During the last decade, several novel members of the Epidermal Growth Factor family of peptide growth factors have been identified. Most prominent among these are the Neuregulins or Heregulins. To date, four different Neuregulin genes have been identified (Neuregulin1-4) and several different splicing isoforms have been identified for at least two of these genes (Neuregulin1 and Neuregulin2). While Neuregulin1 isoforms have been extensively studied, comparatively little is known about Neuregulin3, Neuregulin4, or the Neuregulin2 isoforms. Indeed, there has been no systematic comparison of the activities of these molecules. Here we demonstrate that Neuregulin2alpha and Neuregulin2beta stimulate ErbB3 tyrosine phosphorylation and coupling to biological responses. In contrast, Neuregulin3 and Neuregulin4 fail to activate ErbB3 signaling. Furthermore, Neuregulin2beta, but not Neuregulin2alpha, stimulates ErbB4 tyrosine phosphorylation and coupling to biological responses. Finally, both Neuregulin3 and Neuregulin4 stimulate modest amounts of ErbB4 tyrosine phosphorylation. However, whereas Neuregulin3 stimulates a modest amount of ErbB4 coupling to biological responses, Neuregulin4 fails to stimulate ErbB4 coupling to biological responses. This suggests that there are qualitative as well as quantitative differences in ErbB family receptor activation by Neuregulin isoforms.
The Epidermal Growth Factor (EGF) family of peptide hormones consists of approximately 20 different proteins encoded by at least 10 different genes (Reviewed in Kumar and Vadlamudi, 2000; Gullick, 2001; Yarden and Sliwkowski, 2001). These peptide growth factors are agonists for the four ErbB family receptors, including the Epidermal Growth Factor Receptor (EGFR), ErbB2 (HER2/Neu), ErbB3 (HER3), and ErbB4 (HER4) (Reviewed in Schlessinger, 2000; Gullick, 2001; Yarden and Sliwkowski, 2001). Deregulated signaling by this network has been implicated in the genesis and progression of several types of human malignancies, including tumors of the breast, ovary, prostate, pancreas, lung, and brain (Reviewed in Stern, 2000; Normanno et al., 2001; Ozawa et al., 2001; Yarden and Sliwkowski, 2001).
During the last decade, several novel members of the EGF family have been identified and characterized. Most notable among these proteins are the Neuregulins (NRGs), also known as the Heregulins (HRGs) or Neu Differentiation Factors (NDFs) (Holmes et al., 1992; Wen et al., 1992; Carraway et al., 1997; Chang et al., 1997; Zhang et al., 1997; Harari et al., 1999). Currently, there are four known Neuregulin genes, NRG1 through NRG4. NRG1 and NRG2 encode multiple splicing isoforms; these are denoted as either alpha or beta isoforms depending on the sequence of the EGF homology domain.
Difficulties in the expression and purification of Neuregulin isoforms have hampered efforts to characterize the functions of these ligands. Nonetheless, several fundamental principles have emerged: (1) both the alpha and beta isoforms of NRG1 are ErbB3 ligands (Kita et al., 1994; Lu et al., 1995; Pinkas-Kramarski et al., 1996; Jones et al., 1999); (2) the NRG1 beta isoform is a higher affinity ligand for ErbB4 than is the NRG1 alpha isoform (Tzahar et al., 1994; Jones et al., 1999); (3) NRG3 and NRG4 are ErbB4 ligands (Zhang et al., 1997; Harari et al., 1999). However, some of these experiments have been performed in vitro using recombinant receptor fragments and synthetic hormones or hormones expressed from bacteria. Other experiments have been performed using a variety of cultured cell lines. Thus, it has been difficult to compare the results that appear in different reports. Indeed, there has been no report of a systematic functional comparison of NRG2 alpha, NRG2 beta, NRG3, and NRG4.
Thus, a careful analysis of the published literature reveals a number of fundamental questions concerning NRG function: (1) Do the alpha and beta isoforms of NRG2 behave similarly to the corresponding NRG1 isoforms? (2) Are NRG3 and NRG4 agonists for ErbB3? (3) Given the large number of ErbB4 agonists among the NRGs, are the different agonists for ErbB4 functionally distinct? In this study we describe a novel method for easily expressing and purifying recombinant, bioactive NRGs. We present data indicating that NRG2 alpha (NRG2α) and NRG2 beta (NRG2β) are functionally distinct. We also present data indicating that NRG3 and NRG4 are ErbB4 agonists but do not appear to be ErbB3 agonists. Finally, we present data indicating that the different NRG ErbB4 agonists cause differential coupling of ErbB4 to biological responses. This is some of the most compelling evidence to date that different direct agonists for the same ErbB family receptor may be functionally distinct.
Recombinant NRGs can be expressed in insect cells
Several groups have expressed recombinant EGF family peptide hormones in E. coli. Advantages of this strategy include yield and suitability of the protein for structural analysis by NMR or X-ray crystallography. One disadvantage of this strategy is that the purification and refolding strategies may be cumbersome. Another is that the proteins lack the glycosylation present in proteins expressed in eukaryotic cells. Other groups have generated synthetic EGF family peptide hormones. A significant disadvantage of this strategy is the expense. Thus, we sought to produce recombinant NRG2α, NRG2β, NRG3, and NRG4 using an insect cell expression system (Invitrogen). The advantages of this system are that the yield is reasonable (∼300 μg purified protein/liter of insect cell culture), the expense is modest (∼$200/mg purified protein), the purification strategy is straightforward, there is no need to refold the protein following purification, and the protein is glycosylated.
We began by subcloning a portion of the NRG cDNAs into the insect cell vector pMT-BiP-V5HisB (Invitrogen). Others and we have previously reported the cloning of the NRG2α and NRG2β cDNAs (Carraway et al., 1997; Chang et al., 1997). We isolated the NRG3 cDNA (Zhang et al., 1997) from a human cDNA library and we isolated the NRG4 cDNA (Harari et al., 1999) from a mouse cDNA library. The regions of these cDNAs encoding the EGF homology domain and surrounding sequences (NRG2α: Ser247 to Asp328; NRG2β: Ser247 to Lys314; NRG3: Ser284 to Gln360; NRG4: Thr3 to Asn60) were subcloned into the conditional insect cell expression vector pMT-BiP-V5HisB (Invitrogen). The inserts were cloned in frame with the vector sequences that encode the upstream BiP secretion signal and the downstream V5 and polyhistidine epitope tags (Figure 1a). The predicted sequences of the recombinant NRGs encoded by these expression constructs are shown in Figure 1b. The amino acid sequences of the NRG3 and NRG4 regions are identical to those reported in the literature (Zhang et al., 1997; Harari et al., 1999).
We cotransfected the S2 Schneider insect cell line (American Type Culture Collection) with the NRG constructs and pCoHygro, a plasmid that directs the expression of the hygromycin resistance gene (Invitrogen). Transfected S2 cells were selected using hygromycin and were pooled to generate stable cell lines. A one liter culture of each cell line was expanded to a density of 107 cells/ml and resuspended in serum-free insect cell medium (Gibco/BRL/Life Technologies) supplemented with 1 mM CuSO4 to induce recombinant NRG expression from the pMT-BiP-V5His constructs. The recombinant NRGs were purified and concentrated by ultrafiltration, dialysis, and chromatography using ProBond Ni2+ beads.
We quantified the absolute concentrations of the NRG preparations by immunoblotting using an anti-V5 antibody (Invitrogen). A 53-kDa positive control peptide (Positope–Invitrogen) was used as the standard (Figure 2). Each NRG appeared as a heterogeneous mixture of at least three isoforms with distinct mobilities. Overall, the apparent molecular weights of the NRG isoforms is a little less than those predicted from the amino acid sequences (NRG2α – 12 500 Da; NRG2β – 10 900 Da; NRG3 – 12 200 Da; NRG4 – 9700 Da); however, the relative apparent molecular weights are in agreement with those predicted from the amino acid sequences. (NRG2α and NRG3 have higher apparent molecular weights than NRG2β and NRG2β has a higher apparent molecular weight than NRG4). Finally, the multiple isoforms of each NRG were resolved to a single, tightly focused band by treatment with peptide N-glycosidase F (data not shown). This suggests that these isoforms represent differentially glycosylated species.
We digitized the immunoblots and quantified the bands for each NRG. We quantified all of the bands for those NRGs that exhibited multiple isoforms. We used these values to construct a dose response curve of best fit for each NRG. These curves were used in conjunction with the dose response curve of best fit for the Positope control to calculate the concentration of each NRG stock. We also quantified the relative concentrations of the NRG preparations by ELISA using an anti-V5 antibody (Invitrogen) and the ABC ELISA kit (Pierce). Recombinant NRG yields were typically 300 μg from a one liter culture of insect cells.
Recombinant NRGs differentially stimulate ErbB3 tyrosine phosphorylation
We assessed the interactions of the recombinant NRGs with ErbB family receptors by first assaying induction of ErbB3 tyrosine phosphorylation by the NRG isoforms. ErbB3 lacks tyrosine kinase activity and ErbB2 is an orphan receptor for which there is no known ligand. Consequently, we assayed ligand induction of ErbB3 tyrosine phosphorylation in mouse BaF3 lymphoid cells (which lack endogenous EGFR, ErbB2, and ErbB4 expression) that we had engineered to express ErbB2 and ErbB3 (BaF3/ErbB2+ErbB3) (Riese et al., 1995). The recombinant NRG1β positive control (EGF homology domain; R&D Systems) stimulates abundant ErbB2 and ErbB3 tyrosine phosphorylation (Figure 3). Both NRG2α and NRG2β stimulate more modest amounts of ErbB3 tyrosine phosphorylation, nonetheless indicating that these growth factors are ligands for ErbB3. In contrast, neither NRG3 nor NRG4 stimulate detectable ErbB3 tyrosine phosphorylation.
Recombinant NRGs differentially stimulate ErbB4 tyrosine phosphorylation
We assayed induction of ErbB4 tyrosine phosphorylation by the NRG isoforms using a human CEM lymphoid cell line (which lacks endogenous ErbB receptor expression) engineered to express ErbB4 (Plowman et al., 1993). The NRG1β positive control and NRG2β stimulate abundant ErbB4 tyrosine phosphorylation, whereas NRG4 stimulates a moderate amount of ErbB4 tyrosine phosphorylation and NRG3 stimulates a modest amount of ErbB4 tyrosine phosphorylation (Figure 4). NRG2α fails to stimulate any detectable ErbB4 tyrosine phosphorylation (Figure 4).
Increasing NRG2α concentrations fail to stimulate ErbB4 tyrosine phosphorylation (Figure 4). We were concerned that the failure of NRG2α to stimulate ErbB4 tyrosine phosphorylation was due to a relatively modest difference in the affinities of NRG2α and NRG2β for ErbB4. Consequently, we stimulated CEM/ErbB4 cells with greater concentrations of NRG2α. In Figure 5 we show that 1000 ng/ml NRG2α stimulates little ErbB4 tyrosine phosphorylation. In contrast, ErbB4 tyrosine phosphorylation reaches saturation at a NRG2β concentration of approximately 30 ng/mL and a NRG3 and NRG4 concentration of approximately 300 ng/ml. Thus, the dissociation constant (Kd) of NRG3 and NRG4 for ErbB4 appears to be approximately 10 times greater than the Kd of NRG2β for ErbB4. Furthermore, if the failure of NRG2α to stimulate abundant ErbB4 tyrosine phosphorylation is due to the decreased affinity of NRG2α for ErbB4, the Kd of NRG2α for ErbB4 must be more than 30 times greater than the Kd of NRG2β for ErbB4.
Recombinant NRGs differentially stimulate ErbB family receptor coupling to biological responses
We assayed induction of ErbB3 coupling to biological responses in the BaF3/ErbB2+ErbB3 cell line and the BaF3/EGFR+ErbB4 cell line. BaF3 cells are dependent upon interleukin-3 (IL3) for survival and proliferation. However, we have previously shown that ligands for ErbB3 induce IL3-independent survival, but not proliferation, in BaF3/ErbB2+ErbB3 cells (Riese et al., 1995). Furthermore, we have previously shown that ligands for EGFR or ErbB4 induce IL3-independent proliferation in BaF3/EGFR+ErbB4 cells (Riese et al., 1995, 1996a).
Here we demonstrate that NRG2α and NRG2β, as well as the NRG1β positive control, induce IL3-independent survival in BaF3/ErbB2+ErbB3 cells (Figure 6). Furthermore, NRG3 fails to induce IL3 independence in BaF3/ErbB2+ErbB3 cells and NRG4 induces minimal IL3 independence in these cells. These results are largely consistent with the ErbB3 tyrosine phosphorylation data that suggest that NRG2α and NRG2β are ligands for ErbB3, whereas NRG3 and NRG4 are not ligands for ErbB3 (Figure 3).
We also demonstrate that NRG2β and the NRG1β positive control induce IL3-independent proliferation in BaF3/EGFR+ErbB4 cells (Figure 6). In contrast, NRG3 induces IL3-independent survival (not proliferation) in BaF3/EGFR+ErbB4 cells and both NRG2α and NRG4 induce minimal IL3 independence in these cells. The results of the IL3 independence assays indicate that NRG2β, but not NRG2α, is an ErbB4 agonist, which is in line with the ErbB4 tyrosine phosphorylation data with NRG2α and NRG2β (Figures 4 and 5).
NRG3 and NRG4 fail to stimulate ErbB4 tyrosine phosphorylation in BaF3/EGFR+ErbB4 cell lines
Despite the fact that 100 ng/ml NRG3 or NRG4 stimulates ErbB4 tyrosine phosphorylation (Figures 4 and 5), 100 ng/ml NRG3 or NRG4 fail to stimulate ErbB4 coupling to biological responses to the extent that NRG2β does (Figure 6). Furthermore, despite the fact that identical concentrations of NRG3 and NRG4 stimulate similar levels of ErbB4 tyrosine phosphorylation (Figure 5), NRG3 stimulates a greater level of IL3 independence in the BaF3/EGFR+ErbB4 cell line than does NRG4 (Figure 6). In an attempt to resolve these discrepancies, we stimulated BaF3/EGFR+ErbB4 cells with the various NRG isoforms and assayed both EGFR and ErbB4 tyrosine phosphorylation by receptor immunoprecipitation and antiphosphotyrosine immunoblotting. In Figure 7 we show that 100 ng/ml NRG1β or NRG2β stimulates EGFR and ErbB4 tyrosine phosphorylation, but 100 ng/ml NRG2α, NRG3, or NRG4 does not. Indeed, even 1000 ng/ml NRG3 or NRG4 does not stimulate EGFR or ErbB4 tyrosine phosphorylation (Figure 8). These data are consistent with the IL3 independence data (Figure 6) and suggest that EGFR inhibits stimulation of ErbB4 tyrosine phosphorylation and coupling to downstream signaling events by NRG3 and NRG4.
In this study we demonstrate that recombinant NRGs can be expressed in insect cells and that these molecules retain biological and biochemical activities. This is a significant advance since methods traditionally used to generate EGF family peptide hormones are cumbersome or expensive. Indeed, this methodology will facilitate functional analyses of NRGs by site-directed mutagenesis. While this strategy has been used to analyse the function of some EGF family hormones, most notably EGF itself (Reviewed in Groenen et al., 1994; Boonstra et al., 1995), such analysis of NRGs have been limited to binding studies done using NRGs expressed in phage display systems (Jones et al., 1998; Ballinger et al., 1999). Undoubtedly, studies facilitated by the ready availability of NRG mutants will reveal new insights into the nature of the interactions between EGF family peptide growth factors and their cognate ErbB family receptor tyrosine kinases.
The studies presented here also represent the initial systematic functional comparison of NRG2α, NRG2β, NRG3, and NRG4. Here we show that NRG2α and NRG2β stimulate ErbB3 tyrosine phosphorylation (in the context of ErbB2 and ErbB3 coexpression), whereas NRG3 and NRG4 do not. These results are consistent with the published observation that NRG alpha and beta isoforms are ligands for ErbB3 (Tzahar et al., 1994; Pinkas-Kramarski et al., 1996, 1998, 1999; Jones et al., 1999). However, these results contrast the observation that a recombinant NRG2α fusion protein fails to compete with radiolabeled NRG1β for binding to recombinant ErbB2-ErbB3 heterodimers (Jones et al., 1999). Of course the physiologic relevance of preformed recombinant ErbB2-ErbB3 heterodimers is unclear and it was noted by the authors that the NRG fusion proteins have reduced affinity for their native receptors (Jones et al., 1998). Our results are consistent with the published observation that a recombinant NRG3 fusion protein fails to compete with radiolabeled NRG1β for binding to a recombinant ErbB3 fusion protein (Jones et al., 1999). However, the observation that NRG3 stimulates ErbB2 and ErbB3 tyrosine phosphorylation in 32D cells devoid of endogenous ErbB family receptors (Hijazi et al., 1998) contrasts our results. Of course, coexpression of ErbB2 and ErbB3 in 32D cells permits EGF stimulation of receptor coupling to IL3 independence and mitogenesis (Pinkas-Kramarski et al., 1998). This calls into question the utility of the 32D model system for defining ligand-receptor interactions. Regardless, we conclude that NRG2α and NRG2β are functionally distinct from NRG3 and NRG4.
We also show that NRG2β is a potent agonist of ErbB4 tyrosine phosphorylation, whereas NRG3 and NRG4 stimulate modest levels of ErbB4 tyrosine phosphorylation and NRG2α fails to stimulate ErbB4 tyrosine phosphorylation (Figures 4 and 5). These results are consistent with the observation that NRG beta isoforms are more potent and higher affinity ligands for ErbB4 than are NRG alpha isoforms (Tzahar et al., 1994; Lu et al., 1995; Pinkas-Kramarski et al., 1998, 1999; Jones et al., 1999). These data are also consistent with the observation that NRG3 and NRG4 are both ErbB4 ligands (Zhang et al., 1997; Harari et al., 1999; Jones et al., 1999). However, these data also suggest that NRG2β is a more potent ligand for ErbB4 than are NRG4 and NRG3. It should be noted that ErbB4 tyrosine phosphorylation reaches saturation following stimulation with 30 ng/ml NRG2β, 300 ng/ml NRG3, or 300 ng/ml NRG4. Thus, some of the functional difference between NRG2β and NRG3 or NRG4 appears to be due to the higher affinity of NRG2β for ErbB4. Indeed, the affinity of NRG3 for ErbB4 is reported to be less than one-tenth the affinity of NRG2β for ErbB4 (Jones et al., 1999). Similarly, the affinity of NRG4 for ErbB4 is reported to be approximately one-tenth the affinity of NRG1β for ErbB4 (Harari et al., 1999).
NRG2α and NRG2β stimulate IL3 independent survival in BaF3/ErbB2+ErbB3 cells, whereas NRG3 and NRG4 do not stimulate IL3 independence in these cells (Figure 6). These results are consistent with the observation that NRG alpha and beta isoforms stimulate coupling of ErbB2 and ErbB3 to biological responses (Pinkas-Kramarski et al., 1996, 1998, 1999). These results are also consistent with the tyrosine phosphorylation data shown in Figure 3. NRG3 does not stimulate any IL3 independence in the BaF3/ErbB2+ErbB3 cells (Figure 6), consistent with the tyrosine phosphorylation data shown in Figure 3. NRG4 also fails to stimulate ErbB3 tyrosine phosphorylation in the BaF3/ErbB2+ErbB3 cells (Figure 3). However, NRG4 stimulates a modest amount of IL3 independence in these cells (Figure 6). It is possible that the IL3 independence assay is a more sensitive measure of ligand-induced receptor signaling than is antiphosphotyrosine immunoblotting. Indeed, we have previously shown that the ligand concentration required for saturated levels of ErbB receptor tyrosine phosphorylation in BaF3 cells is approximately 10-fold greater than the ligand concentration required for saturated levels of IL3 independence in the same cell lines (Riese et al., 1995).
NRG2β stimulates IL3-independent proliferation in BaF3/EGFR+ErbB4 cell lines, whereas NRG2α stimulates minimal IL3 independence (Figure 6). This is consistent with the tyrosine phosphorylation data shown in Figures 4 and 5. More intriguing are the observations that 100 ng/ml NRG3 stimulates only IL3-independent survival and that 100 ng/ml NRG4 stimulates minimal IL3 independence (Figure 6). We were concerned that we were not using a sufficient concentration of NRG3 or NRG4 in these IL3 independence assays. However, even 1000 ng/ml NRG3 or NRG4 failed to stimulate IL3-independent proliferation in the BaF3/EGFR+ErbB4 cells (data not shown). Thus, we attempted to explain these results by assaying ligand-induced receptor tyrosine phosphorylation in the BaF3/EGFR+ErbB4 cell (Figures 7 and 8). These experiments reveal that NRG2α, NRG3, and NRG4 stimulate minimal receptor tyrosine phosphorylation in the BaF3/EGFR+ErbB4 cells. This is consistent with the relative inactivity of these ligands in the IL3 independence assay using these cells. Furthermore, the high basal (ligand-independent) level of receptor tyrosine phosphorylation in these cells (Figures 7 and 8) may account for the small amount of IL3-independence stimulated by NRG2α (which is presumably not a potent ErbB4 agonist).
We are left trying to explain why NRG3 and NRG4 stimulate much lower levels of ErbB4 tyrosine phosphorylation (Figures 7 and 8) and ErbB receptor coupling to biological responses in the BaF3/EGFR+ErbB4 cells (Figure 6) than would be expected from the ErbB4 tyrosine phosphorylation data obtained from the CEM/ErbB4 cells (Figures 4 and 5). A simple, non-mechanistic explanation is that EGFR inhibits ligand-induced ErbB4 tyrosine phosphorylation. However, we have previously shown that ErbB2 or ErbB3 expression does not quantitatively modulate ErbB4 tyrosine phosphorylation stimulated by betacellulin or NRG1β (Feroz et al., 2002). Nonetheless, it is possible that inhibition of ligand-induced ErbB4 signaling is specific for EGFR, NRG3, or NRG4.
A more attractive, mechanistic explanation is that the EGFR-ErbB4 heterodimers stimulated by NRG3 and NRG4 treatment are in a different conformation that results in less receptor tyrosine phosphorylation than the EGFR-ErbB4 heterodimers stimulated by NRG2β. There are precedents for differential receptor tyrosine kinase dimerization, tyrosine phosphorylation, and coupling to downstream events. The bovine papillomavirus (BPV) E5 protein is a membrane-bound agonist for the platelet derived growth factor receptor (PDGFR) and stimulates PDGFR dimerization, tyrosine phosphorylation, and PDGFR-dependent malignant growth transformation of fibroblasts (Reviewed in Drummond-Barbosa and Dimaio, 1997). However, there are BPV E5 mutants that stimulate PDGFR tyrosine phosphorylation yet fail to couple to PDGFR-dependent growth transformation (Nilson et al., 1995; Klein et al., 1998). Similarly, mutation of different ErbB2 extracellular juxtamembrane amino acids residues to cysteine results in ErbB2 disulfide-linked dimers that exhibit high levels of ErbB2 tyrosine phosphorylation yet fail to cause malignant growth transformation of fibroblasts (Burke and Stern, 1998).
Thus, we hypothesize that in BaF3/EGFR+ErbB4 cells, NRG3 and NRG4 stimulate EGFR and ErbB4 phosphorylation on different or a smaller number of tyrosine residues than does NRG2β. These hypotheses are consistent with several published observations. In the MDA-MB-453 breast tumor cell line, NRG1β and NRG2β stimulate ErbB2 and ErbB3 tyrosine phosphorylation to similar extents, but only NRG1β causes differentiation of these cells and the two growth factors cause differential recruitment of SH2 domain-containing proteins to the phosphorylated ErbB2 and ErbB3 and differential activation of gene transcription (Sweeney-Crovello et al., 1998; Sweeney et al., 2001). Similarly, betacellulin, NRG1β, NRG2β, and NRG3 induce qualitatively different patterns of ErbB4 tyrosine phosphorylation, as revealed by 2-dimensional peptide mapping (Sweeney et al., 2000).
Our data suggest that regulation of ErbB family receptor signaling by EGF family hormones occurs at multiple levels. NRG2α and NRG2β are more potent ErbB3 agonists than are NRG3 and NRG4 (Figures 3 and 6). Furthermore, NRG2β is a potent ErbB4 agonist, whereas NRG3 and NRG4 are less potent ErbB4 agonists and NRG2α directly stimulates minimal ErbB4 signaling (Figures 4,5,6,7,8). To a first approximation, these differences in activity of the various NRGs and other EGF family hormones (Figure 9) reflect the different affinities of these hormones for ErbB family receptors (Jones et al., 1999; Reviewed in Riese and Stern, 1998; Kumar and Vadlamudi 2000; Gullick, 2001; Yarden and Sliwkowski, 2001). Consequently, differential ligand activation of signaling by a specific ErbB family receptor is in part a function of quantitative differences in the affinities of the various ligands for the particular receptor.
However, this quantitative model cannot explain all of the data presented here. The dissociation constant (Kd) of NRG2α and ErbB4 is reported to be 20–50-fold greater than the Kd of NRG2β and ErbB4 (Jones et al., 1999; Pinkas-Kramarski et al., 1998, 1999). Yet, whereas 3 ng/ml NRG2β stimulates a modest amount of ErbB4 tyrosine phosphorylation, 300 ng/ml NRG2α fails to stimulate ErbB4 tyrosine phosphorylation and 1000 ng/ml NRG2α stimulates only a very small amount of ErbB4 tyrosine phosphorylation (Figures 4 and 5). Furthermore, whereas NRG3 and NRG4 stimulate detectable amounts of ErbB4 tyrosine phosphorylation in CEM/ErbB4 cells, at the same ligand concentrations they fail to stimulate detectable amounts of ErbB4 tyrosine phosphorylation in BaF3/EGFR+ErbB4 cells. In contrast, NRG2β stimulates abundant ErbB4 tyrosine phosphorylation in both cell lines. These observations suggest that there are qualitative differences in activation of ErbB4 signaling and coupling to downstream signaling events by the various NRG isoforms. Published data suggest that these qualitative differences between the NRG isoforms reflect ligand-induced ErbB4 tyrosine phosphorylation on different tyrosine residues and consequent differential receptor coupling to downstream signaling pathways. This would explain the functional differences of the NRG isoforms seen in this study. One of our future challenges will be to formally test whether there are qualitative differences in the activities of the ErbB4 ligands and to identify the mechanism for these differences. Another challenge will be to develop a model that explains the interactions of EGF family hormones with ErbB family receptors and that accounts for these qualitative differences in the activities of the ErbB4 ligands.
Materials and methods
Cell lines and cell culture
The S2 Schneider insect cells were purchased from the American Type Culture Collection. The CEM/ErbB4 cells (Plowman et al., 1993) are a generous gift from Dr Gregory D Plowman, Exelixis Pharmaceuticals. The BaF3/ErbB2+ErbB3 and BaF3/EGFR+ErbB4 cell lines have been described previously (Riese et al., 1995). All cell lines were maintained according to vendor instructions or published procedures (Plowman et al., 1993; Riese et al., 1995; Feroz et al., 2002).
Plasmids and plasmid construction
The insect cell conditional expression vector pMT-BiP-V5HisB and the pCoHygro plasmid were purchased from Invitrogen. We isolated NRG2α, NRG2β, NRG3, and NRG4 clones from human, rat, and mouse cDNA libraries. The regions of the cDNA clones that encode the EGF homology domain of the NRG isoforms were amplified by PCR and were subcloned by standard molecular biology techniques into the BglII and SacII sites of pMT-BiP-V5HisB. The upstream primer used to amplify the rat NRG2α sequences has the following sequence: 5′-CTCGAGAGATCTTCGGGGCACGCCCGGAAGTG-3′. The downstream primer has the following sequence: 5′-CTCGAGCCGCGGATTCAAATCCAAGGTGCTTGG-3′. The amplified sequences encode Ser247 to Asp328 (Carraway et al., 1997; Chang et al., 1997). The upstream primer used to amplify the rat NRG2β sequences has the following sequence: 5′-CTCGAGAGATCTTCGGGGCACGCCCGGAAGTG-3′. The downstream primer has the following sequence: 5′-CTCGAGCCGCGGCTTCTGGTACAGCTCCTC-3′. The amplified sequences encode Ser247 to Lys314 (Carraway et al., 1997; Chang et al., 1997). The upstream primer used to amplify the human NRG3 sequences has the following sequence: 5′-CTCGAGAGATCTTCCGAGCACTTCAAACCCTG-3′. The downstream primer has the following sequence: 5′-CTCGAGCCGCGGCTGCCTTTGATAAACTTCTTCACTCTCC-3′. The amplified sequences encode Ser284 to Gln360 (Zhang et al., 1997). The upstream primer used to amplify the mouse NRG4 sequences has the following sequence: 5′-CTCGAGAGATCTACAGATCACGAGCAGCC-3′. The downstream primer has the following sequence: 5′-CTCGAGCCGCGGATTACTTTCGCTTGGGATGCTGG-3′. The amplified sequences encode Thr3 to Asn60 (Harari et al., 1999). The inserts were subcloned in frame with the upstream BiP secretion signal encoded by pMT-BiP-V5HisB and in frame with the downstream V5 and polyhistidine epitope tags encoded by pMT-BiP-V5HisB.
Generation and purification of recombinant NRGs
The NRG clones were co-transfected into the S2 cells along with the plasmid pCoHygro, which carries the hygromycin resistance gene. Transfections were performed using a calcium phosphate transfection kit (Invitrogen) according to vendor instructions. Transfected cells were selected using 300 U/ml hygromycin B (Cellgro) and stably transfected cells appeared approximately 14 days after the beginning of selection.
Hygromycin-resistant cells were pooled, expanded, and frozen for archival purposes. Transfected cells were seeded in a one liter culture at a density of 2×106 cells/ml. Cells were maintained until they reached a density of 1×107 cells/ml. At that point cells were collected by centrifugation and seeded at a density of 2×107 cells/ml in serum-free medium (Gibco/BRL/Life Technologies) supplemented with 1 mM CuSO4. Cells were maintained for 5 days in serum-free medium to permit recombinant NRG expression and secretion into the culture medium.
The insect cells were collected from the culture media by centrifugation. The conditioned media supernatants were transferred into a fresh container and clarified by filtration through a 0.22 μM filter. The NRGs present in the conditioned medium were concentrated approximately 30-fold by ultrafiltration using a 5000 M.W.C.O. filter (Amicon). The concentrated NRGs were dialyzed against PBS using a 5000 M.W.C.O. membrane (Pierce) to remove low-molecular weight impurities. The NRGs were purified by incubating the samples with ProBond Ni2+ beads (Invitrogen), which bind proteins tagged with polyhistidine. The NRGs were eluted from the beads using 500 mM imidazole. We removed the imidazole from the eluates by dialysis against PBS using a 500 M.W.C.O. membrane (Pierce). The dialyzed proteins were then concentrated to a final volume of 2–5 ml by ultrafiltration using a 5000 M.W.C.O. filter (Amicon).
Anti-V5 immunoblotting was used to quantify the concentrations of the recombinant NRG samples. Samples were resolved by SDS–PAGE using a 20% acrylamide gel. Resolved samples were electroblotted onto nitrocellulose. The blots were probed using an anti-V5 mouse monoclonal antibody (Invitrogen). Primary antibody binding was detected using a goat anti-mouse antibody conjugated to horseradish peroxidase (Pierce). Secondary antibody binding was visualized by chemiluminescence (Amersham). The positope recombinant protein (Invitrogen) was analysed in parallel as a control for V5 immunoblotting and as a standard for quantification.
The resulting immunoblot was digitized using a UMAX Astra 2400S flatbed scanner and the image was cropped using Adobe Photoshop. The bands were quantified using NIH Image for Macintosh software. We generated a dose-response line of best fit for each recombinant NRG using Microsoft Excel. The coefficients of correlation exceeded 0.96. These curves were used to calculate the concentration of each recombinant NRG stock.
The concentration of the NRG2α, NRG3, and NRG4 preparations were determined relative to the concentration of the NRG2β preparation by ELISA using an anti-V5 monoclonal antibody (Invitrogen) and the ABC ELISA kit (Pierce). Polyvinyl chloride (PVC) 96-well assay plates were seeded with 1, 3, and 10 ng/well of NRG2β and 3, 10, and 30 μl/well of several dilutions of the other NRGs in a total volume of 100 μl/well. The plates were incubated for 1 h at room temperature to allow for protein binding to the wells. The wells were then washed three times with 200 μl tris-buffered saline supplemented with 0.05% Tween-20 (TBS-T). Non-specific binding of the antibody to the wells was blocked by incubating the wells for 1 h at room temperature with 100 μl TBS/1% bovine serum albumin (Sigma). Next, 100 μl of the mouse-anti-V5 monoclonal antibody (0.2 μg/ml – Invitrogen) was added to each well and the plates were incubated for 30 min at room temperature. The wells were then washed three times with 200 μl TBS-T and 100 μl of a biotinylated anti-mouse antibody (1.5 μg/ml – Pierce) was added to each well. The plates were incubated for 30 min at room temperature. The wells were washed three times with 200 μl TBS-T. An avidin/biotinylated alkaline phosphatase complex (100 μl) was added to each well and the plates were incubated for 30 min at room temperature. The wells were washed three times with 200 μl TBS-T, after which 100 μl of TBS-T was added to each well and the plates were incubated for 5 min at room temperature. The TBS-T was removed and 100 μl of the alkaline phosphatase substrate, p-nitrophenyl phosphate (1 mg/ml solution dissolved in diethanolamine – Pierce), was added to each well. The plates were incubated until the appropriate amount of substrate had been dephosphorylated, which is evident from the yellow color of the product. The reactions were terminated by adding 25 μl 2 M NaOH to each well. Finally, the amount of product in each well was determined by measuring absorbance at 405 nm using a SpectrFluor Plus plate reader (Tecan).
The amount of product was plotted as a function of sample stock volume for NRG2α, NRG3, and NRG4. These dose-response curves were compared to a standard dose-response curve generated using NRG2β to determine the relative concentration of the NRG2α, NRG3, and NRG4 stocks.
Stimulation and analysis of ErbB family receptor tyrosine phosphorylation
We analysed ligand-induced ErbB family receptor tyrosine phosphorylation in CEM/ErbB4, BaF3/ErbB2+ErbB3, and BaF3/EGFR+ErbB4 cells using procedures published previously (Riese et al., 1995, 1996a,b, 1998; Chang et al., 1997; Feroz et al., 2002). Briefly, approximately 107 cells were stimulated for 7 min on ice with ligand, after which the cells were lysed in an isotonic lysis buffer supplemented with the nonionic detergent NP40. Nuclei and debris were collected from the lysates by centrifugation and the supernatants were transferred to a fresh tube. The protein content of the lysates was analysed using a modified Bradford assay (Pierce). ErbB family receptors were precipitated from the lysates using Concanavlin A-sepharose, which binds to glycoproteins. ErbB family receptors were also precipitated from the lysates using an anti-EGFR mouse monoclonal antibody (Santa Cruz Biotechnology), or an anti-ErbB4 rabbit polyclonal antibody (Santa Cruz Biotechnology).
The precipitates were resolved by SDS–PAGE using a 7.5% acrylamide gel. The resolved samples were electroblotted onto nitrocellulose. The blots were probed using an anti-phosphotyrosine mouse monoclonal antibody (Upstate Biotechnology). Primary antibody binding was detected using a goat anti-mouse antibody conjugated to horseradish peroxidase (Pierce). Secondary antibody binding was visualized by chemiluminescence (Amersham).
Stimulation and analysis of ErbB family receptor coupling to IL3 independence
We analysed ligand-induced ErbB family receptor coupling to IL3 independence in BaF3/ErbB2+ErbB3 and BaF3/EGFR+ErbB4 cells using procedures published previously (Riese et al., 1995, 1996a,b, 1998; Chang et al., 1997). Briefly, cells were seeded in 24-well dishes at a density of 105 cells/ml in medium lacking interleukin3 (IL3), in medium supplemented with IL3, or in medium lacking IL3 but supplemented with a recombinant NRG. Cells were incubated for 96 h, after which viable cells were counted using a hemacytometer. If the viable cell density was greater than 105 cells/ml, the cells were judged to be proliferating. If the viable cell density was between 104 and 105 cells/ml, the cells were judged to be surviving. If the viable cell density was below 104 cells/ml, the cells were judged to be dying.
Ballinger MD, Jones JT, Lofgren JA, Fairbrother WJ, Akita RW, Sliwkowski MX, Wells JA . 1998 J. Biol. Chem. 273: 11675–11684
Boonstra J, Rijken P, Humbel B, Cremers F, Verkeij A, van Bergen en Henegouwen P . 1995 Cell Biol. Intl. 19: 413–430
Burke CL, Stern DF . 1998 Mol. Cell. Biol. 18: 5371–5379
Carraway III KL, Weber JL, Unger MJ, Ledesma J, Yu N, Gassmann M, Lai C . 1997 Nature 387: 512–516
Chang H, Riese II DJ, Gilbert W, Stern DF, McMahan UJ . 1997 Nature 387: 509–512
Drummond-Barbosa D, DiMaio D . 1997 Biochim. Biophys. Acta 1332: M1–M17
Feroz K, Williams E, Riese II DJ . 2002 Cell. Signal. 14: 793–798
Groenen LC, Nice EC, Burgess AW . 1994 Growth Factors 11: 235–237
Gullick WJ . 2001 Endocrine-Related Cancer 8: 75–82
Harari D, Tzahar E, Romano J, Shelly M, Pierce JH, Andrews GC, Yarden Y . 1999 Oncogene 18: 2681–2689
Hijazi MM, Young PE, Dougherty MK, Bressette DS, Cao TT, Pierce JH, Wong LM, Alimandi M, King CR . 1998 Int. J. Oncol. 13: 1061–1067
Holmes WE, Sliwkowski MX, Akita RW, Henzel WJ, Lee J, Park JW, Yansura D, Abadi N, Raab H, Lewis GD, Shepard HM, Kuang W-J, Wood WI, Goeddel DV, Vandlen RL . 1992 Science 256: 1205–1210
Jones JT, Akita RW, Sliwkowski MX . 1999 FEBS Lett. 447: 227–231
Jones JT, Ballinger MD, Pisacane PI, Lofgren JA, Fitzpatrick VD, Fairbrother WJ, Wells JA, Sliwkowski MX . 1998 J. Biol. Chem. 273: 11667–11674
Kita YA, Barff J, Luo Y, Wen D, Brankow D, Hu S, Liu N, Prigent SA, Gullick WJ, Nicolson M . 1994 FEBS Lett. 349: 139–143
Klein O, Polack GW, Surti T, Kegler-Ebo D, Smith SO, DiMaio D . 1998 J. Virol. 72: 8921–8932
Kumar R, Vadlamudi RK . 2000 J. Clin. Ligand Assay 23: 233–237
Lu HS, Chang D, Philo JS, Zhang K, Narhi LO, Liu N, Zhang M, Sun J, Wen J, Yanagihara D, Karunakaran D, Yarden Y, Ratzkin B . 1995 J. Biol. Chem. 270: 4784–4791
Nilson LA, Gottlieb RL, Polack GW, DiMaio D . 1995 J. Virol. 69: 5869–5874
Normanno N, Bianco C, DeLuca A, Salomon DS . 2001 Frontiers Biosci. 6: D685–D707
Ozawa F, Friess H, Tempia-Caliera A, Kleeff J, Buchler MW . 2001 Teratogen. Carcinogen. Mutagen. 21: 27–44
Pinkas-Kramarski R, Shelly M, Glathe S, Ratzkin BJ, Yarden Y . 1996 J. Biol. Chem. 271: 19029–19032
Pinkas-Kramarski R, Shelly M, Guarino BC, Wang LM, Lyass L, Alroy I, Alamandi M, Kuo A, Moyer JD, Lavi S, Eisenstein M, Ratzkin BJ, Seger R, Bacus SS, Pierce JH, Andrews GC, Yarden Y . 1998 Mol. Cell. Biol. 18: 6090–6091
Pinkas-Kramarski R, Shelly M, Guarino BC, Wang LM, Lyass L, Alroy I, Alamandi M, Kuo A, Moyer JD, Lavi S, Eisenstein M, Ratzkin BJ, Seger R, Bacus SS, Pierce JH, Andrews GC, Yarden Y . 1999 Mol. Cell. Biol. 19: 8695
Plowman GD, Green JM, Culouscou J-M, Carlton GW, Rothwell VM, Buckley S . 1993 Nature 366: 473–475
Riese II DJ, Bermingham Y, van Raaij TM, Buckley S, Plowman GD, Stern DF . 1996a Oncogene 12: 345–353
Riese II DJ, Kim ED, Elenius K, Buckley S, Klagsbrun M, Plowman GD, Stern DF . 1996b J. Biol. Chem. 271: 20047–20052
Riese II DJ, Komurasaki T, Plowman GD, Stern DF . 1998 J. Biol. Chem. 273: 11288–11294
Riese II DJ, Stern DF . 1998 Bioessays 20: 41–48
Riese II DJ, van Raaij TM, Plowman GD, Andrews GC, Stern DF . 1995 Mol. Cell. Biol. 15: 5770–5776
Schlessinger J . 2000 Cell 103: 211–225
Stern DF . 2000 Breast Cancer Res. 2: 176–183
Sweeney C, Fambrough D, Huard C, Diamonti AJ, Lander ES, Cantley LC, Carraway III KL . 2001 J. Biol. Chem. 276: 22685–22698
Sweeney C, Lai C, Riese II DJ, Diamonti AJ, Cantley LC, Carraway III KL . 2000 J. Biol. Chem. 275: 19803–19807
Sweeney-Crovello C, Lai C, Cantley LC, Carraway III KL . 1998 J. Biol. Chem. 273: 26954–26961
Tzahar E, Levkowitz G, Karungaran D, Yi L, Peles E, Lavi S, Chang D, Liu N, Yayon A, Wen D, Yarden Y . 1994 J. Biol. Chem. 269: 25226–25233
Wen D, Peles E, Cupples R, Suggs SV, Bacus SS, Luo Y, Trail G, Hu S, Silbiger SM, Ben Levy R, Koski RA, Lu HS, Yarden Y . 1992 Cell 69: 559–572
Yarden Y, Sliwkowski MX . 2001 Nature Revs. Mol. Cell. Biol. 2: 127–137
Zhang D, Sliwkowski MX, Mark M, Frantz G, Akita R, Sun Y, Hillan K, Crowley C, Brush J, Godowski PJ . 1997 Proc. Natl. Acad. Sci. USA 94: 9562–9567
SS Hobbs was supported by an NIH predoctoral training grant (T32GM008737). EM Cameron was supported by undergraduate research fellowships from the Carroll County (Indiana) Cancer Society and the American Foundation for Pharmaceutical Education. EE Williams was supported by an undergraduate research fellowship from the American Association of Colleges of Pharmacy and Merck. RP Hammer was supported by an NIH sabbatical leave fellowship (F33CA085049). We also acknowledge additional support from the NIH (R21CA080770 to DJ Riese) the U.S. Army Medical Research and Materiel Command (DAMD17-00-1-0415 and DAMD17-00-1-0416 to DJ Riese), the Indiana Elks Foundation (to DJ Riese), and the American Cancer Society (IRG-58-006 to the Purdue Cancer Center).
About this article
Cite this article
Hobbs, S., Coffing, S., Le, A. et al. Neuregulin isoforms exhibit distinct patterns of ErbB family receptor activation. Oncogene 21, 8442–8452 (2002). https://doi.org/10.1038/sj.onc.1205960
Molecular Neurobiology (2020)
Scientific Reports (2017)
Acquisition of EMT phenotype in the gefitinib-resistant cells of a head and neck squamous cell carcinoma cell line through Akt/GSK-3β/snail signalling pathway
British Journal of Cancer (2012)
Journal of Investigative Dermatology (2008)