Epidermal growth factor-related peptides activate distinct subsets of ErbB receptors and differ in their biological activities.

Numerous epidermal growth factor (EGF)-related peptide binding members of the ErbB family of receptor tyrosine kinases have been described. While several EGF agonists bind and activate ErbB-1/EGF receptor, neu differentiation factor (NDF) functions as a ligand for ErbB-3 and ErbB-4. However, it is currently unknown which specific subsets of ErbB receptors become activated in response to each of these ligands. The present study addresses this issue using the T47D breast tumor cell line, which expresses moderate levels of all the presently known ErbB receptors. We show that all the EGF agonists, but not NDF, stimulated tyrosine phosphorylation of ErbB-1. In contrast, all the EGF-related factors except amphiregulin were able to induce tyrosine phosphorylation of ErbB-2. The ability to induce tyrosine phosphorylation of ErbB-3 varied dramatically among the different EGF-related peptides. While EGF, transforming growth factor (TGF)-alpha, and amphiregulin only had a moderate effect, NDF dramatically increased the ErbB-3 phosphotyrosine content. Most notably, heparin binding EGF-related growth factor (HB-EGF) and betacellulin (BTC) were more effective than other EGF agonists. Consequently, only NDF, HB-EGF, and BTC significantly stimulated association of phosphatidylinositol kinase activity with ErbB-3. Among the EGF agonists, HB-EGF induced a low level of ErbB-4 tyrosine phosphorylation, while BTC was as efficient as NDF in activating ErbB-4. The BTC activation of ErbB-4 appears to be independent of ErbB-1, as shown by pretreatment of cells with an antibody that inhibits binding of EGF agonists to ErbB-1. As a result of the differential activation of ErbB receptors, most of the EGF-related growth factors had distinguishable biological activities on cultured mammary epithelial cell lines.

Many polypeptide growth factors exert their function by binding cell surface receptors with intrinsic protein tyrosine kinase activity. Ligand binding to the extracellular domain of a receptor tyrosine kinase (RTK) 1 induces dimerization, activation of the intracellular kinase domain, and autophosphorylation by an intermolecular mechanism (1). Tyrosine-phosphorylated residues serve as high affinity binding sites for SH2 or phosphotyrosine binding domain containing proteins and allow for the modulation of intracellular signaling pathways (2,3). RTKs play a fundamental role in the regulation of cell growth and differentiation. A large number of RTK subclasses has been described, among which the type I/ErbB family of RTKs is of particular interest due to their frequent involvement in human cancer. Four members of this family are currently known: epidermal growth factor (EGF) receptor/ErbB-1, ErbB-2, ErbB-3, and ErbB-4 (4 -7). Aberrant expression of ErbB-1 has been observed in various human tumors (8). Overexpression of ErbB-2 in the presence or absence of gene amplification is frequently found in tumors arising at many sites, especially of the breast and ovary where it correlates with poor patient prognosis (9). High levels of ErbB-3 have been described both in breast tumor cell lines (10) and in primary breast cancer (11). In order to fully understand the normal function of ErbB receptors as well as their role in neoplastic transformation it is crucial that their regulation by growth factors be elucidated in detail.
Regulation of ErbB receptor activity appears to be very complex since a large number of ErbB ligands have been described. All of these peptides are derived from transmembrane precursors and are cleaved to give rise to the mature, soluble ligands. Depending on their binding specificities, they can be subdivided into two classes. The first group of ligands binds to the EGFR and includes EGF itself, transforming growth factor (TGF) ␣, amphiregulin (AR), heparin-binding EGF-like growth factor (HB-EGF), and betacellulin (BTC) (12)(13)(14)(15)(16)(17). Each of these peptides competes with EGF for receptor binding (13,14,16,17) and therefore this family of growth factors is referred to as the EGF agonists. Another family of EGF-related peptides is composed of the neu differentiation factors (NDFs)/heregulins (18,19), ligands for ErbB-3 and ErbB-4 (20,21). There are at least 12 different isoforms arising from a single gene by alternative splicing and depending on the sequence of their EGFlike repeat they are classified as either ␣ or ␤ isoforms (22). However, despite the large number of NDFs no differences in receptor binding specificities appear to exist: ErbB-3 functions as a low affinity receptor for all NDF isoforms while ErbB-4 serves as a high affinity receptor (23).
An additional level of complexity in ErbB-mediated signaling has been brought forward by the finding that binding of a specific ligand not only triggers the formation of receptor homodimers, but also heterodimers. For instance, EGF induces tyrosine phosphorylation of ErbB-2 through formation of ErbB-1/ErbB-2 dimers (24,25). Similarly, NDF stimulates tyrosine phosphorylation of ErbB-2, presumably through formation of ErbB-3/ErbB-2 and ErbB-4/ErbB-2 dimers (21,26,27). An important consequence of heterodimerization of EGF and NDF receptors with ErbB-2 is a dramatically increased affinity for the respective ligand (26,28,29). However, heterodimerization may also serve to diversify the nature of the intracellular signal elicited by a specific growth factor. The prototype example supporting this notion is the recent observation that EGF stimulates the association of phosphatidylinositol (PtdIns) 3-kinase with ErbB-3, presumably through EGFR/ErbB-3 heterodimerization followed by cross-phosphorylation (30,31).
In view of the large number of ErbB ligands and the extensive cross-talk occurring between ErbB receptors, the knowledge about functional differences in the signaling capacities of EGF-related peptides has remained fragmentary. Thus, the present study was designed as a comparative analysis of all the currently available EGF-related growth factors with respect to activation of ErbB receptors and biological activity.
Immunoprecipitation and Western Blot Analysis-Cells were solubilized in Triton X-100 extraction buffer (50 mM Tris, pH 7.5, 5 mM EGTA, 150 mM NaCl, 1% Triton X-100, 2 nM sodium orthovanadate, 50 mM sodium fluoride, 10 mM sodium molybdate, 20 M phenylarsine oxide, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin) for 10 min on ice. The lysates were collected, transferred to an Eppendorf tube, and clarified by centrifugation for 10 min at 16,000 ϫ g. For immunoprecipitations, equal amounts of protein (usually 1-2 mg) were incubated with specific antibodies for 2 h on ice. Immune complexes were collected with protein A-Sepharose (Sigma), washed three times with extraction buffer, and once with TNE (50 mM Tris-HCl, pH 7.5, 140 mM NaCl, 5 mM EDTA). Bound proteins were released by heating 10 min at 95°C in sample buffer. Total cell lysates or immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and proteins were blotted to polyvinylidene difluoride membranes. After blocking with 20% horse serum (Life Technologies, Inc.) in TTBS (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween 20), filters were probed with specific antibodies and proteins visualized with peroxidase-coupled secondary antibody using the ECL detection system (Amersham). Filters were stripped in a buffer containing 62.5 mM Tris-HCl, pH 6.8, 2% SDS, and 100 mM ␤-mercaptoethanol for 30 min at 65°C, washed three times in TTBS, blocked, and reprobed with the indicated antibodies.
Assay for ErbB-3-associated PtdIns Kinase Activity-Cells were lysed as described above and ErbB-3 was immunoprecipitated from 500 g of total protein extract. The immune complexes were washed once in phosphate-buffered saline, twice in 0.1 M Tris, pH 6.8, 0.5 M LiCl, once in 10 mM Tris, pH 7.5, 100 mM NaCl, 1 mM EDTA, and resuspended in 30 l of a kinase mixture containing 100 mM Hepes, pH 7.5, 5 mM MgCl 2 , 10 M ATP, 10 Ci of [␥-32 P]ATP, and 10 g of sonicated PtdIns (Sigma). After 10 min at 30°C the reaction was stopped and extracted by the addition of 0.4 ml of 1 M HCl and 0.8 ml of chloroform/methanol (1:1 (v/v) mixture). The lipid containing organic phase was extracted with 0.32 ml of 1 M HCl, methanol (1:1 (v/v) mixture), dried down in a Speed Vac, redissolved in 20 l of chloroform, and resolved on thin-layer chromatography plates (Silica Gel 60; Merck) developed in chloroform, methanol, 4 M ammonium hydroxide (9:7:2 (v/v/v) mixture). Radiolabeled spots corresponding to PtdIns phosphate were quantitated using a PhosphorImager (Molecular Dynamics).
Cell Growth Assays-MCF10A cells were plated in triplicate in se-rum-free medium (Dulbecco's modified Eagle's medium/F-12 containing 1 mg/ml fetuin, 10 g/ml transferrin, 5 g/ml insulin, and 1 M dexamethasone) in the presence or absence of 100 ng/ml EGF-related growth factors (2 ϫ 10 3 cells/well/96-well dish). MDA MB 453 cells were plated in triplicate in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum in the presence or absence of 5 nM EGF-related growth factors (10 4 cells/well/96-well dish). HC11 cells were plated in triplicate in RPMI supplemented with 0.5% fetal calf serum and either 5 g/ml insulin or 100 ng/ml EGF in the presence or absence of 100 ng/ml EGF-related growth factors (2 ϫ 10 3 cells/well/96-well dish). Growth was monitored after 3 days (MCF10A and HC11 cells) or 4 days (MDA MB 453 cells) using the Cell Titer AQ TM kit (Promega). Soft Agar Growth Assay-T47D cells (3 ϫ 10 3 ) were plated in duplicate in 6-cm dishes in 6 ml of culture medium supplemented with 0.35% Noble agar (Difco) overlaying a 0.7% agar layer. For growth stimulation, 1 nM EGF-related growth factor was added to the medium. The cells were incubated for 4 weeks at 37°C, after which colonies were stained by adding 2 ml of phosphate-buffered saline containing 0.5 mg/ml of nitro blue tetrazolium (Sigma) for 24 h. Colonies larger than 200 m were counted with an Artek 880 colony counter (Dynatech Laboratories, Inc.).

EGF-related Growth Factors Differentially Induce Tyrosine Phosphorylation of ErbB-1 and ErbB-2-
The T47D breast tumor cell line has proven to be an invaluable model system for the study of ErbB-mediated signaling, since the cells express moderate levels of all the currently known ErbB receptors (25,27). T47D cells were treated with EGF, TGF␣, HB-EGF, AR, BTC, or NDF, and ErbB-1 was immunoprecipitated and analyzed for its phosphotyrosine content by Western blotting with a specific antibody (Fig. 1A). All the EGF agonists induced tyrosine phosphorylation of ErbB-1. While EGF, TGF␣, HB-EGF, and BTC showed similar activities, AR was remarkably less effective. In striking contrast to the EGF agonists, NDF had no effect on ErbB-1 phosphorylation.
Previously, NDF and several EGF agonists have been shown to induce the tyrosine phosphorylation of ErbB-2 (19, 24, 36). However, BTC-or HB-EGF-induced phosphorylation has not been reported so far. Thus, we compared all the factors with T47D cells were starved for 24 h in serum-free medium and treated for 10 min at room temperature with 100 ng/ml of the indicated factors prior to lysis. ErbB-1 and ErbB-2 were immunoprecipitated with, respectively, a 1:1 mixture of mAb EGFR and mAb 528 or the antiserum 21N, subjected to SDS-PAGE, and analyzed by Western blotting with a phosphotyrosine-specific mAb (upper panels). Filters were stripped and reprobed with, respectively, antiserum 15E or 21N (lower panels).
respect to activation of ErbB-2. With the exception of AR, the EGF agonists, as well as NDF, induced tyrosine phosphorylation of ErbB-2 to a similar extent, presumably due to the ability of both EGF-and NDF-receptors to heterodimerize with this receptor (Fig. 1B).

EGF-related Growth Factors Differentially Induce Tyrosine Phosphorylation of ErbB-3 and Association with PtdIns Kinase
Activity-Both EGF and NDF have been reported to induce tyrosine phosphorylation of ErbB-3. However, little is known about activation of ErbB-3 by the various EGF agonists. Thus, T47D cells were stimulated with each of the peptides and ErbB-3 was immunoprecipitated and analyzed for its phosphotyrosine content by Western blotting with a specific antibody ( Fig. 2A). Treatment with both NDF and EGF led to an increased tyrosine phosphorylation of ErbB-3. Compared to NDF, however, EGF only had a very small effect on ErbB-3. This may be due to the relatively low ErbB-1 levels expressed in these cells. In ErbB-1 overexpressing MDA MB 468 cells a much more pronounced effect on ErbB-3 phosphorylation has been observed (31). The EGF agonists differed significantly in their ability to activate ErbB-3. While TGF␣ and AR were similar to EGF, HB-EGF and BTC had a much stronger effect on ErbB-3 phosphorylation. This observation is of particular interest since it implies a potential for differential activation of ErbB-3-specific pathways by EGF agonists.
ErbB-3 is unique among subclass I RTKs due to the presence of several Tyr-X-X-Met motifs in its intracellular domain, the consensus binding site for the p85 subunit of PtdIns 3-kinase (37). Consequently, both EGF and NDF induce the association of PtdIns 3-kinase activity with ErbB-3 (30,31). Given the differential effects of EGF agonists on ErbB-3 phosphorylation, we analyzed their ability to induce association of PtdIns-kinase activity with this receptor. T47D cells were stimulated with the various peptides, ErbB-3 was immunoprecipitated, and the associated PtdIns kinase activity was measured in an in vitro kinase assay using PtdIns as a substrate (Fig. 2B). In good correlation with the effects on tyrosine phosphorylation of ErbB-3, NDF dramatically induced ErbB-3 association of PtdIns kinase activity, while the EGF agonists were generally less active. Significantly, HB-EGF and BTC stimulated PtdIns kinase activity more strongly than the other EGF agonists. Thus, by differentially inducing tyrosine phosphorylation of ErbB-3, EGF-related growth factors have different abilities to couple to ErbB-3-specific pathways.
EGF-related Growth Factors Differentially Induce Tyrosine Phosphorylation of ErbB-4 -Little is known about the potential interaction of EGF agonists and ErbB-4, another receptor for NDF (21). Thus, T47D cells were stimulated with each of the EGF-related peptides, and ErbB-4 was immunoprecipitated and analyzed for tyrosine phosphorylation by Western blotting (Fig. 3A). As shown before (27), NDF rapidly induces phosphorylation of ErbB-4 in T47D cells. Among the EGF agonists HB-EGF induced a low level of ErbB-4 phosphorylation, while BTC was found to be as efficient as NDF in activating ErbB-4.
There are at least two possibilities for the strong BTC-induced increase in tyrosine phosphorylation of ErbB-4. Either BTC directly binds and activates ErbB-4, or BTC binds to ErbB-1 and induces the formation of ErbB-1/ErbB-4 heterodimers followed by cross-phosphorylation. To test these possibilities, T47D cells were pretreated with 10 g/ml mAb 225, which competes with EGF for binding to ErbB-1, prior to stimulation with EGF or BTC (Fig. 3B). While mAb 225 readily abolished both EGF-and BTC-induced tyrosine phosphorylation of ErbB-1, it only partially inhibited BTC-induced phosphorylation of ErbB-4. Thus, BTC appears to directly bind and activate ErbB-4, although the availability of ErbB-1 seems to enhance the activation of ErbB-4, presumably through ErbB-

FIG. 2. Tyrosine phosphorylation of ErbB-3 and association with PtdIns kinase activity.
After 24 h starvation in serum-free medium, T47D cells were treated for 10 min at room temperature with 100 ng/ml of the indicated factors, lysed, and ErbB-3 was immunoprecipitated with antibody C17. A, immunoprecipitates were subjected to SDS-PAGE and analyzed by Western blotting with a phosphotyrosinespecific mAb (upper panel). Filters were stripped and reprobed with antibody C17 (lower panel). B, associated PtdIns kinase activity was determined in an immune complex kinase assay using PtdIns as a substrate. Bars show the fold stimulation of PtdIns kinase activity in comparison to untreated cells. FIG. 3. A, tyrosine phosphorylation of ErbB-4. T47D cells were starved for 24 h in serum-free medium and treated for 10 min at room temperature with 100 ng/ml of the indicated factors prior to lysis. ErbB-4 was immunoprecipitated with antibody C18, subjected to SDS-PAGE, and analyzed by Western blotting with a phosphotyrosine-specific mAb (upper panel). Filter was stripped and reprobed with antibody C18 (lower panel). B, competition of mAb 225 with EGF and BTC. T47D cells were starved for 24 h in serum-free medium and either left untreated or pretreated with 10 g/ml mAb 225 ( ϩ mAb) for 10 min at room temperature prior to the addition of 1 nM EGF or BTC. ErbB-4 (left panel) and ErbB-1 (right panel) were immunoprecipitated with, respectively, antibody C18 or a 1:1 mixture of mAb EGFR1 and mAb 528, subjected to SDS-PAGE, and analyzed by Western blotting with a phosphotyrosine-specific antibody.

1/ErbB-4 heterodimerization.
EGF-related Growth Factors Differ in Their Biological Activity-To characterize the biological properties of the various EGF-related growth factors, each of the peptides was tested for its ability to stimulate the growth of mammary epithelial cell lines (Fig. 4). First, the factors were tested for their ability to replace EGF in the growth medium of MCF10A cells, a spontaneously immortalized human mammary epithelial cell line that requires EGF for growth (38) (Fig. 4A). Under serum-free conditions, most of the EGF agonists had similar activities and caused approximately a 3-fold stimulation in growth. In contrast, NDF and AR stimulated growth significantly less, showing that the activation of ErbB-1 is important for a mitogenic effect on MCF10A cells. Second, the factors were tested on MDA MB 453 cells, a human breast tumor cell line that does not contain detectable amounts of ErbB-1 (39) (Fig. 4B). Most of the EGF agonists had only a marginal effect on the growth of these cells. In contrast, NDF was significantly more active, obviously due to the fact that NDF signaling does not involve ErbB-1. Most importantly, BTC was comparable to NDF, in agreement with the ability of BTC to elicit ErbB-1-independent signaling (see Fig. 3B).
We have previously shown that NDF, but not EGF, stimulates the anchorage-independent growth of T47D cells (25). To further characterize the biological activities of the EGF-related peptides, we tested their ability to stimulate T47D colony formation under anchorage-independent conditions (Table I). While TGF␣ and AR, like EGF, had no significant effect on growth, NDF stimulated colony formation by almost 6-fold. Similarly, HB-EGF and BTC led to a 2-3-fold increase in the number of colonies. Thus, EGF-related factors show distinguishable biological activities, most likely depending on the subsets of ErbB receptors that become activated.
NDF Cooperates with EGF to Stimulate the Growth of Mammary Epithelial Cells-Many normal mammary epithelial cell lines require the synergistic action of EGF-and insulin-like growth factor-1 receptors for optimal growth under serumrestricted conditions. This is usually achieved by maintaining the cells in EGF and superphysiological concentrations of insulin. HC11 cells are a mouse mammary gland-derived cell line that has retained many important characteristics of mammary epithelial cells (40), including a requirement of insulin and EGF for growth under serum-restricted conditions. To analyze in more detail the biological activities of EGF-related growth factors, their ability to replace EGF or insulin for growth under serum-restricted conditions was evaluated (Fig. 5). As already observed with MCF10A cells, TGF␣, HB-EGF, and BTC were as effective as EGF in stimulating growth in insulin-containing medium (Fig. 5A). Significantly, NDF was almost as good as EGF in stimulating growth of HC11 cells. Thus, all the EGFrelated factors are able to cooperate with insulin for stimulating the growth of HC11 cells. Next, the ability of the different peptides to replace insulin was tested (Fig. 5B). While none of the EGF agonists significantly stimulated growth in EGF-containing medium, NDF could partially replace insulin and reproducibly stimulated growth by approximately 40%. Thus, NDF activity is unique among EGF-related peptides due to its ability to cooperate with both EGF and insulin for stimulating the growth of mammary epithelial cells.

DISCUSSION
In this report we demonstrate that EGF-related growth factors differ in their ability to induce tyrosine phosphorylation of the various ErbB receptors in T47D breast tumor cells (see Table II for a summary). Most notable are the different abilities of EGF agonists to induce ErbB-3 tyrosine phosphorylation, as well as the ability of BTC to efficiently activate ErbB-4 in an apparently ErbB-1-independent manner. We show that in line with their effect on ErbB-3 tyrosine phosphorylation, only NDF, BTC, and HB-EGF significantly stimulated the recruitment of PtdIns kinase activity to this receptor. Furthermore, we demonstrate that the activation of different receptor subsets by EGF-related peptides results in distinct biological activities on cultured mammary epithelial cell lines.
While most of the EGF agonists were similarly effective in inducing tyrosine phosphorylation of ErbB-1 in T47D cells and stimulating growth of MCF10A cells, AR was strikingly less active. This is in agreement with the observation that AR has a significantly lower affinity for ErbB-1 than EGF (14). The lower affinity of AR appears to be due to the fact that the mature form (either AR 1-84 or AR 7-84 , due to alternate N-   terminal cleavage) is truncated at its C terminus and lacks a leucine residue which is highly conserved among EGF agonists. This leucine residue appears to be important for high affinity binding of EGF to ErbB-1 (41). In fact, it has been shown that an artificially C terminally extended recombinant AR incorporating 6 additional amino acid residues predicted from its coding sequence, as well as a methionine to leucine substitution at position 86 (AR 1-90(leu86) ) had a 50 -100-fold higher activity than wild-type AR and was comparable to EGF (42). In line with its low affinity for ErbB-1, AR was the only EGF agonist that failed to induce tyrosine phosphorylation of ErbB-2 in T47D cells. This is in contrast to a previous publication showing AR-induced phosphorylation of ErbB-2 in a number of cell lines with relatively high ErbB-1 levels (36). This discrepancy most likely is due the low levels of ErbB-1 expressed in T47D cells (27).
Despite quite high ErbB-3 and ErbB-4 expression levels in T47D cells (27), we failed to observe any NDF-induced increase in ErbB-1 tyrosine phosphorylation. This is in contrast to a recent publication which shows that coexpression of ErbB-1 with either ErbB-3 or ErbB-4 allows NDF to regulate ErbB-1 tyrosine phosphorylation (43). These results strongly suggest that ErbB receptor heterodimer formation follows a strict hierarchy. If only ErbB-1 is available, ErbB-3 and ErbB-4 may be able to heterodimerize with ErbB-1 in response to NDF. However, if all the four ErbB receptors are present, other heterodimers seem to be preferentially formed (i.e. ErbB-3/ErbB-2, ErbB-4/ErbB-2, or ErbB-3/ErbB-4).
EGF has been found to induce ErbB-3 tyrosine phosphorylation and association with PtdIns 3-kinase in ErbB-1 overexpressing A431 and MDA MB 468 cells, presumably through ErbB-1/ErbB-3 heterodimerization (30,31). However, while NDF strongly induced ErbB-3 phosphorylation and association with PtdIns kinase activity in T47D cells, EGF only had a mild effect, probably due to the low abundance of ErbB-1 in these cells (27). Thus, although in ErbB-1 overexpressing tumor cells ErbB-1/ErbB-3 cross-talk may contribute to the transformed phenotype, our observations raise potential doubts about the relevance of this cross-talk at physiological ErbB-1 expression levels. HB-EGF and BTC were significantly more efficacious in inducing ErbB-3 phosphorylation and association with PtdIns kinase activity than the other agonists. Although the mechanism of the enhanced ErbB-1/ErbB-3 cross-talk in response to HB-EGF and BTC is unclear, it seems possible that different ErbB-1 binding peptides preferentially promote the formation of specific heterodimers. In fact, it appears that AR also preferentially induces ErbB-1/ErbB-3 cross-talk, since it induced a low level of ErbB-3 phosphorylation without having an effect on ErbB-2. EGF agonistic growth factors can be subdivided into heparin binding (AR, HB-EGF, and BTC) and non-heparin binding (EGF and TGF␣) peptides. The presence of heparan sulfate-containing proteoglycans on the cell surface and/or in the extracellular matrix has been shown to be essential for activity of several heparin binding growth factors, including AR (44,45). It is likely that the binding of AR, HB-EGF, and BTC to heparan sulfate-containing proteoglycans influences receptor-ligand interactions, and it may be possible that it facilitates the binding of specific receptor dimers, e.g. ErbB-1/ ErbB-3 heterodimers. Although the precise mechanism remains to be elucidated, the different effect of EGF-related peptides on ErbB-3 phosphorylation seems relevant, since it directly correlates with their effect on the soft agar growth of T47D cells.
BTC was shown to bind ErbB-1 with an affinity similar to that of EGF (46). However, BTC-induced tyrosine phosphorylation of ErbB receptors has not been analyzed previously. Here we show that BTC is unique among the EGF-related peptides, since it has the ability to efficiently activate all the four members of the ErbB receptor family. Most importantly, it was as effective as NDF in inducing ErbB-4 phosphorylation. Binding of a specific EGF agonist to ErbB-1 per se may induce ErbB-4 phosphorylation through the formation of ErbB-1/ErbB-4 heterodimers. In fact, this mechanism may account for the slight increase in ErbB-4 tyrosine phosphorylation we observed in response to HB-EGF. However, BTC appears to directly interact with ErbB-4. First, BTC was similar to NDF in that it was more effective than other EGF agonists in stimulating the growth of MDA MB 453 cells, which do not contain detectable amounts of ErbB-1 (39). Second, pretreatment of T47D cells with mAb 225, which competes with EGF for binding to ErbB-1 (32), completely abolished BTC-induced phosphorylation of ErbB-1 while only partially inhibiting the phosphorylation of ErbB-4. In line with these observations, a recent report shows BTC-induced phosphorylation of ErbB-4 ectopically expressed in Ba/F3 cells, which lack detectable levels of endogeneous ErbB receptors (47). However, our results also point to a role for ErbB-1/ErbB-4 heterodimers in BTC-induced signaling, since ErbB-1 appears to be required for full activation of ErbB-4.
The requirements of HC11 mouse mammary epithelial cells for both EGF and insulin for growth under serum-restricted conditions allowed another comparison of the biological activities of the peptides. While all the EGF agonists and NDF were able to replace EGF in the growth medium, none of the EGF agonists was able to replace insulin. Only NDF could partially replace insulin, thereby cooperating with EGF for the growth stimulation of HC11 cells. NDF has a similar activity on the MCF10A human mammary epithelial cells (48). Thus, it seems that NDF has a dual EGF-and insulin-like activity enabling it to stimulate signaling pathways that normally require two separate growth factors. Activated insulin and insulin-like growth factor-1 receptors induce the rapid tyrosine phosphorylation of insulin receptor substrate-1, which contains multiple binding sites for the p85 subunit of PtdIns 3-kinase (49). Similarly, NDF is the best activator of ErbB-3, the only ErbB receptor involved in efficient recruitment of p85 (30,37). Thus, it is possible that the ability to efficiently activate PtdIns 3-kinase is responsible for NDFs insulin-like activity.
In conclusion, this study represents the first comprehensive analysis of all the presently available EGF-related growth factors with respect to the activation of ErbB receptors. The distinct biological activities described in this report provide a rational for the existence of the great variety of EGF-related peptides.