-Tyr-OH in a post assembly reaction with di-tert-butyl-diethylphos- phoramidite (Aldrich) followed by oxidation with tert-butyl-hydroper-oxide. The peptides were cleaved from the resin by reaction with trifluoroacetic acid and purified by reversed-phase high performance liquid chromatography.

The EGF receptor family of tyrosine kinase growth factor receptors is expressed in a variety of cell types and has been implicated in the progression of certain human adenocarcinomas. The most recent addition to this family of receptors, HER4, was expressed in NIH 3T3 cells to determine its biological and biochemical characteristics. Cells expressing HER4 were responsive to heregulin ␤2 as demonstrated by an increase in HER4 tyrosine phosphorylation and ability to form foci on a cell monolayer. HER4 exhibited in vitro kinase activity and was able to phosphorylate the regulatory subunit of phosphatidylinositol 3-kinase and SHC. Peptide competition studies identified tyrosine 1056 of HER4 as the phosphatidylinositol 3-kinase binding site and tyrosines 1188 and 1242 as two potential SHC binding sites. Interestingly, transfection of HER4 into NIH 3T3 cells conferred responsiveness to EGF with respect to colony formation in soft agar. It was also found that in response to heregulin ␤2, endogenous murine HER1 or transfected human HER1 became phosphorylated when HER4 was present. This demonstrates that HER1 and HER4 can exist in a heterodimer complex and likely activate each other by transphosphorylation.
The human epidermal growth factor receptor (HER) 1 family of type I receptor tyrosine kinases has been linked to the progression of certain human adenocarcinomas. The mechanism by which these receptors function with respect to tumor development is thought to involve overproduction of the gene product resulting from an increase in gene copy number or expression (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Receptor overexpression often correlates with a high level of constitutive tyrosine kinase activity, and this activity is thought to promote signals leading to uncontrolled cell growth (13).
Several ligands that bind to and stimulate the kinase activity of the HER family members have been identified and are classified as EGF-like ligands. EGF, amphiregulin, betacellulin, and transforming growth factor ␣ are specific for HER1 (14). Heregulin and its rat homologue neu differentiation fac-tor, are a subfamily of the EGF-like ligands that have been shown to bind to and activate both HER3 and HER4 (15)(16)(17)(18)(19)(20)(21)(22). The molecular characterization of a HER2 ligand has not been reported; however, a number of factors have been shown to activate HER2 tyrosine kinase activity specifically (23)(24)(25).
Activation of the EGF receptor family of tyrosine kinase receptors, using EGFR as the model, is thought to involve the binding of ligand and subsequent homodimerization of the receptor resulting in a conformational change and activation of the intrinsic tyrosine kinase activity (58). Recent advances suggest that a more complex mechanism of HER activation can occur. Heregulin has been shown to bind and stimulate both HER3 and HER4 (15,26). Thus, the signaling repertoire of any one ligand can be multifaceted depending upon the cellular receptor expression patterns, potentially allowing the same ligand to deliver distinct downstream signals. Further, ligandinduced receptor activation in the HER family is not limited to homodimerization, for it has been demonstrated that many of the HER family members can associate with and activate each other (15,(27)(28)(29)(30). HER2 has been shown to heterodimerize with HER1, HER3, and HER4, and EGF stimulation of HER1 has been shown to activate HER3 signaling (31). Therefore, the signaling of receptors by multifunctional ligands on homoreceptor and heteroreceptor complexes could result in a multiplicity of downstream signaling events.
Activation of growth factor receptors results in the transmission of stimulatory signals from the outside to the inside of the cell. This process makes use of intracellular proteins, or targets, that contain Src homology region 2 (SH2) domains that recognize and bind to the activated receptor (32,33). A number of these proteins have been identified and shown to interact with membrane-associated and cytoplasmic tyrosine kinases (32, 34 -36). The specificity of their binding is dictated by receptor sequences that flank a specific phosphotyrosine residue (33,37). Since the HER family members are similar but not identical in their amino acid sequence, an expected diversity of associations with different SH2-containing proteins has been demonstrated. HER1 has been shown to associate with phospholipase C-␥, SHC, and Grb2 (13, 38 -42). HER2 can associate with phospholipase C-␥, SHC, and Ras-GTPase activating protein (43)(44)(45). HER3 has recently been shown to activate PI 3-kinase (31,41,46).
We have focused on characterizing the newest member of the HER family, HER4, with regard to its biological activity and activation of downstream targets. To this end, a collection of HER4-expressing 3T3 cell lines has been generated and used to assay the biological and enzymatic activities of this receptor.

EXPERIMENTAL PROCEDURES
Cell Culture-NIH 3T3 clone 7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (Life Technologies, Inc.) at 37°in 5% CO 2 (47). Cells expressing receptors * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
were produced by pooling G418 resistant colonies from cultures that had been cotransfected with the pSV2neo plasmid and maintained in 500 g/ml G418. Transfections were done using Lipofectamine (Life Technologies, Inc.). HER2 and HER4 plasmid construction have been described previously (15). Cell lines expressing both HER1 and HER4 were produced by transfection of HER1-expressing plasmid into the HER4 cell line and selecting clones by histidinol selection. Clonal lines were isolated and lines expressing similar amounts of HER4 were used.
Ligands-Heregulin ␤2 used in all assays was synthesized as described previously (48). Briefly, the EGF domain of heregulin ␤2 was fused to the Fc portion of a human IgG1 gene, and this construct was transfected and expressed in COS cells. The heregulin portion of the purified IgG1 fusion was cleaved away from the Fc domain with thrombin and purified.
Focus-forming Assay-Transfected NIH 3T3 clone 7 cells were trypsinized, and approximately 500 cells were mixed with 10 5 progenitor 3T3 clone 7 cells. The following day, EGF or heregulin ␤2 was added at 20 g/ml, and the cells were fed on a 2-day schedule. Cells were stained after 7-10 days with methylene blue/carboyl fuschia.
Construction, Expression, and Purification of PI 3-Kinase SH2-GST and SHC-GST Fusion Proteins-Two glutathione S-transferase (GST) fusion proteins containing each of the two SH2 domains of the p85 subunit of PI 3-kinase were constructed using the pGEX-2T plasmid vector (Pharmacia Biotech Inc.) with some modifications in the multiple cloning sites (49).
Cell Extractions and Immunoprecipitation-To analyze the phosphotyrosine content of receptors from cell extracts, confluent plates were stimulated with 100 ng/ml EGF or heregulin ␤2 and extracted in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 5 mM EDTA, 5% glycerol, 1 mM sodium orthovanadate). Insoluble material was pelleted by a 5-min microfuge spin, and 5% of a 100 mM plate was boiled in SDS loading buffer. For immunoprecipitations, 2 l of antiphosphotyrosine antibodies were added (py20-ICN) with 40 l of a 50% slurry of goat anti-mouse IgG-linked agarose (Sigma), and the mixture was mixed for 60 min at 4°C. Samples were washed twice with lysis buffer and boiled in SDS sample buffer. Analysis of HER4 protein was done by immunoprecipitating HER4 using a HER4-specific antibody, 6-4-11 (12), and immunoblotted using the same antibody. HER2 protein was analyzed by Western blotting cell extracts with the antibody AB3 anti-HER2 from Oncogene Science, Inc. HER1 immunoprecipitates made use of the AB1 anti-EGFR antibody from Oncogene Science. Antibodies for SHC 06-203, PI 3-kinase 06-195 were purchased from Upstate Biotechnology Inc. (UBI). Anti-EGFR blots were done with antibody 06-129 from UBI. All immunoblotting experiments used 125 I-labeled goat anti-mouse or -rabbit second antibody for detection (DuPont NEN).
In Vitro Kinase Assay-Cells were extracted, and the protein complexes were isolated as described above. After the last wash with lysis buffer, the samples were washed once with kinase buffer (20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2 , 150 mM NaCl) and then resuspended in 100 l of kinase buffer containing 10 Ci of [␥-32 P]ATP (DuPont NEN, NEG-002A) and incubated at room temperature for 10 min. The beads were then washed once with kinase buffer and boiled in SDS sample buffer.
Double Precipitation Procedure-In vitro kinased samples were isolated as described except that they were boiled for 2 min in 200 l of solubilization buffer (0.4% SDS, 50 mM triethanolamine chloride (pH 7.4), 100 mM NaCl, 2 mM EDTA (pH 7.4), and 2 mM 2-mercaptoethanol) instead of SDS sample buffer. The Sepharose beads were then pelleted, the supernatant was recovered, and iodoacetamide (final concentration 10 mM) was added, followed by 0.25 volumes of 10% (v/v) Triton X-100. The extract was then precipitated with antibodies specific for each target and incubated at 4°C for 60 min (UBI-PI 3-kinase 06-195, SHC-06-203). Protein A-Sepharose CL4B (40 l of a 50% solution; Pharmacia) was used to precipitate the antibodies. The Sepharose beads were then pelleted, washed twice with lysis buffer, and boiled in SDS sample buffer.
Peptide Inhibition Studies-HER4 was immunoprecipitated and radiolabeled in an in vitro kinase reaction and denatured by boiling in solubilization buffer (0.4% SDS, 50 mM triethanolamine chloride (pH 7.4), 100 mM NaCl, 2 mM EDTA (pH 7.4), and 2 mM 2-mercaptoethanol). GST fusion proteins purified on glutathione-Sepharose beads were used to precipitate HER4 in the absence or presence of 50 M peptide at 4°C for 60 min. Complexes were washed in lysis buffer and analyzed by SDS-PAGE. Radiolabeled HER4 was quantified by analysis on a Phos-phorImager (Molecular Dynamics).
Peptides-Peptides were synthesized by the solid-phase method on a Gilson multiple peptide synthesizer, model AMS422, using Fmoc amino acids. Phosphotyrosine was incorporated into the peptide either by direct coupling of N-Fmoc-O-phosphotyrosine or by coupling of N-Fmoc-Tyr-OH in a post assembly reaction with di-tert-butyl-diethylphosphoramidite (Aldrich) followed by oxidation with tert-butyl-hydroperoxide. The peptides were cleaved from the resin by reaction with trifluoroacetic acid and purified by reversed-phase high performance liquid chromatography.

HER2 and HER4 Expression and Focal
Transforming Activity-To compare the signaling properties of HER4 with that of HER1 and HER2, NIH 3T3 clone 7 cells were transfected with plasmids encoding HER1, HER2, or HER4, resulting in the selection of clonal cell lines. NIH 3T3 clone 7 cells, initially derived from the NIH 3T3 ATCC cell line, are useful in biological assays due to their flatter morphology and ability to remain contact-inhibited when grown to confluence. Immunoprecipitation of cell extracts using receptor-specific antibodies demonstrated that HER1 (47), HER2, and HER4 were overexpressed in the appropriate cell lines (data not shown). HER1, HER2, and HER4 transfectants were seeded onto a monolayer of parental cells and grown in the presence or absence of specific ligand (Fig. 1). Used as a positive control, the HER1expressing 3T3 clone 7 cell line was able to form foci only when stimulated with EGF. The HER2-expressing cell line was able to form foci in the absence of any added ligand and did not respond to either EGF or heregulin ␤2 (50 -52). HER4-expressing cells were able to form foci in the absence of ligand; however, this activity was further stimulated by the addition of recombinant heregulin ␤2. These results demonstrate that HER4 is able to induce cellular transformation similar to that shown for HER1 and HER2.
HER4 Tyrosine Phosphorylation-To determine whether heregulin-induced focal transforming activity in HER4 transfectants correlated with an increase in receptor tyrosine phosphorylation, extracts from ligand-stimulated and -unstimulated HER cell lines were denatured, and the levels of tyrosine phosphorylation were analyzed by immunoblot using an antiphosphotyrosine antibody (Fig. 2A). Analysis of extracts in this manner minimizes the impact of protease, phosphatase, and potential autokinase activities that might occur during precipitation. In the HER2 transfectants, the HER2 receptor was constitutively tyrosine-phosphorylated and unresponsive to ligand. The HER4-expressing line demonstrated an elevated level of tyrosine phosphorylation of a 180-kDa protein in the absence of ligand when compared with control 3T3-7 cells; however, phosphorylation could further be induced by exposure to heregulin ␤2. Immunoprecipitation of HER4 from extracts with a monoclonal antibody to HER4 (6-4-11) confirmed that the band is HER4 (Fig. 2B). Depending upon the extraction conditions, immunoprecipitation of HER4 resulted in the in vitro activation of receptor kinase activity. Lysis buffer containing Mg 2ϩ promoted the in vitro activation of HER4, whereas, in the absence of Mg 2ϩ , HER4 activation was ligand-dependent. The dependence of Mg 2ϩ on HER4 kinase activity is unique to HER4 and has no effect on HER1, HER2, or HER3 (data not shown).
HER4 Autokinase Activity-Since HER4 is tyrosine-phosphorylated in response to ligand we set out to determine whether this change in modification correlated with HER4 in vitro autokinase activity. Antiphosphotyrosine immunoprecipitates from HER4 transfectants demonstrated a heregulin ␤2dependent activation of a 180-kDa protein (Fig. 3). This band is HER4, as demonstrated by precipitation using HER4-specific antisera (data not shown). The untransfected control 3T3-7 cell line was not activated by EGF or heregulin ␤2, but addition of PDGF stimulated the endogenous PDGF receptor. In contrast to HER1 and HER4, immunoprecipitates from HER2-expressing cells had very little associated kinase activity. This finding was unexpected since we know that HER2 is biologically active (Fig. 1) and is tyrosine-phosphorylated ( Fig. 2A). This data identifies that HER4 in vitro kinase activity is activated in response to heregulin ␤2. The phosphorylation of receptors is specific to tyrosine residues as determined by resistance to alkaline hydrolysis (data not shown).
HER4 Association with PI 3-Kinase and SHC-The signaling through growth factor receptors results in the phosphorylation of downstream targets. We set out to identify HER4-associated targets and to determine if there were differences in signaling between HER4 and that of other HER family members. In vitro 32 P-labeled antiphosphotyrosine immune complexes were denatured, and the target of interest was reimmunoprecipitated with specific antibodies. Heregulin ␤2 stimulation of cells expressing HER4 were found to be associated with an increased level of tyrosine-phosphorylated p85, the regulatory subunit of the PI 3-kinase (Fig. 4A), as well as all splicing variants of SHC (Fig. 4B). Similar results were obtained when the initial precipitation was performed with antibody specific for HER4, demonstrating a physical association between HER4 and both PI-3 kinase and SHC (data not shown). EGF stimulation of HER1 induced the phosphorylation of SHC but not 85-kDa PI 3-kinase. HER2-expressing cells did not demonstrate any inducible activation of 85-kDa PI 3-kinase or SHC. In the untransfected control cell line, stimulation with PDGF, but not EGF and heregulin ␤2, resulted in the activation of 85-kDa PI 3-kinase, reflecting the activation of the endogenous mouse PDGF receptor. There was a detectable activation of SHC by EGF in the parental cell line, possibly due to low level activation of the endogenous mouse EGF receptor.

Identification of PI 3-Kinase and SHC Binding
Sites-HER4 contains a single putative binding site for PI 3-kinase (YTPM) at amino acid 1056 and three potential SHC binding sites (NPXY) at amino acids 1188, 1242, and 1284. To investigate whether or not these sites mediate interaction with these targets, phosphotyrosine-containing peptides correlating to these sequences were synthesized and used to inhibit the in vitro binding of PI 3-kinase and SHC to HER4 (Fig. 5A). GST was fused to the N-and C-terminal SH2 domains of PI 3-kinase and full-length SHC and used to precipitate HER4. Peptide 1056, which includes the PI 3-kinase binding motif, was able to inhibit the binding of both PI 3-kinase SH2 fusions to HER4 (Fig. 5B). No inhibition was seen when the unphosphorylated control peptide (1056Y) was used, nor with other peptides containing NPXY motifs. The GST-SHC fusion was also able to associate with HER4, and this binding was inhibited 74% by peptide 1188 and 91% by peptide 1242. These data suggest that there may be two SHC binding sites on HER4 and that 1242 is the primary site. In addition, peptide 1284, which also contains an NPXY motif, did not compete. This demonstrates that sequences adjacent to NPXY must be important for SHC recognition of the receptor.
Heterodimerization between HER4 and HER1-It has previously been demonstrated that certain members of the HER family of receptors can heterodimerize with one another (15,(27)(28)(29)(30)(31). We set out to determine whether or not HER4 was able to heterodimerize with HER1. The first indication that HER1 and HER4 could functionally interact was demonstrated in a growth in soft agar assay. 3T3-and 3T3/HER4-transfected cell lines were plated in agar in the presence of heregulin ␤2 or EGF. The control 3T3 cells did not form colonies under any condition, whereas the HER4 expressing cells formed colonies in response to both heregulin ␤2 and EGF (Fig. 6).
If HER1 and HER4 can functionally interact and elicit a growth response, then stimulation with the appropriate ligand might result in cross-phosphorylation of one receptor by the other. Stimulation of parental 3T3 cells resulted in increased tyrosine phosphorylation of mouse EGFR in response to EGF but not heregulin (Fig. 7A). However, in cells transfected with HER4, tyrosine phosphorylation of the endogenous EGFR was elevated in the absence of ligand and further increased in response to both EGF and heregulin ␤2. Increasing the level of EGFR expression also increased the level of heregulin ␤2induced EGFR tyrosine phosphorylation. In contrast, there was no detectable increase in HER4 phosphorylation in response to EGF (Fig. 7B). These results suggest that HER1 and HER4 can exist in a heterodimer complex and that transphosphorylation can occur in one direction, in that activation of HER4 results in the phosphorylation of HER1 but HER4 is not a substrate for HER1.

DISCUSSION
The activation of HER family members, as well as their respective signaling through downstream targets is different for each receptor. Further, the complexity of these signaling pathways becomes even more diverse due to receptor heterodimerization, which may alter the specificity or strength of response to a given ligand. It has been demonstrated that HER1, HER2, and HER3 can interact with one another (15,(27)(28)(29)(30)(31). It has also been shown that HER4 activation can result in the phosphorylation of HER2. In this paper, we describe some of the properties of HER4 that are similar and different from those of other family members. We were able to demonstrate that HER4 can also complex with HER1, resulting in an enhanced growth signal in agar in response to EGF and the transphosphorylation of EGFR by HER4 in response to heregulin ␤2. The endogenous level of EGFR tyrosine phosphorylation was elevated in cells co-expressing HER4, suggesting that the heterodimer may exist at low levels in the absence of ligand. Increasing the level of EGFR by transfecting HER1 resulted in a higher level of both ligand-independent and heregulin-stim-ulated HER1 phosphorylation, suggesting that the stoichiometry of receptor heterodimers is dependent upon receptor expression levels. We did not observe an EGF-dependent phosphorylation of HER4, which suggests that HER4 is not a substrate for HER1. Nevertheless, EGF could induce colony formation in cells expressing HER4, which suggests that a HER1-HER4 complex is being formed. It remains to be investigated whether or not HER4 and HER3 can form a complex that would complete the possible combinations of currently known HER family member heterodimerizations.
Biologically, we have shown that HER4 can stimulate 3T3 cells to grow and overcome cell-to-cell contact inhibition at a low level in the absence of exogenous ligand and that this  4. HER4 can associate with SHC and PI 3-kinase. A, antiphosphotyrosine immunoprecipitates from control and ligand-stimulated cell lines were labeled in an in vitro kinase assay. Labeled complexes were denatured by boiling in solubilization buffer, and 85-kDa PI 3-kinase (A) and SHC (B) were reimmunoprecipitated with specific antisera. Purified immune complexes were analyzed on SDS-PAGE, and gels were dried down and exposed on a PhosphorImager screen. activity can be further induced by the addition of heregulin ␤2 but not EGF. Conversely, both heregulin ␤2 and EGF were able to induce the growth of colonies in soft agar in HER4-expressing cells but not the control NIH 3T3 cells. It is unclear why there is a difference in EGF responsiveness between the focal transformation assay and the agar assay. Perhaps the heterodimer complex is providing the necessary downstream signals needed for growth in soft agar but not the signals needed to overcome contact inhibition.
In analyzing HER4 tyrosine phosphorylation in rapidly denatured extracts, HER4 is slightly phosphorylated in the absence of ligand and can be further induced by the addition of heregulin ␤2. Immunoprecipitation of HER4 in the presence of Mg 2ϩ results in the in vitro stimulation of HER4 tyrosine phosphorylation and stimulation of HER4 in vitro kinase activity. This activation, in the absence of ligand, is likely a result of concentrating the receptor by immunoprecipitation and is unique to HER4, for it does not occur with HER1, HER2, and HER3. Removal of Mg 2ϩ from the lysis buffer restores ligand responsiveness to our biochemical analysis. While HER1 and HER4 are both ligand-responsive, HER2 is constitutively hyperphosphorylated.
The in vitro kinase activity associated with these receptors provided a means to identify downstream targets activated by these receptors. HER4 was able to induce the phosphorylation of both SHC and 85-kDa PI 3-kinase. HER1 could activate SHC but not 85-kDa PI 3-kinase, which identifies a divergence of activities between HER1 and HER4. HER2 kinase activity and subsequent association with downstream targets could not be determined due to the lack of in vitro kinase activity associated with the receptor. HER2 is constitutively tyrosine-phosphorylated and appears to associate with downstream targets in other types of assays (43-45, 53, 54). The lack of HER2 in vitro kinase activity remains to be investigated.
The association of SH2-containing proteins with growth factor receptors is integral in transmitting growth-stimulatory signals. The elements required for the association of SH2containing proteins to tyrosine kinases involve a phosphorylated tyrosine residue flanked by specific sequences. Specifically, YXXM (34) and NPXY (37) are essential motifs for the binding of PI 3-kinase and SHC, respectively, to their target kinases. The C-terminal domain of HER4 contains one YXXM motif (YTMP). Peptide 1056, corresponding to this motif, was able to inhibit binding of both N-and C-terminal PI 3-kinase SH2 domains. HER1 does not contain a YXXM motif in its C terminus, which correlates with the inability to phosphorylate the 85-kDa subunit of PI 3-kinase in our assay. These data confirm the essential nature of the YXXM motif in HER4 for binding PI 3-kinase and that each of the PI 3-kinase SH2 domains alone is sufficient for association with HER4.
HER4 contains three potential SHC binding sites (NPXY) within its C-terminal coding sequence located at amino acids 1188, 1242, and 1284. Full-length SHC was able to form a complex with HER4, and this interaction was inhibited by peptides 1188 and 1242. This suggests that there are two potential SHC binding sites on HER4. Recent observations have identified a second motif in the N-terminal portion of SHC, distinct from the SH2, that is able to associate with tyrosine-phosphorylated EGFR (55) and a SHC-associated protein, p145 (56). Identification of the domain within SHC required to associate with HER4 is under investigation.
The relationship between the HER family of receptors and their ability to respond to similar ligands resulting in both homo-and heterodimerized complexes lends itself to a very complex system of cell signaling. Since multiple HER members are often expressed concomitantly, the specific combination and relative level of expression of different receptors will determine the response to a given ligand. This is evident in human tumor cell lines in which heregulin can induce either a mitogenic (21,57) or a terminally differentiating signal (17), depending upon the cell type. Thus, altering the pattern of expression of HER family members may offer a selective growth advantage during tumor progression in the presence of HER family ligands. It has been shown that overexpression of several EGF-like growth factors, such as transforming growth factor ␣, amphiregulin, and cripto-1, represents a hallmark feature of many solid tumors. A better understanding of the downstream signals produced from both receptor homo-and heterodimers, as well as the ligands involved, is critical to a better understanding of the biology of HER family signal transduction and its role in tumor progression. FIG. 7. Complex formation between HER1 and HER4. NIH 3T3 cells expressing HER1, HER4, or both were stimulated with either EGF or heregulin ␤2. Receptors were immunoprecipitated following cell lysis and analyzed on SDS-PAGE followed by transfer onto nitrocellulose. The blot was probed with the indicated antibody followed by 125 I-labeled protein A.