Pygo2 Activates MDR1 Expression and Mediates Chemoresistance in Breast Cancer Via the Wnt/Β-Catenin Pathway

ORIGINAL ARTICLE Pygo2 activates MDR1 expression and mediates chemoresistance in breast cancer via the Wnt/β-catenin pathway

Z-M Zhang1,2,5, J-F Wu2,3,5, Q-C Luo1,5, Q-F Liu2,4,5, Q-W Wu1, G-D Ye2,4, H-Q She1 and B-A Li2,4

The Wnt/β-catenin pathway has important roles in chemoresistance and multidrug resistance 1 (MDR1) expression in some cancers, but its involvement in breast cancer and the underlying molecular mechanism are undefined. In this study, we demonstrated that the Wnt/β-catenin pathway is activated in chemoresistant breast cancer cells. Using a Wnt pathway-specific PCR array screening assay, we detected that Pygo2, a newly identified Wnt/β-catenin pathway component, was the most upregulated gene in the resistant cells. Additional experiments indicated that Pygo2 activated MDR1 expression in the resistant cells via the Wnt/β-catenin pathway. Moreover, the inhibition of Pygo2 expression restored the chemotherapeutic drug sensitivity of the resistant cells and reduced the breast cancer stem cell population in these cells in response to chemotherapy. Importantly, these activities induced by Pygo2 were mediated by MDR1. We also determined the effect of Pygo2 on the sensitivity of breast tumors resistant to doxorubicin in a mouse model. Finally, RNA samples from 64 paired patient tumors (before and after chemotherapy) highly and significantly overexpressed Pygo2 and/or MDR1 after treatment, thus underlining a pivotal role for the Pygo2-mediated Wnt/β-catenin pathway in the clinical chemoresistance of breast cancer. Our data represent the first implication of the Wnt/β-catenin pathway in breast cancer chemoresistance and identify potential new targets to treat the recurrence of breast cancer.

Oncogene (2016) 35, 4787–4797; doi:10.1038/onc.2016.10; published online 15 February 2016

INTRODUCTION ubiquitin-mediated proteasomal degradation.8 Stimulation by Wnt Breast cancer is the leading female malignancy and cause of ligands suppresses the phosphorylation and degradation of β- cancer death in developed countries. Despite recent advances in catenin, which then enters the nucleus and binds to the LEF1/TCF combined therapies, recurrent disease due to treatment failure in family of transcription factors. In turn, the β-catenin–LEF1/TCF patients remains a major clinical problem. This recurrence is complex regulates the expression of downstream target genes mainly attributed to the development of chemoresistance during involved in diverse cellular processes.9,10 Although the Wnt/β- treatment.1 Cancer cells can acquire chemoresistance via the catenin pathway has been revealed to have important roles in overexpression of drug efflux transporters, such as multidrug chemoresistance and MDR1 expression in colon cancer11 and resistance 1 (MDR1), and multidrug resistance-associated neuroblastoma,12 its involvement in breast cancer and the proteins,2–5 which facilitate the efflux of therapeutic drugs from underlying molecular mechanism are undefined. cells to prevent the accumulation of the cytotoxic agents in cells. Pygopus has been identified as a novel component of the MDR1 (P-glycoprotein, ABCB1), the best-characterized drug efflux Wingless (Wg) pathway in Drosophila.13,14 Pygo proteins (Pygo1 pump, has been found to participate in the chemoresistance of and Pygo2 in mammals) are thought to promote β-catenin/LEF/ breast cancer.5 However, the molecular mechanisms of the TCF transcriptional activation via the regulation of β-catenin regulation of MDR1 expression in breast cancer remain poorly nuclear retention and/or by facilitating β-catenin to recruit – understood. transcriptional co-activators.15 20 Moreover, multiple findings have The Wnt/β-catenin pathway has a central role in normal demonstrated that Pygo2 promotes β-catenin activity in a – development and tumorigenesis.6,7 In the absence of Wnt gene- and tissue-dependent manner in mammalian cells.21 24 In stimulation, cytoplasmic β-catenin is assembled into the destruc- this study, we demonstrate that the Wnt/β-catenin pathway and tion complex composed of axin, glycogen synthase kinase 3β Pygo2 are activated in chemoresistant breast cancer and that (GSK3β), adenomatous polyposis coli (APC) and CK1-α. β-catenin is Pygo2 activates MDR1 expression in chemoresistant breast cancer phosphorylated in this complex and subsequently targeted for cells via the Wnt/β-catenin pathway. Our study reveals an

1Department of Breast Surgery, The First Affiliated Hospital, Xiamen University, Xiamen, Fujian, China; 2State Key Laboratory of Cellular Stress Biology, Innovation Center for Cell Signaling Network, School of Life Sciences, Xiamen University, Xiamen, Fujian, China; 3Department of Bioengineering, College of Food and Bioengineering, Henan University of Science and Technology, Luoyang, China and 4Engineering Research Center of Molecular Diagnostics, Ministry of Education, School of Life Sciences, Xiamen University, Xiamen, Fujian, China. Correspondence: Dr H-Q She, Department of Breast Surgery, The First Affiliated Hospital, Xiamen University, 55 Zhenhai Rd, Xiamen, 361003 Fujian, China or Dr B-A Li, Department of Biomedical Sciences, Xiamen University School of Life Sciences, D212 Huang Chaoyang Hall, Xiang’an Campus of Xiamen University, Xiamen, 361100 Fujian, China. E-mail: [email protected] or [email protected] 5These authors contributed equally to this work. Received 20 June 2015; revised 24 December 2015; accepted 27 December 2015; published online 15 February 2016 Pygo2 and MDR1 in chemoresistance of breast cancer Z-M Zhang et al 4788 association between MDR1 expression and Pygo2, which high- immunoprecipitation assay, both β-catenin and Pygo2 were found to lights the role of the Wnt/β-catenin pathway in the regulation of occupy the endogenous MDR1 promoter at the LEF1/TCF-binding chemoresistance of breast cancer. site in MCF-7/ADR cells (Figure 2e). However, the knockdown of β-catenin led to a dramatically reduced occupancy by Pygo2 (Figure 2f). Together, the above results suggest that Pygo2 activates RESULTS MDR1 expression in MCF-7/ADR cells via the Wnt/β-catenin pathway. The Wnt/β-catenin pathway and Pygo2 are activated in drug- resistant cell lines Inhibition of Pygo2 expression restores chemotherapeutic drug To determine whether the Wnt/β-catenin pathway is involved in sensitivity of the resistant cells β the chemoresistance of breast cancer, we performed Wnt/ - We hypothesized that Pygo2 activation is required for drug catenin-dependent luciferase assays in three doxorubicin-resistant resistance of the resistant cells. Therefore, we evaluated the effect cell lines (MCF7/ADR, MDA-MB-231/ADR and T-47D/ADR) com- of Pygo2 knockdown on the cell sensitivity to doxorubicin and pared with their respective counterparts (MCF7, MDA-MB-231 and paclitaxel. After Pygo2 was knocked down in MCF-7/ADR cells, the T-47D). The TopFlash assays revealed a much higher TCF cells were treated with the aforementioned drugs, and the cell transcriptional activities in the resistant cell lines than in the viability was measured using the MTS/PMS assay. As shown in non-resistant cell lines (Figure 1a). Consistent with this result, a Figures 3a and b, Pygo2 inhibition significantly restored the western blot analysis revealed that the nuclear expression of sensitivity of MCF-7/ADR cells to doxorubicin and paclitaxel in a β-catenin was increased in the resistant cells compared with the dose-dependent manner. Similar results were achieved when we non-resistant cells (Figure 1b). To define the underlying molecular performed the above experiments in MDA-MB-231/ADR cells mechanism by which β-catenin is activated, we screened the gene (Supplementary Figures S2A and B) and T-47D/ADR cells expression profiles of MCF7 versus MCF7/ADR using a Wnt (Supplementary Figures S2C and D). We next examined the effect pathway-specific PCR array. Notably, the upregulation of Pygo2, of Pygo2 on the drug efflux driven by MDR1. The rhodamine-123 a newly identified Wnt/β-catenin pathway component, was retention assay demonstrated that Pygo2 knockdown significantly highest in MCF7/ADR cells (Supplementary Table S1). Quantitative increased the intracellular accumulation of rhodamine-123 com- PCR and western blot analyses confirmed the overexpression of pared with the control cells (Figure 3c, left). When MDR1 was Pygo2 in all three resistant cell lines (Figure 1c). When Pygo2 was overexpressed in the Pygo2 knockdown cells to a level compar- knocked down in MCF-7/ADR cells, the TopFlash assay showed able with that of the control cells (Figure 3c, right), the knocking significantly reduced TCF transcriptional activity (Figure 1d). As down of Pygo2 no longer increased the intracellular accumulation negative controls, NF-κB and FOXO1/3a signaling pathways, two of rhodamine-123 (Figure 3c, left). Finally, we explored the other key MDR1 (see below) transcription regulators, were not mechanisms involved in drug sensitization after Pygo2 inhibition. affected upon Pygo2 downregulation, when we performed NF-κB- A TUNEL assay showed that Pygo2 knockdown in the MCF-7/ADR and FOXO-luciferase assays in MCF-7/ADR cells (Supplementary cells resulted in apoptotic cells in response to chemotherapy Figures S1A-C). Moreover, the nucleus/cytoplasm ratio of β- (Figure 3d). Consistent with this result, cleaved poly ADP-ribose catenin was found to be reduced in MCF-7/ADR cells in which polymerase, an apoptotic protein, was detected in the TUNEL- Pygo2 was knocked down (Figure 1e). By contrast, the nucleus/ positive cells as revealed by a western blot analysis (Figure 3e). cytoplasm ratio of β-catenin was increased when Pygo2 was The induction of caspase-dependent cell death was also measured overexpressed (Figure 1f). Taken together, these results suggest using a caspase-3-like activity assay. As shown in Figures 3f and g, that the Wnt/β-catenin pathway is activated in drug-resistant cells Pygo2 knockdown in the MCF-7/ADR cells resulted in a higher and that this activity is likely mediated by Pygo2. relative caspase-3-like activity in response to doxorubicin and paclitaxel compared with the control cells. In addition, blocking Pygo2 activates MDR1 expression in the resistant cells through the MDR1 function with the MDR1 inhibitor verapamil did not elicit an Wnt/β-catenin pathway apparent additive effect on cell death in Pygo2-knockdown cells. Previous reports have indicated that the Wnt/β-catenin pathway These results reveal that Pygo2 induces chemoresistance through activates MDR1 expression in colorectal cancer and chemoresis- MDR1 in MCF-7/ADR cells. tant neuroblastoma cells. We evaluated the involvement of this pathway in the activation of MDR1 in chemoresistant breast Inhibition of Pygo2 expression leads to a reduced breast cancer cancer cells. First, we cloned a fragment extending from − 350 to stem cell population in the resistant cells under chemotherapy +31 of the MDR1 genomic sequence, which contains a consensus Because cancer stem cells possess chemoresistant features, we LEF1/TCF-binding site of CTTTGAA, into a PGL3-basic promoter- hypothesized that chemoresistant cancer cells acquire an less luciferase reporter cassette. As shown in Figure 2a, this expanded stem cell population. Stem cells in breast cancer can fragment was sufficient to drive the luciferase reporter activity in be enriched from established cancer cell lines using either MCF-7/ADR cells. When β-catenin was knocked down in MCF-7/ mammosphere cultures or fluorescence-activated cell sorting ADR cells, the MDR1 luciferase reporter activity was significantly (FACS) sorting for a CD44+ CD24− population. Compared with reduced. Consistent with this result, the knockdown of β-catenin MCF-7 cells, MCF-7/ADR cells formed more and larger mammo- resulted in the decreased expression of MRD1, as assessed by spheres (Figure 4A) as well as a larger CD44+ CD24− pool quantitative PCR and western blot analyses (Figure 2b). We next (Figure 4B), which suggested that chemoresistant breast cancer examined the relationship between Pygo2 and MDR1. As shown in contains more stem cells. Using western blot analysis, we detected Figures 2a and c, the knockdown of Pygo2 also resulted in the a higher level of Pygo2 protein in mammospheres of MCF7 cell reduction of both the luciferase activity and expression of MDR1. lines than in their corresponding adherent cultures Finally, we investigated the ability of Pygo2 to regulate MDR1 (Supplementary Figure S3A, top lane). Although MDR1 expression expression via the Wnt/β-catenin pathway. We mutated the LEF1/ was undetectable in MCF7 cells, we found an obvious expression TCF-binding site within the MDR1 promoter and performed the of MDR1 in mammospheres (Supplementary Figure S3A, middle luciferase reporter assay in MCF7/ADR cells. As shown in Figure 2d, lane). The overexpression of Pygo2 in MCF7 cells resulted in overexpression of Pygo2 resulted in increased luciferase activity increased MDR1 expression in mammospheres, but not in the in the no mutated wild-type luciferase reporter; however, adherent cultures (Supplementary Figure S3B). The above results mutation of the LEF1/TCF-binding site resulted in significantly imply that Pygo2 might connect the chemoresistance and breast compromised luciferase activity induced by Pygo2. In a chromatin cancer stem cells.

Figure 1. The Wnt/β-catenin pathway and Pygo2 are activated in drug-resistant cell lines. (a) TCF transcriptional activity was compared between resistant cell lines (MCF7/ADR, MDA-MB-231/ADR and T-47D/ADR) and their respective counterparts (MCF7, MDA-MB-231 and T-47D). Cells were transiently transfected with TopFlash reporter, and the luciferase activity was measured. 231 indicated MDA-MB-231. (b) The relative nuclear and cytoplasmic expression of β-catenin was compared between the resistant cell lines and their respective counterparts using western blotting. (c) Pygo2 expression was compared between the resistant cell lines and their respective counterparts using quantitative PCR (top) and western blotting (bottom). (d) Pygo2 knockdown resulted in reduced Wnt/β-catenin activity in MCF-7/ADR cells. Pygo2 expression was knocked down by short hairpin RNAs in MCF-7/ADR cells, and the luciferase activity was measured. Western blotting indicated that Pygo2 was effectively knocked down. (e and f) The relative nuclear–cytoplasmic distribution of β-catenin was detected upon Pygo2 knockdown (e) or overexpression (f) using western blotting. In a, c and d, data are the means ± s.d. of three independent experiments (***Po0.001; ****Po0.0001).

Next, we knocked down the Pygo2 expression in MCF-7/ADR treatment did not elicit changes in mammosphere formation in cells and found a reduced number and size of mammospheres the control cells, in which Pygo2 was not knocked down (Figures 4C; a versus c and f), which was consistent with our (Figure 4C; a versus b and f). However, when MDR1 was re- previous observation.20 MCF-7/ADR cells were also treated with expressed in the Pygo2 knockdown cells, the doxorubicin-induced 50 μM doxorubicin. The number and size of mammospheres was stem cell reduction phenotype disappeared (Figure 4C; c–g). further reduced in cells in which Pygo2 was knocked down Similar results were achieved when we detected the size of the (Figure 4C; c versus d and f), whereas the same doxorubicin CD44+ CD24− population in the same cells (Figure 4D). We also

Figure 2. Pygo2 activates MDR1 expression in the resistant cells through the Wnt/β-catenin pathway. (a) β-catenin or Pygo2 knockdown resulted in reduced pGL3-MDR1 promoter luciferase activity in MCF-7/ADR cells. (left) β-catenin or Pygo2 expression levels were knocked down by short hairpin RNAs in MCF-7/ADR cells. The luciferase activity of the pGL3-MDR1 promoter (−350 bp to +31 bp) vector containing a putative TCF-binding site was measured. (right) Quantitative PCR indicated that β-catenin and Pygo2 were effectively knocked down. (b) Knockdown of β-catenin resulted in the decreased expression of MRD1. β-catenin expression was knocked down by short hairpin RNAs in MCF-7/ADR cells. (top) Quantitative PCR analysis indicated the relative mRNA expression of MDR1. (bottom) Western blotting indicated the relative expression levels of β-catenin and MDR1. (c) Knockdown of Pygo2 resulted in the decreased expression of MRD1. Pygo2 expression was knocked down by short hairpin RNAs in MCF-7/ADR cells. (top) Quantitative PCR analysis indicated the relative mRNA expression of MDR1. (bottom) Western blotting indicated the relative expression levels of Pygo2 and MDR1. (d) Pygo2 activates the pGL3-MDR1 promoter luciferase activity through the TCF-binding site. (left) Luciferase activity of wild-type or TCF-binding site mutant pGL3-MDR1 vectors in LV-GFP or LV-Pygo2-treated MCF7/ADR cells. (right) Western blotting indicated that Pygo2 was effectively overexpressed. (e) Chromatin immunoprecipitation assay of Pygo2 and β-catenin for the MDR1 promoter using semi-quantitative reverse transcription-PCR (left) and quantitative PCR (right) analyses. Primers amplifying a region 10 kb upstream of the MDR1 promoter were used as a negative control for the semi-quantitative reverse transcription-PCR. (f) β-catenin knockdown resulted in the decreased occupancy of the MDR1 promoter by Pygo2. Chromatin fragments from MCF-7/ADR cells infected with lentiviruses expressing LacZ (control) or β-catenin short hairpin RNAs were immunoprecipitated with the Pygo2 antibodies. DNA association was determined by quantitative PCR. In a–f, data are the means ± s.d. of three independent experiments (**Po0.01).

performed the above experiments in other resistant cell lines and Collectively, these results suggest that inhibition of Pygo2 similar phenotypes were observed for mammosphere in MDA- expression leads to a reduced breast cancer stem cell population MB231/ADR cells (Supplementary Figure S3C) and for CD44+ in the resistant cells under chemotherapy and that this activity is CD24− population in T-47D/ADR cells (Supplementary Figure S3D). mediated by MDR1.

Figure 3. Inhibition of Pygo2 expression restores the chemotherapeutic drug sensitivity of the resistant cells. (a and b) Pygo2 inhibition signiﬁcantly restored the sensitivity of MCF-7/ADR cells to doxorubicin (a) and paclitaxel (b) in a dose-dependent manner. MCF-7/ADR cells infected with lentiviruses expressing LacZ (control) or Pygo2 short hairpin RNA were treated with drugs for 48 h. Cell viability was measured using the MTS/PMS assay and compared with the non-treated cells (100% viability). (c) MCF-7/ADR cells were infected with sh-Pygo2 and simultaneously treated with Ad-MDR1. (left) R-123 retention assay was used to detect the intracellular accumulation of Rhodamine-123. (right) Quantitative PCR analysis was used to detect the relative mRNA expression levels of Pygo2 and MDR1. (d) MCF-7/ADR cells infected with lentiviruses expressing LacZ (control) or Pygo2 short hairpin RNAs were treated with or without 50 μM doxorubicin for 48 h. A TUNEL assay was used to detect the apoptotic cells. (e) Western blotting was used to measure the expression of cleaved poly ADP-ribose polymerase, an apoptotic protein, in the cells described in D. (f and g) Knockdown of Pygo2 induces caspase-dependent apoptosis in the MCF-7/ADR cells in response to the different drugs via MDR1 downregulation. MCF-7/ADR cells infected with lentiviruses expressing LacZ (control) or Pygo2 short hairpin RNAs were treated with 50 μM doxorubicin (f) or 1.0 μM paclitaxel (g) for 48 h in presence or absence of 10 mM verapamil. Control cells were treated with drug vehicle. The caspase-3-like activities relative to non-treated cells were measured for each condition. In a, b, c, f and g, data are the means ± s.d. of three independent experiments (*Po0.05; **Po0.01; ***Po0.001). Scale bar, 100 μm.

Figure 4. Inhibition of Pygo2 expression leads to a reduced breast cancer stem cell population in the resistant cells under chemotherapy. (A)MCF-7/ ADR cells formed more and larger mammospheres than MCF-7 cells. Representative photographsofmammospheresweretakenonday12(left). The number (middle) and size (right) of spheres were quantified. (B) MCF-7/ADR cells formed a larger CD44+ CD24− pool than MCF-7 cells. The percentage of CD44+ CD24− cells was determined using FACS analysis. Representative FACS profiles from a single pair are shown (left), and the diagrams show the mean values for three different pairs (right). (C) Knockdown of Pygo2 expression in MCF-7/ADR cells resulted in a reduced number and size of mammospheres in response to chemotherapy via MDR1 downregulation. MCF-7/ADR cells infected with lentiviruses expressing LacZ (control, a and b) or Pygo2 (c and d) short hairpin RNAs were treated with (b and d) 50 μM doxorubicin or vehicle (a and c) for 48 h. MCF-7/ADR cells infected with lentiviruses expressing Pygo2 short hairpin RNAs were also simultaneously treated with doxorubicin and Ad-MDR1 (e). Representative photographs of mammospheres were taken on day 12 (a–e). The number (f) and size (g) of spheres were quantified. (D)Knockdown of Pygo2 expression in MCF-7/ADR cells resulted in a smaller CD44+CD24− pool in response to chemotherapy via MDR1 downregulation. The experimental procedures were the same as those described in C.RepresentativeFACSprofiles from a single pair are shown (a–e), and the diagrams show the mean values for three different pairs (f). In A and C, data are the means ± s.d. of four independent experiments (***Po0.001). In B and D, data are the means ± s.d. of three independent experiments (***Po0.001, ****Po0.0001). Scale bars, 100 μm.

Figure 5. Inhibition of Pygo2 expression increases the sensitivity of the drug-resistant breast tumors to doxorubicin in a mouse model. MCF-7/ ADR cells were sorted to enrich the CD44+ CD24− population. Nude mice were subcutaneously inoculated with CD44+ CD24− MCF-7/ADR cells (106 cells per mouse) in which Pygo2 was either expressed or knocked down and MDR1 either remained downregulated or was re-expressed. (a) Growth curve of tumors from different groups of nude mice treated as indicated. (b) Diagram of the average weight of the dissected xenograft tumors from different groups of nude mice treated as indicated. (c) Photographs of the dissected xenograft tumors from different groups of nude mice treated as indicated. (d) Protein expression levels of Pygo2 and MDR1 were examined by western blot analysis of tumor tissues from different groups of nude mice. In a and b, data are shown as the means ± s.d. (*Po0.05; **Po0.01).

Inhibition of Pygo2 expression increases the sensitivity of the the mice were intraperitoneally injected with doxorubicin drug-resistant breast tumors to doxorubicin in a mouse model (20 μg/g) every three days for 30 days, and the tumor volumes A tumor xenograft nude mouse model was used to investigate the were monitored and recorded. As indicated in Figures 5a–c, functional role of Pygo2 in regulating the drug resistance of breast xenograft tumors grew much slower in the Sh-Pygo2 group when cancer in vivo. MCF-7/ADR cells were sorted to enrich the CD44+ treated with doxorubicin compared with the sh-LacZ control CD24− population. Nude mice were subcutaneously inoculated group, which was not affected by the treatment. However, when with CD44+ CD24− MCF-7/ADR cells (106 cells per mouse), in which MDR1 was re-expressed in the Pygo2 knockdown cells, the Pygo2 was either expressed or knocked down and MDR1 either doxorubicin no longer effectively inhibited tumor growth. remained downregulated or was re-expressed. Two weeks later, Correspondingly, the silencing efﬁciency of Pygo2 and the extent

© 2016 Macmillan Publishers Limited, part of Springer Nature. Oncogene (2016) 4787 – 4797 Pygo2 and MDR1 in chemoresistance of breast cancer Z-M Zhang et al 4794 of MDR1 re-expression were confirmed in the tumor tissues from higher in the group of relapsed patients than in the non-relapsed the mice by a western blot analysis (Figure 5d). These results group of patients (Figures 6a and b). A strong correlation between suggest that the inhibition of Pygo2 expression can restore the the Pygo2 and MDR1 expression levels after chemotherapy was chemotherapeutic drug sensitivity of the resistant cells in vivo. observed in the relapsed patient samples but not in the non- relapsed patient samples (Figures 6c and d). Among the group of Expression of Pygo2 and MDR1 in patient tumors relapsed patients, 48.4% of patient tumor samples (15/31) showed To investigate the association of Pygo2 with clinical chemoresis- a significant increase in Pygo2 expression after chemotherapy, tance in breast cancer patients, paired tumor samples obtained via and 71.0% of patient tumor samples (22/31) showed an increase in needle biopsy at diagnosis or surgery after 6 cycles of neoadjuvant MDR1 expression after chemotherapy. Notably, a simultaneous chemotherapy in a cohort of 64 patients with locally advanced increase in Pygo2 and MDR1 was measured in 38.7% of the breast cancer (LABC; stage IIB and III) were analyzed. The patients samples from the relapsed patient group (12/31). In the non- were divided into two groups according to their clinical outcome relapsed patient group, only 15.2% of patients (5/33) showed a as follows: group 1, patients who relapsed; and group 2, patients significant increase in Pygo2 expression, and 18.2% of patients who had not relapsed 1 year after surgery. Matched pre-and post- (6/33) showed an increase in MDR1 expression. Only 6.1% of these chemotherapy samples were compared for Pygo2 or MDR1 mRNA samples (2/33) showed a simultaneous increase in Pygo2 and expression by quantitative PCR (Supplementary Table S2). Overall, MDR1 expression (Table 1, Supplementary Table S2). The Pygo2 or MDR1 overexpression after chemotherapy was much percentages of samples that overexpressed Pygo2 and MDR1 or

Figure 6. Expression of Pygo2 and MDR1 in patient tumors. (a and b) The overexpression of Pygo2 or MDR1 after chemotherapy was much higher in the group of relapsed patients (group 1) than the non-relapsed group of patients (group 2). The Pygo2 or MDR1 mRNA expression levels of matched pre-and post-chemotherapy samples were measured by quantitative PCR. The fold changes in the gene expression of Pygo2 (a) or MDR1 (b) after chemotherapy were compared between group 1 and group 2. The P-values are shown in the graphs (two-tailed Student’s t-test. ****Po0.0001). (c and d) A strong correlation was observed between the Pygo2 and MDR1 expression levels after chemotherapy in the relapsed patients (group 1) but not the non-relapsed samples (group 2). The correlations between the Pygo2 and MDR1 mRNA levels were measured in group 1 (c) and group 2 (d). The Pearson’s R correlation coefﬁcient and P-value are shown in the graphs.

Table 1. The expression status of Pygo2 and MDR1 after chemotherapy in relapsed versus non-relapsed patients

Relapsed Non-relapsed χ2-test

χ2 df P

Pygo2 Overexpressed 15 (48.4%) 5 (15.2%) 8.218 1 0.0041 Non-overexpressed 16 (51.6%) 28 (84.8%) MDR1 Overexpressed 22 (71.0%) 6 (18.2%) 18.10 1 o0.0001 Non-overexpressed 9 (29.0%) 27 (81.8%) Pygo2/MDR1 Co-overexpressed 12 (38.7%) 2 (6.1%) 9.970 1 0.0016 Non-co-overexpressed 19 (61.3%) 31 (93.9%)

Oncogene (2016) 4787 – 4797 © 2016 Macmillan Publishers Limited, part of Springer Nature. Pygo2 and MDR1 in chemoresistance of breast cancer Z-M Zhang et al 4795 co-overexpressed Pygo2/MDR1 after chemotherapy were all not directly induce chemoresistance in breast cancer is unclear. significantly higher in the relapsed patients than in the non- We deduced that MDR1 expression requires higher Wnt/β-catenin relapsed patients (Table 1). These results strongly indicate that output levels and additional factors, which can only be gained by Pygo2 activation is involved in chemoresistance in breast cancer cancer cells subjected to chemotherapy. Previous studies have patients. indicated that Pygo2 is required for the maximal transactivation of Wnt/β-catenin-specific target genes.20,21 Therefore, the Wnt/β-catenin output may be sufficient for MDR1 expression, DISCUSSION but only when Pygo2 is further overexpressed in breast cancer The Wnt/β-catenin pathway has a central role in the development cells that are exposed to chemotherapy. Because almost all breast of normal tissues by regulating cell proliferation and cancer patients are subjected to chemotherapy before the tumor differentiation.6 The aberrant activation of the Wnt/β-Catenin is resected, we speculate that the Pygo2 overexpression found in a pathway has been detected in the carcinogenesis of a wide range previous study was at least partially due to the chemotherapy. of tumors of intestinal, mammary gland, skin, hepatic and Importantly, another study has also reported that hypermethyla- hematopoietic cell origin.7 Moreover, the activation of this tion of the MDR1 gene promoter is observed in MCF-7 cells but pathway may induce MDR1 expression and chemoresistance in that hypomethylation is found in MCF-7/ADR cells. In addition, H3 tumors, such as colon cancer and neuroblastoma.11,12 However, and H4 histone acetylation levels of MCF-7/ADR cells are Wnt/β-catenin activity is triggered by a variety of factors that significantly higher than those of MCF-7 cells.36 These results depend on cellular context and are tissue specific. In this study, we suggest that before transcriptional activation by Pygo2, epigenetic detected enhanced Wnt/β-catenin activity in chemoresistant regulation is required to open the chromatin of MDR1 for the breast cancer cells compared with non-chemoresistant breast development of chemoresistance in breast cancer cells. cancer cells. Moreover, MDR1 was found to be upregulated by the Recent advances in the characterization of stem cells have Wnt/β-catenin pathway in these cells. In our search for the linked cancer stem cells to chemoresistance. Numerous studies factor(s) responsible for the activation of β-catenin signaling in have indicated that cancer stem cells may be involved in chemoresistant breast cancer cells using a PCR array system, we carcinogenesis, invasion and metastasis as well as in the resistance identified Pygo2, a component of the Wnt/β-catenin pathway, to various forms of chemotherapy.37 One of the important as the most upregulated gene in the resistant cells. Pygo2 is a features of cancer stem cells is that they contain high levels of mammalian homolog of Drosophila Pygopus protein, which was ABC transporter proteins, which are responsible for protecting initially identified as an essential component of the Wingless (Wg) cells from drug damage via efflux pumping mechanisms. pathway. The biological functions of Pygo proteins in mammalian Consequently, cancer stem cells are resistant to drug treatment, cells are complex. For example, Pygo2 is required for the including chemotherapeutic drugs. Because cancer stem cells Wnt/β-catenin output in skin and mammary gland tissues possess chemoresistant features, we hypothesized that chemo- but not in other tissues, such as the intestines, where Wnt/β- resistant cancer cells acquire more stem cells. Indeed, we 25 catenin signaling is active. Therefore, Pygo2 promotes observed an expanded stem cell population in MCF-7/ADR cells Wnt/β-catenin activity in a tissue-dependent manner in mamma- compared with MCF-7 cells. Moreover, we demonstrated that the lian cells. In addition, previous studies have also revealed that not inhibition of Pygo2 expression further reduced the breast cancer all Wnt/β-catenin target genes are affected by Pygo2 in skin and stem cell population in the resistant cells in response to 21,24,26 mammary gland tissues, which suggests that Pygo2 chemotherapy and that this activity was mediated by MDR1. promotes β-catenin activity in a gene-dependent manner in Our previous studies indicated that Pygo2 expands breast cancer mammalian cells. Thus, although MDR1 is the downstream target stem cell population through other Wnt/β-catenin targets such as gene of the Wnt/β-catenin pathway in colon cancer and c-myc and Lin28-let7.20,38 That is why we observed in this study neuroblastoma, MDR1 is not necessarily regulated by Pygo2 in that MDR1 overexpression in Pygo2 knockdown cells only re- breast cancer. To date, Pygo2 has been found to be overexpressed establish sphere number and cells/spheres when compared with 27 28 29 30 in breast cancer, ovarian cancer, lung cancer, glioma and DOX-untreated Pygo2 knockdown cells, but not the wild-type 31 esophageal squamous cell carcinoma, which implies that Pygo2 cells. Nevertheless, our study revealed a reciprocal link between has a pivotal role in the carcinogenesis of these tumors. However, stem cell properties and chemoresistance in cancer cells. the ability of Pygo2 to regulate the chemoresistance of tumors In summary, we identified a mechanism that controls MDR1 remains unknown. In the current study, we demonstrated that expression in vitro and in vivo, which sheds new light on the Pygo2 was overexpressed in both a chemoresistant breast cancer regulation of chemoresistance in breast cancer. The modulation of line and chemoresistant breast cancer specimen compared with this molecular process may be a method of inhibiting breast cancer their non-chemoresistant counterparts. We also found that Pygo2 recurrence by targeting Pygo2 to block Wnt/β-catenin signaling. was required for the transduction of Wnt/β-catenin signaling and the activation of MDR1 in chemoresistant breast cancer cells, suggesting that Pygo2 mediates the chemoresistance of breast MATERIALS AND METHODS cancer. The results of our mechanistic study indicated that Pygo2 Cell culture β regulated the -catenin nuclear/cytoplasmic distribution in Human breast cancer MCF-7, MDA-MB-231 and T-47D cell lines were chemoresistant breast cancer cells, which was consistent with obtained from ATCC (American Type Culture Collection, Manassas, VA, 15,16 the previously reported nuclear retention function of Pygo2. USA). Doxorubicin-resistant cell lines, MCF-7/ADR, MDA-MB-231/ADR and In human breast cancer, overt mutations of Wnt/β-catenin T-47D/ADR, were established from maternal MCF-7, MDA-MB-231 and pathway components are rare. Instead, several members of this T-47D cells by stepwise increasing concentrations (double for each pathway are frequently altered by other means. Several Wnt passage) of doxorubicin treatment. Once the cells could survive in the ligands are overexpressed in human breast cancers and breast highest concentration of the drug for three passages, the stable cell line cancer cell lines.32,33 A loss of sFRP1 expression has been detected was obtained and maintained in medium containing this concentration of doxorubicin. The three paired cell lines were checked for 16 short tandem in breast cancers and found to be associated with poor fi 34 fi repeats DNA pro les, within which nine short tandem repeats are prognosis. Dvl1 is ampli ed and overexpressed in some breast published on ATCC website. When searched the short tandem repeat cancers.35 More recently, Pygo2 has also been found to be 27 database, we found that all the three maternal cell lines are the same with overexpressed in breast cancers. The above changes may those in ATCC. The % match of MCF7 versus MCF-7/ADR, MDA-MB-231 contribute to the activation of β-catenin and carcinogenesis of versus MDA-MB-231/ADR and T-47D versus T-47D/ADR are, respectively, breast cancer. However, why the initially activated β-catenin does 92.8, 85.71 and 100% (According to ATCC standard, cell lines with ⩾ 80%

© 2016 Macmillan Publishers Limited, part of Springer Nature. Oncogene (2016) 4787 – 4797 Pygo2 and MDR1 in chemoresistance of breast cancer Z-M Zhang et al 4796 match are considered to be related; derived from a common ancestry). adenovirus was 1 × 1011 PFU/ml. The purified adenoviruses were used to These cell lines were grown in Dulbecco’s modified Eagle’s medium infect cells. supplemented with 10% fetal calf serum (Gibco, Grand Island, NY, USA), μ 100 U/ml penicillin and 100 g/ml streptomycin at 37 °C in a 5% CO2 Mammosphere culture humidified atmosphere. PCR tests for mycoplasma were negative. The cells (5000 cells per 1 ml) were cultured in ultra-low attachment plates in serum-free Dulbecco’s modified Eagle’s medium/F12 (Invitrogen, Antibodies Waltham, MA, USA) supplemented with B-27 (1:50; Invitrogen), 20 ng/ml The following antibodies were used: anti-β-catenin (Cell Signaling epidermal growth factor (BD Biosciences, San Jose, CA, USA), 20 ng/ml Technology, Danvers, MA, USA, cat. no. 9562), anti-β-actin (Sigma-Aldrich, basic fibroblast growth factor (bFGF; BD Biosciences) and 4 mg/ml insulin St Louis, MO, USA cat. no. A1978), anti-Lamin B (Santa Cruz Biotechnology, (Sigma-Aldrich) and the cells were fed every 3 days. The mammosphere Dallas, TX, USA, cat. no. sc-6216), anti-MDR1 (Santa Cruz Biotechnology, cat. number and size were measured on day 12. no. sc-733514), anti-Pygo2 (R&D Systems, Minneapolis, MN, USA, cat. no. AF3616), anti-CD44-PE-Cyanine5 (eBioscience, San Diego, CA, USA, cat. no. Cell analysis and sorting of CD44+CD24− population 15-0441) and anti-CD24-PE (eBioscience, cat. no. 12-0247). Confluent cells were washed once with phosphate-buffered saline and then harvested with 0.05% trypsin/0.025% EDTA. Detached cells were Quantitative real-time PCR washed with phosphate-buffered saline containing 2% FBS (FACS buffer), Full details are available in Supplementary Materials and Methods. and resuspended in the FACS buffer. Combinations of fluorochrome- conjugated monoclonal antibodies against human CD44-PE-Cyanine5 and CD24-PE or their respective isotype controls were added to the cell Chromatin immunoprecipitation assay suspension at concentrations recommended by the manufacturer and Full details are available in Supplementary Materials and Methods. incubated on ice in the dark for 30 min. The labeled cells were washed and resuspended in the FACS buffer. For Cell analysis of CD44+CD24− 6 Cell viability assays population, 10 cells were analyzed using a Beckman EPICS XL instrument at 10 000 events. For Cell sorting of CD44+CD24− population, cells were Full details are available in Supplementary Materials and Methods. sorted with Beckman MoFlo XDP instrument. The cutoff point was weak or absent expression of CD24 in conjunction with high expression of CD44. A Caspase-3 activity assay small amount of cells were reanalyzed for CD44+/CD24− purity. Full details are available in Supplementary Materials and Methods. Xenografted breast cancer model in mice Rhodamine-123 retention assay CD44+CD24− MCF-7/ADR cells were plated on day 1 into 6 cm dishes with Full details are available in Supplementary Materials and Methods. complete medium (Dulbecco’smodified Eagle’s medium/F12 supplemented with 10% fetal bovine serum and 1 × antibiotic-mycotic) and incubated overnight. The medium was replaced on day 2 with fresh complete Luciferase assay medium supplemented with 5 μg/ml polybrene (Sigma-Aldrich) and 107 Full details are available in Supplementary Materials and Methods. infectious units of lentivirus and 107 infectious units of adenovirus. The plate was shaken to mix the virus particles and incubated overnight at Lentivirus-mediated RNA interference or gene transfer 37 °C in 5% CO2. The cells were harvested on day 3 and were used for further experiments for evaluating tumorigenesis in nude mice. The Plko.1 lentiviral vector was used to express short hairpin RNA directed Animal care and handling procedures were performed in accordance with β β against -catenin (sh- -catenin), Pygo2 (sh-Pygo2) or LacZ control the Guide for the Care and Use of Laboratory Animals, and the animal study β ’ sequence. The following sequences were used: sh- -catenin-1, 5 -TGG protocol was approved by the Institutional Animal Care and Use Committee TTAATAAGGCTGCAGTTATTTCAAGAGAATAACTGCAGCCTTATTAACCTTTTTT of Xiamen University (Reference No.XMULAC20150166). The animals used ’ β ’ C-3 ; sh- -catenin-2,5 - TGGATGTGGATACCTCCCAAGTTTCAAGAGAACTTG to test the treatment were 5-week-old female BALB/c athymic nude mice. ’ ’ GGAGGTATCCACATCCTTTTTTC -3 ; Sh-Pygo2-1, 5 -TGGGATTTGGTCCCATGA Briefly, 30 nude mice were randomly allocated to five groups and ’ ’ − TCTCTTCAAGAGAGAGATCATGGGACCAAATCCCTTTTTTC-3 ; Sh-Pygo2-2, 5 - subcutaneously inoculated with the above CD44+CD24 MCF-7/ADR cells AAAAGGATGTGGATACCTCCCAAGTTTGGATCCAAACTTGGGAGGTATCCACA (106 cells for each mouse, n=6 mice per group) with different treatment. ’ ’ TCC-3 and Sh-LacZ, 5 -GTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACA Two weeks later, the mice were intraperitoneally injected with doxorubicin ’ CGTTCGGAGACTTTTTTC-3 . (20 μg/g) every three days for 30 days, and the tumor volumes were The lentiviral pBobi vector was used to express the full-length Pygo2, monitored and recorded every five days. The investigator who monitored ’ and the following paired primers were used: 5 GGGAATTCCATATGATGGCC the tumor volumes was blinded to group allocation. The tumor volumes ’ ’ ’ GCCTCGGCGCCGCC 3 and 5 GCTCTAGATCACCCATCGTTAGCAGCCA 3 . were estimated using the following formula: 0.5 × length × width2. The generation of lentivirus vectors was performed by co-transfecting Plko.1 or pBobi carrying the expression cassette with helper plasmids pVSV-G and pHR into 293 T cells using Lipofectamine 3000 (Invitrogen Life Patients Technology, Waltham, MA, USA). The viral supernatant was collected 48 h The cohort of patients included 64 patients with locally advanced breast after transfection, and cells at 60–70% confluency were infected with viral cancer (LABC; stage IIB and III) diagnosed from 2010 to 2013. The Pygo2 supernatants containing 10 mg/ml polybrene for 24 h, after which fresh and MDR1 expression levels were analyzed in the paired tumor samples medium was added to the infected cells. obtained by needle biopsy at diagnosis or surgery after six cycles of neoadjuvant chemotherapy. This sample size (n420 for each group) Adenovirus-mediated gene transfer ensured an equal variance between the two compared groups and a normal distribution of the detected data within each group. The tumor The MDR1-coding sequence was cloned into the PAdtrack-CMV shuttle material was collected after informed consent and in agreement with the ’ vector using the following paired primers: 5 -ACGCGTCGACGCCACCATG Institutional Review Board of The First Affiliated Hospital of Xiamen ’ ’ ’ GATCTTGAAGGGGACCG-3 and 5 -CCTCGAGTCACTGGCGCTTTGTTCCAG-3 . University. The investigator who performed the examination of Pygo2 and The resultant plasmid was linearized by digestion with the PmeI restriction MDR1 expression was blinded to the information about all the samples. endonuclease and subsequently co-transformed into BJ5183 E.coli cells with the pAdEasy-1 adenoviral backbone plasmid. Recombinants were selected for kanamycin resistance, and recombination was confirmed by a Statistical analyses restriction endonuclease analysis. Finally, the linearized recombinant Statistical analyses were performed using GraphPad Prism 4.0 (La Jolla, CA, plasmid was transfected into adenovirus-packaging 293 cells. Recombinant USA). All data are presented as the means ± s.d. Student’s t-test, the Chi- adenoviruses were typically generated within 5–7 days. A high titer stock square test, and Pearson’s R correlation test were used to compare and was prepared using 2 × 108 293 cells. The adenoviruses were enriched via calculate P-values. All the statistical tests were two-sided. Po0.05 was cesium chloride gradient centrifugation. The titer of the generated considered significant.

Oncogene (2016) 4787 – 4797 © 2016 Macmillan Publishers Limited, part of Springer Nature. Pygo2 and MDR1 in chemoresistance of breast cancer Z-M Zhang et al 4797 CONFLICT OF INTEREST 18 Carrera I, Janody F, Leeds N, Duveau F, Treisman JE. Pygopus activates wingless The authors declare no conflict of interest. target gene transcription through the mediator complex subunits Med12 and Med13. Proc Natl Acad Sci USA 2008; 105: 6644–6649. 19 Wright KJ, Tjian R. Wnt signaling targets ETO coactivation domain of TAF4/TFIID ACKNOWLEDGEMENTS in vivo. Proc Natl Acad Sci USA 2009; 106:55–60. 20 Chen J, Luo Q, Yuan Y, Huang X, Cai W, Li C et al. Pygo2 associates with MLL2 This work was supported by grants from the ‘973’ Project of the Ministry of Science histone methyltransferase and GCN5 histone acetyltransferase complexes to and Technology (grant numbers 2013CB530600 to B-AL); the National Natural augment Wnt target gene expression and breast cancer stem-like cell expansion. Science Foundation of China (grant numbers U1205023, 81472457, 81201616, Mol Cell Biol 2010; 30:5621–5635. 81201617 and 81272384 to B-AL, Q-CL and Q-FL); the Health-Education Joint 21 Gu B, Sun P, Yuan Y, Moraes RC, Li A, Teng A et al. Pygo2 expands mammary Research Project of Fujian Province (grant numbers WKJ-FJ-23 to B-AL and Z-MZ); the progenitor cells by facilitating histone H3 K4 methylation. J Cell Biol 2009; 185: Major Project of Science and Technology from the Department of Education 811–826. (grant number 313051 to B-AL); the Natural Science Foundation of Fujian Province 22 Cantu C, Valenta T, Hausmann G, Vilain N, Aguet M, Basler K. The (grant numbers 2011D017 to Z-MZ); the Science and Technology Program of Pygo2-H3K4me2/3 interaction is dispensable for mouse development and Wnt Xiamen (grant numbers 3502Z20114003 to H-QS) and ‘Project 111’ sponsored by the signaling-dependent transcription. Development 2013; 140: 2377–2386. State Bureau of Foreign Experts and Ministry of Education (grant number B06016). 23 Watanabe K, Fallahi M, Dai X. Chromatin effector Pygo2 regulates mammary tumor initiation and heterogeneity in MMTV-Wnt1 mice. Oncogene 2014; 33: – REFERENCES 632 642. 24 Sun P, Watanabe K, Fallahi M, Lee B, Afetian ME, Rheaume C et al. Pygo2 regulates 1 Zardawi SJ, O'Toole SA, Sutherland RL, Musgrove EA. Dysregulation of Hedgehog, beta-catenin-induced activation of hair follicle stem/progenitor cells and skin Wnt and Notch signalling pathways in breast cancer. Histol Histopathol 2009; 24: hyperplasia. Proc Natl Acad Sci USA 2014; 111: 10215–10220. 385–398. 25 Li B, Rheaume C, Teng A, Bilanchone V, Munguia JE, Hu M et al. Developmental 2 Haber M, Bordow SB, Haber PS, Marshall GM, Stewart BW, Norris MD. phenotypes and reduced Wnt signaling in mice deficient for pygopus 2. Genesis The prognostic value of MDR1 gene expression in primary untreated neuro- 2007; 45:318–325. blastoma. Eur J Cancer 1997; 33: 2031–2036. 26 Gu B, Watanabe K, Sun P, Fallahi M, Dai X. Chromatin effector Pygo2 mediates 3 Blanc E, Goldschneider D, Ferrandis E, Barrois M, Le Roux G, Leonce S et al. MYCN Wnt-notch crosstalk to suppress luminal/alveolar potential of mammary stem and enhances P-gp/MDR1 gene expression in the human metastatic neuroblastoma basal cells. Cell Stem Cell 2013; 13:48–61. IGR-N-91 model. Am J Pathol 2003; 163: 321–331. 27 Andrews PG, Lake BB, Popadiuk C, Kao KR. Requirement of Pygopus 2 in 4 Munoz M, Henderson M, Haber M, Norris M. Role of the MRP1/ABCC1 multidrug breast cancer. Int J Oncol 2007; 30: 357–363. transporter protein in cancer. IUBMB Life 2007; 59:752–757. 28 Popadiuk CM, Xiong J, Wells MG, Andrews PG, Dankwa K, Hirasawa K. Antisense 5 Han CY, Cho KB, Choi HS, Han HK, Kang KW. Role of FoxO1 activation in MDR1 suppression of pygopus2 results in growth arrest of epithelial ovarian cancer. Clin expression in adriamycin-resistant breast cancer cells. Carcinogenesis 2008; 29: Cancer Res 2006; 12: 2216–2223. 1837–1844. 29 Liu Y, Dong QZ, Wang S, Fang CQ, Miao Y, Wang L et al. Abnormal expression of 6 Wodarz A, Nusse R. Mechanisms of Wnt signaling in development. Annu Rev Cell Pygopus 2 correlates with a malignant phenotype in human lung cancer. BMC Dev Biol 1998; 14:59–88. Cancer 2013; 13: 346. 7 Polakis P. Wnt signaling and cancer. Genes Dev 2000; 14: 1837–1851. 30 Wang ZX, Chen YY, Li BA, Tan GW, Liu XY, Shen SH et al. Decreased pygopus 2 8 Lustig B, Behrens J. The Wnt signaling pathway and its role in tumor develop- expression suppresses glioblastoma U251 cell growth. J Neurooncol 2010; 100: ment. J Cancer Res Clin Oncol 2003; 129: 199–221. 31–41. 9 Liu C, Li Y, Semenov M, Han C, Baeg GH, Tan Y et al. Control of beta-catenin 31 Moghbeli M, Abbaszadegan MR, Farshchian M, Montazer M, Raeisossadati R, phosphorylation/degradation by a dual-kinase mechanism. Cell 2002; 108: Abdollahi A et al. Association of PYGO2 and EGFR in esophageal squamous cell 837–847. carcinoma. Med Oncol 2013; 30: 516. 10 MacDonald BT, Tamai K, He X. Wnt/beta-catenin signaling: components, 32 Bafico A, Liu G, Goldin L, Harris V, Aaronson SA. An autocrine mechanism for mechanisms, and diseases. Dev Cell 2009; 17:9–26. constitutive Wnt pathway activation in human cancer cells. Cancer Cell 2004; 6: 11 Yamada T, Takaoka AS, Naishiro Y, Hayashi R, Maruyama K, Maesawa C et al. 497–506. Transactivation of the multidrug resistance 1 gene by T-cell factor 4/beta-catenin 33 Milovanovic T, Planutis K, Nguyen A, Marsh JL, Lin F, Hope C et al. Expression of complex in early colorectal carcinogenesis. Cancer Res 2000; 60: 4761–4766. Wnt genes and frizzled 1 and 2 receptors in normal breast epithelium and 12 Flahaut M, Meier R, Coulon A, Nardou KA, Niggli FK, Martinet D et al. The Wnt infiltrating breast carcinoma. Int J Oncol 2004; 25: 1337–1342. receptor FZD1 mediates chemoresistance in neuroblastoma through activation of 34 Veeck J, Niederacher D, An H, Klopocki E, Wiesmann F, Betz B et al. Aberrant the Wnt/beta-catenin pathway. Oncogene 2009; 28:2245–2256. methylation of the Wnt antagonist SFRP1 in breast cancer is associated with 13 Kramps T, Peter O, Brunner E, Nellen D, Froesch B, Chatterjee S et al. Wnt/wingless unfavourable prognosis. Oncogene 2006; 25: 3479–3488. signaling requires BCL9/legless-mediated recruitment of pygopus to the nuclear 35 Nagahata T, Shimada T, Harada A, Nagai H, Onda M, Yokoyama S et al. beta-catenin-TCF complex. Cell 2002; 109:47–60. Amplification, up-regulation and over-expression of DVL-1, the human 14 Thompson B, Townsley F, Rosin-Arbesfeld R, Musisi H, Bienz M. A new nuclear counterpart of the Drosophila disheveled gene, in primary breast cancers. Cancer component of the Wnt signalling pathway. Nat Cell Biol 2002; 4: 367–373. Sci 2003; 94: 515–518. 15 Townsley FM, Cliffe A, Bienz M. Pygopus and Legless target Armadillo/beta- 36 Huang C, Cao P, Xie Z. Relation of promoter methylation of mdr-1 gene and catenin to the nucleus to enable its transcriptional co-activator function. Nat Cell histone acetylation status with multidrug resistance in MCF-7/Adr cells. Zhong Biol 2004; 6:626–633. Nan Da Xue Xue Bao Yi Xue Ban 2009; 34:369–374. 16 Krieghoff E, Behrens J, Mayr B. Nucleo-cytoplasmic distribution of beta-catenin is 37 Abdullah LN, Chow EK. Mechanisms of chemoresistance in cancer stem cells. regulated by retention. J Cell Sci 2006; 119: 1453–1463. Clin Transl Med 2013; 2:3. 17 Andrews PG, He Z, Popadiuk C, Kao KR. The transcriptional activity of Pygopus is 38 Cai WY, Wei TZ, Luo QC, Wu QW, Liu QF, Yang M et al. The Wnt-beta-catenin enhanced by its interaction with cAMP-response-element-binding protein pathway represses let-7 microRNA expression through transactivation of Lin28 to (CREB)-binding protein. Biochem J 2009; 422: 493–501. augment breast cancer stem cell expansion. J Cell Sci 2013; 126: 2877–2889.

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