inhibition of ciliogenesis promotes hedgehog signaling, … · signal transduction inhibition of...

Signal Transduction

Inhibition of Ciliogenesis Promotes HedgehogSignaling, Tumorigenesis, and Metastasis inBreast CancerNadia B. Hassounah1, Martha Nunez1, Colleen Fordyce2, Denise Roe1,3,Ray Nagle1,4, Thomas Bunch1, and Kimberly M. McDermott1,5,6

Abstract

Primary cilia are chemosensors that play a dual role to eitheractivate or repress Hedgehog signaling, depending on presenceor absence of ligand, respectively. While inhibition of ciliogen-esis has been shown to be characteristic of breast cancers, thefunctional consequence is unknown. Here, for the first time,inhibition of ciliogenesis led to earlier tumor formation, fastertumor growth rate, higher grade tumor formation, andincreased metastasis in the polyoma middle T (PyMT) mousemodel of breast cancer. In in vitro model systems, inhibition ofciliogenesis resulted in increased expression of Hedgehog-tar-get genes through a mechanism involving loss of the repressorform of the GLI transcription factor (GLIR) and activation ofHedgehog target gene expression through cross-talk with TGF-alpha (TGFA) signaling. Bioinformatics analysis revealed that

increased Hedgehog signaling is frequently associated withincreased TGFA signaling in patients with triple-negative breastcancers (TNBC), a particularly aggressive breast cancer subtype.These results identify a previously unrecognized role for inhi-bition of ciliogenesis in breast cancer progression. This studyidentifies inhibition of ciliogenesis as an important event foractivation of Hedgehog signaling and progression of breastcancer to a more aggressive, metastatic disease.

Implications: These findings change the way we understand howcancer cells turn on a critical signaling pathways and a providerationale for developing novel therapeutic approaches to targetnoncanonical Hedgehog signaling for the treatment of breastcancer. Mol Cancer Res; 15(10); 1421–30. �2017 AACR.

IntroductionBreast cancer is the leading cause of cancer-related deaths in

womenworldwide (1). Analysis of human samples demonstratesthat Hedgehog signaling is increased in 20%–30% of breastcancers. Experimental inhibition of Hedgehog activity in breastcancer models reduces tumor growth and metastasis in vivosuggesting that targeting this pathway is a promising therapeuticstrategy. Inhibition of the Hedgehog pathway (SMO inhibitors)has been shown to be effective in the treatment of basal cellcarcinoma; however, clinical trials using SMO inhibitors were notsuccessful in the treatment of many other Hedgehog-positivecancers including breast cancer. While canonical Hedgehog sig-

naling is dependent on SMO, it is now clear that cancer cells canactivate Hedgehog signaling via alternativemechanisms indepen-dent of SMO activity and therefore render them resistant to SMOinhibitors. A better understanding of how Hedgehog signaling isactivated in breast cancers is needed to design novel therapeuticstrategies to target this important oncogenic pathway.

Primary cilia are specialized immotilemicrotubule-based orga-nelles that protrude from the plasma membrane of many verte-brate cell types including mammary epithelial cells (2). Thefunctional significance of primary cilia has been made clear bythe discovery that defects in ciliogenesis or their function plays acausal role in a range of severe diseases and developmentaldisorders. Primary cilia are essential for both activation andrepression of Hedgehog signaling, depending on the presence orabsence of Hedgehog ligand, respectively. Key proteins thatregulate Hedgehog signaling, including PTCH, SMO, and GLItranscription factors, localize to the primary cilium (3). Canonicalactivation of the Hedgehog pathway occurs through the bindingof Hedgehog ligands to the PTCH receptor, thereby relievinginhibition of SMO (4). Activated SMO then translocates to thecilium resulting in posttranslational processing of GLI into itsactivated form (GLIA). GLIA then translocates to the nucleus fortranscriptional activation of Hedgehog target genes. In theabsence of Hedgehog ligand, GLI is posttranslationally modifiedfor transcriptional repression (GLIR) and then translocates fromthe cilium to the nucleus where it represses expression of Hedge-hog target genes. We, along with others, have demonstrated thatciliogenesis is inhibited in breast cancer cells (5–7); however, itsbiological consequence is not understood. The findings in thisreport demonstrate for the first time that inhibition of ciliogenesis

1TheUniversity of Arizona Cancer Center, University of Arizona, Tucson, Arizona.2Department of Biochemistry and Molecular Biology, University of New MexicoHealth Sciences Center, Albuquerque, New Mexico. 3Department of Epidemi-ology and Biostatistics, University of Arizona, Tucson, Arizona. 4Department ofPathology, University of Arizona, Tucson, Arizona. 5Department of Medicine,University of Arizona, Tucson, Arizona. 6Bio5 Institute, University of Arizona,Tucson, Arizona.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Current address for N. B. Hassounah: Dana-Farber Cancer Institute, Boston,Massachusetts.

Corresponding Author: Kimberly M. McDermott, University of Arizona, MedicalResearch Building, 1656 E Mabel St. Room #120, PO Box 245215, Tucson, AZ85724. Phone: 520-626-8231; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-17-0034

�2017 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 1421

on March 19, 2021. © 2017 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst June 13, 2017; DOI: 10.1158/1541-7786.MCR-17-0034

http://crossmark.crossref.org/dialog/?doi=10.1158/1541-7786.MCR-17-0034&domain=pdf&date_stamp=2017-9-15

http://mcr.aacrjournals.org/

promotes breast cancer progression to a more aggressive andmetastatic disease. Furthermore, our findings facilitate a betterunderstanding of noncanonical activation of Hedgehog signalingin breast cancer by demonstrating for thefirst time the importanceof inhibition of ciliogenesis for Hedgehog target gene expression.

Materials and MethodsMice

Mice were bred andmaintained in our colony at the Universityof Arizona Animal Facility. The mice used were Ift88fl/fl mice (Dr.Bradley Yoder, University of Alabama, Birmingham, AL; ref. 8),Kif3afl/fl mice (Dr. Jeremy Reiter, University of California, SanFrancisco; ref. 9), andMMTV-PyMT and Ift20fl/flmice (Dr. GregoryPazour, University of Massachusetts Medical Center and Jacksonlabs). We generated Ift88fl/fl/MMTV-PyMT and Ift20fl/fl/MMTV-PyMT mice by crossing Ift88fl/fl or Ift20fl/fl to MMTV-PyMT mice.Recipient mice for mammary fat pad transplant surgeries wereSCID/Beige mice (Charles River). All procedures were performedin accordance with our Institutional Animal Care and Use Com-mittee–approved protocol.

Mammary epithelial cell preparationCell preparations were performed with fresh mammary

glands, or alternatively, with glands that had been frozen in90% FBS and 10% DMSO, and were thawed upon epithelial cellharvest. Mammary epithelial cells were isolated on the basis ofthe method described by Prater and colleagues (10). In short,harvested glands from a single mouse were digested for 14–16hours at 37�C in collagenase/hyaluronidase (Stemcell Technol-ogies, catalog no. 07919) in DMEM/F12/H (Gibco, catalog no.11330-032) supplemented by 50 mg/mL gentamicin (Gibco,catalog no. 15750-060). Digested tissue was resuspended incold HF [HBSS (Gibco, catalog no. 14025-092), with 10mmol/L HEPES (Gibco, catalog no. 15630), and 2% heat-inac-tivated FBS; Gibco, catalog no. 16140-071)] and NH4Cl (Stem-cell Technologies, catalog no. 07800) to lyse red blood cells.After centrifugation to remove the supernatant, trypsin (StemcellTechnologies, catalog no. 07400) was added for 1.5 minutes. Todilute the trypsin, 10 mL of cold HF was added, and the mixturewas spun to remove the supernatant. Further protease digestionwas performed with trypsin (Stemcell Technologies, catalog no.07400) followed by treatment with dispase (Stemcell Technol-ogies, catalog no. 07913) and 1 mg/mL DNAse (Sigma, catalogno. DN25). The cell prep was then filtered through a 40-mm cellstrainer (BD Falcon, catalog no. 352340). The Mouse EpithelialEnrichment Kit (Stemcell Technologies, catalog no. 19758) wasused for enrichment of epithelial cells according to manufac-turer's instructions. Isolated epithelial cells were directly used foradenovirus treatment for subsequent mammary fat pad trans-plantations (Ift88fl/fl/PyMTþ and Ift20fl/fl/PyMTþ studies). Alter-natively, epithelial cells were plated for a week according to Liuand colleagues (11) to expand cell number (Kif3afl/fl and Ift88fl/fl

studies) and then trypsinized for adenovirus treatment andmammary fat pad transplantation.

Adenovirus treatmentAdenovirus transduction was performed at 20 MOI using Ad5-

CMVeGFP (Ad-GFP, 2 � 1010 TU/mL) or Ad5-CMV-Cre-eGFP(Ad-Cre-GFP, 5 � 1010 TU/mL; University of Iowa Gene TransferVector Core). Viraductin Adenovirus kit (Cell Biolabs Inc., catalog

no. Ad-201) was used to infect 1 � 105 epithelial cells in HBSScontaining 10 mmol/L HEPES.

Mammary fat pad transplantationSCID/Beige 3-week-old pups, between 8 and 11 g, were used as

recipients for transplant surgeries. Both inguinal fat pads permouse were cleared surgically prior to epithelial cell transplan-tation. For tumor growth studies, Ad-GFP–treated epithelial cellswere injected into one fat pad, and Ad-Cre-GFP–treated cells wereinjected into the contralateral fat pad of the same mouse. Formetastasis studies, Ad-GFP–treated or Ad-Cre-GFP–treated epi-thelial cells were injected into both fat pads of the same mouse.Adenovirus-treated epithelial cells were resuspended in 65%HBSS, 25% growth factor–reduced Matrigel (BD Biosciences,catalog no. 356231), and 10% Trypan blue (Thermo ScientificHyclone, catalog no. SV3008401) and injected into the clearedinguinal fat pad using a Hamilton microliter syringe. For Ift88fl/fl

and Kif3afl/fl surgeries, 5 � 104 cells were injected per transplant.For Ift88fl/f/MMTV-PyMT and Ift20fl/fl/MMTV-PYMT surgeries, 3�104 cells were injected per transplant.

Tumor growthInguinal mammary glands were palpated twice per week

postsurgery, and measurements were collected using a digitalcaliper upon detection of a mass that did not regress uponsuccessive palpation. The two longest diameters were mea-sured, and tumor volume (mm3) was calculated with a mod-ified ellipsoid equation: (width2)�(height/2). Mice were eutha-nized upon reaching maximum tumor burden (10% of bodyweight). Alternatively, mice were euthanized 23 weeks post-surgery (Ift88fl/fl/MMTV-PyMT mice) or 25 weeks postsurgery(Ift20fl/fl/MMTV-PyMT mice). For premalignant studies, mam-mary glands were harvested at 5 weeks postsurgery. All mam-mary tumors were cut into thirds for separate analyses:immersed in RNA-later RNA Stabilization Reagent (catalog no.76106, Qiagen) and frozen at �80�C for real-time PCR, flashfrozen for protein isolation, and fixed in 4% paraformaldehyde(Electron Microscopy Sciences, catalog no. 15714-S) at 4�Covernight and paraffin embedded for IHC and immunofluo-rescence. For metastasis studies, the lungs were harvested after20 weeks for whole mounting and paraffin embedding.

Kaplan–Meier curves were generated, and log-rank tests wereperformed on tumor-free survival data using SPSS 21 (StatisticalPackage for the Social Sciences; IBM Corporation). Differences intumor growth rateswere analyzedwith STATA 12. Tumor volumesfor tumor growth rate statistics were first transformed to inducenormality. A mixed model with random effects to account forrepeated measures per mouse within treatment groups was thenused.Mammary glandwholemountswereperformedasdescribedpreviously (2).

Immunocytochemistry and IHCCell and tissue stainingwere performed as previously described

(2, 7, 12). Adenovirus-treated MECs were plated on coverslips(Fisherbrand 12-545-81, 12CIR-1.5) and grown to confluencyand then serum starved in 0.1% serum-containing media for 24hours. BT-549 cells were serum starved in 0.5% serum-containingmedia for 24 hours.

For all immunofluorescent staining, the primary antibodiesused included: ARL13b (1:400, mouse monoclonal IgG2a, UC,Davis/NIH NeuroMab Facility, clone N295B/66), acetylated

Hassounah et al.

Mol Cancer Res; 15(10) October 2017 Molecular Cancer Research1422




tubulin (1:1,000, mouse monoclonal IgG2B, Sigma, catalog no.T7451, clone 6-11B-1), g-tubulin (mouse monoclonal IgG1,Sigma, catalog no. T5326, clone GTU-88), CK5 (1:400, rabbitpolyclonal, Abcam, catalog no. ab53121). Secondary antibodiesused included: Alexa Fluor 555–labeled goat anti-mouse-IgG2B

(Invitrogen, catalog no. A21147), Alexa Fluor 633–labeled goatanti-mouse-IgG1 (Invitrogen, catalog no. A21126), Alexa Fluor546–labeled goat anti-mouse-IgG2a (Invitrogen, catalog no.A21133) and Alexa Fluor 488–labeled goat anti-rabbit-IgG (Invi-trogen, catalog no. A11034). Hoechst 33342 (catalog no. H3570,Invitrogen) was used as a counterstain. Slides were mounted with1.5 coverslips (0.16–0.19 mm thickness, Fisher Scientific, catalogno. 12-544B) using Prolong Gold Antifade mounting media(catalog no. P36934, Invitrogen).

For IHCusing the primary antibody recognizing SLUG (SNAI2;1:400, C19G7, Cell Signaling Technology, rabbit monoclonal,catalog no. 9585), the primary antibody was detected with auniversal anti-rabbit, anti-mouse secondary antibody conjugatedto peroxidase (ImmPress Universal Reagent, Vector Laboratories,catalog no. MP-7500). 3,30-diaminobenzidine (DAB, Dako, cat-alog no. K3467) was used as the peroxidase substrate. The tissueswere counterstained with Hematoxylin 1 (Thermo Scientific,catalog no. 7221). Tissue slides with secondary antibody onlywere used as a negative control. Slides were mounted withFaramount Aqueous Mounting Media (Dako, catalog no.S3025) using 1.5 coverslips (0.16–0.19 mm thickness, FisherScientific, catalog no. 12-544B).

Immunofluorescence imaging and analysisIHC was performed as described previously (2, 7, 12). At least

three images were acquired per tissue to quantitate cilia, with atleast 150 total nuclei per sample. The presence of cilia in adjacentstroma, as well as centrosomes in the tumor tissue (g-tubulin),was used as internal staining quality control. Primary cilia werequantified manually frommaximum projections. The percentageof ciliated cells was quantified for basal cells (CK5þ) and luminalepithelial cells (CK5�) separately.

RNA isolation and real-time PCRTumor samples that were used for PCR analysis were selected as

pairs where Ciliaþ/PyMTþ and Cilia�/PyMTþ tumors from thesamemouse had an equal percentage of cancer cells (ratio of 1.0�0.2). This selection criteria served as a control for variations ingene expression due to disproportionate ratios of cancer cells tothe stroma. The RNeasy Lipid TissueMini kit (Qiagen, catalog no.74804) was used for RNA isolation according to manufacturer'sinstructions. Mammary glands frozen in RNAlater RNA Stabili-zation reagent were thawed and homogenized in Qiazol LysisReagent (Qiagen, catalog no. 79306) using a Powergen Model1000Homogenizer (Thermo Fisher Scientific). For RNA isolationfrom cell lines, RNAwas harvested from cells in 24wells using theRNeasy mini kit (Qiagen, catalog no. 74104) according to man-ufacturer's instructions. RNAwas on-columnDnase-treated usingan RNase-Free Dnase set (Qiagen, catalog no. 79254). All RNAsamples were checked for concentration and A260/280 andA260/230 ratios using a NanoDrop 1000 (Thermo Scientific).RNA samples were run on a polyacrylamide gel to check forribosomal bands and RNA degradation. Only samples of highquality (260/280 ratios of 1.8–2.1 and 260/230 ratios of above1.5, and minimal degradation) were used for further analysis.First-strand cDNA was generated with 1 mg of RNA using the

Maxima First Strand cDNA Synthesis Kit or the Ambion Retro-script Kit for RT-qPCR (Thermo Scientific, catalog no. K1641 and#AM1710). Real-time PCR was performed using the EXPRESSSYBR GreenER qPCR Supermix Universal kit (Invitrogen, catalogno. A10314). A total of 20 ng of cDNA was used per reaction,and reactions were performed in duplicates. A final primer con-centration of 300 nmol/L was used.

The primer sequences for mouse used were:

* mGli1-F: 50-GCGGAAGGAATTCGTGTGCC-30,* mGli1-R: 50-CGACCGAAGGTGCGTCTTGA-30,* mGli2-F: 50-TTTGCCGATTGACATGAGACA-30,* mGli2-R: 50-GGTGGGAGGCCCGTGTAC-30,* mPtch1-F: 50-CTCTGGAGCAGATTTCCAAGG-30,* mPtch1-R: 50-TGCCGCAGTTCTTTTGAATG-30,* mSnai2-F: 50-GGCTGCTTCAAGGACACATT-30,* mSnai2-R: 50-TTGGAGCAGTTTTTGCACTG-30,* mb-Actin-F: 50-CACAGCTTCTTTGCAGCTCCTT-30,* mb-Actin-R: 50-CGTCATCCATGGCGAACTG-30.

The following primer sequences for human were used:* hGLI1-F: 50-GAAGTCATACTCACGCCTCGAA-30,* hGLI1-R: 50- CAGCCAGGGAGCTTACATACAT-30,* hb-ACTIN-F: 50-AGAGCTACGAGCTGCCTGAC-30,* hb-ACTIN-R: 50- AGCACTGTGTTGGCGTACAG-30.

Real-time PCRwas run on a standard ABI 7500 or a Roche LightCycles 480 II. RelativemRNA fold changewas calculated using the2�DDCt method, using b-actin as a loading control. Fold changeswere calculated between paired tumors from a single mousetreated with Ad-Cre-GFP and Ad-GFP (grown in separate mam-mary glands in the same mouse). Average fold changes and SEwere calculated per gene acrossmice. Two-sided paired t tests wereperformed on paired fold changes.

Protein isolation and Western blottingFlash-frozen tissue was submerged in RIPA buffer (150

mmol/LNaCl, 1%NP-40, 0.5% sodiumdeoxycholate, 0.1% SDS,50 mmol/L Tris, pH 8.0) with protease inhibitors (Roche#04693159001) and phosphatase inhibitors (0.37 mg/mL sodi-um orthovanadate, 0.42 mg/mL sodium fluoride, 0.012 mg/mLammonium molybdate) and homogenized using an electrichomogenizer (Kinematica Polytron). Homogenized lysates werebriefly sonicated, spun at 4,000 rpm for 20minutes at 4�C, and thesupernatant was aliquoted to a new tube. Cell lysates from celllines were harvested by incubating RIPA buffer on plated cells.Lysates were then briefly sonicated, spun at 12,000 � g, and thesupernatant was aliquoted to a new tube. Protein concentrationswere measured using a Pierce BCA assay (Thermo Scientific).Protein lysates were run on an SDS-PAGE gel and blotted to anImmobilon-FL membrane (Millipore #IPFL00010). The mem-brane was blocked with Odyssey blocking buffer (LI-COR # 927-4000) and then probed with antibodies overnight at 4�C dilutedin Odyssey blocking buffer. Antibodies used were against: phos-pho-Akt (p-Akt, Ser 473, clone 587F11,mouse monoclonal, CellSignaling Technology, catalog no. 4051; 1:1,000), phospho-ERK1/2 (p-ERK1/2, Thr202/Tyr204, Clone D13.14.48, rabbitmonoclonal, Cell Signaling Technology, catalog no. 4370;1:2,000), GLI3 (R&D Systems, catalog no. AF3690, 1:500), andb-actin (clone AC-15, mouse monoclonal, Sigma, catalog no.A1978; 1:10,000). Membranes were then incubated with LI-CORfluorescent secondary antibodies (1:10,000), and membraneswere imaged using an LI-COR Odyssey Imaging System.

Primary Cilia Loss Promotes Aggressive Breast Cancer

www.aacrjournals.org Mol Cancer Res; 15(10) October 2017 1423




Maintenance of cell linesBT-549 cells were a kind gift from the laboratory of Ghassan

Mouneimne.NIH3T3 cellswere obtained from theATCC. BT-549cells were maintained in RPMI1640 with L-glutamine (CellGro,catalog no. 10-040-CV), sodium pyruvate (Gibco, catalog no.11360–070), 1 mg/mL bovine insulin (Sigma, catalog # I5500),10% FBS, 10 mmol/L HEPES (Gibco, catalog. # 15630-080), andantibiotic. NIH 3T3 cells were maintained in DMEMwith 4.5 g/Lglucose, L-glutamine and sodium pyruvate (CellGro, catalog no.10-013-CV), 10% FBS (HyClone, Lot no. AUE34991, catalog no.SH30541.03), and antibiotic (Gibco, catalog no. 15240-062).This cell line was authenticated by the ATCC testing service, andmycoplasma testing was performed immediately prior to use ofthis cell line for experiments described in this study.

Ligand treatmentsBT-549 cells were seeded in a 24-well plate as to be at subcon-

fluency at the time of lysate harvest (4� 104 for NIH 3T3 cells, 2�104 for BT549 cells). Twenty-four hours postseeding, completemediawere replacedwith serum-reducedmedia (0.5%FBS). Aftertwenty-four hours of serum starvation, cells were treated with 200nmol/L SAG or 100 ng/mL TGFA (R&D Systems, catalog no. 239-A-100 mg) in serum-reducedmedia for 24 hours. Cell lysates werethen collected for RNA harvesting.

Drug treatmentsFor Forskolin treatments, BT-549 cells were seeded in a 6-well

plate so as to be at confluency at time of lysate harvest (1.5� 105).Forty-eight hours postseeding, complete media were replacedwith serum-reduced media (0.5% FBS), or with complete media.After twenty-four hours of serum starvation, cells were treatedwith 100 mmol/L Forskolin (Sigma, catalog no. F6886) in serum-reduced or complete media for 24 hours. Cell lysates were thencollected for RNA and protein harvesting.

Statistical analysisThe remaining statistics performedwerewithpaired two-sided t

tests usingMicrosoft Excel. AP value less than0.05was consideredsignificant.

ResultsThe goal of the current study was to investigate whether

inhibition of primary cilia contributes to tumor formation. Wefirst assessed whether inhibiting ciliogenesis in the mammarygland is sufficient to cause tumor formation. Loss of expressionof the Kif3A or Ift88 genes, which are required for intraflagellartransport (IFT), have been shown to inhibit ciliogenesis (13, 14).Primary mammary epithelial cells (MEC) from mice harboringfloxed Kif3A or Ift88 genes were isolated and treated with anadenovirus expressing GFP-tagged Cre-recombinase (Ad-Cre-GFP) to delete the Kif3A (Kif3A-) or Ift88 (Ift88-) alleles, or cellswere treated with an adenovirus expressing GFP alone (Ad-GFP)as a control (Kif3Aþ or Ift88þ). Two models were utilized tovalidate that any phenotypes observed were due to inhibition ofciliogenesis and not due to noncilia-related functions of KIF3A orIFT88. We confirmed that ciliogenesis was inhibited in epithelialcells ofKif3a- and Ift88- glands (Supplementary Fig. S1Aand S1B).Next, mice were followed for 16–17 weeks posttransplant andglands were found to be morphologically normal, with no indi-cation of tumor growth (Supplementary Fig. S1C). Glands lacked

histologic abnormalities such as hyperplasia, dysplasia, or carci-noma in situ (Supplementary Fig. S1D). Others have also reportedthat loss of Ift88 in the mammary gland in the mouse does notresult in tumor formation, even up to 18 months of age (15).Together, these data indicate that loss of primary cilia alone is notsufficient to cause tumor formation.

Given that cancer is a multi-hit disease and that inhibition ofciliogenesis is a common/early event in human breast cancer, wehypothesized that inhibition of ciliogenesis synergizes with otheroncogenic events to promote tumor progression. We tested thishypothesis in the polyomamiddle T antigen (PyMT) breast cancermodel (16). To inhibit ciliogenesis in vivo, we crossed MMTV-PyMT mice with two different mouse models that allow for Cre-specific removal of genes required for ciliogenesis (mice carryingfloxed alleles of either Ift88- or Ift20-treated with Ad-Cre-GFP orAd-GFP). We validated that ciliogenesis was inhibited in epithe-lial cells of Ift88�/PyMTþ and Ift20-�/PyMTþ glands (Fig. 1A andB; Supplementary Fig. S2).

We investigatedwhether tumor onset is enhanced by inhibitionof ciliogenesis in the context of PyMT expression. IFT88/PyMTþ

and IFT20/PyMTþ transplants treatedwithAd-Cre-GFPorAd-GFPwere analyzed for tumor growth.We observed significantly earliertumor formation in cilia-negative glands (Cilia�/PyMTþ) com-paredwith cilia-positive glands (Ciliaþ/PyMTþ), upondeletion ofIft88 and Ift20 (Fig. 1C). Palpable tumors formed 4–6 weeksearlier in Cilia�/PyMTþ glands versus Ciliaþ/PyMTþ glands (Sup-plementary Table S1 and S2). These data support the conclusionthat inhibition of ciliogenesis decreases the time to tumor for-mation in the PyMT model.

To determine whether inhibiting ciliogenesis leads to earliertumor formation by increasing the frequency of premalignantlesions, we harvested mammary glands 5 weeks posttransplanta-tion and examined premalignant lesions in Ift88/PyMTþ andIft20/PyMTþ transplants treated with Ad-Cre-GFP or Ad-GFP.Tissue sections from the glands were stained with hematoxylinand eosin (H&E) to quantify the percent of epithelium as normal,hyperplasia, and carcinoma in situ (CIS). Inhibition of ciliogenesisdid not increase the frequency of hyperplasia or CIS in Ift88�/PyMTþ and Ift20�/PyMTþ transplants (Supplementary Fig. S3).These findings suggest that inhibition of ciliogenesis in the PyMTmodel enhances the growth of malignant lesions, rather thanincreasing progression from normal to premalignant lesions.

We also analyzed tumor growth rate in Ift88/PyMTþ and Ift20/PyMTþ transplants treated with Ad-Cre-GFP or Ad-GFP. Thetumor growth rate was found to be significantly increased inIft88�/PyMT tumors (Fig. 1D and E). The tumor growth rate wasalso increased in Ift20�/PyMT tumors, but thiswas not statisticallysignificant, likely due to a smaller sample size. These data indicatethat inhibition of ciliogenesis increases the rate of tumor growthby cooperating with the PyMT oncogene.

To determine whether inhibition of ciliogenesis altered thehistologic features of the tumors, we harvested end-stage tumorsand analyzed tissue sections with H&E staining. Tumors wereblindly graded by a certified pathologist based on increaseddisorganization of tissue architecture and increased nuclear pleo-morphism (premalignant lesion in situ and invasive cancer grade1, 2, and3; Fig. 1F).We termed grade1 as lowgrade,while grades 2and 3 were grouped as high grade. While we observed in situ onlyCiliaþ/PyMTþ tumors, we did not observe any in situ only Cilia�/PyMTþ tumors (Fig. 1G). A similar frequency of grade 1 tumorswas observed in Cilia�/PyMTþ and the respective Ciliaþ/PyMT

Hassounah et al.





Figure 1.

Inhibition of ciliogenesis increases tumorigenesis, rate of tumor growth, and tumor grade in the PyMT breast cancer model. A, ICC staining of Ift88fl/fl/PyMTþ

mammary tissues sections for ARL13b (cilia marker; red), g-tubulin (centrosome marker; green), CK5 (myoepithelial cell marker; white), and Hoechst(nuclei). The lumen (lum) and stromal (str) regions of the mammary gland are separated by a dashed line. Arrowheads point to primary cilia in Ad-GFP–treatedglands, and to centrosomes without cilia in Ad-Cre-GFP–treated glands. B, The bar graph represents quantification of the percent ciliated epithelial cells[basal (bas), luminal (lum), total epithelial cells (all)] from Ift88fl/fl/PyMTþ [Ad-GFP (blue), n ¼ 3; Ad-Cre-GFP (red), n ¼ 3] and Ift20fl/fl/PyMTþ [Ad-GFP,n ¼ 4; Ad-Cre-GFP, n ¼ 4] mammary glands. Error bars indicate SE; P values were determined by the Student t test (paired, two-tailed); � , P < 0.05. Kaplan–Meiercurves for tumor-free survival (C) and average tumor volume versus time graphed for transplants of Ift88fl/fl/PyMTþ (n ¼ 9) and Ift20fl/fl/PyMTþ (n ¼ 6)mammary epithelial cells treated with Ad-Cre-GFP (red) and Ad-GFP (blue; C). Error bars indicate SE; log-rank test and mixed model P value displayed.Representative images of whole mammary tumors (E), and H&E staining (F) of mammary tumor sections from end-stage mammary tumors. H&E staining depicts agrade 1 tumor from an Ad-GFP–treated transplant and a grade 3 tumor from an Ad-Cre-GFP–treated transplant. G, The bar graph represents the percent oftumors that were graded as in situ, grade 1, 2, or 3. Ift88fl/fl/PyMTþ (Ad-GFP, n ¼ 9; Ad-Cre-GFP, n ¼ 9) and Ift20fl/fl/PyMTþ (Ad-GFP, n ¼ 6; Ad-Cre-GFP,n ¼ 6).






tumors (Fig. 1G). However, Cilia�/PyMTþ tumors had a higherfrequency of high-grade tumors compared with Ciliaþ/PyMTþ

tumors (Fig. 1G). Taken together, these data suggest that inhibi-tion of ciliogenesis is an important step in the progression tomoreaggressive, invasive disease.

The PyMT oncoprotein is known to constitutively activate theRAS/MAPK and PI3K/AKT pathways (17). We validated activa-tion of these pathways in Ciliaþ/PyMTþ, and Cilia�/PyMTþ

derived tumors by Western blotting for p-ERK1/2 and p-AKTusing tumor lysates (Fig. 2A; Supplementary Fig. S5A). Weobserved that inhibition of ciliogenesis did not increase theamount of p-ERK1/2 or p-AKT in the presence of PyMT. Asmentioned, PyMT results in constitutive activation of thesepathways; therefore, it is not surprising to find that inhibitionof ciliogenesis does not increase signaling of PI3K/AKT orRAS/MAPK in this model.

The known role of primary cilia in Hedgehog signalingprompted us to investigate Hedgehog signaling in our Ciliaþ/PyMTþ and Cilia�/PyMTþmodel. Primary cilia inhibit Hedgehogpathway activity via processing of GLI transcription factors intotheir repressor forms (GLIR; refs. 4, 18, 19). Phosphorylation ofGLI3 by protein kinase A (PKA) and the partial proteolyticcleavage via the bTrCP/Cul1 complex is required for endogenousGLI3 processing into GLI3R (4). We found the protein level of theproteolytically cleaved GLI3R was decreased when ciliogenesiswas inhibited in PYMTþ mammary epithelial cells (MEC; Fig. 2B;Supplementary Figs. S4 and S5B). These data demonstrate that

inhibition of ciliogenesis decreases processing of GLI3 into itsrepressor form in the PyMT model of breast cancer.

Next, we tested the hypothesis that inhibition of ciliogenesis,anddecreased expressionofGLI3R, results in increased expressionof Hedgehog target genes in the PyMT breast cancer model. Geneexpression of downstream Hedgehog target genes was analyzedby qPCR from RNA isolated from whole tumors (in vivo) fromCiliaþ/PyMTþ, and Cilia�/PyMTþ derived tumors (analysis ofboth IFT88/PyMTþ and IFT20/PyMTþ tumors). Specifically, rel-ative gene expression of Gli1, Gli2, Ptch1, and the promigratorygene Slugwere analyzed. We observed that inhibition of ciliogen-esis resulted in an increase in expression of Gli1, Ptch1, Gli2, andSlug (Fig. 2C). While the increased transcription ofGli1 and Ptch1was not statistically significant, there was a trend towards anincrease in expression when ciliogenesis was inhibited (P ¼0.08 and P ¼ 0.23). Immunostaining of tumor tissues for expres-sion of SLUG protein confirmed that expression of SLUG wasincreased when ciliogenesis was inhibited (Fig. 2D). We have notfound antibodies that work for IHC with specificity for murineGLI1, GLI2, and PTCH1 protein.

Given that breast cancer cells do not have the primary cilianeeded for activation of the Hedgehog pathway via Hedgehogligands or SMO activation, we hypothesized that loss of GLIRleaves Hedgehog target genes vulnerable to activation by anotherpathway (noncanonical activation). To determinewhich pathwaymay be responsible for activation of Hedgehog-target genes inhuman breast cancer we utilized a bioinformatic approach. PyMT

Figure 2.

Inhibition of ciliogenesis decreases production of GLIR and increases Hedgehog signaling in PyMT-positive tumors. Western blot of lysates isolated from Ift88fl/fl/PyMTþ or Ift20fl/fl/PyMTþ mammary tissue treated with Ad-Cre-GFP or Ad-GFP for phospho-ERK1/2 (p-ERK1/2) or phospho-AKT (p-AKT; representativeblot from three mice; A) and cell lysates from primary mammary epithelial cells from Ift88fl/fl/PyMTþ or Ift20fl/fl/PyMTþ glands treated with Ad-Cre-GFP orAd-GFP blotted for GLI3R protein levels normalized to b-actin levels (B). C, Bar graph represents the combined analysis of quantitative real-time PCR of Hedgehog-target genes (Gli1, Ptch1, Gli2, Slug) of RNA isolated from Ift88fl/fl/PyMTþ and Ift20fl/fl/PyMTþ mammary tumors. Number of tumors analyzed for each geneare: Gli1 (Ift88fl/fl/PyMT, n¼ 7; Ift20fl/fl/PyMTþ, n¼ 5); Ptch1 (Ift88fl/fl/PyMT, n¼ 7; Ift20fl/fl/PyMTþ, n¼ 3); Gli2 (Ift88fl/fl/PyMT, n¼ 7; Ift20fl/fl/PyMTþ, n¼ 3); Slug(Ift88fl/fl/PyMT, n ¼ 5; Ift20fl/fl/PyMTþ, n ¼ 5). Average fold change compared to Ad-GFP is plotted. P values were determined by the Student t test (paired,two-tailed); �, P < 0.05. Error bars represent SE. D, Representative images of IHC staining of tumor tissues for SLUG. Arrows point to SLUG-positive nuclei.

Hassounah et al.





is amembrane scaffold protein that is a valuablemodel of humanbreast cancer because it functions like constitutively activatedreceptor tyrosine kinases (RTKs; e.g., EGFR and ERBB2) to activatethe PI3K/AKT and RAS/MAPK pathways. Overactivation of RTKs,including EGFR and ERBB2, are observed frequently in humanbreast cancer (20). To determine whether the PI3K/AKT andRAS/MAPK pathways are associated with Hedgehog target geneexpression in human breast cancers, we analyzed these signalingpathways in a published breast cancer microarray dataset (21).First, we identified patients with high levels of expression ofHedgehog target genes as well as patients with high levels ofexpression of genes associated with PI3K/AKT and/or RAS/MAPKsignaling (Fig. 3A). To accomplish this, we utilized gene expres-sion signatures that represent activation of signaling of the

Hedgehog, PI3K/AKT, and RAS/MAPK pathways (SupplementaryTable S3). The gene signatures are based on comprehensivetranscriptional outputs that have been validated by others andprovide a tool to identify patients with increased signaling in therespective pathways (22–24). Analysis showed a positive corre-lation between patients with activated Hedgehog signaling andpatients with activated PI3K/AKT (Fig. 3A). Using this bioinfor-matics approach, we determined that 39 of 295 breast cancerpatients (13%) have both high Hedgehog and PI3K signaling. Ofthe 39 patients with high Hedgehog and PI3K/AKT signatures, 16are also positive for the RAS/MAPK signature. To determine whatRTK may be responsible for PI3K/AKT and RAS/MAPK signaling,we analyzed the 39 patients with high Hedgehog and PI3K/AKTsignaling (Hedgehogþ/PI3Kþ) to determine which of the known

Figure 3.

Inhibition of ciliogenesis resulted indecreased processing of GLI3R andinduction by TGFA for expression ofHedgehog target genes. A, Scatterplot represents the analysis of geneexpression signature for PI3K/AKTand Hedgehog pathways. Each circlerepresents a single patient of the 295breast cancer patients analyzed.Thirteen percent of breast cancershave transcriptional programsconsistent with high PI3K/AKT andhigh Hedgehog signaling (box, upperright quadrant; Pearson Correlationanalysis). Dashed box represents the39 patients positive for both theHedgehog and PI3K/AKT signatures.B, Bar graph represents quantitationof percent ciliated BT-549 cellstreated with 0.5% serum or 10% FBSfor 24 hours. Bar graphs representaverage fold change (as comparedwith untreated cells) in expression ofGli1 based on quantitative real time-PCR analysis of BT-549, NIH-3T3, andcultured mammary epithelial cellstreated with C, SAG (fold change from–SAG; black bars –SAG, white barsþSAG;), D, and BT-549, NIH-3T3treatedwith 0.5% serum, 10% serumor0.5% serumþ TGFA (fold change fromBT-549 0.5% serum) and 10% FBS or10% serum þ forskolin (fold changefrom 10% serum alone). E, Westernblot analysis for GLI3R protein levelsnormalized to b-actin levels in BT-549cells treated with 10% serum, 10%serum þ forskolin, 0.5% serum or0.5% serum þ forskolin. F, Bar graphof average fold change for analysisof Gli1 expression based onquantitative real-time PCR analysisof BT-549 cells. Error bars, SE; P valueswere determined by the Studentt test (paired, two-tailed);� , P < 0.05. All experimentsperformed in duplicate (B–F).






RTKs and their ligands are coexpressed. We observed that expres-sion of EGFR, as well as the EGFR ligand TGFA, was the most fre-quently expressed in this subset of patients (80% of Hedgehogþ/PI3Kþ patients; Fig. 3A). Further analysis of the 39 patients withhigh Hedgehog and PI3K/AKT signaling revealed that thesepatients are negative for expression of ER and ERBB2 indicatingthat these patients are among the triple-negative breast cancer sub-type (TNBC). This bioinformatics data indicates that Hedgehogsignaling is active in a subset of breast cancers and further supportspublished findings by others that Hedgehog signaling is increas-ed in patients with TNBC (25, 26). This bioinformatics analysisalso supports the use of the PyMTmodel toprovide insight into theinteraction/crosstalk between the PI3K/AKT and RAS/MAPK sig-naling pathways with that of the Hedgehog signaling pathway.

To investigate whether Hedgehog target genes can be inducedby TGFA in the absence of cilia, we utilized the human TNBC cellline, BT-549. We analyzed expression of primary cilia in BT-549cells and found that expression of primary cilia on these cells israre (Fig. 3B) indicating that ciliogenesis is inhibited in BT-549cells. To investigate whether Hedgehog signaling can be inducedin these cells via the canonical pathway (Hedgehog ligand), wetreated BT-549 cells with a potent Hedgehog receptor agonist,Smo agonist (SAG) and performed qPCR to examine the tran-scriptional expression of Gli1 as a reporter of activated Hedgehogsignaling. As a positive control, we also treated normal, Ciliaþ

cells (NIH-3T3)with SAG. As expected, SAG induces expression ofGli1 in NIH-3T3 cells (fold change ¼ 234.5; Fig. 3C). In contrast,our results demonstrated that treatment of BT-549 cells with SAGdid not increase expression of Gli1 (fold change ¼ 1; Fig. 3C). Inaddition, Ift88�/PyMTþ MECs did not respond to SAG as com-pared with Ift88þ/PyMTþ MECs (fold change ¼ 13.7; Fig. 3C).These results indicate that Hedgehog signaling is not activated incilia minus BT-549 and MEC cells via SMO activation.

Next,we treatedBT-549 cellswith 10%serumorwith TGFAandobserved a statistically significant increase inGli1 expression (Fig.3D). In contrast, TGFA did not induce Gli1 expression in Ciliaþ

NIH-3T3 cells (data not shown). It was previously reported thatTNF- induces Gli1 expression through the mTOR/S6K1 pathwayin esophageal adenocarcinoma cells (27).However, TNFadidnotinduce expressionofGli1 in theBT-549 cells (data not shown).Wealso did notfindhigh gene expression of TNFa in theHedgehogþ/PI3Kþ breast cancer patients in our bioinformatics analysis,suggesting that TNFa is irrelevant toHedgehog signaling in TNBC.The observation that TGFA inducesGli1 expression in BT-549 cellsis consistent with our hypothesis and in vivo data suggesting thatactivation of PI3K and/or RAS signaling induces expression ofHedgehog target genes when ciliogenesis is inhibited.

As discussed, phosphorylation of GLI3 by PKA and its subse-quent proteolytic processing is enhanced by localization of theinvolved proteins to the primary cilium. Consistent with ourfindings in the Cilia�/PyMTþ mouse model (Fig. 2B), GLI3Rlevels were also found to be low in the Cilia� BT-549 cells (Fig.3E; Supplementary Fig. S6). Forskolin is known to activate PKAbyincreasing cAMP levels. We treated BT-549 cells with Forskolin todetermine whether this can induce processing of GLI3 into GLI3Rand inhibit expression ofHedgehog target genes. Treatment of BT-549 cells with Forskolin resulted in increased levels of GLI3Rprotein (Fig. 3E). Treatment of BT-549 cells with Forskolin alsoinhibited induction of Gli1 expression (Fig. 3F). These datasupport the hypothesis that loss of GLIR leaves Hedgehog-targetgenes vulnerable to noncanonical activation.

The increased expression of the Hedgehog target gene Slug,which has been demonstrated to be a promigratory gene (28),prompted us to further investigate whether inhibiting ciliogenesisresults in increased metastasis. Ift88�/PyMTþ and Ift88þ/PyMTþ

MECs were injected into cleared fat pads of recipient mice. Wethen followed mice for 20 weeks and harvested the lungs fromtumor-bearing mice to measure lung metastases. The number ofmetastatic lesions was quantitated and found to bemore frequentin lungs from mice with Cilia�/PyMTþ tumors (Fig. 4A and B).These data indicate that inhibiting ciliogenesis in the PyMTmodelresults in a more metastatic phenotype.

DiscussionWe and others have documented that ciliogenesis is inhibited

in breast cancer; however, its biological consequence is notunderstood. The findings in this report demonstrate for the firsttime that inhibition of ciliogenesis promotes breast cancer pro-gression to amore aggressive andmetastatic disease. In support ofthe functional importance of inhibition of ciliogenesis in humanbreast cancer, we found that inhibition of ciliogenesis resulted inincreased Hedgehog target gene expression.

Figure 4.

Inhibition of ciliogenesis increases metastasis. A, Representative images ofwhole mount and H&E staining of lungs from Ift88fl/fl/PyMTþ (Ad-GFP, n ¼ 3;Ad-Cre-GFP, n ¼ 3) tumor-bearing mice. Arrowhead points to examples oftumor nodules; dashed outline highlights the whole mount tumor nodules.B, Bar graph represents quantitation of tumor nodules that metastasized fromthe mammary gland to the lung. Error bars indicate SE; P values weredetermined by the Student t test (paired, two-tailed); � , P < 0.05. C, Modelrepresenting the synergy by which inhibition of ciliogenesis cooperates withRTK signaling to activate expression of Hedgehog target genes.

Hassounah et al.





While it is well established that Hedgehog signaling plays arole in the progression of breast cancers to a more aggressive,metastatic disease, and is associated with TNBC and decreasedsurvival, it is not fully understood how breast cancers activateHedgehog target gene expression in the absence of cilia.Indeed, consistent with findings in pancreatic, prostate, andesophageal cancer (29–31), we showed that breast cancer cellsdo not activate Hedgehog target gene expression through thecanonical pathway, demonstrating that Hedgehog target geneexpression was not dependent on SMO activity. Using agenetically engineered mouse model of breast cancer, a humanbreast cancer cell line, pharmacologic treatment and interro-gation of patient datasets, we provide data that suggests apossible mechanism for how breast cancers activate Hedgehogtarget gene expression (Model, Fig. 4C). We have identifiedthat when ciliogenesis was inhibited, breast cancer cells losethe ability to process the repressor form of the GLI transcrip-tion factor (GLIR). Our data also demonstrated that pharma-cologic rescue of GLI3R expression resulted in decreasedexpression of Hedgehog target genes in breast cancer cells.Our results are consistent with results published from mousemodels of basal cell carcinoma and medulloblastoma (32, 33).In these studies, the authors demonstrated that exogenousexpression of a constitutively activated form of GLIA resultedin increased tumorigenesis when ciliogenesis was inhibited.They further demonstrated that GLIR expression was decreasedwhen ciliogenesis was inhibited and concluded that the offsetbalance between GLIA and GLIR was responsible for increasedexpression of Hedgehog target genes and the observed increasein tumorigenesis. Further studies are needed to determinewhether GLIA is required for activation of Hedgehog targetgenes in breast cancer.

In this study, we further investigated the mechanism ofactivation of noncanonical Hedgehog signaling in breast can-cer. Unlike the cilia/Hedgehog studies modeled in basal cellcarcinoma and medulloblastoma, expression of a mutant,constitutively active form of GLI has not been reported inhuman breast cancer and is therefore unlikely to account forincreased Hedgehog signaling in breast cancer. Our bioinfor-matic analysis revealed that the PI3K and RAS signaling path-ways, as well as their upstream activators EGFR and TGFA, arecoexpressed in Hedgehog-positive human breast cancers. Onthe basis of our genetic data and its relevance in humandisease, we hypothesized that PI3K and RAS signaling providethe necessary cross-talk for increased expression of Hedgehogtarget genes. We utilized cilia-negative human breast cancercells and demonstrated that TGFA induced Hedgehog targetgene expression. Importantly, TGFA could not induce Hedge-hog target gene expression in cilia-positive cells. This findingfurther supports the biological importance of inhibition ofciliogenesis in breast cancer. What is not yet clear is if theTGFA-induced expression of Hedgehog target genes is depen-dent on activation of GLI1 or GLI2. However, our results doindicate that Gli2 expression was increased in the Cilia�/PyMTþ tumors. It is widely accepted in the literature that GLI2require posttranslational modification to become activatedand that in normal cells cilia are essential for this posttrans-lational modification to occur (18, 19). It is possible that TGFAactivates the PI3K and/or RAS pathway and this, in turn, leadsto posttranslational modification of GLI into its activatedform. Indeed, treatment of breast cancer cell lines with specific

chemical and genetic inhibitors of the PI3K/AKT pathwayresulted in a significant decrease in Hedgehog signaling(34). A second article demonstrated that in esophageal ade-nocarcinoma, the PI3K/AKT pathway promoted the posttrans-lational processing (phosphorylation) of GLI transcriptionfactors into their activated form and that this was dependenton mTOR/S6K1 activation (27). Our findings facilitate a betterunderstanding of noncanonical activation of Hedgehog sig-naling in breast cancer by demonstrating for the first time theimportance of inhibition of ciliogenesis for Hedgehog targetgene expression. A deeper exploration of the mechanism bywhich TGFA induces Hedgehog target gene expression in cilia-negative cells is ongoing.

As discussed, we provide evidence that inhibition of ciliogen-esis resulted in noncanonical activation of Hedgehog signaling inbreast cancer. The observed increased expression of Hedgehogtarget genes in breast cancer cells was not dependent on theclassical expression of Hedgehog ligands or the activation ofSMO. This is an important finding because the majority ofHedgehog pathway inhibitors that have been designed and thatare being used clinically target canonical Hedgehog signalingthrough inhibition of Hedgehog ligands and SMO (3). As dis-cussed, recent literature supports the conclusion that targetingcanonical Hedgehog signaling is ineffective in the treatment ofHedgehog-positive breast cancer cells. Understanding themechanisms by which inhibition of ciliogenesis cooperates withPI3K and/or RAS pathways to activate noncanonical Hedgehogsignaling will provide novel treatment strategies. Pharmacolog-ically reexpressing primary cilia may be a novel approach toinhibition of the Hedgehog pathway in breast cancers. Indeed, invitro studies have shown that pharmacologic inhibition of theAurora A Kinase/HDAC6 cascade can result in reexpression ofcilia in cholangiocarcinoma and ovarian cancer cell lines (35,36). Others have also demonstrated that ciliogenesis can beturned on in breast cancer cell lines by knockdown of Nek2 orKif24 (37). Our results also invite speculation regarding Hedge-hog and PI3K/RAS combination therapy for the treatment ofTNBC and for avoiding drug resistance frequently seen withsingle-agent targeting.

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

Authors' ContributionsConception and design: N.B. Hassounah, K.M. McDermottDevelopment of methodology:N.B. Hassounah, T.A. Bunch, K.M. McDermottAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.):N.B. Hassounah, M. Nunez, C.A. Fordyce, R.B. Nagle,K.M. McDermottAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): N.B. Hassounah, M. Nunez, D.J. Roe, R.B. Nagle,K.M. McDermottWriting, review, and/or revision of themanuscript:N.B. Hassounah, D.J. Roe,K.M. McDermottAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): M. Nunez, C.A. Fordyce, K.M. McDermottStudy supervision: K.M. McDermott

AcknowledgmentsThe authors would like to acknowledge the University of Arizona

Cancer Center Core Facilities including the Tissue Acquisition andCellular/Molecular Analysis Service for tissue processing, and H&Estaining, the Genomics Shared Service for assistance with real-time






RT-PCR and the Experimental Mouse Shared Service for help with invivo studies.

Grant SupportThis work was supported by National Cancer Institute of the NIH award

numbers P30 CA023074 (to K.M. McDermott) and T32 CA009213 (to N.B.Hassounah), National Institute of Child Health and Human Development ofthe NIH award number R00 HD056965 (to K.M. McDermott), AACR Institu-tional Research Grant number 74-001-34-IRG from the American Cancer

Society (to K.M. McDermott), and the Better Than Ever Foundation (to K.M.McDermott).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received January 19, 2017; revised May 26, 2017; accepted June 8, 2017;published OnlineFirst June 13, 2017.

References1. American Cancer Society. Global cancer facts and figures, 2nd edition.

Atlanta, GA: American Cancer Society; 2011.2. McDermott KM, Liu BY, Tisty TD, Pazour GJ. Primary cilia regulate

branchingmorphogenesis duringmammary gland development. Curr Biol2010;20:731–7.

3. HassounahNB, Bunch TA,McDermott KM.Molecular pathways: the role ofprimary cilia in cancer progression and therapeutics with a focus onHedgehog signaling. Clin Cancer Res 2012;18:2429–35.

4. Briscoe J, Therond PP. Themechanisms ofHedgehog signaling and its rolesin development and disease. Nat Rev Mol Cell Biol 2013;14:416–29.

5. Nobutani K, Shimono Y, Yoshida M, Mizutani K, Minami A, Kono S, et al.Absence of primary cilia in cell cycle-arrested human breast cancer cells.Genes Cells 2014;19:141–52.

6. Yuan K, FrolovaN, Xie Y,WangDZ, Cook L, Kwon YJ, et al. Primary cilia aredecreased in breast cancer: analysis of a collection of human breast cancercell lines and tissues. J Histochem Cytochem 2010;58:857–70.

7. Menzl I, LeBeau L, Pandey R, Hassounah NB, Li FW, Nagle R, et al. Loss ofprimary cilia occurs early in breast cancer development. Cilia 2014;3:7.

8. Haycraft CJ, Zhang Q, Song B, Jackson WS, Detloff PJ, Serra R, et al.Intraflagellar transport is essential for endochondral bone formation.Development 2007;134:307–16.

9. Lin F, Hiesberger T, Cordes K, Sinclair AM, Goldstein LS, Somlo S, et al.Kidney-specific inactivation of the KIF3A subunit of kinesin-II inhibitsrenal ciliogenesis andproduces polycystic kidney disease. ProcNat Acad SciU S A 2003;100:5286–91.

10. Prater M, Shehata M, Watson CJ, Stingl J. Enzymatic dissociation, flowcytometric analysis, and culture of normal mouse mammary tissue. Meth-ods Mol Biol 2013;946:395–409.

11. Liu X, Ory V, Chapman S, Yuan H, Albanese C, Kallakury B, et al. ROCKinhibitor and feeder cells induce the conditional reprogramming of epi-thelial cells. Am J Pathol 2012;180:599–607.

12. Hassounah NB, Nagle R, Saboda K, Roe DJ, Dalkin BL, McDermott KM.Primary cilia are lost in preinvasive and invasive prostate cancer. PLoS One2013;8:e68521.

13. Marszalek JR, Ruiz-Lozano P, Roberts E, Chien KR, Goldstein LSB. Situsinversus and embryonic ciliary morphogenesis defects in mouse mutantslacking the KIF3A subunit of kinesin-II. Proc Nat Acad Sci U S A 1999;96:5043–8.

14. Cano DA, Murcia NS, Pazour GJ, Hebrok M. orpk mouse model ofpolycystic kidney disease reveals essential role of primary cilia in pancreatictissue organization. Development 2004;131:3457–67.

15. Mitchell EH, Serra R.Normalmammary development and function inmicewith Ift88 deleted in MMTV- and K14-Cre expressing cells. Cilia 2014;3:4.

16. Lin EY, Jones JG, Li P, ZhuUY,Whitney KD,MullerWJ, et al. Progression tomalignancy in the polyoma middle T oncoprotein mouse breast cancermodel provides a reliable model for human diseases. Am J Pathol2003;163:2113–26.

17. Fluck MM, Schaffhausen BS. Lessons in signaling and tumorigenesis frompolyomavirus middle T antigen. Microbiol Mol Biol Rev 2009;73:542–63.

18. Liu A, Wang B, Niswander LA. Mouse intraflagellar transport proteinsregulate both the activator and repressor functions of Gli transcriptionfactors. Development 2005;132:3103–11.

19. May SR, AshiqueAM,KarlenM,Wang B, Shen Y, Zarbalis K, et al. Loss of theretrograde motor for IFT disrupts localization of Smo to cilia and preventsthe expression of both activator and repressor functions of Gli. Dev Biol2005;287:378–89.

20. Templeton AJ, Diez-Gonzalez L, Ace O, Vera-Badillo F, Seruga B, Jordan J,et al. Prognostic relevance of receptor tyrosine kinase expression in breastcancer: A meta-analysis. Cancer Treat Rev 2014;40:1048–55.

21. Rodriguez-Viciana P, Collins C, Fried M. Polyoma and SV40 proteinsdifferentially regulate PP2A to activate distinct cellular signaling pathwaysinvolved in growth control. Proc Natl Acad Sci U S A 2006;103:19290–5.

22. Loboda A, Nebozhyn M, Klinghoffer R, Frazier J, Chastain M, Arthur W,et al. A gene expression signature of RAS pathway dependence predictsresponse to PI3K and RAS pathway inhibitors and expands the populationof RAS pathway activated tumors. BMC Med Genomics 2010;3:26.

23. Loboda A, Nebozhyn M, Cheng C, Vessey R, Huang P, Dai H, et al.Biomarker discovery: identification of a growth factor gene signature. ClinPharmacol Ther 2009;86:92–6.

24. Shi T,Mazumdar T, Devecchio J, Duan ZH, AgyemanA, AzizM, et al. cDNAmicroarray gene expression profiling of hedgehog signaling pathwayinhibition in human colon cancer cells. PLoS ONE 2010;5:pii: e13054.

25. Tao Y, Mao J, Zhang Q, Li L. Overexpression of Hedgehog signalingmolecules and its involvement in triple-negative breast cancer. Oncol Lett2012;2:995–1001.

26. Im S, Choi HJ, Yoo C, Jung JH, Jeon YW, Suh YJ, et al. Hedgehog relatedprotein expression in breast cancer: gli-2 is associated with poor overallsurvival. Korean J Pathol 2013;47:116–23.

27. Wang Y, Ding Q, Yen CJ, Xia W, Izzo JG, Lang JY, et al. The crosstalk ofmTOR/S6K1 and Hedgehog pathways. Cancer Cell 2012;21:374–87.

28. Chen H, Zhu G, Li Y, Padia RN, Dong Z, Pan ZK, et al. Extracellular signal-regulated kinase signaling pathway regulates breast cancer cell migrationby maintaining slug expression. Cancer Res 2009;69:9228–35.

29. Nolan-Stevaux O, Lau J, Truitt ML, Chu GC, Hebrok M, Fernandez-ZapicoME, et al. GLI1 is regulated through Smoothened-independent mechan-isms in neoplastic pancreatic ducts and mediates PDAC cell survival andtransformation. Genes Dev 2009;23:24–36.

30. Zhang JX, Lipinski R, Shaw A, Gipp J, Bushman W. Lack of demonstrableautocrine hedgehog signaling in human prostate cancer cell lines. J Urol2007;177:1179–85.

31. Zhou J, Zhu G, Huang J, Li L, Du Y, Gao Y, et al. Non-canonical GLI1/2activation by PI3K/AKT signaling in renal cell carcinoma: a novel potentialtherapeutic target. Cancer Lett 2016;370:313–23.

32. Wong SY, Seol AD, So PL, Ermilov AN, Bichakjian CK, Epstein EH Jr, et al.Primary cilia can both mediate and suppress Hedgehog pathway-depen-dent tumorigenesis. Nat Med 2009;15:1055–61.

33. Han YG, KimHJ, Dlugosz AA, EllisonDW, Gilbertson RJ, Alvarez-Buylla A.Dual and opposing roles of primary cilia in medulloblastoma develop-ment. Nat Med 2009;15:1062–5.

34. Ramaswamy B, Lu Y, Teng KY, Nuovo G, Li X, Shapiro CL, et al.Hedgehog signaling is a novel therapeutic target in tamoxifen-resistantbreast cancer aberrantly activated by PI3K/AKT pathway. Cancer Res2012;72:5048–59.

35. Egeberg D, Lethan M, Manguso R, Schneider L, Awan A, Jorgensen T, et al.Primary cilia and aberrant cell signaling in epithelial ovarian cancer. Cilia2012;1:15.

36. Gradilone S, Radtke B, Bogert P, Huang B, Gajdos G, Larusso N. HDAC6inhibition restores ciliary expression and decreases tumor growth. CancerRes 2013;73:2259–70.

37. Kim S, Lee K, Choi JH, Ringstad N, Dynlacht BD. Nek2 activation of Kif24ensures cilium disassembly during the cell cycle. Nat Commun 2015;6:8087.


Hassounah et al.




2017;15:1421-1430. Published OnlineFirst June 13, 2017.Mol Cancer Res Nadia B. Hassounah, Martha Nunez, Colleen Fordyce, et al. Tumorigenesis, and Metastasis in Breast CancerInhibition of Ciliogenesis Promotes Hedgehog Signaling,

Updated version

10.1158/1541-7786.MCR-17-0034doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2017/07/12/1541-7786.MCR-17-0034.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/15/10/1421.full#ref-list-1

This article cites 36 articles, 12 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/15/10/1421.full#related-urls

This article has been cited by 6 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://mcr.aacrjournals.org/content/15/10/1421To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-17-0034

http://mcr.aacrjournals.org/content/suppl/2017/07/12/1541-7786.MCR-17-0034.DC1

http://mcr.aacrjournals.org/content/15/10/1421.full#ref-list-1

http://mcr.aacrjournals.org/content/15/10/1421.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/15/10/1421


inhibition of ciliogenesis promotes hedgehog signaling, … · signal transduction inhibition of...

Documents