Oncogene (2011) 30, 4990–4998 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE The Ron receptor promotes prostate tumor growth in the TRAMP mouse model

MN Thobe1, JK Gray1, D Gurusamy1, AM Paluch1, PK Wagh1, P Pathrose1, AB Lentsch2 and SE Waltz1,3

1Department of Cancer and Cell Biology, University of Cincinnati College of Medicine, Cincinnati, OH, USA; 2Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, OH, USA and 3Cincinnati Veterans Affairs Medical Center, Cincinnati, OH, USA

The Ron receptor (TK) is overexpressed in Introduction many cancers, including prostate cancer. To examine the significance of Ron in prostate cancer in vivo,we Prostate cancer remains one of the most commonly utilized a genetically engineered mouse model, referred to diagnosed cancers among men in the United States as TRAMP mice, that is predisposed to develop prostate and is a frequent cause of death from cancer in men. tumors. In this model, we show that prostate tumors from Receptor tyrosine kinases (TKs) are emerging as 30-week-old TRAMP mice have increased Ron expression potential therapeutic targets in prostate cancer. For compared with age-matched wild-type prostates. Based on example, inhibition of the the upregulation of Ron in human prostate cancers and in receptor results in decreased prostate cell proliferation this murine model of prostate tumorigenesis, we hypo- (Vicentini et al., 2003), and inhibition of the -like thesized that this receptor has a functional role in the growth factor 1 receptor (de Bono et al., 2007; Kojima development of prostate tumors. To test this hypothesis, et al., 2009) or Her2/neu receptor (Agus et al., 1999) we crossed TRAMP mice with mice that are deficient reduces tumor growth in mouse prostate cancer cell in Ron signaling (TKÀ/À). Interestingly, TKÀ/À xenograft models. These data suggest that receptor TKs TRAMP þ mice show a significant decrease in prostate have an important role in prostate tumor growth and tumor mass relative to TRAMP mice containing func- should be considered in the development of new tional Ron. Moreover, TKÀ/À TRAMP þ prostate treatments in prostate cancer. tumors exhibited decreased tumor vascularization relative One such potential receptor, the Ron receptor, is a to TK þ / þ TRAMP þ prostate tumors, which corre- member of the Met family of receptor TKs. Ron is a lated with reduced levels of the angiogenic molecules heterodimeric glycoprotein consisting of an extracellular vascular endothelial growth factor and CXCL2. Although a-chain and a transmembrane b-chain (Ronsin et al., Ron loss did not alter tumor cell proliferation, a 1993). Binding of ligand, -like significant decrease in cell survival was observed. Simi- , to Ron results in receptor dimerization and larly, murine prostate cancer cell lines containing a Ron phosphorylation on important tyrosine residues within deficiency exhibited decreased levels of active nuclear the TK domain (Gaudino et al., 1994). Trans-autopho- factor-jB, suggesting that Ron may be important in sphorylation also occurs on two tyrosine residues within regulating prostate cell survival at least partly through the carboxyl-domain, resulting in the formation of this pathway. In total, our data show for the first time that docking sites for adaptor molecules, including phos- Ron promotes prostate tumor growth, prostate tumor phoinositide 3-kinase, phospholipase C-g, Grb2 and angiogenesis and prostate cancer cell survival in vivo. Shc. Ron is also upstream of the nuclear factor-kB (NF- Oncogene (2011) 30, 4990–4998; doi:10.1038/onc.2011.205; kB) transcription factor, although the outcome of this published online 30 May 2011 regulation is dependent on cell type and context. NF-kB activity is increased in prostate cancer cells compared Keywords: Ron receptor; hepatoctye growth factor-like with non-transformed prostate cells, and has been protein; prostate cancer; MST1R; receptor tyrosine shown to be an important regulator of angiogenic kinase chemokine production (Shen and Lentsch, 2004). Based on the numerous pathways activated by Ron, this receptor has been shown to elicit a multitude of biological responses, including cellular proliferation, migration, scattering, survival and branching morpho- Correspondence: Dr SE Waltz, Department of Cancer and Cell genesis (Iwama et al., 1996; Wagh et al., 2008; Meyer Biology, Vontz Building for Molecular Studies, University of et al., 2009b). The Ron receptor TK has also been Cincinnati College of Medicine, 3125 Eden Avenue, Cincinnati, OH implicated in the initiation and progression of several 45267-0521, USA. E-mail: [email protected] human cancers, including those of the breast, colon, skin Received 4 February 2011; revised 22 April 2011; accepted 28 April 2011; and prostate (Maggiora et al., 1998; Chen et al., 2000; published online 30 May 2011 O’Toole et al., 2006; Thomas et al., 2007). In human Ron signaling in TRAMP mice MN Thobe et al 4991 prostate cancer cells, Ron is an important regulator of appreciable differences were observed in any of the lobes angiogenic chemokine production. Moreover, the reg- of the prostates of TKÀ/À mice compared with ulation of angiogenic chemokines by the Ron receptor controls. Representative prostate tissue from TK þ / þ has been shown to at least partially be dependent on and TKÀ/À mice is depicted in Figure 1a for the the NF-kB transcription factor (Thobe et al., 2010). anterior prostate. Additional pictures are provided in Furthermore, Ron-knockdown PC-3 cells orthotopi- Supplementary Figure S1. cally transplanted into prostates of nude mice resulted in decreased prostate tumor size and vascularization Ron is highly expressed in TRAMP prostate tumors compared with PC-3 control cells (Thobe et al., 2010). To determine if similar to human prostate disease Although this study suggests the importance of Ron in (O’Toole et al., 2006; Thobe et al., 2010) Ron expression promoting prostate tumor cell growth in vivo, this model is increased in murine prostate tumors, prostates taken does not explain the genetic importance of Ron in the from either 30-week-old wild-type (TK þ / þ ) or age- development and progression of prostate cancer. matched TK þ / þ TRAMP þ mice were stained by Mice harboring a deletion of the Ron TK domain, immunohistochemistry for Ron expression (Figure 1b). termed Ron TKÀ/À mice, are overtly normal, although Compared with the minimal expression of Ron observed they exhibit alterations in inflammatory responses in the TK þ / þ mouse prostate, TK þ / þ TRAMP þ following stress (Waltz et al., 2001). We have previously prostate tumors have significantly elevated levels of published in a skin model of chemically induced Ron. No Ron expression was detected in TKÀ/À carcinogenesis that loss of Ron results in decreased prostates (data not shown) or in prostates from TKÀ/À growth of benign skin papillomas, as well as the percent TRAMP þ mice (Figure 1b). In addition, we analyzed of papillomas that progress to malignancy (Chan et al., Ron mRNA levels by quantitative real-time PCR and 2005). In addition, Ron TKÀ/À mice crossed with mice similarly observed increased Ron transcript levels in that are predisposed to develop mammary tumors TK þ / þ TRAMP þ prostates compared with TK þ / þ display decreased mammary tumor initiation, growth prostates (Figure 1c). No detectable levels of Ron mRNA and vascularization (Peace et al., 2005). To elucidate the werefoundinTKÀ/À prostates or TKÀ/À TRAMP þ role of the Ron receptor in prostate tumorigenesis prostates (data not shown). Further examination of Ron in vivo, we generated TRAMP mice that are deficient in expression by western analyses shows elevated Ron protein Ron signaling (TKÀ/À TRAMP þ mice). expression in TK þ / þ TRAMP þ prostates relative to TRAMP mice have expression of the simian virus wild-type TK þ / þ prostates (Figure 1d). 40 early driven by a prostate-specific rat probasin promoter, specifically expressed in the dorsolateral and ventral lobes of the prostate (Greenberg et al., 1995). Male Ron TKÀ/À TRAMP þ mice have decreased TRAMP mice develop prostate hyperplasia as early as 10 genitourinary complex and prostate tumor size weeks of age, and as young as 18 weeks of age can have Thirty-week-old TRAMP mice crossed with mice invasive adenocarcinoma of the prostate. Interestingly, all harboring a targeted deletion of the TK domain of mice eventually develop metastasis, mainly to the lymph Ron (TKÀ/À TRAMP þ ) showed a significant reduc- nodes and lungs (Gingrich et al., 1996). tion in genitourinary (GU) complex and prostate size Utilizing cell lines derived from TRAMP tumors, it has compared with TRAMP mice expressing wild-type Ron been shown that inhibition of the receptor TK endothelial (TK þ / þ TRAMP þ ; Figures 2a and b). The average results in decreased cell migration, size of TKÀ/À TRAMP þ prostates was 0.3 g and the although there is no impact on cellular proliferation GU complex was 1.3 g at 30 weeks of age. TK þ / þ (Kassis et al., 1999). Adult TRAMP male mice treated TRAMP þ prostates weighed on average 2.1 g and the with Genistein, a phytoestrogen found in soy food known GU complex weighed 3.6 g at this same time point. For to have TK inhibitory effects, have more well-differen- comparison, prostates and GU weights from TK þ / þ tiated prostate tumors compared with control mice (Wang mice were 0.18±0.012 and 0.47±0.035 g, respectively, et al., 2007). Although these data suggest that receptor and did not differ significantly from TKÀ/À mice. TKs are important in prostate tumorigenesis in the As depicted in Figure 2c, TK þ / þ TRAMP þ pros- TRAMP mouse model, the role of the Ron receptor has tates from 30-week-old mice were composed of poorly- not been established. Therefore, in this study we sought to differentiated prostate cancer characterized by anaplas- determine the role of Ron in prostate tumor initiation and tic sheets of cells with irregular nuclei and very little growth using the TRAMP model. cytoplasm. Normal glands could be observed within the sheets of anaplastic cells along with associated hemor- rhagic areas. A spectrum of pathology ranging from Results prostatic epithelial neoplasia to well- and poorly- differentiated adenocarcinoma with lesions staining Wild-type mouse prostates are similar to Ron sporadically for cytokeratin, E-cadherin and synaptho- TKÀ/À prostates physin were observed in the TK þ / þ TRAMP þ Prostates taken from 30-week-old wild-type mice prostates at this time point. In addition, the TK þ / þ (TK þ / þ ) or age-matched TKÀ/À mice were analyzed TRAMP þ mice had large tumors with the tumors by hematoxylin and eosin staining to determine if loss of encompassing all or a large fraction of the individual Ron results in any abnormalities in the prostate. No prostate lobes. In contrast, the majority of the TKÀ/À

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4992

Figure 1 Ron expression in wild-type and TRAMP mouse prostates. (a) Prostates from 30-week-old wild-type (TK þ / þ )orTKÀ/À mice are histologically similar. Representative sections of the anterior prostates of TK þ / þ and TKÀ/À mice are shown. (b) Ron is highly expressed in 30-week-old TK þ / þ TRAMP þ prostates compared with TK þ / þ prostate tissue. No Ron expression is detected in TKÀ/À TRAMP þ prostates. Representative tissue sections were stained with a Ron-specific antibody or with an isotype control immunoglobulin G (IgG). TK þ / þ prostates have minimal Ron expression, whereas TK þ / þ TRAMP þ prostate tumors have significantly elevated Ron expression. (c) Quantitative real-time PCR analysis shows increased Ron mRNA levels in TK þ / þ TRAMP þ prostate tumors (n ¼ 5) relative to TK þ / þ prostates (n ¼ 6). Data are expressed as means±s.e. No detectable Ron mRNA expression was observed in the prostates of TKÀ/À or TKÀ/À TRAMP þ mice. (d) Top: western analysis of Ron protein levels in a TK þ / þ prostate and in a prostate tumor from a TK þ / þ TRAMP þ mouse. Bottom: Densitometry analysis of relative Ron protein levels in TK þ / þ (n ¼ 4) and TK þ / þ TRAMP þ (n ¼ 5) prostates. Data are expressed as means±s.e. *Po0.05 compared with TK þ / þ group.

TRAMP þ mice had enlarged prostates with the patho- although Ron has been shown to regulate macrophage logy, including normal, prostatic epithelial neoplasia activation in acute injury models in vivo (Waltz et al., and well-differentiated adenocarcinoma, and contained 2001; Leonis et al., 2002; Wilson et al., 2008), we a defined glandular architecture with the maintenance of examined the extent of macrophage recruitment in the an underlying stroma (Figure 2c). TK þ / þ TRAMP þ and TKÀ/À TRAMP þ prostates Lungs from these mice were also analyzed for and found no appreciable differences in the number of metastasis by histological analysis. Three out of seven F4/80 þ cells (Supplementary Figure S2). Although the (43%) TK þ / þ TRAMP þ lungs had detectable lung recruitment of F4/80 þ cells may be similar between metastases as depicted in Figure 2d. None of the five genotypes, this does not preclude differences in macro- TKÀ/À TRAMP þ lungs examined contained metas- phage activation as having a role in the reduced tumor tases and a section of a typical lung from these mice is growth observed in the TKÀ/À TRAMP þ prostates. depicted in Figure 2d. Although the presence of lung metastases is more in the TK þ / þ TRAMP þ mice compared with the TKÀ/À TRAMP þ mice, the data Murine TRAMP prostates lacking functional Ron display are not significantly different (Po0.09). The lack of decreased microvessel density significance may be related to the small sample size and To determine the extent of prostate vascularization, warrants further detailed investigation. Interestingly, prostates taken from TRAMP mice containing wild-

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4993

Figure 2 TRAMP mice deficient in functional Ron have decreased GU complex and prostate tumor size. (a) A 30-week-old TKÀ/À TRAMP þ mice (n ¼ 12) exhibit decreased GU complex and prostate tumor mass relative to TK þ / þ TRAMP þ mice (n ¼ 16). Data are expressed as means±s.e. (b) Gross examination of GU complexes from TK þ / þ TRAMP þ mice and TKÀ/À TRAMP þ mice shows decreased GU complex size in the TKÀ/À TRAMP þ mice. (c) Representative histological analysis of TK þ / þ TRAMP þ and TKÀ/À TRAMP þ prostates. The asterisk depicts the large tumor bearing area of the TK þ / þ TRAMP þ prostate, whereas the arrow indicates one of the several neoplastic regions of TKÀ/À TRAMP þ prostate. (d) Representative image of a lung section containing metastatic foci (arrows) from a TK þ / þ TRAMP þ mouse and a corresponding section from a TKÀ/À TRAMP þ lung. Lungs and prostates were stained with hematoxylin and eosin. type Ron (TK þ / þ TRAMP þ ) or prostates taken were any differences in cell proliferation or cell death. from TKÀ/À TRAMP þ mice were stained with a Interestingly, both TK þ / þ TRAMP þ and TKÀ/À CD31-specific antibody. TKÀ/À TRAMP þ prostates TRAMP þ tumors had similar amount of proliferation exhibited a decrease in the number of vessels per unit as determined by bromodeoxyuridine (BrdU) incorpora- area compared with TK þ / þ TRAMP þ prostates tion and immunohistochemistry (Figures 4a and c). (Figures 3a and b). This decrease in microvessel density Staining with a proliferating cell nuclear antigen anti- was correlated with a significant decrease in the levels of body showed similar results as those based on BrdU vascular endothelial growth factor (VEGF) (Figure 3c). incorporation (Supplementary Figure S2B). There was, Moreover, the levels of the pro-angiogenic chemokine however, a significant increase in the number of TdT- CXCL2 were also decreased in TKÀ/À TRAMP þ mediated dUTP nick end labeling (TUNEL)-positive prostates relative to TK þ / þ TRAMP þ prostates cells in the TKÀ/À TRAMP þ tumors, suggesting (Figure 3d), although these data were not statistically that a survival advantage may exist in the TK þ / þ significant. TRAMP þ prostate cancer cells (Figures 4b and c).

TRAMP prostate tumors from TKÀ/À mice have similar Ron activates NF-kB in prostate cancer rates of cellular proliferation, but decreased cellular Ron inhibition in prostate cancer cells results in decreased survival activation of NF-kB(Thobeet al., 2010), and NF-kBis Because of the significant difference in tumor size constitutively active in TRAMP mouse prostate tumors between TK þ / þ TRAMP þ and TKÀ/À TRAMP þ (Shukla et al., 2005). To elucidate a potential mechanism prostate tumors, we sought to determine if there bywhichRonmayberegulatingtheviabilityofthe

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4994

Figure 3 Prostate tumors from TKÀ/À TRAMP þ mice are less vascularized than prostates from TRAMP mice expressing wild- type Ron. (a) TRAMP mice lacking functional Ron (TKÀ/À TRAMP þ ) have decreased vascularization as determined by immunohistochemistry with a CD31-specific antibody. (b) Mean vessel density per area was determined with n ¼ 4 mice for the TK þ / þ TRAMP þ group and n ¼ 3 mice for the TKÀ/À TRAMP þ group. Data are expressed as means±s.e. *P 0.05 o Figure 4 Prostate cell proliferation and apoptosis in TK þ / þ compared with TK þ / þ TRAMP þ group. (c, d) Quantitative TRAMP þ and TKÀ/À TRAMP þ mice. (a) BrdU immunostain- real-time PCR was performed on RNA isolated from prostate ing was performed on prostates from 30-week TK þ / þ tumors from TK þ / þ TRAMP þ and TKÀ/À TRAMP þ mice. TRAMP þ and TKÀ/À TRAMP þ mice. (b) Detection of Expression of the angiogenic factors VEGF (c) and CXCL2 (d) are TUNEL-positive cells in prostate tissue of TK þ / þ TRAMP þ shown. Data are expressed as means±s.e. with n ¼ 4 prostates per and TKÀ/À TRAMP þ mice at 30 weeks of age. (c) Prostates from group. *P 0.05 compared with TK þ / þ TRAMP þ group. o 30-week-old TK þ / þ TRAMP þ and TKÀ/À TRAMP þ mice did not exhibit differences in BrdU staining, but did contain significant differences in the extent of TUNEL-positive cell staining. Data are expressed as means±s.e. Five separate areas were counted from four independent specimens per group, TRAMP prostate epithelial cells, we analyzed levels of and representative images are shown. *Po0.05 compared with NF-kB activity. Figure 5 shows that prostate epithelial TK þ / þ TRAMP þ group. Arrows depict a few of the positive cells derived from TKÀ/À TRAMP þ mice have a signi- staining cells. ficant reduction of NF-kB reporter activity compared with prostate epithelial cells derived from TK þ / þ TRAMP þ mice (Figure 5a). To substantiate decreased NF-kB transcriptional activity in the TKÀ/À TRAMP þ pros- may provide a survival advantage. To further examine if tates, nuclear extracts were generated from prostate tissue cells expressing Ron have increased viability ex vivo,we of 30-week-old TKÀ/À TRAMP þ and TK þ / þ isolated prostate tumor cells from 30-week-old TK þ / þ TRAMP þ mice and examined by western analysis. A TRAMP þ and TKÀ/À TRAMP þ mice. These epithe- representative experiment is depicted in Figure 5b. The lial cells were then analyzed for plasma membrane TK þ / þ TRAMP þ prostates consistently displayed integrity by propidium iodide staining and for apoptosis increased amounts of nuclear NF-kBrelativetoTKÀ/À by Annexin V staining utilizing flow cytometry. In TRAMP þ prostate tissue, further supporting more NF-kB examining the tumor cells in complete media or under activity in the prostates of TK þ / þ TRAMP þ mice. serum deprivation, no appreciable Annexin V staining was observed under our experimental conditions (data not shown). Similarly, although the TKÀ/À TRAMP þ Cells derived from TRAMP prostate tumors lacking tumors trended towards increased cleaved caspase-3 functional Ron have decreased viability staining, there was no significant difference compared The increased number of TUNEL-positive cells in Ron- with TK þ / þ TRAMP þ tumors (Supplementary deficient TRAMP tumors suggests that Ron signaling Figure S3). However, consistent with our TUNEL

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4995 necessary for mammary tumor growth in a polyoma virus middle T (pMT) antigen model of breast cancer. pMT mice deficient in functional Ron were shown to have a longer tumor latency, decreased mammary tumor size and decreased metastasis. However, in the pMT mouse model, a significant decrease in cellular prolifera- tion in Ron-deficient pMT tumors compared with Ron- expressing pMT tumors was observed (Peace et al., 2005). In contrast, no appreciable differences in cellular proliferation were observed between 30-week-old TKÀ/À TRAMP þ and TK þ / þ TRAMP þ prostate tumors. These differences may reflect the aggressiveness of the respective tumors, with significant tumor burden observed in the pMT-induced mammary tumor model Figure 5 Ron loss in the TRAMP prostate results in decreased occurring in o2 months, whereas the TRAMP model levels of active NF-kB. (a) NF-kB reporter assays performed with was evaluated at 8 months of age. prostate cells derived from TK þ / þ TRAMP þ and TKÀ/À Although the difference in prostate tumor size in the TRAMP þ prostates. Decreased NF-kB reporter activity was TK þ / þ TRAMP þ and TKÀ/À TRAMP þ mice did observed in TKÀ/À TRAMP þ cells compared with TK þ / þ TRAMP þ cells. Relative NF-kB activity for each sample is not correlate with differences in tumor cell proliferation, depicted. Data are expressed as the mean±s.e. (b) Nuclear extracts a significant difference in tumor cell viability was from 30-week-old TK þ / þ TRAMP þ and TKÀ/À TRAMP þ observed between groups. Marked increases in the prostates were isolated and examined by western analysis. TKÀ/À number of TUNEL-positive cells were observed in TRAMP þ prostates exhibit decreased levels of nuclear NF-kB prostates from TK / TRAMP mice compared with p65 relative to TK þ / þ TRAMP þ prostates. PARP is shown as a À À þ nuclear loading control. *Po0.05. controls, suggesting the novel finding that Ron may be acting as a pro-survival factor in this model. In addi- tion, prostate epithelial cells derived from TKÀ/À TRAMP þ prostates were less viable ex vivo than cells derived from TK þ / þ TRAMP þ prostate, supporting staining, we did observe a consistent and significant Ron as a regulator of cellular survival, although at increase in the number of dead cells, which lost plasma present the exact mechanism for these differences membrane integrity and were permeable to propidium remains unclear. Moreover, although no changes were iodide staining, in the TKÀ/À TRAMP þ group com- observed in macrophage recruitment in the TK þ / þ pared with the TK þ / þ TRAMP þ group. The average TRAMP þ and TKÀ/À TRAMP þ prostates at 30 number of dead TK þ / þ TRAMP þ cells was 6.6± weeks, this does not preclude the involvement of 1.36, whereas the average number of dead TKÀ/À differential macrophage activation in the TKÀ/À mice TRAMP þ cells was 13.7±1.32 (Supplementary Figure S4). as a mechanism that may regulate prostate tumor growth in the TRAMP model. In addition, our studies do not exclude the possibility of the extracellular portion of Ron remaining in the TKÀ/À mice and acting as a Discussion dominant negative or a sink for Ron ligand, which may also ultimately impact prostate tumor growth in the An analysis of Ron expression in prostate tissue from TKÀ/À TRAMP þ mice. either wild-type (TK þ / þ ) or age-matched TRAMP The NF-kB transcription factor has been shown to be mice (TK þ / þ TRAMP þ ) showed that the Ron constitutively active in TRAMP mouse prostate tumors, receptor is highly expressed in prostate tumors relative and correlates with altered IkBa expression and IkB to normal murine prostate tissue. These data are consis- kinase activity (Shukla et al., 2005). Furthermore, Ron tent with several reports documenting increased Ron inhibition in prostate cancer cells results in decreased expression in prostate cancer compared with normal NF-kB activity and a decrease in pro-angiogenic prostate tissue (Dhanasekaran et al., 2005; O’Toole chemokine production (Thobe et al., 2010). We ob- et al., 2006; Lapointe et al., 2007; Thobe et al., 2010). served that cells derived from TKÀ/À TRAMP þ The overexpression of Ron is an important finding as prostates have decreased levels of active NF-kB previous studies have shown that exogenous Ron compared with TK þ / þ TRAMP þ -derived cells. overexpression is sufficient to induce the formation of Further analysis of the activation of NF-kB in TRAMP both lung and mammary tumors in mice (Chen et al., prostate specimens showed decreased levels of nuclear 2002; Zinser et al., 2006), and Ron knockdown has been NF-kBinTKÀ/À TRAMP þ prostates relative to shown to inhibit activation of NF-kB in prostate cancer TK þ / þ TRAMP þ prostates, suggesting that Ron cell lines (Thobe et al., 2010). may be acting as a pro-survival factor through the Our data show for the first time that TRAMP mice regulation of this pathway. deficient in Ron receptor signaling have decreased TKÀ/À TRAMP þ tumors also show a reduction in prostate tumor growth. Similar studies from our tumor vascularization compared with TK þ / þ laboratory have documented that the Ron receptor is TRAMP þ tumors, suggesting that differences in

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4996 prostate tumor neovascularization may be important for tumor angiogenesis and prostate cancer survival, and the differences in prostate tumor growth observed in this may therefore lead to a reduction in prostate tumor size. study. Prostate tumor angiogenesis is important in the Although deletion studies such as those presented TRAMP mouse model as TRAMP mice crossed with in this report do not necessarily reflect the response of mice containing a knockout of the angiogenic chemo- tumors following targeting Ron pharmacologically, our kine receptor, CXCR2, have decreased prostate tumor findings provide the first insight into the Ron receptor growth and time to tumor formation (Shen et al., 2006). signaling system as a potential new therapeutic target in In addition, TRAMP mice treated with soluble VEGF prostate cancer. receptor-2 showed decreased prostate tumor volume correlating with decreased vasculature (Becker et al., 2002). Our previous studies on prostate cancer cell Materials and methods lines showed that all prostate cell lines tested exhibited VEGF expression, although the VEGF levels did not Mice correlate with Ron expression (Thobe et al., 2010). C57Bl/6 transgenic adenocarcinoma of the mouse prostate In addition, we also showed that a Ron knockdown in mice was purchased from The Jackson Laboratory (Bar DU145 or PC-3 prostate cancer cells did not alter Harbor, ME, USA). Mice containing a germline deletion of VEGF expression in these cells. A knockdown of Ron in the Ron TK domain (TKÀ/À) have been characterized pancreatic cell lines also did not lead to changes in previously (Waltz et al., 2001). TRAMP þ mice were crossed VEGF levels (Logan-Collins et al., 2010). However, the with TKÀ/À mice to generate TK þ /À mice with TRAMP. latter studies reported that exogenous administration of These latter mice were crossed with TK þ /À mice to obtain the Ron ligand in pancreatic cells was able to stimulate wild-type Ron TRAMP mice (TK þ / þ TRAMP þ ) and Ron- deficient TRAMP mice (TKÀ/À TRAMP þ ). Only hemizy- VEGF expression at least at one of the time points gous TRAMP male mice (here-on referred to as TRAMP þ ) examined (Logan-Collins et al., 2010). These studies were used for experimental analyses. All animals used were in suggest that ligand-induced activation of Ron may the C57Bl/6 genetic background. The use and maintenance of influence at least temporal VEGF levels, but that animals was performed under protocols approved by the decreases in the level of Ron in cancer cell lines, as in Institutional Animals and Use Committee of the University of the case of the knockdown or knockout cells, does not Cincinnati. For obtaining weight measurements for the GU appear to alter VEGF production. Interestingly, how- complex and the prostate, mice were euthanized and the GU ever, studies by Logan-Collins and co-workers show complex (consisting of prostate, seminal vesicles, urethra and increased CD31 microvessel density staining in human bladder) was first removed and weighed. If the bladder was full pancreatic xenographs when Ron was knocked down. In upon dissection, it was punctured to ensure that the weight was not including urine. The prostate was then dissected from the contrast, the studies in the accompanying report and GU complex and weighed. Measurements are displayed as published studies consistently showed a significant weight in grams. decrease in CD31 vessel staining in tumors lacking Ron expression (Peace et al., 2005; Thobe et al., 2010). In each case however, all tumors containing decreased Prostate histology and immunohistochemistry Immunohistochemistry of prostate sections from paraffin- Ron (prostate or pancreatic) exhibited increases in cell embedded 30-week-old wild-type (TK þ / þ ), TKÀ/À,TKþ / þ death as indicated by TUNEL staining. Taken together, TRAMP þ or TKÀ/À TRAMP þ prostates were processed as these data suggest that the effects of Ron signaling on described previously (Peace et al., 2005; Zinser et al., 2006). VEGF expression and vascularization are complex and For Ron staining, the Ron C-20 antibody (1:50; Santa Cruz may be influenced possibly by cell type and/or ligand Biotechnology, Santa Cruz, CA, USA) was applied to tissues levels and receptor activation. However, a common overnight at 41C. The next day, goat anti-rabbit secondary theme across all findings is the diminished growth of antibody (Vector Laboratories, Burlingame, CA, USA) was Ron-deficient tumors, which is associated with inc- added to the tissues, and following amplification with a VECTASTAIN ABC (Vector Laboratories, Burlingame, reases in TUNEL staining. Finally, although we 0 did observe a decrease in VEGF mRNA expression in CA, USA), positive cells were visualized with 3,3 -diaminod- benzidine substrate (Vector Laboratories). The sections were TKÀ/À TRAMP þ tumors compared with TK þ / þ counterstained with hematoxylin, dehydrated and mounted TRAMP þ tumors, this model represents a loss of Ron with permount (Fisher Scientific, Pittsburgh, PA, USA). CD31 in all cell types. VEGF modulation in this model might staining was performed and quantified as described previously be influenced by the loss of Ron expression in other (Thobe et al., 2010), and cleaved caspase-3 staining was cell types (such as macrophages) or by the interaction of performed according to the manufacturer’s instructions (1:100 cell types in the tumor microenvironment that contain a Catalog no. 9664; Cell Signaling Technology, Danvers, MA, loss of Ron. USA). F4/80 and proliferating cell nuclear antigen staining Our studies support the Ron receptor TK as being an was performed as described previously (Mallakin et al., important factor for prostate tumor growth in vivo. 2006; Meyer et al., 2009a). For the analysis of proliferation, Further studies are needed to elucidate the potential role 2 h before being killed, 30-week-old TK þ / þ TRAMP þ or TKÀ/À TRAMP þ mice were injected intraperitoneally of Ron in the growth of prostate cancers, and to with BrdU and sections from paraffin-embedded prostate determine if advanced prostate cancers upregulate Ron tissue were stained utilizing a BrdU staining kit (Amersham, expression to bypass androgen dependence. Our data Piscataway, NJ, USA) according to the manufacturer’s support the notion that targeting the Ron receptor in instructions. Image J software was used to determine the prostate cancer has the potential to regulate prostate number of positive cells normalized to the area of prostate

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4997 tissue. For TUNEL analyses, prostate sections from 30-week- epithelial cells, the pellet was resuspended in Dulbecco’s old mice were stained using an In situ Cell Death Detection modied Eagle’s medium/F12 containing 2 U/ml DNase for Kit, AP (Roche Applied Science, Indianapolis, IN, USA) 5 min with vigorous shaking. DNase was inactivated with according to the manufacturer’s directions, with a 1:2 dilution equal volume of Dulbecco’s modied Eagle’s medium/F12 of TdT . The number of TUNEL-positive cells was containing 5% fetal bovine serum. Cells were centrifuged for determined by counting as described previously (Peace et al., 5 min at 800 r.p.m., supernatant was removed and pellet was 2005). resuspended in 1 Â PBS plus 5% fetal bovine serum. To remove fibroblasts, cells were shaken vigorously, then pulse Lung metastasis spun for 10 s to a maximum of 1000 r.p.m.; supernatant was For the analysis of lung metastases, lungs from 30-week-old removed and spins were repeated six additional times. Isolated TK þ / þ TRAMP þ or TKÀ/À TRAMP þ mice were primary tumor cells were utilized before 10 passages and were harvested, fixed in 10% neutral-buffered formalin overnight, positive for both cytokeratin expression and androgen receptor processed and were paraffin-embedded. Sections (4 mm) were status. Purity ranged from 80 to 90%. cut at 50-mm intervals, and were examined by hematoxylin and eosin staining for metastases. Mice were counted as positive NF-kB analyses for lung metastasis if at least one metastatic focus was NF-kB reporter assays, on isolated prostate cancer epithelial observed. cells from TK þ / þ TRAMP þ and TKÀ/À TRAMP þ mice, were performed as described previously (Thobe et al., 2010). Quantitative real-time PCR Briefly, 70 000 cells were plated in triplicate in complete media. RNA was isolated from frozen prostate tissue from 30-week- The next day, cells were transfected using Lipofectamine 2000 old mice using the Trizol method (Invitrogen, Carlsbad, CA, (Invitrogen) with either an NF-kB reporter (pNF-kBluc) or an USA). The high-capacity complementary DNA kit (Applied empty vector (pTAL-luc) construct and a control plasmid Biosystems, Foster City, CA, USA) was used to convert RNA expressing Renilla (pRL-TK). At 48 h after transfection, the to cDNA as per the manufacturer’s instructions, and SYBR cells were lysed and subjected to a dual luciferase assay green incorporation (Applied Biosystems) was used to measure according to the manufacturer’s protocol (Promega Corpora- . Primer sequences used were as follows: tion, Madison, WI, USA). Results are displayed as relative VEGFa forward, GCAGAAGTCCCATGAAGTGA and rev- NF-kB reporter activity, where NF-kB luminescence in each erse, TCCAGGGCTTCATCGTTA; CXCL2 forward, AGTG sample was normalized to the extent of Renilla expression. AACTGCGCTGTCAATGC and reverse, AGGCAAACTT TTTGACCGCC. Primers for Ron were described previously Flow cytometry (Meyer et al., 2009b). Gene expression values were normalized At 48 h after plating cells, media were spun to collect any dead to b-glucuronidase (Gus) (forward, TTGAGAACTGGTAT or apoptotic cells not adhered to the dish, and adherent cells AAGACGCATCAG; reverse, TCTGGTACTCCTCACTGA were collected by trypsinization. The cell pellets were washed ACATGC). Relative gene expression values are shown. with phosphate-buffered saline and binding buffer. Annexin V and propidium iodide were added to the cells according to the Western analyses manufacturer’s instructions (ApoAlert Annexin V-FITC For total protein analyses, frozen prostate tissues were Apoptosis Kit, Clontech, Mountain View, CA, USA). The homogenized in 1 Â Laemmeli buffer and were briefly samples were analyzed using a Coulter Epics XL instrument sonicated. Nuclear and cytoplasmic extracts were isolated as (Beckman Coulter, Miami, FL, USA). described previously (Wagh et al., 2011). Protein concentra- tions were determined using the MicroBCA kit (Pierce Biotechnology, Rockford, IL, USA) according to the manu- Statistical analyses facturer’s instructions. A total of 200 mg of protein (for Ron) Data are expressed as mean±standard error (s.e.). Student’s was loaded into a 6% polyacrylamide gel as described t-test was used for pair-wise comparisons and Mann–Whitney previously (Thobe et al., 2010). Primary antibodies used were: -sum tests were applied when applicable using the Ron C-20 (1:200 Santa Cruz Biotechnology), anti-total p65 SigmaPlot 11.0 software (Systat Software Inc., San Jose, CA, (Cell Signaling, Boston, MA, USA) and actin (1:40 000 a gift USA). Differences between groups were accepted as significant from Dr James Lessard, Cincinnati Children’s Hospital when Po0.05. Medical Center, Cincinnati, OH, USA). Membranes were then incubated with a peroxidase-conjugated anti-rabbit or an anti-mouse secondary antibody (Jackson ImmunoResearch Conflict of interest Laboratories, West Grove, PA, USA), followed by detection of the antibody with the ECL Plus reagent (Amersham The authors declare no conflict of interest. Biosciences, Piscataway, NJ, USA). Protein bands were detected using autoradiography.

Prostate cell isolations Acknowledgements Prostate tumor cells were isolated from prostates of 30-week- old mice and placed into 25 ml digestion media containing We thank Sandy Schwemberger for her assistance with the Dulbecco’s modied Eagle’s medium/F12, 1 mg/ml collagenase flow cytometry experiments as well as Gina Ciovacco for her (Worthington Biochemical, Lakewood, NJ, USA), penicillin/ technical contributions. This work was supported by Public streptomycin and 2 mg/ml bovine serum albumin (Sigma, Health Services Grant CA-125379 (SEW) from the National St Louis, MO, USA) for approximately 2 h at 37 1C with Institutes of Health, and by Grant Project No. PC060821 shaking at 200 r.p.m. Cells were then centrifuged for 5 min at (MNT) from the Department of Defense Congressionally 1000 r.p.m. and supernatant removed. To further dissociate Directed Medical Research Program.

Oncogene Ron signaling in TRAMP mice MN Thobe et al 4998 References

Agus DB, Scher HI, Higgins B, Fox WD, Heller G, Fazzari M et al. mice during nickel-induced acute lung injury. Am J Respir Cell Mol (1999). Response of prostate cancer to anti-Her-2/neu antibody in Biol 34: 15–27. androgen-dependent and -independent human xenograft models. Meyer SE, Waltz SE, Goss KH. (2009a). The Ron receptor tyrosine Cancer Res 59: 4761–4764. kinase is not required for adenoma formation in Apc(Min/+) mice. Becker CM, Farnebo FA, Iordanescu I, Behonick DJ, Shih MC, Mol Carcinogen 48: 995–1004. Dunning P et al. (2002). Gene therapy of prostate cancer with the Meyer SE, Zinser GM, Stuart WD, Pathrose P, Waltz SE. (2009b). soluble vascular endothelial growth factor receptor Flk1. Cancer The Ron negatively regulates mammary Biol Ther 1: 548–553. gland branching morphogenesis. Dev Biol 333: 173–185. Chan EL, Peace BE, Collins MH, Toney-Earley K, Waltz SE. (2005). O’Toole JM, Rabenau KE, Burns K, Lu D, Mangalampalli V, Ron tyrosine kinase receptor regulates papilloma growth and Balderes P et al. (2006). Therapeutic implications of a human malignant conversion in a murine model of skin carcinogenesis. neutralizing antibody to the macrophage-stimulating protein Oncogene 24: 479–488. receptor tyrosine kinase (RON), a c-MET family member. Cancer Chen YQ, Zhou YQ, Angeloni D, Kurtz AL, Qiang XZ, Wang MH. Res 66: 9162–9170. (2000). Overexpression and activation of the RON receptor tyrosine Peace BE, Toney-Earley K, Collins MH, Waltz SE. (2005). Ron receptor kinase in a panel of human colorectal carcinoma cell lines. Exp Cell signaling augments mammary tumor formation and metastasis in Res 261: 229–238. a murine model of breast cancer. Cancer Res 65: 1285–1293. Chen YQ, Zhou YQ, Fisher JH, Wang MH. (2002). Targeted Ronsin C, Muscatelli F, Mattei MG, Breathnach R. (1993). A novel expression of the receptor tyrosine kinase RON in distal lung putative receptor protein tyrosine kinase of the met family. epithelial cells results in multiple tumor formation: oncogenic Oncogene 8: 1195–1202. potential of RON in vivo. Oncogene 21: 6382–6386. Shen H, Lentsch AB. (2004). Progressive dysregulation of transcrip- de Bono JS, Attard G, Adjei A, Pollak MN, Fong PC, Haluska P et al. tion factors NF-kappa B and STAT1 in prostate cancer cells causes (2007). Potential applications for circulating tumor cells expres- proangiogenic production of CXC chemokines. Am J Physiol Cell sing the insulin-like growth factor-I receptor. Clin Cancer Res 13: Physiol 286: C840–C847. 3611–3616. Shen H, Schuster R, Lu B, Waltz SE, Lentsch AB. (2006). Critical and Dhanasekaran SM, Dash A, Yu J, Maine IP, Laxman B, Tomlins SA opposing roles of the chemokine receptors CXCR2 and CXCR3 in et al. (2005). Molecular profiling of human prostate tissues: insights prostate tumor growth. Prostate 66: 1721–1728. into gene expression patterns of prostate development during Shukla S, Maclennan GT, Marengo SR, Resnick MI, Gupta S. (2005). puberty. FASEB J 19: 243–245. Constitutive activation of P I3 K-Akt and NF-kappaB during Gaudino G, Follenzi A, Naldini L, Collesi C, Santoro M, Gallo KA prostate cancer progression in autochthonous transgenic mouse et al. (1994). RON is a heterodimeric tyrosine kinase receptor model. Prostate 64: 224–239. activated by the HGF homologue MSP. EMBO J 13: 3524–3532. Thobe MN, Gurusamy D, Pathrose P, Waltz SE. (2010). The Ron Gingrich JR, Barrios RJ, Morton RA, Boyce BF, DeMayo FJ, receptor tyrosine kinase positively regulates angiogenic chemokine Finegold MJ et al. (1996). Metastatic prostate cancer in a transgenic production in prostate cancer cells. Oncogene 29: 214–226. mouse. Cancer Res 56: 4096–4102. Thomas RM, Toney K, Fenoglio-Preiser C, Revelo-Penafiel MP, Greenberg NM, DeMayo F, Finegold MJ, Medina D, Tilley WD, Hingorani SR, Tuveson DA et al. (2007). The RON receptor Aspinall JO et al. (1995). Prostate cancer in a transgenic mouse. tyrosine kinase mediates oncogenic phenotypes in pancreatic cancer Proc Natl Acad Sci USA 92: 3439–3443. cells and is increasingly expressed during pancreatic cancer Iwama A, Yamaguchi N, Suda T. (1996). STK/RON receptor tyrosine progression. Cancer Res 67: 6075–6082. kinase mediates both apoptotic and growth signals via the multi- Vicentini C, Festuccia C, Gravina GL, Angelucci A, Marronaro A, functional docking site conserved among the HGF receptor family. Bologna M. (2003). Prostate cancer cell proliferation is strongly EMBO J 15: 5866–5875. reduced by the epidermal growth factor receptor tyrosine kinase Kassis J, Moellinger J, Lo H, Greenberg NM, Kim HG, Wells A. inhibitor ZD1839 in vitro on human cell lines and primary cultures. (1999). A role for phospholipase C-gamma-mediated signaling in J Cancer Res Clin Oncol 129: 165–174. tumor cell invasion. Clin Cancer Res 5: 2251–2260. Wagh PK, Gray JK, Zinser GM, Vasiliauskas J, James L, Monga SP Kojima S, Inahara M, Suzuki H, Ichikawa T, Furuya Y. (2009). et al. (2011). beta-Catenin is required for Ron receptor-induced Implications of insulin-like growth factor-I for prostate cancer mammary tumorigenesis. Oncogene 30: 3694–3704. therapies. Int J Urol 16: 161–167. Wagh PK, Peace BE, Waltz SE. (2008). Met-related receptor tyro- Lapointe J, Li C, Giacomini CP, Salari K, Huang S, Wang P et al. sine kinase Ron in tumor growth and metastasis. Adv Cancer Res (2007). Genomic profiling reveals alternative genetic pathways of 100: 1–33. prostate tumorigenesis. Cancer Res 67: 8504–8510. Waltz SE, Eaton L, Toney-Earley K, Hess KA, Peace BE, Ihlendorf Leonis MA, Toney-Earley K, Degen SJ, Waltz SE. (2002). Deletion of JR et al. (2001). Ron-mediated cytoplasmic signaling is dispensable the Ron receptor tyrosine kinase domain in mice provides for viability but is required to limit inflammatory responses. J Clin protection from endotoxin-induced acute liver failure. Hepatology Invest 108: 567–576. 36: 1053–1060. Wang J, Eltoum IE, Lamartiniere CA. (2007). Genistein chemopre- Logan-Collins J, Thomas RM, Yu P, Jaquish D, Mose E, French R vention of prostate cancer in TRAMP mice. J Carcinogen 6:3. et al. (2010). Silencing of RON receptor signaling promotes Wilson CB, Ray M, Lutz M, Sharda D, Xu J, Hankey PA. (2008). apoptosis and gemcitabine sensitivity in pancreatic cancers. Cancer The RON receptor tyrosine kinase regulates IFN-gamma Res 70: 1130–1140. production and responses in innate immunity. J Immunol 181: Maggiora P, Marchio S, Stella MC, Giai M, Belfiore A, De Bortoli M 2303–2310. et al. (1998). Overexpression of the RON gene in human breast Zinser GM, Leonis MA, Toney K, Pathrose P, Thobe M, Kader SA carcinoma. Oncogene 16: 2927–2933. et al. (2006). Mammary-specific Ron receptor overexpression Mallakin A, Kutcher LW, McDowell SA, Kong S, Schuster R, induces highly metastatic mammary tumors associated with beta- Lentsch AB et al. (2006). Gene expression profiles of Mst1r-deficient catenin activation. Cancer Res 66: 11967–11974.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene