NIH Public Access Author Manuscript Oncogene

NIH Public Access Author Manuscript Oncogene. Author manuscript; available in PMC 2014 May 09.

NIH-PA Author ManuscriptPublished NIH-PA Author Manuscript in final edited NIH-PA Author Manuscript form as: Oncogene. 2014 May 8; 33(19): 2464–2477. doi:10.1038/onc.2013.203.

Parathyroid hormone-related protein inhibits DKK1 expression through c-Jun-mediated inhibition of β-Catenin activation of the DKK1 promoter in prostate cancer

H. Zhang1,6, C. Yu1,2,6, J. Dai1, JM. Keller1, A. Hua1, JL. Sottnik1, G. Shelley1, CL. Hall1, SI. Park3, Z. Yao2, J. Zhang4, LK. McCauley3,5, and ET. Keller1,5 1Department of Urology, School of Medicine, University of Michigan, Ann Arbor, MI 48109 2Department of Immunology, Tianjin Key Laboratory of Cellular and Molecular Immunology, Key Laboratory of Educational Ministry, Tianjin Medical University, Tianjin, China; Tianjin 300070, China 3Department of Periodontics and Oral Medicine, School of Dentistry, University of Michigan, MI 48109 4Center for Translational Medical Research, Guangxi Medical University, Guangxi, China 5Department of Pathology, School of Medicine, University of Michigan, MI 48109

Abstract Prostate cancer bone metastases are unique in that that majority of them induce excessive mineralized bone matrix, through undefined mechanisms, as opposed to most other cancers that induce bone resorption. Parathyroid hormone-related protein (PTHrP) is produced by prostate cancer cells and intermittent PTHrP exposure has bone anabolic effects suggesting PTHrP could contribute to the excess bone mineralization. Wnts are bone productive factors produced by prostate cancer cells and the Wnt inhibitor DKK1 has been shown to promote prostate cancer progression. These findings, in conjunction with the observation that PTHrP expression increases and DKK1 expression decreases as prostate cancer progresses led to the hypothesis that PTHrP could be a negative regulator of DKK1 expression in prostate cancer cells, and hence allow the osteoblastic activity of Wnts to be realized. To test this, we first demonstrated that PTHrP downregulated DKK1 mRNA and protein expression. We then found through multiple mutated DKK1 promoter assays that PTHrP, through c-Jun activation, downregulated the DKK1 promoter through a TCF-response element site. Furthermore, chromatin immunoprecipitation (ChIP) and reChIP assays revealed that PTHrP-mediated this effect through inducing c-Jun to bind to a transcriptional activator complex consisting of β-catenin that binds the most proximal DKK1 promoter TCF-response element. Together, these results demonstrate a novel signaling linkage between PTHrP and Wnt signaling pathways that results in downregulation of a Wnt inhibitor allowing for Wnt activity that could contribute the osteoblastic nature of prostate cancer.

To whom correspondence should addressed: Evan T. Keller, Department of Urology, University of Michigan Medical School, 5308 CC, 1500 East Medical Center Drive, Ann Arbor, MI 48109. Phone 734-615-0280. [email protected]. 6these authors contributed equally to this project. Zhang et al. Page 2

Keywords

NIH-PA Author Manuscript NIH-PA Author ManuscriptProstate cancer; NIH-PA Author Manuscript DKK1; PTHrP; Wnt/β-catenin signaling; skeletal metastasis

Introduction Prostate cancer (PCa), a frequently occurring cancer of men in Western countries, leads to bone metastases in over 90% of patients with progressive disease. In contrast to the majority of bone metastatic tumors that produce primarily osteolytic (i.e., bone resorptive) lesions, PCa produces primarily osteoblastic (i.e., bone productive) lesions (reviewed in 1, 2). The mechanisms through which PCa promotes osteoblastic lesions are not clearly defined. Elucidation of these underlying mechanisms holds promise for the identification of therapeutic targets that will benefit PCa patients.

Cellular signaling pathways are important in onset, early progression, and metastasis of multiple cancers and thus offer a fertile area for research. The Wnt signaling pathway has been shown to play a major role in the early events of tumor initiation in colon cancer. However, less is known regarding the role of Wnts once a tumor is established. The Wnt family contains 19 secreted homologue Wnt proteins, 10 Frizzled (Fzd) receptors; two low density lipoprotein-related proteins 5 and 6(LRP5/6) and inhibitory antagonists such as WIF, Dickkopfs (DKK) and secreted frizzled-related protein (sFRP) (3, 4). Wnt signaling is divided into two pathways, the canonical pathway that invokes activation of β-catenin/TCF- target gene activation and the non-canonical pathway that induces signaling independent of β-catenin (5). The key role that Wnt signaling plays in bone biology suggests that the Wnt pathway may have a role in PCa bone metastasis. Along these lines, we have previously identified that PCa produces many Wnts as well as the Wnt inhibitor dickkopf-1 (DKK1) which diminishes PCa-induced osteoblastic activity through blocking Wnt activity (6). More recently, we identified that as PCa progresses from primary tumors to metastases in clinical PCa tissues, DKK1 expression decreases, which could allow Wnt activity to manifest and account for the osteoblastic nature of PCa (7). However, the mechanisms that account for decreasing DKK1 expression in PCa bone metastases are unknown.

Parathyroid hormone (PTH)-related protein (PTHrP), a peptide hormone that shares biologic activity with PTH, has been shown to mediate both bone catabolic (i.e., bone resorptive) and bone anabolic (i.e., bone productive) effects in multiple cancers including PCa (8–11). PTHrP is produced by PCa cells and its expression increases as PCa progresses to the metastatic state (12, 13). PTHrP has been associated with cancer progression, including metastasis (14). The observation that PTHrP expression increases and DKK1 expression decreases during progression of PCa to osteoblastic metastases led us to explore whether the osteoblastic activity of PCa is dependent on PTHrP-mediated regulation of DKK1 expression.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 3

Results

NIH-PA Author ManuscriptPTHrP NIH-PA Author Manuscript modulates NIH-PA Author Manuscript DKK1 expression in prostate cancer cells While the expression of DKK1 is high in primary PCa it decreases as PCa metastasizes to bone (7). In contrast, PTHrP expression in PCa metastases becomes elevated as PCa histological grade progresses(13, 15). To explore if PTHrP could regulate DKK1 expression to account for the inversely related expression of PTHrP and DKK1, we treated PC-3 cells and PC-3M cells, a highly metastatic variant PC-3, with PTHrP and checked the DKK1 expression. PTHrP (1–34) (10−7M) decreased DKK1 mRNA expression by approximately 60% in PC-3 and 85% in PC-3M cells by 8 hours, with the initial decrease observed as early as 2 hours in both cell lines (Figure 1A). Similarly, PTHrP decreased DKK1 protein expression over 16 hours in both cell lines (Figure 1B). Additionally PTHrP (10−8M – 10−6M) induced a dose-responsive decrease of DKK1 protein expression (Figure 1C and 1E; top row). Similarly, PTHrP (10−8M) treatment decreased DKK1 expression in Du145 prostate cancer cells (Supplementary Figure 1A).

PTHrP modulates transcriptional activity of genes through the AP-1 transcription factor composed of heterodimers and homodimers of c- Jun and c-Fos family members (16). This suggests that PTHrP may modulate DKK1 expression through c-Jun and c-Fos family members. Additionally, β-catenin, which modifies activity of several transcription factors, induces DKK1 expression; which gives rise to the possibility that PTHrP downregulates DKK1 expression through modulation of β-catenin. Accordingly, we examined if PTHrP regulates the c-Jun family of transcription factors and β-catenin in the prostate cancer cells. PTHrP (10−8M – 10−6M) increased c-Jun protein expression but had no impact on Jun B, Jun D or β-catenin expression in PC-3 cells (Figure 1C and 1E) or Du145 cancer cells (Supplementary Figure 1A). Similarly, PTHrP (10−7M) increased c-Jun protein expression as soon as 4 hours after treatment and maintained the increase through 16 hours (Figure 1G). To extend our findings to an additional cancer type that commonly metastasizes to bone; we also evaluated the MDA-MB231 breast cancer cell line. PTHrP induced similar effects in the MB231 cells on DKK1 and c-Jun (Supplementary Figure 1B). In summary, PTHrP induced an increase in c-Jun expression that was associated with a decrease in DKK1 expression in multiple cell lines.

Although PTHrP did not alter β-catenin levels, to exclude the possibility that PTHrP induced translocation of β-catenin from cytoplasm to nucleus, protein extract from cytoplasmic and nuclear fraction were performed in PC-3, PC-3M, DU145 and MDA MB231 cell lines. The results showed no significant change in the amount of β-catenin from either cytoplasmic or nuclear fraction between the different time course of PTHrP treatment (Figures. 1D and 1F, and Supplementary Figures. 1C and 1D). These results indicated that PTHrP decreased DKK1 expression independent of β-catenin translocation from the cytoplasm to nucleus.

PTHrP mediates its activity, in part, through induction of cAMP. Forskolin can be used to mimic PTHrP activity through cAMP induction (17, 18). To provide support for the possibility that PTHrP mediates downregulation of DKK1 expression through this classical cAMP mediated pathway, we treated PC-3 cells with forskolin and evaluated DKK1 mRNA

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 4

expression. Forskolin diminished DKK1 mRNA expression (Figure 1H) consistent with the concept that DKK1 expression is down regulated via classical PTHrP-mediated PKA/cAMP

NIH-PA Author Manuscript NIH-PA Author Manuscriptsignaling. NIH-PA Author Manuscript

The results, thus far, indicated that PTHrP was functioning in the PCa cell lines. However, in a previous study, PTH/PTHrP receptor (PPR) was not found by Northern blot and it was observed that PTH(1–34) fragment did not bind to PC-3 cells suggesting that the PPR was not present (12). In contrast, several other groups identified PTHrP is functional in PC-3 cells and identified PPR in clinical prostate cancer tissues (19, 20). To clarify the expression of PTHrP receptor in the cell lines used, we first authenticated the cell lines we used by subjecting them to short tandem repeat (STR) DNA analysis which confirmed their identity (Supplementary Table 1). We then evaluated the cells for PPR protein expression using HEK293T cells derived from human embryonic kidney cells as a positive control (21). PPR was identified as only one band of around molecular weight 82kDa in all cell lines (Supplementary Figure 1E). In addition to confirming its presence, we wanted to confirm the PPR was functional based on classic criteria of PPR activation, i.e., induction of cAMP, in these cell lines. To perform this, cells were exposed to PTHrP and induction of cAMP was measured. After normalization for unstimulated cAMP levels using a control group, PTHrP induced cAMP levels in all cell lines in a dose-responsive fashion (Supplementary Figure 1F). Taken together, the presence of PPR protein expression and induction of cAMP indicate the PCa cells express functional PTH/PTHrP receptor.

PTHrP knockdown increased DKK1 not β-catenin or TCF4 As presence of functional PPR was confirmed, we further explored the mechanism through which PTHrP modulated DKK1 expression. Endogenous PTHrP was knocked down in PC-3M cells and c-Jun and DKK1 expression evaluated. Knockdown of PTHrP, using two different shRNAs, decreased c-Jun and increased DKK1 protein levels with no impact on total β-catenin levels (Figure 2A). To determine if knocking PTHrP down impact the intracellular localization of β-catenin, cytoplasmic and nuclear extract was isolated from the stably-transfected scrambled and PTHrP shRNA cell lines. Knockdown of PTHrP had no impact on cytoplasmic or nuclear β-catenin levels (Figure 2B). To address the importance of c-Jun in PTHrP-mediated down-regulation of DKK1, we transfected the PC-3M cells with a c-Jun expression vector to rescue Jun expression when PTHrP was knocked down Re- expression of c-Jun alone in parental PC-3M cells decreased DKK1 expression (Figure 2C, Scram/JUN vs. Scram/V lanes). Knockdown of PTHrP induced DKK1 expression (Figure 2B Scram/V vs. PTHrPSh1/v) but re-expression of c-Jun blocked the increased DKK1 expression induced by knockdown of PTHrP (Figure 2C, PTHrPSh2/JUN vs. PTHrPSh2/v lanes). These results demonstrate that PTHrP downregulates DKK1 expression through c- Jun.

The observation that DKK1 expression was decreased through PTHrP-mediated induction of c-Jun suggested that PTHrP has an effect on DKK1 promoter activity as DKK1 expression was impacted at the mRNA level and c-Jun is a transcription factor. Accordingly, we assessed if PTHrP inhibited DKK1 expression through repression of the DKK1 promoter. A 3Kb DNA fragment of the DKK1 promoter 5′ upstream of the DKK1 transcription initiation

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 5

site was cloned (pGL3-DKK1p-Luc reporter) using PC-3M genomic DNA as template then used to evaluate the effect of PTHrP on its activity. Consistent with their effect on DKK1

NIH-PA Author Manuscript NIH-PA Author Manuscriptexpression, NIH-PA Author Manuscript both PTHrP and forskolin inhibited basal DKK1 promoter reporter activity by approximately 70% (Figure 2D).

As PTHrP repressed basal DKK1 promoter activity, we next wanted to determine if PTHrP could repress induced DKK1 promoter activity. Canonical Wnt signaling mediates its transcriptional activity through promoting β-catenin-mediated activation of transcription factor activity. Thus, we first evaluated whether β-catenin could induce DKK1 promoter activation. To perform this, a β-catenin expression plasmid was co-transfected with pGL3- DKK1p-Luc reporter in 293T cells (the degree of ectopic expression of β-catenin is shown in Supplementary Figure 2C). β-catenin induced promoter activity by approximately 4-fold within 24 hours (Figure 2E). These results indicated that β-catenin activates the DKK1 promoter.

The observation that PTHrP was able to block DKK1 promoter activation (Figure 2D) combined with the observation that it does not alter β-catenin expression (Figure 2A) suggests that PTHrP mediates its activity on the DKK1 promoter independent of regulating β-catenin expression. To provide further support for the ability of PTHrP to impact the promoter activity independent of a requirement for changing β-catenin expression, we stably transfected β-catenin into PC-3M cells and evaluated the impact of PTHrP on DKK1 promoter activity in these cells. Overexpression of β-catenin in PC-3M increased both DKK1 and c-JUN protein level; whereas, PTHrP increased c-Jun but decreased DKK1 level in both PC-3M/v and PC-3M/β-catenin cell (Figure 2F). Overexpression of β-catenin induced DKK1 reporter activity by more than 3.5-fold compared to control-transfected cell lines (Figure 2G). Addition of PTHrP inhibited β-catenin-induced DKK1 promoter activity by approximately 38% (Figure 2G). These results indicated that even in excess of β-catenin expression PTHrP can suppress DKK1 promoter activity. PTHrP may achieve this effect through either impacting the DKK1 promoter independently of any effect on β-catenin or through inhibiting β-catenin’s function on the promoter.

PTHrP inhibits DKK1 expression through WRE motifs in the DKK1 promoter The above findings are consistent with PTHrP inhibiting β-catenin-mediated activation of the DKK1 promoter independent of changes in β-catenin levels. This may occur either through PTHrP promoting binding of inhibitory transcription factors to the promoter, inducing release of activating transcription factors from the promoter or competing with β- catenin on Wnt-responsive elements (WRE). To explore for this latter possibility, we initially set out to identify potential cis-acting sites through which PTHrP could interfere with β-catenin-induced activation of the DKK1 promoter. Accordingly, the 3Kb DKK1 promoter was serially deleted to 1.8Kb, 1.0Kb and 0.5Kb (Figure 3A). Additionally, transcription element analysis (http://www.cbil.upenn.edu/cgi-bin/tess/tess) was performed to identify putative WRE (i.e. TCF/β-catenin binding sites) in the 3Kb DKK1 promoter fragment. Five putative sites were identified and then mutated using site-directed mutagenesis (sites indicated in Figure 3A). All mutant promoter reporters were co- transfected with β-catenin plasmid (or empty vector control) and promoter activity was

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 6

quantified. The largest deletion (delK/KA construct) resulted in reduction of β-catenin- induced promoter activity by 60%. It should be noted this deletion did not include the most

NIH-PA Author Manuscript NIH-PA Author Manuscriptproximal NIH-PA Author Manuscript two putative WRE sites (sites 1 and 2). Mutation specifically of site 2 (Mut2 construct) reduced β-catenin-induced promoter activity by approximately 90% (Figure 3A); whereas, mutation of site 1 reduced promoter activity by approximately 70%. These results indicate site 2 is the main mediator of β-catenin-induced DKK1 promoter activation, albeit other sites contribute to the overall activity.

TCF4 is a Wnt-responsive transcription factor that binds onto its WRE in a non-active fashion. Upon stimulation by Wnts, β-catenin translocates to the nucleus and promotes transcriptional activation through changing the conformation of TCF4 to an active state. To further explore if DKK1 expression is mediated through this classical Wnt pathway, we knocked down TCF4 expression. Knockdown of TCF4 decreased DKK1 expression in PC-3M cells (Figure 3B), indicating that DKK1 is regulated by TCF4. To determine if β- catenin-induced DKK1 expression was mediated through TCF activation of the DKK1 promoter, we co-expressed β-catenin and TCF4 or dominant negative TCF and measured activation of the DKK1 promoter (TCF protein levels are shown in Supplementary Figure 2A). TCF4 stimulated β-catenin-induced expression of the DKK1 promoter; whereas, dominant negative TCF4 blocked β-catenin-induced activation of the DKK1 promoter (Figure 3C). In contrast, neither β-catenin, TCF4 nor dominant negative TCF had an effect on the WRE-2 mutant promoter (Figure 3C). Taken together, these results indicate that TCF4/β-catenin mediates activation of the DKK1 promoter through TCF4 activity on the WRE-2 element.

To determine if PTHrP inhibits the DKK1 promoter through activity on the WRE-2 element, PC-3M/v and PC-3M/β-catenin stable cell lines (as shown in Figure 2F) were transfected with wildtype or the WRE mutant reporters then treated with 10−7M PTHrP for 24 hours. β- catenin activated the wild type and all mutant reporters except the WRE-2 mutant. PTHrP decreased β-catenin-induced DKK1 promoter activation by 1.5 to 2 fold except in the WRE-2 mutant (Figure 3D). This indicates that PTHrP mediates its inhibitory effect on the DKK1 promoter through the WRE-2 site.

c-Jun downregulates DKK1 expression through WRE-2 in DKK1 promoter The observation that PTHrP induces c-Jun, while decreasing DKK1 expression, is consistent with the possibility that PTHrP mediates repression of DKK1 expression through c-Jun. Thus, to determine if PTHrP mediates its activity on the DKK1 promoter through c-Jun, we examined the effect of several mutated forms of c-Jun on their ability to impact DKK1 promoter reporter activity (Supplementary Figure 2B). The c-Jun mutated forms used were: TAM67, a dominant negative form of c-Jun due to deletion of the transactivation domain; DBM3, which contains a mutated DNA binding domain due to insertion of 5 amino acids (EEGIE); and LZM, that contains a deletion of the leucine zipper domain which destroys the ability of c-Jun to dimerize (22, 23). To test effects of c-Jun and its mutants on DKK1 reporter activities, the plasmid, pGL3-DKK1p-luc with 3Kb DKK1 promoter-driven luciferase, was used as a reporter. We used the transcription factor SOX9, a known inhibitor of β-catenin signaling pathway (24) as a control for inhibition of the DKK1 promoter

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 7

reporter. We used SP1, which has no known interaction with β-catenin signaling, as a negative control. We first confirmed SP1, SOX9 and β-catenin overexpression NIH-PA Author Manuscript NIH-PA Author Manuscript(Supplementary NIH-PA Author Manuscript Figure 2C). SP1 had no effect on β-catenin-induced activation of the DKK1 promoter in 293T cells; whereas, SOX9 and c-Jun reduced β-catenin-mediated activation of the DKK1 promoter by approximately 50% and 80%, respectively (Supplementary Figure 2D). These results indicate that both SOX9 and that c-Jun inhibit β-catenin-induced activation of the DKK1 promoter.

To confirm the ability of c-Jun to inhibit DKK1 promoter activity, 293T cells were transfected with plasmids that express the various c-Jun forms. Proteins were of the expected sizes (Supplementary Figure 2E). The 293T cells were co-transfected with either pcDNA3-β-catenin or pcDNA3 empty vector combined with the various c-Jun constructs. c- Jun inhibited basal promoter activity by 60%; whereas, Tam67 and LZM constructs had no effect and DBM3 actually enhanced reporter activity 2.2-fold compared to empty vector control (Figure 4A). In addition to impacting basal promoter activity, c-Jun inhibited; whereas, DBM3 and LZM promoted β-catenin-induced DKK1 promoter activation. Tam67 had minimal impact on DKK1 basal or induced activity. Taken together, these results provide further evidence that c-Jun inhibits both basal and β-catenin-mediated activation of the DKK1 promoter.

To determine if c-Fos, another transcription factor that can be a component of AP-1, plays a role in DKK1 expression, c-Fos was overexpressed (as were c-Jun and Tam67 as positive and negative controls, respectively) in 293T cells transfected with the DKK1 3Kb promoter reporter. We first confirmed that c-Fos protein was expressed (Supplementary Figure 2C). We then confirmed that c-Jun and c-Fos were functional, by using a pAP1-luc reporter assay in 293T cells. Both JUN and FOS constructs activated the AP-1 reporter, while Tam67 did not (Supplementary Figure 2F). Neither c-Fos nor Tam67 had an impact on β-catenin- mediated activation; whereas, c-Jun inhibited the activation (Supplementary Figure 2G). Taken together, these results confirm that the c-Jun, but not the c-Fos component of AP-1, inhibits DKK1 promoter activity.

To delineate the specific response elements on the promoter where c-Jun mediates its activity we used the promoter reporter mutants (as originally described in Figure 3A) and tested their activity using various combinations of control vectors or β-catenin expression vectors with JUN or Tam67 expression vectors. c-Jun inhibited β-catenin-induced activation of all three progressively deleted DKK1 promoter mutants (Figure 4B). However, while c- Jun inhibited β-catenin-induced activation of promoters containing point mutations at sites 1, 3, 4, and 5, it had no effect on the promoter containing point mutation at site 2 (Figure 4C). These results demonstrate that c-Jun mediates its inhibitory activity through site 2 of the DKK1 promoter. These results also correspond with our earlier finding that site 2 is critical for β-catenin mediated induction of the DKK1 promoter (Figure 3).

In order to determine if the ability of c-Jun to inhibit β-catenin-mediated activation of DKK1 was specific to the DKK1 promoter or a more general effect, we tested if c-Jun could block β-catenin-mediated activation of the TCF binding promoter that is known to be activated by β-catenin. To perform this we used the Topflash construct that contains 3 WRE element

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 8

sites and Fopflash, in which these sites are mutated so as to be nonfunctional, as a negative control. c-Jun inhibited β-catenin-mediated activation of Topflash by approximately 80%; NIH-PA Author Manuscript NIH-PA Author Manuscriptwhereas, NIH-PA Author Manuscript Tam67 had no effect (Figure 4D). Neither c-Jun nor Tam67 had an impact on Fopflash (not shown). These results indicate that c-Jun’s ability to inhibit β-catenin- mediated activation of promoters is a general effect on WREs.

As β-catenin promotes activation through forming a complex with TCF4 on the WRE, we wanted to determine if c-Jun inhibited the ability of TCF4, in addition to β-catenin, to activate the DKK1 promoter. First we tested the effect of c-Jun on the delKp-luc DKK1 reporter that contains the wildtype site 2. c-Jun blocked activation caused by either TCF4 or β-catenin alone (Figure 4E); but mutation of WRE motif site 2 eliminated the activation caused by either TCF4 and/or β-catenin, and also blocked the ability of c-Jun to inhibit the promoter (Figure 4F). These results demonstrated that c-Jun inhibition of DKK1 activation is a broad effect. This led us to explore the mechanism underlying the crosstalk of c-Jun and TCF4/β –catenin signal transduction.

PTHrP inhibits DKK1 expression through c-Jun in PCa To confirm that c-Jun’s activity on the DKK1 expression extended to PCa cell lines and through use of stable expression of proteins, as opposed to just transient expression, PC-3M cells were transduced with empty retroviral vector or retroviral vector encoding JUN, DBM3 or Tam67 cDNA. c-Jun markedly inhibited; whereas, DBM3 and Tam67 increased basal DKK1mRNA (Supplementary Figure 3A) and protein expression (Figure 5A). None of these factors altered total cellular β-catenin expression. We also noted that endogenous c-Jun was decreased in the Tam67-transduced cells. The reason for this is likely that c-Jun upregulates its own expression (22, 23) thus blocking its activity with Tam67 results in lower c-Jun expression. These results were consistent with previous reporter assays in 293T cells and indicate that c-Jun inhibits endogenous DKK1 expression in PC-3M cells.

To determine if c-Jun impacts β-catenin-induced DKK1 promoter activation in PCa cells, control vector or β-catenin were co-transfected into the stably transformed PC-3M cell lines. c-Jun inhibited; whereas, blocking endogenous c-Jun with DBM3 or Tam67 promoted both endogenous and β-catenin-induced DKK1 promoter activity (Figure 5B). To further confirm that c-Jun inhibits expression of DKK1 at the mRNA and protein level in PCa cells we transduced PC-3M cells with myc-vector, myc-tagged c-Jun, myc-tagged Tam67. Again, c- Jun expression decreased and Tam67 increased DKK1 mRNA (Supplementary Figure 3B) and protein expression, respectively (Figure 5C). To evaluate the effect of Jun on intracellular localization of β-catenin, cytoplasmic and nuclear extracts of these stably- transfected cell lines were compared. Neither Jun nor Tam67 didn’t altered the β-catenin level in the cytoplasm or nucleus (Figure 5D). We also evaluated the impact of c-Jun on DKK1 protein expression in another PCa cell line, Du145 cells, and observed results similar to those as in PC-3M (Supplementary Figure 3C and D). Taken together, these results indicate that c-Jun inhibits basal and β-catenin-induced DKK1 promoter activity, and its resulting mRNA production and protein expression, in PCa cells and is consistent with the results observed in the 293T cells.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 9

Although it was demonstrated that c-Jun inhibits DKK1 expression at the transcriptional level, it was not clear that PTHrP inhibited DKK1 expression through c-Jun. To confirm that

NIH-PA Author Manuscript NIH-PA Author ManuscriptPTHrP NIH-PA Author Manuscript downregulates DKK1 expression through c-Jun, the stably transduced cell lines PC-3M/V, PC-3M/Myc-JUN and PC-3M/Myc-Tam67 were treated with PTHrP. C-Jun moderately enhanced PTHrP-mediated inhibition of DKK1 mRNA expression; whereas, Tam67 completely blocked it (Figure 5E). To further evaluate the requirement for c-Jun to mediate PTHrP’s inhibitory effect, we first knocked down c-Jun expression in PC-3M cells, which resulted in an increase of endogenous DKK1 mRNA (Supplementary Figure 3E) and protein (Figure 5F). In the presence of control siRNA, PTHrP inhibited DKK1 mRNA expression; whereas, when c-Jun was knocked down, PTHrP had no effect on DKK1 mRNA expression (Figure 5G). Taken together, these results provide strong evidence that PTHrP inhibits β-catenin mediated activation of the DKK1 promoter through c-Jun in PCa cells. This finding provides the mechanism for the results from the earlier experiments (Figure 2C) in which we observed that PTHrP inhibits DKK1 expression in PC-3M cells through c-Jun and extends the finding regarding PTHrP and its activity on the promoter from 293T cells to PCa cells.

c-Jun interacted with β-catenin/TCF to inhibit DKK1 expression The observations that c-Jun inhibited β-catenin-mediated activation of both the DKK1 promoter and Topflash reporter indicates the possibility that c-Jun interacts directly with β- catenin. To assess for this possibility we performed co-immunoprecipitation (CO-IP) and chromatin- immunoprecipitation (ChIP) assays. Initially, endogenous co- immunoprecipitation was performed using PC-3M cells with or without PTHrP treatment. When using anti-β-catenin as primary antibody, endogenous c-Jun co-immunoprecipitated with β-catenin (Figure 6A). Conversely, when anti-c-Jun was used as the primary antibody, β-catenin co-immunoprecipitated with c-Jun. PTHrP enhanced the complex formation of c- Jun and β-catenin compared to that in vehicle control group (left panel in Figure 6A). To determine if there was a difference between c-Jun and Tam67 association with β-catenin, stable cell lines expressing either flag-tagged c-Jun or flag-tagged Tam67 were used for CO- IP. More β-catenin protein was coprecipitated with flag-c-Jun than with Flag-Tam67 while the blotting indicated amount of input protein for all samples was similar (Figure 6B). These results demonstrate that c-Jun binds β-catenin; whereas, Tam67 has less affinity for β- catenin.

To determine if c-Jun and β-catenin interact on the DKK1 promoter, ChIP assay was performed with different treatments. Primers to amplify −200bp to −100bp region in DKK1 promoter, were used to check the DNA fragments after ChIP. This fragment covered Site 2, which we determined above is a critical WRE. Additionally, a pair of primers to amplify 100bp of fragment in the ORF was used as a control. Both c-Jun and β-catenin antibodies resulted in ChIP of the 100bp DKK1 WRE-2 promoter fragment compared to the ORF control (Figure 6C and 6D). Stable c-Jun overexpression decreased that amount of DKK1 WRE-2 promoter fragment ChIPed using β-catenin antibody (Figure 6D; compare PC-3M/v vs. PC-3M/JUN). Inhibition of c-Jun with stable expression of Tam67 resulted in a marked increase of WRE-2 promoter fragment ChIPed using β-catenin antibody (Figure 6D; compare PC-3M/v vs. PC-3M/v and PC-3M/JUN). These results indicate, in combination

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 10

with the finding that c-Jun does not alter β-catenin levels, that c-Jun inhibits β-catenin binding to the transcription factor:WRE-promoter fragment complex. NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript PTHrP promotes c-Jun association with the DKK1 promoter/β-catenin complex To next determine if PTHrP promotes association of c-Jun with the DKK1 promoter/β- catenin complex, we treated cells that had expression of the various mutated c-Jun constructs with PTHrP and evaluated for the changes in co-precipitation of the DKK1 WRE-2 promoter fragment with β-catenin. PC-3M/V, PC-3M/JUN and PC-3M/Tam67 cell lines were either untreated or treated with PTHrP for 4 hours, followed by ChIP with β- catenin antibody or control IgG. PTHrP decreased β-catenin binding to the DKK1 WRE-2 promoter fragment in PC-3M/V (Figure 7A). The presence of exogenous c-Jun further decreased the PTHrP-mediated decrease of β-catenin binding to the target DNA; whereas, blocking c-Jun with Tam67 rescued the binding of β-catenin to the target DNA from PTHrP-induced repression (Figure 7A). We next wanted to determine if PTHrP inhibition of β-catenin activity extends to other promoters besides DKK1. AXIN2 is an inhibitor of the Wnt signaling pathway through its ability to promote degradation of β-catenin resulting in a negative feedback loop because β-catenin activates the AXIN2 promoter (25). Additionally, MYC is a known β-catenin target gene (26). Accordingly, the ChIP assay was repeated using the AXIN2 and MYC WREs as target DNAs. PTHrP inhibited ChIP of both AXIN2 (Figure 7B) and MYC (Supplementary Figure 4A) WREs when β-catenin antibody was used. Similar to the ChIP studies using the DKK1 DNA promoter target, c-Jun overexpression increased and Tam67 overexpression blocked the PTHrP-mediated inhibition of β-catenin binding to both the AXIN2 and MYC WREs. These results indicate that PTHrP promotes binding of c- Jun to the β-catenin/WRE complex of multiple promoters in PCa cells.

In order to fully activate a target promoter, β-catenin binds to promoter-bound transcription factor TCF4. This raises the possibility that c-Jun inhibits DKK1 promoter activity through competing with β-catenin for binding to TCF4. To test this possibility, re-ChIP assay was performed using anti-TCF4 as first ChIP antibody followed by anti-c-Jun or anti-β-catenin antibodies in PC-3M cells. Using TCF4 as the first antibody we found that re-ChIP with either anti-β-catenin or anti-Jun resulted in ChIP of the DKK1 WRE (Figure 7C). This indicated that both β-catenin and c-Jun bind the same TCF4/WRE complex. To confirm the interaction of c-Jun and the β-catenin/TCF4 complex happened on TCF binding motifs in DKK1 promoter region, ChIP assays were employed with Jun, β-catenin or TCF4 knockdown PC-3 cell lines (Knockdowns shown in Figure 3B and 5E). Knockdown of c-Jun resulted in more anti-β-catenin ChIPed DKK1 WRE-2 promoter fragments compared to that in siLuc (control), β-catenin or TCF4 knockdown groups; TCF4 knockdown decreased the amount of DKK1 WRE elements ChIPed by either β-catenin or c-Jun (Figure 7D). This indicates that c-Jun does interfere with β-catenin/TCF4 on the WRE-2 of DKK1. To confirm that PTHrP-mediated inhibition of DKK expression through the β-catenin/TCF4/WRE complex was not specific to the PC-3M cell line, we performed ChIP assays using Du145 and MDAMB231 cells following PTHrP treatment. PTHrP decreased both ChIPed DKK1 and AXIN2 compared to control-treated groups (Supplementary Figures 4B and C, respectively). These results indicate that c-Jun was dynamically binding with β-catenin on the TCF4/DNA platform. This observation, along with the previous result that both PTHrP

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 11

administration, which increases c-Jun, and overexpression of c-Jun result in decreased binding of β-catenin to the WRE-2 promoter fragment (Figure 6D and 7A) indicate that c- NIH-PA Author Manuscript NIH-PA Author ManuscriptJun competes NIH-PA Author Manuscript with β-catenin for binding to the TCF4/WRE complex in multiple cell lines.

Discussion Prostate cancer has underlying osteolytic activity (2), yet it ultimately develops an overall osteoblastic phenotype. We previously reported that DKK1 expression decreases in clinical bone metastases and that loss of DKK1 contributes to the development of osteoblastic bone lesion (6). However, the mechanism through which DKK1 expression decreases was unknown. The current study provides novel insight into the regulation of DKK1expression in prostate cancer and demonstrates novel crosstalk between PTHrP and the Wnt pathways to modulate DKK1 expression. Specifically, PTHrP was found to inhibit DKK1 expression through c-Jun-mediated inhibition of β-catenin activity on the DKK1 promoter.

Overexpression of PTHrP was first identified in squamous cell carcinoma and associated with inducing marked bone resorption and hypercalcemia of malignancy (27). Similarly, PTHrP expression has been shown to increase with prostate cancer progression in bone (12, 28, 29). However, unlike squamous cell carcinoma, increased PTHrP in prostate cancer is not associated with hypercalcemia, but rather is associated with classical osteoblastic lesions (9)). Thus, it is possible that depending on the tumor context PTHrP may induce a bone anabolic effect. Although induced overexpression of PTHrP in PCa cells induces osteolytic lesions (30), it is unclear what role of endogenous PTHrP expression plays on bone remodeling in the context of the prostate cancer. It is plausible that endogenous PTHrP expression in clinical PCa has a bone anabolic effect, similar to PTH. This possibility is consistent with the observation that PTHrP expression increases as prostate cancer progresses to bone metastases, which are osteoblastic (13, 31). Our findings suggest that upregulation of PTHrP expression results in inhibition of DKK1 expression at the bone metastatic site, which allows for Wnt-mediated osteoblastic activity.

Our observation that PTHrP down-regulates DKK1 expression in PCa cells is consistent with the action of two other bone anabolic factors, endothelin-1 and transforming growth factor-beta-1, which were shown to decrease DKK1 expression in primary murine osteoblasts and human endometrial stromal cells, respectively (32, 33). The molecular mechanism resulting in decreased DKK1 expression in these models is unknown. Our study sheds light on a novel molecular mechanism through which a bone anabolic factor transcriptionally regulates a Wnt pathway inhibitor to allow the bone anabolic activity of Wnt to be realized.

In addition to our finding that PTHrP, through c-Jun, inhibits DKK1 promoter activity, several other mechanisms have been identified that repress DKK1 expression including epigenetic silencing in gastrointestinal, colonic and ovarian cancers (34–36) and through the activity of miR-335-5p (37) in normal bone. Additionally, several oncogenic transcription factors have been shown to repress DKK1 expression including c-MYC in mammary epithelial cells (38), MYCN in neuroblastoma (39), GATA6 in pancreatic cancer (40), and EWS/FL1 in Ewing’s sarcoma (41). Our data are the first to link the PTHrP pathway and its

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 12

downstream proto-oncogene effector, c-Jun, with Wnt pathway modulation through regulation of the DKK1 promoter. NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript In contrast to our findings, c-Jun has been previously shown to increase DKK1 expression during apoptosis that occurs concomitant with vertebrate limb development (42). In that study, overexpression of c-Jun through electroporation into limb buds induced DKK1 expression based on in situ hybridization. Additionally, DKK1 expression was decreased in mouse embryonic fibroblasts derived from c-Jun knockout mice compared to wildtype mice. Although it is not clear how c-Jun can mediate DKK1 expression in this instance as opposed to inhibiting, as we observed in prostate cancer, it is possible that either tissue specificity or the context of development versus cancer accounts for this difference. Additionally, technical differences may account for the conflicting results. For example, in the previous report evaluation of DKK1 was semiquantitative and not confirmed at protein or promoter levels. Our findings are consistent with previous reports that have demonstrated c-Jun can inhibit transcription activation complexes, such as those containing MyoD (43) or Smad3 (44), and identifies a novel interaction of c-Jun as an inhibitor of the β-catenin/TCF4 transcriptional activation complex.

The majority of cancers that metastasize to bone induce osteolytic lesions; whereas, prostate cancer is unique as it induces primarily osteoblastic lesions (1). Although not clearly defined, multiple mechanisms have been demonstrated to contribute to promotion of osteoblast activity. Prostate cancer produces a variety of factors that stimulate osteoblast growth and differentiation including, but not limited to, fibroblast growth factor (45), bone morphogenetic proteins (46, 47), endothelin-1 (48) and vascular endothelial growth factor (49). In addition to these factors, we had previously demonstrated that prostate cancer produces Wnts, which contribute to prostate cancer-mediated osteoblastic activity (6). We also identified that their endogenous Wnt activity was inhibited as they also expressed DKK1. Analysis of clinical prostate cancer tissues demonstrated that DKK1 expression decreased as prostate cancer progressed towards osteoblastic metastases (7). The mechanism through which DKK1 expression decreased was unknown. The current study provides a mechanism that accounts for this clinical observation. Namely, PTHrP mediates downregulation of DKK1, which then allows for Wnts to be active and induce osteoblast growth and differentiation.

In summary, our study offers a novel paradigm underlying PTHrP and DKK1 and their role in prostate cancer bone metastasis. We identified novel crosstalk between PTHrP and the Wnt pathways in prostate cancer and the molecular mechanism that achieves this activity. Specifically, the crosstalk requires PTHrP-mediated induction of c-Jun, which then interacts with β-catenin and blocks transcription of DKK1. Defining mechanisms through which prostate cancer changes the tumor microenvironment is critical towards defining therapeutic targets for prevention and therapy of prostate cancer bone metastases. Our findings indicate a new mechanism to account for how prostate cancer induces a reactive bone microenvironment and these findings offer several signaling pathways to exploit for therapeutic benefits.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 13

Materials and methods

NIH-PA Author ManuscriptCell NIH-PA Author Manuscript lines NIH-PA Author Manuscript PC-3, PC-3M, Du145 PCa cells and MDA MB231 cell were maintained in 10% FBS, RPMI 160 medium at 37 °C in a 5% CO2/95% air atmosphere. The PC-3M cell line was derived by injecting PC-3 cells into the spleen of a nude mouse and isolating a cell line from a liver metastasis (50). The identify of all cell lines used in this study, were confirmed using short tandem repeat(STR) profiling (Supplementary Table 1)to ensure they STR profiles matched previous reports or the ATCC collection (51, 52). The 293T cells were maintained in 10% FBS, DMEM with high glucose at 37 °C in a 5% CO2/95% air atmosphere.

Stable cell lines were produced via transduction of cells using retrovirus or lentivirus followed by antibiotic resistance selection for 2–4 wks. Pooled surviving cells were checked by confirming expression of the gene of interest. PC-3/V, PC-3/Myc-JUN and PC-3/Myc- TAM67 cells were derived from transduction of PC-3 cells. Du145-3/V, Du145/Myc-JUN and Du145/Myc-TAM67 were derived from Du145 cells. PC-3M/V, PC-3M/Myc-JUN, PC-3M/Myc-TAM67, PC-3M/Flag-V, PC-3M/Flag-JUN, PC-3M/Flag-TAM67, PC-3M/ Flag-DBM3, PC-3M/β-Catenin cells were derived from transduction of PC-3M cells.

siRNA knockdown In addition to utilizing shRNA lentiviral vectors to establish the stable cell lines, double strands of short RNAs (ON-Targetplus Smartpooled dsRNA) (Dharmacon) for JUN (L-003268), β-catenin (L-003482), TCF4 (L-004594) as well as non-target luc control (D-001210-04) were used to knockdown the target genes in cell lines. Briefly, 100nM– 200nM dsRNA were used to transfect cells using Oligofectamine (Invitrogen) for 48–72hr, and followed by validation of expected change in expression of the target gene using Western blotting or for further experiments.

RNA extraction and real-time PCR PTHrP (1–34) (H-9095, Bachem) was dissolved in 4mM HCl-1%BSA, stock solution and kept at −80°C. Forskolin (F3917, Sigma) was dissolved in DMSO. Cells (106) in culture media were treated with PTHrP indicated time and/or dose in figures. After incubation with PTHrP, total RNA was extracted using Qiagen RNeasy Kit (Qiagen, Ltd, CA) followed the manufacturer’s directions. Three micrograms of total RNA was used for reverse transcription using the Superscript first strand DNA synthesis kit (Invitrogen, MD). Real time PCR was performed with Qiagen SYBR Green Master Mix Kit using a Roche Light Cycler. The results of real time PCR were analyzed using the −ΔΔCt method.

Reporter Luciferase assays Cells plated in 12-well plates and incubated overnight were transfected with plasmids using Lipofectamine 2000 (Invitrogen) reagents according to the manufacturer’s protocol. For DKK1 promoter reporter assay, 100ng DKK1 promoter luciferase reporter were used per well. For Topflash or Fopflash reporter assay, 50ng plasmids were used. To keep plasmid amount constant in each well, pCMV or pCDNA3 empty vector was added as appropriate. Luciferase and Renilla activity of total cell lysates were determined using the Dual Luc Kit

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 14

(Promega); and β-galactosidase activity of total cell lysates were determined using the β-Gal Kit (Clontech). NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript ChIP, re-ChIP Assay ChIP assays were performed using a ChIP assay system (Upstate Biotechnology) according to the manufacturer’s protocol. Cells were incubated with a dimethyl 3, 3′- dithiobispropionimidate-HCl (DTBP; Pierce) solution (5 mmol) for 30 min on ice followed by formaldehyde treatment. For each ChIP reaction, 2 × 106 cells were used. All precipitated DNA samples were quantified using real-time PCR. Data are expressed as the percentage of input DNA. For re-ChIP assays, 1 × 107 cells were used for the first-step ChIP, then quantitative PCR was used to determine the difference between the samples using primers. Anti-c-Jun (sc-44), anti-β-catenin (07-1653, Millipore, CA) and anti-TCF4 (Clone 6H5-3, Millipore, CA) were used for ChIP and re-ChIP assays.

Western blotting, IP and antibodies Cells were split and maintained in full medium 24 hours before treatment. Cells were treated with 10−8M, 10−7M, or 10−6M PTHrP for 8 hours (for dose response experiments) or 10−7M PTHrP for 0, 4, 8, 16 hours for time courses. Cells were washed with cold PBS and the cell pellet was lysed with RIPA buffer (25 mM Tris-HCl pH 8.0, 50 mM NaF, 1 % Igepal CA-630, 1% Sodium Deoxycholic Acid, 10 mM NaHPO4, 150 mM NaCl, 2 mM EDTA, 0.1% SDS, 0.2mM Na3VO4, 1mM Phenylmethylsulfonyl Fluoride (PMSF), 1X proteinase inhibitor cocktail) (Sigma). Protein concentration was measured using the BCA Kit (Pierce). Protein samples were prepared with 3X loading buffer (10% glycerol, 6%SDS, 5% 2- Mercaptoethanol, 0.1% bromophenol blue), and boiled at 100°C for 5 minutes before being loaded onto SDS-PAGE gels. Semi-dry transfer blotting was performed using the manufacturer’s recommendations. PVDF membranes were blocked with 5% nonfat milk – TBST buffer or 5% BSA-TBST buffer for 1hr. Primary antibody was added to block buffer overnight on a shaker at 4°C. HRP-conjugated secondary antibodies were added and incubated for one hour at room temperature. ECL was performed according the standard protocol with the Supersignal Detection Kit (Pierce).

Endogenous co-immunopreciptation was performed to determine the interaction of proteins. PC-3 or PC-3M and their derivative cells were plated in 150mm dishes, 2–5×107 cells were collected, and lysed in lysis buffer (25mM Hepes, pH7.9, 0.5% Chaps, 0.5% NP-40, 150 mM NaCl, 20mM Na3VO4, 1mM PMSF, 1mM PIC) on ice for 2 hours. Supernatants were collected after 14000rpm 10 minutes centrifuge, and diluted with IP buffer (25mM Hepes, pH7.9, 0.5% Chaps, 150 mM NaCl, 1mM PMSF, 1mM PIC). The diluted supernatant was precleared with 60μl 50% slurry Protein A/G agarose beads (Pirece) for 2hrs. Total 0.5 to 1 mg proteins were incubated with 2–5μg antibody for each IP reaction on rotated routera overnight at 4°C. Fifty microliter of 50% slurry Protein A/G agarose beads were added to the reaction tube and rotated for another 2hr and then washed with IP buffer twice, PBS twice, and finally 50μl 3x SDS loading buffer was added to elute for Western blotting.

The following antibodies were used for co-immunoprecipitation or Western blotting: Anti-c- Jun (sc-44), anti-β-catenin (mAb sc-7199 and pAb, sc-7200), anti- PTHrP (sc-9680), anti-

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 15

MYC (SC-789), anti-hnRNP(SC-15386) anti-FOS (sc-253, Santa Cruze Biotech); anti-Jun (9165, Cell Signaling), anti-β-catenin (610154, BD Transduction Lab), anti-Flag M2 NIH-PA Author Manuscript NIH-PA Author Manuscript(F3165), NIH-PA Author Manuscript Anti-Myc (M4439) and Anti-Flag (M2)-conjugated agarose (Sigma), anti-α- tubulin (T6074, Sigma), anti-HA mAb (HA11.1) (Covance), anti-DKK1 (goat, R&D system); Anti-rabbit IgG light chain HRP and anti-mouse IgG light chain HRP (Jackson Lab), goat anti-IgG HRP (Santa Cruz Biotech), rabbit whole anti-IgG HRP and mouse whole anti-IgG HRP (GE Health System).

Statistical analyses A one-way ANOVA analysis was used with Bonferroni’s post hoc analysis for comparison between multiple groups. A Student’s t test was used for comparison between two groups. Significance was defined as a P value of <0.05.

Supplementary Material

Refer to Web version on PubMed Central for supplementary material.

Acknowledgments

This work was supported by the NIH PO1-CA939000 (ETK and LKM), and the US Department of Defense W81XWH-10-1-0546 (SIP).

References 1. Weilbaecher KN, Guise TA, McCauley LK. Cancer to bone: a fatal attraction. Nat Rev Cancer. 2011 Jun; 11(6):411–25. Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, Non-P.H.S. Review. [PubMed: 21593787] 2. Keller ET, Brown J. Prostate cancer bone metastases promote both osteolytic and osteoblastic activity. J Cell Biochem. 2004 Mar 1; 91(4):718–29. [PubMed: 14991763] 3. Robinson DR, Zylstra CR, Williams BO. Wnt signaling and prostate cancer. Curr Drug Targets. 2008 Jul; 9(7):571–80. Research Support, U.S. Gov’t, Non-P.H.S. Review. [PubMed: 18673243] 4. Hall CL, Keller ET. The role of Wnts in bone metastases. Cancer metastasis reviews. 2006 Dec; 25(4):551–8. [PubMed: 17160558] 5. MacDonald BT, Tamai K, He X. Wnt/beta-catenin signaling: components, mechanisms, and diseases. Dev Cell. 2009 Jul; 17(1):9–26. Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t Review. [PubMed: 19619488] 6. Hall CL, Bafico A, Dai J, Aaronson SA, Keller ET. Prostate cancer cells promote osteoblastic bone metastases through Wnts. Cancer Res. 2005 Sep 1; 65(17):7554–60. [PubMed: 16140917] 7. Hall CL, Daignault SD, Shah RB, Pienta KJ, Keller ET. Dickkopf-1 expression increases early in prostate cancer development and decreases during progression from primary tumor to metastasis. Prostate. 2008 Sep 15; 68(13):1396–404. [PubMed: 18561248] 8. Lu Y, Cai Z, Xiao G, Keller ET, Mizokami A, Yao Z, et al. Monocyte chemotactic protein-1 mediates prostate cancer-induced bone resorption. Cancer Res. 2007 Apr 15; 67(8):3646–53. [PubMed: 17440076] 9. Liao J, Li X, Koh AJ, Berry JE, Thudi N, Rosol TJ, et al. Tumor expressed PTHrP facilitates prostate cancer-induced osteoblastic lesions. Int J Cancer. 2008 Nov 15; 123(10):2267–78. [PubMed: 18729185] 10. Guise TA, Yoneda T, Yates AJ, Mundy GR. The combined effect of tumor-produced parathyroid hormone-related protein and transforming growth factor-alpha enhance hypercalcemia in vivo and bone resorption in vitro. J Clin Endocrinol Metab. 1993; 77(1):40–5. [PubMed: 8325957]

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 16

11. Uy HL, Guise TA, De La Mata J, Taylor SD, Story BM, Dallas MR, et al. Effects of parathyroid hormone (PTH)-related protein and PTH on osteoclasts and osteoclast precursors in vivo. Endocrinology. 1995; 136(8):3207–12. [PubMed: 7628353] NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript 12. Dougherty KM, Blomme EA, Koh AJ, Henderson JE, Pienta KJ, Rosol TJ, et al. Parathyroid hormone-related protein as a growth regulator of prostate carcinoma. Cancer Res. 1999; 59(23): 6015–22. [PubMed: 10606251] 13. Iwamura M, di Sant’Agnese PA, Wu G, Benning CM, Cockett AT, Deftos LJ, et al. Immunohistochemical localization of parathyroid hormone-related protein in human prostate cancer. Cancer Res. 1993 Apr 15; 53(8):1724–6. [PubMed: 8467485] 14. Kremer R, Li J, Camirand A, Karaplis AC. Parathyroid Hormone Related Protein (PTHrP) in Tumor Progression. Adv Exp Med Biol. 2011; 720:145–60. [PubMed: 21901625] 15. Asadi F, Farraj M, Sharifi R, Malakouti S, Antar S, Kukreja S. Enhanced expression of parathyroid hormone-related protein in prostate cancer as compared with benign prostatic hyperplasia. Hum Pathol. 1996 Dec; 27(12):1319–23. Comparative Study Research Support, U.S. Gov’t, Non-P.H.S. [PubMed: 8958305] 16. Qin L, Qiu P, Wang L, Li X, Swarthout JT, Soteropoulos P, et al. Gene expression profiles and transcription factors involved in parathyroid hormone signaling in osteoblasts revealed by microarray and bioinformatics. J Biol Chem. 2003 May 30; 278(22):19723–31. Research Support, U.S. Gov’t, P.H.S. [PubMed: 12644456] 17. Azarani A, Goltzman D, Orlowski J. Structurally diverse N-terminal peptides of parathyroid hormone (PTH) and PTH-related peptide (PTHRP) inhibit the Na+/H+ exchanger NHE3 isoform by binding to the PTH/PTHRP receptor type I and activating distinct signaling pathways. J Biol Chem. 1996 Jun 21; 271(25):14931–6. Research Support, Non-U.S. Gov’t. [PubMed: 8663042] 18. Koh AJ, Beecher CA, Rosol TJ, McCauley LK. 3′,5′-Cyclic adenosine monophosphate activation in osteoblastic cells: effects on parathyroid hormone-1 receptors and osteoblastic differentiation in vitro. Endocrinology. 1999 Jul; 140(7):3154–62. Research Support, U.S. Gov’t, P.H.S. [PubMed: 10385409] 19. Lupp A, Klenk C, Rocken C, Evert M, Mawrin C, Schulz S. Immunohistochemical identification of the PTHR1 parathyroid hormone receptor in normal and neoplastic human tissues. Eur J Endocrinol. 2010 May; 162(5):979–86. [PubMed: 20156969] 20. Tovar Sepulveda VA, Falzon M. Parathyroid hormone-related protein enhances PC-3 prostate cancer cell growth via both autocrine/paracrine and intracrine pathways. Regul Pept. 2002 May 15; 105(2):109–20. [PubMed: 11891011] 21. Muller M, Gagiannis S, Nawroth PP, Brune M, Schilling T. Activation of the receptor for parathyroid hormone and parathyroid hormone related protein induces apoptosis via the extrinsic and intrinsic signaling pathway. Int J Mol Med. 2009 Sep; 24(3):373–80. [PubMed: 19639230] 22. Brown PH, Chen TK, Birrer MJ. Mechanism of action of a dominant-negative mutant of c-Jun. Oncogene. 1994 Mar; 9(3):791–9. [PubMed: 8108121] 23. Thompson EJ, Gupta A, Stratton MS, Bowden GT. Mechanism of action of a dominant negative c- jun mutant in inhibiting activator protein-1 activation. Mol Carcinog. 2002 Dec; 35(4):157–62. Research Support, U.S. Gov’t, P.H.S. [PubMed: 12489106] 24. Akiyama H, Lyons JP, Mori-Akiyama Y, Yang X, Zhang R, Zhang Z, et al. Interactions between Sox9 and beta-catenin control chondrocyte differentiation. Genes Dev. 2004 May 1; 18(9):1072– 87. Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S. [PubMed: 15132997] 25. Li J, Wang CY. TBL1-TBLR1 and beta-catenin recruit each other to Wnt target-gene promoter for transcription activation and oncogenesis. Nat Cell Biol. 2008 Feb; 10(2):160–9. Research Support, N.I.H., Extramural. [PubMed: 18193033] 26. Muncan V, Sansom OJ, Tertoolen L, Phesse TJ, Begthel H, Sancho E, et al. Rapid loss of intestinal crypts upon conditional deletion of the Wnt/Tcf-4 target gene c-Myc. Mol Cell Biol. 2006 Nov; 26(22):8418–26. Research Support, Non-U.S. Gov’t. [PubMed: 16954380] 27. McCauley LK, Martin TJ. Twenty-five years of PTHrP progress: from cancer hormone to multifunctional cytokine. J Bone Miner Res. 2012 Jun; 27(6):1231–9. [PubMed: 22549910]

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 17

28. Deftos LJ, Barken I, Burton DW, Hoffman RM, Geller J. Direct evidence that PTHrP expression promotes prostate cancer progression in bone. Biochem Biophys Res Commun. 2005 Feb 11; 327(2):468–72. [PubMed: 15629138] NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript 29. Chirgwin JM, Mohammad KS, Guise TA. Tumor-bone cellular interactions in skeletal metastases. J Musculoskelet Neuronal Interact. 2004 Sep; 4(3):308–18. Research Support, Non-U.S. Gov’tResearch Support, U.S. Gov’t, Non-P.H.S. Research Support, U.S. Gov’t, P.H.S. Review. [PubMed: 15615499] 30. Rabbani SA, Gladu J, Harakidas P, Jamison B, Goltzman D. Over-production of parathyroid hormone-related peptide results in increased osteolytic skeletal metastasis by prostate cancer cells in vivo. Int J Cancer. 1999 Jan 18; 80(2):257–64. [PubMed: 9935208] 31. Perez-Martinez FC, Alonso V, Sarasa JL, Nam-Cha SG, Vela-Navarrete R, Manzarbeitia F, et al. Immunohistochemical analysis of low-grade and high-grade prostate carcinoma: relative changes of parathyroid hormone-related protein and its parathyroid hormone 1 receptor, osteoprotegerin and receptor activator of nuclear factor-kB ligand. J Clin Pathol. 2007 Mar; 60(3):290–4. [PubMed: 16775117] 32. Clines GA, Mohammad KS, Bao Y, Stephens OW, Suva LJ, Shaughnessy JD Jr, et al. Dickkopf homolog 1 mediates endothelin-1-stimulated new bone formation. Mol Endocrinol. 2007 Feb; 21(2):486–98. [PubMed: 17068196] 33. Kane N, Jones M, Brosens JJ, Saunders PT, Kelly RW, Critchley HO. Transforming growth factor- beta1 attenuates expression of both the progesterone receptor and Dickkopf in differentiated human endometrial stromal cells. Mol Endocrinol. 2008 Mar; 22(3):716–28. [PubMed: 18032694] 34. Aguilera O, Fraga MF, Ballestar E, Paz MF, Herranz M, Espada J, et al. Epigenetic inactivation of the Wnt antagonist DICKKOPF-1 (DKK-1) gene in human colorectal cancer. Oncogene. 2006 Jul 6; 25(29):4116–21. Research Support, Non-U.S. Gov’t. [PubMed: 16491118] 35. Lee J, Yoon YS, Chung JH. Epigenetic silencing of the WNT antagonist DICKKOPF-1 in cervical cancer cell lines. Gynecol Oncol. 2008 May; 109(2):270–4. [PubMed: 18377964] 36. Maehata T, Taniguchi H, Yamamoto H, Nosho K, Adachi Y, Miyamoto N, et al. Transcriptional silencing of Dickkopf gene family by CpG island hypermethylation in human gastrointestinal cancer. World J Gastroenterol. 2008 May 7; 14(17):2702–14. [PubMed: 18461655] 37. Zhang J, Tu Q, Bonewald LF, He X, Stein G, Lian J, et al. Effects of miR-335-5p in modulating osteogenic differentiation by specifically downregulating Wnt antagonist DKK1. J Bone Miner Res. 2011 Aug; 26(8):1953–63. [PubMed: 21351149] 38. Cowling VH, D’Cruz CM, Chodosh LA, Cole MD. c-Myc transforms human mammary epithelial cells through repression of the Wnt inhibitors DKK1 and SFRP1. Mol Cell Biol. 2007 Jul; 27(14): 5135–46. [PubMed: 17485441] 39. Koppen A, Ait-Aissa R, Hopman S, Koster J, Haneveld F, Versteeg R, et al. Dickkopf-1 is downregulated by MYCN and inhibits neuroblastoma cell proliferation. Cancer Lett. 2007 Oct 28; 256(2):218–28. [PubMed: 17643814] 40. Zhong Y, Wang Z, Fu B, Pan F, Yachida S, Dhara M, et al. GATA6 Activates Wnt Signaling in Pancreatic Cancer by Negatively Regulating the Wnt Antagonist Dickkopf-1. PLoS One. 2011; 6(7):e22129. [PubMed: 21811562] 41. Navarro D, Agra N, Pestana A, Alonso J, Gonzalez-Sancho JM. The EWS/FLI1 oncogenic protein inhibits expression of the Wnt inhibitor DICKKOPF-1 gene and antagonizes beta-catenin/TCF- mediated transcription. Carcinogenesis. 2009 Mar; 31(3):394–401. [PubMed: 20019092] 42. Grotewold L, Ruther U. The Wnt antagonist Dickkopf-1 is regulated by Bmp signaling and c-Jun and modulates programmed cell death. Embo J. 2002 Mar 1; 21(5):966–75. [PubMed: 11867524] 43. Henderson E, Stein R. c-jun inhibits transcriptional activation by the insulin enhancer, and the insulin control element is the target of control. Mol Cell Biol. 1994 Jan; 14(1):655–62. Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S. [PubMed: 8264634] 44. Dennler S, Prunier C, Ferrand N, Gauthier JM, Atfi A. c-Jun inhibits transforming growth factor beta-mediated transcription by repressing Smad3 transcriptional activity. J Biol Chem. 2000 Sep 15; 275(37):28858–65. Research Support, Non-U.S. Gov’t. [PubMed: 10871633]

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 18

45. Li ZG, Mathew P, Yang J, Starbuck MW, Zurita AJ, Liu J, et al. Androgen receptor-negative human prostate cancer cells induce osteogenesis in mice through FGF9-mediated mechanisms. J Clin Invest. 2008 Aug; 118(8):2697–710. [PubMed: 18618013] NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript 46. Dai J, Keller J, Zhang J, Lu Y, Yao Z, Keller ET. Bone morphogenetic protein-6 promotes osteoblastic prostate cancer bone metastases through a dual mechanism. Cancer Res. 2005 Sep 15; 65(18):8274–85. [PubMed: 16166304] 47. Lee YC, Cheng CJ, Bilen MA, Lu JF, Satcher RL, Yu-Lee LY, et al. BMP4 Promotes Prostate Tumor Growth in Bone through Osteogenesis. Cancer Research. 2011 Aug 1; 71(15):5194–203. [PubMed: 21670081] 48. Nelson JB, Nabulsi AA, Vogelzang NJ, Breul J, Zonnenberg BA, Daliani DD, et al. Suppression of prostate cancer induced bone remodeling by the endothelin receptor A antagonist atrasentan. J Urol. 2003 Mar; 169(3):1143–9. [PubMed: 12576870] 49. Kitagawa Y, Dai J, Zhang J, Keller JM, Nor J, Yao Z, et al. Vascular endothelial growth factor contributes to prostate cancer-mediated osteoblastic activity. Cancer Res. 2005 Dec 1; 65(23): 10921–9. [PubMed: 16322239] 50. Pettaway CA, Pathak S, Greene G, Ramirez E, Wilson MR, Killion JJ, et al. Selection of highly metastatic variants of different human prostatic carcinomas using orthotopic implantation in nude mice. Clin Cancer Res. 1996; 2(9):1627–36. [PubMed: 9816342] 51. Azari S, Ahmadi N, Tehrani MJ, Shokri F. Profiling and authentication of human cell lines using short tandem repeat (STR) loci: Report from the National Cell Bank of Iran. Biologicals. 2007 Jun; 35(3):195–202. [PubMed: 17254797] 52. van Bokhoven A, Varella-Garcia M, Korch C, Johannes WU, Smith EE, Miller HL, et al. Molecular characterization of human prostate carcinoma cell lines. Prostate. 2003 Nov 1; 57(3): 205–25. [PubMed: 14518029]

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 19 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1. PTHrP decreases DKK1 and increases c-Jun expression in PCa cells PC-3 and PC-3M cells were treated with 10−7M PTHrP for 0, 2, 4, 8 and 16 hours in normal medium with 10%FBS. (A)Real-time PCR results calculated by −ΔΔCt. The experiments were repeated three times (mean±SEM). *: P<0.05 vs. 0 hr control group. (B) Whole cell lysate was loaded and separated on 10%SDS-PAGE gels. The blot membrane was probed using anti-DKK1 and anti-α-Tubulin antibodies. (C) PC-3M cells were treated with 0 (vehicle control), 10−8M, 10−7M and 10−6M PTHrP for 8 hours. Total cell lysate was separated using SDS-PAGE, and membranes were probed with antibodies targeting the indicated proteins. (D) PC-3M cells were treated with 10−7M PTHrP for the indicated times followed by cytoplasmic extract (CE) and nuclear extract (NE) isolation. The CE and NE

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 20

were loaded and separated on 10%SDS-PAGE gels. The membrane was probed using the indicated antibodies. (E) PC-3cells were treated with 0(vehicle control), 10−8M, 10−7M and

NIH-PA Author Manuscript NIH-PA Author Manuscript10−6 NIH-PA Author Manuscript M PTHrP for 8 hours. Total cell lyste was separated by SDS-PAGE, and membranes were probed with antibodies as labeled. (F) PC-3 cells were treated with 10−7M PTHrP for the indicated times followed by cytoplasmic extract (CE) and nuclear extract (NE) isolation. The CE and NE were loaded and separated on 10%SDS-PAGE gels. The membrane was probed using the indicated antibodies. (G) PC-3 cells were treated with 10−7M PTHrP for 0, 4, 8, 16 hours; followed by Western blotting. Blotting membranes were probed with antibodies as labeled. (H) PC-3 cells were treated with 10−5M forskolin, followed by qPCR. The experiments were repeated three times (mean±SEM), *: P<0.05 vs. 0 hr control group.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 21 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 2. PTHrP knockdown increases DKK1, not β-catenin or TCF4 in PCa cells (A) Stable PC-3M cells were transduced with lentiviral shRNA PTHrP Sh1 and PTHrP Sh2; membranes were probed with anti-PTHrP, c-Jun, DKK1 and α-tubulin. (B) CE and NE were isolated from the stably-transduced PC-3M cells as indicated in Figure 2A. 20μg protein of CE and NE were separated on SDS-PAGE gel, followed by blotting and probing with anti- β-catenin, anti-α-tubulin and anti-hnRNP. (C)To rescue the c-Jun, PC-3M/PTHrPSh2 cells infected with retroviral Myc-JUN or Myc empty vector as well as a scramble control cell line. Membranes were probed with antibodies as labeled. (D) 293T cells (8×104) were transfected with 100ng pGL3 -DKK1p-Luc and 5ng control pRL-Renilla reporters, then treated with 10−7M PTHrP, 10−5M forskolin for 24hr. The experiments were repeated three times (mean±SEM). *: P<0.05 vs. control group. (E) 100ng pGL3-DKK1p-luc, 5ng pRL- Renilla reporters and 50ng pCDNA3 empty vector or pCDNA3-β-catenin were co-infected to 293T cell for 36 hours. The experiments were repeated three times (mean±SEM). *: P<0.05 vs. pCDNA3 group. (F) Stably-transfected cell lines PC-3M/v and PC-3M/β–

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 22

Catenin were treated with or with 10−7M PTHrP for 16 hours, and then protein expression was examined using Western blotting. Blot membranes were probed with antibodies to the NIH-PA Author Manuscript NIH-PA Author Manuscriptindicated NIH-PA Author Manuscript targets. (G) Stably-transfected cell lines PC-3M/v and PC-3M/β-Catenin were transfected with 100ng pGL3-DKK1p-luc, followed by treatment with or without 10−7M PTHrP for 24hr. The experiments were repeated three times (mean±SEM). *: P<0.05.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 23 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 3. DKK1 promoter is activated by β-catenin/TCF4, but inhibited by PTHrP (A) A progressive deletion series of DKK1 promoter(3Kb) reporters were achieved by with restriction enzymes (Kpn I)-Kpn I – Bsu36 I-Age I-(XhoI) sites. Half arrows indicated potential TCF sites in the DKK1 promoter. Arrows indicate positions of ChIP1F and ChIP1R primers. Mutation series of five potential WRE sites located in 1.8kb of DKK1 promoter region (labeled as delKp) were constructed. Mut1&2 was completed based on Mut1 and Mut2. Luciferase reporter (100ng), renilla (5ng) control plasmids combined with pcDNA3 or pcDNA3-β-Catenin were co-transfected into 293Tcells. Luciferase activity was measured with Dual Luc Kit. The experiments were repeated three times (mean±SEM). *: P<0.05 vs. pCDNA3 group. (B) dsRNA for TCF4 (or control siLuc) was used in PC-3M cells to knock down TCF4 followed by Western blotting. Antibodies were used as labeled. (C) 1.8kb DKK1 promoter wild type delKp-luc and site2 mutant luciferase reporters were combined with TCF4 or dominant TCF(Δ N) and pcDNA3 or pcDNA3-β-Catenin co- transfected intoPC-3M cell for 36 hours, followed by luciferase measurement. The experiments were repeated three times (mean±SEM). *: P<0.05 vs. pCDNA3 group. **: P<0.01 vs other groups. (D) delKp-luc and the series of mutant luciferase reporters were transfected into PC-3M/v or PC-3M/β-catenin stable cell lines for 24 hours, then treated with 10−7M PTHrP or vehicle control for another 24 hours. The experiments were repeated three times (mean±SEM). *P<0.05 vs. other groups. **P<0.05 vs. PC-3M/v with or without PTHrP.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 24 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 4. c-Jun inhibits the activation of β-catenin on DKK1 expression through WRE-2 site (A) The 3Kb promoter pGL3-DKK1p-luc reporter, pCMV- JUN or its variants, renilla control plasmids were combined with pcDNA3 or pcDNA3-β-catenin and co-transfected into 293T cells. Experiments were repeated three times (mean±SEM).*: p<0.05 compared with vector, Tam67, DBM3 and LZM in pCDNA3 group; **p<0.01 compared with other groups with β-catenin respectively. (B) Deletion series of DKK1 promoter delKp (−1.8Kb), Bap (−1.0Kb) and KAp (−0.5Kb) luciferase reporters, pCMV- JUN or pCMV-Tam67 or pCMV vectors, renilla control plasmids were combined with pcDNA3 or pcDNA3-β-catenin and were co-transfected into 293T cells. The experiments were repeated three times(mean ±SEM). *: p<0.05 vs. β-catenin with or without Tam67 groups. (C) Mutant reporters of

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 25

DKK1 promoter delKp 100ng pCMV- JUN or pCMV-Tam67 or pCMV vectors, renilla control plasmids were combined with pcDNA3 or pcDNA3-β-catenin and were co- NIH-PA Author Manuscript NIH-PA Author Manuscripttransfected NIH-PA Author Manuscript into 293T cells. The experiments were repeated three times(mean±SEM). *: p<0.05 vs. β-catenin with or without Tam67 groups. (D) pGL3-Topflash reporter, pCMV- JUN or pCMV-Tam67 or pCMV empty vector, renilla control plasmids were combined with pcDNA3 or pcDNA3-β-catenin and were co-transfected into 293T cells. The experiments were repeated three times (mean±SEM).*: p<0.05 vs. vector or Tam67 combined with β- catenin group. (E) To confirm c-JUN interacted with TCF4/b-catenin complex, wild type delKp-luc, pCMV- JUN, pCDNA3-TCF4 or dn-TCF4 plasmids were combined with pcDNA3 or pcDNA3-β-catenin and were co-transfected into 293T cells. The experiments were repeated three times (mean±SEM). *: p<0.05. (F) WRE site2 mutation reporter Mut2p- luc, pCMV- JUN, pCDNA3-TCF4 or dn-TCF4 plasmids were combined with pcDNA3 or pcDNA3-β-catenin and were co-transfected into 293T cells. The experiments were repeated three times.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 26 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 5. PTHrP mediates inhibition of DKK1 expression through c-Jun (A) The cell lysates of stable cell lines with JUN, Tam67 and DBM3 overexpression, were separated using SDS-PAGE gels, and blotting membranes were probed with anti-β-catenin (sc-7199), anti-DKK1(R&D), and tubulin as loading control. (B) pGL3-DKK1p-luc, renilla control plasmids combined with pcDNA3 or pcDNA3- β –catenin plasmids were co- transfected into PC-3M/v, PC-3M/fc-JUN, PC-3M/f-Tam67 cells. The experiments were repeated three times (mean±SEM). *: p<0.05 vs. vector or JUN group with pCDNA3. (C) The expression levels of DKK1 and c-Jun were evaluated by Western blotting in stable PC-3M/v, PC-3M/Myc-JUN and PC-3M/myc-Tam67 cell lines. (D) CE and NE were isolated from stably-transfected PC-3M cells as indicated in Figure 5C. 20μg protein of CE and NE were separated on an SDS-PAGE gel, followed by blotting and probing with anti-β- catenin, anti-α-tubulin and anti-hnRNP. (E) Stable PC-3M/v, PC-3M/Myc-JUN and PC-3M/ Myc-Tam67 cell lines were treated with 10−7M PTHrP for 0, 2, 4, 8hr. Total RNA was extracted and subjected to real time PCR. The experiment was repeated three times (mean ±SEM).*: p<0.05 vs. 0hr group. (F) dsRNAs for Luc, JUN or β-catenin were used to knockdown target genes in PC-3M cells. After transfection for 48hr, cell lysate was harvested and probed for target proteins. (G). dsRNAs for Luc, JUN were used to knock down target genes in PC-3M cells. After transfection for 36hrs, cells were treated with

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 27

10−7M PTHrP for 0, 2, 4, 8hrs. Total RNAs were extracted and subjected to real time PCR. Experiments were repeated three times (mean±SEM). *: p<0.05 vs. 0hr group. NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 28 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 6. c-Jun interacts with the β –catenin in PCa cells (A) To evaluate if PTHrP impacts the endogenous JUN interaction with β-catenin, PC-3M was treated with or without 10−7M PTHrP for 8 hours. Then 1mg of whole cell lysate protein in 0.5% Chaps buffer was immunoprecipitated with anti-β-catenin, anti-c-JUN (SC-44) and IgG as control. 50μg protein of whole lysates was loaded as input control. Membranes were probed with anti- β-catenin and anti-c-Jun(SC-44). (B) To confirm the different binding abilities of JUN and TAM67, PC-3M/v, PC-3M/f-JUN and PC-3M/f- Tam67 stably-transfected cells were utilized. Cell lysates were immunoprecipitated with anti-Flag (M2) conjugated agarose. Cells were treated with 20mM LiCl for 18hr, 500ug protein were used for each pull-down reaction. Anti-β-catenin (E5, mAb) and anti-Flag was used for detection. Lower panel shows protein expression level in 1/20 of input for IP, tubulin as loading control. (C) ChIP was performed using anti-c-Jun or anti- β-catenin in pull down complexes followed by PCR with primers for DKK1 WRE or ORF (as control). (D)To determine the interaction of JUN and β-catenin/TCF, 106 PC-3M/Myc-v, PC-3M/ Myc-JUN and PC-3M/Myc-Tam67 cells were ChIPed with IgG, anti- β-catenin, and anti-c- Jun, followed by real time PCR to amplify the DKK1 WRE or ORF (as control). The experiment was repeated twice.

Oncogene. Author manuscript; available in PMC 2014 May 09. Zhang et al. Page 29 NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 7. c-Jun interacts with β-catenin/TCF4 complex on the DKK1 promoter (A) and (B) PC-3M/Myc-v, PC-3M/Myc-JUN and PC-3M/Myc-Tam67 cell lines were treated with 10−7 M PTHrP for 4hrs followed by ChIP chromatin complex using anti-β- catenin and PCR primers DKK1 WRE and DKK1 ORF (A), AXIN2 WRE and AXIN2 ORF set (B) for real time PCR. The experiments were repeated three times(mean±SEM),*:p<0.05 vs 0hr point in PC-3M/v group and that in all PC3M/Tam67 group, **: p<0.05 vs 0hr in PC-3M/JUN group and PC-3M/v but all in PC3M/Tam67 group. (C) re-ChIP. PC-3M cell lysate was ChIPed with anti-TCF4 first, the elutes were followed to ChIP with IgG, anti-β- catenin and anti-c-Jun, PCR primers DKK1 WRE and DKK1 ORF were used for real time PCR. The experiments were repeated twice. (D) 100nM dsRNA were transfected into PC-3M cell by Oligofectamine for 48 hours, followed by ChIP procedures and real time PCR. The experiments were repeated twice.

Oncogene. Author manuscript; available in PMC 2014 May 09.