(2011) 30, 4656–4665 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE TP63 P2 promoter functional analysis identifies b-catenin as a key regulator of DNp63 expression

C Ruptier1,2,3,10,11, A De Gaspe´ris1,2,3,10,SAnsieau1,2,3,AGranjon1,2,3,12,PTanie` re4,13, I Lafosse1,2,3,11, HShi4,14,APetitjean4,15, E Taranchon-Clermont4,16, V Tribollet1,2,3,17,TVoeltzel1,2,3, J-Y Scoazec1,2,3,5, V Maguer-Satta1,2,3, A Puisieux1,2,3,6,7, P Hainaut4, C Cavard8,9 and C Caron de Fromentel1,2,3

1Universite´ de Lyon, Lyon, France; 2INSERM U1052, Centre de Recherche en Cance´rologie de Lyon, Lyon, France; 3CNRS5286, Centre de Recherche en Cance´rologie de Lyon, Lyon, France; 4International Agency for Research on , Lyon, France; 5Hospices Civils de Lyon, Hoˆpital E. Herriot, Service central d’Anatomie et de Cytologie Pathologiques, Lyon, France; 6Centre Le´on Be´rard, Lyon, France; 7Institut Universitaire de France, Paris, France; 8Institut Cochin, De´partement Endocrinologie Me´tabolisme et Cancer, Universite´ Paris Descartes, CNRS (UMR 8104), Paris, France and 9INSERM U567, Paris, France

The DNp63 , a product of the TP63 that lacks Introduction the N-terminal domain, has a critical role in the maintenance of self renewal and progenitor capacity in is the founding member of a family of that several types of epithelial tissues. DNp63 is frequently also comprises p63 and . All three proteins are overexpressed in squamous carcinoma (SCC) and in transcription factors that control partially overlapping some other epithelial tumours. This overexpression may and interconnected genomic programs, such as cell-cycle contribute to tumour progression through dominant- arrest or (Moll and Slade, 2004). These negative effects on the transcriptionally active (TA) proteins are expressed as multiple isoforms due to isoforms of the p53 family (TAp63, TAp73 and p53), as alternative splicing and promoter usage (Yang et al., well as through independent mechanisms. However, the 1998; Yang and McKeon, 2000; Marcel and Hainaut, molecular basis of DNp63 overexpression is not fully 2009). Most of the isoforms possess a central DNA- understood. Here, we show that the expression of DNp63 binding domain, but differ in their N- and C-terminal is regulated by the Wnt/b-catenin pathway in human regions. The N-terminal region contains a transactiva- hepatocellular carcinoma (HCC) and SCC cell lines. This tion domain present in the so-called transcriptionally regulation operates in particular through TCF/LEF sites active (TA) isoforms, but absent in the truncated (DN) present in the P2 promoter of TP63. In addition, we show ones. In the case of TP63, TA isoforms are produced that DNp63 and b-catenin are frequently coexpressed and from P1 promoter, whereas P2 promoter, located in accumulated in oesophageal SCC, but not in HCC. These intron 3, generates DN isoforms (Yang et al., 1998) results suggest that activation of the b-catenin pathway (Figure 1). By competing with TA isoforms for the may contribute to overexpression of DNp63 during binding to their responsive elements (REs), DN isoforms tumour progression, in a cell type-specific manner. prevent target-gene activation, thus behaving as trans- Oncogene (2011) 30, 4656–4665; doi:10.1038/onc.2011.171; dominant-negative isoforms (Yang et al., 1998; Petitjean published online 6 June 2011 et al., 2008). Given the similarities in DNA-binding properties between p63, p73 and p53, non-physiological Keywords: b-catenin; p53 family; promoter; transcrip- upregulation of DNp63 or DNp73 expression may tional regulation; tumour progression contribute to the functional inactivation of p53, thereby promoting tumourigenesis in the absence of mutation in p53. DNp63 or DNp73 upregulation can also turn down Correspondence: Dr C Caron de Fromentel, UMR INSERM U1052 TAp63 and TAp73 activity, favouring tumour progres- CNRS5286, 151 cours Albert Thomas, 69008 Lyon, France. sion independently of p53 inactivation. The notion that E-mail: [email protected] the three factors operate through partially independent 10These authors contributed equally to this work. mechanisms is supported by the cooperative effects of 11Current address: Genoway, 181 avenue Jean Jaure` s, 69007 Lyon, France. 12Current address: CNRS UMR5123 Physiologie Inte´grative Cellulaire TP63, TP73 and TP53 gene inactivation on tumour et Mole´culaire, 43 bd du 11 novembre 1918, 69622 Villeurbanne formation in transgenic mice (Flores et al., 2002; Flores Cedex, France. et al., 2005). Thus, the transcriptional activity of each 13Current address: Department of Cellular Pathology, The Medical p53 family member may depend upon a delicate balance School, University of Birmingham, Edgbaston, Birmingham B115 between TA and DN isoforms. 2TT, UK. 14Current address: Knobelstrasse 3, 5611 Anglikon, Switzerland. Despite their structural and functional similari- 15Current address: Re´sidence Roche Arnaud, 27 rue Emile Reynaud, ties, TP53, TP63 and TP73 present different genetic 43000 Le Puy en Velay, France. alterations in human cancer. Although p53 is frequently 16Current address: CEPH, 27 rue Juliette Dodu, 75010 Paris, France. inactivated by mutation and loss of heterozygosity in 17Current address: CNRS UMR5242- IGFL, ENS Lyon, 46 alle´e d’Italie, 69364 Lyon cedex 07, France. many cancer types, mutations in TP73 and TP63 Received 31 July 2010; revised and accepted 8 April 2011; published are rare. In contrast, deregulated expression of TP63, online 6 June 2011 resulting in an increased DNp63/TAp63 ratio, is common b-catenin-DNp63 axis in tumour progression C Ruptier et al 4657 γ P1 P2 Here, we demonstrate that the longest DN p63

TP63 gene α isoform (DNp63a) and b-catenin bind to and modulate β the TP63 P2 promoter activity. Moreover, we show STAT3 TBE 2 SP1 that b-catenin-mediated P2 activation implies multiple (-678/-659) (-347/-340) (-55/-49) +1 non-canonical TBEs. Finally, in line with these results, P2 we describe that DNp63 overexpression correlates with (-2 928) TBE p53RE CAAT boxes TATA (+139) (-2 083/-2 074) (-615/-593) (-164/-159) box b-catenin delocalization in oesophageal SCC. These (-112/-107) (-27/-23) results provide novel insights into the mechanism by (-78/-73) which TP63 expression is altered in tumours. Figure 1 Structure of the P2 promoter. Schematic representation of TP63 gene organization and of the P2 promoter REs reported in the literature or identified by our in silico analysis. TP63 gene, white, UTR; black, exon 30 coding region; dotted light grey, DNA binding domain; Results hatched, oligomerization domain; dotted dark grey, SAM domain; light grey, post SAM; a, b, g, C-terminal alternative splicing products. Structure of P2 promoter The putative -binding sites present in a 3000 bp region of P2 promoter were identified by using in epithelial tumours, including all forms of squamous the Matinspector (http://www.genomatix.de) software. cell carcinomas (SCCs) and some forms of adeno- Figure 1 shows the structure of the promoter and the carcinoma (Moll and Slade, 2004). In these tissues, the position of putative RE (detailed in Supplementary expression of TP63 is a hallmark of normal proliferating Figure 1). Several regulatory elements located within the cells of the basal or myoepithelial layers. In stratified TP63 P2 promoter, including a p53RE (Lanza et al., , p63 is essential for the self renewal of 2006), a TATA box, three CAAT boxes and a SP1/SP3 progenitors, as well as their rate of progression towards et al. differentiation (Yang et al., 1999). Thus, in tumours, the site (Romano , 2006), were confirmed in our in silico analysis. However, we failed to identify the increased DN/TAp63 ratio is likely to provide growth STAT3 RE previously reported (Chu et al.,2008).Onthe and survival advantages by maintaining cells in a proli- ferative status that retains some of the characteristics of other hand, we identified two putative TBEs and four Pitx2-binding sites (PBSs). The presence of p53RE and progenitors (Senoo et al., 2007; Wu et al., 2005; Rocco TBEs/PBSs suggests that expression of DNp63 may be et al., 2006). However, the molecular mechanisms that controlled through two mechanisms of particular rele- lead to DNp63 overexpression are only partially charac- vance in carcinogenesis, one involving auto- and hetero- terized. DNp63 overexpression has been associated with regulatory loops by the p53 family members and the 3q27–29 amplification, a region encompassing the TP63 other involving the Wnt/b-catenin pathway and TCF/ gene, in about 20–25% of SCC, but it is also observed in without such an amplification, suggesting that LEF factors. On the basis of these considerations, we set out to assess the functionality of TBEs and of the p53RE. other mechanisms may operate to upregulate this gene in a pathological context (Taniere et al., 2001). In the present study, we carried out a survey of Regulation of P2 promoter by DNp63a protein itself TP63 P2 promoter in search of candidate factors for We previously reported that DNp63 and p53 expression transcriptional regulation of DNp63. In silico analysis are inversely correlated in hepatocytes and hepatocel- revealed the presence of a potential p53RE, previously lular carcinoma (HCC) cell lines (Petitjean et al., 2005), identified by others (Waltermann et al., 2003; Lanza suggesting that p53 might repress DNp63. In fact, et al., 2006), and two TCF/LEF-binding elements repression of the P2 promoter by p53 has further been (TBEs). TCF/LEF transcription factors are activated confirmed in keratinocytes (Lanza et al., 2006). As through a signalling cascade involving b-catenin, a DNp63 and p53 share similar DNA-binding specificities key protein of the Wnt/b-catenin pathway frequently (Yang et al., 1998), we investigated the possibility that deregulated in many epithelial cancers, including SCC DNp63a regulates its own promoter. We first transfected (Ninomiya et al., 2000; Kudo et al., 2007). In normal Hep3B cells with a DNp63a-expression vector and cells, b-catenin is present in the cytoplasm at low levels evaluated the resulting effect on the expression of the due to its phosphorylation and its subsequent protea- endogenous DNp63. As shown in Figure 2a, an increase some-dependent degradation (Novak and Dedhar, 1999; of the endogenous DNp63 expression was observed at Ninomiya et al., 2000; Kudo et al., 2007). In tumours, the mRNA level as compared with control cells, deregulation of the Wnt pathway leads to increased indicating that DNp63a could activate its own promo- stability of b-catenin, which accumulates and translo- ter. Then, ChIP assay was used to evaluate the ability of cates into the nucleus (Korinek et al., 1997; Morin et al., DNp63a to bind to P2 promoter. A transfected wild- 1997). Nuclear b-catenin binds to members of the TCF/ type p53 and the promoter of WAF1/CIP1 (a well- LEF family, resulting in the transactivation of characterized p53-target gene), on which DNp63a is able containing TBE in their promoter, such as CCND1 to bind, were used as positive controls; the GAPDH (Shtutman et al., 1999), AXIN2 (Jho et al., 2002; Lustig promoter was used as a negative one. As expected, in et al., 2002) or PITX2, a paired-related gene p53-transfected Hep3B cells, both WAF1 and P2 fragments required for the development of several tissues (Kioussi were specifically amplified from the fraction immuno- et al., 2002; Ai et al., 2007). precipitated with the anti-p53 antibody (Figure 2b, left

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4658 5 p = 0.02 4 25 3 Control Ig IP expression 2 20 Anti-p63 IP

Np63 1

 15 Relative endogenous 0 Exogenous - + 10 ΔNp63α 5 α-p53 IgG Input α

-p63 IgG Input Fold enrichment / Control Ig 0 TP63 P2 promoter TP63 P2 WAF1 GAPDH promoter promoter promoter WAF1 promoter

GAPDH promoter

p53RE - - -2 928/+139 +- - +

- - -740/+139 + - -+ p = 0.04 p = 0.008 - - -740/+139 p53RE +- -+

- - -404/+139 +- -+

0 0.5 1 1.5 2 Relative luciferase activity

CAAT boxes - - -740/+139 +- - +

- - -740/+139 CAAT2 +- - +

- - -740/+139 CAAT3 +- -+

- - -740/+139 CAAT2/3 +- -+

0 1 2 3 4 Relative luciferase activity Figure 2 DNp63a protein binds to P2 promoter and regulates its expression via the CAAT boxes and independently of the p53RE. Experiments were performed in the p53-null Hep3B cell line transfected with either a wild-type p53- or a DNp63a-expression vector. (a) Reverse transcription–qPCR analysis of endogenous DNp63 expression 48 h after transfection with 5 mgofDNp63a-expression vector by using primers amplifying the DNp63 50UTR. Results are expressed as endogenous DNp63 expression in the presence of exogenous DNp63a ( þ ) relative to empty vector (À) and are mean±s.e.m. from three independent experiments. (b) Binding of p53 (positive control) or DNp63a to TP63 P2 and WAF1 promoters, assessed by ChIP assay using anti-p53 (a-p53) or anti-p63 (a-p63) polyclonal antibodies for p53- and DNp63a-transfected cells, respectively. IgG, non-immune serum. Input, 4% of total DNA extracted before immunoprecipitation. GAPDH promoter amplification was used as negative control. (c) Quantification of DNp63a binding on TP63 P2, WAF1 and GAPDH promoters assessed by ChIP assay and qPCR. Results are presented as fold enrichment with an anti-p63 antibody relative to control immunoglobulin (see materials and methods for details). (d, e) Ability of exogenous DNp63a protein to activate various TP63 P2 promoter constructs. Luciferase activity is presented as fold induction with p53 or DNp63a ( þ ) relative to the empty vector (À) and error bars mean±s.e.m. from three independent experiments.

panel). A similar amplification was also obtained in et al., 2005), these results demonstrated that exogenous DNp63a-transfected Hep3B cells, with an antibody against DNp63a binds both to WAF1 and P2 promoters, the all p63 isoforms (Figure 2b, right panel). The latter result latter suggesting that DNp63 expression might be was confirmed by quantitative PCR (qPCR) (Figure 2c). controlled by an autologous feedback mechanism. As the TAp63 isoform is expressed at very low levels in We next determined the resulting effect of DNp63a the Hep3B cell line under basal conditions (Petitjean binding on its own promoter and localized the involved

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4659 binding sites by testing full-length P2 promoter or DNp63 expression is activated by b-catenin deletion mutants by luciferase assay. Although p53 was Two putative TBEs have been identified in our in silico able to repress the activity of P2 promoter, as already analysis, thus suggesting a cross-talk between DNp63 described (Lanza et al., 2006), an enhanced P2 activity and the Wnt/b-catenin pathway. To test this hypothesis, was observed in the presence of DNp63a (Figure 2d). we overexpressed an exogenous stabilized b-catenin However, these effects did not require the presence of (DN89 deletion mutant) and then quantified DNp63 the putative p53RE, as p53RE-deleted (À740/ þ 139 Dp expression by reverse transcription–qPCR. Endogenous 53RE) and -truncated (À404/ þ 139) P2 fragments were DNp63 expression was significantly upregulated (up to still activated by DNp63a. Furthermore, although the 25-fold) by DN89 b-catenin (Figure 3a) more strongly À740/ þ 139- and À404/ þ 139-truncated fragments dif- than induction of the b-catenin-target gene AXIN2 fer by the presence of a STAT3 RE required for the (Supplementary Figure 2). This result was in agreement self-activation of DNp63 expression (Chu et al., 2006; with luciferase assay, in which the increased expression Chu et al., 2008), they exhibit similar levels of activity. of a P2-driven luciferase gene was observed in the Actually, DNp63a induction was still observed with a presence of this b-catenin mutant (Supplementary deletion mutant lacking the distal 75% of the P2 Figure 3A). Moreover, in ChIP assay, endogenous promoter, but retaining a proximal segment that stabilized b-catenin, but not the wild-type one, was contains the CAAT boxes, the SP1 site and the TATA found to bind to the P2 promoter (Figure 3b). Indeed, box (À404/ þ 139 fragment). qPCR analysis of ChIP assays revealed that immuno- Lanza et al. (2006) identified the CAAT boxes and not precipitation with an anti-b-catenin antibody results in the p53RE as the DNA sequences of P2 promoter essen- an enrichment of P2 sequences of 10-, two- and one fold tial for p53 repression. We thus tested whether the CAAT (for HepG2, Hep3B and HuH7, respectively) compared boxes are required for DNp63-mediated activation of the with immunoprecipitation with non-specific immuno- P2 promoter. Deletion of both the CAAT boxes 2 and 3 globulin G (data not shown). Finally, knocking down resulted in a modest decrease in both p53 repression and b-catenin expression by siRNA resulted in a statistically DNp63 activation of the P2 promoter (Figure 2e). significant reduction of DNp63, both at mRNA (30%) Overall, our results indicate that TP63 P2 promoter and protein levels (Figure 3c). In agreement with these activation by DNp63a is independent of the identified results, b-catenin extinction also led to a decreased P2 p53RE and probably through the CAAT boxes. promoter activity, as shown by luciferase assay (Sup-

a b ChIP assay c 35 1.2 p=0.03 p=0.01 p=0.0006 30 p=0.05 p=0.002 Control 1.0 siRNA β-catenin β-catenin 25 β-catenin β-catenin - - status: α IgG Input α IgG Input siRNA 1 expression 0.8 20 β HepG2 -catenin stabilized siRNA 2 Np63 15 Hep3B 0.6  10 wild-type HuH7 TP63 P2 promoter GAPDH promoter 0.4 5 Relative

0 expression Relative mRNA 0.2 Western Blot ΔN89 2 β-catenin - 0.0 Endogenous HepG Hep3B HuH7 Np63 AXIN2 β-catenin FL FL β-catenin Exogenous ΔN siRNA ΔNp63 ΔN89 Ku80 β-catenin Ku80 mock control 1 2 p63 β-catenin Ku80

Figure 3 DNp63 is a b-catenin-target gene. (a) Reverse transcription–qPCR and western blot analyses of DNp63 expression in the presence of increasing amounts (0, 5, 10, 15 or 20 mg) of DN89 b-catenin mutant in wild-type b-catenin HuH7 HCC cells. qPCR results are expressed as DNp63 expression in the presence of stabilized b-catenin relative to empty vector (À) and are mean±s.e.m. from three independent experiments. (b) b-catenin binding to TP63 P2 promoter in various HCC cell lines differing in their b-catenin status, by ChIP assay using an anti-b-catenin antibody (a-b-catenin). IgG, non-specific antibody. Input, 4% of total DNA extracted before immunoprecipitation. GAPDH promoter amplification was used as negative control. The intracellular level of endogenous b-catenin in all the three cell lines used was showed. (c) Reverse transcription–qPCR analysis of endogenous DNp63 and AXIN2 expression, after silencing of b-catenin expression by RNA interference, in the human oesophageal SCC TE-10 cell line containing a stabilized b-catenin (data not shown). Results are expressed as DNp63 or AXIN2 expression in the presence of b-catenin-specific siRNA relative to control siRNA and are mean±s.e.m. from three independent experiments. The efficiency of b-catenin siRNA1 and 2 was evaluated by western blot. Ku80, loading control for western blot; FL, full-length b-catenin; DN, endogenous truncated b-catenin; DN89, exogenous b-catenin mutant. Mock, non-transfected cells; control, irrelevant siRNA.

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4660 plementary Figure 3B). Altogether, these results are ence on P2 promoter activation by b-catenin (Figure 4a). consistent with the identification of DNp63 as a novel However, the expression of a TCF4 dominant-negative b-catenin-target gene. It should be noted that in vitro, variant (DNTCF4) reduced the effect of b-catenin on b-catenin is able to activate DNp63 expression in both DNp63 expression in a manner similar to its effect the cell types used, that is, HCC and SCC. on AXIN2 expression (Figure 4b). This raised the We next investigated whether the b-catenin-driven possibility that P2 promoter activation involves non- effect was mediated through the two canonical TBEs canonical TBEs (Supplementary Figure 1). Actually, the located within the P2 promoter. Neither the deletion of combined deletion of the proximal TBE and of non- the distal site (À740/ þ 139 fragment) nor the mutation canonical TBEs (À740/ þ 139 DTBE2,7 and À740/ þ 139 or deletion of the proximal one (À740/ þ 139 TBE2mut DTBE2,3,4,5,7 fragments) significantly decreased repor- and À740/ þ 139 DTBE2 fragments) had a significant influ- ter activation (Figure 4c). These results highlight a

TBE TBE 2 Stabilized (-2 083) (-340) β-catenin - -2 928/+139 + - -740/+139 + - -740/+139 TBE2mut +  -

-740/+139 TBE2 p = 0.02 + - -2 928/-1 900 + 0 2 4 6 8 10 Relative luciferase activity

1.5 p=0.04 Np63 AXIN2 p=0.08

1

0.5 Relative mRNA expression 0 Stabilized + + + β-catenin ΔNTCF4 -

Stabilized β-catenin 1 2 345 6 7 - -740/+139 +  - -740/+139 TBE2 + -  -740/+139 TBE7 + p = 0.03

- p = 0.02  -740/+139 TBE2,7 + -  -740/+139 TBE2,3,4,5,7 +

0 5 10 15 20 Relative luciferase activity Figure 4 DNp63 expression is activated by b-catenin through TCF factors and multiple TBEs. (a, c) Ability of DN89 b-catenin variant to modulate the activity of TP63 P2 reporter constructs harbouring deletion or mutation of the canonical (a) or non-canonical (c) TBEs, as depicted on the left. The experiments were performed in the non-activated b-catenin-HuH7 cell line. Luciferase activity is presented as fold induction with b-catenin ( þ ) relative to empty vector (À) and are mean±s.e.m. from three independent experiments. Basal activity of the different constructs is showed in Supplementary Figure 4. (b) A transdominant-negative variant of TCF4 (DNTCF4) alleviates b-catenin-mediated activation of DNp63 and AXIN2. Hep3B cells were cotransfected with 10 mgofDN89 b-catenin and 0, 5 or 10 mgofDNTCF4-expression vectors, and total RNA extracted 24 h later. Results of reverse transcription–qPCR are expressed as DNp63 or AXIN2 expression in the presence of DNTCF4 relative to empty vector (À) and are mean±s.e.m. from three independent experiments. 1, 3–7, non-canonical TBEs; 2, canonical TBE.

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4661 b-catenin-mediated effect through synergistic action of Concordance between deregulation of b-catenin and several binding sites rather than through a single DNp63 in human tumours canonical one. Nevertheless, P2 promoter also contains The results above suggest that activation of b-catenin four potential Pitx2-binding sites. PITX2 has been may contribute to the increased DNp63 expression in reported as a b-catenin-target gene (Kioussi et al., tumours. In tumours, the presence of an activated 2002). Moreover, in the presence of b-catenin, Pitx2 b-catenin is revealed by changes in its localization, protein is converted from repressor to activator in a that is, a loss of membranous staining and a gain of manner similar to TCF/LEF. Therefore, we tested cytosolic and sometimes nuclear staining. To evaluate whether DNp63 expression is regulated by Pitx2. For the relevance of the b-catenin–DNp63 cross-talk in this, we silenced Pitx2 and b-catenin, separately or cancer, we determined whether the b-catenin delocaliza- together, in the Hep3B cell line. We first verified the tion is frequently associated with overexpression of silencing efficiency both at mRNA (Figure 5a) and DNp63 in human oesophageal SCC, a type of cancer in protein (Figures 5b and c) levels and then checked for which overexpression of DNp63 is a common event. In the expression of DNp63. The extinction of PITX2 normal oesophageal epithelium, a membrane-restricted expression led to a significant decrease of DNp63, b-catenin and a nuclear DNp63 staining were found in similar to the one obtained with b-catenin inhibition cells of basal and parabasal layers (Figure 6, upper panel). (Figure 5d). No synergistic effect was observed with the A total of 43 oesophageal SCC were tested for b-catenin double knockdown, in agreement with the cross-talk and DNp63 staining by . Strong between Pitx2 and b-catenin already described (Kioussi nuclear accumulation of DNp63 or change in b-catenin et al., 2002). localization was each observed in 35 cases, 33 of them Overall, our results identify DNp63 as a novel b-catenin- showing both events (Table 1). Thus, about 76% of target gene and demonstrate the implication of both oesophageal SCC displayed a concordant alteration of canonical and non-canonical TBEs and of Pitx2 in the b-catenin localization and strong DNp63 expression, b-catenin-mediated activation of DNp63 expression. as illustrated in Figure 6, middle and lower panels, strengthening the notion that alteration of the Wnt/b- siRNA 1.2 catenin pathway might contribute to DNp63 upregulation. p=0.06 mock Pitx2 Control 1 Pitx2 0.8 Ku80 expression 0.6 siRNA -catenin p63 Pitx2 0.4 staining staining 0.2 sp c sp

Relative mock Control 0 β Normal -catenin b mock Pitx2 Control b epithelium c siRNA Actin

2 p=0.01

1.5 p=0.04 expression 1 Squamous Cell Np63  Carcinoma 0.5 Relative 0 st st mock β-catenin 3 Pitx2 β-catenin 3 Control + Pitx2 siRNA Figure 6 DNp63 and b-catenin co-localize in human oesophageal Figure 5 Variation of DNp63 expression after b-catenin and/or SCCs. Immunohistochemistry of p63 and b-catenin in normal and Pitx2 siRNA silencing. Hep3B cells were transfected with siRNA tumour oesophageal epithelium. In normal epithelium (upper targeting b-catenin and/or Pitx2. (a) The anti-Pitx2 siRNA panel), b-catenin staining is detected at the membrane of cells efficiency was evaluated both by reverse transcription–qPCR and from basal and parabasal layers (b), but not in the differentiated (b) by western blot. Two different Pitx2 siRNAs were tested but cells of the suprabasal layers (sp). p63 exhibits a strong nuclear only the most efficient one is shown. (c) The b-catenin siRNA3 staining in cells from basal and parabasal layers that dramatically efficiency was evaluated by western blot. (d) DNp63 expression was decreases in the suprabasal layers. None of them are expressed evaluated by reverse transcription–qPCR. Results are expressed as in the conjunctive tissue (c). In SCC (middle and lower panels), DNp63 expression in transfected cells relative to the mock most of tumour cells present a loss of membranous staining condition after an internal normalization on reference gene values associated with a gain of cytosolic and sometimes nuclear staining and error bars mean±s.e.m. of three independent experiments. for b-catenin and a strong nuclear staining for p63, whereas Mock, non-transfected cells; control, irrelevant siRNA. Ku80 and stromal cells (st) are negative for both proteins. Magnification, Actin, loading controls for western blot. upper and lower panels, Â 200; middle panel, Â 100.

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4662 Table 1 p63 and b-catenin staining in oesophageal SCC tumour and DNp63 expression. The specificity of DNp63 samples activation through the b-catenin pathway was confirmed b-catenin staining Total both by silencing b-catenin expression and by inhibiting the b-catenin–TCF transcription factor interaction. Positive Negative However, promoter-deletion studies showed that the effect of b-catenin was mediated through multiple p63 staining Positive 33 2 35 canonical and non-canonical TBEs rather than a single Negative 2 6 8 canonical one. The presence of multiple functional TBEs Total 35 8 43 in some putative b-catenin-target genes as AXIN2 has been previously documented (Hatzis et al., 2008), thus Abbreviation: SCC, . supporting our results. P 5.5 10À6. ¼ Â We observed that DNp63 expression was not com- pletely abolished in the presence of DNTCF4 and that Discussion the decrease of P2 promoter-deletion mutant activity only reached a maximum of 60%. This suggests that There is growing evidence for a major role of DNp63 in other transcription factors remained active. One possi- tumour progression. Indeed, beyond its dominant- ble candidate was Pitx2, as four Pitx2-binding sites are negative effect on p53 suppressor function, DNp63 was present in the P2 promoter, which were not removed in found to favour angiogenesis and chemoresistance the deletion mutants. More than a direct b-catenin- (Senoo et al., 2002; Wu et al., 2005; Rocco et al., target gene, Pitx2 is also a transcription factor that is 2006) and to regulate cell-adhesion processes (Carroll converted from repressor to activator in the presence of et al., 2006; Yang et al., 2006). Conversely, several b-catenin. Indeed, the extinction of PITX2 expression by studies reported a loss of p63 expression in metastatic RNA interference resulted in a decrease of DNp63 stages (Barbieri et al., 2006; Adorno et al., 2009). These expression similar to the one obtained upon b-catenin observations suggest a dual role of DNp63 in tumouri- silencing. Moreover, the double silencing of b-catenin genesis with enhanced expression in the early stages and PITX2 expression did not provide any additional of progression during which the epithelial cell fate effect, indicating that DNp63 expression required both is maintained, followed by loss of expression when of these two proteins. Overall, the converging data that cells acquire mesenchymal properties associated with we obtained by various technical approaches indicate metastatic potential. that DNp63 is a robust transcriptional target of the Given this context, identifying the transcription Wnt/b-catenin pathway through both TCF/LEF and factors that may contribute either to enhance or repress Pitx2 transcription factors. DNp63 expression is of particular interest. To this end, Two reports about the effect of DNp63 on b-catenin an in silico analysis of the P2 promoter that controls accumulation and activity have been published. One DNp63 expression was performed and one p53RE and study suggests that overexpressed DNp63 is able to two TBEs were identified. We next determined whether induce b-catenin accumulation by interacting with the either DNp63 itself or b-catenin could regulate the complex responsible for b-catenin phosphorylation P2 promoter and whether their effects involved the (Patturajan et al., 2002). The second one reports an identified binding sites. increase of Wnt/b-catenin activity in the presence of We observed a regulation of the P2 promoter by overexpressed exogenous DNp63, but an inhibition DNp63a itself, which does not implicate the p53RE. with endogenous DNp63a, without detectable change Both negative and positive regulation were reported in b-catenin level or phosphorylation (Drewelus et al., (Harmes et al., 2003; Waltermann et al., 2003; Chu 2010). We were not able to detect any increase in et al., 2006; Romano et al., 2006), but our results argue b-catenin activity after DNp63 overexpression in our in favour of a positive regulation of P2 by DNp63a. experiments (data not shown). Altogether, these appar- It should be noted that activation of P2 by DNp63a was ent contradictory results strongly suggest that DNp63– observed in cell lines lacking p53, demonstrating that b-catenin cross-talk may depend on experimental this activation is not due to a dominant-negative effect models and conditions. of DNp63a on p53-dependent P2 repression. The fact All our analyses were performed in two different that both factors exert independent and opposite effects epithelial contexts represented by HCC and SCC cell leads to multiple regulatory possibilities, complicated lines, as 20–30% of these two types of tumours show by the observation that P2 promoter may be repressed alterations in the Wnt/b-catenin pathway (Ninomiya in the presence of small amounts of DNp63a and et al., 2000; Kudo et al., 2007; Cavard et al., 2008). activated with increasing amounts (Chu et al., 2006). We equally used HCC and SCC cell lines with different This high complexity may explain the controversy b-catenin status and did not observe any discrepancy highlighted above. between the two models. But, most of the experiments The most novel observation in our in silico study is the were performed with high amounts of exogenous presence of two canonical TBEs within the P2 promoter. protein. This approach can provide important informa- Therefore, we checked whether DNp63 expression was tion on molecular mechanisms but, as discussed above, regulated by b-catenin. We showed that overexpression does not reflect in vivo conditions. For this reason, we of a stabilized b-catenin increases P2 promoter activity searched whether the deregulation of the b-catenin

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4663 Normal cells epithelial stem cells, a process in which both factors have p53 -catenin important roles. In a physiological context, b-catenin and

Np63 promoter DNp63 regulate proliferation or differentiation of epider- mal and mammary progenitor/stem cells (Senoo et al., 2007; Ambler and Maatta, 2009; Pece et al., 2010). Np63 In tumours, they have been implicated in the maintenance Tumour cells of the stemness of cancer stem cells (Katoh and Katoh, HCC SCC 2007; Zucchi et al., 2008). These observations suggest that the cross-talk between DNp63 and b-catenin may represent -catenin -catenin a major regulator checkpoint in the initiation, progression -catenin p53 -catenin p53 -catenin p53 ? -catenin and dissemination of tumours of epithelial origin. -catenin ? ?

Materials and methods Np63 Np63 Np63   Np63 Np63 Plasmids and mutagenesis Figure 7 Proposed model of DNp63 regulation by p53, DNp63 DNp63a, DN89 b-catenin mutant and DNTCF4-expression and b-catenin. vectors have been previously described (Korinek et al., 1997; Cavard et al., 2006; Petitjean et al., 2008). pCMV–p53- expression vector was from Clontech (Mountain View, CA, pathway may contribute to sustain DNp63 expression USA). The pGL3-P2 luciferase reporter plasmid encompassing in vivo. b-catenin is accumulated in about 30% of HCC the –2928/ þ 139 TP63 P2 promoter sequences was kindly (Cavard et al., 2008). However, we found, by both DNA provided by Dr A Constanzo (Lanza et al., 2006). Deletions and mutations of p53RE, TBEs and CAAT boxes were and tissue microarray studies, that DNp63 overexpres- performed by using the Quick change II mutagenesis sion is a very rare event in HCC (data not shown), (Agilent, Massy, France). Primer sequences are described in suggesting that the frequent alterations of the b-catenin Supplementary Table I. pathway do not lead to DNp63 overexpression in this type of tumour. As b-catenin alterations and TP53 gene Cell culture mutations are often mutually exclusive in HCC (Laur- Hep3B and HepG2 HCC cells were cultured as previously ent-Puig and Zucman-Rossi, 2006), we may hypothesize described (Petitjean et al., 2005). HuH7 HCC cells were that, in HCC cells, DNp63 activation by b-catenin is cultured in RPMI medium supplemented with penicillin/ counteracted by the repression exerted by wild-type p53. streptomycin/glutamine, 10% foetal calf serum, 1% non- Oesophageal SCC frequently exhibit Wnt/b-catenin essential amino acids and 0.01 M Hepes Buffer (Life Techno- pathway alterations, TP53 mutation and DNp63 expres- logies, Illkirch, France). TE-6 and TE-10 oesophageal SCC cell sion (Ninomiya et al., 2000; Taniere et al., 2001; Kudo lines were cultured in Dulbecco’s modified Eagle medium et al., 2007). Interestingly, of 43 tumours without TP63 supplemented with penicillin/streptomycin/glutamine and 10% gene amplification analysed, 33 exhibited a concordant foetal calf serum (Life Technologies). References and p53 and b-catenin status of cell lines are reported in Supplementary expression of DNp63 and delocalization of b-catenin. Table II. These results provide pathological support in favour of a role of b-catenin in DNp63 expression in SCC. However, no correlation between these two alterations Transient transfections and the p53 status (wild type, deleted or mutated) was Expression vectors were introduced by using Lipofectamine 2000 (Life Technologies) and siRNA with Hiperfect trans- observed, supporting a mechanism independent of the fection reagent (Roche, Meylan, France) or Lipofectamine p53 status in this tumour type. Thus, our results suggest RNAiMax (Life Technologies). first, a link between DNp63 and b-catenin by placing RNA interference experiments were performed by transfect- DNp63 downstream from the Wnt/b-catenin pathway ing 2 Â 105 TE-10, HepG2 or Hep3B cells with 6 pmol of and second, a cell type-dependent balance between p53 siRNA. Total RNA was extracted 24 or 48 h later with and b-catenin opposite effects on DNp63 expression. TriREAGENT (Sigma, Lyon, France), according to manu- In conclusion, our results show that expression of facturer’s conditions. The following siRNA oligonucleotides 0 DNp63 is dependent upon the status of two main were used: control siRNA (sense) 5 -GGCAUAGAUGUAGC UGUAA-30, b-catenin siRNA 1 (sense) 50-GGUGCUAUCU proteins, p53 and b-catenin (Figure 7). In normal 0 differentiated cells, the presence of a functional p53 GUCUGCUCUAGUAAU-3 , b-catenin siRNA 2 (sense) 50-CCACAGCUCCUUCUCUGAGUGGUAA-30, b-catenin and a non-activated Wnt/b-catenin pathway should lead siRNA 3 (sense) 50-AGCUGAUAUUGAUGGACAG-30,Pitx2 to the inhibition of DNp63 expression. In contrast, in siRNA (sense) 50-CCAGGCUAUUCCUACAACA-30. tumour cells containing a non-functional p53 and an activated b-catenin pathway, DNp63 expression may Luciferase assays increase and thus favour tumour progression. It must be Luciferase assays were performed as previously described considered that this scenario is strongly influenced by (Petitjean et al., 2008), by using 1.8 mg of p53, DNp63a, wild- the type of cell and tumour. In any case, the cross-talk type or mutant b-catenin-expression vectors, 1 mg of luciferase between b-catenin and DNp63 may be of major reporter plasmid, and 0.1 mg of renilla luciferase control relevance in the maintenance of normal and cancer plasmid.

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4664 Western blot input. Then, the resulting values obtained for the anti-p63 Cell pellets were lysed on ice for 30 min in NP-40 buffer (50 mM antibody were normalized to those obtained for the control Tris, pH 8.0; 150 mM NaCl; 5 mM EDTA; 1% NP-40) with Immunoglobulin conditions. Primers and conditions are de- protease inhibitors (Complete mini, EDTA free, Roche). scribed in Supplementary Table III. Protein lysates were collected after a 15-min centrifugation at 15 700 g at 4 1C. Then, proteins were separated by electro- phoresis on an SDS–polyacrylamide (30:0.8, Euromedex, Immunohistochemistry Souffelweyersheim, France) gel, transferred onto a PVDF Immunohistochemistry was performed as previously described membrane (Biorad, Marnes-la-Coquette, France) and probed (Taniere et al., 2001). Anti-p63 (clone AB1, 1/200, Fisher overnight with antibodies against p53 (Clone DO7, 1/600, Scientific, Illkirch, France) and anti-b-catenin (clone E-5, Dako, Versailles, France), p63 (4A4, 1/500, Santa Cruz, 1/2000, Santa Cruz) monoclonal antibodies were used. Heidelberg, Germany), b-catenin (Clone 14, 1/500, BD Bioscience, Le Pont de Claix, France) and Pitx2 (C16, 1/600, Santa Cruz) proteins. Membranes were then incubated with a Abbreviations peroxidase-conjugated secondary antibody against rabbit (1/5000, Jackson, Suffolk, UK), goat (1/3000, Dako) or mouse CBF, CAAT-binding factor; HCC, hepatocellular carcinoma; (1/3000, Dako). The reaction was revealed by chemilumines- PBS, Pitx2-binding site; RE, responsive element; SCC, squamous cence with the ECL kit (Santa Cruz). cell carcinoma; TA, transcriptionally active; TBE, TCF/LEF- binding element; UTR, untranslated region. qPCR In all, 2 mg of RNA retrotranscribed with random hexamers and Superscript II enzyme (Life Technologies) were four-fold Conflict of interest diluted and used as template for PCR amplification using the iTaq SYBR Green Supermix with ROX (Biorad) in an ABprism The authors declare no conflict of interest. qPCR apparatus (Applied Biosystems, Cergy, France) or with the QuantiFast SYBR Green Mix (Qiagen, Courtaboeuf, France) in a Light Cycler 480 apparatus (Roche). Primers and PCR Acknowledgements conditions are described in Supplementary Table III. We thank Drs W Dinjens (University Medical Center, ChIP assays Rotterdam, The Netherlands) and A Costanzo (University ChIP assay was done as described elsewhere (Auriol et al., 2005). of Rome, Italy) for the generous gift of TE-10 cells and pGL3- Nuclear protein extract was incubated overnight at 4 1Cwith4mg P2 plasmid, respectively; Drs J Zucmann-Rossi (INSERM of 31.4 (home-made anti-p53 rabbit polyclonal serum), anti-b- U674, Paris, France) and B Terris (Institut Cochin, Paris, catenin clone 14 monoclonal antibody (BD Biosciences) or anti- France) for the analysis of DNp63 expression in HCC samples; p63 H137 rabbit polyclonal antibody (SC-2044, Santa Cruz). A Dr B Abedi for her help for IHC experiments; Dr G Martel- p63 pan antibody was used because TAp63 protein is not Planche and A Masquelet and Ms A Durand and A Charnay detectable in the Hep3B cell line used in ChIP assay (Petitjean for technical assistance; Drs S Sentis, G Hinkal and S and et al., 2005). Immune complexes were then collected by adding D Cox for critical reading of the manuscript. This project was 60 ml of salmon sperm DNA—protein A or G—agarose or supported by INSERM and the Association pour la Recherche sepharose beads for 1 h. PCR amplifications were performed contre le Cancer (grant number 3117). CR and HS were using the Platinum Taq DNA polymerase (Life Technologies). funded by the French Ligue Nationale Contre le Cancer, AP qPCR were made using the QuantiFast SYBR Green Mix by the French Ministe` re de l’Enseignement et de la Recherche, (Qiagen) in a Light Cycler 480 apparatus (Roche). First, for each and ET, PT, and VT by the International Agency for Research promoter, the fold enrichment was calculated relative to the on Cancer.

References

Adorno M, Cordenonsi M, Montagner M, Dupont S, Wong C, Cavard C, Colnot S, Audard V, Benhamouche S, Finzi L, Hann B et al. (2009). A Mutant-p53/Smad complex opposes p63 to Torre C et al. (2008). Wnt/beta-catenin pathway in hepatocellular empower TGFbeta-induced metastasis. Cell 137: 87–98. carcinoma pathogenesis and liver physiology. Future Oncol 4: Ai D, Wang J, Amen M, Lu MF, Amendt BA, Martin JF. (2007). 647–660. Nuclear factor 1 and T-cell factor/LEF recognition elements Cavard C, Terris B, Grimber G, Christa L, Audard V, Radenen- regulate Pitx2 transcription in pituitary development. Mol Cell Biol Bussiere B et al. (2006). Overexpression of regenerating islet-derived 27: 5765–5775. 1 alpha and 3 alpha genes in human primary liver tumors with beta- Ambler CA, Maatta A. (2009). Epidermal stem cells: location, catenin mutations. Oncogene 25: 599–608. potential and contribution to cancer. J Pathol 217: 206–216. Chu WK, Dai PM, Li HL, Chen JK. (2008). Transcriptional activity of Auriol E, Billard LM, Magdinier F, Dante R. (2005). Specific binding the DeltaNp63 promoter is regulated by STAT3. J Biol Chem 283: of the methyl binding domain protein 2 at the BRCA1-NBR2 . 7328–7337. Nucleic Acids Res 33: 4243–4254. Chu WK, Lee KC, Chow SE, Chen JK. (2006). Dual regulation of the Barbieri CE, Tang LJ, Brown KA, Pietenpol JA. (2006). Loss of p63 DeltaNp63 transcriptional activity by DeltaNp63 in human leads to increased cell migration and up-regulation of genes involved nasopharyngeal carcinoma cell. Biochem Biophys Res Commun in invasion and metastasis. Cancer Res 66: 7589–7597. 342: 1356–1360. Carroll DK, Carroll JS, Leong CO, Cheng F, Brown M, Mills AA Drewelus I, Gopfert C, Hippel C, Dickmanns A, Damianitsch K, et al. (2006). p63 regulates an adhesion programme and cell survival Pieler T et al. (2010). p63 antagonizes Wnt-induced transcription. in epithelial cells. Nat Cell Biol 8: 551–561. 9: 580–587.

Oncogene b-catenin-DNp63 axis in tumour progression C Ruptier et al 4665 Flores ER, Sengupta S, Miller JB, Newman JJ, Bronson R, Crowley D Patturajan M, Nomoto S, Sommer M, Fomenkov A, Hibi K, et al. (2005). Tumor predisposition in mice mutant for p63 and p73: Zangen R et al. (2002). DeltaNp63 induces beta-catenin nuclear evidence for broader tumor suppressor functions for the p53 family. accumulation and signaling. Cancer Cell 1: 369–379. Cancer Cell 7: 363–373. Pece S, Tosoni D, Confalonieri S, Mazzarol G, Vecchi M, Ronzoni S Flores ER, Tsai KY, Crowley D, Sengupta S, Yang A, McKeon F et al. (2010). Biological and molecular heterogeneity of breast et al. (2002). p63 and p73 are required for p53-dependent apoptosis cancers correlates with their cancer stem cell content. Cell 140: in response to DNA damage. Nature 416: 560–564. 62–73. Harmes DC, Bresnick E, Lubin EA, Watson JK, Heim KE, Curtin JC Petitjean A, Cavard C, Shi H, Tribollet V, Hainaut P, Caron de FC. et al. (2003). Positive and negative regulation of deltaN-p63 (2005). The expression of TA and DeltaNp63 are regulated by promoter activity by p53 and deltaN-p63-alpha contributes different mechanisms in liver cells. Oncogene 24: 512–519. to differential regulation of p53 target genes. Oncogene 22: Petitjean A, Ruptier C, Tribollet V, Hautefeuille A, Chardon F, 7607–7616. Cavard C et al. (2008). Properties of the six isoforms of p63: p53-like Hatzis P, van der Flier LG, van Driel MA, Guryev V, Nielsen F, regulation in response to genotoxic stress and cross talk with Denissov S et al. (2008). Genome-wide pattern of TCF7L2/TCF4 DeltaNp73. Carcinogenesis 29: 273–281. chromatin occupancy in colorectal cancer cells. Mol Cell Biol 28: Rocco JW, Leong CO, Kuperwasser N, DeYoung MP, Ellisen LW. 2732–2744. (2006). p63 mediates survival in squamous cell carcinoma by Jho EH, Zhang T, Domon C, Joo CK, Freund JN, Costantini F. suppression of p73-dependent apoptosis. Cancer Cell 9: 45–56. (2002). Wnt/beta-catenin/Tcf signaling induces the transcription of Romano RA, Birkaya B, Sinha S. (2006). Defining the regulatory Axin2, a negative regulator of the signaling pathway. Mol Cell Biol elements in the proximal promoter of DeltaNp63 in keratinocytes: 22: 1172–1183. potential roles for Sp1/Sp3, NF-Y, and p63. J Invest Dermatol 126: Katoh M, Katoh M. (2007). and stem cell 1469–1479. signaling network. Clin Cancer Res 13: 4042–4045. Senoo M, Matsumura Y, Habu S. (2002). TAp63gamma (p51A) and Kioussi C, Briata P, Baek SH, , Rose DW, Hamblet NS, Herman T dNp63alpha (p73L), two major isoforms of the p63 gene, exert et al. (2002). Identification of a Wnt/Dvl/beta-Catenin —4 Pitx2 opposite effects on the vascular endothelial (VEGF) pathway mediating cell-type-specific proliferation during develop- . Oncogene 21: 2455–2465. ment. Cell 111: 673–685. Senoo M, Pinto F, Crum CP, McKeon F. (2007). p63 is essential Korinek V, Barker N, Morin PJ, van WD, de WR, Kinzler KW for the proliferative potential of stem cells in stratified epithelia. et al. (1997). Constitutive transcriptional activation by a beta- Cell 129: 523–536. catenin-Tcf complex in APCÀ/À colon carcinoma. Science 275: Shtutman M, Zhurinsky J, Simcha I, Albanese C, D’Amico M, 1784–1787. Pestell R et al. (1999). The D1 gene is a target of the beta- Kudo J, Nishiwaki T, Haruki N, Ishiguro H, Shibata Y, Terashita Y catenin/LEF-1 pathway. Proc Natl Acad Sci USA 96: 5522–5527. et al. (2007). Aberrant nuclear localization of beta-catenin without Taniere P, Martel-Planche G, Saurin JC, Lombard-Bohas C, Berger F, genetic alterations in beta-catenin or Axin genes in esophageal Scoazec JY et al. (2001). TP53 mutations, amplification of P63 cancer. World J Surg Oncol 5: 21. and expression of cell cycle proteins in squamous cell carcinoma Lanza M, Marinari B, Papoutsaki M, Giustizieri ML, D’Alessandra Y, of the oesophagus from a low incidence area in Western Europe. Chimenti S et al. (2006). Cross-talks in the p53 family: deltaNp63 is an Br J Cancer 85: 721–726. anti-apoptotic target for deltaNp73alpha and p53 gain-of-function Waltermann A, Kartasheva NN, Dobbelstein M. (2003). Differential mutants. Cell Cycle 5: 1996–2004. regulation of p63 and p73 expression. Oncogene 22: 5686–5693. Laurent-Puig P, Zucman-Rossi J. (2006). Genetics of hepatocellular Wu G, Osada M, Guo Z, Fomenkov A, Begum S, Zhao M et al. tumors. Oncogene 25: 3778–3786. (2005). DeltaNp63alpha up-regulates the Hsp70 gene in human Lustig B, Jerchow B, Sachs M, Weiler S, Pietsch T, Karsten U et al. cancer. Cancer Res 65: 758–766. (2002). Negative feedback loop of Wnt signaling through upregula- Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V et al. tion of conductin/axin2 in colorectal and liver tumors. Mol Cell Biol (1998). p63, a p53 homolog at 3q27-29, encodes multiple products 22: 1184–1193. with transactivating, death-inducing, and dominant-negative activ- Marcel V, Hainaut P. (2009). p53 isoforms—a conspiracy ities. Mol Cell 2: 305–316. to kidnap p53 tumor suppressor activity? Cell Mol Life Sci 66: Yang A, McKeon F. (2000). P63 and P73: P53 mimics, menaces and 391–406. more. Nat Rev Mol Cell Biol 1: 199–207. Moll UM, Slade N. (2004). p63 and p73: roles in development and Yang A, Schweitzer R, Sun D, Kaghad M, Walker N, Bronson RT tumor formation. Mol Cancer Res 2: 371–386. et al. (1999). p63 is essential for regenerative proliferation in limb, Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, craniofacial and epithelial development. Nature 398: 714–718. Vogelstein B et al. (1997). Activation of beta-catenin-Tcf signaling Yang A, Zhu Z, Kapranov P, McKeon F, Church GM, Gingeras TR in colon cancer by mutations in beta-catenin or APC. Science 275: et al. (2006). Relationships between p63 binding, DNA sequence, 1787–1790. transcription activity, and biological function in human cells. Mol Ninomiya I, Endo Y, Fushida S, Sasagawa T, Miyashita T, Cell 24: 593–602. Fujimura T et al. (2000). Alteration of beta-catenin expression in Zucchi I, Astigiano S, Bertalot G, Sanzone S, Cocola C, Pelucchi P esophageal squamous-cell carcinoma. Int J Cancer 85: 757–761. et al. (2008). Distinct populations of tumor-initiating cells derived Novak A, Dedhar S. (1999). Signaling through beta-catenin and Lef/ from a tumor generated by rat mammary cancer stem cells. Proc Tcf. Cell Mol Life Sci 56: 523–537. Natl Acad Sci USA 105: 16940–16945.

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

Oncogene