M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

available at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/molonc

DNp63a induces quiescence and downregulates the BRCA1 pathway in estrogen receptor-positive luminal breast cancer cell line MCF7 but not in other breast cancer cell lines Ruhul Amina,b, Yuiko Morita-Fujimuraa,c, Hiroshi Tawarayamaa, Kentaro Sembad, Natsuko Chibae, Manabu Fukumotob, Shuntaro Ikawaa,* a

Department of Project Programs, Institute of Development, Aging and Cancer (IDAC), Tohoku University, Sendai, Japan b Department of Pathology, Institute of Development, Aging and Cancer (IDAC), Tohoku University, Sendai, Japan c Frontier Research Institute for Interdisciplinary Sciences (FRIS), Tohoku University, Sendai, Japan d Department of Life Science and Medical Bioscience, Waseda University, Tokyo, Japan e Department of Cancer Biology, Institute of Development, Aging and Cancer (IDAC), Tohoku University, Sendai, Japan

A R T I C L E

I N F O

A B S T R A C T

Article history:

Despite apparent resection of tumors, breast cancer patients often suffer relapse due to

Received 22 September 2015

remnant dormant tumor cells. Although quiescence of cancer stem cells is thought as

Received in revised form

one of the mechanisms regulating dormancy, the mechanism underlying quiescence is un-

10 October 2015

clear. Since DNp63a, an isoform of p51/p63, is crucial in the maintenance of stem cells

Accepted 16 November 2015

within mammary epithelium, we investigated its roles in the regulation of dormancy in

Available online 24 November 2015

normal and malignant breast cells. Inducible expression of DNp63a in MCF7 estrogen re-

Keywords:

progenitor-like properties. Judging from mRNA-microRNA microarray analysis, activation

p63

of bone morphogenetic protein (BMP) signaling and inhibition of Wnt signaling emerged

Breast cancer

as prominent mechanisms underlying DNp63a-dependent induction of quiescence and

Stem cell

acquisition of stemness in MCF7. More interestingly, through Ingenuity Pathway analysis,

Quiescence

we found for the first time that BRCA1 pathway was the most significantly downregulated

ceptor positive (ERþ) luminal breast cancer cells led to quiescence and acquisition of

Dormancy

pathway by DNp63a expression in quiescent MCF7 cells, where miR-205 was a downstream mediator. Furthermore, DNp63a-expressing MCF7 cells exhibited resistance to paclitaxel and doxorubicin. Expression of DNp63a in normal MCF10A basal cells increased proliferation and stemness, but did not affect more aggressive luminal (T47D) and basal (MDA-MB231) cells with p53 mutation. Gene expression datasets analyses suggested that DNp63 expression is associated with relapse-free survival of luminal A/B-type patients, but not of the other subtypes. Our results established a cell type-specific function of DNp63a in

Abbreviations: ER, estrogen receptor; BMP, bone morphogenetic protein; DTC, disseminated tumor cell; CTC, circulating tumor cell; TNBC, triple negative breast cancer; BLBC, basal-like breast cancer; BCSC, breast cancer stem cell; GO, gene ontology; IPA, ingenuity pathway analysis; Dox, doxycycline. * Corresponding author. Department of Project Programs, Institute of Development, Aging and Cancer (IDAC), Tohoku University, 4-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan. Tel.: þ81 22 717 8471; fax: þ81 22 717 8476. E-mail address: [email protected] (S. Ikawa). http://dx.doi.org/10.1016/j.molonc.2015.11.009 1574-7891/ª 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

576

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

induction of quiescence and downregulation of the BRCA1 pathway which suggested a role of DNp63a in the dormancy of luminal breast cancers. ª 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1.

Introduction

Many breast cancer patients suffer relapse years after removal of the primary tumor. Dormancy in disseminated tumor cells (DTCs) presumably critically contributes to relapse (Aguirre-Ghiso, 2007). Quiescence, a reversible, non-dividing state of adult stem cells, is thought to contribute to tumor dormancy and poses therapeutic challenges since conventional therapies mainly target proliferating cells (Aguirre-Ghiso, 2007; Cheung and Rando, 2013). Although the mechanisms of cellular quiescence have been studied in different physiological systems (Li and Clevers, 2010), the mechanisms regulating quiescence in breast cancer cells remain unclear. Mammary tissue undergoes several developmental stages including embryonic, pubertal, pregnancy, lactation, and post-lactation stages and undergoes repetitive cycles of maturation and involution in each pregnancy (Hennighausen and Robinson, 2001). Therefore, it inevitably contains multiple types of stem and progenitor cells, which undergo cycles of quiescence and proliferation upon appropriate hormonal signals including estrogen, progesterone and prolactin. Hence, mammary tissue is susceptible to carcinogenesis. Breast cancer is a heterogeneous disease encompassing several molecular subtypes with specific disease progression mechanisms. Based on gene expression profiling, breast cancers are classified into molecular subtypes including luminal A/B, basal-like, epidermal growth factor receptor 2 (ERBB2/ HER2)-overexpressing, and normal or claudin-low (Perou et al., 2000). Nonetheless, this exhaustive expression analysis failed to separate the clinically important class triplenegative breast cancer (TNBC) lacking estrogen receptor (ER), progesterone receptor (PR), and HER2. Furthermore, additional TNBC subtypes including basal-like (BL1 and BL2), immune modulatory (IM), mesenchymal (M), mesenchymal stem-like (MSL), and luminal androgen receptor (LAR) subtypes have been described, suggesting a heterogeneous nature of TNBCs (Chen et al., 2012; Chiorean et al., 2013; Jezequel et al., 2015; Lehmann et al., 2011). In addition, TNBCs and basal-like breast cancer (BLBC) significantly overlap (Rody et al., 2011) and are the most aggressive types, prone to relapse, metastasis, and early death. Furthermore, the BRCA1/2 genes have been identified as genes predisposing women for breast and ovarian cancers (Welcsh and King, 2001). Germline mutation or loss of BRCA1, which is a crucial regulator of DNA repair pathways, is associated with highly aggressive basal-like breast cancer (Deng and Wang, 2003). Adding more complexity to the classification is the identification of breast cancer stem cells (BCSCs). These cells, which can be a population distinct from tissue stem cells, are critical for cancer management because they are capable of

self-renewal and contribute to therapeutic resistance, recurrence, and metastasis (Al-Hajj et al., 2003; Kreso and Dick, 2014). Two distinct types of BCSCs have been proposed; one consists of epithelial-like BCSCs, which are proliferative and are thought to be localized in primary tumor or macrometastases, and the other are mesenchymal-like BCSCs, which are quiescent and probably occur in circulating tumor cells (CTCs) and bone marrow micrometastases (Liu et al., 2014). However, the existence of these distinct types of BCSCs is still debated and further verification is necessary. Recently, breast cancers have additionally been categorized according to the differentiation status of mammary cells, which tumor cells originate from. In normal mammary epithelial differentiation, stem cells differentiate into bipotential progenitors, which undergo further differentiation into basal/myoepithelial and luminal progenitor cells. Luminal progenitors have been identified as the origin of basal-like and HER2þ breast cancer, whereas claudin-low tumors originate from mammary stem cells (Prat and Perou, 2009). Recent evidence suggests that quiescent stem cells (PKH26þ) in normal mammary gland and breast cancer are positive for p63, a p53 tumor suppressor family member (Pece et al., 2010). p63 (initially named “p51” by us) plays an unprecedented role in various developmental processes and cancer (Osada et al., 1998; Yang et al., 1998). DNp63, an Nterminally truncated isoform, is predominantly expressed in the basal/myoepithelial layer of the mammary gland (Nylander et al., 2002). It has been reported to be involved in regulating mammary stem cell quiescence (Li et al., 2008), mammary epithelial integrity (Carroll et al., 2006), lactation (Forster et al., 2014), BLBC prevention (Buckley et al., 2011), and metastasis suppression (Bergholz et al., 2014). Furthermore, DNp63 is shown as a potential BLBC marker (RibeiroSilva et al., 2005) and prosurvival factor in HER2 tumorigenesis (Yallowitz et al., 2014). Although p63 function has been extensively studied, the role of DNp63a in luminal ERþ breast cancers, in which tumor dormancy is prevalent, is unclear (Zhang et al., 2013). Elucidation of the precise functions of p63 in each type of breast cancer cell line, especially in luminal ERþ breast cancer cells, may give valuable information on breast cancers given the heterogeneity of their origin. Here, we studied the role of p63 in luminal, ERþ breast cancer cells; we inducibly expressed DNp63a in MCF7 luminal, ERþ breast cancer cells to study its roles in proliferation, stemness, and therapeutic response. We used various other cell lines to elucidate p63 function in distinct cell types and to obtain additional information on different breast cancer subtypes. In addition, we analyzed DNp63 expression in public gene expression datasets and correlated our findings with the clinical outcomes of patients with different subtypes of breast cancer.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

2.

Materials and methods

2.1.

Cell culture

Human breast adenocarcinoma cell line, MCF7 was obtained from the Cell Resource Center for Biomedical Research, Tohoku University, Japan. MCF10A, T47D, and MDA-MB231 cells were purchased from American Type Culture Collection (ATCC) (Rockville, MD, USA). These cell lines have been thoroughly tested in the cell bank and ATCC. MCF7 and MDA-MB-231 were grown in RPMI1640 (Wako) supplemented with tetracycline-free 10% Fetal bovine serum (FBS) (Clontech laboratories). The MCF10A, immortalized human mammary epithelial cells were cultured in DMEM/Ham’s-F12 with 5% horse serum, 0.5 mg/ml hydrocortisone (Wako), 10 mg/ml Insulin (Wako) and 20 ng/ml epidermal growth factor (EGF) (Chemicon). T47D cells were grown in DMEM with 10% FBS and 10 mg/ ml insulin. All the cells were grown in the presence of 1% penicillin/streptomycin.

2.2. Cloning, virus production and generation of stable cell lines cDNA encoding human DNp63a was cloned into pRetroXTight-Pur vector by using a standard technique. For retroviral production, the RetroX-TetON-advanced system (Clontech) was used according to the manufacturer’s instructions. To generate doxycycline-inducible stable cell lines, first, MCF7 and MCF10A cells were infected with pRetroX-TetOnAdvanced virus using 10 mg/ml polybrene final concentration. MCF7 and MCF10A cells with stably integrated retroviral vector were selected with 5 mg/ml blasticidin S. pRetroX-TetOnAdvanced stable cell lines were then infected with pRetroXTight-Pur vector encoding DNp63a or TAp63g and selected with 2 mg/ml puromycin for 3 days. Doxycycline (Dox; 2 mg/ ml) was used for induction of DNp63a or TAp63g in all the experiments.

2.3.

2.5. Combined mRNA and microRNA expression profiling Total RNAs prepared using the miRNeasy mini kit (Qiagen) were sent to Toray Industries Inc, Japan for custom microarray analysis using the 3D-Gene Human Oligo chip 25k and Rat miRNA Oligo chip (Toray, Japan).

2.6.

Microarray data analysis

Microarray data analysis was performed using different strategies for cross validation. Gene ontology (GO) analysis was performed using the GENECODIS program (Carmona-Saez et al., 2007). Significant enrichment of biological pathways was analyzed using the MAPPfinder (Doniger et al., 2003) and Ingenuity Pathway Analysis (IPA) programs (Qiagen). MicroRNA-mRNA network analysis was performed using IPA’s microRNA target filter program. Functional annotation clustering was performed using DAVID (The Database for Annotation, Visualization and Integrated Discovery; Huang da et al., 2009).

2.7. Real-time quantitative reverse transcription PCR (qRT-PCR) for miRNA and mRNA Total RNA was isolated using the miRNeasy Mini kit or RNeasy Plus kit (Qiagen). cDNA was synthesized using Superscript III reverse transcriptase (Invitrogen) with Oligo(dT) primer. qRT-PCR was carried out using GoTaq Sybr Green qPCR master mix (Promega) with the CFX96 Touch real-time PCR detection system (BIO-RAD). Expression of mRNAs was normalized using b-actin or GAPDH. For microRNA qRT-PCR, total RNA containing miRNAs was isolated using miRNeasy Mini kit (Qiagen). Mature miRNA cDNAs were synthesized from 500 ng to 1 mg of total RNAs using Superscript III reverse transcriptase (Invitrogen) with miRNA-specific stem-loop RT primer (Chen et al., 2005). The expression level of small RNA U6 (RNU6B) was used as an internal control. The sequences of the primers are listed in Table S1.

Transfection 2.8.

Transient transfection was performed using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s instructions. miR-205 mimic (Dharmacon) was transfected at concentration of 50 nM using Lipofectamine RNAiMAX reagent (Life Technologies) for 48 h. Synthesized doublestranded RNA oligo was used as negative control.

2.4.

577

Mammosphere formation assay

Mammosphere formation assay was performed as described previously (Dontu et al., 2003). Briefly, for primary mammosphere generation, single cells were cultured on ultra-low attachment plates (Corning Costar) with phenol-red free DMEM/Ham’s-F12 (Wako) in the presence of B27 supplement (Gibco), 20 ng/ml EGF, and 20 ng/ml bFGF (BioVision).

Protein isolation and western blotting

Preparation of cell lysates and western blotting were performed as described previously (Udden et al., 2013). The following antibodies were used: anti-GAPDH rabbit polyclonal (GeneTex), anti-b-actin mouse monoclonal (Sigma), anti-p63 mouse monoclonal (clone: 4A4) (Santa Cruz Biotechnology), anti-Ki-67 rabbit monoclonal (clone: SP6) (Spring Bioscience), and anti-E-cadherin rabbit monoclonal (24E10) (Cell Signaling). The dilution and incubation time for the different antibodies are given in Table S2.

2.9.

Cell invasion assay

Cell invasion assays were performed in Corning BioCoat Matrigel Invasion Chamber (Corning) according to the manufacturer’s instructions. Briefly, 48 h after induction of

578

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

DNp63a, equal numbers of cells (5  104) from both Dox and Doxþ conditions suspended in serum-free RPMI medium were seeded into a matrigel-coated transwell chamber (8-mm pore size) and the lower chamber was filled with RPMI medium containing 10% FBS as chemoattractant. After incubation for 36e48 h at 37  C, invaded cells were fixed and stained with methanol and 1% toluidine blue. After rinsing and air-drying of the inserts, invaded cells were observed under an Olympus IX71 microscope and counted.

For the measurement of cell proliferation, 2000 MCF7 and 1000 MCF10A cells seeded into 96-well plates were induced for DNp63a expression for 48 h. Then, the cells were treated with varying concentrations of either the DNA damaging drug, doxorubicin (Sigma) or the anti-mitotic drug paclitaxel (Sigma) for an additional 48 h and analyzed with an MTT (3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) cell growth assay kit (Millipore). Quadruplicate wells were assayed for each condition.

2.10.

2.14.

Cell cycle analysis

Cells were trypsinized, washed with PBS, fixed with 70% icecold ethanol, centrifuged, treated with 10 mg/ml propidium iodide (BioLegend), 100 mg/ml ribonuclease A (SigmaeAldrich) and then analyzed with a Cytomics FC500 (Beckman Coulter) flow cytometer. The percentage of cells in each phase of the cell cycle was determined using FlowJo Version 7.6.5 (TreeStar) with the Dean-Jett-Fox statistical model.

2.11.

Flow cytometry

Mammospheres were collected by centrifugation and dissociated into single cells with trypsin-PBS-EDTA. The cells were stained with Alexa Fluor 647-conjugated anti-human CD24 antibody (clone ML5) and mouse IgG2a k isotype control, and with FITC-conjugated anti-mouse/human CD44 antibody (Clone: IM7) and rat IgG2b k isotype control. For CD49f and EpCAM staining, the cells were stained with Alexa Fluor 647-conjugated anti-human/mouse CD49f antibody (clone G0H3) and rat IgG2a, k isotype control and with FITCconjugated anti-human EpCAM antibody (clone 9C4) and mouse IgG2b k isotype control. Antibodies were purchased from BioLegend and used at 1:40 dilution. Cytometric data collection was done using the Cytomics FC500 (Beckman Coulter) and data analysis was performed by using the CXP system software.

We used The Cancer Genome Atlas (TCGA) breast invasive carcinoma (BRCA) exon-specific RNAseq data to analyze the expression of the p63 (TA and DN isoform) and Ki-67 (MKI67) genes in different subtypes of breast cancer (TCGA, 2012). Normalized exon-specific RNAseq data (TCGA_BRCA_exp_HiSeqV2_exon) were obtained from the UCSC Cancer Genomics Browser (Zhu et al., 2009). This dataset presents the transcription estimates at exon level in RPKM (Reads Per Kilobase of exon model per Million mapped reads) values. Normalized RPKM values of isoform-specific exons were used to determine the expression level of a particular isoform. Correlations between DNp63 and Ki-67 expression in different subtypes of breast cancer were analyzed by calculating the Pearson’s correlation coefficient (r) using the IBM SPSS statistics 20 program. Analysis of mRNA microarray gene expression data from normal and breast cancer patients was performed using the Oncomine database (Rhodes et al., 2004). KaplaneMeier survival analyses on RNAseq data were performed using SPSS. For the microarray data, KaplaneMeier survival analyses were done using Km-plotter, an online survival analysis tool on the basis of microarray data of 1809 breast cancer patients (Gyorffy et al., 2010). Log-rank tests were performed to determine statistical significance. A P-value < 0.05 was considered significant.

2.15. 2.12.

Statistical analysis

BrdU-Ki-67 double immunostaining

Cells were cultured on poly-D-lysine-coated cover glasses in a 4-well tissue culture plate. The cells were treated with 10 ng/ ml 5-bromo-20 -deoxyuridine (BrdU; Abcam) for 3 h, 2% paraformaldehyde (PFA) for 30 min, washed with 1 PBST and blocked with 5% donkey serum. The cells were incubated overnight with rabbit anti-human Ki-67 antibody (1:1000) at 4  C, donkey Cy3-conjugated anti-rabbit-IgG (Jackson ImmunoResearch) secondary antibody (1:500) for 2 h. After a second fixation with 4% PFA and denaturation with 2N HCl, the cells were treated overnight with anti-rat monoclonal antibody to BrdU (1:500) at 4  C and donkey Alexa Fluor 488-conjugated anti-rat-IgG (Invitrogen) secondary antibody (1:500). After washing with PBS, the cell nuclei were stained with 1 mg/ml DAPI for 10 min. Ki-67 and BrdU-positive cells were counted using a Biorevo BZ-9000 fluorescence microscope (Keyence).

2.13.

Clinical dataset analysis

Cell proliferation assay

Cells were counted using the Countess Automated Cell Counter (Invitrogen) using trypan blue dye exclusion method.

The results were presented as the mean  SEM and Student’s t-tests were performed for comparison of different sample groups.

3.

Results

3.1. DNp63a induced quiescence in MCF7, ERþ luminal breast cancer cells To investigate the effects of DNp63a expression on luminal breast cancer cells, we expressed DNp63a in MCF7 cells using a doxycycline-inducible expression system. MCF7 retains luminal epithelial character and is used as a model system for its estrogen-responsiveness (Neve et al., 2006). First, we checked the DN and TAp63 isoform transcript levels in MCF7 cells using qRT-PCR and found that DN was the major isoform expressed in these cells (Figure 1A). Doxycycline-inducible expression of DNp63a (Figure 1B) significantly reduced proliferation (Figure 1C and D), which was confirmed by MTT assay (Figure 1E). Cell cycle analysis after 48 h and 96 h expression of

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

579

580

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

DNp63a revealed an increase in G0/G1 cells and a dramatic reduction in S and G2/M cells (Figure 1F). The level of proliferation marker Ki-67 was dramatically reduced at 72 h (Figure 1G). A significant reduction in DNA synthesis in DNp63a-overexpressing cells was observed in the BrdU incorporation assay (Figure 1H). Taken together, these results indicated that DNp63a expression strongly reduced proliferation in MCF7 cells. Furthermore, since the TA transcript was almost undetectable, the effects of DNp63a expression can be ascribed to its intrinsic function, not to its dominantenegative activity toward TAp63. To strengthen the above assumption further, we investigated the effect of the TA isoforms of p63 in MCF7 cells. Induction of TAp63g in MCF7 cells led to dramatic G0/G1 growth arrest, along with an increased sub-G1 population, suggesting its involvement in an apoptotic response (Figure S1A and B). Although DNp63a and TAp63g exhibited similar effect on the proliferation of MCF7 cells, overexpression of the TAp63a isoform in these cells did not have any significant effect on cell proliferation (Figure S2A-D). This result indicated a functional difference between the TAp63g and TAp63a isoforms in MCF7 cells. To replenish loss of differentiated cells, stem cells are capable of re-entering the cell cycle upon physiological demands (Cheung and Rando, 2013). Therefore, stem cells are in non-cycling, G0/quiescent state, which differs from cellular senescence (G0) characterized by eternal growth arrest. To determine the reversibility of growth arrest by DNp63a expression, MCF7 cells were cultivated with Dox for 8 days to sustain DNp63a expression, and then without Dox for 8 days. Prolonged DNp63a expression led to dramatic growth arrest with concomitant morphological alterations (elongation, branching); however, upon removal of Dox, the cells quickly resumed proliferation and returned to their original morphology (Figure 1I), indicating that DNp63a induced quiescence in MCF7 cells.

3.2. DNp63a regulated quiescence-related gene expression in MCF7 cells Given the importance of DNp63a for MCF7 cell status, we performed genome-wide mRNA and microRNA expression profiling of cells incubated either with or without Dox for 48 h. We identified 1178 up- and 1171 downregulated transcripts with more than 2-fold change (Figure 2A). The results were confirmed by qRT-PCR for some genes whose expression was dramatically influenced by DNp63a expression (Figure 2B and C). The ID genes, which are bone morphogenetic protein

(BMP) signaling pathway target genes, were upregulated (ID1; 3.85-fold, ID2; 2.41-fold, and ID3; 11.74-fold), whereas BAMBI, a BMP-pathway inhibitor, was downregulated (0.13fold), indicating that the BMP pathway was activated in DNp63a-expressing MCF7 cells, leading to quiescence and acquisition of stemness. DKK1, an inhibitor of Wnt signaling, was upregulated (3.31-fold), suggesting that DNp63a antagonized Wnt-dependent cell proliferation, leading to quiescence. GO analysis indicated that stem cell quiescencerelated genes involved in cell cycle progression (CCNA2, CCNB1, CCNE2, BIRC5, ANLN, and SGOL1), DNA replication and chromosome segregation (MCM4, PCNA, RRM2, and TOP2A), and RNA processing (DDX39) were downregulated (Figure 2D), most likely as a secondary effect of DNp63adependent cell cycle withdrawal.

3.3. BRCA1-dependent DNA damage response pathway was found to be the most downregulated pathway in DNp63a expressing MCF7 cells through ingenuity pathway analysis To gain further insight into the biological pathways regulated in DNp63a-expressing MCF7 cells, we performed pathway enrichment analysis. IPA revealed that the BRCA1dependent DNA damage response pathway was the pathway most significantly downregulated by DNp63a expression (Figure 2E). In accordance, many DNA damage response and homologous recombination pathway genes downstream of BRCA1 were downregulated (Figure S3). Additionally, hereditary breast cancer signaling, cell cycle control of chromosomal replication, ATM signaling, and estrogen-mediated S-phase entry pathways were downregulated (Figure 2E). In contrast, HER2 signaling in breast cancer, prolactin signaling, HGF signaling, mouse embryonic stem cell pluripotency, and EGF, integrin, and ERBB4 signaling were upregulated. Furthermore, several integrin-mediated signaling regulators were upregulated as indicated by MAPP and GeneCodis analysis (Tables S3 and S4). These results suggested that DNp63a regulated diverse biological pathways associated with normal and breast cancer development and employed various pathways to induce quiescence (Figure 2F). To exclude the possibility that the DNp63a-regulated quiescence gene expression profile was simply due to indirect effects of cell cycle arrest, we compared it with that of cell cycle arrest achieved by overexpression of cyclin-dependent kinase (CDK) inhibitor p21 (CDKN1A). DNp63a by itself was capable of inducing p21 5.36-fold in MCF7 cells (Figure S4A). When p21 was overexpressed in MCF7 cells (Figure S4B), cell

Figure 1 e DNp63a induces quiescence in MCF7 cells. (A) Transcript levels of the DN and TA isoforms of the p63 gene in MCF7 cells. (B) Scheme of the retroviral inducible expression system. DNp63a expression was induced in MCF7 cells using a doxycycline (Dox)-inducible expression system. Increased DNp63a expression was confirmed at 48 h after induction using western blotting. (C) Morphology of the MCF7 cells after induction of DNp63a. Scale bars, 100 mm. (D) The number of cells upon induction of DNp63a in MCF7 cells. 48 h after induction of DNp63a, cells were trypsinized and counted using trypan blue dye exclusion method, *P < 0.05, n [ 3. (E) Viable MCF7 cell count by MTT assay, **P < 0.01, n [ 3. (F) Flow cytometric cell cycle analysis of MCF7 cells at 48 and 96 h after induction. DNp63a induces G0/G1 cell cycle arrest in MCF7 cells. (G) Western blot analysis of Ki-67 protein levels in MCF7 cells at 72 h after induction. (H) Ki-67 and BrdU doubleimmunostaining of MCF7 cells at 48 h after induction (left panels). Fractions of Ki-67D and BrdUD populations at 48 and 96 h after induction of DNp63a. (I) Light microscopic analysis of MCF7 cells at indicated time points after induction. DNp63a induces reversible growth arrest in MCF7 cells.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

581

Figure 2 e DNp63a regulates stem cell quiescence and DNA damage response pathways in MCF7 cells. Microarray gene expression analysis at 48 h after induction of DNp63a expression in MCF7 cells. (A) Pie chart of up- and downregulated genes. (B) Scatter plot showing changes in gene expression by DNp63a upregulation. Some major cancer-related genes are indicated. (C) qRT-PCR analysis of major cancer-related genes. Expression was normalized to GAPDH or b-ACTIN, *P < 0.05, **P < 0.01, n [ 3. (D) Quiescence-related genes are enriched after induction of DNp63a. (E) Top canonical pathways enriched in response to DNp63a analyzed by using the Ingenuity Pathway Analysis (IPA) program. (F) Schematic model of the DNp63a action in MCF7 cells. proliferation decreased (Figure S4C and S4D) and cells were arrested at G0/G1 as determined by cell cycle analysis (Figure S4E). Nevertheless, qRT-PCR analysis revealed that p21 overexpression did not induce gene expression changes

similar to those induced by DNp63a (Figure S4F). These results indicated that DNp63a-induced quiescence gene expression was not due to the cell cycle arrest-inducing property of DNp63a.

582

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

3.4. microRNA network in DNp63a-expressing MCF7 cells favored downregulation of the canonical BRCA1 pathway and induction of quiescence through ingenuity pathway analysis To evaluate the involvement of microRNAs (Bartel, 2004), integrated microRNA-mRNA expression analysis was conducted after induction of DNp63a. We identified 298 up- and 106 downregulated microRNAs (at least 1.5-fold) (Figure 3A). miR-205 was the most highly upregulated (Figure 3B), which was confirmed by quantitative stem-loop RT-PCR detecting mature miR-205 in DNp63a-expressing cells (21.5-fold) (Figure 3C). Transfection of miR-205 mimic into MCF7 cells decreased cell proliferation (Figure 3D and E), and reduced the number of cells in S phase and increased those in G0/G1 phase (Figure 3F). Furthermore, the number of Ki-67-positive cells was dramatically reduced by miR-205 mimic transfection, suggesting that miR-205 is capable of inducing cells into G0 phase (Figure 3G). qRT-PCR analysis confirmed that BRCA1 and BAMBI were downregulated by miR-205 expression (Figure 3H) whereas it upregulated DKK1, ID1, and ID3 expression. These results suggested that miR-205-mediated gene expression changes were similar to the effect of DNp63a in MCF7 cells, indicating that miR-205 indeed acts as a downstream mediator of DNp63a in the induction of quiescence in MCF7 cells. Moreover, using IPA’s microRNA Target Filter, numerous potential miR-205 targets were identified by searching for downregulated mRNAs that contained miR205-binding site in the 3’-untranslated region (Figure S5), which included BRCA1 and BAMBI. The miR-205 predicted targets were mainly involved in chromosome segregation and condensation, and DNA replication (Table S5), implying that these pathways were downregulated by DNp63a expression, inducing quiescence. Integrated network analysis of microRNAs and mRNAs revealed a highly interconnected network of DNp63a-regulated pathways with key node microRNA families (Figure S6). Interestingly, many of these microRNA targets are involved in cell cycle control or quiescence (Ki-67/MKI67, p21/CDKN1A, CCNA2, MCM10, BRCA1, and SMAD3).

3.5. cells

583

MCF10A and MCF7 cells. We found that expression of DNp63 is higher in MCF10A than in MCF7 cells; high-level DNp63 expression is a common characteristic of basal/myoepithelial cells (Figure 4B). However, protein of DNp63 was undetectable in MCF10A cells as indicated by western blot analysis (Figure 4C). This might be explained by the use of earlypassage MCF10A cells; similar observations have been reported previously (Chua et al., 2007). DNp63a expression in MCF10A (Figure 5C) led to slightly increased cell proliferation (Figure 4D, E and F) and a significant increase in both S and G2/M cell fractions (Figure 4G). Next, we investigated whether induction of DNp63a in MCF10A cells has any effect on the genes regulated in quiescence of MCF7 cells. Induction of DNp63a in MCF10A cells did not have any significant effect on the expression of BRCA1, BAMBI, ID3, and miR-205 (Figure 4H). However, the expression levels of DKK1 and ID1 were decreased by DNp63a in these cells. This result suggested that DNp63a acts on DKK1 and ID1 in a context-dependent manner to regulate proliferation of these cells. Further, we investigated the function of the TA isoform of the p63 gene in MCF10A cells. Interestingly, in contrast to the DNp63a effect, induction of TAp63g decreased the proliferation of MCF10A cells (Figure S7A, B, and C). Cell cycle analysis revealed that TAp63g induced G0/G1 cell cycle arrest in these cells (Figure S7D). In addition, induction of TAp63g in MCF10A cells upregulated miR-205 and downregulated the BRCA1 transcript level (Figure S7E). Although miR-205 is upregulated by TAp63g in MCF10A cells, overexpression of miR-205 did not exert any significant effect on proliferation (Figure S8), suggesting that the effect of TAp63g in growth arrest is independent of miR-205. The expression levels of ID1/2/3 were also slightly, albeit insignificantly, increased. Like DNp63a, TAp63g decreased the expression of DKK1 in MCF10A cells. These results implied that TAp63g might induce quiescence in MCF10A cells by regulating genes similar to those regulated by DNp63a in quiescence of MCF7 cells. However, overexpression of TAp63a in MCF10A cells did not exert any significant effect on cell proliferation (Figure S9). Thus, p63 might regulate quiescence in cellular context- and isoform-dependent manners.

DNp63a increased proliferation in basal MCF10A

Next, we examined the effect of DNp63a in MCF10A, a frequently used non-transformed mammary epithelial cell line that sustains basal epithelial gene expression (Neve et al., 2006). First, we checked the transcript levels of the DN and TA isoforms of the p63 gene in MCF10A cells and found that DNp63 was the major isoform expressed (Figure 4A). Then, we compared the expression level of DNp63 in

3.6. DNp63a likely promoted stem cell or progenitor properties, depending on the cell type Accumulating evidence showing that p63 acts as a stem cell regulator in various tissues led us to investigate the effects of DNp63a expression on stemness by the in vitro mammosphere culture assay (Dontu et al., 2003). MCF7 and MCF10A cells were cultivated with or without Dox under conditions suitable for mammosphere formation for 5 days. The numbers

Figure 3 e MicroRNA-regulated networks are induced by DNp63a in quiescent MCF7 cells. Microarray microRNA expression analysis at 48 h after induction of DNp63a expression in MCF7 cells. (A) Pie chart of differentially regulated microRNAs. (B) Scatter plot showing that miR-205 is the most upregulated microRNA by DNp63a upregulation. (C) microRNA-specific stem-loop RT-PCR of the most highly upregulated microRNA (miR-205). Expression was normalized to RNU6B, **P < 0.01, n [ 3. (D) Micrograph of MCF7 cells 48 h after overexpression of miR-205 mimic, NC [ negative control, scale bar [ 100 mm. (E) Number of MCF7 cells decreased after overexpression of miR-205 for 48 h (F) miR-205-induced G0/G1 cell cycle arrest at 48 h, ***P < 0.001, n [ 3. (G) Ki-67D MCF7 cells are decreased 48 h after expression of miR-205, ***P < 0.001, n [ 4. (H) qRT-PCR analysis of quiescence-associated genes after expression of miR-205 for 48 h, *P < 0.05, **P < 0.01, n [ 3, NS [ non-significant.

584 M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

Figure 4 e DNp63a increases proliferation in MCF10A cells. (A) Transcript level of DN and TA isoforms of p63 in MCF10A cells. (B) Comparison of DNp63 transcript level between MCF10A and MCF7 cells. (C) Western blot showing doxycycline-induced expression of DNp63a in MCF10A after 48 h. (D) Micrographs of the cells 48 h after induction of DNp63a. Scale bar, 100 mm. (E) Number of MCF10A cells after induction of DNp63a, NS [ Non-significant. (F) MTT assay indicates that cell proliferation is increased by DNp63a expression, **P < 0.01, n [ 3. (G) Flow cytometric analysis of cell populations in MCF10A cells at 48 h after induction of DNp63a expression, *P < 0.05, **P < 0.01, ***P < 0.001, n [ 3. (H) DNp63a induction in MCF10A cells exhibited differential effect on the expression of quiescence-related genes. *P < 0.05, n [ 3, NS [ non-significant.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

of mammospheres were significantly increased in DNp63aexpressing MCF7 and MCF10A cells compared to uninduced cells (Figure 5A and B), indicating that DNp63a might promote acquisition of stem cell or progenitor properties. Indeed, it is generally accepted that non-adherent mammospheres are enriched for stem or progenitor cell populations in the mammosphere assay utilized in this study (Dontu et al., 2003). We used fluorescence-activated cell sorting (FACS) to analyze the expression of the cell surface markers CD49f and EpCAM, which are commonly used to determine the differentiation status of the cells (Lim et al., 2009); CD49fþ/EpCAM, CD49fþ/EpCAMþ, and CD49/EpCAMþ cells represent cell fractions enriched for mammary stem cells, luminal progenitors, and differentiated luminal cells, respectively. DNp63a expression in MCF7 cells induced a dramatic increase in CD49þ/ EpCAMþ cells implying that DNp63a might convert luminal cancer cells into luminal progenitor-like cells (Figure 5C and K). In MCF10A cells, the CD49fþ/EpCAM mammary stem cell-enriched fraction was increased upon DNp63a expression (Figure 5D and K), implying elevated mammary stem-like cells. Interestingly, no significant changes in the surface markers were observed in the more aggressive p53 mutationharboring luminal T47D (Figure 5E) and basal MDA-MB-231 (Figure 5F) cell lines upon DNp63a expression. This might be due to high expression of CD49fþEpCAM (MDA-MB-231) and CD49fþEpCAMþ (T47D) prior to DNp63a induction and/or inhibition of DNp63a activity by mutant p53, which can inhibit p63 function (Adorno et al., 2009; Muller et al., 2009).

3.7. DNp63a induced cancer cell stemness markers but decreased invasiveness in MCF7 cells, and induced both cancer cell stemness marker and invasiveness in MCF10A cells Mammosphere cells were subjected to FACS analysis using CD24 and CD44 surface markers to analyze cancer cell stemness. We observed a dramatic increase in the fraction of CD24/CD44þ cells in DNp63a-expressing MCF7 and MCF10A (Figure 5G and H). Interestingly, mammosphere cells with this surface marker expression profile have been shown to be highly tumorigenic with BCSC-like properties (Al-Hajj et al., 2003). Since cancer cell stemness is often linked to increased invasiveness (Liu et al., 2014), we determined the invasiveness of DNp63a-expressing MCF7 and MCF10A. DNp63a expression severely compromised the invasion ability of MCF7 cells, which showed anti-migratory characteristics of normal luminal progenitor cells (Figure 5I). Consistent with this, E-cadherin expression, which is usually lost during epithelialemesenchymal transition (EMT), was maintained in DNp63a-expressing MCF7 cells (Figure S10). On the other hand, DNp63a expression strongly enhanced invasiveness in basal epithelial MCF10A cells that showed migratory characteristics of normal mammary stem cells (Figure 5J). This result implied that invasiveness was dictated by the nature of the cell of origin. DNp63a expression converted luminal cells to luminal stem/progenitor cells and restored low invasiveness, characteristic for normal luminal progenitor cells, while basal epithelial cells gained mammary stem cell nature and were more invasive. We believe that the observed effects of DNp63a on invasion in MCF7 and MCF10A cells were unlikely

585

due to its effects on cell proliferation. That is to say that the assay reflects the alteration in the invasiveness of the cells provoked by DNp63a expression; firstly, 48 h after induction of DNp63a, equal number of cells from both Dox- and Doxþ condition were seeded onto upper chamber of transwell, secondly, Dox was not administered to induce DNp63a in cells incubated inside transwell which should lesson DNp63amediated effects on proliferation (MCF7 cells should resume proliferation after removal of Dox as shown in Figure 1I), and thirdly, serum-free medium was used in the upper chamber of transwell plate which was primarily used as a chemoattractant decreasing cell proliferation.

3.8. DNp63a conferred differential response to paclitaxel and doxorubicin Because DNp63a-expressing MCF7 cells exhibited downregulation of BRCA1-dependent DNA damage response genes, we examined the effect of DNp63a expression on drug sensitivity. MCF7 and MCF10A cells were exposed to paclitaxel and doxorubicin (a DNA double-strand break-inducing agent). DNp63a expression rendered MCF7 cells more resistant to paclitaxel (Figure 6A). It is reasonable that, since DNp63a-expressing MCF7 cells are quiescent, anti-mitotic activity of paclitaxel may not exert additional anti-mitotic activity. DNp63aexpressing MCF7 cells were resistant to low (0.1e1 mM) but sensitive to high (5 mM) concentrations of doxorubicin (Figure 6A). As DNp63a-expressing MCF7 cells had acquired low proliferative capability due to quiescence, it is possible that further drug treatment did not have any effect in these cells. MCF10A cells became more sensitive to paclitaxel as well as doxorubicin upon DNp63a expression (Figure 6B). These results suggested that the therapeutic response of normal and breast cancer cells depended on expression of DNp63a in a cell type-specific manner.

3.9. High DNp63 expression was a prognosis factor for luminal A/B-type breast cancers, but not for other types To generalize our in vitro findings, we evaluated the possible roles of DNp63a in human breast cancer in vivo by analyzing the expression of both p63 isoforms in different breast cancer subtypes. We first looked into the TCGA breast adenocarcinoma RNAseq dataset to analyze isoform-specific expression of p63 gene. We found that DNp63 was the major isoform expressed in luminal A/B, basal, TNBC, Her2þ and normallike/claudin-low type breast cancer patients (Figure 7AeF). The highest level of DNp63 was observed in normal-like/ claudin-low type breast cancer patients (Figure 7F). Since DNp63a was negatively correlated with Ki-67 expression in MCF7 cells, we analyzed the correlation between the expression of DNp63 and that of Ki-67 in various breast cancer subtypes. When we analyzed the entire RNAseq dataset (n ¼ 1215) combining all subtypes, expression of DNp63 was found to be negatively correlated with that of Ki-67 (Pearson’s r ¼ 0.456, P < 0.001) (Figure 7G). Upon classification of the patients into the subtypes, negative correlation of DNp63 with Ki-67 was observed in luminal (r ¼ 0.204, P < 0.001), TNBC (r ¼ 0.451, P < 0.001), Her2þ (r ¼ 0.508, P < 0.001), and normal-like (r ¼ 0.229, P < 0.05) (Figure 7H, J, K, L), but not

586

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

Figure 5 e DNp63a induces mammary stem or luminal progenitor-like markers in a cell type-specific manner. (A) MCF7 and (B) MCF10A were cultured in mammosphere-inducing conditions with or without doxycycline for 5 days to induce DNp63a. The number of mammospheres was significantly increased upon induction with DNp63a in both cell types, *P < 0.05, n [ 3, scale bar, 100 mm. (C) Flow cytometric analysis of human mammary epithelial markers CD49f and EpCAM in MCF7 cells after inducing DNp63a for 5 days. DNp63a dramatically enriches the

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

in the basal subtype (r ¼ 0.017, P ¼ 0.83) (Figure 7I). The results of the database analyses strongly supported our findings on the negative correlation of DNp63 with Ki-67 in certain breast cancer subtypes. We also analyzed the expression of p63a/b isoforms and Ki-67 in publicly available microarray gene expression datasets of invasive ductal breast carcinoma patients from the Oncomine database. The p63a/b isoforms in these datasets are most likely DN isoforms, since we showed that DNp63 is the major isoform expressed in different breast cancer subtypes (Figure 7). Based on data from three different datasets (TCGA dataset, Curtis dataset, Richardson breast2 dataset), p63a/b expression was lower in invasive ductal breast carcinoma than in normal mammary gland, while Ki67 expression was higher (Figure 8A, B and C), suggesting an inverse association between p63a/b and Ki-67 expression in invasive ductal breast carcinoma. Since the majority of these carcinomas are of the luminal type (Weigelt et al., 2010), this association is in agreement with the inducibility of quiescence by DNp63a in MCF7 luminal cells. A similar association between p63a/b and Ki-67 expression was evident in ERþ and ER breast cancer patients (Figure 8D). Next, we used the KaplaneMeier method to analyze the impact of p63a/b expression on relapse-free survival using public breast cancer microarray datasets. High p63a/b (Figure 8E and F) expression was associated with better relapse-free survival of patients with luminal A/B-type breast cancers, but not others such as basal-type and HER2þ cancers (Figure 8G and H). Furthermore, although the number of ERtumors in luminal A/B type was not large enough for statistical analysis, no difference was observed between high p63a/b expression and ER status (Figure S11). Finally, we analyzed the survival of breast cancer patients with different subtypes from the RNAseq dataset. We found that high expression of DNp63 increased overall survival of luminal A-type breast cancer patients (n ¼ 216, P ¼ 0.04) (Figure S12A). Although similar trends were observed for some other subtypes, these were not statistically significant (Figure S12BeF). Large numbers of patients are required to allow better interpretation of the association of DNp63 with patient survival in other subtypes.

4.

Discussion

4.1. DNp63a induced quiescence only in MCF7 luminal, ERþ breast cancer cells Although accumulating evidence suggests the existence of quiescence in mammary gland stem cells (Harmes and DiRenzo, 2009), studies on quiescence in breast cancer cells

587

are limited. We showed that DNp63a induces reversible growth arrest in MCF7 cells, suggesting cellular quiescence. Microarray analysis revealed upregulation of the BMP signaling target genes ID1/2/3 and the Wnt pathway antagonist DKK1 in DNp63a-expressing MCF7 cells. We speculate that DNp63a-dependent BMP signal enhancement and Wnt signal inhibition led to quiescence and withdrawal from proliferation. Since activation of BMP and inhibition of Wnt pathways have been observed in quiescent adult stem cells (Li and Clevers, 2010; Mira et al., 2010), DNp63a probably employs similar mechanisms in regulating quiescence of MCF7 cells. Interestingly, BMP signaling is also implicated in mouse TNBC cell dormancy (Gao et al., 2012), corroborating our hypothesis. MicroRNA network analysis of DNp63a-expressing MCF7 cells revealed miR-205 as the most highly upregulated miRNA. Although the function of miR-205 in mammary epithelial cells is controversial (Greene et al., 2010), our study suggested that miR-205 is the downstream mediator of DNp63a-induced quiescence in MCF7 cells. Notably, upregulation of miR-483 and downregulation of miR-520 that target Ki67 (Bertero et al., 2011) and p21 (Wu et al., 2010), respectively, were observed in DNp63a-overexpressing MCF7 cells. Downregulation of Ki-67 and upregulation of p21 are generally associated with quiescence. Because DNp63a-expressing MCF7 cells were enriched for stem cell quiescence genes, we conducted a series of stem cell assays. Although DNp63a induced quiescence in adherent MCF7 cells, it increased mammosphere-formation efficiency in those cells characterized as more prematured state. Thus, it is conceivable that the ability of DNp63a to induce quiescence in adherent cells could be counterbalanced by the acquisition of stem/progenitor-like state of these cells under mammosphere conditions. Cell surface marker analysis confirmed that MCF7 luminal cancer cells entered into a luminal progenitor-like and cancer stem-like state upon induction of DNp63a expression, indicating that DNp63a-dependent enhancement of BMP signaling contributed to the acquisition of the stem cell phenotype. Such plasticity of mammary epithelial cells has been observed in several studies in which differentiated luminal cells acquired basal/mammary stem or luminal progenitor properties (Guo et al., 2012; Skibinski et al., 2014; Yalcin-Ozuysal et al., 2010). In contrast, induction of DNp63a increased proliferation and stemness markers in basal MCF10A cells. It is noteworthy that DNp63a enhanced mammosphere formation in MCF10A cells in non-adherent mammosphere culture conditions, but did not induce high-level proliferation in adherent culture conditions. We speculate that DNp63a probably does not allow proliferation of differentiated cells under adherent conditions,

luminal progenitor-like (CD49fDEpCAMD) population in MCF7 cells, **P < 0.01 (n [ 3). (D) DNp63a dramatically enriches the mammary stem-like population (CD49fDEpCAML) in MCF10A cells, **P < 0.01, n [ 3. (E) and (F) DNp63a does not induce stemness in mutant p53containing cell lines T47D and MDA-MB-231, n [ 3, P-value non-significant. Breast cancer stem cell-like markers (CD24LCD44D) are dramatically induced by DNp63a in both MCF7 (G) and MCF10A (h) cells, **P < 0.01, n [ 3. (I) Less invasive MCF7 cells dramatically lose invasion ability by DNp63a after 48 h, ***P < 0.001, n [ 3, scale bar, 100 mm. (J) Invasion ability was drastically increased in DNp63a-induced MCF10A cells after 36 h, ***P < 0.001, n [ 3, scale bar, 100 mm. (K) Normal differentiation hierarchy of mammary epithelial cells. Mammary stem cells differentiate into bipotential progenitor cells, which further differentiate into basal/myoepithelial and luminal progenitor cells. Luminal progenitor cells give rise to mature luminal cells. In our experiment, DNp63a converts MCF7 luminal cancer cells into luminal progenitor-like cells whereas normal basal MCF10A cells enter into mammary stem-like state by DNp63a.

588

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

Figure 6 e Chemotherapeutic response in DNp63a-expressing cells. (A) Cell proliferation rate measured by MTT assay, indicating that DNp63aexpressing MCF7 cells are unresponsive to paclitaxel and doxorubicin (0.1e1 mM) but become more sensitive to high concentration of doxorubicin (5 mM) (ADR). Cell proliferation rate as measured by MTT assay, *P < 0.05, **P < 0.01, ***P < 0.001, n [ 4, NS [ Non-significant. (B) DNp63a-expressing MCF10A cells were slightly more sensitive to paclitaxel at high concentration, but became highly sensitive to doxorubicin, *P < 0.05, **P < 0.01, ***P < 0.001, n [ 4, NS [ non-significant.

but rather only allows proliferation of undifferentiated stem/ progenitor-like cells enriched under mammosphere conditions. DNp63a overexpression did not increase proliferation and mammary stem-like cells in the highly aggressive mutant p53-expressing MDA-MB-231 basal/claudin-low and T47D luminal breast cancer cells. The MDA-MB-231 and T47D cell lines already possessed higher proportions of mammary stem-like and luminal progenitor-like cells, respectively. Since a large fraction of patients with highly aggressive breast cancer harbor mutant p53, which promotes oncogenesis by

antagonizing p63 (Adorno et al., 2009; Muller et al., 2009), we speculate that the ability of DNp63a to regulate stem or progenitor activity might be confined to normal developmental processes and/or less aggressive breast cancers.

4.2. Downregulation of the BRCA1 pathway by DNp63a expression in MCF7 cells Our results, for the first time, demonstrated downregulation of the canonical BRCA1 pathway by DNp63a expression through

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

589

Figure 7 e Correlation of DNp63 expression with Ki-67 expression in different subtypes of breast cancer patients. (AeF) Box-and-whisker plots of DNp63 and TAp63 expression in different subtypes of breast cancer patients on the basis of TCGA RNAseq data. Correlation of DNp63 with Ki67 (G) TCGA whole dataset (n [ 1215), (H) luminal (n [ 628), (I) basal (n [ 142), (J) TNBC (n [ 125), (K) Her2D (n [ 184) and (L) normallike/claudin-low type (n [ 119) breast cancer patients.

Figure 8 e Expression of p63 correlates with patient outcome of luminal A-type breast cancer. Oncomine database analysis of p63 and Ki-67 expression in normal and breast cancer patients. (AeD) Box and whisker plots of p63a/b and Ki-67 expression in various public datasets. (A) p63a/b (Agilent probe: A_32_P114475) and Ki-67 (A_23_P202232) expression in the TCGA dataset. (B) p63a/b (Illumina probe: ILMN_2138801) and Ki-67 (ILMN_1734827) expression in the Curtis dataset. (C) p63a/b (Affymetrix probe: 209863_s_at) and Ki-67 (212022_s_at) expression in the Richardson Breast2 dataset. (D) p63a/b (209863_s_at) and Ki-67 (212022_s_at) expression separated by ER status in the Lu dataset. KaplaneMeier survival analysis was performed using the Km-plotter database with the Affymetrix probe id 209863_s_at for p63a/b. P-values were calculated by using a log-rank test. (E) p63a/b expression in the relapse-free survival of luminal-A type breast cancer patients. (F) p63a/b expression in the relapse-free survival of luminal B-type breast cancer patients. (G) p63a/b expression in the relapse-free survival of basal type breast cancer patients. (H) p63a/b expression in the relapse-free survival of HER2D breast cancer patients.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

Ingenuity Pathway analysis, and the putative role of miR-205 therein. miR-146a/b, which targets BRCA1, was upregulated, and, thus, may contribute to the downregulation of BRCA1 (Garcia et al., 2011). BRCA1, highly expressed in luminal epithelial cells, has been shown to contribute to the commitment of luminal fate (Buckley and Mullan, 2012) and to promote differentiation of mammary stem cells. Furthermore, loss of BRCA1 increased stem or progenitor activity (Liu et al., 2008). These findings are consistent with our observation that in MCF7 cells, DNp63a induces stemness and downregulates BRCA1, which would otherwise maintain the differentiated state. In addition, the necessity of BRCA1 function for luminal cell commitment may explain why BRCA1 mutations are rarely observed in luminal cancers. Interestingly, DNp63 was shown to be a downstream transcriptional target of BRCA1 in basal mammary epithelium (Buckley et al., 2011), implying the possibility of DNp63 and BRCA1 constituting a negative feedback loop mechanism, which needs to be verified in future. Lastly, downregulation of the repair pathway in stem cells may be relevant to the observed reduced DNA damage response in hematopoietic stem cells upon entry into quiescence during aging (Beerman et al., 2014).

4.3. High p63 expression was a prognosis factor for luminal A/B-type breast cancers, but not for other types High p63a/bdmost likely DNp63adexpression indicated better prognosis in luminal A/B-type breast cancer patients, but not in others. Together with our finding that DNp63a-expressing MCF7 cells exhibit a quiescent progenitor-cell phenotype and reduced invasiveness, this suggests that p63-expressing luminal tumors might be less aggressive because of their quiescent nature. Thus, these luminal tumor cells might retain some characters intrinsic to the cells they originated from. Interestingly, DNp63a-expressing MCF7 cells expressed CD24CD44þ BCSC markers, implying that they can serve as cancer stem cells capable of forming relapsed tumors in the long run. This is consistent with the fact that CD24CD44þ cells usually are regarded as quiescent cancer stem cells found in the circulation and bone marrow micrometastases (Liu et al., 2014). p63a/b expression did not influence prognosis of ERþ and ER-tumors, implying that tumor cell origin and ER positiveness have an influence independent from p63 expression. Furthermore, we found no association of p63a/b expression with the survival of patients in highly aggressive basal and HER2þ tumors, which often harbor p53 mutations; this was in agreement with the fact that DNp63a expression had no observable effect in tumor cells with the p53 mutation (Adorno et al., 2009; Muller et al., 2009).

591

DTCs is also associated with tumor/metastasis suppression, DNp63a-induced quiescence might be linked to tumor/metastatic dormancy of luminal breast cancer, implying tumor/ metastasis suppressor function. In contrast, DNp63a increased cell proliferation, invasion, and cancer stem-like cells in normal, basal epithelial MCF10A cells, suggesting oncogenic function. However, overexpression of DNp63a did not affect stemness in more aggressive luminal and basal breast cancer cells with mutant p53, suggesting a contextspecific function of DNp63a. Further investigations are required to support the context-specific role of DNp63a in patients with different breast cancer subtypes.

Accession number Microarray data have been submitted to the NCBI Gene Expression Omnibus (GEO) database under the accession number GSE64955 (URL: http://www.ncbi.nlm.nih.gov/geo/ query/acc.cgi?token¼cxabikwijdwtjed&acc¼GSE64955).

Conflicts of interest The authors declare no conflict of interests.

Author’s contributions RA designed the study, performed the experiments, analyzed the data and wrote the manuscript. YMF, SI, HT analyzed the data and wrote the manuscript. KS, NC, MF provided material and administrative support.

Acknowledgments The authors thank Ms. Hiromi Yoshida for technical assistance. The study was partially supported by the Creative Interdisciplinary Research Division from Frontier Research Institute for Interdisciplinary Sciences (FRIS), Tohoku University, Japan. Ruhul Amin received a fellowship from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.molonc.2015.11.009.

5.

Conclusions

The present study demonstrated that induction of DNp63a in luminal, ERþ MCF7 breast cancer cells promoted cellular quiescence and a progenitor-like state. DNp63a-dependent downregulation of the BRCA1 pathway was demonstrated for the first time. In addition, high p63a/b expression was found to be associated with better prognosis for luminal A/ B-type breast cancer patients. Since prolonged dormancy in

R E F E R E N C E S

Adorno, M., Cordenonsi, M., Montagner, M., Dupont, S., Wong, C., Hann, B., Solari, A., Bobisse, S., Rondina, M.B., Guzzardo, V., Parenti, A.R., Rosato, A., Bicciato, S., Balmain, A., Piccolo, S., 2009. A mutant-p53/Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell 137, 87e98.

592

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

Aguirre-Ghiso, J.A., 2007. Models, mechanisms and clinical evidence for cancer dormancy. Nat. Rev. Cancer 7, 834e846. Al-Hajj, M., Wicha, M.S., Benito-Hernandez, A., Morrison, S.J., Clarke, M.F., 2003. Prospective identification of tumorigenic breast cancer cells. Proc. Natl. Acad. Sci. USA 100, 3983e3988. Bartel, D.P., 2004. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281e297. Beerman, I., Seita, J., Inlay, M.A., Weissman, I.L., Rossi, D.J., 2014. Quiescent hematopoietic stem cells accumulate DNA damage during aging that is repaired upon entry into cell cycle. Cell Stem Cell 15, 37e50. Bergholz, J., Zhang, Y., Wu, J., Meng, L., Walsh, E.M., Rai, A., Sherman, M.Y., Xiao, Z.X., 2014. DeltaNp63alpha regulates Erk signaling via MKP3 to inhibit cancer metastasis. Oncogene 33, 212e224. Bertero, T., Gastaldi, C., Bourget-Ponzio, I., Imbert, V., Loubat, A., Selva, E., Busca, R., Mari, B., Hofman, P., Barbry, P., Meneguzzi, G., Ponzio, G., Rezzonico, R., 2011. miR-483-3p controls proliferation in wounded epithelial cells. FASEB J. 25, 3092e3105. Buckley, N.E., Conlon, S.J., Jirstrom, K., Kay, E.W., Crawford, N.T., O’Grady, A., Sheehan, K., Mc Dade, S.S., Wang, C.W., McCance, D.J., Johnston, P.G., Kennedy, R.D., Harkin, D.P., Mullan, P.B., 2011. The DeltaNp63 proteins are key allies of BRCA1 in the prevention of basal-like breast cancer. Cancer Res. 71, 1933e1944. Buckley, N.E., Mullan, P.B., 2012. BRCA1econductor of the breast stem cell orchestra: the role of BRCA1 in mammary gland development and identification of cell of origin of BRCA1 mutant breast cancer. Stem Cell Rev. 8, 982e993. Carmona-Saez, P., Chagoyen, M., Tirado, F., Carazo, J.M., PascualMontano, A., 2007. GENECODIS: a web-based tool for finding significant concurrent annotations in gene lists. Genome Biol. 8, R3. Carroll, D.K., Carroll, J.S., Leong, C.O., Cheng, F., Brown, M., Mills, A.A., Brugge, J.S., Ellisen, L.W., 2006. p63 regulates an adhesion programme and cell survival in epithelial cells. Nat. Cell Biol. 8, 551e561. Chen, C., Ridzon, D.A., Broomer, A.J., Zhou, Z., Lee, D.H., Nguyen, J.T., Barbisin, M., Xu, N.L., Mahuvakar, V.R., Andersen, M.R., Lao, K.Q., Livak, K.J., Guegler, K.J., 2005. Realtime quantification of microRNAs by stem-loop RT-PCR. Nucleic Acids Res. 33, e179. Chen, X., Li, J., Gray, W.H., Lehmann, B.D., Bauer, J.A., Shyr, Y., Pietenpol, J.A., 2012. TNBCtype: a subtyping tool for triplenegative breast Cancer. Cancer Inform. 11, 147e156. Cheung, T.H., Rando, T.A., 2013. Molecular regulation of stem cell quiescence. Nat. Rev. Mol. Cell Biol. 14, 329e340. Chiorean, R., Braicu, C., Berindan-Neagoe, I., 2013. Another review on triple negative breast cancer. Are we on the right way towards the exit from the labyrinth? Breast 22, 1026e1033. Chua, H.L., Bhat-Nakshatri, P., Clare, S.E., Morimiya, A., Badve, S., Nakshatri, H., 2007. NF-kappaB represses E-cadherin expression and enhances epithelial to mesenchymal transition of mammary epithelial cells: potential involvement of ZEB-1 and ZEB-2. Oncogene 26, 711e724. Deng, C.X., Wang, R.H., 2003. Roles of BRCA1 in DNA damage repair: a link between development and cancer. Hum. Mol. Genet. 12, R113eR123. Spec No 1. Doniger, S.W., Salomonis, N., Dahlquist, K.D., Vranizan, K., Lawlor, S.C., Conklin, B.R., 2003. MAPPFinder: using gene ontology and GenMAPP to create a global gene-expression profile from microarray data. Genome Biol. 4, R7. Dontu, G., Abdallah, W.M., Foley, J.M., Jackson, K.W., Clarke, M.F., Kawamura, M.J., Wicha, M.S., 2003. In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells. Genes Dev. 17, 1253e1270.

Forster, N., Saladi, S.V., van Bragt, M., Sfondouris, M.E., Jones, F.E., Li, Z., Ellisen, L.W., 2014. Basal cell signaling by p63 controls luminal progenitor function and lactation via NRG1. Dev. Cell 28, 147e160. Gao, H., Chakraborty, G., Lee-Lim, A.P., Mo, Q., Decker, M., Vonica, A., Shen, R., Brogi, E., Brivanlou, A.H., Giancotti, F.G., 2012. The BMP inhibitor Coco reactivates breast cancer cells at lung metastatic sites. Cell 150, 764e779. Garcia, A.I., Buisson, M., Bertrand, P., Rimokh, R., Rouleau, E., Lopez, B.S., Lidereau, R., Mikaelian, I., Mazoyer, S., 2011. Down-regulation of BRCA1 expression by miR-146a and miR146b-5p in triple negative sporadic breast cancers. EMBO Mol. Med. 3, 279e290. Greene, S.B., Herschkowitz, J.I., Rosen, J.M., 2010. The ups and downs of miR-205: identifying the roles of miR-205 in mammary gland development and breast cancer. RNA Biol. 7, 300e304. Guo, W., Keckesova, Z., Donaher, J.L., Shibue, T., Tischler, V., Reinhardt, F., Itzkovitz, S., Noske, A., Zurrer-Hardi, U., Bell, G., Tam, W.L., Mani, S.A., van Oudenaarden, A., Weinberg, R.A., 2012. Slug and Sox9 cooperatively determine the mammary stem cell state. Cell 148, 1015e1028. Gyorffy, B., Lanczky, A., Eklund, A.C., Denkert, C., Budczies, J., Li, Q., Szallasi, Z., 2010. An online survival analysis tool to rapidly assess the effect of 22,277 genes on breast cancer prognosis using microarray data of 1,809 patients. Breast Cancer Res. Treat. 123, 725e731. Harmes, D.C., DiRenzo, J., 2009. Cellular quiescence in mammary stem cells and breast tumor stem cells: got testable hypotheses? J. Mammary Gland Biol. Neoplasia 14, 19e27. Hennighausen, L., Robinson, G.W., 2001. Signaling pathways in mammary gland development. Dev. Cell 1, 467e475. Huang da, W., Sherman, B.T., Lempicki, R.A., 2009. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44e57. Jezequel, P., Loussouarn, D., Guerin-Charbonnel, C., Campion, L., Vanier, A., Gouraud, W., Lasla, H., Guette, C., Valo, I., Verriele, V., Campone, M., 2015. Gene-expression molecular subtyping of triple-negative breast cancer tumours: importance of immune response. Breast Cancer Res. 17, 43. Kreso, A., Dick, J.E., 2014. Evolution of the cancer stem cell model. Cell Stem Cell 14, 275e291. Lehmann, B.D., Bauer, J.A., Chen, X., Sanders, M.E., Chakravarthy, A.B., Shyr, Y., Pietenpol, J.A., 2011. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J. Clin. Invest. 121, 2750e2767. Li, L., Clevers, H., 2010. Coexistence of quiescent and active adult stem cells in mammals. Science 327, 542e545. Li, N., Singh, S., Cherukuri, P., Li, H., Yuan, Z., Ellisen, L.W., Wang, B., Robbins, D., DiRenzo, J., 2008. Reciprocal intraepithelial interactions between TP63 and hedgehog signaling regulate quiescence and activation of progenitor elaboration by mammary stem cells. Stem Cells 26, 1253e1264. Lim, E., Vaillant, F., Wu, D., Forrest, N.C., Pal, B., Hart, A.H., Asselin-Labat, M.L., Gyorki, D.E., Ward, T., Partanen, A., Feleppa, F., Huschtscha, L.I., Thorne, H.J., Fox, S.B., Yan, M., French, J.D., Brown, M.A., Smyth, G.K., Visvader, J.E., Lindeman, G.J., 2009. Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat. Med. 15, 907e913. Liu, S., Cong, Y., Wang, D., Sun, Y., Deng, L., Liu, Y., MartinTrevino, R., Shang, L., McDermott, S.P., Landis, M.D., Hong, S., Adams, A., D’Angelo, R., Ginestier, C., Charafe-Jauffret, E., Clouthier, S.G., Birnbaum, D., Wong, S.T., Zhan, M., Chang, J.C., Wicha, M.S., 2014. Breast cancer stem cells transition between epithelial and mesenchymal states

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 5 7 5 e5 9 3

reflective of their normal counterparts. Stem Cell Rep. 2, 78e91. Liu, S., Ginestier, C., Charafe-Jauffret, E., Foco, H., Kleer, C.G., Merajver, S.D., Dontu, G., Wicha, M.S., 2008. BRCA1 regulates human mammary stem/progenitor cell fate. Proc. Natl. Acad. Sci. USA 105, 1680e1685. Mira, H., Andreu, Z., Suh, H., Lie, D.C., Jessberger, S., Consiglio, A., San Emeterio, J., Hortiguela, R., Marques-Torrejon, M.A., Nakashima, K., Colak, D., Gotz, M., Farinas, I., Gage, F.H., 2010. Signaling through BMPR-IA regulates quiescence and longterm activity of neural stem cells in the adult hippocampus. Cell Stem Cell 7, 78e89. Muller, P.A., Caswell, P.T., Doyle, B., Iwanicki, M.P., Tan, E.H., Karim, S., Lukashchuk, N., Gillespie, D.A., Ludwig, R.L., Gosselin, P., Cromer, A., Brugge, J.S., Sansom, O.J., Norman, J.C., Vousden, K.H., 2009. Mutant p53 drives invasion by promoting integrin recycling. Cell 139, 1327e1341. Neve, R.M., Chin, K., Fridlyand, J., Yeh, J., Baehner, F.L., Fevr, T., Clark, L., Bayani, N., Coppe, J.P., Tong, F., Speed, T., Spellman, P.T., DeVries, S., Lapuk, A., Wang, N.J., Kuo, W.L., Stilwell, J.L., Pinkel, D., Albertson, D.G., Waldman, F.M., McCormick, F., Dickson, R.B., Johnson, M.D., Lippman, M., Ethier, S., Gazdar, A., Gray, J.W., 2006. A collection of breast cancer cell lines for the study of functionally distinct cancer subtypes. Cancer Cell 10, 515e527. Nylander, K., Vojtesek, B., Nenutil, R., Lindgren, B., Roos, G., Zhanxiang, W., Sjostrom, B., Dahlqvist, A., Coates, P.J., 2002. Differential expression of p63 isoforms in normal tissues and neoplastic cells. J. Pathol. 198, 417e427. Osada, M., Ohba, M., Kawahara, C., Ishioka, C., Kanamaru, R., Katoh, I., Ikawa, Y., Nimura, Y., Nakagawara, A., Obinata, M., Ikawa, S., 1998. Cloning and functional analysis of human p51, which structurally and functionally resembles p53. Nat. Med. 4, 839e843. Pece, S., Tosoni, D., Confalonieri, S., Mazzarol, G., Vecchi, M., Ronzoni, S., Bernard, L., Viale, G., Pelicci, P.G., Di Fiore, P.P., 2010. Biological and molecular heterogeneity of breast cancers correlates with their cancer stem cell content. Cell 140, 62e73. Perou, C.M., Sorlie, T., Eisen, M.B., van de Rijn, M., Jeffrey, S.S., Rees, C.A., Pollack, J.R., Ross, D.T., Johnsen, H., Akslen, L.A., Fluge, O., Pergamenschikov, A., Williams, C., Zhu, S.X., Lonning, P.E., Borresen-Dale, A.L., Brown, P.O., Botstein, D., 2000. Molecular portraits of human breast tumours. Nature 406, 747e752. Prat, A., Perou, C.M., 2009. Mammary development meets cancer genomics. Nat. Med. 15, 842e844. Rhodes, D.R., Yu, J., Shanker, K., Deshpande, N., Varambally, R., Ghosh, D., Barrette, T., Pandey, A., Chinnaiyan, A.M., 2004. ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia 6, 1e6.

593

Ribeiro-Silva, A., Ramalho, L.N., Garcia, S.B., Brandao, D.F., Chahud, F., Zucoloto, S., 2005. p63 correlates with both BRCA1 and cytokeratin 5 in invasive breast carcinomas: further evidence for the pathogenesis of the basal phenotype of breast cancer. Histopathology 47, 458e466. Rody, A., Karn, T., Liedtke, C., Pusztai, L., Ruckhaeberle, E., Hanker, L., Gaetje, R., Solbach, C., Ahr, A., Metzler, D., Schmidt, M., Muller, V., Holtrich, U., Kaufmann, M., 2011. A clinically relevant gene signature in triple negative and basallike breast cancer. Breast Cancer Res. 13, R97. Skibinski, A., Breindel, J.L., Prat, A., Galvan, P., Smith, E., Rolfs, A., Gupta, P.B., Labaer, J., Kuperwasser, C., 2014. The Hippo transducer TAZ interacts with the SWI/SNF complex to regulate breast epithelial lineage commitment. Cell Rep. 6, 1059e1072. TCGA, N., 2012. Comprehensive molecular portraits of human breast tumours. Nature 490, 61e70. Udden, S.M., Morita-Fujimura, Y., Satake, M., Ikawa, S., 2013. cABL tyrosine kinase modulates p53-dependent p21 induction and ensuing cell fate decision in response to DNA damage. Cell Signal 26, 444e452. Weigelt, B., Geyer, F.C., Reis-Filho, J.S., 2010. Histological types of breast cancer: how special are they? Mol. Oncol. 4, 192e208. Welcsh, P.L., King, M.C., 2001. BRCA1 and BRCA2 and the genetics of breast and ovarian cancer. Hum. Mol. Genet. 10, 705e713. Wu, S., Huang, S., Ding, J., Zhao, Y., Liang, L., Liu, T., Zhan, R., He, X., 2010. Multiple microRNAs modulate p21Cip1/Waf1 expression by directly targeting its 3’ untranslated region. Oncogene 29, 2302e2308. Yalcin-Ozuysal, O., Fiche, M., Guitierrez, M., Wagner, K.U., Raffoul, W., Brisken, C., 2010. Antagonistic roles of Notch and p63 in controlling mammary epithelial cell fates. Cell Death Differ. 17, 1600e1612. Yallowitz, A.R., Alexandrova, E.M., Talos, F., Xu, S., Marchenko, N.D., Moll, U.M., 2014. p63 is a prosurvival factor in the adult mammary gland during post-lactational involution, affecting PI-MECs and ErbB2 tumorigenesis. Cell Death Differ. 21, 645e654. Yang, A., Kaghad, M., Wang, Y., Gillett, E., Fleming, M.D., Dotsch, V., Andrews, N.C., Caput, D., McKeon, F., 1998. p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol. Cell 2, 305e316. Zhang, X.H., Giuliano, M., Trivedi, M.V., Schiff, R., Osborne, C.K., 2013. Metastasis dormancy in estrogen receptor-positive breast cancer. Clin. Cancer Res. 19, 6389e6397. Zhu, J., Sanborn, J.Z., Benz, S., Szeto, C., Hsu, F., Kuhn, R.M., Karolchik, D., Archie, J., Lenburg, M.E., Esserman, L.J., Kent, W.J., Haussler, D., Wang, T., 2009. The UCSC Cancer genomics browser. Nat. Meth. 6, 239e240.