C-Jun Enhancement of Androgen Receptor Transactivation Is Associated with Prostate Cancer Cell Proliferation

Home , Androgen receptor

Oncogene (2006) 25, 7212–7223 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE c-Jun enhancement of androgen receptor transactivation is associated with prostate cancer cell proliferation

S-Y Chen1,2, C Cai2, CJ Fisher3, Z Zheng3, J Omwancha4, C-L Hsieh and L Shemshedini

Department of Biological Sciences, University of Toledo, Toledo, OH, USA

Androgens and the androgen receptor (AR) are involved Introduction in the growth and progression of prostate cancer. Our previous studies suggest that the proto-oncoprotein c-Jun Androgen signaling via the androgen receptor (AR) is is an AR coactivator that stimulates AR transactivation critical for the development of both normal prostate and by mediating receptor dimerization and subsequent DNA prostate cancer. In normal prostate development, binding. To study the physiological relevance of this c-Jun androgen signaling plays mainly a differentiative role, activity on AR, we have generated stable LNCaP cell lines whereas in prostate cancer the role switches to a expressing different levels of c-Jun. These cell lines proproliferative one (reviewed in Jenster, 1999). This exhibit a direct correlation between endogenous c-Jun molecular switch makes the growth of prostate cancer levels and AR transcriptional activity and expression of cells dependent on androgens and AR (reviewed in endogenous androgen-regulated genes. Disruption by Jenster, 1999; Arnold and Isaacs, 2002). Indeed, there is antisense RNA of endogenous c-Jun expression in increasing evidence that not only prostate cancer LNCaP cells strongly compromises the androgen- initiation but also its progression to the lethal andro- dependent proliferation of these cells. In contrast, gen-independent and metastatic form are all dependent expression of a c-Jun mutant, which is fully active in on androgens and AR (Isaacs, 1999; Arnold and Isaacs, coactivation of AR but deficient in AP-1 transactivation, 2002). It was recently reported that AR overexpression significantly enhances androgen-dependent proliferation. converts prostate cancer from androgen-sensitive to This finding indicates that the coactivation function of c- androgen-insensitive (Chen et al., 2005). This is further Jun is sufficient for regulating androgen-induced growth supported by previous findings that functional AR is of LNCaP cells. c-Jun also enhances AR transactivtion in found in most prostate cancer cells, including late-stage androgen-independent LNCaP cells, which closely mimic cells that become androgen-independent (Culig et al., hormone-refractory prostate cancer cells in gene expres- 2000), suggesting that the AR pathway may function sion and growth behavior. Importantly, siRNA-mediated throughout prostate carcinogenesis. AR mutations are repression of endogenous c-Jun expression results in frequent in prostate cancer, which often broaden AR markedly reduced growth of these cells, strongly suggest- ligand specificity and convert AR antagonists to ing an important biological role for c-Jun in hormone- agonists (reviewed in Linja and Visakorpi, 2004). refractory prostate cancer. Significantly, a recent study demonstrated that mutation Oncogene (2006) 25, 7212–7223. doi:10.1038/sj.onc.1209705; of the AR is sufficient for initiation and progression of published online 29 May 2006 prostate cancer in transgenic mice (Han et al., 2005). Like other nuclear receptors, the AR is a modular Keywords: androgen receptor; c-Jun; coactivation protein consisting of functionally distinct and conserved proliferation domains. The amino-terminal region contains transcriptional activation function-1 (AF-1) and the carboxy- terminal E region, which harbors the ligand-binding domain (LBD), has activation function-2 (AF-2) that is conserved in other receptors (Durand et al., 1994). AR transcriptional activity depends on the functions of its Correspondence: Associate Professor L Shemshedini, Department of two AFs, which are targets of the concerted actions of Biological Sciences, University of Toledo, 2801, W. Bancroft, Avenue, cofactors. Cofactors can either inhibit (co-repressors or Toledo, OH 43606, USA. repressors) or enhance (coactivators) AR activity E-mail: [email protected] (reviewed in Acevedo and Kraus, 2004). We have 1Current Address: Harvard Medical School, 330 Brookline Avenue, Boston, MA, USA. previously published that c-Fos (Tillman et al., 1998) 2These authors contributed equally to this work. and p53 (Shenk et al., 2001) can disrupt the amino-to- 3Current Address: Medical University of Ohio, 3000 Arlington carboxy terminal (N-to-C) interaction of AR. This Avenue, Toledo, OH, USA. interaction, which may be intermolecular or intramole- 4 Current Address: Johns Hopkins University, 615 North Wolfe St., cular, is thought responsible for forming a transcrip- Baltimore, MD, USA. Received 16 February 2006; revised 5 April 2006; accepted 8 April 2006; tional competent AR dimer (Langley et al., 1998). published online 29 May 2006 Interestingly, the dimerization of AR, like other nuclear c-Jun enhances androgen-induced proliferation S-Y Chen et al 7213 receptors, is driven by the D box in the second zinc- cells is antiproliferative, suggesting that high c-Jun finger of the DNA-binding domain (Haelens et al., transactivation reduces the growth of prostate cancer 2003). Other AR repressors include histone deacetylase cells. 1 (HDAC1) (Korkmaz et al., 2004), Hey1 (Belandia et al., 2005), PIASg (Gross et al., 2004), silencing mediator for retinoic acid and thyroid hormone receptor (SMRT) (Dotzlaw et al., 2002), and HBO1 (Sharma Results et al., 2000). AR coactivators include SRC-1 (Onate et al., 1998), TIF2 (Berrevoets et al., 1998), SRC-3 (also Exogenous c-Jun enhances the transcriptional activity of known as RAC3 or AIB1) (Zhou et al., 2005), and CBP endogenous AR in LNCaP cells (Aarnisalo et al., 1998), all of which are overexpressed in Our previous data suggesting that c-Jun can act as an prostate cancer (Culig et al., 2004). Interestingly, SRC-1 AR coactivator prompted us to investigate a possible (Spencer et al., 1997) and CBP (Martinez-Balbas et al., regulatory role for c-Jun on the endogenous AR in 1998) have histone acetylase activity. The coactivator LNCaP cells. Transient transfection experiments p300 is implicated in ligand-independent AR activation showed that c-Jun has the same positive effect on (Debes et al., 2002). SUMO-3 can enhance AR endogenous AR in LNCaP cells that has been seen on transcriptional activity independent of its sumoylation exogenous AR in other mammalian cells (Figure 1a). activity (Zheng et al., 2006). LSD1 is a histone In addition, a transactivation-deficient mutant of c-Jun, demethylase that recently was shown to act on the AR c-Jun(Ala63/73) (Wise et al., 1998), was fully able and be overexpressed in prostate cancer (Metzger et al., to coactivate the endogenous AR in LNCaP cells 2005). The AR is unique in its interaction with c-Jun, a proto-oncoprotein that can heterodimerize with c-Fos to form AP-1 (activator protein-1) (reviewed in Angel and Karin, 1991). While c-Jun represses the transcriptional activity of the AR-related glucocorticoid receptor (GR) (Schule et al., 1990), as well as other nuclear receptors (reviewed in Pfahl, 1993), it strongly potentiates AR- mediated transcription (Shemshedini et al., 1992; Bubulya et al., 1996). Several studies from our lab have characterized this coactivator function of c-Jun on AR (Bubulya et al., 1996, 2000, 2001; Tillman et al., 1998; Wise et al., 1998). Perhaps the most intereting finding is that c-Jun’s function as coactivator can be separated from that of a transactivator, as a transactivation- deficient mutant, c-Jun(Ala63/73), is still active in coactivation (Wise et al., 1998). Moreover, whereas c-Fos strongly stimulates c-Jun’s transactivation properties, it is equally able to repress its coactivation of AR (Tillman et al., 1998). These results together strongly suggest that the transactivation and coactivation functions of c-Jun are distinct and thus may operate via different mechanisms. More recently, we have published that c-Jun can mediate AR N-to-C interaction and thereby enhance DNA binding (Bubulya et al., 2001), an ability that is also observed with other AR coactivators, including SRC-1, TIF2, and CBP (Ikonen et al., 1997). To study the physiological relevance of c-Jun’s coactivation of AR transactivation, we have generated stable LNCaP cells expressing different levels of c-Jun. Our data show that overexpression of the c-Jun mutant cJun(Ala63/73) leads to enhanced androgen-regulated AR transactivation, gene expression, and cell proliferation, whereas in the cells overexpressing antisense-c-Jun Figure 1 c-Jun enhances, whereas antisense c-Jun represses, the effects are opposite. Importantly, short interference androgen-dependent AR transactivation in LNCaP cells. LNCaP RNA (siRNA)-mediated downregulation of c-Jun in cells were transfected with 2 mg MMTV-CAT or 2 mg TRE-tk-CAT androgen-independent LNCaP cells also has a negative and 1 or 5 mg c-Jun (a), c-Jun(Ala63/73) (b), or antisense c-Jun (c), effect on their growth, strongly suggesting that c-Jun as indicated. DHT (100 nM) was used as indicated. In all cases, the cells were grown in the absence (À) or presence ( þ ) of 100 nM coactivation of AR is important in both the androgen- DHT. All CAT activities are relative to the activity of the first responsive and -unresponsive stages of prostate cancer. condition, which was transfection of an empty expression plasmid, Finally, overexpression of wild-type c-Jun in LNCaP and this activity was set to 1.

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7214 (Figure 1b), suggesting that the endogenous level of c-Jun in LNCaP cells is limiting for AR-dependent transactivation. To determine if this endogenous c-Jun may be involved in AR transactivation, we used antisense technology to diminish c-Jun expression. LNCaP cells were transiently transfected with an AP- 1-responsive plasmid and a plasmid expressing a full- length c-Jun RNA in the antisense orientation (antisense c-Jun). As is shown in Figure 1c, antisense c-Jun repressed AP-1-dependent transcription in a dose- dependent manner, suggesting that the antisense tran- script blocked expression of endogenous c-Jun. When the same experiment was repeated with MMTV-CAT, an androgen-responsive reporter, antisense c-Jun had the same negative effect, severely inhibiting DHT-induced transcription (Figure 1c).

Generation of stable LNCaP cell lines having altered c-Jun expression levels The data above strongly suggest that endogenous c-Jun in LNCaP cells can enhance AR-dependent transactivation and provide a novel approach to up- or down- regulate the transcriptional activity of the AR in LNCaP cells, by altering the endogenous levels of c-Jun in the cells. Therefore, we generated stable transfectants of LNCaP cells expressing c-Jun, c-Jun(Ala63/73),or antisense c-Jun. Multiple cell lines were selected for each transfection, with each line having different level of AR transcriptional activity. Importantly, all the cell lines expressing either wild-type c-Jun (Figure 2a) or c-Jun(Ala63/73) protein (Figure 2b) exhibited variable, but elevated, AR transcriptional activity, as compared to untransfected LNCaP cells or C14 cells that were stably transfected with empty vector. The cell lines expressing antisense c-Jun have AR transcriptional activity that is reduced when compared to C14 Figure 2 Screen of stable LNCaP lines with altered endogenous c-Jun levels. Cells stably expressing (a) c-Jun (J cell lines), (b) (Figure 2c). These stable transfection results support c-Jun(Ala63/73) (M cell lines), or (c) antisense c-Jun (AJ cell lines) the transient transfection experiments (see Figure 1) and were selected on neomycin-containing medium and screened for provide us with LNCaP cells that have significantly DHT-dependent transactivation by transiently transfecting 1 mg elevated or reduced AR transcriptional activity. MMTV-CAT. 100 nM DHT was used to activate the AR in all cases. Note that the C14 line represents a stable transfection with Two lines each of wild-type c-Jun, c-Jun(Ala63/73), or the empty pCI-Neo vector. All activities are relative to the activity antisense c-Jun have been selected for further analysis. of either (a and b) untransfected LNCaP cells or (c) C14 cells, and As shown in Figure 3a, the cells expressing the mutant this activity was set to 1. c-Jun protein (M37, M26) exhibit significantly higher AR transcriptional than either untransfected LNCaP or C14 cells. Similar results have been obtained with cells To determine the phosphorylation state of c-Jun in our overexpressing wild-type c-Jun (S Chen and L Shem- stable LNCaP cell lines, Western blotting was per- shedini, unpublished results). In contrast, androgen- formed using an anti-phospho-Ser63-c-Jun antibody. As dependent AR transcriptional activity is almost com- shown in Figure 3d, the AJ81 cells have barely pletely abolished in the two antisense c-Jun lines (AJ6, detectable levels of phosphorylated c-Jun. In contrast, AJ81) (Figure 3a). Also note that the c-Jun(Ala63/73) J63 cells express high amounts of phosphorylated c-Jun lines have little activity on an AP-1-responsive reporter, (Figure 3d). As expected, M37cells have significantly whereas as expected, the antisense c-Jun cells have less phosphorylated c-Jun than do J63 cells, even though greatly reduced AP-1 activity (Figure 3b). Western blot the total c-Jun expression levels are similar in these two analysis showed that M37cells have elevated c-Jun cell lines (Figure 3d). Importantly, these stably trans- protein levels, whereas AJ81 cells have no detectable fected LNCaP cells express AR protein to levels c-Jun (Figure 3c). c-Jun is phosphorylated by JNK (Jun comparable to what is expressed in C14 cells N-terminal kinase) at serines 63 and 73, which is (Figure 3e), demonstrating that changes in endogenous necessary for its transactivation function (Smeal et al., c-Jun do not significantly alter the expression or stability 1991) but not its coactivation of AR (Wise et al., 1998). of AR protein.

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7215

Figure 3 Stable LNCaP cell lines expressing c-Jun(Ala63/73) or antisense c-Jun exhibit altered transcriptional activity. LNCaP cell lines were transfected with (a)1mg MMTV-CAT or (b)2mg TRE-tk-CAT. DHT (100 nM) was used as indicated. All activities are relative to the activity in (a) LNCaP cells without DHT or (b) C14, and this activity was set to 1. Western blotting was used to measure the expression of c-Jun (c), phosphrylated c-Jun (d), or AR (e) in stable LNCaP cell lines. Note that P-c-Jun was detected with an anti- phsopho-Ser63 c-Jun antibody and that b-actin was used as a control for protein loading. As indicated, the cells were grown in the absence (À) or presence ( þ ) of 100 nM DHT.

c-Jun regulates the expression of endogenous androgen- (Tomlins et al., 2005). Both Northern blot (Figure 4a) regulated genes in LNCaP cells and RT–PCR analyses (Figure 4d) showed that DHT While the above experiments (see Figure 3a) demon- induced the expression of TMPRSS2 mRNA in our strate that modulation of endogenous c-Jun levels in LNCaP stable lines, consistent with previous reports LNCaP cells affects AR transactivation of a transiently (Lin et al., 1999). Importantly, DHT-induced expression transfected reporter plasmid, they do not monitor of TMPRSS2 is increased in M37cells (Figure 4a) and changes in endogenous gene expression. To address significantly decreased in AJ6 cells (Figure 4d). North- this, we performed Northern blot, reverse transcriptase- ern blot analysis has shown two TMPRSS2 transcripts, polymerase chain reaction (RT–PCR), and Western blot one B3.8 kb and the other B2.0 kb (Paoloni-Giacobino analyses with RNA or proteins extracted from the stable et al., 1997). Our data demonstate that both transcripts cell lines to measure the expression of PSA and hKLK2. are induced by DHT in LNCaP cells (Figure 4a). These two androgen-regulated genes are known to be Interestingly, the RT–PCR reactions yielded two differ- upregulated in both proliferating LNCaP cells and ent-sized products and showed that only the lower prostate cancer (Gleave et al., 1992; Murtha et al., product is DHT induced (Figure 4d). It is unclear 1993). As shown in Figure 4a and b, dihydrotestosterone whether these two PCR products reflect the two (DHT) induced the expression of both PSA and hKLK2 transcripts observed on the Northern blot (Figure 4a). in LNCaP cell lines. The same DHT induction was These results, together with the PSA and hKLK2 data, observed on prostate-specific antigen (PSA) protein clearly demonstrate that altering the endogenous levels levels (Figure 4c). Importantly, the expressions of of c-Jun in LNCaP cells results in changes in androgen- hKLK2 mRNA and PSA mRNA and protein are regulated gene expression, arguing for a biological role significantly higher in M37cells than in C14 cells for c-Jun in AR activity in LNCaP cells. (Figure 4a–c). In contrast, the AJ81 cells, which have significantly reduced AR activity, yielded no detectable DHT induction of PSA mRNA (Figure 4b) and protein c-Jun coactivation mediates and transactivation inhibits (Figure 4c). androgen-dependent proliferation of LNCaP cells To determine if the expression of other androgen- To determine if this altered AR transcriptional activity regulated genes may be altered in the stable LNCaP affects a biological function of AR, androgen-dependent lines, we studied transmembrane protease serine 2 cellular proliferation was measured. However, first, the (TMPRSS2), a known androgen-regulated gene (Lin conditions for cellular proliferation were optimized as et al., 1999). Like PSA and hKLK2, TMPRSS2 is androgens have both stimulatory and inhibitory effects overexpressed in prostate cancer cells (Lin et al., 1999) on LNCaP proliferation (Horoszewicz et al., 1983; Lee and in a majority of prostate cancer patients (Vaarala et al., 1995), depending on the concentration of the et al., 2001). Recently, this gene was reported to undergo steroid hormone and culture medium conditions. There- translocation with two ETs oncogenes in prostate cancer fore, the growth properties of LNCaP cells were studied

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7216

Figure 5 Dose–response effects of DHT and R1881 on the Figure 4 Stable LNCaP cell lines expressing c-Jun(Ala63/73) or proliferation of androgen-dependent LNCaP cells. LNCaP cells antisense c-Jun exhibit altered androgen-regulated gene expression were grown in 2% DCC-stripped FBS under (a) different that is directly related to AR transcriptional activity. LNCaP stable concentrations of DHT (0, 0.1, 1, 10, and 100 nM) for 6 days, lines were grown in the presence ( þ ) or absence (À) of 100 nM (b) in the presence ( þ ) or absence (À) of 100 nM DHT for 1, 2, and DHT. Total RNA or whole-cell protein extracts were subjected to 6 days, as indicated, or (c) under different concentrations (1, 10, (a) Northern blot analysis to measure PSA, hKLK2,orTMPRSS2 and 100 nM) of DHT or R1881 for 6 days. The bar graphs represent expression, (b, d) RT–PCR to measure PSA or TMPRSS2 three independent experiments plus standard deviations. (d) R1881 expression, or (c) Western blot analysis to measure PSA protein is more effective at stabilizing AR protein than DHT, as measured expression. Note that the loading control was glyceraldehyde 3- by Western blotting of LNCaP cells treated with different phosphate dehydrogenase (GAPDH)(a, b, d)orb-actin (c). concentrations (1, 10, and 100 nM) of either DHT or R1881 for 2 days. Note that b-actin was the loading control.

in the presence of varying concentrations of DHT and in As expected, cells grown in 2% FBS medium showed a low (2%) or high (10%) fetal bovine serum (FBS). DHT-dependent increase in cell number with increasing LNCaP cells grown for 6 days in medium containing time (Figure 5b), whereas those grown in 10% FBS- low FBS (2%) extracted with dextran-coated charcoal containing medium showed a decrease in proliferation (DCC) exhibited a dose-dependent increase in cell (S Chen and L Shemshedini, unpublished results). To number with increasing concentration of DHT (from conﬁrm that our growth conditions allowed us to 0.1 to 100 nM) (Figure 5a), in agreement with earlier observe androgen-induced proliferation, we compared results (Horoszewicz et al., 1983), whereas those grown the effect of DHT to that of R1881, a synthetic analog in the presence of 10% FBS showed a slight decrease (S of DHT that is more stable than DHT (Schuurmans Chen and L Shemshedini, unpublished results). To study et al., 1988). As shown in Figure 5c, 1 nM R1881 is the effect of 100 nM DHT on proliferation, LNCaP cells sufﬁcient to induce optimal proliferation of LNCaP were grown in the presence of either carrier (ethanol) or cells, as reported previously (Song et al., 2004). DHT for 1, 2, and 6 days in medium with 2 or 10% FBS. Importantly, 1 nM R1881 is as active, if not slightly

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7217 more active, on proliferation as is 100 nM DHT cells and the coactivation function of exogenously (Figure 5c). To determine if this difference between expressed c-Jun is proproliferative and its transactiva- DHT and R1881 could be explained by AR protein tion function is antiproliferative. Note that AR expres- stability, which is strongly stimulated by androgens sion is not significantly changed in these stable LNCaP (Yeap et al., 1999), AR protein levels were measured by cell lines (see Figure 3e; C Cai and L Shemshedini, Western blotting. DHT had a significant stabilizing unpublished results). effect on AR protein only at a concentration of 100 nM, whereas R1881 was equally effective from 1 to 100 nM c-Jun mediates the proliferation of androgen-independent (Figure 5d). These results demonstrate that, under our LNCaP cells culture conditions, 100 nM DHT has the same effect as To determine the importance of c-Jun in androgen- 1nM R1881 on LNCaP cellular proliferation. independent LNCaP cells, we studied androgen-inde- Thus, the above growth conditions (2% FBS medium pendent LNCaP cells that were established through and 100 nM DHT) were used to compare the prolifera- continuous passage in culture (Lin et al., 1998; Igawa tion of the stable LNCaP cell lines. M37and M26 cells et al., 2002). These cells, C81, closely mimic hormone- had enhanced DHT-dependent proliferation, compared refractory prostate cancer and thus proliferate in an to the control C14 cells (Figure 6a). In contrast, the AJ6 androgen-unresponsive manner, whereas C33 are the and AJ81 cells lost most of their proliferative response parental androgen-dependent cells (Igawa et al., 2002). to DHT (Figure 6b). Surprisingly, cell overexpressing The C81 LNCaP cells express c-Jun mRNA (Figure 7a) wild-type c-Jun protein grew as poorly as the antisense c- and protein (Figure 7b) and AR protein (Figure 7c) at Jun-expressing cells (Figure 6c). Importantly, none of comparable levels that are expressed in androgen- the stable cell lines differed significantly from C14 cells dependent LNCaP cells. Although the total c-Jun in growth under DHT-minus conditions, demonstrating protein levels are similar, C81 cells have a higher level that overexpression of wild-type or mutant c-Jun or of phosphorylated c-Jun than either C33 or LNCaP cells under-expression of endogenous c-Jun did not affect the (Figure 7b). Interestingly, both androgen-dependent and inherent growth rate of the cells. Together, these results -independent LNCaP cells exhibit enhanced AR protein strongly suggest that endogenous c-Jun plays an levels in the presence of DHT (Figure 7c), owing to the important role in DHT-induced proliferation of LNCaP well positive effect androgens have on AR protein

Figure 6 c-Jun coactivation of AR enhances the proliferation of androgen-dependent LNCaP cells. Stable LNCaP lines (C14, M37, and AJ81) were grown for 0, 2, 4, and 6 days in 2% DCC-stripped FBS in the presence of 100 nM DHT (gray bars) or ethanol carrier (white bars) as a control. The cell lines express (a) c-Jun (Ala63/73) mutant protein, (b) antisense c-Jun RNA, or (c) wild-type c-Jun protein. The bar graphs represent three independent experiments plus standard deviations.

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7218 In view of the importance of c-Jun in the growth of androgen-independent LNCaP cells, we were interested in exploring a possible proliferative role for c-Jun in androgen-independent cells. We utilized siRNA- mediated repression of c-Jun, which was effective on both c-Jun mRNA (Figure 8a) and protein (Figure 8b) levels. These siRNA-transfected cells were then studied for growth, which expectedly is androgen-independent. As shown in Figure 8c, the androgen-independent proliferation of C81 cells is dramatically reduced by c-Jun siRNA treatment, as compared to control siRNA. As the siRNA pool used above consisted of two distinct siRNAs targeting different regions of c-Jun, we tested each siRNA separately to ensure the speciﬁcity of the approach. As shown in Figure 8d, both siRNAs are able to inhibit c-Jun expression but they differ in magnitude. This difference in magnitude on c-Jun expression correlated directly with the siRNA effectiveness on cellular proliferation (Figure 8e). Collectively, these data illustrate an important role for endogenous c-Jun in the androgen-independent growth of LNCaP cells.

c-Jun can coactivate AR in PC-3 cells The data described above demonstrate that c-Jun can enhance AR transactivation in both androgen-dependent and androgen-independent LNCaP cells. PC-3 cells were used to determine if c-Jun coactivation of AR can occur in androgen-independent prostate cancer cells that do not express endogenous AR. These cells were stably transfected with AR, giving rise to A103 cells (Salas et al., 2004), which express AR protein levels compar- Figure 7 c-Jun enhances AR transactivation in androgen-inde- able to LNCaP cells (S Chen and L Shemshedini, pendent LNCaP cells. (a, b) c-Jun is equally expressed in androgen- unpublished results) and thus exhibit androgen-induced dependent and androgen-independent LNCaP cells. The expression AR transactivation (Figure 9a). Interestingly, both wild- of c-Jun was measured by semiquantitative RT–PCR (a)or Western blotting (b) in LNCaP, C33, and C81 cells. Note that type c-Jun and c-Jun(Ala63/73) can enhance AR P-c-Jun was detected with an anti-phsopho-Ser63 antibody. DHT transactivation in A103 cells (Figure 9a), showing that increases the expression levels of AR protein (c)orPSA and c-Jun coactivation of AR can occur in different prostate TMPRSS2 mRNA (d) in both androgen-dependent (LNCaP) and cancer cell lines. To determine if the endogenous c-Jun is androgen-independent (C81) LNCaP cells. Note that GAPDH or b-actin was used as a control for the amount of RNA or protein, involved in A103 cellular proliferation, this c-Jun respectively. (e) c-Jun stimulates androgen-dependent AR transac- expression was repressed by siRNA (Figure 9b). A103 tivation in androgen-independent LNCaP cells. C81 cells were cellular proliferation is weakly but reproducibly inhib- transfected with 2 mg MMTV-CAT and 0.1 or 0.5 mg c-Jun or ited by DHT (Figure 9c), as it has been shown for other c-Jun(Ala63/73), as indicated. In all cases, the cells were grown in AR-expressing PC-3 cells (Heisler et al., 1997). More the absence (À) or presence ( þ ) of 100 nM DHT. All activities are relative to AR activity in the absence of transfected c-Jun or interestingly, transfection with c-Jun siRNA markedly c-Jun(Ala63/73), both of which were set to 1. reduced A103 proliferation (Figure 9c), suggesting an endogenous role for c-Jun in PC-3 cellular proliferation.

c-Jun expression is elevated in prostate cancer tissues stability (Yeap et al., 1999). RT–PCR analysis was used Collectively, our cellular proliferation data above to demonstrate that DHT can induce the expression of strongly suggest that c-Jun coactivation of AR is PSA and TMPRSS2 mRNA in both C33 and, more involved in the two distinct growth stages of LNCaP weakly, C81 cells (Figure 7d). When AR transactivation cells: androgen-dependent and androgen-independent. was measured using MMTV-CAT (Bubulya et al., This ﬁnding implicates the possible involvement of c-Jun 1996), both wild-type c-Jun and c-Jun(Ala63/73) were in androgen-refractory prostate cancer, the lethal form able to stimulate AR transcriptional activity (Figure 7e), of the disease. Support for a possible c-Jun involvement but with a lower capacity than what was observed in in prostate carcinogenesis depends on demonstrating its androgen-dependent LNCaP cells (see Figure 1). These expression in human tumor tissues. We have used RT– results demonstrate that c-Jun coactivation of AR can PCR to measure c-Jun expression in several prostate occur in both androgen-dependent and androgen- cancer cell lines and prostate tissues. As shown in Figure independent LNCaP cells. 10a, c-Jun mRNA is expressed similarly in LNCaP and

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7219

Figure 8 Endogenous c-Jun is involved in the proliferation of androgen-independent LNCaP cells. (a, b, d) Transfection of 25 nM c-Jun siRNA into C81 signiﬁcantly reduces c-Jun expression. C81 cells were transfected with a c-Jun siRNA pool, the two individual siRNAs, or an unrelated control siRNA, and the expression of c-Jun was measured by semiquantitative RT–PCR (a) or Western blotting (b, d). Note that GAPDH or b-actin were used as controls for the amount of RNA or protein, respectively. (c, e) Transfection of c-Jun siRNA signiﬁcantly reduces the growth of androgen-independent LNCaP (C81) cells. In all cases, the cells were grown in the absence (À) or presence ( þ ) of 100 nM DHT. The bar graphs represent three independent experiments plus standard deviations.

PC-3 cells, without a significant effect by DHT Interestingly, at least with regard to AR (Wise et al., treatment. The tissues studied were purchased from 1998), PU.1 (Behre et al., 1999), and ETV1 (C Cai and the Cooperative Human Tissue Network (CHTN), L Shemshedini, unpublished results), c-Jun’s function as which identifies these tissues as normal, benign prosta- a coactivator is separable from its transactivation tic hyperplasia (BPH), or malignant prostate cancer function, as the c-Jun(Ala63/73) mutant is deficient in (MPC). Interestingly, c-Jun mRNA is only detected in transactivation but fully active in coactivation. This our two MPC and one BPH samples (Figure 10b). separation of AR coactivation from AP-1 transactiva- Importantly, these three tissues also express elevated tion has been demonstrated in a variety of cell lines, levels of PSA, AR, and EZH2 (Varambally et al., 2002) including Cos (Bubulya et al., 2001), androgen-depen- mRNAs (Figure 10b). dent and -independent LNCaP, and AR-expressing PC-3 cells. In addition, c-Fos strongly stimulates c- Jun’s transactivation properties but represses its coactivation of AR (Tillman et al., 1998). These results Discussion together strongly suggest that the transactivation and coactivation functions of c-Jun are distinct and thus The c-Jun proto-oncoprotein can activate transcription may operate via different mechanisms. In this paper, we through at least two mechanisms: transactivation and provide evidence that, for prostate cancer cells, c-Jun coactivation. The transactivation properties of c-Jun are coactivation of AR is proliferative, whereas, surpris- well understood, having clearly demonstrated that this ingly, c-Jun transactivation is antiproliferative. protein is found together with c-Fos in AP-1, which is LNCaP is a lymph node-derived prostate cancer cell able to transcriptionally regulate the expression of genes line that expresses AR, c-Jun, and c-Fos (Henttu and important for cell division (reviewed in Shaulian and Vihko, 1998). Because this cell line has many of the same Karin, 2001). In contrast, c-Jun’s coactivation function androgen-regulated properties that are found in andro- is less well understood. This c-Jun action appears to be gen-sensitive, early-stage prostate cancer (reviewed in limited to only a few transcriptional activators, includ- Jester, 1999), LNCaP cells are used as a model for ing AR and the three ETS transcription factors ERM studying androgen-sensitive prostate cancer. These (Nakae et al., 1995), PU.1 (Behre et al., 1999), and properties include androgen-induced cellular prolifera- ETV1 (C Cai and L Shemshedini, unpublished results). tion and gene expression. LNCaP cells were, therefore,

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7220

Figure 10 c-Jun is overexpressed in prostate cancer tissues. (a) Semiquantitative RT–PCR was used to measure expression of c-Jun mRNA in C14 (LNCaP) and A103 (PC-3). (b) Total RNA was isolated from prostate tissues (acquired from the Cooperative Human Tissue Network) that are normal (N1), benign prostatic hyperplasia (B1-B3), or malignant prostate cancer (C1, C2) and subjected to semiqantitative RT–PCR to measure the expression of c-Jun, PSA, hAR, and EZH2 mRNA. In both cases, GAPDH was used as a control for the RNA amount.

Figure 9 c-Jun enhances androgen-dependent AR transactivation are completely consistent with our earlier data using in AR-expressing PC-3 cells. (a) A103 cells (PC3 cells stably transient transfection experiments (Bubulya et al., 1996, expressing AR) were transfected with 1 mg PSA-Luc and 1 mg c-Jun, 2000, 2001; Tillman et al., 1998; Wise et al., 1998) and or c-Jun(Ala63/73), as indicated. DHT (100 nM) was used as indicated. In all cases, the cells were grown in the absence (À)or strongly suggest that endogenous levels of c-Jun in presence ( þ ) of 100 nM DHT. All luciferase activities are relative LNCaP cells are both required and limiting for to the activity of the ﬁrst condition, which was transfection of an androgen-regulated gene expression. Although it is not empty expression plasmid, and this was set to 1. A103 cells were known how widespread the involvement of c-Jun transfected with a c-Jun siRNA pool, or control siRNA, and coactivation on the expression of androgen-regulated subjected to (b) Western blotting or (c) proliferation using cells grown for 0, 2, and 4 days in the absence (À) or presence ( þ )of genes is, it is important to note that all three androgen 100 nM DHT. The bar graphs represent three independent target genes analysed in this study appear to be experiments plus standard deviations. regulated by c-Jun coactivation. One important biological consequence of androgen- regulated gene expression is induction of prostate cancer used in this study to determine the importance of c-Jun cell proliferation, which is clearly demonstrated by in AR-dependent proliferation of prostate cancer cells, LNCaP cells. Indeed, our data with the stable LNCaP an important biological function for AR in prostate cell lines implicate a role for c-Jun coactivation in this cancer. Stable LNCaP cells were generated that ex- property. In the cell lines expressing c-Jun(Ala63/73) pressed different levels of either wild-type c-Jun or protein or antisense c-Jun, androgen-induced cellular c-Jun(Ala63/73) protein, or antisense c-Jun. As expected, proliferation is directly proportional to the magnitude of expression of either c-Jun protein resulted in enhanced AR transcriptional activity. These results demonstrate AR transactivation of an androgen-regulated reporter that alteration of AR transcriptional activity in response gene and, more importantly, known endogenous andro- to changes in endogenous c-Jun levels leads to changes gen-regulated genes. In contrast, expression of antisense in cellular proliferation. Our data with the antisense c-Jun had a negative effect. These transcription results c-Jun-expressing cells show that reducing endogenous

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7221 c-Jun expression has a larger negative effect on DHT- c-Jun coactivation strongly enhances AR-dependent induced LNCaP cellular proliferation than on prolifera- gene expression and androgen-induced cellular prolif- tion without DHT. This suggests that the proliferative eration. Additionally and perhaps more importantly, effect of endogenous c-Jun on prostate cancer cells may endogenous c-Jun is required for the proliferation of be largely mediated by its coactivation function on AR. androgen-independent LNCaP cells. Utilizing LNCaP Additional evidence comes from the cell lines expressing cells that were cultured to grow independent of c-Jun(Ala63/73), which exhibit significantly elevated androgens and thereby mimic the hormone-refractory growth of LNCaP cells in the presence of DHT, but stage of prostate cancer (Lin et al., 1998; Igawa et al., no effect in its absence. This c-Jun mutant protein is 2002), we have been able to block the growth of these fully able to coactivate androgen-dependent transactiva- cells, in either the absence or presence of androgens, by tion, but is deficient in AP-1 transactivation (Wise et al., siRNA-mediated diminishing of endogenous c-Jun 1998). It has been previously reported that c-Jun can expression. As these cells exhibit c-Jun coactivation of mediate G1 cell-cycle progression of fibroblasts inde- AR-dependent transcription and we have evidence that pendent of phosphorylation of its serines 63 and 73 their proliferation is dependent on AR-regulated gene (Wisdom et al., 1999). Therefore, it is possible that the c- expression (C Cai and L Shemshedini, unpublished Jun mutant may be having a similar effect in LNCaP results), it is likely that the proliferative role of cells. Nevertheless, our finding that the c-Jun(Ala63/73) endogenous c-Jun in androgen-independent LNCaP mutant affects only androgen-dependent proliferation cells is mediated via its coactivation function on AR, strongly suggests that this effect is mediated via AR irrespective of whether androgen is present or not. coactivation. Future studies can determine if c-Jun coactivation of In contrast to cells expressing c-Jun(Ala63/73), the AR regulates the same proproliferation genes in LNCaP cell lines expressing wild-type c-Jun exhibit androgen-dependent and -independent LNCaP cells. greatly reduced cellular proliferation. In fact, the proliferative ability of c-Jun-overexpressing cells is as weak as that of the antisense c-Jun-expressing cells. Materials and methods These results demonstrate that altering endogenous Plasmids levels of c-Jun, either up or down, is antiproliferative To make c-Jun/pCI-Neo, c-Jun/pSG5 (Bubulya et al., 2001) in LNCaP cells, suggesting that the endogenous c-Jun was digested with BsaAI and XbaI, and the restriction level in LNCaP cells is optimal for proliferation. As the fragment was filled in with DNA polymerase klenow fragment. two cell lines transfected with antisense or sense c-Jun This fragment was then inserted into EcoRV-digested pTRE2 express widely different c-Jun protein levels, it is likely vector (Clontech, Mountain View, CA, USA) to get c-Jun/ that diminished and elevated c-Jun levels inhibit cell pTRE2. c-Jun/pTRE2 was digested with SalI and XbaI and proliferation via different mechanisms. The antisense inserted into pCI-Neo (Promega, Madison, WI, USA) digested c-Jun cells exhibit greatly reduced AR transcriptional with XhoI and XbaI, yielding c-Jun/pCI-Neo. activity and thus may have compromised androgen- To make antisense-c-Jun/pCI-Neo, c-Jun/pTRE2 was di- regulated gene expression that is necessary for proli- gested by NheI and XbaI and inserted into pCI-Neo in reverse orientation. feration. On the other hand, whereas the c-Jun- To make c-Jun(Ala63/73)/pCI-Neo, c-Jun(Ala63/73)/RSV overexpressing cells yield significantly enhanced AR (Bubulya et al., 2001) was digested with BsaAI and StyI and transactivation comparable to what is observed in filled in with DNA polymerase klenow fragment. This c-Jun(Ala63/73)-expressing cells, their growth is fragment was then inserted into EcoRV-digested pTRE2 to strongly inhibited. This observation suggests that high get c-Jun(Ala63/73)/pTRE2. c-Jun(Ala63/73)/pTRE2 was di- levels of c-Jun in LNCaP cells leads to activation of gested with SalI and XbaI and inserted into pCI-Neo digested antiproliferation genes that may dominantly negate the with XhoI and XbaI, yielding c-Jun(Ala63/73)/pCI-Neo. proproliferation effects of increased expression of The reporter plasmids MMTV-CAT, TRE-tk-CAT, and androgen-regulated genes. Our data demonstrate that PSA-Luc have been previously described (Bubulya et al., 1996; the c-Jun cells express significantly more phosphorylated Shenk et al., 2001). Transfection efficiency was standardized by measuring the b-galactosidase activity, originating from the c-Jun than do the c-Jun(Ala63/73) cells, which would cotransfected plasmid pCH110. allow a higher expression of c-Jun target genes in c-Jun cells than in c-Jun(Ala63/73) cells. These findings may Generation of stable cell lines represent a novel function for c-Jun transactivation in LNCaP cells were grown to 60–70% confluence in RPMI 1640 prostate cancer cells- that of antiproliferation. It is also complete medium (Sigma, St Louis, MD, USA) containing possible that the reduced cellular growth in c-Jun- 10%FBS (from Hyclone, Logan, UT, USA). Then, the expressing prostate cancer cells is owing to increased medium was changed to DMEM complete medium (Sigma) apoptosis, since c-Jun transactivation is known to for overnight incubation. Cells were transfected with 2 mg each induce apoptosis in neurons, lymphocytes, and hepato- of pCI-Neo vector (negative control from Promega), c-Jun/ cytes (reviewed in Ameyar et al., 2003). Future work will pCI-Neo, antisense c-Jun/pCI-Neo, or c-Jun(Ala63/73)/pCI- Neo, using calcium phosphate precipitation (CaPO4) method determine if overexpression of c-Jun in LNCaP cells (Bubulya et al., 2001). At 24 h after transfection, the cells were leads to decreased cell-cycle progression and/or in- again incubated in RPMI 1640 complete medium. After 48 h, creased apoptosis. LNCaP cells were selected in RPMI 1640 complete medium Whatever the mechanism of action of c-Jun transac- containing 0.9 mg/ml neomycin. The medium was refreshed tivation is in LNCaP cells, our data strongly suggest that every 4 days until single colonies appeared. The colonies were

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7222 screened by CAT assay of transiently transfected cells with Western blot analysis MMTV-CAT. The generation of A103 cells, which stably To prepare cell extracts for Western blot analysis, cells were express AR, has been previously described (Salas et al., 2004). boiled for 5–10 min in sodium dodecyl sulfate (SDS)-sample buffer (63 mM Tris, pH 6.8; 20% glycerol; 2% SDS; 5% Transient transfections and reporter gene assays b-mercaptoethanol). Proteins were separated by SDS-poly- Transient transfections were carried out using the CaPO4 acrylamide gel electrophoresis (SDS–PAGE) and then electro- method as described above in LNCaP cells and stable lines of transferred onto nitrocellulose. After blocking with nonfat dry LNCaP cells, in which case the medium contained 0.1 mg/ml milk, the nitrocellulose blots were probed with the anti-c-Jun neomycin. A103 cells were transfected using FuGENE 6 antibody sc-45 (Santa Cruz Biotechnology), anti-phospho- (Roche, Basel, Switzerland). Cells were incubated with or Ser63 c-Jun antibody sc-822 (Santa Cruz Biotechnology), the without 100 nM DHT for 24 h, after which whole-cell extracts anti-AR antibody PA1-110 (Affinity BioReagents, Golden, were prepared and then subjected to CAT or luciferase assays CO, USA), the anti-PSA antibody sc-7638 (Santa Cruz as described (Bubulya et al., 1996; Shenk et al., 2001). CAT or Biotechnology), or the anti-b-actin antibody ab6276 (Abcam, luciferase assays were standardized according to the measured Cambridge, MA, USA) as a control for protein loading. The b-galactosidase activity as previously described (Bubulya et al., blots were developed using the chemiluminescence detection 1996). For all transfections, we used different amounts of kit from Amersham, Piscataway,New Jersey, USA. expression plasmid, 1 mg reporter plasmid (MMTV-CAT, TRE-tk-CAT, or PSA-Luc), 2 mg pCH110, and enough pTL1 to bring the final plasmid amount to 10 mg per dish. For all Northern blot analysis transfections, empty vector (empty expression plasmid and/or RNA was isolated using the TRIZOL Reagent (Invi- promoterless reporter plasmid) was used to ensure equal trogen, Carlsbad, CA, USA) as recommended by the amounts of each kind of vector. CAT assay results were manufacturer and run on an agarose–formaldehyde gel. In quantified using the BioRad Molecular Imager FX of at least total, 20 mg of total RNA was used. GAPDH was used as a three repeats for each transfection. Thus, each CAT assay or control. luciferase value represents the average of three to four repeats plus standard deviation. Semiquantitative RT–PCR analysis siRNA transfection RNA was isolated using the TRIZOL Reagent and then C33 or C81 LNCaP cells were grown in RPM1 1640 medium subjected to reverse transcription using M-MLV Reverse with 0.05 mg/ml gentamicin (Gibco, Carlsbad, CA, USA) and Transcriptase (Fisher) following the manufacturer’s protocol. supplemented with 5% FBS. A103 cells were grown in F12K The cDNA product was used for PCR using the PCR Master medium (Sigma) supplemented with 5% FBS. Cells were Mix (Promega). The PCR products were separated by agarose transfected with the c-Jun siRNA pool sc-44201 (Santa Cruz gel-electrophoresis and stained with ethidium bromide. The Biotechnology, Santa Cruz, CA, USA), either of two c-Jun following PCR primers were used: c-Jun (upstream) 50- siRNAs (sc-44201A or sc-44201B, Santa Cruz Biotechnology) TGACTGGAAAGAT-GGAAACG-30 and (downstream) 50- or negative control siRNA 4611G (Ambion, Austin, TX, CCGTTGCTGGACTGGA-TTAT-30 TMPRSS2 (upstream) USA). The X-tremeGENE siRNA transfection reagent was 50-CACTGTGCATCACCTTGACC -30 and (downstream) 50- used following the prescribed protocol (Roche). The final ACACACCGATTCTCGTCCTC-30; AR (upstream) 50-CAA 0 0 siRNA concentration was 25 or 50 nM for each transfection. TGAGTACCGCATGCAC-3 and (downstream) 5 -GCCCA TCCACTGGAATAATG-30; PSA (upstream) 50-GCAGCAT 0 0 Cell proliferation assay TGAACCAGAGGAG-3 and (downstream) 5 -CCCATGA 0 0 50 000 cells were grown in a 24-well-plate with 500 ml RPMI CGTGATACCTTGA-3 ; GAPDH (upstream) 5 -CGACCA 0 0 1640 medium (LNCaP cells) or F12K medium (A103 cells) CTTTGTCAAGCTCA-3 and (downstream) 5 -AGGGGA 0 containing 2% DCC-stripped FBS. After a 2-day incubation, GATTCAGTGTGGTG-3 . either ethanol or 100 nM DHT was added into the wells. Then, the MTT (3-[4,5]dimethylthiazol-2,5-diphenyltetrazolium bromide) (Sigma) assay was performed after 0–6 days (LNCaP Acknowledgements cells) or 0–4 days (A103 cells) of incubation following the manufacturer’s protocol. The neomycin concentration was We thank Dr S Leisner for critical reading of the manuscript. 0.1 mg/ml for the LNCaP stable cell lines and 0.2 mg/ml for This work was supported by grants from National Institutes of A103 cells. Health and Ohio Cancer Research Associates.

References

Aarnisalo P, Palvimo JJ, Janne OA. (1998). Proc Natl Acad Sci Belandia B, Powell SM, Garcia-Pedrero JM, Walker MM, USA 95: 2122–2127. Bevan CL, Parker MG. (2005). Mol Cell Biol 25: Acevedo ML, Kraus WL. (2004). Essays Biochem 40: 73–88. 1425–1436. Ameyar M, Wisniewska M, Weitzman JB. (2003). Biochimie Berrevoets CA, Doesburg P, Steketee K, Trapman J, 85: 747–752. Brinkmann AO. (1998). Mol Endocrinol 12: 1172–1183. Angel P, Karin M. (1991). Biochim et Biophys Acta 1072: Bubulya A, Chen S-Y, Fisher CJ, Zheng Z, Shen X-Q, 129–157. Shemshedini L. (2001). J Biol Chem 276: 44704–44711. Arnold JT, Isaacs JT. (2002). Endocrine-Related Cancer 9: Bubulya A, Wise SC, Shen X-Q, Burmeister LA, Shemshedini 61–73. L. (1996). J Biol Chem 271: 24583–24589. Behre G, Whitmarsh AJ, Coghlan MP, Hoang T, Carpenter Bubulya A, Zhou X-F, Shen X-Q, Fisher CJ, Shemshedini L. CL, Zhang DE et al (1999). J Biol Chem 274: 4939–4946. (2000). Endocrine 13: 55–62.

Oncogene c-Jun enhances androgen-induced proliferation S-Y Chen et al 7223 Chen CD, Welsbie DS, Tran C, Baek SH, Chen R, Vessella R Metzger E, Wissmann M, Yin N, Muller JM, Schneider R, et al (2005). Nat Med 10: 33–39. Peters AH et al. (2005). Nature 437: 436–439. Culig Z, Comuzzi B, Steiner H, Bartsch G, Hobisch A. (2004). Murtha P, Tindall DJ, Young CY. (1993). Biochemistry 32: J Steroid Biochem Mol Biol 92: 265–271. 6459–6464. Culig Z, Hobisch A, Batsch G, Klocker H. (2000). Microscop Nakae K, Nakajima K, Inazawa J, Kitaoka T, Hirano T. Res Tech 51: 447–455. (1995). J Biol Chem 270: 23795–23800. Debes JD, Schmidt LJ, Huang H, Tindall DJ. (2002). Cancer Onate SA, Boonyaratanakornkit V, Spencer TE, Tsai SY, Res 62: 5632–5636. Tsai MJ, Edwards DP et al (1998). J Biol Chem 273: Dotzlaw H, Moehren U, Mink S, Cato AC, Iniguez Lluhi JA, 12101–12108. Baniahmad A. (2002). Mol Endocrinol 16: 661–673. Paoloni-Giacobino A, Chen H, Peitsch MC, Rossier C, Durand B, Saunders M, Gaudon C, Roy B, Losson R, Antonarakis SE. (1997). Genomics 44: 309–320. Chambon P. (1994). EMBO J 13: 5370–5382. Pfahl M. (1993). Enodcr Rev 14: 651–658. Gleave ME, Hsieh JT, Wu HC, von Eschenbach AC, Chung Salas TR, Kim J, Vakar-Lopez F, Sabichi AL, Troncoso P, LW. (1992). Cancer Res 52: 1598–1605. Jenster G et al (2004). J Biol Chem 279: 19191–19200. Gross M, Yang R, Top I, Gasper C, Shuai K. (2004). Schule R, Rangarajan P, Kliewer S, Ransone LJ, Bolado J, Oncogene 23: 3059–3066. Yang N et al (1990). Cell 62: 1217–1226. Haelens A, Verrijdt G, Callewaert L, Christiaens V, Schau- Schuurmans AL, Bolt J, Mulder E. (1988). Prostate 12: waers K, Peeters B et al (2003). Biochem J 369: 141–151. 55–63. Han G, Buchanan G, Ittmann M, Harris JM, Yu X, Sharma M, Zarnegar M, Li X, Lim B, Sun Z. (2000). J Biol DeMayo FJ et al (2005). Proc Natl Acad Sci USA 102: Chem 275: 35200–35208. 1151–1156. Shaulian E, Karin M. (2001). Oncogene 20: 2390–2400. Heisler LE, Evangelou A, Lew AM, Trachtenberg J, Shemshedini L, Knauthe R, Sassone-Corsi P, Pornon A, Elsholtz HP, Brown TJ. (1997). Mol Cell Endocrinol 126: Gronemeyer H. (1992). EMBO J 10: 3839–3849. 59–67. Shenk JL, Fisher CJ, Chen S-Y, Zhou X-F, Tillman K, Henttu P, Vihko P. (1998). Biochem Biophys Res Commun 244: Shemshedini L. (2001). J Biol Chem 276: 38742–38749. 167–171. Smeal T, Binetruy B, Mercola DA, Birrer M, Karin M. (1991). Horoszewicz JS, Leong SS, Kawinski E, Karr JP, Rosenthal Nature 354: 494–496. H, Chu TM et al (1983). Cancer Res 43: 1809–1818. Song LN, Coghlan M, Gelmann EP. (2004). Mol Endocrinol Igawa T, Lin FF, Lee MS, Karan D, Batra SK, Lin MF. 18: 70–85. (2002). Prostate 50: 222–235. Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J, Mizzen Ikonen T, Palvimo JJ, Janne OA. (1997). J Biol Chem 272: CA et al (1997). Nature 389: 194–198. 29821–29828. Tillman K, Oberﬁeld JL, Shen X-Q, Bubulya A, Shemshedini Isaacs JT. (1999). Urol Clin N Am 26: 263–273. L. (1998). Endocrine 9: 193–200. Jenster G. (1999). Semin Oncol 26: 407–421. Tomlins SA, Rhodes DR, Perner S, Dhanasekaran SM, Mehra Korkmaz CG, Fronsdal K, Zhang Y, Lorenzo PI, Saatcioglu R, Sun XW et al (2005). Science 310: 644–648. F. (2004). J Endocrinol 182: 377–389. Vaarala MH, Porvari K, Kyllonen A, Lukkarinen O, Vihko P. Langley E, Kemppainen JA, Wilson EM. (1998). J Biol Chem (2001). Int J Cancer 94: 705–710. 273: 92–101. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Lee C, Sutkowski DM, Sensibar JA, Zelner D, Kim I, Amsel I Kumar-Sinha C, Sanda MG et al (2002). Nature 419: 624–629. et al (1995). Endocrinology 136: 796–803. Wisdom R, Johnson RS, Moore C. (1999). EMBO J 18: Lin B, Ferguson C, White JT, Wang S, Vessella R, True LD 188–197. et al (1999). Cancer Res 59: 4180–4184. Wise SC, Burmeister LA, Zhou XF, Bubulya A, Oberﬁeld JL, Lin MF, Meng TC, Rao PS, Chang C, Schonthal AH, Lin FF. Birrer MJ et al (1998). Oncogene 16: 2001–2009. (1998). J Biol Chem 273: 5939–5947. Yeap BB, Krueger RG, Leedman PJ. (1999). Endocrinol 140: Linja MJ, Visakorpi T. (2004). J Steroid Biochem Mol Biol 92: 3282–3291. 255–264. Zheng Z, Cai C, Omwancha J, Chen S-Y, Baslan T, Martinez-Balbas MA, Bannister AJ, Martin K, Haus-Seuffert Shemshedini L. (2006). J Biol Chem 281: 4002–4012. P, Meisterernst M, Kouzarides T. (1998). EMBO J 17: Zhou HJ, Yan J, Luo W, Ayala G, Lin SH, Erdem H et al 2886–2893. (2005). Cancer Res 65: 7976–7983.

Oncogene