Src Family Kinases and HER2 Interactions in Human Breast Cancer Cell Growth and Survival

Src family kinases and HER2 interactions in human breast cancer cell growth and survival

Allison P Belsches-Jablonski1,2, Jacqueline S Biscardi1, Dena R Peavy1, David A Tice1, Davis A Romney1 and Sarah J Parsons*,1

1Department of Microbiology and Cancer Center, Box 441, University of Virginia Health Sciences Center, Charlottesville, Virginia, VA 22908, USA; 2Biology Program, Lynchburg College, 1501 Lakeside Drive, Lynchburg, Virginia, VA 24501, USA

Evidence from murine ®broblast models and human (EGFR or HER1), HER2/neu, HER3 and HER4. breast cancer cells indicates that c-Src and human EGF These receptors are involved in regulation of a number receptor (HER1) synergize to enhance neoplastic growth of dierent cellular processes, including mitogenesis, of mammary epithelial cells. To investigate whether tumorigenesis, and dierentiation (Alroy and Yarden, interactions between c-Src and other HER family 1997; Pinkas-Kramarski et al., 1998). HER1 has been members may also play a role in breast tumor shown to bind and become activated by dierent progression, we characterized 13 human breast carcino- members of the EGF family of ligands, including EGF, ma cell lines and 13 tumor samples for expression of TGF-a, epiregulin and betacellulin (Linsley and Fox, HER family members and c-Src and examined a subset 1980; Massague, 1983; Barnard et al., 1994; Komur- of the cell lines for Src-dependent, heregulin (HRG)- asaki et al., 1997; Riese et al., 1998; Shelly et al., 1998). augmented, anchorage-dependent and independent HER3 functions as a receptor for neuregulin 1 (also growth. By immunoblotting, we found that all cell lines known as heregulin [HRG] and neu dierentiation overexpressed one or more HER family member, and factor) and neuregulin 2 (Carraway and Cantley, 1994; 60% overexpressed c-Src. Seventy-®ve per cent of the Hynes and Stern, 1994; Crovello et al, 1998). HER4 tumor tissues overexpressed HER2, while 64% over- also serves as a receptor for HRG, as well as for expressed c-Src. Colony formation in soft agar was betacellulin (Barnard et al., 1994) and epiregulin (Riese enhanced by HRG in three of ®ve cell lines tested, a et al., 1998). response that correlated with the presence of a c-Src/ No ligand has been identi®ed for HER2. Activation HER2 heterocomplex. This result suggests that HRG of HER2 most likely occurs through overexpression may act through both HER2 and c-Src to facilitate and homodimerization, or through heterodimerization anchorage-independent growth. In contrast, HRG had with ligand-activated HER1, HER3 or HER4. The little eect on anchorage-dependent growth in any of the HER2/3 heterodimer is preferred and most tumorigenic cell lines tested. PP1, a Src family kinase inhibitor, of the possible combinations (Wallasch et al., 1995; reduced or ablated HRG-dependent and independent soft Karunagaran et al., 1996; Pinkas-Kramarski et al., agar growth or anchorage dependent growth, and 1996; Tzahar et al, 1997). Dimerization is followed by triggered apoptosis in all cell lines tested. The apoptotic trans-phosphorylation between paired receptor mono- eect of PP1 could be partially or completely reversed mers and stimulation or recruitment to the receptor by HRG, depending on the cell line. These results complex of signaling molecules, including PI-3 kinase suggest that while Src family kinases may cooperate with (phosphatidylinositol-3 kinase), SHC, GRB2, and HRG to promote the survival and growth of human GRB7 (Segatto et al., 1993; Sepp-Lorenzino et al., breast tumor cells, they also function independently of 1996; Dankort et al., 1997; Fiddes et al., 1998). In the HER2/HRG in these processes. Oncogene (2001) 20, HER2/3 heterodimer, HER3, which exhibits reduced 1465 ± 1475. intrinsic catalytic activity (Guy et al., 1994) appears to provide ligand binding activity and the ability to Keywords: Src family kinases; heregulin; breast cancer; associate with PI-3 kinase, while HER2 contributes HER2/neu; cell growth; apoptosis kinase activity and the ability to phosphorylate downstream substrates (Carraway and Cantley, 1994; Solto et al., 1994; Gamett et al., 1995). Introduction HER2 has potent oncogenic activity in cultured ®broblasts when overexpressed as a wild type protein, The HER family of transmembrane tyrosine kinases when activated by a single point mutation in the consists of four members, the human EGF receptor transmembrane domain, or by deletion of the ®rst 621 amino acids from the N-terminus (human [Pierce et al., 1991]; rat [DiFiore et al., 1987]). Increased expression of HER2 is observed in a number of human cancers, *Correspondence: SJ Parsons Received 3 July 2000; revised 18 December 2000; accepted 3 January including breast (Slamon et al., 1989), ovarian 2001 (Berchuck et al., 1990), gastric (Kameda et al., 1990), c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1466 lung (Kern et al., 1990) and prostate (Ware et al., Results 1991). In breast cancers, overexpression and/or ampli®cation of HER2, which occurs in *30% of Overexpression of c-Src and HER family members in the cases, has been correlated with poor prognosis human breast carcinoma cell lines (Slamon et al., 1987, 1989). Overexpression of HER2 can function as a causal factor for breast cancer A panel of 13 human breast carcinoma cell lines was development, as has been demonstrated by targeted analysed by Western blotting for expression levels of c- overexpression of HER2/neu and development of Src and HER family members (a representative series mammary tumors in mice (Guy et al., 1992). More- of blots is shown in Figure 1). Results of the scanning over, HRG treatment of cells overexpressing HER2 densitometry analysis of the immunoblots is shown in causes increases in cell proliferation and tumorigenesis Table 1. Eight of the cell lines (*60%) expressed c-Src (Aguilar et al., 1999), suggesting that HRG indirectly greater than ®vefold above the immortalized, non- acts upon HER2 to increase downstream signaling. tumorigenic human breast epithelial cell line, Recently, several laboratories have reported the Hs578Bst. All tumor cell lines tested overexpressed detection of elevated levels of c-Src protein/activity or one or more of the four HER family members, in physical complexes between members of the HER various combinations. Within any given cell line, receptor tyrosine kinase family (including HER2) and however, overexpression of HER1 and HER2 appeared c-Src in human and mouse mammary tumors (Ottenh- to be mutually exclusive. This ®nding agrees with o-Kal et al., 1992; Sierke et al., 1993; Luttrell et al., published literature, which indicates that HER2 is 1994; Muthuswamy et al., 1994; Muthuswamy and typically expressed in earlier stage breast carcinomas Muller, 1995; Maa et al., 1995; Sato et al., 1995; (DePotter et al., 1990; Ramachandra et al., 1990; Campbell et al., 1996; Biscardi et al., 1998). Studies Maguire et al., 1992;), while HER1 is expressed in done in HER2-expressing mammary epithelial cells later, more aggressive tumors (Sainsbury et al., 1987; (Sheeld, 1998) or in transgenic mouse systems Toi et al., 1991). In addition, HER2 and HER3 were expressing HER2 (Muthuswamy et al., 1994) demon- found to be co-overexpressed (twofold or greater above strated that overexpression of HER2 leads to an endogenous levels) in the majority of cell lines activation of c-Src kinase activity in mammary cells. (*60%), thereby enhancing the probability of forma- These ®ndings suggest that functional interactions between c-Src and HER family members may play a role in regulating human breast tumor formation and progression, but only in the case of HER1 and c-Src has biological synergism between the two kinases been demonstrated (Maa et al., 1995; Biscardi et al., 1998, 1999a, 2000; Tice et al., 1999). To investigate whether c-Src is capable of function- ally interacting with other HER family members as it does with HER1, a panel of human breast tumor cell lines and tumor tissues was ®rst examined for levels of HER family members and c-Src, and for evidence of physical interactions between HER2 and c-Src. Using a subset of these lines (selected for various patterns of HER family member and c-Src overexpression and HER2/c-Src association), HRG was used as a ligand to activate HER2 through heterodimerization with HER3 or HER4. We observed that HRG increased colony formation in soft agar above that seen in serum alone and that enhanced anchorage-independent growth appeared to correlate with the ability to detect c-Src/ HER2 heterocomplexes. In such cell lines, HRG- Figure 1 Overexpression of HER family members and c-Src in enhanced soft agar colony formation was partially or human breast carcinoma cell lines. Lysates from indicated cell lines were prepared in RIPA buer, and 100 mg protein was completely inhibited by PP1, a Src family kinase loaded per well in a 7% polyacrylamide gel. Western blots were inhibitor (Hanke et al., 1996). Anchorage-dependent probed with antibodies to c-Src (2 ± 17); HER1 (3A and 4A); growth in low serum, which was mostly unaected by HER2 (Santa Cruz SC-284); HER3 (Santa Cruz SC-285); and HRG treatment, was likewise inhibited by PP1. HER4 (Santa Cruz SC-283/UBI #06-572). Primary antibodies were visualized by 125I-protein A, except for the HER4 antibodies, Interestingly, extended PP1 treatment of adherent cells which were detected with 125I-goat anti-rabbit IgG. Exposure time in low serum was found to induce apoptosis, and HRG of the ®lm was selected to re¯ect the ranges and pattern of rescued this eect. Taken together, these data provide expression and not visualization of the lowest levels of detectable evidence for cooperation between c-Src and HER2 in expression. This ®gure is representative of four dierent mediating HRG-induced growth and survival of human experiments, whose combined results are described in Table 1. Portions of this ®gure are reproduced from Biscardi et al., breast tumor cells, and also demonstrate that c-Src can Copyright 1998, Wiley-Liss, Inc., a subsidiary of John Wiley & function independently of HER2/HRG in these events. Sons, Inc

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1467 Table 1 Human breast carcinoma cell lines. Relative levels of HER family members and c-Src{ Association of c-Src with: Cell line HER1 HER2 HER3 HER4 c-Src ER HER1 HER2

MDA-MB175 + ++++ ++ + 77 UACC-893 + ++++ ++ ++7ND ND UACC-812 + ++++ ++ ++77 + SK-BR-3 +++ + +++++ 7 + 7 MDA-MB361 + +++ ++ + ++++ + 7 + MDA-MB453 + +++++77 + MDA-MB468 ++++ +++++ 7 + 7 ZR75-B +++++ ++ + 77 BT-474 + ++++ +++ ++ + + 77 BT-549 ++ ++++++ 7 + 7 MDA-MB231 ++ ++++ 7 + 7 BT-20 + + ++ + +++ 7 + 7 MCF-7 +++ + ++ + 77 Hs578Bst1 ++++++ 77

+, represents basal or undetectable levels by immunoblotting; +, 2 ± 5-fold above basal levels; ++, 5 ± 10-fold; +++, 10 ± 25-fold; ++++, 425-fold, ND indicates no data. {Autoradiographs from Figure 1 and repetitions of this experiment were analysed by densitometry, and values for each of the HER family members and c-Src from the tumor lines were compared to Hs578Bst1, an immortalized normal human mammary epithelial cell line. `+' designations represent ranges of values since measurable levels for each of the proteins varied slightly from experiment to experiment. Also included in the table are estrogen receptor (ER) status of the cell lines, as determined by Western blotting (Biscardi et al., 1998), and physical assocation between c-Src and either HER1 or HER2, as determined by co-immunoprecipitation with monoclonal antibodies to c-Src. Portions of this Table are re-printed with permission of Molecular Carcinogenesis tion of the preferred, more tumorigenic HER2/3 185 kDa phosphoprotein in UACC812 cells was heterodimer. Under our immunoblotting conditions, ascertained to be HER2 by stripping the blot and only two cell lines, ZR75B and BT-474, were observed reprobing with HER2-speci®c antibodies (data not to have signi®cantly elevated levels of HER4. In shown). Physical association of c-Src and HER1 in summary, 100% of the breast tumor cell lines over- MDA-MB-468 cells was also seen under these condi- expressed one or more HER family member, while tions, as was described previously in Biscardi et al. 60% overexpressed c-Src. (1998); however, no co-precipitating HER2, which migrated at a position distinct from that of HER1, was detectable. Physical association of c-Src with HER2 in human breast carcinoma cell lines Association of c-Src with HER2 in human breast tumor To investigate whether there were potential structural samples and functional interactions between c-Src and HER2 such as exists between c-Src and HER1 (Maa et al., A panel of 13 tumor tissues, obtained from the 1995; Biscardi et al., 1998, 1999a), we analysed the University of Virginia and University of Michigan panel of human breast carcinoma cell lines for evidence tumor banks, was analysed by immunoprecipitation of in vivo association between c-Src and HER2. Figure and Western blotting with antibodies to HER2 and c- 2a depicts results from two representative cell lines: one Src in a fashion similar to that used for the breast positive for co-immunoprecipitation of HER2 and c- tumor cell lines. Figure 2c shows examples of this Src (MDA-MB-361) and one negative. MDA-MB-468 analysis. Tumor UVA263 exhibited c-Src/HER2 asso- was selected as a negative control cell line since it ciation, whereas normal breast tissue did not. In contains undetectable levels of HER2. The outcome of Tumor UVA263, a faint band that immunoblots the analysis is summarized in Table 1, which indicates speci®cally with c-Src antibodies and could be observed that a stable complex between HER2 and c-Src was in the HER2 immunoprecipitates, suggesting a reci- observed in three of the 13 cell lines tested (MDA-MB- procal co-precipitation between c-Src and HER2 361, MDA-MB-453, and UACC-812). It is possible (Figure 2c). In addition, HER2/1 complexes could be that c-Src/HER2 association did occur in the other cell seen, indicating constitutive heterodimerization be- lines as well, but was undetectable perhaps due to tween HER family members in this tumor. Consistent limitations in sensitivity of the assay. Complex with the data from the cell lines, association in the formation was con®rmed in other experiments in which tumors was constitutive, suggesting the presence of an c-Src was immunoprecipitated with a dierent c-Src autocrine loop. Normal breast tissue contained detect- speci®c antibody (GD11 vs 2 ± 17) and immunoblotted able levels of c-Src but undetectable levels of HER2, with HER2-speci®c antibodies. Figure 2b shows and no association between HER2 and c-Src, HER1, association between a tyrosine phosphorylated protein or HER3 was observed (Figure 2c and data not of 185 kDa and c-Src in UACC-812 cells, which shown). overexpress HER2, but not in MDA-MB-468 or Table 2 summarizes the results of an analysis of all MCF7 cells, which contain little or no HER2. The 13 tumor samples and describes tumor type and grade,

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1468 estrogen receptor (ER) status, lymph node involvement, relative levels of c-Src and HER2 protein, and the presence or absence of a stable complex between c- Src and HER2 for each tumor. In all, stable association between c-Src and HER2 was seen in three out of 13 samples, the same frequency as seen for the breast tumor cell lines. This is the ®rst demonstration of an in vivo association in human breast carcinoma samples.

HRG-enhanced soft agar colony formation is ablated by the Src family inhibitor, PP1 In human breast tumor cell lines, HRG treatment has been demonstrated to increase tyrosine phosphorylation of many cellular proteins that are involved in mitogenic signaling pathways (Carraway et al., 1995; Belsches-Jablonski, unpublished observations). It was therefore of interest to determine if HRG treatment aected the mitogenic or tumorigenic properties of these cell lines. To investigate this question, we chose a subset of the cell lines described in Table 1, based on their varying expression of HER family members and c-Src. Three cell lines (MDA-MB-361, MDA-MB-453, and UACC-812) were selected because they co-overexpressed modest to high levels of HER2 and HER3 and exhibited HER2/c-Src association. Of these, only MDA-MB-361 overexpressed c-Src, while the other two contained basal levels. MCF-7 and MDA-MB-468 cells were selected as control cell lines on the basis of their relatively low levels of HER2/HER3 (basal or undetectable), moderate levels of c-Src (5 ± 10-fold over basal), and lack of detectable HER2/c-Src complexes. Figure 3 shows that only the cell lines that co- overexpressed HER2 and HER3 (MDA-MB-361, MDA-MB-453, UACC-812) responded to HRG in a soft agar colony formation assay. HRG-induced increases in soft agar growth correlated with the presence of a preformed heterocomplex between Figure 2 Physical association between c-Src and HER2 in HER2 and c-Src in these cells. HRG treatment human breast carcinoma cell lines and in human breast tumor potentiated anchorage-independent growth of these tissue. (a) MDA-MB-361 and MDA-MB-468 cells were serum- starved for 17 h and either stimulated with 200 ng/ml HRG-a (+) cell lines 2 ± 3-fold over that in 10% serum. In contrast, or incubated in fresh serum-free media (7) for 5 min. Cells were HRG had no signi®cant eect on serum-induced then lysed, and 500 mg lysate protein from each cell line were colony formation of MCF7 or MDA-MB-468 cell incubated with either anti-Src (S), anti-HER2 (HER2) or mouse lines, which was most likely due to the fact that these IgG negative control [(7)C] antibodies. Precipitated proteins were separated by SDS ± PAGE and transferred to a nylon cells express low to undetectable levels of HER2/ membrane. The upper half of the membrane was probed with HER3. These ®ndings suggest that HER2, HRG, and anti-HER2 antibody, and the lower half was probed with anti-Src c-Src may cooperate to increase tumorigenic growth. (2 ± 17) antibody. Both primary antibodies were visualized by 125I- To test this hypothesis further, sister cultures were protein A. (b) MDA-MB-468, MCF7, and UACC812 cells were treated with the Src family tyrosine kinase inhibitor serum-starved for 17 h, lysed in RIPA buer, and 500 mg lysate protein from each cell line was incubated with either anti-Src (S) PP1 in the presence or absence of HRG and in the (GD11 mAb), anti-HER2 (HER2) or mouse IgG [(7)C]. continual presence of 10% serum. Figure 3 shows that Precipitated proteins were separated by SDS ± PAGE and PP1 partially or completely inhibited HRG-enhanced transferred to a nylon membrane. The upper half of the soft agar growth in MDA-MB-361, MDA-MB-453, membrane was probed with anti-phosphotyrosine primary antibody, and the lower half was probed with anti-Src (2 ± 17) and UACC-812 cells. Furthermore, PP1 had a striking primary antibody. Both primary antibodies were visualized by inhibitory eect on the ability of all cell lines to form 125I-protein A. (c) Normal breast or breast tumor samples were homogenized, lysed in RIPA detergent buer, and clari®ed by centrifugation as described in Materials and methods. Cellular proteins were immunoprecipitated with anti-Src (S; 2 ± 17 MAb), SDS ± PAGE and blotted as in (a) and (b). The dotted arrow anti-HER1 (H1), anti-HER2 (H2), or mouse IgG (7) from indicates a protein that co-precipitates with HER2 and co- 250 mg lysate protein. Precipitated proteins were separated by migrates with c-Src

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1469 Table 2 Human breast tumors tested for physical association of HER2 and c-Src{ Src/HER2 Relative levels Tumor Tumor type ER Lymph node association c-Src HER2

UVA103 IVDCa, grade 3/3 77 7ND ND UVA156 IDCb, grade 2/3 7 + + +++ +++ UVA226 IDAc, grade 2/3 + ND + ++ +++ UVA263 IDC, grade 3/3 ND + + +++ +++++ UVA387 ILCd +ND 7 +* UVA399 IDC, grade 3/3 7 ND 7 + +++ UVA454 IDC, grade 3/3 7 + 7 ++ ++ UVA616 IVDC, grade 3/3 7 + 7 ND ND MICHN1 medullary, grade 3/3 ND ND 7 +++ MICHP1 # ND ND 7 *++ MICHP2 # ND ND 7 *+ MICHP3 # ND ND 7 *+ MICHP4 # ND ND 7 *+

{Human breast tumor samples obtained from the UVA Tissue Procurement Facility and the University of Michigan Tumor Bank were analysed by immunoprecipitation and Western blotting for physical association between HER2 and c-Src. Tumor type and grade, ER and lymph node status, and relative levels of HER2 and c-Src protein as listed for each tumor. aIVDC, invasive ductal carcinoma; bIDC, in®ltrating ductal carcinoma; cIDA, in®ltrating ductal adenocarcinoma; dILC, in®ltrating lobular carcinoma; #, data currently not available; ND, no data; *, below limits of detection

HRG-potentiated anchorage-dependent growth is partially dependent on Src family kinases Because HRG treatment led to an increase in anchorage-independent growth of the three cell lines that co-overexpressed HER2 and HER3 and contained constitutive detectable c-Src/HER2 complexes, we wished to determine whether HRG also aected anchorage-dependent growth of these and other cell lines. An MTT assay was used to follow cell number as a function of incubation time in various treatment conditions. No eect of HRG on the cell number of any of the cell lines was observed in 10% serum, even when assessed up to 2 weeks after plating (data not shown). However, Figure 4 shows that HRG treatment resulted in small, but signi®cant increases in numbers of MDA-MB-361, UACC-812, and MCF7 cells, when they were maintained in reduced serum (0.5%) over a 5 day period (Figure 4a,b,c; compare `day 5' to `day 5+H'). PP1 treatment prevented growth or reduced the cell number of all cell lines maintained in low serum alone Figure 3 Eects of HRG and the Src kinase family inhibitor, (Figure 4; compare `day 5+P' to `day 5' in each panel), PP1, on anchorage-independent growth of human breast tumor but the eect of PP1 on cells treated with HRG varied cells. MDA-MB-361, MDA-MB-453, UACC-812, MCF7, and from cell line to cell line. HRG-treated MDA-MB-361 MDA-MB-468 cells were plated in soft agar as described in Materials and methods and incubated at 378C in a humidi®ed, and MCF7 lines were only slightly reduced in cell 5% CO2 atmosphere for 2 weeks in DMEM containing the number by PP1 treatment (Figure 4a,c; compare `day following: 10% serum+DMSO vehicle control; 10% ser- 5+H+P' to `day 5+H'), whereas UACC-812 was um+40 ng/ml HRG-a; 10% serum+40 ng/ml HRG-a+10 mM reduced to levels equal to or below that of cells PP1; or 10% serum+10 mM PP1. Colonies were stained, counted, maintained in 0.5% serum alone (Figure 4b). In this and expressed as the mean colony number+s.e.m. of three independent experiments, each performed in triplicate. *Signi®- regard, PP1 could exert its eects either by inhibiting cantly dierent than 10% serum alone (P-values 40.05). Mean anchorage-dependent growth or by decreasing cell number of colonies in serum alone were 25 for MDA-MB-361; survival. These results suggest that c-Src and/or related 100 for MDA-MB-453; 120 for UACC-812, 720 for MCF7, and family members are capable of cooperating with HRG 525 for MDA-MB-468 cells to promote anchorage-dependent growth and/or survival of the breast cancer cell lines tested, but the extent of this cooperation varies from cell line to cell line. colonies in serum alone, suggesting that PP1 is capable Results similar to those for MCF7 cells were found for of aecting other growth pathways in addition to those MDA-MB-468 and SK-BR3 cells. MDA-MB-453 cells that are HRG-induced. did not respond to HRG in this assay (data not shown).

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1470

Figure 4 Eects of HRG and PP1 on anchorage-dependent, reduced-serum growth and survival of human breast cancer cells. MDA-MB-361 (a), UACC-812 (b), and MCF7 (c) cells were plated in DMEM with the following additives: 0.5% serum+DMSO vehicle control; 0.5% serum+40 ng/ml HRG-a or -b+DMSO vehicle control (+H); 0.5% serum+HRG+10 mM PP1 (H+P); or 0.5% serum+10 mM PP1 (+P), as described in Materials and methods. The cultures were incubated for 5 days, with replenishment of media containing the appropriate additives at day 2.5. Cell number at days 0 and 5 was determined by the MTT assay and is expressed as mean fold-change+s.e.m. relative to day 0. Day 0 density was determined 12 h after plating. Results are pooled from four independent experiments, each performed in triplicate. #Signi®cantly dierent than day 5, 0.5% serum alone. cSigni®cantly dierent than day 5, 0.5% serum+HRG. P-values 40.05

PP1 has previously been shown to inhibit the Src except that the extents of reversal diered slightly family members, Fyn and Lck (Hanke et al., 1996). To (Figure 5b,c). In all cell lines, the anti-apoptotic eects determine whether PP1 also inhibited c-Src kinase of HRG were found to be insensitive to Wortmannin, activity, immune complexes of c-Src that had been suggesting a PI-3 kinase-independent mechanism of prepared from MDA-MB-361 cells were preincubated action by HRG (data not shown). Results of the DAPI for 15 min with varying concentrations of PP1 (ranging analysis were veri®ed by the TUNEL assay (data not from 0.1 to 10 mM) before the auto- and trans- shown). These ®ndings suggest that HRG can stimulate phosphorylating activities were assessed in an in vitro survival pathways independently of PI-3 kinase in these immune complex kinase assay, using heat-denatured cells, and that overexpression of HER2/3 is not enolase as an exogenous substrate and/or [g-32P]-ATP. required for the eect. Both the auto- and trans-phosphorylating activities of c- Src were reduced by PP1 in a dose- and time-dependent manner (data not shown). The IC50 was 0.3 ± 1.0 mM Discussion PP1 in a 10 min incubation. These results veri®ed that PP1 is an ecient inhibitor of c-Src kinase activity. Both HER family members (Sainsbury et al., 1987; Slamon et al., 1987, 1989; Biscardi et al., 1998; this study) and c-Src (Ottenho-Kal et al., 1992; Camp- Heregulin reverses the apoptotic effect of PP1 in reduced bell et al., 1996; Biscardi et al., 1998; this study) are serum overexpressed at high frequency in human breast PP1 has previously been shown to induce apoptosis in cancers and have been implicated in the genesis and NIH3T3 cells (Karni et al., 1999). Because PP1 had progression of the disease. Ottenho-Kal et al. (1992) such a deleterious eect on the growth and viability of found elevated levels of c-Src kinase activity were the majority of breast cancer cell lines tested (which present in 70% of human breast tumors examined. In was partially or completely reversed by HRG), we transgenic mouse studies where overexpressed HER2 examined three of the cell lines (MDA-MB-361, was targeted to mammary epithelium, the resulting UACC-812, and MCF7) for possible evidence of an mammary tumors overexpressed c-Src as well, and c- apoptotic eect of PP1 and an anti-apoptotic eect of Src and HER2 were found to physically associate HRG. Cells were cultured for 3 days in the presence or (Muthuswamy et al., 1994; Muthuswamy and Muller, absence of HRG and/or PP1 and analysed for evidence 1995). Ampli®cation/overexpression of HER2, espe- of apoptosis, using DAPI-staining of DNA to detect cially, has been associated with a poor patient abnormally shaped nuclei and/or condensed chromatin prognosis (Slamon et al., 1987, 1989). Synergistic (Villaneuva et al., 1984). Figure 5a shows that PP1 interactions between c-Src and HER1 in tumor treatment of MDA-MB-361 cells induced apoptosis in formation have recently been reported (Maa et al., 10 ± 15% of the cells (vs *2% in 0.5% serum alone) 1995; Biscardi et al., 1998, Tice et al., 1999), and c-Src and that PP1-induced apoptosis was almost completely has been shown to be required for EGF-induced reversed by HRG. UV-induced apoptosis was used as a mitogenesis (Wilson et al., 1989; Roche et al., 1995), positive control (485%). Similar results were obtained suggesting that similar interactions may exist between when UACC-812 and MCF7 cells were examined, c-Src and other HER family members.

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1471

Figure 5 Anti-apoptotic eects of HRG and Src family kinases in human breast tumor cell lines. Cells were plated onto glass coverslips (56104 cells/coverslip) and grown in one of the following, DMEM plus: 0.5% serum alone+DMSO vehicle control; 0.5% serum+40 ng/ml HRG-a+DMSO vehicle control (H); 0.5% serum+10 mM PP1 (P); or 0.5% serum+40 ng/ml HRG- a+10 mM PP1 (H+P), and incubated for 3 days. Cells were then ®xed in 4% paraformaldehyde and permeabilized, and assessed for nuclear condensation by DAPI staining. At least 400 cells per treatment group were analysed for the presence of apoptotic nuclei in two separate experiments. A 2-min UV-B treatment and 48 h recovery in serum-free media was used as a positive control for apoptosis. (a) MDA-MB-361; (b) UACC-812; (c) MCF7. *Signi®cant dierence between 0.5% serum, H, or H+P and P (P-values 40.05)

Here we report a physical association between HER2 that Src family kinases participate in HRG-regulated and c-Src in a subset of human breast cancer cell lines cellular processes and vice-versa, but the extent to and tumors and provide evidence for cooperative which this occurs is likely to be determined by the interaction between c-Src and HRG/HER2/3 mediated unique signaling circuitry of each tumor cell line. These pathways in controlling cell growth and survival. results are also suggestive of Src-independent as well as Aguilar et al. (1999) have demonstrated that HRG Src-dependent mechanisms for HRG action. has growth promoting properties and increases cell Further evidence for the requirement of c-Src for the proliferation, soft agar colony formation and tumor- growth of breast tumor cells comes from studies igenicity in a variety of breast cancer cells. These attempting to stably express a kinase-inactive form of investigators also showed that the stimulatory eects of c-Src in human breast tumor cells. MDA-MB-361 cells HRG are enhanced in cells with endogenous or were transfected with a dominant negative, kinase engineered overexpression of HER2. Our results are inactive form of chicken c-Src (Belsches-Jablonski and consistent with this ®nding in that treatment of certain S Parsons, unpublished data) and selected for breast cancer cell lines with HRG caused an increase in neomycin resistance. Although short-term selected anchorage-independent growth. In addition, we showed populations of transfected cells were found to over- that this enhanced growth correlates with the presence express kinase-defective c-Src (as measured by im- of a preformed c-Src/HER2 heterocomplex (Figure 3). munoblotting), no long-term stable clones could be The Src family inhibitor PP1 inhibited both HRG- obtained. These results suggest a striking dependence and serum-dependent colony formation and adherence- on c-Src kinase activity for growth and survival of dependent proliferation (Figures 3 and 4), while long these cells, even in full serum conditions. term PP1 treatment triggered apoptosis in all cell lines One mechanism by which HRG-regulated pathways tested (Figure 5). Interestingly, this PP1-induced may interact with Src family members is through inhibition of growth or induction of apoptosis could physical association of HER2 with c-Src. In such a be partially or completely rescued by treatment with scenario, c-Src might be thought to function by HRG, with the degree of rescue varying among the facilitating formation of heterocomplexes between various cell lines. The inhibitory eects of PP1 HER2 and other HER family members or by occurred in cell lines with both endogenous and phosphorylating HER2 or one of its receptor partners, overexpressed levels of c-Src. Likewise, the pro-growth as has been shown for HER1 (Maa et al., 1995; eects of HRG were not dependent on HER2 over- Biscardi et al., 1999a). Other researchers have shown expression, suggesting that endogenous levels of HER2 that c-Src and HER2 can physically associate and HER3/4 can mediate HRG eects. PP1 has been (Muthuswamy and Muller, 1995; Luttrell et al., shown to speci®cally inhibit Src family kinases and is 1994); however, the biological consequences of this at least an order of magnitude less ecient at association are unclear. Our data suggest that c-Src/ inhibiting the EGFR, ZAP70, JAK2, or protein kinase HER2 interaction may contribute to regulation of A (Hanke et al., 1996). Therefore, our results suggest anchorage-independent growth and survival signals,

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1472 although a stable physical association between the two fected by HRG treatment. Given the many ligands that kinases was linked only with anchorage-independent have been shown to bind to HER family members, it is growth. This ®nding does not preclude the possibility reasonable to speculate that autocrine production of that heterocomplexes do exist in survival signaling, ligands could cause constitutive activation of the only that they cannot be detected by co-immunopre- receptors and thus association with c-Src. In support cipitation, possibly due to their low abundance. What of this is our ®nding that p-Tyr-containing proteins in factors regulate formation of the heterocomplexes are the 170 ± 185 kDa range are seen in p-Tyr immunoblots unclear at the present time. These issues warrant from lysates derived from each cell line which contains further investigation in a non-tumorigenic cell system, the c-Src/HER2 complex (data not shown). Treatment where the constituents of the complex can be regulated of these cell lines with EGF or HRG did not increase by overexpression and tumor-speci®c modi®cations to the overall amount of p-Tyr signal, consistent with the the receptor are eliminated. notion that the endogenous receptors were already The ability to detect c-Src/HER2 complexes in the activated. This issue warrants further investigation in various breast tumor cell lines correlated with the model systems where the presence or levels of HER2 ability of HRG to promote colony formation in soft and c-Src can be regulated. agar, suggesting that the complex may participate in The correlation between the ability to detect a c-Src/ anchorage-independent growth. Involvement of c-Src HER2 complex in certain cell lines and to observe a in such a process would be consistent with its biological eect on HRG-induced growth in soft agar documented roles in regulating focal adhesion forma- suggests that this complex may be important in tion (cell-substratum interactions) and inter-cellular promoting malignant transformation. The functional interactions via cadherins (Malik and Parsons, 1996; linkage between Src family kinases and HRG in the reviewed in Parsons and Parsons, 1997). Indeed, we anchorage-independent growth assays provides addi- have observed complete inhibition of p190RhoGAP, tional support for the hypothesis that c-Src and HER2 p130CAS, and cortactin tyrosine phosphorylation in cooperate at many levels to increase tumorigenicity, as ®ve dierent human breast tumor cell lines treated with we and others have demonstrated for c-Src and HER1 PP1 (D Tice and S Parsons, unpublished observations). (Maa et al., 1995; Belsches et al., 1997; Biscardi et al., These proteins are all demonstrated substrates of c-Src 1998, 1999a, 2000; Tice et al., 1999). However, and involved in signaling pathways that regulate the available evidence suggests that the mechanism of actin cytoskeleton (reviewed in Belsches et al., 1997). cooperativity between c-Src and HER2 may be While the presence of HER2 in c-Src immunopreci- dierent from that of c-Src and HER1. For example, pitates from MDA-MB-361, MDA-MB-453, and we have not observed a common downstream signaling UACC-812 cells was reproducibly seen, the amount molecule that is activated in all the breast tumor cell of HER2 co-associating with c-Src was low, represent- lines we stimulated with HRG, while SHC and MAP ing less than 5% of the total HER2 in the cell. The kinase have been routinely found to be activated in amount of association in the reciprocal immunopreci- those same cell lines by EGF (Biscardi et al., 1998; pitation occurred at similar levels, but was much less Belsches-Jablonski and S Parsons, unpublished). Cur- reproducible. This result could re¯ect the low abun- rent eorts are directed toward understanding the dance of the complex or an interaction that is weak in nature of the physical interaction between c-Src and nature. The fact that c-Src/HER2 complexes were not HER2 (direct or indirect) and the signaling pathways observed in all cell lines does not necessarily argue that convey their overlapping and unique directives to against their existence, since they may be present but the nucleus. undetectable by our assay system. (Maa et al., 1995). HER1 (EGF receptor) was previously shown to associate with c-Src in breast cancer cells and ®broblasts, and this interaction also occurred at a low stoichiometry (Maa et al., 1995; Biscardi et al., Materials and methods 1998). However, the fact even a low abundance or weak association may be physiologically relevant is Cell lines evidenced by the HER1/c-Src complex, for which a Tumor cell lines MDA-MB-468, MDA-MB-231, MCF7 and functional consequence has now been demonstrated ZR75-1 were provided by N Rosen (Sloan-Kettering, NY, (Biscardi et al., 1999a; Tice et al., 1999). USA), while MDA-MB-175, UACC-893, UACC-812, SK- Despite the structural similarities between HER1 and BR-3, MDA-MB-361, MDA-MB-453, BT-474, BT-549, BT- HER2 and their involvement in the genesis of breast 20 and HS578Bst were obtained from American Type Culture cancer and other human tumors, there is a striking Collection (Rockville, MD, USA). Tumor cells were main- dierence between c-Src/HER1 and c-Src/HER2 com- tained in Dulbecco's modi®ed Eagle medium (DMEM) (Life Technologies, Bethesda, MD, USA) in 10% fetal calf serum plexes. In both a murine ®broblast model system and (Atlanta Biologicals, Atlanta, GA, USA). Where indicated, in a subset of human breast tumor cell lines, physical cells were stimulated with recombinant human HRG-a, EGF- association of c-Src with HER1 is EGF-dependent like domain (HRG-a) (200 ng/ml) (Sigma Chemical, St. (Maa et al., 1995; Biscardi et al., 1998), while in human Louis, MO, USA) or HRG-b 1 (extracellular domain) tumor cell lines, we found that the formation of (NeoMarkers, CA, USA). HRG-a and HRG-b1 gave nearly HER2/c-Src complexes to be constitutive and unaf- identical results and were used interchangeably.

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1473 continuous rocking at 48C for 2 h and then for an additional Growth assays hour at 48C with protein A-sepharose beads (Sigma Tumor cell lines were plated in triplicate in DMEM plus Chemical). Beads were pelleted and washed three times with 0.5% fetal calf serum at a density of either 26104 cells/well cold RIPA buer. Pellets were resuspended in 26 sample (MDA-MB-468, MCF7) or 56104 cells/well (UACC-812, buer (125 mM Tris-HCl, 4% SDS, 10% glycerol, 0.02% SK-BR-3, MDA-MB-361, MDA-MB-453) in 24-well tissue bromophenol blue, 4% b-mercaptoethanol) and boiled for culture clusters (Corning, NY, USA) and allowed to adhere 5 min. Eluted proteins were separated by electrophoresis overnight at 378C in a humidi®ed, 5% CO2 atmosphere through a 7% SDS-polyacrylamide gel and immunoblotted. before treatment protocols were initiated. Cells were then For direct Western blots, 100 mg whole-cell lysate was incubated for 5 days with one of the indicated preparations in assayed per lane. Binding of primary antibodies was DMEM. Media with additives was replenished on the third visualized by 125I-protein A, used at 1 mCi/ml (New England day. Cell number was determined by the MTT (3-(4,5- Nuclear, USA), or by enhanced chemiluminescence (ECL) dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide) (Amersham, Buckinghamshire, UK). assay as per Kaspers et al. (1991). The Src family kinase For Western blots of HER4, proteins were transferred inhibitor, PP1 (Hanke et al., 1996), was obtained from from gels to Immobilon membranes using the semi-dry CalBiochem, San Diego, CA, USA. apparatus. Membranes were blocked in phosphate-buered Soft agar assays were performed as described by Maa et al. saline (PBS) containing 4% non-fat dry milk and incubated (1995). Cells were plated at a density of 16104 cells in 60 mm overnight at 48C with a mixture of the two primary anti- dishes in triplicate, in the presence or absence of HRG-a HER4 antibodies at a ®nal concentration of 2 mg/ml. (40 ng/ml), and in the presence or absence of PP1 (10 mM). Membranes were washed ®ve times in deionized water and The Src family kinase inhibitor, PP1 (Hanke et al., 1996), was incubated with 125I-goat anti-rabbit IgG in 4% non-fat dry obtained from CalBiochem, San Diego, CA, USA. In the milk in PBS for 90 min at room temperature. Five washes in absence of PP1, DMSO was used as a vehicle control. Plates water, a single wash in PBS, and ®ve additional washes in were incubated for 2 weeks, with replenishment of appro- water followed. The blot was exposed to ®lm overnight at priate media every 4 days, and stained overnight at 378C, 5% 7708C. CO2 in a solution of 1 mg/ml iodonitrotetrazolium salt (Sigma Chemical, St. Louis, MO, USA) in water. Colonies were Apoptosis assays counted using EagleSight analysis software (Stratagene, La Jolla, CA, USA). Apoptosis was quanti®ed by determining the percentage of cells that exhibited deformed nuclei and/or condensed chromatin following DAPI (di-[amidinophenyl]-indole) stain- Antibodies ing of treated cells, as described by Villaneuva et al. (1984). C-Src-speci®c antibodies used in this study include mouse Brie¯y, cells were plated onto glass coverslips in 6-well dishes monoclonal antibodies (mAb) 2 ± 17, directed against amino (Corning, NY, USA), and incubated for 3 days with DMEM acids 2 ± 17 (Quality Biotech, Camden, NJ, USA); mAbs containing the indicated additives. UV-B treatment (2 min, GD11 and EB8, both directed against residues 92 ± 128 in the followed by 48 h recovery in serum-free media) was used as a SH3 domain (Parsons et al., 1984, 1986); and mAb 327, positive control for apoptosis. After treatment, cells were which recognizes the SH3 domain (gift from J Brugge). serum-starved overnight, ®xed in 4% paraformaldehyde, HER2-speci®c antibody (rabbit polyclonal), directed toward quenched in 0.25 mg/ml H3BO4 and 20 mg/ml glycine, and residues 1169 ± 1186, and HER3-speci®c antibody, directed permeabilized in 0.5% Triton-X on ice. DAPI (1.0 mg/ml) toward residues 1307 ± 1323, were obtained from Santa Cruz (Sigma Chemical, St. Louis, MO, USA) was added for Biotechnologies, Santa Cruz; CA, USA. Antibodies speci®c speci®c DNA staining, and nuclear/chromatin morphology for HER1, mAbs 3A and 4A, were provided by D McCarley was imaged by photomicroscopy using phase contrast and and R Schatzman of Syntex Research, Palo Alto, CA, USA. immuno¯uorescence optics (Leica, Rijswijk, The Nether- The epitopes of these antibodies map to amino acid residues lands). 889 ± 944 and 1052 and 1134, respectively (Maa et al., 1995). Anti-human c-erbB-4 antibodies were obtained from Upstate Biotechnology, Inc. (UBI), Lake Placid, NY, USA (catalog #06-572) and Santa Cruz Biotechnologies (catalog #SC- 283). Anti-phosphotyrosine antibody was obtained from Transduction Laboratories (Lexington, KY, USA). Negative control antibodies included puri®ed normal rabbit or mouse immunoglobulin (Jackson Immunoresearch Laboratories, West Grove, PA, USA). Acknowledgments The authors thank Dr S Ethier of the University of Immunoprecipitation and Western immunoblotting Michigan and A Ingram from the University of Virginia Cells were lysed in RIPA detergent buer (10 mM Tris-HCl Tissue Procurement Facility for tumor tissue, and members pH 7.2, 1% Triton-X, 0.5% sodium deoxycholate, 150 mM of the T Parsons, M Weber, and S Parsons laboratories for NaCl, 1 mM EDTA, 1 mM sodium orthovanadate, 50 mg/ml their critical evaluation of this work. These studies were leupeptin and 0.5% aprotinin). Tumor samples were minced supported by grants CA71449 of the NIH and 4621 of the with a scalpel and ground in a Dounce homogenizer in RIPA Council for Tobacco Research, awarded to SJ Parsons. AP buer. All procedures were performed at 4 ± 108C. Lysates Belsches-Jablonski and JS Biscardi were recipients of post- were cleared by centrifugation for 15 min at 15 000 g in a doctoral fellowships (DAMD 17-96-6098 and DAMD 17- microcentrifuge, and protein concentration was determined 97-7329, respectively), and DA Tice is a past recipient of a by BioRad protein reagent. For immunoprecipitation, 5 mg pre-doctoral fellowship (DAMD 17-96-6126), all from the antibody was incubated with 500 mg cell lysate protein with Army Breast Cancer Initiative, D.O.D.

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1474 References

Aguilar Z, Akita RW, Finn RS, Ramos BL, Pegram MD, Komurasaki T, Toyoda H, Uchida D and Morimoto S. Kabbinavar FF, Pietras RJ, Pisacane P, Sliwkowski MX (1997). Oncogene, 15, 2841 ± 2848. and Slamon DJ. (1999). Oncogene, 18, 6050 ± 6062. Linsley PS and Fox CF. (1980). J. Supramol. Struct., 14, Alroy I and Yarden Y. (1997). FEBS Lett., 410, 83 ± 86. 441 ± 459. Barnard JA, Graves-Deal R, Pittelkow MR, DuBois R, Luttrell DK, Lee A, Lansing TJ, Crosby RM, Jung KD, Cook P, Ramsey GW, Bishop PR, Damstrup L and Coey Willard D, Luther M, Rodriguez J, Berman J and Gilmer RJ. (1994). J. Biol. Chem., 269, 22817 ± 22822. TM. (1994). Proc.Natl.Acad.Sci.USA,91, 83 ± 87. Belsches AP, Haskell MD and Parsons SJ. (1997). Frontiers Maa M-C, Leu T-H, McCarley DJ, Schatzman RC and in Biosc., 2, 501 ± 518. Parsons SJ. (1995). Proc. Natl. Acad. Sci. USA, 9, 6981 ± BerchuckA,KamelA,WhitakerR,KernsB,OltG,Kinney 6985. R, Soper JT, Dodge R, Clarke-Pearson DL, Marks P, Maguire Jr HC, Hellman ME, Greene MI and Yeh I. (1992). McKenzie S, Yin S and Bast RCJ. (1990). Cancer Res., 50, Pathobiology, 60, 117 ± 121. 4087 ± 4091. Malik RK and Parsons JT. (1996). Biochim. Biophys. Acta., Biscardi JS, Belsches AP and Parsons SJ. (1998). Mol. 1287, 73 ± 76. Carcinog., 21, 261 ± 272. Massague J. (1983). J. Biol. Chem., 258, 13614 ± 13620. BiscardiJS,MaaM-C,TiceDA,CoxME,LeuT-Hand Muthuswamy SK, Siegel PM, Dankort DL, Webster MA Parsons SJ. (1999a). J. Biol. Chem., 274, 8335 ± 8343. and Muller WJ. (1994). Mol. Cell Biol., 14, 735 ± 743. Biscardi JS, Tice DA and Parsons SJ. (1999b). Adv. In Cancer Muthuswamy SK and Muller WJ. (1995). Oncogene, 11, Res., 76, 61 ± 119. 271 ± 279. Biscardi JS, Ishizawar RC, Silva CM and Parsons SJ. (2000). Ottenho-Kal AE, Rijksen G, van Beurden EACM, Breast Cancer Res., 2, 203 ± 210. Hennipman A, Michels AA and Staal GEJ. (1992). Cancer Campbell DH, deFazio A, Sutherland RL and Daly RJ. Res., 52, 4773 ± 4778. (1996). Int. J. Cancer, 68, 485 ± 492. Parsons SJ, McCarley DJ, Ely CM, Benjamin DC and Carraway III KL, Solto SP, Diamonti AJ and Cantley LC. Parsons JT. (1984). J. Virology, 51, 272 ± 282. (1995). J. Biol. Chem., 270, 7111 ± 7116. Parsons SJ, McCarley DJ, Raymond VW and Parsons JT. Carraway III KL and Cantley LC. (1994). Cell, 78, 5±8. (1986). J. Virology, 59, 755 ± 758. Crovello CS, Lai C, Cantley LC and Carraway III KL. Parsons JT and Parsons SJ. (1997). Curr. Op. Cell Biol., 9, (1998). J. Biol. Chem., 273, 26954 ± 26961. 187 ± 192. Dankort DL, Wang Z, Blackmore V, Moran MF and Muller Pierce JH, Arnstein P, DiMarco E, Artrip J, Kraus MH, WJ. (1997). Mol. Cell Biol., 17, 5410 ± 5425. Lonardo F, DiFiore PP and Aaronson SA. (1991). De Potter CR, Beghin C, Makar AP, Vandekerckhove D and Oncogene, 6, 1189 ± 1194. Roels HJ. (1990). Int. J. Cancer, 45, 55 ± 58. Pinkas-Kramarski R, Soussan L, Waterman H, Levkowitz DiFiore PP, Pierce JH, Kraus MH, Segatto O, King CR and G, Alroy I, Klapper L, Lavi S, Seger R, Ratzkin BJ, Sela Aaronson SA. (1987). Science, 237, 178 ± 182. M and Yarden Y. (1996). EMBO J, 15, 2452 ± 2467. FiddesRJ,CampbellDH,JanesPW,SivertsenSP,SasakiH, Pinkas-Kramarski R, Shelly M, Guarino CB, Wang LM, Wallasch C and Daly RJ. (1998). J. Biol. Chem., 273, Lyass L, Alroy I, Alamandi M, Kuo A, Moyer JD, Lavi S, 7717 ± 7724. Eisenstein M, Ratzkin BJ, Seger R, Bacus SS, Pierce JH, Gamett DC, Greene T, Wagreich AR, Kim HH, Koland JG Andrews GC and Yarden Y. (1998). Mol. Cell Biol., 18, and Cerione RA. (1995). J. Biol. Chem., 270, 19022 ± 6090 ± 6101. 19027. Ramachandra S, Machin L, Ashley S, Monaghan P and Guy CT, Webster MA, Schaller M, Parson TJ, Cardi RD Gusterson BA. (1990). J. Pathol., 161, 7±14. and Muller WJ. (1992). Proc. Natl. Acad. Sci. USA, 89, Riese DJ II, Komurasaki T, Plowman GD and Stern DF. 10578 ± 10582. (1998). J. Biol. Chem., 273, 11288 ± 11294. Guy PM, Platko JV, Cantley L, Cerione RA and Carraway Roche S, Koegl M, Barone MV, Roussel MF and KL III. (1994). Proc. Natl. Acad. Sci. USA, 91, 8132 ± Courtneidge SA. (1995). Mol. Cell Biol., 15, 1102 ± 1109. 8136. Sainsbury JRC, Farndon JR, Needham GK, Malcolm AJ Hanke JH, Gardner JP, Dow RL, Changelian PS, Brissette and Harris AL. (1987). Lancet, 1, 1398 ± 1402. WH, Weringer EJ, Pollok BA and Connelly PA. (1996). J. Sato K-I, Sato A, Aoto M and Fukami Y. (1995). Biochem. Biol. Chem., 271, 695 ± 701. Biophys. Res. Commun., 210, 844 ± 851. Hynes NE and Stern DR. (1994). Biochim. Biophys. Acta, Segatto O, Pelicci G, Giuli S, Digiesi G, DiFiore PP, 1198, 165 ± 184. McGlade J, Pawson T and Pelicii PG. (1993). Oncogene, Kameda T, Yasui W, Yoshida K, Tsujino T, Nakayama H, 8, 2105 ± 2112. Ito M, Ito H and Tahara E. (1990). Cancer Res., 50, 8002 ± Sepp-Lorenzino L, Eberhard I, Zhenping M, Cho C, Serve 8009. H, Liu F, Rosen N and Lupu R. (1996). Oncogene, 12, Karni R, Jove R and Levitzki A. (1999). Oncogene, 18, 1679 ± 1687. 4654 ± 4662. Sheeld LG. (1998). Biochem. Biophys. Res. Commun., 259, Karunagaran D, Tzahar E, Beerli RR, Chen X, Graus-Porta 27 ± 31. D, Ratzkin BJ, Seger R, Hynes NE and Yarden Y. (1996). Shelly M, Pinkas-Kramarski R, Guarino BC, Waterman H, EMBO J., 15, 254 ± 264. Wang L-M, Lyass L, Alimandi M, Kuo A, Bacus S, Pierce Kaspers GJL, Pieters R, Van Zantwijk CH, De Laat PAJM, JH, Andrews GC and Yarden Y. (1998). J. Biol. Chem., De Waal FC, Van Wering ER and Veerman AJP. (1991). 273, 10496 ± 10505. Br. J. Can., 64, 469 ± 474. Sierke SL, Longo GM and Koland JG. (1993). Biochem. Kern JA, Schwartz DA, Nordberg JE, Weiner DA, Greene Biophys. Res. Commun., 191, 45 ± 54. MI, Torney L and Robinson RA. (1990). Cancer Res., 50, Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A and 5184 ± 5187. McGuire WL. (1987). Science, 235, 177 ± 182.

Oncogene c-Src interaction with HER2/neu in human breast cancer cells AP Belsches-Jablonski et al 1475 Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Villaneuva A, Stockert JC and Armas-Portela R. (1984). Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A and Histochem., 81, 103 ± 104. Press MF. (1989). Science, 244, 707 ± 712. Wallasch C, Weiss FU, Niederfellnew G, Jallal B, Issing W Solto SP, Carraway KL, Prigent SA, Gullick WG and and Ullrich A. (1995). EMBO J., 14, 4267 ± 4275. Cantley LC. (1994). Mol. Cell Biol., 14, 3550 ± 3558. Ware JL, Maygarden SJ, Koontz WW and Strom SC. (1991). Tice DA, Biscardi JS, Nickles AL and Parsons SJ. (1999). Hum. Pathol., 22, 254 ± 258. Proc. Natl. Acad. Sci. USA, 96, 1415 ± 1420. Wilson LK, Luttrell DK, Parsons JT and Parsons SJ. (1989). Toi M, Osaki A, Yamada J and Toge T. (1991). Eur. J. Mol. Cell Biol., 9, 1536 ± 1544. Cancer, 27, 977 ± 980. Tzahar E, Pinkas-Kramarski R, Moyer JD, Klapper LN, Alroy I, Levkowitz G, Shelly M, Henis S, Eisenstein M, Ratzkin BJ, Sela M, Andrews GC and Yarden Y. (1997). EMBO J., 16, 4938 ± 4960.

Oncogene