TGF-b Signaling in Prostate Cancer Mouse Models...

Tumor and Stem Cell Biology

FGFR1–WNT–TGF-b Signaling in Prostate Cancer MouseModels Recapitulates Human Reactive Stroma

Julienne L. Carstens1,3, Payam Shahi1, Susan Van Tsang2, Billie Smith3, Chad J. Creighton3, Yiqun Zhang3,Amber Seamans3, Mamatha Seethammagari1, Indira Vedula1, Jonathan M. Levitt1, Michael M. Ittmann2,3,4,David R. Rowley1,3, and David M. Spencer1,3

AbstractThe reactive stroma surrounding tumor lesions performs critical roles ranging from supporting tumor cell

proliferation to inducing tumorigenesis and metastasis. Therefore, it is critical to understand the cellularcomponents and signaling control mechanisms that underlie the etiology of reactive stroma. Previous studieshave individually implicated fibroblast growth factor receptor 1 (FGFR1) and canonicalWNT/b-catenin signalingin prostate cancer progression and the initiation and maintenance of a reactive stroma; however, both pathwaysare frequently found to be coactivated in cancer tissue. Using autochthonous transgenic mouse models forinducible FGFR1 (JOCK1) and prostate-specific and ubiquitously expressed inducible b-catenin (Pro-Cat andUbi-Cat, respectively) and bigenic crosses between these lines (Pro-Cat � JOCK1 and Ubi-Cat � JOCK1), wedescribeWNT-induced synergistic acceleration of FGFR1-driven adenocarcinoma, associated with a pronouncedfibroblastic reactive stroma activation surrounding prostatic intraepithelial neoplasia (mPIN) lesions foundboth in in situ and reconstitution assays. Both mouse and human reactive stroma exhibited increasedtransforming growth factor-b (TGF-b) signaling adjacent to pathologic lesions likely contributing to invasion.Furthermore, elevated stromal TGF-b signaling was associated with higher Gleason scores in archived humanbiopsies, mirroring murine patterns. Our findings establish the importance of the FGFR1–WNT–TGF-b signalingaxes as driving forces behind reactive stroma in aggressive prostate adenocarcinomas, deepening their relevanceas therapeutic targets. Cancer Res; 74(2); 609–20. �2013 AACR.

IntroductionCancer growth and metastasis require proliferation, cell

migration, and stromal remodeling—functions also criticalduring organogenesis andwound repair. The fibroblast growthfactors (FGF) and their receptors (FGFR) play pivotal rolesin development, wound healing, and tumorigenesis (1).FGFR1 is a receptor tyrosine kinase that signals through theRAF–mitogen-activated protein/extracellular signal–regulat-ed kinase (MEK)–mitogen-activated protein kinase (MAPK)–extracellular signal–regulated kinase-1/2 (ERK1/2) kinase cas-cade and the phosphoinositide 3-kinase (PI3K)–AKT axis, both

of which are well-described oncogenic pathways, shown topromote androgen independence in prostate cancer (2, 3).WNT signaling is vital for embryogenesis, homeostasis of adulttissues, and wound repair, as well as being associated withmalignancy (4–11). In the absence of WNT ligands, the GSK-3b–containing "destruction complex" phosphorylates b-cate-nin, targeting it for proteasomal degradation (4). CanonicalWNT signaling involves WNT ligands binding to the corecep-tors Frizzled and lipoprotein receptor-related protein (LRP)-5/6, facilitating disruption of the destruction complex, allowingb-catenin to translocate to the nucleus.

There are many examples of FGF and WNT signaling coop-erating indevelopment and tumorigenesis (12–14). Inmammarytissue, FGFR1 rapidly accelerates Wnt-1–induced carcinomas(15), whereas the FGFR1 targets, MAPK–pERK, are known tomodulate WNT signaling and colorectal tumorigenicity (16). Inaddition,WNTsignaling cooperateswithactivatedK-RAS, and isupregulated in FGFR1-driven prostate cancer (10, 17).Moreover,recent searches of two independent databases revealed a mod-ulation of either FGF and/or canonical Wnt signaling membersin 81% and 92% of completely characterized prostate adeno-carcinoma cases, respectively. Modulation of both pathwayswithin the same case occurred in 38% and 72% of the cases,respectively (18). Furthermore, tumor-associated reactive stro-ma is similar to hyperproliferative stroma found during woundrepair. Appropriately, mechanical injurymodulates the FGF andWNT, as well as TGF-b, signaling pathways (19–24). Finally, the

Authors' Affiliations: Departments of 1Pathology and Immunology and2Molecular andCellular Biology; and 3Dan LDuncan Cancer Center, BaylorCollege of Medicine, Houston; 4M.E. DeBakey, Department of VeteransAffairs Medical Center, Houston, Texas

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Current address for D.M. Spencer: Bellicum Pharmaceuticals, Houston,Texas; current address for P. Shahi, University of California, San Francisco,California; and current address for J.L. Carsten, Dept of Cancer Biology,MDACC, Houston, Texas.

Corresponding Authors: David M. Spencer, Baylor College of Medicine,One Baylor Plaza/MS245, Houston, TX 77030. E-mail: [email protected];and David R. Rowley, Phone: 713-798-6220; Fax: 713-790-1275; E-mail:[email protected]

doi: 10.1158/0008-5472.CAN-13-1093

�2013 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 609

on March 3, 2020. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst December 4, 2013; DOI: 10.1158/0008-5472.CAN-13-1093

http://cancerres.aacrjournals.org/

expression of FGF, WNT, and TGF-b signaling primes bonemetastatic niches. These intricate developmental and wound-repair interactions support the synergistic cooperation of thesesignaling pathways in carcinogenesis that includes a poorlyunderstood, complementary stromal component.

We hypothesized that FGFR1 and canonical Wnt signaling,within the same epithelial cells (intracellular), should acceleratetumor progression, given the previously described data (dis-cussed below). However, it was unclear whether (intra/inter-cellular) paracrine signals derived from associated stromaltissue would further accelerate tumor progression, producinga second potential target for therapeutic intervention. In orderto study epithelial FGFR1 signaling, we previously generated thesynthetic ligand-inducible FGFR1 transgenic mouse model,JOCK1, via use of the prostate epithelium-specific, compositeprobasin promoter, ARR2PB (25). The addition of a lipid-per-meable chemical inducer of dimerization (CID), AP20187,causes the intracellular signaling domains of FGFR1 to oligo-merize, inducing downstream signaling (26, 27). We previouslydemonstrated that prolonged induction (�60 weeks) of FGFR1results in a step-wise progression to invasive prostate cancerwith distant metastasis (17). Furthermore, we recently demon-strated that induced aggregation of the WNT coreceptor LRP5,is sufficient to induce b-catenin stabilization, similar to canon-ical WNT signaling (28). We subsequently developed two noveltransgenic mouse models wherein WNT signaling can bespecifically activated in the prostate epithelium, yielding"Pro-Cat" mice (Prostatic Activator of b-catenin; ARR2PB pro-moter) or ubiquitously, in "Ubi-Cat" mice (Ubiquitous Activatorof b-catenin; H-2Kb promoter; ref. 28). Interestingly, Pro-Catmice do not progress beyond prostatic hyperplasia; however,after a year of induction, by which time murine prostaticepithelia can undergo natural wounding (29), two out of sixUbi-Cat mice developed adenocarcinoma (28). Furthermore,WNT signaling derived from Ubi-Cat adult stroma enhancedprostate tissue reconstitutions, further supporting an addition-al role for stromal WNT signaling in tumorigenesis (28).

In this report,we delineate synergismbetweenFGFandWNTpathways across different subtissue layers in prostate carcino-genesis by crossing Pro-Cat and Ubi-Cat lines onto the JOCK1background. We observe that WNT signaling acceleratesFGFR1-driven adenocarcinoma, through both intracellularcross-talk (Pro-Cat � JOCK1) and intra/intercellular cross-talk(Ubi-Cat � JOCK1) by significantly accelerating adenocarcino-ma initiation, whereas intra/intercellular cross-talk acceleratestumor progression. In addition, we observed that WNT signal-ing drives an inflammatory, fibroblastic reactive stroma in bothendogenous tumor and reconstitutions assays, in contrast tothe primarily myofibroblastic reactive stroma seen in humanprostate cancer (Table 1). Finally, we have identified increasedstromal TGF-b signaling in bothhuman andmurine tumors as apotential mechanism for accelerated tumorigenesis.

Materials and MethodsAnimals and treatment

Mice were housed in a pathogen-free facility by followingapproved Institutional Animal Care and Use Committee

protocols. Generation and genotyping of transgenic mice havebeen previously described (27, 28). Mice were treated biweekly,starting at 6 weeks of age, by intraperitoneal (i.p.) injections ofAP20187 (Ariad Pharmaceuticals) at 2 mg/kg in drug diluent(1,2-propanediol, 22.5% PEG400, 1.25% Tween 80). The 8-week-old nu/nu mice were procured from Charles River Laborato-ries. See Table 2 for N values.

Human tissue microarraysTissue microarrays have been described previously (30).

Radical prostatectomy tissues were collected with informedconsent, as approved by the Baylor College of Medicine Insti-tutional Review Board, by the Prostate Cancer SpecializedProgram of Research Excellence (SPORE). All patients under-went surgery for clinically localized prostate cancer and hadreceived no prior treatment.

Histology and immunohistochemistryTissues were fixed in either 10% buffered formalin (JOCK1,

Ubi-Cat, and Ubi-Cat � JOCK1) or 4% paraformaldehyde (Pro-Cat andPro-Cat� JOCK1) overnight, embedded inparaffin, and5-mm sections were stained with hematoxylin and eosin (H&E)or the Accustain Trichrome Stain system (Sigma-Aldrich).

Immunohistochemical (IHC) analysis of 5-mm sections wasperformed, as described, for each antibody using standard light

Table 1. Similarities and differences betweenmouse and human prostatic stroma

Similarities Differences

* Both stain strongly forfibroblastic markersvimentin, collagen, andtenascin c

* Both reactive stroma exhibitimmune infiltration, nerveinnervation, andneovasculature

* Both have similar keysignaling molecules* FGF* WNT* EGF* TGF-b

* Both exhibit a partial loss ofTGF-bRII but an overallincrease in TGF-b signalingas a driving force behindneoplasia andtumorigenesis

* Humans have densenormal stroma, whereasmice have sparse stroma

* Human reactive stroma ispredominatelymyofibroblastic (stainingstrongly for SMA), andmurine stroma arepredominantlyfibroblastic

NOTE: List of similarities and differences between humanand murine prostatic stroma under normal and reactiveconditions. Information provided is collected from this workand the current understanding of the literature.

Carstens et al.

Cancer Res; 74(2) January 15, 2014 Cancer Research610




microscopy. SMA: no antigen retrieval; 1:400 overnight, 4�Cmouse monoclonal smooth muscle-a actin (SMA; SigmaA2547); secondary—BioCare's MM AP Polymer 30 minutes atroom temperature; Vector red substrate for 20minutes at roomtemperature. Vimentin: 20 minutes heated EDTA buffer pH 8.0antigen retrieval; 1:50 overnight, 4�C goat polyclonal vimentin(sc-7557); secondary—BioCare's goat HRP detection kit for 15minutes at room temperature; substrate—DAB, enhanced byBioCare's DAB Sparkle. Tenascin C: heated EDTA buffer, pH 8.0antigen retrieval; 1:10 overnight, 4�C rabbit monoclonal tenas-cin C (ab108930); secondary—1:500 Anti-rabbit, 1 hour at roomtemperature; Vector red substrate, 30 minutes at room tem-perature. Phospho-Smad2: heated citrate buffer, pH 6.0 (Diag-nostic BioSystems), 20 minutes antigen retrieval; 1:100 over-night, 4�C, pSmad2 (Millipore, AB3849); secondary—Rabbit onRodent (BioCareMedical), for 30minutes at room temperature;substrate—DAB, enhanced by BioCare's DAB Sparkle. Ki-67:heated citrate buffer, pH 6.0 (Diagnostic BioSystems), 20 min-utes antigen retrieval; 1:100 overnight, 4�C, anti-Ki-67 (SP6Neomarkers RM-9106-SO); secondary—Rabbit on Rodent (Bio-CareMedical), for 30minutes at room temperature; substrate—DAB, enhanced by BioCare's DAB Sparkle.

Prostate reconstitutionEmbryonic stroma: prostates from 7- to 13-week-old

untreated JOCK1, Pro-Cat� JOCK1, or Ubi-Cat� JOCK1 micewere isolated and sorted for basal progenitors (LSC), as pre-viously described (31). In total, 100,000 LSC cells were mixedwith 100,000 rat urogenital sinus mesenchyme (UGSM) in 100mLMatrigel and transplanted subcutaneously into nu/numiceand treated � CID for 8 weeks or 30 weeks (JOCK1 grafts).Adult stroma: prostates from adult untreated, wild-type

(WT), JOCK1, or Ubi-Cat mice were isolated, and sorted fortwo populations: LSC (Lin�, Sca1þ, CD49fþ) and stroma (Lin�,Sca1þ, CD49f�). In total, 100,000 cells were mixed from eachpopulation from different genotypes in 100 mL Matrigel, trans-planted subcutaneously into nu/nu mice and treated � CID(16 weeks).

Laser-capture microdissection and whole-genomearraysFrozen samples were collected in ornithine carbamyl trans-

ferase (OCT) medium and stored at �80�C. mPIN lesions andadjacent reactive stromawere captured from 5-mmsections on

the Pixcell Laser Capture machine. Total RNA was isolatedusing the CapSure Macro LCM caps (Applied BiosystemsLCM0211) and amplified using Arcturus RiboAmp PLUS(Applied Biosystems KIT0501). Amplified RNA was labeledusing the Invitrogen cDNA labeling system (L1015-06) andhybridized to the AgilentWhole Mouse Genome Oligo (4� 44)Microarray (G4122) chip for analysis.

Statistical analysesPathologic grading. Histologic analysis was performed by

a board-certified pathologist (M.M. Ittmann) in a blindedfashion. The penetrance of each histologic phenotype through-out the prostate (whole anterior, dorsal/lateral, and ventrallobes) was graded on a scale of 0 to 3 (0, none; 1,�25%; 2,�50%;3, �75%). Scores for each phenotype were averaged across allmice in a particular group, giving the mean group score. Groupphenotype scores were thenweighted summed [(2� PIN score)þ (3� adenocarcinoma score)þ (4� sarcomatoid carcinomascore) þ (5 � sarcoma score)] to derive the total pathologicscore for each group (plotted with SEM). In addition, tissueswere scored for reactive stroma penetrance (disorganized stro-ma, characterized by fibroblasts, neovasculature, and inflam-mation) on a scale of 0 to 3 (0, none; 1,�25%; 2,�50%; 3,�75%).Finally, the score for each mouse was averaged to get the meanscore for the group (reported with SEM). See Table 2 for thenumber of mice in each group. Statistical significance wasdetermined by two-way ANOVA with Bonferroni multiplecomparisons within the interaction effect. Two-tailed post-hocpower analysis for an a of 0.05 was also performed.

To determine the ratio of p-Smad2 in reactive stromaadjacent to increasing epithelia grades, 20 random high-power(�60 optic) fields of each mouse group and 387 normal, 272PIN, and 343 tumor fields of patient samples were examined.Positive cells were defined as any stromal cells (outside thebasement membrane: includes fibroblasts, muscle cells, andendothelial cells; excludes immune cells and epithelial cells)with clear nuclear staining (brown IHC over a faint bluecounterstain). The total number of p-Smad2þ stromal nucleiwere divided by the total number of stromal cells per field, andthese values were averaged to determine the mean of eachgroup with significance determined by two-way ANOVA withBonferroni multiple comparisons. The significance of specifichuman subsets was determined by a t test or Mann–WhitneyRank sum test, as needed.

Table 2. Mice per group exhibiting reported histologic phenotype at 50% tissue penetrance

JOCK1 Pro-Cat Pro-Cat � JOCK1 Ubi-Cat Ubi-Cat � JOCK1

12 weeks 5/5a 3/3 3/3a 3/3 6/7a

24 weeks 5/5a 7/7 5/6a 3/3 3/4a

40 weeks 4/9a 4/4 4/6a 3/3 3/3a

60 weeks 9/14a 6/6 4/6a 2/6 NA

NOTE: Data represent number of mice exhibiting reported phenotype (i.e., adenocarcinoma) with at least 50% (a score of 2) tissuepenetrance out of the number of mice in each group (treatment time with CID in each genotype).aCID vs. no CID post hoc power analysis greater than 0.995 (a ¼ 0.05, directional); Pro-Cat and Ubi-Cat mice were not significantlydifferent than the null.

FGFR1–WNT–TGF-b Signaling in Prostate Cancer Stroma

www.aacrjournals.org Cancer Res; 74(2) January 15, 2014 611




The percentage of Ki-67þ stromal cells was quantified byInform software analysis of 10 random �10 micrographsimaged by a Nuance camera. Percentages for each micrographwere calculated, and averaged to give a mean for each mouse.Themean score for eachmouse was averaged to give themeanscore for the group (reported value).

Prostate reconstitution assays. Groups were comparedfor statistical significance with one-tailed Student t tests.

Gene expression analysis. Expression arrayswere proces-sed and loess-normalized using BioConductor, and heat mapswere generated using JavaTreeView (32). Array data have beendeposited in the Gene Expression Omnibus (GEO; accessionGSE41344).

ResultsWNT signaling accelerates FGFR1-driven prostateadenocarcinoma

Both murine and human prostate tumorigenesis advancesthrough well-characterized pathologic "hallmarks": epithelialhyperplasia, mPIN, adenocarcinoma, and transitional sarco-matoid. Transitional sarcomatoid lesions have been recentlyreclassified as sarcomatoid carcinoma and are considered ahistologic manifestation of epithelial–mesenchymal transition(EMT; Supplementary Fig. S3; ref. 33). Our previous workdemonstrated that JOCK1 mice steadily progress throughthese hallmarks, similar to human prostate cancer, and reachan adenocarcinoma state by 40 weeks of CID treatment (Fig. 1;ref. 17). To determine whether canonical WNT signalingaccelerates JOCK1 progression, therebymodelingmore aggres-sive human cancer, we crossed Pro-Cat (inducible WNT sig-naling in prostate epithelium) and Ubi-Cat mice (ubiquitousexpression of inducible WNT signaling) with JOCK1 animals(Supplementary Figs S1 and S2). Both strains were treated withCID biweekly for 12, 24, 40, or 60 weeks. Tissues were collectedand analyzed for the presence and penetrance of progressionhallmarks. Seminal vesicle occlusion correlated with prostatetransformation (Supplementary Fig. S5); all other tissues werehistologically normal (data not shown). Pro-Cat mice did notprogress beyond minor hyperplasia. When combined withFGFR1 signaling, all Pro-Cat � JOCK1 mice developed adeno-carcinoma significantly faster thanmonogenic JOCK1mice (24weeks vs. 40weeks). However, tumors inPro-Cat� JOCK1micedid not progress until 60 weeks of treatment, similar to JOCK1,consistent with an age-dependent and self-limiting progres-sion threshold or "brake". Similar to Pro-Cat animals,most Ubi-Cat mice did not progress beyond hyperplasia; however, after 1year of CID treatment, two out of six mice developed adeno-carcinoma (28), also suggesting an age-dependent, self-limitingstep. When crossed to JOCK1, all Ubi-Cat � JOCK1 micedisplayed the same significant acceleration in initiation asPro-Cat � JOCK1 mice; however, Ubi-Cat � JOCK1 micecontinued to progress, resulting in tissues containing 50%sarcomatoid carcinoma by 40 weeks, 20 weeks earlier thanJOCK1 or Pro-Cat � JOCK1 mice. Prostate tumors phenocop-ied the JOCK1 histopathology with heterogeneous, invasivehyperchromatic epithelial cells and loss of the basal cell layer,as previously described. These data suggest that the addition of

WNT signaling changed the rate of progression: intracellularcross-talk (Pro-Cat � JOCK1) accelerated FGFR1-driven ade-nocarcinoma initiation, whereas intra- and intercellular cross-talk (Ubi-Cat � JOCK1) accelerated both progression andinitiation.

Transgenic progenitors reconstitute mPIN andadenocarcinoma

We next investigated the ability of basal progenitor cells,known as LSC (Lin�, Sca1þ, CD49fþ), to reconstitute prostatepathology seen in the endogenous tumors. LSC cells fromJOCK1, Pro-Cat � JOCK1, and Ubi-Cat � JOCK1 mice wereisolated, transplanted subcutaneously along with inductive raturogenital sinus mesenchyme in Matrigel into recipient mice,and treated � CID for 8 weeks (28). Consistent with theautochthonous tumor progression, all transgenic LSC graftsdisplayed hyperplasia within 8 weeks. Interestingly, JOCK1grafts also displayed mPIN, like the double-transgenic lines(Fig. 2A), indicating loss of a tumor-suppressing function,typically distinguishing WNT-accelerated FGFR1-drivenpathology. To test whether JOCK1 grafts also achieved accel-erated adenocarcinoma, we extended the CID treatment andobserved reconstitution of adenocarcinoma-like lesions by 30weeks, 10 weeks earlier than autochthonous JOCK1 tumors(Fig. 2B). This accelerated tumorigenesis could be attributableto the use of pro-inductive embryonic stroma, which, whencombinedwith oncogenic initiation (FGFR1), recapitulated thepronounced, tumor-associated reactive stroma observed indouble-transgenic mice. Thus, in the presence of inductivestroma, reconstituted grafts, containing FGFR1 and FGFR1/WNT signaling-activated progenitor cells, can progress tosimilar levels of pathology as in autochthonous transgenictissues.

WNT drives an inflammatory, fibroblastic reactivestroma

To further characterize the reactive stroma, we performedimmunohistochemical analysis of several reactive stromamar-kers in histologically similar tumors from JOCK1 (i.e., following40 weeks of CID treatment) and Pro-Cat� JOCK1 and Ubi-Cat� JOCK1 (24weeks of treatment). The reactive stromawas verycollagen-rich (Fig. 3A and Supplementary Fig. S4; Trichrome,blue staining)withUbi-Cat� JOCK1 being themost dense. TheSMA layer surrounding the acini in normal stroma is greatlyreduced in reactive stroma. This correlates with an overalldecrease in SMA; however, JOCK1 tissues still displayed focalareas of dense SMA staining (pictured). In addition, weobserveddense expressionof thefibroblasticmarker, vimentin,in all reactive stroma (Fig. 3A; Supplementary Fig. S4). Fur-thermore, focal expression of the reactive stroma marker,Tenascin C, closely correlated with invasive lesions, as previ-ously observed in human prostate cancer (30). Finally, overallproliferation across genotypes (Fig. 3A; Ki-67) correlated withthe rate of progression. Interestingly, the reactive stroma ofUbi-Cat � JOCK1 tumors revealed considerably more Ki-67þ

cells in the stroma. Thus, b-catenin stabilization acceleratedFGFR1-driven prostate cancer, producing a heterogeneous,proliferative reactive stroma that is predominantlyfibroblastic,

Carstens et al.





as opposed to a myofibroblastic phenotype reported in humanprostate cancer.To further understand the relationship between the reactive

stroma and cancer progression, we quantified the stromalpathologic score across each group (Fig. 3B). The developmentof reactive stroma was CID-dependent and consistently arosein conjunction with mPIN, prior to adenocarcinoma. Thisobservation suggests the presence of reactive stroma could bedriving prostate cancer development, as previously suggestedin humans (30). Finally, there was significantly more reactivestroma in the double-transgenic mice over time with the mostpronounced difference seen in the Ubi-Cat � JOCK1 animals(Fig. 3B). Interestingly, in the two Ubi-Cat mice that developedCID-dependent tumors, reactive stroma was observed only atfocal sites of adenocarcinoma and was not observed in normalormPIN tissue, suggesting canonicalWNT signaling in autoch-thonous epithelium and stroma (without oncogenic stimuli) isnot sufficient to drive a reactive stroma. Taken together, thesedata indicate thatWNTsignalingdrives stromal expansion and,when accompanied by appropriate oncogenic signaling (e.g.,FGFR1), correlates with aggressive prostate cancer.

Stromal WNT signaling induction induces a fibroblasticreactive stroma with lymphocyte infiltrationAlthough we observed cooperative effects of FGFR1 and

WNT signaling, limitations of autochthonous tumor modelsand the lack of an all-inclusive stromal promoter precluded afull assessment of whether WNT signaling in the stromaalone is sufficient to synergize with FGFR1 signaling in theepithelium. To address this, we utilized the ubiquitous iLRP5expression in Ubi-Cat mice for prostate reconstitutionassays. LSC cells and stromal cells (Lin�, Sca1þ, CD49f�)from adult WT, JOCK1, or Ubi-Cat mice were isolated. Cellpopulations from different genotypes were mixed 1:1 inMatrigel and transplanted subcutaneously into athymicmice. Ubi-Cat stroma mixed with WT LSC cells in theabsence of CID had very poor engraftment (in reference tovery small acini size and number and not "take" rate), andonly a small, single-cell layer acini (28). However, in thepresence of canonical WNT signaling (þCID), acini weremore numerous and developed, although their overall sizewas small. Interestingly, WNT-induced stroma was thickerwith spindle cell-like morphology and contained vascular(Fig. 4A) components. Likewise, in the absence of CID, JOCK1LSC and WT stroma had poor engraftment. The induction ofFGFR1 signaling in the epithelium was sufficient to improveengraftment, without aberrant pathology, as was seen whencombined with embryonic stroma (Fig. 2). Without CID, Ubi-Cat stroma and JOCK1 LSC grafts had similar poor engraft-ment rates and small acini (Fig. 4A and 4B). However, in thepresence of WNT signaling, acini contained proliferativelesions and were significantly larger than CID-treated WTLSC þ Ubi-Cat stroma (Supplementary Fig. S6). As shownin Fig. 4A, a reactive host response by histologic parametersis apparent in CID-treated JOCK1 LSC þ Ubi-Cat stromagrafts. Meanwhile, control grafts (i.e., minus CID or CID-treated WT LSC þ Ubi-Cat stroma) exhibited very littlereactive stroma (Fig. 4A). These experiments demonstrate

that WNT signaling alone in the stroma is associated with areactive stroma phenotype in synergy with FGFR1-inducedproliferation and premalignant transformation.

Aggressive adenocarcinoma correlates with modulatedreactive stroma and TGF-b signaling

To elucidate the mechanism behind the WNT-acceleratedFGFR1 adenocarcinoma, we performed laser capture microdis-section (LCM) to separate precancerous (hyperplasia andmPIN)epithelia and adjacent reactive stroma cells. We utilized frozensamples from the same timepoints as described earlier, andperformed gene expression profiling (Fig. 5A). Comparing themoreaggressivePro-Cat� JOCK1precancerous tissue to JOCK1,we discoveredmodulation of the FGF-axis receptors (i.e., FGFR1,2, 3, and 4) and ligands (i.e., FGFs 3, 7 and 8) in both theepithelium and stroma, demonstrating a feed-forward loopthat could lead to accelerated CID-independence, previouslydescribed in JOCK1-driven tumors (17, 26). Furthermore, weobservedanupregulationof variousmetastasis-associated genes(Snail2, Twist1, andR-Ras) in the epithelium of both bigenic linesversus JOCK1 (Supplementary Fig. S7). Interestingly, we alsoobserved a significant upregulation of TGF-b ligands in theepithelium of double-transgenic mice compared with JOCK1mice and a downregulation of TGF-b receptors in both theepithelium and stroma (Fig. 5B). This is consistent with therecent discovery that loss of TGF-bRII in a subset of stromal cells(but not all) promotes a more aggressive disease (34). AlthoughPro-Cat� JOCK1 and Ubi-Cat� JOCK1 tissues were compared,no notable differences were observed at this particular timepoint of "accelerated initiation"; however, later time points indisease progression might reveal gene expression changes thatcould distinguish an "accelerated progression" signature.

To explore which cells were responding to the increasedTGF-b ligands, we examined nuclear localization of phosphor-ylated-Smad2 (p-Smad2) as a read-out for TGF-b receptoractivation in JOCK1, Pro-Cat � JOCK1, and Ubi-Cat � JOCK1tissues, and in a human tissue array, consisting of normal,mPIN and tumor samples frommatched donors. Approximate-ly 90% of the normal prostate epithelium darkly stained for p-Smad2. In light of the established antiproliferative role of TGF-b, this supports a tightly controlled signaling mechanism tomaintain epithelial homeostasis (Fig. 5C). Interestingly, mPINlesions showed high expression of nuclear p-Smad2, whereashyperplasic areas showed p-Smad2 heterogeneity, suggestingthat a change in homeostatic signaling permits neoplasia.Furthermore, most of the nuclei in normal stroma faintly tomodestly stain for nuclear p-Smad2, whereas reactive stromastainedmore intensely and heterogeneously. To investigate theassociation and potential role of this heterogeneity, we ana-lyzed the percentage of nuclear p-Smad2þ stromal cells adja-cent to various types of epithelium. Overall, the expression oftumor-associated stromal p-Smad2 correlated with increaseddisease aggressiveness in the mouse models (Fig. 5D). Thissuggests that the canonical WNT-driven partial loss of TGF-breceptors and increase in ligands identified by LCM couldinitiate the reactive stroma, which collaborates with neoplasticepithelium to drive accelerated adenocarcinogenesis in dou-ble-transgenic mice. Moreover, the higher overall TGF-b






Ubi-Cat × JOCK1Pro-Cat × JOCK1JOCK1

********

****

****

**** ****

****

**

NA

** ********

****

********

****

****

**

****

**

****

** P < 0.01

**** P < 0.0001

0

5

10

15

20

25

Weig

hte

d p

ath

olo

gic

al score

A

PIN

JOCK1

Pro-Cat

Ubi-Cat

Pro-Cat × JOCK1

Ubi-Cat × JOCK1

Adenocarcincoma

Transitional sarcomatoids

12 wks 24 wks 40 wks 60 wks

??

Pro

gre

ssio

n

B

Ub

i-C

at

24 wks12 wks

WT

JO

CK

1

No CID

Pro

-Ca

t

Pro

-Ca

t

× J

OC

K1

40 wks

Ub

i-C

at

× J

OC

K1

40 wks

C

–CID

+CID

8 wks

12 w

ks

24 w

ks

40 w

ks

60 w

ks

8 wks

12 w

ks

24 w

ks

40 w

ks

60 w

ks

8 wks

12 w

ks

24 w

ks

40 w

ks

60 w

ks

Figure1. Progression of tumor phenotypic hallmarks over time in fivemousemodels. A, graphical representation of tumor progression kinetics. (See Table 2 forN value ofmice per group exhibiting the reported phenotype). B, ventral lobe, H&E staining. Scale bar, 100 mm.Column 1, control (30weeks). Transgenic lineshave minor background hyperplasia that is consistent throughout the observed time points. (Continued on the following page.)

Carstens et al.





signaling observed in Ubi-Cat � JOCK1 stroma suggests thatslight changes in TGF-b signaling in an already primed pro-neoplastic background elicits more rapid progression.When the patient data was grouped according to pathologic

scores, we discovered that Gleason 7 relative to Gleason 5 to 6tumors and stage T2/T3a relative to T3b or N1 tumors revealeda strong trend (P ¼ 0.073 and P ¼ 0.112, respectively) towardhigher TGF-b signaling in the stroma immediately adjacent tothe tumors; however, the relatively limited cohort size (n¼ 49)may have tempered statistical significance. Interestingly,tumor-associated stroma containing greater than 60% nuclearpSmad2þ cells is significantly associated with more aggressivetumors than stroma that was less than 60% positive (P ¼0.003; Fig. 5D). Finally, four patients had recorded biochemicalrecurrence with a trend in higher tumor-associated TGF-b;however, the small number of patients did not lend statisticalsignificance. Together, these data demonstrate that bothbigenic mouse lines model several key attributes of humanprostate cancer progression: (i) pathologic progression, (ii)expression of many important gene signatures, and (iii) altera-

tions in stromal TGF-b signaling. This last point is particularlyrelevant, as patients with higher stromal TGF-b signalingpresent with more aggressive disease, which is paralleled inthe more aggressive Ubi-Cat� JOCK1 compared with Pro-Cat� JOCK1 or monogenic JOCK1 tumors.

Novel WNT signaling signature in human grade 3Although the role of FGFR1 in human prostate carcino-

genesis is well established, a regulatory role of canonicalWNT signaling during therapy resistance in response totreatment-induced wound healing is only beginning to beunraveled, and WNT signaling in the pretreated stroma hasnot been investigated. To this end, we used the ontogenicanalysis tool in DAVID Bioinformatics Resources 6.7 tointerrogate the microarray data collected by Dakhova andcolleagues, who compared normal human stroma to reactivestroma grade 3 (RSG3, characterized by at least a 1:1 ratiobetween reactive stroma and cancer epithelium; ref. 35).We discovered a distinct WNT signature, consisting of theupregulation (�1.2-fold and P < 0.001) of 15 genes implicated

JOCK1 Ubi-Cat × JOCK1Pro-Cat × JOCK1

No

CID

CID

CIDNo CID

JO

CK

1

A

B

Figure 2. Reconstituted prostatesdevelop hyperplasia, mPIN, andadenocarcinoma. A, H&E analysisof basal progenitor cells (LSC)mixed with UGSM prostate grafts.Activation of CID-inducibletransgenes for 8 weeks inducedhyperplasic acini and mPIN lesion.Scale bar, 100 mm (n ¼ 4). B,activation of JOCK1 grafts for 30weeks resulted in adenocarcinoma,10 weeks earlier than inautochthonous JOCK1 tumors(n ¼ 12).

(Continued.) Column 2, 12 weeks CID-treated. Ubi-Cat animals exhibit a thickened stroma, whereas Pro-Cat animals have minor hyperplasia and slightlythickened stroma. JOCK1, Ubi-Cat � JOCK1, and Pro-Cat � JOCK1 animals have marked hyperplasia and focal areas of mPIN. Column 3, 24 weeksCID-treated. Pro-Cat animals have areas of hyperplasia whereas Ubi-Cat animals have abnormal proliferation and thickened stroma. Ubi-Cat� JOCK1 andPro-Cat � JOCK1 animals have marked hyperplasia, high-grade mPIN, and adenocarcinoma; significantly faster than JOCK1, which did not obtainthis level of progression until 40 weeks of treatment. Column 4, 40 weeks CID-treated. JOCK1 and Pro-Cat � JOCK1 have adenocarcinoma whereasUbi-Cat � JOCK1 mice have advanced adenocarcinoma with sarcomatoid carcinomas. Column 5, ventral prostate at 40 w CID. Scale bar, 500 mm. C,weighted pathologic score for each JOCK1, Pro-Cat� JOCK1, and Ubi-Cat� JOCK1 time point�CID. P values determined by two-way ANOVA. Post-hocpower analysis (a¼ 0.05, directional) for CID versus no CID and 24-week CID double transgenic vs. JOCK1 was greater than 0.995; 40-week CID Pro-Cat�JOCK1 versus JOCK1 and Ubi-Cat � JOCK1 versus JOCK1 was 0.70 and 0.99, respectively.






in a WNT signaling cascade (Supplementary Fig. S8A; refs 36,37). Although the WNT signature observed in the CAT �JOCK transgenic animals contains different misregulatedgenes (Supplementary Fig. S8B), it, nonetheless, validatesthe importance of stromal WNT signaling in progression ofhuman carcinogenesis. Furthermore, this WNT signaturereveals the modulation of several WNT ligands (Wnt7a,Wnt3, and Wnt6) in Pro-Cat � JOCK1 and a dramaticupregulation of Wnt3 in Ubi-Cat � JOCK1, indicating thatinduced FGF and canonical WNT signaling could induce theproduction of more WNT ligands, as seen in development,

that could act on the stroma needed for cancer initiation(Supplementary Fig. S6; ref. 38).

DiscussionTumor-associated epithelial cells are surrounded by sup-

porting cells and matrices, together termed "stroma". Thestroma contains vasculature, smooth muscle, nerves, fibro-blasts, and immune cells (33). Following wounding, this diver-sity further expands in response to various stimuli (e.g., TGF-b,FGF and WNT signaling) to facilitate fibroblast migration,

8 wks

12 wks

24 wks

40 wks

60 wks

WT

JO

CK

1

Pro

-Cat

Ubi-C

at

Pro

-Cat×

JO

CK

1

Ubi-C

at×

JO

CK

1

WT

JO

CK

1

Pro

-Cat

Ubi-C

at

Pro

-Cat×

JO

CK

1

Ubi-C

at×

JO

CK

1

No CID CID

***

***

***

***

NA*** P < 0.001

2.5

2.0

1.5

1.0

0.5

0

Pe

ne

tra

nce

B

AK

i-67

Tenascin

CV

imentin

SM

ATri

chro

me

JOCK1Pro-Cat

× JOCK1

Ubi-Cat

× JOCK1WT

0.5% 7.0% 6.9% 12.3%

Figure 3. FGFR1 and WNT induce afibroblastic reactive stroma. A,adenocarcinoma lesions of JOCK140 weeks CID; Pro-Cat � JOCK1and Ubi-Cat � JOCK1 24 weeksCID. Scale bar, 200 mm; n ¼ 3.Row 1, trichrome; row 2, smoothmuscle a-actin (SMA; muscle andmyofibroblast marker); row 3,vimentin (fibroblast marker); row 4,tenascin C (reactive stromamarker); row 5, Ki-67 (proliferationmarker); percentage of Ki-67þ

stromal cells in the bottom left-hand corner. The induced reactivestroma is collagen- and vimentin-rich with focal areas of tenascin Cstaining accompanied by a loss ofsmooth muscle—hallmarks of afibroblastic reactive stroma. B,inset, H&E representativemicrographs of reactive stroma.Scale bar, 200 mm. Inset bar,10 mm. Each lobe was scored forthe presence and penetrance ofreactive stroma on a scale of 0 to 3(0, none; 1, �25%; 2, �50%; 3,�75%); mean score for each groupis plotted � SEM. Reactive stromais a CID-dependent stromaldisorganization, characterized byfibroblasts, neovasculature, andinflammation. Penetrance ofreactive stroma significantlyincreases over time in both bigenicmice, with the Ubi-Cat � JOCK1increase being more pronounced.Significance determined by two-way ANOVA. (See Table 2 for Nvalue in each group; denominator).

Carstens et al.





myofibroblast differentiation, epithelial cell proliferation, andhoming of inflammatory cells to close the breach and eliminatepathogens (33). These same innate functions are capable ofresponding to basement membrane breaches caused by neo-plastic growth. However, growth stimuli further exacerbatetumorigenicity, eliciting a classic "chicken-or-egg" dilemma, asto which tissue-type evolves first and promotes/recruits theother. This report supports tumor-promoting mutations in theepithelial cells occur first, which subsequently induce reactivestroma growth. In turn, reactive stroma is then required tocoevolve with a dysplastic epithelium to drive malignanttransformation to adenocarcinoma.We and others have previously shown the transforming

ability of chronic FGFR1 stimulationwhenexpressed inprostateepithelia cells (17, 39). Using the Pro-Cat� JOCK1 and Ubi-Cat� JOCK1 mice, we show that canonical WNT signaling accel-erates FGFR1-driven adenocarcinoma, due largely to an accel-erated conversion to a fibroblastic reactive stroma. Althoughthe mechanisms whereby WNT accelerates this wound-likereactive stroma are unknown, migratory and metastatic rolesfor theWNTpathwayhavebeen reported. For example, FGFandcanonical WNT signaling are highly expressed and result in theosteoblastic phenotype of AR-negative bone metastasis (40).Furthermore, anticancer therapeutics promote canonical WNTsignaling in prostate cancer stroma via WNT16B, leading tocarcinoma progression and therapeutic resistance (24). As a

correlate with the role of WNT signaling, we have identified thedecrease of TGF-bRII and an increase in TGF-b ligands in thesebigenic tumors as the focus of future study.

Others have shown that the decrease in TGF-bRII andconcomitant increase in overall signaling is found in morethan 60% of human prostate cancers (41), and is sufficient toinduce malignant transformation in prostate reconstitutionsand mouse models (42–45). Using known prostate-associatedoncogenes, we serendipitously observed a similar increase inTGF-b signaling, which correlates strongly with the coevolvingneoplasia (mPIN) and reactive stroma, as a likely mechanismfor progression from mPIN to adenocarcinoma. Likewise, in ahuman tissue array, we observed the same pattern of Smad2activation. Furthermore, tumorswith greater than 60% stromalTGF-b signaling significantly correlated with higher Gleasongrades, resembling the signaling pattern of themore aggressiveUbi-Cat� JOCK1mice. Thesefindings agreewith the "two-stepcomputational model", whereby TGF-b–nonresponsive stro-mal cells are required for the first (epithelial) transformationalstep, and TGF-b–responsive cells are needed for the second(invasion) step of prostate cancer progression (46). Interest-ingly, this model demonstrated that canonical WNT signalingwas a vital paracrine factor secreted by the TGF-bRII–deficientstromal cells as a mechanism for epithelial transformation.Although this model does not rule out other explanations, thisbipartite TGF-bmodel could further explain the higher levels of

No

CID

CID

WT LSC +

Ubi-Cat stroma

JOCK1 LSC +

Ubi-Cat stromaA JOCK1 LSC +

WT stroma

B

No CID

CID

WT LSC

Ubi-Cat stroma

JOCK1 LSC

Ubi-Cat stroma

JOCK1 LSC

WT stroma

8

6

4

2

** P < 0.01

Avera

ge a

cin

i are

a

(pix

els

2 ×

1,0

00)

Average acini area

**

0

Figure 4. WNT activation in thestroma alone is sufficient to inducefibroblastic reactive stroma. A,H&E analysis of WT or JOCK1basal progenitor cells (LSC) mixedwith adult Ubi-Cat stromal cellgrafts. Activation of stromal WNTfor 16 weeks increased aciniformation and a thickened thefibroblastic reactive stroma,compared with no CID. Activationof FGFR1 in the epithelium andWNT in the stroma also increasedacini number, size, and hyperplasiawith a fibroblastic reactive stroma.Scale bar, 100 mm. B, average aciniarea for each transplant type. n¼3.Significance determined byt test.






nuclear p-Smad2 staining seen in Ubi-Cat� JOCK1 animals, asenforced transgenic WNT signaling in the stromal compart-ment could circumvent the initial requirement of TGF-b–nonresponsive cells to secrete Wnt3a for the proliferationstep, allowing TGF-b–responsive (p-Smad2þ) cells to efficient-ly promote epithelial invasion and accelerated progression.

The observed epithelial phenotype of our double-transgenicmice correlates with progression hallmarks identified inpatients, even to the extent that sarcomatoid carcinomas arenow accepted as histologic manifestations of EMT (17, 33, 47).The tumor-associated stroma, however, reveals both similari-ties and distinct differences between mouse and human dis-eased tissue (Table 1). The reactive stroma seen in these mice"closely resembles the stroma of [the] aggressive human pros-tate cancer subgroup (RSG3)" (33). Furthermore, these modelsdisplay a fibroblastic reactive stroma (i.e., high-level collagenand vimentin staining, a decrease in SMA, and focal areas oftenascin c), all of which indicate a fibroblastic reactive stroma.In humans, these fibroblasts, more commonly referred to ascancer-associated fibroblasts (CAF), make up approximately

50% of reactive stroma in Gleason 3 foci and are progressivelyreplaced by myofibroblasts in Gleason 4 to 5 foci (30). In moreadvanced foci, myofibroblasts are the majority and share bothfibroblastic (i.e., tenascin c and vimentin) and smooth musclemarkers (i.e., SMA). This difference correlates with the dissim-ilarities seen in the normal stroma ofmice and humans. Normalmurine stroma is sparse,with only a thin layer of smoothmusclesurrounding the acini, whereas human stroma consists of denseareas of smoothmusclefilling the space surrounding the glands.Therefore, it is easy to conceptualize that the characterizationof the tumor-associated fibroblasts found in each species willcorrelate with the characteristics of the normal stroma. Incontrast, the predominant myofibroblasts may be derived fromCAFs or fromanother stem/progenitor cell types. Therefore, thelack of myofibroblasts in murine prostate cancer may not bedue to the startingmaterial, but a difference in progression and/or development. Despite this difference, these murine modelsclearly model CAFs, infiltration of immune cells, and upregula-tion of key signaling pathways, like those initiated by TGF-b,FGFs, and WNTs, which are all present in human cancers.

Tumor-associated stromal pSmad2

0

50

100

***

B

a ecb d

Pro

-Cat ×

JO

CK

1 E

p(C

ID)

Pro

-Ca

t ×

JO

CK

1 S

tr(C

ID)

Ubi-C

at ×

JO

CK

1 E

p(C

ID)

Ubi-C

at ×

JO

CK

1 S

tr(C

ID)

JO

CK

1 E

p(C

ID)

JO

CK

1 E

p(C

ID)

JO

CK

1 E

p(C

ID)

JO

CK

1 S

tr(C

ID)

JO

CK

1 S

tr(C

ID)

JO

CK

1 S

tr(C

ID)

JO

CK

1 S

tr(

No C

ID)

JO

CK

1 E

p(

No

CID

)C

ba

dc

fe

hg

f

A

Fold

change

(big

enic

vs.

JO

CK

1)

4

2

0

–2

Epith.Epith. StromaStroma

Pro-Cat ×

JOCK1

Ubi-Cat ×

JOCK1

TGF-βI

TGF-βIITGF-βIIITGF-βRITGF-βRIITGF-βRIII

D

% S

ma

d2

+ n

ucle

i

Pro-Cat×

JOCK1

Ubi-Cat×

JOCK1

GleasonJOCK1

<7

Gleason

7

Figure 5. FGFR1 and WNT promoteTGF-b signaling. A, selected genes(centered on centroid mean ofgroups) from epithelium/mPIN andreactive stoma LCM. a and d,JOCK1CID versus no CID; b and e,Pro-Cat � JOCK CID versusJOCK1 CID; c and f, Ubi-Cat �JOCK1CID versusJOCK1CID. Theexpression patterns of thepromoters correlate with the similarpatterns between Pro-Cat �JOCK1 and Ubi-Cat � JOCK1epithelium and the less similarpatterns between Pro-Cat �JOCK1 and Ubi-Cat � JOCK1stroma. There is a markeddownregulation of TGF-b receptorsand an upregulation of TGF-bligands in both the epithelium andstroma of Pro-Cat � JOCK1 andUbi-Cat � JOCK1 mice. B, qRT-PCR validation for TGF-b ligandsand receptors. C, pSmad2 IHC.Scale bar, 20 mm. a, normal murineepithelium; b, mPIN; c, normalhuman epithelium; d, PIN; e, normalmurine stroma; f, mouse reactivestroma, negative nuclei (blackarrow) immediately adjacent topositive nuclei (white arrowhead);g, normal human stroma; and h,human reactive stroma. D, meanpercentage of pSmad2þ stromalnuclei adjacent to tumor epithelium(n ¼ 20 for each murine group,n > 350 for each human; plotted þSD). No significant differencebetween JOCK1 and Pro-Cat �JOCK1. Significant increase in Ubi-Cat � JOCK1 in pSmad2þ stromalcells. Correlative pattern in higher(7 vs. <7) Gleason Score patienttumors. Two-way ANOVA, P < 0.05; , P < 0.01.

Carstens et al.





In conclusion, we have demonstrated the ability of canonicalWNT signaling to accelerate FGFR1-driven carcinogenesis, bypromoting a fibroblastic reactive stroma, characterized byincreased TGF-b heterogeneity. As reactive stroma has nowbeen shown to be predictive of prostate cancer-specific death(32), this work provides a powerful new model to test andvalidate the importance of reactive stroma in human prostatecancer progression.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: J.L. Carstens, P. Shahi, S. Tsang, J.M. Levitt, D.R.Rowley, D.M. SpencerDevelopment of methodology: J.L. Carstens, P. Shahi, B. Smith, A. Seamans,D.R. Rowley, D.M. SpencerAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): J.L. Carstens, P. Shahi, A. Seamans, M. Seethamma-gari, I. Vedula, M.M. Ittmann, D.R. RowleyAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): J.L. Carstens, P. Shahi, S. Tsang, C.J. Creighton,Y. Zhang, J.M. Levitt, M.M. Ittmann, D.M. Spencer

Writing, review, and/or revision of the manuscript: J.L. Carstens, C.J.Creighton, J.M. Levitt, M.M. Ittmann, D.R. Rowley, D.M. SpencerAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): J.L. Carstens, P. Shahi, S. Tsang, B.Smith, A. Seamans, J.M. Levitt, M.M. Ittmann, D.R. Rowley, D.M. SpencerStudy supervision: J.L. Carstens, J.M. Levitt, D.R. Rowley, D.M. Spencer

AcknowledgmentsThe authors thankW.Decker, K. Chan, J. Rosen, L. Xin, LDonehower, and their

laboratory staff for their technical expertise, generous sharing of reagents, andinvaluable discussions. AP20187 was generously provided by ARIAD Pharma-ceuticals (Cambridge, MA).

Grant SupportThis work was financially supported by the mouse models for human cancer

consortium (MMHCC) grant U01-CA84296, NIH R01 CA58093; the Cytometryand Cell Sorting Core at Baylor College of Medicine with funding from the NIH(AI036211, CA125123, and RR024574); and the National Cancer Institute P30Cancer Center support grant (P30 CA125123) for support of the Human TissueAcquisition and Pathology and Genetically Engineered Mouse Cores.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received April 15, 2013; revised October 8, 2013; accepted November 8, 2013;published OnlineFirst December 4, 2013.

References1. Cronauer M, Schulz W, Seifert H-H, Ackermann R, Burchardt M.

Fibroblast growth factors and their receptors in urological cancers:basic research and clinical implications. Eur Urol 2003;43:309–19.

2. Acevedo VD, Ittmann M, Spencer DM. Paths of FGFR-driven tumor-igenesis. Cell Cycle 2009;8:580–8.

3. Hardy KM, Yatskievych TA, Konieczka JH, Bobbs AS, Antin PB. FGFsignaling throughRAS/MAPK andPI3K pathways regulates cell move-ment and gene expression in the chicken primitive streak withoutaffecting e-cadherin expression. Dev Biol 2011;11:20.

4. Nusse R. Wnt signaling in disease and in development. Cell Res2005;15:28–32.

5. Chesire DR, Ewing CM, Sauvageot J, Bova GS, Isaacs WB. Detectionand analysis of beta-catenin mutations in prostate cancer. Prostate2000;45:323–34.

6. Huang S, Ting W, Chen L, Huang L, Liu C. Association analysis of wntpathway genes on prostate-specific antigen recurrence after radicalprostatectomy. Ann Surg Oncol 2010;17:312–22.

7. Akiyama H, Lyons JP, Mori-Akiyama Y, Yang X, Zhang R, Zhang Z,et al. Interactions between sox9 and beta-catenin control chondrocytedifferentiation. Genes Dev 2004;18:1072–87.

8. CarothersAM,Rizvi H,HassonRM,Heit YI, Davids JS,BertagnolliMM,et al. Mesenchymal stromal cell mutations and wound healing con-tribute to the etiology of desmoid tumors. Cancer Res 2012;72:346–55.

9. Chesire DR, Ewing CM, Gage WR, Isaacs WB. In vitro evidence forcomplex modes of nuclear b-catenin signaling during prostate growthand tumorigenesis. Oncogene 2002;21:2679–94.

10. Pearson HB, Phesse TJ, Clarke AR. K-ras andWnt signaling synergizeto accelerate prostate tumorigenesis in the mouse. Cancer Res 2009;69:94–101.

11. Ilyas M. Wnt signalling and the mechanistic basis of tumour develop-ment. J Pathol 2005;205:130–44.

12. Ciruna B, Rossant J. Mesoderm cell fate specification and morpho-genetic movement at the primitive streak. Dev Cell 2001;1:37–49.

13. Matsuyama M, Mizusaki H, Shimono A, Mukai T, Okumura K, Abe K,et al. A novel isoform of vinexin, vinexin gamma, regulates Sox9 geneexpression through activation ofMAPK cascade inmouse fetal gonad.Genes Cells 2005;10:421–34.

14. Mori-Akiyama Y, AkiyamaH, Rowitch DH, DeCrombrugghe B. Sox9 isrequired for determination of the chondrogenic cell lineage in thecranial neural crest. Proc Natl Acad Sci U S A 2003;100:9360–5.

15. PondAC,Herschkowitz JI, Schwertfeger KL,WelmB,ZhangY,YorkB,et al. Fibroblast growth factor receptor signaling dramatically accel-erates tumorigenesis and enhances oncoprotein translation in themouse mammary tumor virus–Wnt-1 mouse model of breast cancer.Cancer Res 2010;70:4868–79.

16. Horst D, Chen J, Morikawa T, Ogino S, Kirchner T, Shivdasani RA.Differential WNT activity in colorectal cancer confers limited tumori-genic potential and is regulated byMAPK signaling. Cancer Res 2012;72:1547–56.

17. Acevedo VD, Gangula RD, Freeman KW, Li R, Zhang Y, Wang F, et al.Inducible FGFR-1 activation leads to irreversible prostate adenocar-cinoma and an epithelial-to-mesenchymal transition. Cancer Cell2007;12:559–71.

18. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al.The cBio cancer genomics portal: an open platform for exploringmultidimensional cancer genomics data. Cancer Discov 2012;2:401–4.

19. Dell'accio F, De Bari C, Eltawil NM, Vanhummelen P, Pitzalis C.Identification of the molecular response of articular cartilage to injury,by microarray screening: Wnt-16 expression and signaling after injuryand in osteoarthritis. Arthritis Rheum 2008;58:1410–21.

20. Cheon SS, Nadesan P, Poon R, Alman BA. Growth factors regulateb-catenin-mediated TCF-dependent transcriptional activation in fibro-blasts during the proliferative phase of wound healing. Exp Cell Res2004;293:267–74.

21. Yokoyama H, Ogino H, Stoick-Cooper CL, Grainger RM, Moon RT.Wnt/beta-catenin signaling has an essential role in the initiation of limbregeneration. Dev Biol 2007;306:170–8.

22. Lin G, Slack JMW. Requirement for Wnt and FGF signaling in xenopustadpole tail regeneration. Dev Biol 2008;316:323–35.

23. Aman A, Piotrowski T. Wnt/beta-catenin and Fgf signaling controlcollective cell migration by restricting chemokine receptor expression.Dev Cell 2008;15:749–61.

24. Sun Y, Campisi J, Higano C, Beer TM, Porter P, Coleman I, et al.Treatment-induced damage to the tumor micro-environment pro-motes prostate cancer therapy resistance through WNT16B. Nat Med2012;18:1359–68.

25. Zhang J, Thomas TZ, Kasper S, Matusik RJ. A small compositeprobasin promoter confers high levels of prostate-specific geneexpression through regulation by androgens and glucocorticoids invitro and in vivo. Endocrinology 2000;141:4698–710.






26. Freeman KW, Welm BE, Gangula RD, Rosen JM, Ittmann M, Green-berg NM, et al. Inducible prostate intraepithelial neoplasia with revers-ible hyperplasia in conditional FGFR1-expressing mice. Cancer Res2003;63:8256–63.

27. Freeman KW, Gangula RD, Welm BE, Ozen M, Foster BA, RosenJM, et al. Conditional activation of fibroblast growth factor receptor(FGFR) 1, but not FGFR2, in prostate cancer cells leads to increasedosteopontin induction, extracellular signal-regulated kinase activa-tion, and in vivo proliferation. Cancer Res 2003;63:6237–43.

28. Shahi P, Park D, Pond AC, Seethammagari M, Chiou S-H, Cho K,et al. Activation of Wnt signaling by chemically induced dimeriza-tion of LRP5 disrupts cellular homeostasis. PLoS ONE 2012;7:e30814.

29. Barron DA, Strand DW, Ressler SJ, Dang TD, Hayward SW, Yang F,et al. TGF-b1 induces an age-dependent inflammation of nerve gangliaand fibroplasia in the prostate gland stroma of a novel transgenicmouse. PLoS ONE 2010;5:e13751.

30. Tuxhorn JA, Ayala GE, Smith MJ, Smith VC, Dang TD, Rowley DR.Reactive stroma in human prostate cancer: induction of myofibroblastphenotype and extracellular matrix remodeling. Clinical Cancer Res2002;9:2912–23.

31. Xin L, Lukacs RU, Lawson DA, Cheng D, Witte ON. Self-renewal andmultilineage differentiation in vitro from murine prostate stem cells.Stem Cells 2007;25:2760–9.

32. Saldanha AJ. Java Treeview–extensible visualization of microarraydata. Bioinformatics 2004;20:3246–8.

33. Ittmann M, Huang J, Radaelli E, Martin P, Signoretti S, Sullivan R,et al. Animal models of human prostate cancer: the consensusreport of the New York meeting of the mouse models of humancancers consortium prostate pathology committee. Cancer Res2013;73:2718–36.

34. Li X, Sterling JA, FanK-H, Vessella RL, Shyr Y,HaywardSW, et al. Lossof TGF-b responsiveness in prostate stromal cells alters chemokinelevels and facilitates the development ofmixed osteoblastic/osteolyticbone lesions. Mol Cancer Res 2012;10:494–503.

35. Dakhova O, Ozen M, Creighton CJ, Li R, Ayala G, Rowley D, et al.Global gene expression analysis of reactive stroma in prostate cancer.Clin Cancer Res 2009;15:3979–89.

36. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichmenttools: paths toward the comprehensive functional analysis of largegene lists. Nucleic Acids Res 2009;37:1–13.

37. Huang DW, Sherman BT, Lempicki RA. Systematic and integrativeanalysis of large gene lists using DAVID bioinformatics resources. NatProtoc 2009;4:44–57.

38. Yin Y, Wang F, Ornitz DM. Mesothelial- and epithelial-derived FGF9have distinct functions in the regulation of lung development. Devel-opment 2011;138:3169–77.

39. Elo TD, Valve EM. Stromal activation associated with development ofprostate cancer in prostate-targeted fibroblast growth factor 8b trans-genic mice. Neoplasia 2010;12:915–27.

40. Wan X, Liu J, Lu J-F, Tzelepi V, Yang J, Starbuck MW, et al. Activationof b-catenin signaling in androgen receptor-negative prostate cancercells. Clin Cancer Res 2012;18:726–36.

41. Li X, Placencio V, Iturregui JM, Uwamariya C, Sharif-Afshar A-R,KoyamaT, et al. Prostate tumor progression ismediated by a paracrineTGF-beta/Wnt3a signaling axis. Oncogene 2008;27:7118–30.

42. FrancoOE, JiangM,StrandDW,Peacock J, FernandezS, JacksonRS,et al. AlteredTGF-b signaling in a subpopulation of humanstromal cellspromotes prostatic carcinogenesis. Cancer Res 2011;71:1272–81.

43. Bhowmick NA, Chytil A, Neilson EG, Moses HL. TGF-b signaling infibroblasts modulates the oncogenic potential of adjacent epithelia.Science 2004;303:848–851.

44. Li X, Wang Y, Sharif-Afshar A-R, Uwamariya C, Yi A, Ishii K, et al.Urothelial transdifferentiation to prostate epithelia is mediated byparacrine TGF-beta signaling. Differentiation 2009;77:95–102.

45. Kiskowski MA, Jackson RS, Banerjee J, Li X, Kang M, Iturregui JM,et al. Role for stromal heterogeneity in prostate tumorigenesis. CancerRes 2011;71:3459–70.

46. MulhollandDJ,KobayashiN,RuscettiM, Zhi A, Tran LM,Huang J, et al.Pten loss and RAS/MAPK activation cooperate to promote EMT andmetastasis initiated from prostate cancer stem/progenitor cells. Can-cer Res 2012;72:1878–89.

47. Ayala GE, Muezzinoglu B, Hammerich KH, Frolov A, Liu H, ScardinoPT, et al. Determining prostate cancer-specific death through quan-tification of stromogenic carcinoma area in prostatectomy specimens.Am J Pathol 2011;178:79–87.

Carstens et al.





2014;74:609-620. Published OnlineFirst December 4, 2013.Cancer Res Julienne L. Carstens, Payam Shahi, Susan Van Tsang, et al. Recapitulates Human Reactive Stroma

Signaling in Prostate Cancer Mouse ModelsβTGF-−WNT−FGFR1

Updated version

10.1158/0008-5472.CAN-13-1093doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2013/12/04/0008-5472.CAN-13-1093.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/2/609.full#ref-list-1

This article cites 47 articles, 18 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/74/2/609.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/74/2/609To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-13-1093

http://cancerres.aacrjournals.org/content/suppl/2013/12/04/0008-5472.CAN-13-1093.DC1

http://cancerres.aacrjournals.org/content/74/2/609.full#ref-list-1

http://cancerres.aacrjournals.org/content/74/2/609.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/74/2/609


TGF-b Signaling in Prostate Cancer Mouse Models...

Documents

Transcript of TGF-b Signaling in Prostate Cancer Mouse Models...