Sam68 Haploinsufficiency Delays Onset of Mammary Tumorigenesis

Oncogene (2008) 27, 548–556 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Sam68 haploinsufﬁciency delays onset of mammary tumorigenesis and metastasis

S Richard1,2,3, G Vogel1,2,3, M-E´ Huot1,2,3, T Guo1,2,3, WJ Muller2,4,5 and KE Lukong1,2,3

1Terry Fox Molecular Oncology Group and the Bloomfield Center for Research on Aging, Sir Mortimer B Davis Jewish General Hospital, Lady Davis Institute for Medical Research, Montreál, Que´bec, Canada; 2Department of Medicine, McGill University, Montreál, Que´bec, Canada; 3Department of Oncology, McGill University, Montreál, Que´bec, Canada; 4Molecular Oncology Group, McGill University Health Centre, Montreál, Que´bec, Canada and 5Department of Biochemistry, McGill University, Montreál, Que´bec, Canada

The Src-associated substrate in mitosis Sam68 is a KH domain, proline-rich SH3 (Src homology domain 3)- type RNA-bindingprotein known to be a substrate of binding and SH2-interacting tyrosine-rich motifs (Lu- numerous tyrosine kinases, and often referred to as a kong and Richard, 2003). Sam68 is a substrate of breast STAR (signal transduction activator of RNA) protein. tumor kinase (BRK; Derry et al., 2000), a kinase Herein, we observed that Sam68-null mice display overexpressed in about 60% of breast tumors (Kamalati mammary gland and the uterine development defects. et al., 1996). The tyrosine phosphorylation of Sam68 Moreover, we report that Sam68 haploinsufficiency may be a marker for certain cancers where tyrosine impedes mammary tumor onset in vivo driven by the kinases are induced. Indeed, we have shown that Sam68 potent mammary-targeted polyoma middle T-antigen is a substrate downstream of the epithelial growth factor (MMTV-PyMT) oncogene. The effect was cell autono- (EGF; Lukong et al., 2005) and Sam68 tyrosine mous as the Sam68 knockdown in PyMT-transformed cell phosphorylation is elevated in human breast tumors lines also delayed tumorigenesis and metastasis formation tissues and cell lines (Babic et al., 2004; Lukong et al., in nude mice. Interestingly, tumor extracts isolated from 2005). Consistent with these findings is the elevated PyMT/Sam68 þ /À mice compared with PyMT/Sam68 þ / þ expression of Sam68 observed in prostate cancer tissues mice contained activated Src and FAK kinases. These (Busa et al., 2007). These findings suggest that tyrosine findings suggest that Sam68 may be a modulator of phosphorylation of Sam68 and its expression are tyrosine kinase activity in vivo and a signaling require- required for tumorigenesis. Conversely, the reduced ment for mammary tumorigenesis and metastasis. expression of Sam68 in NIH3T3 cells using a gene trap Oncogene (2008) 27, 548–556; doi:10.1038/sj.onc.1210652; strategy and antisense-induced neoplastic transfor- published online 9 July 2007 mation suggests Sam68 may have tumor suppressor properties (Liu et al., 2000). However, the Sam68À/À Keywords: Sam68; RNA binding; mammary tumorigenesis; mice live to old age (B2 years) and were not prone to Src; metastasis; signaling tumor formation, suggesting that Sam68 is not a tumor suppressor in vivo (Richard et al., 2005).

Results Introduction Sam68À/À mice have defects in breast and Sam68, the Src-Associated substrate during Mitosis of uterine development 68 kDa, is a member of the STAR (Signal Transduction Herein, we investigated whether the absence of Sam68 Activator of RNA) family of RNA-binding proteins prevented or accelerated tumorigenesis in vivo, and we proposed to link signaling cascades to RNA metabolism employed the mammary-targeted polyoma middle T (Vernet and Artzt, 1997; Lukong and Richard, 2003). antigen (MMTV-PyMT) mouse model because PyMT These proteins share at least three functional motifs utilizes Src for signaling and induces rapid breast tumors including a K (hnRNP K) homology (KH) RNA- that resemble human breast cancers (Lin et al., 2003). binding domain embedded in a larger domain termed Whole-mount analysis and hematoxylin/eosin staining the GSG (GRP33, Sam68, GLD-1) or the STAR of sections of the fourth inguinal mammary glands of 6- and 12-week-old virgin females from Sam68 þ / þ , Correspondence: Professor S Richard, Segal Cancer Centre, Lady Sam68 þ /À and Sam68À/À mice were performed to Davis Institute for Medical Research, 3755 Coˆ te Ste-Catherine Road, examine mammary development (Figure 1). Both wild- Montre´ al, Que´ bec, Canada H3T 1E2. E-mail: [email protected] type and heterozygous Sam68 mice displayed normal Received 27 September 2006; revised 31 May 2007; accepted 1 June 2007; and comparable ductal outgrowth (Figure 1a, top and published online 9 July 2007 middle panels) consistent with Sam68 þ /À females bearing Sam68 modulates signaling in mammary tumorigenesis S Richard et al 549

Figure 1 Sam68À/À mice display a mammary gland defect. (a) Whole-mount analysis of mammary glands from Sam68 mice is shown (n>4). Scale bar 2 mm and the lymph node (LN) is depicted. (b) Terminal-end buds and number of branchings were counted and expressed as the mean7s.d. of the mean. *P-valueo0.01. (c) Histological examination by hematoxylin/eosin of 6- and 12-week-old mammary glands from Sam68 þ / þ , Sam68 þ /À and Sam68À/À mice. Scale bar, 100 mm.

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 550 normal litters and lactating normally (data not shown). MMTV-PyMT transgenic as a model to assess the effect In contrast, Sam68À/À mice had a striking defect in the of Sam68 haploinsufficiency in mammary tumor deve- development of the ductal outgrowth observed in 6- lopment and tumor incidence. Sam68À/À mice displayed week-old mice, and this difference was less obvious and an overt mammary defect (Figure 1), and, therefore, partially recovered in 12-week-old mice (Figure 1a, could not be used in these studies. Therefore, Sam68 þ /À lower panels). The 6-week-old Sam68À/À mice had mice were bred with MMTV-PyMT transgenic mice and approximately three times less terminal end buds 6-week-old female offspring palpated twice a week in the compared with Sam68 þ / þ and Sam68 þ /À littermate mice mammary glands for the development of tumors. The (Figure 1b). Adult virgin female mice of 12-week-old animals were killed as soon as one tumor reached mice had fully developed ductal trees; however, the 1.5 cm3 in size and the data expressed using a Kaplan– ductal outgrowth density was lower in whole mounts of Meier analysis. The absence of one allele of Sam68 Sam68À/À mice compared to Sam68 þ / þ and Sam68 þ /À in PyMT/Sam68 þ /À mice (n ¼ 14) delayed tumor deve- mice, as assessed by counting the number of branches in lopment to 122 days as opposed to 80 days observed a given area (Figure 1b). Histological analysis of the in PyMT/Sam68 þ / þ (Figure 3a). The tumor onset in mammary glands showed a normal epithelium with wild-type background was approximately 11 weeks, fewer cross-sections of mammary ducts in the 6-week- consistent with a C57BL6/FVB mice mixed back- old Sam68À/À female mice compared to littermate ground (Cheng et al., 1998). To assess tumor multi- controls or 12-week-old Sam68À/À mice, as predicted plicity, cohorts of PyMT/Sam68 þ / þ (n ¼ 15) and with the reduced ductal outgrowth in 6-week-old PyMT/Sam68 þ /À (n ¼ 10) female mice were analysed. Sam68À/À mice (Figure 1c). We also noted a marked decrease in tumor multiplicity We next investigated the gross morphological analysis in PyMT/Sam68 þ /À mice compared to PyMT/Sam68 þ / þ of the ovaries and uteri derived from 6- and 12-week-old mice, with an average of 2.5 and 4.92 tumors per Sam68 mice. The gross morphology of the uteri from 6- mouse, respectively (Figure 3b). These data indicate that week-old Sam68À/À females appeared atrophic and less Sam68 haploinsufficiency results in both increased developed compared with Sam68 þ / þ and Sam68 þ /À tumor latency and decreased tumor multiplicity in mice; however, this difference was not visible in 12- the MMTV-driven PyMT mice. These findings suggest week-old Sam68À/À mice (Figure 2a). Histological that Sam68 is required for PyMT-induced mammary analysis revealed a normal endometrium with a well- tumorigenesis. organized outer layer of smooth muscle in Sam68 mice (Figure 2b). However, cross-sectioning of the uteri revealed that the uteri of Sam68À/À mice are smaller Src and FAKare activated in PyMT/Sam68 þ /À tumors than those observed in Sam68 þ / þ and Sam68 þ /À mice PyMT-mediated tumorigenesis involves recruitment (Figure 2b). Histologically, no discernable difference is and/or activation of several signaling molecules includ- observed between ovaries from 6- and 12-week-old ing c-Src, phosphatidylinositol 3-kinase (PI-3K), Akt Sam68 mice (Figure 2c). Primary follicles and corpora and Ras (Andrechek and Muller, 2000). As an lutea are present, consistent with the fact that Sam68À/À intracellular target of Src, Sam68 itself may be a female mice are fertile and give birth to reduced litter downstream effector of the PyMT pathway. To deter- sizes (Richard et al., 2005). Taken together, our mine whether the decreased tumorigenicity in PyMT/ observations show that Sam68 deficiency delays the Sam68 þ /À mice is associated with altered expression normal development of the uterus. and/or activation of selected PyMT-coupled signaling molecules, we performed immunoblotting on tumor tissues from PyMT/Sam68 þ / þ and PyMT/Sam68 þ /À Sam68 haploinsufficiency delays mammary tumor onset mice. Tumor extracts from PyMT/Sam68 þ /À mice and multiplicity in PyMT transgenic mice showed B50% decrease in the level of Sam68 expression Mammary tumorigenesis phenotypically similar to human consistent with heterozygosity (Figure 4, top panel). breast cancers has been induced in mice by targeted Interestingly, partial loss of Sam68 consistently resulted overexpression of oncogenes including the PyMT driven in B50% reduced expression of PyMT (Figure 4). We by the MMTV promoter (Lin et al., 2003). PyMT leads analysed the expression and activation of PI-3K, AKT to the activation of signaling cascades that parallel the and c-Src as well as the Src substrate FAKfrom tumor activation of Neu/Erb-B2 in human mammary carcino- cell lysates of PyMT/Sam68 þ / þ and PyMT/Sam68 þ /À mas (Muller et al., 1988; Guy et al., 1992; Lucchini et al., mice. We examined at least six tumors for each genotype 1992). All female virgin MMTV-PyMT transgenic mice and representative immunoblots are shown demonstrat- rapidly develop multifocal mammary adenocarcinomas ing that PyMT/Sam68 þ /À mice did not affect the with 100% penetrance and short latency. Other studies constitutive expression or activation of PI-3Kand Akt have indicated that PyMT signals through Src in poten- (Figure 4). Interestingly, the phosphorylation of Src tiating mammary tumor development (Courtneidge (pY416), FAK(pY397, pY576 and pY577) was elevated et al., 1991; Muthuswamy and Muller, 1994). Sam68 in all PyMT/Sam68 þ /À tumor lysates compared to itself is an Src substrate and probably plays an PyMT/Sam68 þ / þ tumor lysates (Figure 4). FAKY397 adapter role downstream of the Src pathway (Lukong is the autophosphorylation site that drives the initial and Richard, 2003). To examine whether Sam68 is a activation of the kinase, while Y576 and Y577 are signaling requirement for PyMT, we employed the preferentially phosphorylated by Src, resulting in

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 551

Figure 2 Uterine and ovarian phenotypes in Sam68À/À female mice. (a) Uteri from Sam68 mice (n>4). Scale bar, 5 mm. (b) Histological examination by hematoxylin/eosin of uteri from Sam68 mice. The smooth muscle layer (SM) and lumen (LU) are shown. Scale bar, 100 mm. (c) Histological examination by hematoxylin/eosin of 6- and 12-week-old ovaries from Sam68 mice. The follicle (F) and the corpus luteum (CL) are shown. Scale bar, 100 mm. maximal FAKactivation (Cohen and Guan, 2005). Sam68 knockdown limits PyMT-induced tumor growth b-Actin and PRMT1 immunoblots were used as controls Our findings that haploinsufficiency of Sam68 impedes for equal loading. These findings show that the mammary tumor onset and tumor multiplicity in reduction of Sam68 in an in vivo setting leads to the PyMT mice prompted us to examine whether this increased presence of active marks on Src (pY416) and phenotype is cell autonomous. The Db7 cell line is a FAK(pY397, pY576, pY577) tyrosine kinases. serially transplantable, MMTV-PyMTY315F/Y322F mammary

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 552

Figure 3 Effect of Sam68 heterozygosity on tumor development and incidence. (a) Kaplan–Meier analysis indicating that PyMT/ Sam68 þ /À (n ¼ 14) female mice have signiﬁcantly delayed onset of palpable mammary tumors compared to PyMT/ Sam68 þ / þ (n ¼ 17). (b) Tumor multiplicity was reduced in PyMT/Sam68 þ /À (n ¼ 10) compared to PyMT/Sam68 þ / þ mice (n ¼ 15). *P-valueo0.0015.

adenocarcinoma-derived cell line (Borowsky et al., Sam68 knockdown results in decreased lung metastasis 2005). Using short-hairpin RNA that targets mouse We next extended our studies to investigate the Sam68 (Sam68sh), we achieved over 90% knockdown of involvement of Sam68 in PyMT-induced metastasis. Sam68 expression, as evidenced by immunoblotting To this end, we used Met-1 cells known to have an analysis with b-actin and c-Src as loading controls elevated metastatic potential (Borowsky et al., 2005). (Figure 5a). Next, we injected 106 cells of either Db7- We thus generated control (pSUPER) and Sam68 pSUPER or Db7-Sam68sh cells into the mammary fat knockdown (Sam68sh) cells in Met-1 cells. A knock- pad of athymic mice (n>11). After 3 weeks, tumors down of B90% was obtained using b-actin and PRMT5 were isolated and weighed. We observed an B35% as loading controls (Figure 6a). The levels of PyMT decrease in tumor size and weight generated from Db7- protein remained equivalent between pSUPER and Sam68sh-injected mice, as compared to the Db7- Sam68sh Met-1 cells and the addition of the proteasome pSUPER control (Figure 5b). Immunoblot analysis of inhibitor (MG132) caused the appearance of a slower the lysates from the Db7 tumors conﬁrmed that Sam68 migrating form of PyMT, but there was no difference knockdown was maintained during the development of between pSUPER and Sam68sh cells (Figure 6a). The the tumors except for one (Figure 5c, lane 7). PRMT1 ﬁndings suggest that a slower migrating PyMT protein was used as a loading control in this experiment. These may represent ubiquitylated PyMT. To examine results indicate that depletion of Sam68 in Db7 cells whether the reduced levels of Sam68 affect the meta- limits tumor burden in mice, and demonstrate that static potential of the Met-1 cells, we performed an Sam68 functions in cell autonomous manner to regulate in vivo metastatic analysis. Athymic mice were hence PyMT-induced mammary tumorigenesis. injected i.v. with pSUPER Met-1 cells (n ¼ 12) or

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 553

Figure 5 Sam68 depletion in PyMT-derived cells delays tumor growth in vivo.(a) Immunoblot analysis of stable Db7 cell lines transfected with pSUPER or pSam68sh. Stably transfected cells were subjected to immunoblot analysis for Sam68, b-actin and Src expression. (b) Db7-pSUPER or Db7-Sam68sh cells injected in the fourth inguinal mammary fat pad of athymic mice were monitored for tumorigenesis over 3 weeks. The error bars indicate the s.d. of the mean (n>11). *P-valueo0.01. (c) Four representative tumor extract lysates were immunoblotted with anti-Sam68 and anti- PRMT1 (loading control) antibodies.

metastasis observed in vivo is due to a lack in cell migration in Sam68-deﬁcient Met-1 cells.

Discussion

Sam68À/À female mice displayed mammary gland development and uterine atrophy at prepuberal age where estrogen, growth hormone, growth factors (IGF-1) and corticoids play a major role (Fendrick et al., 1998; Hovey et al., 2002). The phenotype observed in the Sam68À/À mice could be explained by a lack of estrogen Figure 4 p-Src and p-FAKare elevated in tumor extracts from at the prepuberal age. However, measurement of the PyMT/Sam68 þ /À mice. Portions of tumors isolated from 3-month- plasma estrogen levels revealed that estrogen levels were old PyMT/Sam68 þ / þ and PyMT/Sam68 þ /À mice were lysed and significantly elevated (Po0.01 Sam68À/À versus Sam68 þ / subjected to immunoblot analysis as indicated. Sam68 and PyMT þ ; n ¼ 4) in 6-week-old Sam68À/À mice (29.7374.71 pg/ expression were quantitated by densitometric scanning, and the þ / þ 7 relative values shown below the immunoblots. b-Actin, PRMT1, ml) compared to Sam68 (17.57 2.28 pg/ml) and Src and AKT served as controls for equal loading. Sam68 þ /À (23.1875.64 pg/ml) mice. Estrogen levels remained elevated in 4-month-old Sam68À/À female mice but the difference was not statistically significant (Richard et al., 2005). These observations are counter- Sam68sh Met-1 cells (n ¼ 12) and 40 days later, the lungs intuitive but may reflect a lack of estrogen responsive- were harvested and foci counted and scored under a ness as observed in estrogen receptor-a null mice that stereomicroscope. The number of foci in the lungs of have increased plasma estrogen levels and display mice injected with the Sam68sh Met-1 cells was mammary gland and uterine development defects significantly reduced by B5-fold compared to wild-type (Lubahn et al., 1993; Schomberg et al., 1999). Future Met-1 and pSUPER Met-1 cells (Figure 6b). A decrease experiments are required to determine the hormonal in metastatic potential is often correlated with a defect responsiveness of Sam68À/À mice. c-Src null mice in cell migration and haptotactic migration was analysed (Soriano et al., 1991) known to regulate ERa expression using a Transwell chamber assay. Sam68sh Met-1 also display defects in mammary glands, ovarian and showed a significantly impaired migration compared to uterine development (Kim et al., 2005), demonstrating the parental Met-1 cell line and the pSUPER Met-1 cells that Src and Sam68 null mice have similar target tissues. (Figure 6c). These findings suggest that the lack of Our findings suggest that in humans a delay in the onset

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 554

Figure 6 Reduction of Sam68 expression decreases PyMT-induced lung metastasis. (a) Stably selected Met-1 cells with control pSUPER or Sam68sh were treated with the proteasome inhibitor MG132 for 6 h and immunoblotted with b-actin (loading control), Sam68, PRMT5 (loading control) and PyMT antibodies. (b) Phase-contrast images of pSUPER and Sam68sh Met-1 cells. (c) Nude mice were injected i.v. with 2 Â 105 pSUPER or Sam68sh Met-1 cells (n>12). Lung foci were counted and expressed as mean þ s.d. of the mean. (d) Haptotactic migration assays of pSUPER and Sam68sh Met-1 cells.

of puberty may harbor loss-of-function mutations in the cells suggests that PyMT is not a direct protein or RNA sam68 gene. target of Sam68. The reduced levels of PyMT expression in tumor How is it possible to have activated Src and FAK extracts from PyMT/Sam68 þ /À may explain the reduced and observe reduced tumorigenesis and metastasis? mammary gland tumorigenesis observed. However, Although c-Src is required for efficient mammary reduced PyMT expression was not observed in PyMT tumorigenesis, multifocal mammary tumors and metas- cell lines harboring Sam68sh, and yet tumor formation tasis (Guy et al., 1994), mammary epithelial expression and metastasis in nude mice was significantly delayed. of an activated c-Src oncogene in transgenic mice Collectively, these findings suggest that Sam68 is likely induces mammary hyperplasia but not multifocal a signaling requirement for PyMT-mediated tumori- mammary tumors in PyMT strains (Webster et al., genesis and metastasis. It is possible that the increased 1998). It appears from our studies therefore that active Src activity observed in tumor extracts from PyMT/ Src alone is not sufficient for tumor induction, and Sam68 þ /À mice necessitates an in vivo tumor environ- may require Sam68 for a comprehensive induction of ment to promote the ubiquitylation and degradation of tumorigenesis. Thus, in the case of PyMT/Sam68 þ /À PyMT by the protesome pathway. The fact that we tumors in the present study, Src activation alone may observed slower migrating PyMT species with the not be sufficient to induce rapid tumor progression. proteasome inhibitor MG132 makes this a viable How can the haploinsufficiency of Sam68 result in possibility. It is well documented that Src family kinases active Src kinase activity in a PyMT expressing cell? promote the ubiquitylation and proteasomal degrada- Sam68 is hypothesized to play an adapter role in a tion of certain of their substrates (Fujita et al., 2002; Bao multitude of signaling pathways since it have been et al., 2003). The fact that the protein levels of PyMT shown to form SH2- and/or SH3-domain interactions did not change between pSUPER and Sam68sh in Met-1 with various signaling molecules including GRB2

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 555 (Lukong and Richard, 2003). Several studies have TACTTCACAAGGG-30; reverse, 50-GGAAAGTCACTAG shown that disruption of certain genes leads to GAGCAGGG-30. Once a tumor reached a volume of abnormal mammary development and tumorigenesis 1.5 cm3, assessed by a caliper, the mice were killed. Mammary, (Andrechek and Muller, 2000). For example, the ovarian and uterine tissues were fixed in 10%-buffered haploinsufficiency of GRB2 delays PyMT-induced formalin and embedded in paraffin for sectioning. The sections were then stained with hematoxylin and eosin staining by the tumor progression (Cheng et al., 1998), suggesting that Department of Pathology at the Jewish General Hospital Sam68 may indeed fulfill a signaling requirement for (Montreal, Quebec, Canada). PyMT-mediated tumorigenesis. As Sam68 is a substrate and interacting partner of Src, it may function in a similar fashion as was reported for RACK1 (Mamidi- Cell culture pudi et al., 2007), and it remains to be examined whether For mammary fat pad injection, inguinal (n ¼ 4) mammary 6 Sam68 RNA-binding activity and that of the splicing pads of athymic mice were injected with 10 Db7pSUPER or regulator SRm160 are required for this function and Db7Sam68sh cells and tumors weighed after 4 weeks. For the metastasis analysis, 2 Â 105 pSUPERMet-1 or Sam68shMet-1 whether CD44 alternative splicing contributes to the cells were injected i.v. in female athymic mice, and after 40 observed phenotype (Matter et al., 2002; Cheng and days lung foci assessed. Haptotactic migration was performed Sharp, 2006). with the Corning Transwell with a serum gradient and the cells The mechanism for Src activation and inhibition of were permitted to migrate overnight, stained with crystal violet PyMT expression in PyMT/Sam68 þ /À tumors is unclear. and counted. RNA interference for Sam68sh was performed as But it seems to necessitate an in vivo tumor environment described previously (Richard et al., 2005). since asynchronously growing PyMT-transformed cells demonstrate normal levels of PyMT (Figure 6). How do Antibodies and estrogen levels all these events account for the delay in the onset of Antibodies recognizing Sam68, PRMT1, PRMT5, p85 subunit þ /À mammary tumors in the PyMT/Sam68 ? The middle of PI-3 kinase, phospho-AKT, AKT, phospho-Src (Y416) ‘T’ has been tightly associated with transforming were from Upstate/Millipore (Billerica, MA, USA). The b- capabilities, and is able to engage both the PI-3 kinase actin antibody was purchased from Sigma (St Louis, MO, and MAPKcascades (Dilworth, 2002). We speculate USA) and the phospho-FAK(Y576; Y577; Y397) and first that the increase activation of Src is not paralleled phospho-paxillin were from Biosource (Camarillo, CA, with an activation of the survival pathways (AKT). USA). The estrogen levels in plasma were measured in Second, as a modulator of mesenchymal cell differentia- duplicate from four mice of each genotype using Estradiol tion (Richard et al., 2005), Sam68 may be involved in Coat-A-Count from Diagnostic Products Corporation (Los the epithelial–mesenchymal transition, characterized by Angeles, CA, USA). altered cellular morphology, loss of cell adhesion and gain of migratory ability (Barbera et al., 2004; Larue Acknowledgements and Bellacosa, 2005; Christiansen and Rajasekaran, 2006). We thank Dr Robert D Cardiff (University of California- Davis) and Dr Stephen M Dilworth (Imperial College London, UK) for the generous gifts of PyMT-derived cells and Materials and methods antibody, respectively. This work was supported by Grant MT-13377 from the Canadian Institutes of Health Research Mouse models (CIHR). MEH is the recipient of a postdoctoral fellowship The genotyping of PyMT transgenic mice was preformed by from Cancer Research Society Inc. SR is an investigator of the PCR with the following primers: forward, 50-GGAAGCAAG CIHR.

References

Andrechek ER, Muller WJ. (2000). Tyrosine kinase signalling in breast Busa R, Paronetto MP, Farini D, Pierantozzi E, Botti F, Angelini D, cancer: tyrosine kinase-mediated signal transduction in transgenic Attisani F et al. (2007). The RNA-binding protein Sam68 mouse models of human breast cancer. Breast Cancer Res 2: contributes to proliferation and survival of human prostate cancer 211–216. cells. Oncogene (e-pub ahead of print). Babic I, Jakymiw A, Fujita DJ. (2004). The RNA binding protein Cheng AM, Saxton TM, Sakai R, Kulkarni S, Mbamalu G, Vogel W Sam68 is acetylated in tumor cell lines, and its acetylation correlates et al. (1998). Mammalian Grb2 regulates multiple steps in embryonic with enhanced RNA binding activity. Oncogene 23: 3781–3789. development and malignant transformation. Cell 95: 793–803. Bao J, Gur G, Yarden Y. (2003). Src promotes destruction of c-Cbl: Cheng C, Sharp PA. (2006). Regulation of CD44 alternative splicing implications for oncogenic synergy between Src and growth factor by SRm160 and its potential role in tumor cell invasion. Mol Cell receptors. Proc Natl Acad Sci USA 100: 2438–2443. Biol 26: 362–370. Barbera MJ, Puig I, Dominguez D, Julien-Grille S, Guaita-Esteruelas Christiansen JJ, Rajasekaran AK. (2006). Reassessing epithelial to S, Peiro S et al. (2004). Regulation of Snail transcription during mesenchymal transition as a prerequisite for carcinoma invasion and epithelial to mesenchymal transition of tumor cells. Oncogene 23: metastasis. Cancer Res 66: 8319–8326. 7345–7354. Cohen LA, Guan JL. (2005). Mechanisms of focal adhesion kinase Borowsky AD, Namba R, Young LJ, Hunter KW, Hodgson JG, regulation. Curr Cancer Drug Targets 5: 629–643. Tepper CG et al. (2005). Syngeneic mouse mammary carcinoma cell Courtneidge SA, Goutebroze L, Cartwright A, Heber A, Scherneck S, lines: two closely related cell lines with divergent metastatic Feunteun J. (1991). Identiﬁcation and characterization of the behavior. Clin Exp Metastasis 22: 47–59. hamster polyomavirus middle T antigen. J Virol 65: 3301–3308.

Oncogene Sam68 modulates signaling in mammary tumorigenesis S Richard et al 556 Derry JJ, Richard S, Valderrama Carvajal H, Ye X, Vasioukhin V, Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS, Smithies Cochrane AW et al. (2000). Sik (BRK) phosphorylates Sam68 in the O. (1993). Alteration of reproductive function but not prenatal nucleus and negatively regulates its RNA binding ability. Mol Cell sexual development after insertional disruption of the mouse Biol 20: 6114–6126. estrogen receptor gene. Proc Natl Acad Sci USA 90: 11162–11166. Dilworth SM. (2002). Polyoma virus middle T antigen and its role in Lucchini F, Sacco MG, Hu N, Villa A, Brown J, Cesano L et al. identifying cancer-related molecules. Nat Rev Cancer 2: 951–956. (1992). Early and multifocal tumors in breast, salivary, harderian Fendrick JL, Raafat AM, Haslam SZ. (1998). Mammary gland growth and epididymal tissues developed in MMTY-Neu transgenic mice. and development from the postnatal period to the postmenopause: Cancer Lett 64: 203–209. ovarian steroid receptors ontogeny and regulation in the mouse. Lukong KE, Larocque D, Tyner AL, Richard S. (2005). Tyrosine J Mammary Gland Biol Neoplasia 3: 7–21. phosphorylation of sam68 by breast tumor kinase regulates intra- Fujita Y, Krause G, Scheffner M, Zechner D, Leddy HE, Behrens J nuclear localization and cell cycle progression. J Biol Chem 280: et al. (2002). Hakai, a c-Cbl-like protein, ubiquitinates and 38639–38647. induces endocytosis of the E-cadherin complex. Nat Cell Biol 4: Lukong KE, Richard S. (2003). Sam68, the KH domain-containing 222–231. superSTAR. Biochim Biophys Acta 1653: 73–86. Guy CT, Cardiff RD, Muller WJ. (1992). Induction of mammary Mamidipudi V, Dhillon NK, Parman T, Miller LD, Lee KC, tumors by expression of polyomavirus middle T oncogene: a Cartwright CA. (2007). RACK1 inhibits colonic cell growth transgenic mouse model for metastatic disease. Mol Cell Biol 12: by regulating Src activity at cell cycle checkpoints. Oncogene 26: 954–961. 2914–2924. Guy CT, Muthuswamy SK, Cardiff RD, Soriano P, Muller WJ. Matter N, Herrlich P, Konig H. (2002). Signal-dependent regulation of (1994). Activation of the c-Src tyrosine kinase is required for the splicing via phosphorylation of Sam68. Nature 420: 691–695. induction of mammary tumors in transgenic mice. Genes Dev 8: Muller WJ, Sinn E, Pattengale PK, Wallace R, Leder P. (1988). Single- 23–32. step induction of mammary adenocarcinoma in transgenic mice Hovey R, Trott J, Vonderhaar B. (2002). Establishing a framework for bearing the activated c-neu oncogene. Cell 54: 105–115. the functional mammary gland: from endocrinology to morphology. Muthuswamy SK, Muller WJ. (1994). Activation of the Src family of J Mammary Gland Biol Neoplasia 7: 17–38. tyrosine kinases in mammary tumorigenesis. Adv Cancer Res 64: Kamalati T, Jolin HE, Mitchell PJ, Barker KT, Jackson LE, Dean CJ 111–123. et al. (1996). Brk, a breast tumor-derived non-receptor protein- Richard S, Torabi N, Franco GV, Tremblay GA, Chen T, Vogel G tyrosine kinase, sensitizes mammary epithelial cells to epidermal et al. (2005). Ablation of the Sam68 RNA binding protein protects growth factor. J Biol Chem 271: 30956–30963. mice from age-related bone loss. PLoS Genet 1: e74. Kim H, Laing M, Muller W. (2005). c-Src-null mice exhibit defects in Schomberg DW, Couse JF, Mukherjee A, Lubahn DB, Sar M, Mayo normal mammary gland development and ERalpha signaling. KE et al. (1999). Targeted disruption of the estrogen receptor-alpha Oncogene 24: 5629–5636. gene in female mice: characterization of ovarian responses and Larue L, Bellacosa A. (2005). Epithelial-mesenchymal transition in phenotype in the adult. Endocrinology 140: 2733–2744. development and cancer: role of phosphatidylinositol 3’ kinase/AKT Soriano P, Montgomery C, Geske R, Bradley A. (1991). Targeted pathways. Oncogene 24: 7443–7454. disruption of the c-src proto-oncogene leads to osteopetrosis in Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ et al. (2003). mice. Cell 64: 693–702. Progression to malignancy in the polyoma middle T oncoprotein Vernet C, Artzt K. (1997). STAR, a gene family involved in signal mouse breast cancer model provides a reliable model for human transduction and activation of RNA. Trends Genet 13: 479–484. diseases. Am J Pathol 163: 2113–2126. Webster MA, Hutchinson JN, Rauh MJ, Muthuswamy SK, Anton M, Liu K, Li L, Nisson PE, Gruber C, Jessee J, Cohen SN. (2000). Tortorice CG et al. (1998). Requirement for both Shc and Neoplastic transformation and tumorigenesis associated with sam68 phosphatidylinositol 3’ kinase signaling pathways in polyomavirus protein deﬁciency in cultured murine ﬁbroblasts. J Biol Chem 275: middle T-mediated mammary tumorigenesis. Mol Cell Biol 18: 40195–40201. 2344–2359.

Oncogene