Oncogene (2007) 26, 2386–2394 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ONCOGENOMICS Ubiquitin E3 ligase WWP1 as an oncogenic factor in human prostate cancer

C Chen1,2, X Sun1, P Guo1, X-Y Dong1, P Sethi1, W Zhou1, Z Zhou2, J Petros1,3,4, HF Frierson Jr5, RL Vessella6, A Atfi7 and J-T Dong1,3,8

1Winship Cancer Institute and Department of Hematology and Oncology, Emory University School of Medicine, Atlanta, GA, USA; 2Center for Cell Biology and Cancer Research, Albany Medical College, Albany, NY, USA; 3Department of Urology, Emory University School of Medicine, Atlanta, GA, USA; 4Atlanta VA Medical Center, Decatur, GA, USA; 5Department of Pathology, University of Virginia Health System, Charlottesville, VA, USA; 6Department of Urology, University of Washington, Seattle, WA, USA; 7INSERM U482, Hoˆpital St-Antoine, Paris, France and 8Department of Genetics and Cell Biology, Nankai University College of Life Sciences, Tianjin, China

The gene for E3 ubiquitin ligase WWP1 is located at Introduction 8q21, a region frequently amplified in human cancers, including prostate cancer. Recent studies have shown that Prostate cancer is one of the most common solid tumors WWP1 negatively regulates the TGFb tumor suppressor in men. The molecular basis of prostate cancer is not pathway by inactivating its molecular components, well understood, although it has been recognized that including Smad2, Smad4 and TbR1. These findings identifying molecular alterations underlying the deve- suggest an oncogenic role of WWP1 in carcinogenesis, lopment and progression of prostate cancer will be but direct supporting evidence has been lacking. In this necessary for improving its detection and treatment. study, we examined WWP1 for gene dosage, mRNA Copy number gain or loss is a common genetic expression, mutation and functions in a number of human alteration in solid tumors. Many chromosomal regions prostate cancer samples. We found that the WWP1 gene have been identified for such changes in cancer by had copy number gain in 15 of 34 (44%) xenografts and comparative genomic hybridization (CGH) and other cell lines from prostate cancer and 15 of 49 (31%) clinical approaches, yet the underlying target genes remain to be prostate cancer samples. Consistently, WWP1 was over- identified and characterized for most cancers. The q21 expressed in 60% of xenografts and cell lines from band of chromosome 8 (8q21) shows frequent copy prostate cancer. Mutation of WWP1 occurred infre- number gain in prostate cancer, and the gain at 8q21 is quently in prostate cancer. Functionally, WWP1 over- often associated with aggressive behaviors of prostate expression promoted colony formation in the 22Rv1 cancer. For example, in men with clinically localized prostate cancer cell line. In PC-3 prostate cancer cells, prostate cancer, gain of 8q21 is associated with poor WWP1 knockdown significantly suppressed cell prolifera- outcome and metastasis (Nupponen et al., 1998; van tion and enhanced TGFb-mediated growth inhibition. Dekken et al., 2003; Rubin et al., 2004). Therefore, 8q21 These findings suggest that WWP1 is an oncogene that is believed to harbor an important oncogene for prostate undergoes genomic amplification at 8q21 in human cancer. prostate cancer, and WWP1 overexpression is a common Many important molecules involved in cell proli- mechanism involved in the inactivation of TGFb function feration, differentiation and carcinogenesis are tightly in human cancer. controlled by different mechanisms, including ubiquitin Oncogene (2007) 26, 2386–2394. doi:10.1038/sj.onc.1210021; proteasome pathway (UPP)-mediated protein degrada- published online 2 October 2006 tion. Many E3 ubiquitin ligases in the UPP have been implicated in cell cycle control and uncontrolled cell ONCOGENOMICS Keywords: WWP1; amplification; overexpression; prostate proliferation. Some E3 ligases such as MDM2 and cancer; KLF5; TGFb SKP2, which target p53 and CDKN1A/CDKN1B, are considered oncoproteins (Leite et al., 2001; Lu et al., 2002; Drobnjak et al., 2003), while some other E3 ligases are considered tumor suppressors. WWP1 is an E3 ubiquitin ligase that contains four tandem WW domains and a HECT domain. Published studies suggest that WWP1 plays a role in different biological processes Correspondence: Dr JT Dong, Winship Cancer Institute, Emory including regulation of epithelial sodium channels, University School of Medicine, 1365 Clifton Road, Room C4080, viral budding, receptor trafficking and transcription Atlanta, GA 30322, USA. E-mail: [email protected] (Malbert-Colas et al., 2003; Ingham et al., 2004; Received 27 April 2006; revised 26 June 2006; accepted 30 June 2006; Zhang et al., 2004). With regard to a role in published online 2 October 2006 cell proliferation and carcinogenesis, it has been WWP1 in prostate cancer C Chen et al 2387 demonstrated that WWP1 negatively regulates the Results TGFb tumor suppressor pathway by mediating the ubiquitination and degradation of multiple components Frequent copy number gain of the pathway, including Smad2 (Seo et al., 2004), The WWP1 gene is located in the q21 band of Smad4 (Moren et al., 2005), and TGFb receptor 1 chromosome 8 (8q21), which is frequently amplified in (TbR1) proteins (Komuro et al., 2004). WWP1 appears human prostate cancer, including the PC-3 prostate to be regulated by androgen, and is highly expressed cancer cell line (Porkka et al., 2002; Wang et al., 2004). in androgen-independent prostate cancer cells in the To evaluate if WWP1 undergoes genomic amplification LAPC-9 model (Gu et al., 2005). In our recent in prostate cancer, we first examined WWP1’s DNA study, we found that WWP1 also mediates the ubi- copy number in 34 prostate cancer cell lines/xenografts quitination and degradation of the KLF5 transcription by using real-time polymerase chain reaction (PCR) factor, which was identified as a candidate tumor assay, with normal human DNA as a normal control suppressor gene in prostate and breast cancers (Chen and PC-3 prostate cancer cell line as a control with et al., 2002, 2003a, 2005a). Consistent with a previous doubled WWP1 genome. Fifteen of the 34 cases (44%) study showing upregulation of WWP1 in multiple showed an increased copy number by at least one fold carcinoma cell lines (Komuro et al., 2004), we found (Figure 1a). Among the samples examined, LuCaP 23.1, that WWP1 is overexpressed in some prostate and LuCaP 23.12 and LuCaP 23.8 were from different breast cancer cell lines. Furthermore, WWP1 over- metastases of one patient. LuCap35 and LuCaP35V expression leads to excessive degradation of KLF5 in were also from one patient. Cell line 22Rv1 was derived these cancer samples (Chen et al., 2005a, b). These from the CWR22 xenograft. As all these samples were in findings suggest that WWP1 could have an oncogenic the non-amplified group, the actual amplification rate role in prostate cancer. for WWP1 could be higher than described in this study. In this study, we performed genetic and functional Conversely, none of the four un-transformed cell lines analyses to evaluate the role of WWP1 in human (PZ-HPV-7, PWR-1E, RWPE-1 and BPH55T) showed prostate cancer. Both copy number gain and over- a copy number change at WWP1. In agreement with expression were frequently detected in human prostate previous findings, the PC-3 prostate cancer cell line had cancer samples, although mutations were rare. Func- a doubled copy number for WWP1 when compared to tionally, WWP1 overexpression appeared to promote normal samples. Increased copy number at WWP1 was cell proliferation by antagonizing TGFb’s inhibitory also validated in six of the 34 samples by duplex PCR effect on cell proliferation. combined with gel electrophoresis (Figure 1a, inset),

Figure 1 Frequent copy number gain at WWP1 in prostate cancer. (a) WWP1 copy numbers in cell lines and xenografts from prostate cancer, as detected by real-time PCR. The KAI1 gene serves as a control for normalizing input DNA. Black bars indicate samples with WWP1 copy number gain. The empty bars indicate non-transformed prostate samples, while gray bars indicate tumor samples without obvious copy number gains. The inset shows WWP1 copy number gain in six prostate cancer samples detected by duplex PCR and gel electrophoresis. (b) Examples of WWP1 copy number gain in clinical prostate cancer specimens detected by duplex PCR assay. Case numbers are shown at the top. N, normal; T, tumor.

Oncogene WWP1 in prostate cancer C Chen et al 2388 with normal human DNA and BPH55T untransformed cell line as negative controls. To determine if WWP1 also has copy number gains in clinical prostate cancer samples, we examined 49 cases of prostate cancer and their matched normal cells by performing duplex PCR and gel electrophoresis, which was developed and used to detect gene dosage in our previous studies (Dong et al., 2000; Chen et al., 2003a). As shown in Figure 1b, the ratios between WWP1 signals and control KAI1 signals were significantly greater (>2) than that in the normal DNA in 15 of 49 (31%) tumors examined. It is worth noting that eight of 49 (16%) tumors appeared to have copy number losses at WWP1. No significant association was found between copy number change at WWP1 and age at diagnosis, tumor grade, or tumor stage, which could be due to the smaller sample size.

WWP1 expression is frequently upregulated in prostate cancer Copy number gain often causes overexpression for a gene. To test whether WWP1 is overexpressed in prostate cancer, we first examined WWP1 mRNA expression by real time PCR in 30 cell lines and xenografts from prostate cancer. Compared to the average level of WWP1 expression in six non-trans- formed prostate samples, the level of WWP1 expression was increased by at least onefold in 18 of 30 (60%) prostate cancer samples, including the PC-3 prostate Figure 2 Increased expression of WWP1 RNA in prostate cancer. cancer cell line (Figure 2a). (a) Expression of WWP1 mRNA in prostate cell lines and Northern blot analysis was performed to confirm xenografts detected by real time PCR. GAPDH was used as a WWP1 overexpression in prostate cancer. Whereas a control to normalize the cDNA input. Black bars identify the normal prostate tissue and five non-transformed pros- samples with WWP1 overexpression, while empty bars indicate non-transformed prostate samples and gray bars indicate tumor tate epithelial cell lines expressed little WWP1 RNA, samples without WWP1 overexpression. (b) Expression of WWP1 four of six prostate cancer cell lines, including DU145, RNA in prostate cancer cell lines detected by Northern blot PC-3, BRF-41T and LNCaP/C4-2, showed an increased analysis. Lanes 1–6 are non-transformed prostate samples, and the expression of WWP1 (Figure 2b). Except for BRF-41T, rest are prostate cancer cell lines. b-Actin serves as a RNA loading expression levels of WWP1 in these samples were control. (c) Protein expression of WWP1, TbR1 and Smad4 in cell lines from the prostate, as detected by western blot analysis. consistent between real-time PCR assay and northern b-Actin serves as a loading control. blot analysis. BRF-41T showed strong signals for both WWP1 and b-actin in northern blot analysis, but its WWP1 expression did not show a significant increase in these cell lines, because they have been demonstrated to real-time PCR analysis in which GAPDH was used as be negatively regulated by WWP1 (Komuro et al., 2004; the normalizing control. In addition, a smaller isoform Moren et al., 2005). In all but two (PZ-HPV-7 and of WWP1, shown as a band below the wild-type WWP1 LNCaP) of the six cell lines analysed, an inverse mRNA at 4.2 Kb, was clearly detected in some samples. correlation was noted between WWP1 expression and However, the isoform was not associated with cancer, the level of both TbRI and Smad4 (Figure 2c). because it was present in both cancer samples and non- In combination with the results of copy number gain tumor controls (Figure 2b). at WWP1, 11 of 15 cancer cell lines with an increased We also measured WWP1 protein levels, along with WWP1 copy number showed WWP1 overexpression, TbRI and Smad4, in one untransformed and five whereas only four of 18 samples without WWP1 copy prostate cancer cell lines, using Western blot analysis. number gain showed WWP1 overexpression (Figures 1 Consistent with mRNA expression results (Figure 2b), and 2) (Fisher exact test, P ¼ 0.004). WWP1 protein level is higher in prostate cancer cell lines PC-3, LNCaP and DU 145, but lower in the untrans- formed PZ-HPV7 cell line and 22Rv1 prostate cancer Mutation of WWP1 is rare in prostate cancer cell line (Figure 2c). An extra band below normal To determine if WWP1 is mutated in prostate cancer, WWP1 band was detected in PC-3 and LAPC-4 prostate the coding region of WWP1 was amplified by PCR with cancer cells but not in the rest of the cell lines. Protein two pairs of primers, and the overlapping PCR products expression of TbR1 and Smad4 was also analysed in were sequenced in 30 cell lines and xenografts from

Oncogene WWP1 in prostate cancer C Chen et al 2389 prostate cancer. Two sequence alterations were detected To further evaluate the effect of WWP1 on cell growth, in xenografts. One was 2393A-T (Glu798Val) in we treated PC-3 cells with siRNA for either WWP1 or CWR91 and the other was 721A-T (Thr241Ser) in the luciferase gene (Figure 3b and c). The latter served as LuCaP35. Both changes were heterozygous and had not a negative control. Cell growth was analysed also by the been reported in the SNP database at NCBI. SRB staining method. Real-time PCR assay showed that the siRNA silenced WWP1 expression by up to 80% (Figure 3c). As shown in Figure 3b, cells with WWP1 promotes cell proliferation reduced WWP1 expression grew significantly slower at To test whether WWP1 affects cell proliferation, we day 6 when compared to the control group (P ¼ 0.009). transfected both wild-type WWP1 and a dominant- These results also suggest that WWP1 plays a growth- negative mutant WWP1 (WWP1C886S) (Chen et al., promoting role in prostate cancer cells. 2005a) into 22Rv1 prostate cancer cells, which express a lower level of WWP1 (Figure 2). Ectopic expression of wild-type and mutant WWP1 in 22Rv1 cells was verified WWP1 antagonizes TGFb in suppressing cell by Western blot analysis (Figure 3a). As expected, wild- proliferation in PC-3 cells type WWP1 increased but mutant WWP1 suppressed Previous studies showed that WWP1 mediates the colony formation in 22Rv1 cells when compared to degradation of three components of the TGFb pathway, empty vector control (Figure 3a). Measurement of cell including Smad2, Smad4 and TbR1 (Komuro et al., numbers by SRB staining in different groups showed 2004; Seo et al., 2004; Moren et al., 2005). It is thus significant differences between both vector control and possible that WWP1 inactivates TGFb in epithelial cells. wild-type WWP1 (P ¼ 0.02) and between vector control To test this hypothesis, we analysed DNA synthesis and mutant WWP1 (P ¼ 0.02) (Figure 3a). These rates in PC-3 cells treated with different concentrations findings suggest that WWP1 promotes cell proliferation of TGFb in combination with siRNA for WWP1 or in 22Rv1 prostate cancer cells. luciferase control. Consistent with colony formation In our previous study, a siRNA for WWP1 was results, knockdown of WWP1 by siRNA transfection shown to efficiently knockdown WWP1 expression in significantly inhibited the DNA synthesis rate, which the PC-3 prostate cancer cell line (Chen et al., 2005a). indicates cell proliferation, in the presence or absence of

Figure 3 WWP1 promotes cell proliferation in different prostate cancer cell lines. (a) Compared to vector control (V), ectopic expressions of wild-type and mutant WWP1 respectively promoted or suppressed cell growth in 22Rv1 prostate cancer cells. Expression of WWP1 was verified by western blot analysis (lower panel). The P-values for the difference between vector control and wild-type WWP1 and that between vector control and mutant WWP1 were both 0.02. (b) Knockdown of WWP1 by RNAi (WWP1siRNA) slowed cell proliferation in PC-3 prostate cancer cells, as determined by growth curve analysis. The siRNA for luciferase (LucsiRNA) was used as a control. The P-value for the difference at day 6 is 0.009 (Student’s t-test). (c) Verification of RNAi-reduced WWP1 expression by real-time PCR assay. SiRNA treatments were in duplicate.

Oncogene WWP1 in prostate cancer C Chen et al 2390 TGFb (Figure 4a). Furthermore, whereas PC-3 cells molecules’ expression was not detectable when exoge- treated with control siRNA did not show an obvious nous TGFb was present (Figure 4b). In the induction of response to TGFb‘s inhibitory effect on cell prolifera- p15 expression, however, the effects of TGFb and tion, siRNA-mediated WWP1 knockdown significantly WWP1 knockdown were additive, which is consistent enhanced such an effect at both TGFb concentrations with an additive inhibitory effect on cell proliferation tested (Po0.05, Student’s t-test). These findings suggest (Figures 3b and 4a). that knockdown of WWP1 sensitizes cells to TGFb’s effect. To explore the mechanisms responsible for WWP1’s Discussion effect on TGFb activity, we examined protein expression of several components of the TGFb signaling pathway, In this study, we present multiple lines of evidence for an including Smad2, Smad3, Smad4 and TbR1, by Western oncogenic role of the WWP1 E3 ubiquitin ligase in blot analysis (Figure 4b). A downstream target molecule human prostate epithelial cells. Genetically, WWP1 is of TGFb, the CDK inhibitor p15, was also examined for located at 8q21, which contains a frequently amplified expression. Without exogenous TGFb, knockdown of locus in human prostate cancer, and showed frequent WWP1 by siRNA transfection increased expression copy number gain in cell lines, xenografts and clinical levels of Smad2, Smad4 and TbR1 in PC-3 cells, which samples from prostate cancer. WWP1 is also frequently is consistent with published studies implicating WWP1 overexpressed in prostate cancer samples, and its in the degradation of these molecules. As expected, overexpression is significantly associated with its knockdown of WWP1 also increased the expression of copy number gain. Functionally, ectopic expression of p15 but not the expression of Smad3 in the same cells. WWP1 enhanced and knockdown of WWP1 suppressed On the other hand, addition of exogenous TGFb to the cell proliferation. Furthermore, WWP1 antagonizes culture medium increased expression of Smad2, Smad4 TGFb’s inhibitory effect on cell proliferation, likely by and TbRI, but the effect of WWP1 knockdown on these mediating the degradation of key components of TGFb pathway including Smad2, Smad4 and TbRI as indi- cated in previous studies (Komuro et al., 2004; Seo et al., 2004; Moren et al., 2005). Therefore, WWP1 is a potential oncogene in prostate cancer. Because copy number gain at 8q21 has been detected in prostatic intraepithelial neoplasia (PIN) of both low and high grades (Zitzelsberger et al., 1998), which is considered a precursor of prostate cancer, it is likely that WWP1 overexpression could also play a role in early stages of prostate cancer development. The mutation of WWP1 also occurs in prostate cancer, although the frequency could be low. In 30 cancer samples examined, two showed sequence altera- tions. Although it is not clear if the change of Thr241Ser has any functional consequence, the change of Glu798- Val in the CWR91 xenograft should be a function- altering mutation, because the Glu798 codon is located in the HECT domain of WWP1, it is conserved in all HECT domains among different molecules, it partici- pates in intermolecular interaction between N-lobe and C-lobe to maintain protein conformation, and mutation of this codon from Glu to Ala significantly decreases WWP1’s ubiquitin transfer activity (Verdecia et al., 2003). WWP1 has multiple isoforms resulting from alter- native splicing (Flasza et al., 2002). Different isoforms have different domain structures and thus may function differently. Although a previous study of the breast cancer cell line T-47D suggests the presence of tumor- specific splicing of WWP1 (Flasza et al., 2002), our northern blot analysis showed that a smaller isoform of Figure 4 WWP1 antagonizes TGFb function in PC-3 cells. WWP1 is present in both normal and cancer cells, and (a) RNAi-mediated WWP1 knockdown (WWP1si) sensitized PC-3 this isoform is not related to cancer (Figure 2b). In to TGFb’s inhibitory effect on cell proliferation when compared addition, wild-type WWP1 is the dominant form that is to the negative control (Lucsi), as determined by DNA synthesis assay. (b) Knockdown of WWP1 increased protein levels of Smad2, overexpressed in some cancer samples. Smad4, TbR1 and p15 in PC-3 cells. All TGFb treatments were for In addition to WWP1, at least two other genes from 24 h. b-Actin served as a loading control. 8q21 have been shown to undergo copy number gain

Oncogene WWP1 in prostate cancer C Chen et al 2391 and overexpression in prostate cancer, including PrLZ/ E3 ubiquitin ligases are a group of scaffold proteins TPD52 (Rubin et al., 2004; Wang et al., 2004) and that recognize and conjugate ubiquitins to specific TCEB1 (Elongin C or transcription elongation factor B) substrates. There are hundreds of E3 ligases actively (Porkka et al., 2002). Although neither PrLZ/TPD52 functioning in mammalian cells (Pickart, 2001). Several nor TCEB1 has been shown to affect cell proliferation as E3 ligases, including Mdm2, Skp2 and b-TrCP, have WWP1, it is possible that each of these molecules plays been demonstrated to play an oncogenic role in human a role in prostatic carcinogenesis. A more definitive carcinogenesis; and frequent gene amplification and approach such as transgenic overexpression in mice is overexpression of these E3 ligases have been detected in needed to better define the role of these molecules in different types of human cancers. Our findings in this prostate cancer. Currently we are testing if prostate- study present WWP1 as another oncogenic E3 ligase. specific overexpression of WWP1 in the prostate can These E3 ligases can be therapeutic targets in the cause prostate cancer in mice. It is also necessary to treatment of cancer. For example, targeting Mdm2 can analyse protein expression of WWP1 in prostate cancer restore the apoptotic response of cells to androgen to determine whether WWP1 is also overexpressed in deprivation therapy in prostate cancer (Zhang et al., clinical samples and whether WWP1 overexpression 2003; Mu et al., 2004). It is possible that inactivating is associated with clinicopathologic characteristics of WWP1 can sensitize cells to TGFb’s inhibitory effect on prostate cancer. In addition, the MYC oncogene is cell proliferation, and thus targeting WWP1 could be a located in 8q24, another chromosomal region that useful therapeutic approach in treating cancers with undergoes frequent copy number gain in prostate WWP1 overexpression. We are currently testing this cancer. The role of MYC in prostate cancer has been possibility. well documented (Dong, 2006). It is possible that both In our previous studies, we identified the KLF5 WWP1 and MYC are overexpressed in some prostate transcription factor as a candidate tumor suppressor cancers, and thus have a cooperative role in prostatic gene that is inactivated by genomic deletion and carcinogenesis. transcriptional downregulation in human prostate and It is well known that TGFb is a potent tumor breast cancers (Chen et al., 2002, 2003a). We have suppressor in the development of cancer, but it becomes recently established WWP1 as an E3 ligase for a promoter of metastasis during cancer progression posttranslational regulation of KLF5 (Chen et al., (Massague et al., 2000; Roberts and Wakefield, 2003). 2005a). In cancer cell lines, we have shown that The TGFb signaling pathway involves a series of overexpression of WWP1 leads to excessive degradation molecules. Among them are the TGFb receptor 1 of KLF5 (Chen et al., 2005a, b). Taken together, the (TbR1), Smad2 and Smad4, which are negatively finding of copy number gain and overexpression of regulated by WWP1 at the protein level (Komuro et al., WWP1 in human prostate cancer in this study provides 2004; Seo et al., 2004; Moren et al., 2005). Consistently, another mechanism for inactivation of KLF5 in human we found that protein levels of TbR1, Smad2 and Smad4 cancer, that is, excessive protein degradation caused were increased when WWP1 expression was knocked by WWP1 overexpression. Interestingly, our ongoing down by siRNA in PC-3 cells. Furthermore, one of the studies suggest that KLF5 also mediates TGFb’s well-established TGFb targets involved in cell cycle function in the control of cell proliferation (unpublished control, the CDK inhibitor p15, was significantly induced data), which provides another line of evidence for at the protein level upon the knockdown of WWP1. WWP1’s antagonizing role in regulating the TGFb When exogenous TGFb was added, the expression of signaling pathway. Smad2, Smad4 and TbRI was increased with or without In summary, we found that the E3 ubiquitin ligase WWP1 knockdown, but WWP1 knockdown did not WWP1 undergoes frequent copy number gain and further increase their expression. For the expression of overexpression in human prostate cancer. WWP1 is p15, however, the effects of TGFb and WWP1 knock- also mutated in some prostate cancers. Functionally, down are additive. These results further suggest that, in WWP1 promotes cell proliferation, likely by antagoniz- addition to increased PI3K/AKT signaling, androgen ing TGFb’s inhibitory function. These results suggest receptor overexpression, and reduced TGFb receptors I that WWP1 could be a potential biomarker and and II which all have been shown to inhibit TGFb therapeutic target in the detection and treatment of signaling in prostate cancer, overexpression of WWP1 is prostate cancer. another mechanism that impairs TGFb’s inhibitory effect on cell proliferation. Escape from TGFb inhibition is considered an early event in prostate cancer. Several mechanisms have Materials and methods been found to overcome the growth inhibitory effect of TGFb, including increased PI3K/AKT/mTOR sig- Tumor specimens, cell lines and xenografts naling (van der Poel, 2004), overexpression of androgen In total, we analysed 83 prostate cancer samples, including 49 clinical tumor specimens and 34 cell lines and xenografts receptor (Danielpour, 2005), and loss of TGFb receptor derived from prostate cancer. Among the 49 tumor specimens (Kim et al., 1996). Whereas WWP1 overexpression with matched non-neoplastic cells, 40 were primary tumors, could lead to downregulation of TbRI, it remains to be six were hormone refractory local cancers and three determined whether WWP1 overexpression also affects were metastases. Patient age at diagnosis ranged from 50 to PI3K/AKT signaling and AR activity. 79 years (average 63.3). The Gleason scores for the tumors

Oncogene WWP1 in prostate cancer C Chen et al 2392 ranged from five to nine, with 23 having a Gleason score of five 50-AAGAATGGCATAGCACAAAC-30 and 50-TTCTGCAC or six, 22 having seven, and four having eight or nine. All TGGTAGAAGGAA-30. The sequence for this probe was tumor and matched normal cells were collected from H&E unique to WWP1 among different gene family members. The stained sections of formalin-fixed, paraffin-embedded tissues probe was labeled with a-32P-dCTP using a random primer by manual microdissection, as described in our previous study DNA labeling system (Invitrogen). For each sample, 15 mgof (Dong et al., 2000). All prostatic epithelial cell lines and total RNA were loaded into a denaturing gel, which was xenografts were described in our previous studies (Chen et al., electrophoresed. Nylon membrane with transferred RNA was 2003a, b; Sun et al., 2005). hybridized with labeled probe at 651C for 3 h in the QuikHyb hybridization solution (Stratagene, La Jolla, CA, USA), washed following the manufacturer’s instructions, and exposed Detection of copy number gain for WWP1 to X-ray film. WWP1 probe was stripped from the blot and the Two approaches were used to detect gene copy number membrane was re-hybridized with the b-actin probe. changes. One was real-time PCR with the SYBR green dye using the ABI 7000 Sequence Detection System (PE Applied Biosystems, Foster City, CA, USA). Primer sequences for Mutation detection human WWP1 were 50-CAGGAGGTTGACTTGGCAGA-30 In 30 cell lines and xenografts from prostate cancer, the and 50-AAACTGTGTCCAAAAGCAGTCTC-30; and those for WWP1 coding region was amplified by RT–PCR with two 0 control gene KAI1 were 50-CAGGTG GGCACGGGTTT-30 pairs of primers. Primers sequences were 5 -GAAAGAGGGA 0 0 and 50-TCCTCGCGACTGCTGTTGTA-30. Briefly, 10 ng of ATCGTGTCTTAC-3 and 5 -ACGATCATCAACTCTTCT 0 0 genomic DNA were amplified in a volume of 25 ml with TTCCC-3 for one pair, and 5 -GTATGGATCCTGTACGG 0 0 0 primers for either WWP1 or KAI1 following the manufac- CAGCA-3 and 5 -TCACTGTACAGAATGAACAGCTT-3 turer’s protocols. The KAI1 gene serves as a non-deleting for the other pair. After purification from gels using Qiagen’s normalizing control. All PCRs were performed in duplicate Gel Purification Kit, the PCR products were sequenced with 40 cycles using standard program with an annealing and with each of the PCR primers and an additional primer 0 0 extension temperature of 601C. Relative DNA copy number 5 -ACCTCCTGCATGCCACACAAC-3 by Macrogene (Korea). was determined by the DDCt method, and WWP1 copy number gains were indicated by the ratio of the WWP1 Colony formation assay reading to the KAI1 reading, with normal samples defined as 1. The 22Rv1 prostate cancer cell line was maintained in RPMI- Copy number gain was defined in a sample when the ratio of 1640 medium supplemented with 5% fetal bovine serum, WWP1 to KAI1 was equal or greater than two, based on the HEPES (0.1 M), sodium pyruvate (1 mM), sodium bicarbonate reading for the PC-3 cell line, which has doubled the WWP1 (0.15%), glucose (0.45%) and penicillin and streptomycin genome. (1%). Myc-tagged wild-type and mutant WWP1 constructs, For all clinical specimens and some cell lines/xenografts, we that is, myc-WWP1 and myc-WWP1C886S, have been also performed duplex PCR to detect copy number gain for described in our previous study (Chen et al., 2005a). In six- WWP1 following our established procedures (Dong et al., well tissue culture plates, 2 Â 105 cells were seeded into each 2000; Chen et al., 2003a). Genomic DNA is severely degraded well. After 1 day, WWP1 plasmids were transfected into cells in formalin-fixed clinical tissue specimens, so a pair of primers using the Lipofectamine 2000 reagent following the manufac- amplifying a small fragment of WWP1 (103 basepairs) were turer’s manual (Invitrogen). Selection medium containing used. Primer sequences for WWP1 were 50-GTATGGATCCT 1 mg/ml G418 was replaced 24 h later. At 2 weeks after growth GTACGGCAGCA-30 and 50-AACAGAGTTGCTACTTAA in the selection medium, cells were fixed, stained with 0.4% ATGCAAGTCAG-30). Again, KAI1 was used as a normaliz- SRB (Sulphorhodamine-B), and washed by 1% acetic acid. ing control, and the primer sequences were 50-GGTGGGA Cell densities were measured with a spectrometer at a 500 nm TGGTGCAGAGCGG-30, and 50-CAGAGACAACCCAGG wavelength, following a published procedure (Sun et al., 1997). AGGAC-30. Gel images were scanned, and band intensities were measured by using the IMAGE J computer program Western blot from NIH (http://rsb.info.nih.gov/ij). The ratio of signal Western blot analysis was performed as described in our intensity for WWP1 to that for the KAI1 control in a tumor previous study (Chen et al., 2005b). Antibodies for Smad2 and sample was compared to the ratio in normal human DNA to Smad3 were purchased from Zymed Laboratories, and define a DNA copy number change. antibodies for Smad4, TbRI and p15 were purchased from Cell Signaling. Anti-WWP1 antibody was described in a WWP1 expression analyses previous study (Seo et al., 2004). Total RNA was extracted using the Trizol reagent following its standard protocol (Invitrogen). The cDNA was prepared by Knockdown of WWP1 and related cell assays using the Script kit according to the manufacturer’s protocol A siRNA with the sequence of 50-GAGTTGATGATCGTAG (Bio-Rad). Real-time RT–PCR with SYBR green dye was AAG-30 was used to target WWP1. A siRNA for the luciferase performed as described above with different primers. Primer gene (50-CTTACGCTGAGTACTTCGA-30) was used as a sequences for WWP1 were 50-GTATGGATCCTGTACGGC negative control. PC-3 cells were transfected with 200 nM of AGCA-30 (forward) and 50-GTTGTGGTCTCTCCCATGTG chemically synthesized siRNA (Dharmacon, Chicago, IL, GT-30 (reverse); and those for GAPDH were 50-GTGGTCCA USA) using Oligofectamine (Invitrogen) in 12-well plates. GGGGTCTT ACTC-30 and 50-AACGGGAAGCTCACTGG RNA was collected 48 h after transfection. C-30. All real-time PCRs were performed in duplicate. The For growth curve analysis, PC-3 cells were plated into 12-well average ratio of WWP1 to GAPDH control in PZ-HPV7 was plates at 4 Â 104 cells per well. Transfection was conducted 24 h defined as 1, and the ratios for other samples were normalized later. At 2, 4 and 6 days after transfection, cells were fixed with accordingly. WWP1 overexpression was defined in a sample 10% TCA and stained with SRB as described above. when the ratio of WWP1 to GAPDH was X2. For DNA synthesis assay, PC-3 cells were seeded into For the northern blot analysis, a probe was prepared 12-well plates at 4 Â 104 cells per well, and labeled with (14C)- from the WWP1 coding region by PCR with primers thymidine (0.02 mCi/ml) overnight. SiRNAs for WWP1 and

Oncogene WWP1 in prostate cancer C Chen et al 2393 control were transfected into cells for 24 h. Cells were then Two-tail Student’s t-test was used to determine statistical treated with TGFb overnight, and labeled with 3H-thymidine significance between experimental and control groups in cell (1 mCi/ml) for 4 h. Upon the completion of treatment, cells proliferation assay, and a P-value smaller than 0.05 was were washed with PBS and treated with 10% TCA to release considered statistically significant. genomic DNA. Radiolabeled DNA was solubilized by 120 ml of 0.3 N NaOH and transferred to glass fiber filter membranes, and radioactivity for 3H and 14C was measured using a Acknowledgements scintillation counter. The DNA synthesis rate was defined by the ratio of incorporated 3Hto14C. C Chen is an AFUD/AUA Research Scholar. This work was supported in part by grants from the National Cancer Institute Statistics analysis (Grant # CA87921), from the Department of Defense Prostate Fisher’s exact test was used to analyse the correlation between Cancer Research Program (Grant # DAMD17-03-2-0033), copy number or expression change with age at diagnosis, from the Georgia Cancer Coalition, and from the Susan G tumor grade or tumor stage in all prostate cancer samples. Komen Breast Cancer Foundation.

References

Chen C, Bhalala HV, Qiao H, Dong JT. (2002). A possible forming growth factor-beta (TGF-beta) signaling by tumor suppressor role of the KLF5 transcription factor in WW domain-containing protein 1 (WWP1). Oncogene 23: human breast cancer. Oncogene 21: 6567–6572. 6914–6923. Chen C, Bhalala HV, Vessella RL, Dong JT. (2003a). KLF5 is Leite KR, Franco MF, Srougi M, Nesrallah LJ, Nesrallah A, frequently deleted and down-regulated but rarely mutated in Bevilacqua RG et al. (2001). Abnormal expression prostate cancer. Prostate 55: 81–88. of MDM2 in prostate carcinoma. Mod Pathol 14: Chen C, Hyytinen ER, Sun X, Helin HJ, Koivisto PA, 428–436. Frierson HF et al. (2003b). Deletion, mutation, and loss of Lu L, Schulz H, Wolf DA. (2002). The F-box protein SKP2 expression of KLF6 in human prostate cancer. Am J Pathol mediates androgen control of p27 stability in LNCaP human 162: 1349–1354. prostate cancer cells. BMC Cell Biol 3: 22. Chen C, Sun X, Guo P, Dong XY, Sethi P, Cheng X et al. Malbert-Colas L, Fay M, Cluzeaud F, Blot-Chabaud M, (2005a). Human Kruppel-like factor 5 is a target of the E3 Farman N, Dhermy D et al. (2003). Differential expression ubiquitin ligase WWP1 for proteolysis in epithelial cells. and localisation of WWP1, a Nedd4-like protein, in J Biol Chem 280: 41553–41561. epithelia. Pflugers Arch 447: 35–43. Chen C, Sun X, Ran Q, Wilkinson KD, Murphy TJ, Massague J, Blain SW, Lo RS. (2000). TGFbeta signaling in Simons JW et al. (2005b). Ubiquitin-proteasome degradation growth control, cancer, and heritable disorders. Cell 103: of KLF5 transcription factor in cancer and untransformed 295–309. epithelial cells. Oncogene 24: 3319–3327. Moren A, Imamura T, Miyazono K, Heldin CH, Moustakas A. Danielpour D. (2005). Functions and regulation of transform- (2005). Degradation of the tumor suppressor Smad4 by WW ing growth factor-beta (TGF-beta) in the prostate. Eur J and HECT-domain ubiquitin ligases. J Biol Chem 280: Cancer 41: 846–857. 22115–22123. Dong JT. (2006). Prevalent mutations in prostate cancer. Mu Z, Hachem P, Agrawal S, Pollack A. (2004). Antisense J Cell Biochem 97: 433–447. MDM2 oligonucleotides restore the apoptotic response of Dong JT, Chen C, Stultz BG, Isaacs JT, Frierson Jr HF. prostate cancer cells to androgen deprivation. Prostate 60: (2000). Deletion at 13q21 is associated with aggressive 187–196. prostate cancers. Cancer Res 60: 3880–3883. Nupponen NN, Kakkola L, Koivisto P, Visakorpi T. (1998). Drobnjak M, Melamed J, Taneja S, Melzer K, Wieczorek R, Genetic alterations in hormone-refractory recurrent prostate Levinson B et al. (2003). Altered expression of p27 and Skp2 carcinomas. Am J Pathol 153: 141–148. proteins in prostate cancer of African-American patients. Pickart CM. (2001). Mechanisms underlying ubiquitination. Clin Cancer Res 9: 2613–2619. Annu Rev Biochem 70: 503–533. Flasza M, Gorman P, Roylance R, Canfield AE, Baron M. Porkka K, Saramaki O, Tanner M, Visakorpi T. (2002). (2002). Alternative splicing determines the domain structure Amplification and overexpression of Elongin C gene of WWP1, a Nedd4 family protein. Biochem Biophys Res discovered in prostate cancer by cDNA microarrays. Lab Commun 290: 431–437. Invest 82: 629–637. Gu Z, Rubin MA, Yang Y, Deprimo SE, Zhao H, Horvath S Roberts AB, Wakefield LM. (2003). The two faces of et al. (2005). Reg IV: a promising marker of hormone transforming growth factor beta in carcinogenesis. Proc refractory metastatic prostate cancer. Clin Cancer Res 11: Natl Acad Sci USA 100: 8621–8623. 2237–2243. Rubin MA, Varambally S, Beroukhim R, Tomlins SA, Ingham RJ, Gish G, Pawson T. (2004). The Nedd4 Rhodes DR, Paris PL et al. (2004). Overexpression, ampli- family of E3 ubiquitin ligases: functional diversity fication, and androgen regulation of TPD52 in prostate within a common modular architecture. Oncogene 23: cancer. Cancer Res 64: 3814–3822. 1972–1984. Seo SR, Lallemand F, Ferrand N, Pessah M, L’Hoste S, Kim IY, Ahn HJ, Zelner DJ, Shaw JW, Sensibar JA, Kim JH Camonis J et al. (2004). The novel E3 ubiquitin ligase Tiul1 et al. (1996). Genetic change in transforming growth factor associates with TGIF to target Smad2 for degradation. beta (TGF-beta) receptor type I gene correlates with EMBO J 23: 3780–3792. insensitivity to TGF-beta 1 in human prostate cancer cells. Sun SY, Yue P, Dawson MI, Shroot B, Michel S, Lamph WW Cancer Res 56: 44–48. et al. (1997). Differential effects of synthetic nuclear retinoid Komuro A, Imamura T, Saitoh M, Yoshida Y, Yamori T, receptor-selective retinoids on the growth of human non- Miyazono K et al. (2004). Negative regulation of trans small cell lung carcinoma cells. Cancer Res 57: 4931–4939.

Oncogene WWP1 in prostate cancer C Chen et al 2394 SunX,FriersonHF,ChenC,LiC,RanQ,OttoKBet al. androgen-responsive gene of the TPD52 family, amplified in (2005). Frequent somatic mutations of the transcription factor chromosome 8q21.1 and overexpressed in human prostate ATBF1 in human prostate cancer. Nat Genet 37: 407–412. cancer. Cancer Res 64: 1589–1594. van Dekken H, Alers JC, Damen IA, Vissers KJ, Krijtenburg Zhang X, Srinivasan SV, Lingrel JB. (2004). WWP1-depen- PJ, Hoedemaeker RF et al. (2003). Genetic evaluation of dent ubiquitination and degradation of the lung Kruppel- localized prostate cancer in a cohort of forty patients: gain of like factor, KLF2. Biochem Biophys Res Commun 316: distal 8q discriminates between progressors and nonpro- 139–148. gressors. Lab Invest 83: 789–796. Zhang Z, Li M, Wang H, Agrawal S, Zhang R. (2003). van der Poel HG. (2004). Mammalian target of rapamycin and Antisense therapy targeting MDM2 oncogene in prostate 3-phosphatidylinositol 3-kinase pathway inhibition enhances cancer: Effects on proliferation, apoptosis, multiple gene growth inhibition of transforming growth factor-beta1 in expression, and chemotherapy. Proc Natl Acad Sci USA 100: prostate cancer cells. J Urol 172: 1333–1337. 11636–11641. Verdecia MA, Joazeiro CA, Wells NJ, Ferrer JL, Zitzelsberger H, Kulka U, Lehmann L, Walch A, Smida J, Bowman ME, Hunter T et al. (2003). Conformational Aubele M et al. (1998). Genetic heterogeneity in a prostatic flexibility underlies ubiquitin ligation mediated by the WWP1 carcinoma and associated prostatic intraepithelial neoplasia HECT domain E3 ligase. Mol Cell 11: 249–259. as demonstrated by combined use of laser-microdissection, Wang R, Xu J, Saramaki O, Visakorpi T, Sutherland WM, degenerate oligonucleotide primed PCR and comparative Zhou J et al. (2004). PrLZ, a novel prostate-specific and genomic hybridization. Virchows Arch 433: 297–304.

Oncogene