45

Oncogenic Activating Mutations in the neu/erbB-2 Oncogene Are Involved in the Induction of Mammary Tumors

RICHARD CHAN,a,c WILLIAM J. MULLER,a,b,c,d AND PETER M. SIEGELa,c

aInstitute for Molecular Biology and Biotechnology, bDepartment of Pathology, cDepartment of Biology, McMaster University, Hamilton, Ontario, Canada, L8S 4K1

ABSTRACT: Amplification and overexpression of erbB-2/neu is an important de-terminant in the initiation and progression of human breast cancer. Indeed,transgenic mice that overexpress the neu proto-oncogene heritably developmammary adenocarcinomas. Tumorigenesis in these transgenic strains is asso-ciated with activation of the intrinsic catalytic activity of Neu. In many of thesetumors, activation of Neu occurs as a result of somatic mutations located withinthe transgene itself. Examination of the altered neu transcripts revealed thepresence of in-frame deletions that encode aberrant Neu receptors lacking 5 to12 amino acids within the extracellular domain proximal to the transmem-brane region of Neu. In addition to these deletion mutants we have also detect-ed single point mutations within this juxtatransmembrane region. Themajority of the mutations analyzed affect the one of several conserved cysteineresidues present within this region. Introduction of these activating mutationsinto the wild-type neu cDNA results in its oncogenic conversion. Taken togeth-er, these observations suggest that this cysteine-rich region plays an importantrole in regulating the catalytic activity of Neu.

INTRODUCTION

The progression of a primary mammary epithelial cell to the malignant pheno-type is thought to involve multiple genetic events, including activation of dominant-acting oncogenes and the loss of specific tumor supressor genes. Activation ofcertain tyrosine kinases have been implicated in the malignant progression of a sig-nificant proportion of human breast cancers. For example, amplification and overex-pression of the erbB-2/neu oncogene has been implicated in the initiation andprogression of a large percentage of primary breast cancers.1–4 The neu oncogeneencodes a receptor tyrosine kinase (RTK) belonging to the epidermal growth factorreceptor (EGFR) family. In addition to neu, the EGFR family comprises three otherclosely related genes, including EGFR, erbB-3, and erbB-4.5–7 Interestingly, mem-

dAddress correspondence to W.J. Muller, Institute for Molecular Biology and Biotechnology,McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada, L8S 4K1. Voice:905-525-9140, ext. 2730; fax: 905-521-2955.

e-mail: muller@mcmail.cis.mcmaster.ca

46 ANNALS NEW YORK ACADEMY OF SCIENCES

bers of the EGFR family can interact with each other. For example, Neu is a substratefor the activated EGFR, following stimulation of cells with EGF,8–10 or activatedErbB-4, following stimulation with heregulin.11,12 The observed tyrosine phospho-rylation of Neu by EGFR is thought to be mediated by heterodimerization betweenNeu and EGFR family members, resulting in a high affinity receptor for these mito-genic ligands.9,13 In addition to activation of ErbB-2/Neu by a variety of physiolog-ical ligands, overexpression or mutation of the receptor can result in its oncogenicactivation.14

Direct evidence for the involvement of Neu in the induction of mammary tumorsstems from observations made with transgenic mice expressing a constitutively ac-tivated version of Neu15 in the mammary epithelium.16,17 In several of these trans-genic strains, high level expression of activated Neu resulted in the development ofmultifocal mammary adenocarcinomas that affected every female carrier.16 Consis-tent with these observations, infection of the mammary epithelium of rats with a ret-roviral vector bearing the activated neu cDNA15 also resulted in the rapiddevelopment of multifocal mammary tumors.18 Taken together, these observationssuggest that activated Neu can act as a potent oncogene in the mammary epithelium.

In human breast cancer, examination of primary breast cancer samples has thusfar failed to reveal a comparable transmembrane mutation in ErbB-2.19 These obser-vations argue that elevated expression of the erbB-2 proto-oncogene is the primarymechanism by which ErbB-2 induces malignant transformation. To directly test theoncogenic potential of wild-type Neu in the mammary epithelium, several indepen-dent strains of transgenic mice carrying an MMTV-driven neu proto-oncogene havebeen established.20 In contrast to transgenic mice expressing the activated version ofneu, mammary gland–specific expression resulted in the appearance of focal mam-mary tumors after a long latency period. Biochemical analyses of these mammarytumors revealed that tumorigenesis in these mice was correlated with increased Neuintrinsic tyrosine kinase activity and the appearance of several tyrosine phosphory-lated proteins.20

Because activation of the catalytic activity of Neu can occur through the mutationof a single amino acid in the transmembrane domain,15 we investigated whether thismutation could be detected during tumor progression in these transgenic strains. Tothis end, total RNA from both breast tumors and adjacent morphologically normalmammary epithelium was subjected to reverse transcription followed by the poly-merase chain reaction (RT-PCR). The resulting PCR products were then hybridizedto oligonucleotides corresponding to both the wild-type sequence and transmem-brane point mutation in Neu. The results showed that although there was no evidenceof the point mutation, many of the PCR products derived from the Neu-induced tu-mors possessed deletions.21 Significantly, sequence analyses of the altered productsrevealed that these deletions were located in a confined region of the extracellulardomain and that they encoded functional Neu proteins due to maintenance of theprotein's reading frame. Furthermore, it was demonstrated that these in-frame dele-tions resulted in constitutive activation of the kinase activity of Neu. The observationthat the identified activating mutations reside in a region of Neu not previouslyknown to be involved in its oncogenic activation raises the intriguing possibility thatcomparable mutations within the ErbB-2 protein might be functionally involved inhuman breast cancer.

47CHAN et al.: MAMMARY TUMORS

RESULTS AND DISCUSSION

Previous studies with tumors induced by the expression of neu revealed that 65%of the tumors analyzed displayed evidence of altered transcripts (35 samplesanalyzed21). In addition, these deletions were noted in mammary tumors arising inthree independent transgenic strains expressing the MMTV/neu fusion gene.21 Be-cause the initial riboprobe used in these analyses spans both the extracellular and in-tracellular region, it was unclear whether these RNase protection analyses woulddetect single base pair mutations. To test this possibility, specific antigen riboprobescomplimentary to the extracellular and transmembrane region were hybridized toRNA derived from tumors from the MMTV/neu mice and subjected to RNase pro-tection analyses under stringent conditions. In addition to the previously describeddeletions, this stringent RNase protection analysis revealed additional tumor sam-ples that had not previously demonstrated evidence of alterations (unpublishedobservations).

To precisely define the nature of these mutations, these altered RNA transcriptswere subjected to RT-PCR and the PCR products inserted into plasmid vectors andsubjected to DNA sequence analyses. For each matched pair of samples, several in-dependent subclones were analyzed. The results of these alterations are summarizedin FIGURE 1. Sequence analyses of these PCR products revealed that two types of sin-gle base pair mutations were detected. In one class of mutations, cysteine residueslocated at positions 635, 639, and 647 were converted to a variety of different aminoacids (FIG. 1). A second class of mutations resulted in the conversion of a tryptophanresidue located at position 619, a tyrosine residue at position 621 and a serine residueat 638 to cysteine residues. In both classes of mutations, the observed alterations re-sulted in either a net increase or net loss of a single cysteine residue. The observationthat these point mutations affected the balance of cysteine residues like the previous-ly described deletions suggested21 that these alterations may be directly involved inthe activation of the intrinsic catalytic domain of Neu. Indeed, previous observationshave demonstrated that insertion of several of these deletions in an otherwise wild-type neu cDNA resulted in its oncogenic activation.21 To confirm that the otherdeletion and insertion mutants behaved in a similar manner, three of the deletion mu-tants and the single insertion mutant were placed into a wild-type neu expressioncassette and tested for their capacity to transform Rat-1 fibroblasts. Consistent withprevious observations,21 expression of these altered Neu receptors was capable oftransforming Rat 1 fibroblasts whereas expression of the wild-type neu cDNA wasnot (unpublished observations).

Because the transforming activity of Neu is closely correlated with its activationof intrinsic tyrosine kinase activity,23–25 we also measured the state of tyrosine phos-phorylation of Neu in cell lines transformed by these altered cDNAs. The resultsshowed that tyrosine-phosphorylated Neu was detected in cell lines expressing thesealtered neu cDNAs but was not detected in cell lines expressing elevated levels ofwild-type Neu.21 Thus the increase in the levels of tyrosine-phosphorylated Neu inthese cell lines likely reflects the catalytic activation of these altered receptors. Theabove observations suggest that activation of the Neu tyrosine kinase is a pivotal stepin the initiation of mammary tumorigenesis and occurs primarily through somaticmutations in this transgenic mouse model of human breast cancer. The observation

48 ANNALS NEW YORK ACADEMY OF SCIENCES

that the identified mutations are located in a region of Neu previously not known tobe involved in its oncogenic activation raises the intriguing possibility that compa-rable mutations might be detected in human breast cancers.

It is striking that the mutations we have detected reside in a relatively narrow re-gion of Neu located outside the transmembrane domain (FIG. 1). In fact, this regionin Neu appears to be conserved among the other members of the EGFR family, in-cluding EGFR, ErbB-3, and ErbB-4.22 Specifically, the region of all four knownfamily members contains five cysteine residues that are perfectly conserved.22

Moreover, the amino acids located adjacent to these cysteine residues exhibit a highdegree of similarity.22 Interestingly, with the exception of the deletion detected in asingle tumor, the alterations that have been sequenced thus far have removed or al-tered at least one of the conserved cysteine residues.22 Furthermore, in the mutantNeu molecule possessing the three-amino acid insertion, one of the inserted aminoacids is a cysteine residue. These data suggest that the balance of cysteine residueswithin this region of Neu may play an important role in activation of the receptor.

Although these observations suggest that these cysteine residues may be involvedin the oncogenic activation of the Neu receptor, the precise molecular mechanism bywhich this occurs remains to be addressed. In this regard, it has been reported thattransforming Neu mutants possessing the point mutation in the transmembrane do-

FIGURE 1. Activation of Neu can occur through single amino acid mutations in thejuxtatransmembrane region. Amino acid sequence alignment of altered Neu receptors. Ineach case, the balance of cysteine residues is directly affected by either conversion of cys-teine residue to another amino acid or conversion of unrelated amino acid to a cysteineresidue.

49CHAN et al.: MAMMARY TUMORS

main demonstrate an increased propensity to homodimerize.22,26–28 Indeed, immu-noprecipitation analyses of this mutant Neu receptor under both reducing and non-reducing conditions revealed that the activated Neu species could be detected as amultimeric complex under nonreducing conditions, which was converted to a mono-meric species under reducing conditions.25 By contrast, the wild-type Neu receptorremained a monomer species under both reducing and nonreducing conditions.These observations suggest that receptor dimerization is occurring through the for-mation of disulfide bonds.

Given that many of the deletions and insertions mutants affect different cysteineresidues, it is conceivable that, like the transmembrane point mutation, alteration ofthese cysteine residues in the mutants promotes receptor dimerization through theformation of cysteine disulfide bonds. Indeed we have recently demonstrated thatthese altered receptors dimerize in a manner dependent on the formation of disulfidebonds.22 It is possible that these deletions and insertions of cysteine residues disruptthe normal cysteine pairing that occurs in the wild-type receptor. As a consequencethis unpaired cysteine residue would be free to participate in disulfide bond forma-tion with another altered receptor.

Consistent with this proposed model, it has recently been demonstrated that in in-herited forms of endocrine neoplasia type 2A, a single mutation in a cysteine residuelocated in the cysteine rich juxtatransmembrane domain of the Ret RTK is responsi-ble for its oncogenic activation. Similar to the Neu deletion mutants, the generationof this cysteine imbalance results in the constitutive dimerization of the Ret RTK ina manner dependent on the formation of disulfide bonds.29,30 In addition to thesestudies, dimerization of the hematopoietic/cytokine receptor superfamily is thoughtto involve cysteine disulfide bonding. For example, replacement of specific aminoacids with cysteine residues in the juxtatransmembrane domain of either the throm-bopoietin receptor or the erythropoietin receptor results in the formation of disul-fide-linked dimers that constitutively activate these receptors.31–33 In addition, it hasrecently been demonstrated that disruption of disulfide bonding by the administra-tion of reducing agents to the external media of cultured cells can interfere with sig-naling from the thrombopoietin receptor by preventing dimerization of thereceptor.30 Indeed, we have recently demonstrated that addition of 2-mercaptoetha-nol to the external media of cells can interfere with the transforming potential ofthese altered Neu receptors in a dose-dependent manner.22 Taken together these ob-servations suggest that this mode of receptor dimerization may be shared by otherreceptor systems.

The data generated thus far suggests that the occurrence of these deletions withinNeu in this transgenic mouse model are likely directly involved in the induction ofmammary tumors expressing the MMTV/wild-type neu fusion gene. To directly as-sess the significance of these activating mutations in Neu, we have recently estab-lished transgenic mice that express certain of these activated neu alleles under thetranscriptional control of the MMTV promoter. Although we are still in the earlystages of characterization of these lines, preliminary observations suggest that incontrast to the parental strains expressing the wild-type neu gene these mice developmultifocal mammary tumors with a relatively shorter latency period (Siegel andMuller, unpublished observations). However, unlike the transgenic mice expressingthe transmembrane point mutation that exhibits rapid progression phenotype,16 these

50 ANNALS NEW YORK ACADEMY OF SCIENCES

mice possess preneoplastic lesions. One potential explanation for these observationsis that in the latter strains the level of Neu-associated tyrosine kinase activity is be-low a certain threshold required for the rapid transformation of the mammary epithe-lial cell, whereas the rapid progression strains exceed this critical threshold. Inaddition to the elevated expression these activated neu alleles, we have frequentlydetected elevated expression of the erbB-3 (Siegel and Muller, unpublished observa-tions). Indeed, erbB-3 is tyrosine phosphorylated in these tumors (Siegel and Muller,unpublished observations). Given the fact that erbB-3 does not possess catalytic ac-tivity, it is likely that the observed tyrosine phosphorylation of ErbB-3 derives fromtransphosphorylation by these activated neu alleles. Future experiments with thesevarious strains should allow this hypothesis to be rigorously tested.

The frequent occurrence of activating mutations in the juxtatransmembrane ofNeu raises the possibility that comparable mutations in ErbB-2 might also be in-volved in the genesis of human breast cancer. We are currently examining humanbreast cancer biopsies for the occurrence of these mutations to test this hypothesis.Given that activating mutations in Ret RTK can result in the hereditary predisposi-tion to development of endocrine neoplasias, it is conceivable that a comparablegerm-line mutation in erbB-2 may, in part, be responsible the development of hered-itary breast and ovarian cancers.

Finally, the observation that the region that is deleted in Neu appears to be highlyconserved among the different EGFR family members raises the possibility that mu-tational activation of these other members may also be involved in the genesis of hu-man breast cancer. Indeed, there is considerable evidence to suggest that these otherEGFR family members are overexpressed in human breast cancer.6,7 Although no di-rect evidence exists for mutational activation of these EGFR family members in hu-man cancers, it has recently been demonstrated that the Caenorhabditis elegansEGFR homologue, LET-23, can be activated by mutation of a single cysteine residuein the extreme amino terminus of the protein.34 Given the potential of the variousEGFR family members to heterodimerize, it is conceivable that mutational activa-tion of any one of the EGFR family can activate the other family members. Futurestudies should allow these issues to be addressed.

REFERENCES

1. KING, C.R., M.H. KRAUS & S.A. AARONSON. 1985. Science 229: 974–976.2. SLAMON, D.J., G.M. CLARK, S.G. WONG, W.J. LEVIN, A. ULLRICH & W.L. MCGUIRE.

1987. Science 235: 177–182.3. SLAMON, D.J., W. GODOLPHIN, L.A. JONES, J.A. HOLT, S.G. WONG, D.E. KEITH, W.J.

LEVIN, S.G. STYART, G. UDOVE, A. ULLRICH & M.F. PRESS. 1989. Science 244:707–712.

4. VAN DE VIJER, M., J. PEYERSE, W.J. MOOI, P. WISMAN, J. LOMANS, O. DALESIO & R.NUSSE. 1988. N. Engl. J. Med. 319: 1239–1245.

5. ULLRICH, A., L. COUSSENS, J.S. HAYFLICK, T.J. DULL, A. GRAY, A.W. TAM, J. LEE,Y. YARDEN, T.A. LIBERMANN, J. SCHLESSINGER, J. DOWNWARD, E.L.V. MAYES, N.WHITTLE, M.D. WATERFIELD & P.H. SEEBURG. 1984. Nature 309: 418–425.

6. KRAUS, M.H., I. ISSING, T. MIKI, N.C. POPESCU & S.A. AARONSON. 1989. Proc. Natl.Acad. Sci. USA 86: 9193–9197.

7. PLOWMAN, G., G. WHITNEY, M. NEUBAUER, J. GREEN, V. MCDONALD, G. TODARO &M. SHOYAB. 1990. Proc. Natl. Acad. Sci. USA 87: 4905–4909.

8. STERN, D.F. & M.P. KAMPS. 1988. EMBO J. 7: 995–1001.

51CHAN et al.: MAMMARY TUMORS

9. WADA, T., X. QIAN & M.I. GREENE. 1990. Cell 61: 1339–1347.10. QIAN, X., S.J. DECKER & M.I. GREENE. 1992. Proc. Natl. Acad. Sci. USA 89: 1330–

1334.11. RIESE II, D.J., T.M. VAN RAAJI, G.D. PLOWMAN, G.C. ANDREWS & D.F. STERN. 1995.

Mol. Cell. Biol. 15: 5770–5776.12. BEERLI, R.R., D. GRAUS-PORTA, K. WOODS-COOK, X. CHEN, Y. YARDEN & N.E.

HYNES. 1995. Mol. Cell. Biol. 15: 6496–6505.13. GOLDMAN, R., R. BEN LEVY, E. PELES & Y. YARDEN. 1990. Biochemistry 29: 11024–

11028.14. HYNES, N.E. & D.F. STERN. 1994. Biochim. Biophys. Acta 1198: 165–184.15. BARGMANN, C.I., M-C. HUNG & R.A. WEINBERG. 1986. Cell 45: 649–657.16. MULLER, W.J., E. SINN, R. WALLACE, P.K. PATTENGALE & P. LEDER. 1988. Cell 54:

105–115.17. BOUCHARD, L., L. LAMARRE, P.J. TREMBLAY, P. JOLICOEUR. 1989. Cell 57: 931–936.18. WANG, B., W.J. KENNAN, J. YASUKAWA-BARNES, M.J. LINDSTROM & M.N. GOULD.

1991. Cancer Res. 51: 5649–5654.19. LEMOINE, N.R., S. STADDON, C. DICKSON, D.M. BARNES & W.J. GULLICK. 1990.

Oncogene 5: 237–239.20. GUY, C.T., M.A. WEBSTER, M. SCHALLER, T.J. PARSONS, R.D. CARDIFF & W.J.

MULLER. 1992. Proc. Natl. Acad. Sci. USA 89: 10578–10582.21. SIEGEL, P.M., D.L. DANKORT, W.R. HARDY & W.J. MULLER. 1994. Mol. Cell. Biol.

14: 7068–7077.22. SIEGEL. P.M. & W.J. MULLER. 1996. Proc. Natl. Acad. Sci. USA 93: 8878–8883.23. BARGMANN, C.I. & R.A. WEINBERG. 1988. Proc. Natl. Acad. Sci. USA 85: 5394–

5398.24. STERN, D.F., M.P. KAMPS & H. CAO. 1988. Mol. Cell. Biol. 8: 3969–3973.25. WEINER, D.B., Y. KOKAI, T. WADA, J.A. COHEN, W.V. WILLIAMS & M.I. GREENE.

1989. Oncogene 4: 1175–1183.26. WEINER, D.B., J. LIU, J.A. COHEN, W.V. WILLIAMS & M.I. GREENE. 1989. Nature

339: 230–231.27. BRANDT-RAUF, P.W., S. RACKOVSKY & M.R. PINCUS. 1990. Proc. Natl. Acad. Sci.

USA 87: 8660–8664.28. CAO, H., L. BANGALORE, B.J. BORMANN & D.F. STERN. 1992. EMBO J. 11: 923–932.29. SANTORO, M., F. CARLOMAGNO, A. ROMANO, D.P. BOTTARO, N.A. DATHAN, M.

GRIECO, A. FUSCO, G. VECCHIO, B. MATOSKOVA, M.H. KRAUS & P.P. DIFIORE.1995. Science 267: 381–383.

30. ASAI, N., T. IWASHITA, M. MATSUYAMA & M. TAKAHASHI. 1995. Mol. Cell. Biol. 15:1613–1619.

31. ALEXANDER, W.S., D. METCALF & A.R. DUNN. 1995. EMBO J. 14: 5569–5578.32. WATOWICH, S.S., A. YOSHIMURA, G.D. LONGMORE, D.J. HILTON, Y. YOSHIMURA &

H.F. LODISH. 1992. Proc. Natl. Acad. Sci. USA 89: 2140–2144.33. WATOWICH, S.S., D.J. HILTON & H.F. LODISH. 1994. Mol. Cell. Biol. 14: 3535–3549.34. KATZ, W.S., G.M. LESA, D. YANNOUKAKOS, T.R. CLANDININ, J. SCHLESSINGER &

P.W. STERNBERG. 1996. Mol. Cell. Biol. 16: 529–537.