Oncogene (2014) 33, 2674–2680 & 2014 Macmillan Publishers Limited All rights reserved 0950-9232/14 www.nature.com/onc

SHORT COMMUNICATION

Evasion of p53 and G2/M checkpoints are characteristic of Hh-driven basal cell carcinoma ZJ Li1,2, SC Mack1,3,4, TH Mak1, S Angers5,6, MD Taylor1,3,4 and C-C Hui1,2 Basal cell carcinoma (BCC), the most common type of cancer, is characterized by aberrant Hedgehog (Hh) pathway activity. Mutations in pathway components, such as PATCHED1 (PTCH1), are commonly found in BCC. While the tumor suppressor role of PTCH1 in BCC is well established, how Hh pathway activation disrupts normal skin homeostasis to promote BCC formation remains poorly understood. Like Ptc1, Sufu is a major negative regulator of the Hh pathway. Previously, we showed that inactivation of Sufu in the skin does not result in BCC formation. Why loss of Ptc1, but not Sufu, in the epidermis induces BCC formation is unclear. In this report, we utilized gene expression profiling to identify biological pathways and processes that distinguish Sufu from Ptc1 mutants, and discovered a novel role for Sufu in cell cycle regulation. We demonstrated that the Hh pathway activation in Sufu and Ptc1 mutant skin is associated with abnormal cell cycle entry, ectopic expression of D-type cyclins and increased DNA damage. However, despite the presence of DNA damage, p53 stabilization was impaired in the mutant skin. Alternative mechanism to halt genomic instability is the activation of G2/M cell cycle checkpoint, which can occur independent of p53. We found that while Ptc1 mutant cells continue to cycle, which would favor genomic instability, loss of Sufu results in G2/M cell cycle arrest. This finding may explain why inactivation of Sufu is not sufficient to drive BCC formation. Taken together, these studies revealed a unique role for Sufu in G2/M phase progression, and uncovered the molecular and cellular features associated with Hh-driven BCC. Oncogene (2014) 33, 2674–2680; doi:10.1038/onc.2013.212; published online 10 June 2013 Keywords: Sufu; Ptc1; Hedgehog; basal cell carcinoma; cell cycle; p53

INTRODUCTION Basal cell carcinoma (BCC) of the skin, the most common human malignancies, is characterized by genomic instability,1,2 TP53 mutations3 and the Hedgehog (Hh) pathway activation.4,5 Mutations in PATCHED1 (PTCH1), which encodes the Hh-binding receptor and a major negative regulator of the pathway, are the predominant genetic alteration found in human BCC. Consistent with this, inactivation of Ptc1 (the mouse counterpart) in the mouse skin results in BCC formation.6–8 Ptc1 acts negatively by inhibiting the activity of Smoothened. Activation of Smoothened promotes the dissociation of inhibitory Sufu–Gli complexes, allowing Gli transcription factors to translocate into the nucleus for Hh target gene expression.9–11 Among the three Gli proteins, Gli2 is the major mediator of Hh response in the skin,12 and overexpression of Gli1 or Gli2 in the epidermis induces BCC formation.13,14 While it is well established that abnormal Hh pathway activation leads to BCC, the downstream molecular event(s) that lead to epidermal transformation are largely unknown. Several studies indicated that Hh signaling impinges on the cell cycle machinery to promote G1/S transition and to inhibit cell cycle withdrawal. Genetic studies in mice revealed that Hh is a potent mitogen, promoting the expression of cell cycle regulators (D-type cyclins, N-myc and E2Fs) during cerebellar12,15–17 and skin development.12,15 Hh signaling induces the degradation of the tumor suppressor p53 in mouse embryonic fibroblasts18 and

can override differentiation-induced signals for cell cycle arrest in cultured human keratinocytes.19 These in vitro studies suggest that Hh signaling drives tumorigenesis via evasion of p53 tumor surveillance and cell cycle arrest, however, their physiological relevance remains undefined. Like Ptc1, Sufu is a major negative regulator of the Hh pathway. Sufu  /  mice exhibit ectopic Hh pathway activation, and display neural and cardiac defects similar to those of Ptc1  /  mice.20–22 However, SUFU mutations are rare in human BCC and often accompanied by PTCH1 and/or TP53 mutations,3,23 suggesting that SUFU mutations might have different impact from PTCH1 mutations in BCC tumorigenesis. Consistent with this, we recently showed that inactivation of Sufu in the mouse skin does not result in BCC formation.11 In this study, we demonstrated that Hh pathway activation leads to similar molecular defects, including increased D-type cyclin expression, sustained G1/S progression and elevated DNA damage. We utilized the distinction between Ptc1 and Sufu knock out (KO) keratinocytes in their susceptibility to undergo transformation, and performed comparative transcriptome analysis to identify molecular events critical for BCC formation. By Gene Set Enrichment Analysis, we uncovered molecular pathways that are characteristic of Sufu and distinct from Ptc1 mutants, including those involved in DNA repair, chromosomal maintenance and mitosis, suggesting that Sufu has a unique role in cell cycle regulation. Consistent with this prediction, we showed that Sufu, but not Ptc1, KO keratinocytes

1 Program in Developmental & Stem Cell Biology, The Hospital for Sick Children, Toronto, Ontario, Canada; 2Department of Molecular Genetics, University of Toronto, Toronto, Ontario, Canada; 3Division of Neurosurgery, Arthur and Sonia Labatt Brain Tumor Research Centre, The Hospital for Sick Children, Toronto, Ontario, Canada; 4Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada; 5Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada and 6 Department of Biochemistry, University of Toronto, Toronto, Ontario, Canada. Correspondence: Dr C-C Hui, Program in Developmental & Stem Cell Biology, The Hospital for Sick Children, 101 College Street, TMDT East Tower, Room 13-314, Toronto, Ontario M5G 1L7, Canada. E-mail: [email protected] Received 12 December 2012; revised 27 April 2013; accepted 3 May 2013; published online 10 June 2013

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2675 exhibit a cell cycle arrest at G2/M, thus halting tumor progression. Loss of G2/M checkpoint renders p53 response critical for tumor suppression. Interestingly, we found that DNA damage-mediated p53 stabilization is uncoupled in these mutants. Together, these results uncover the importance of p53-mediated response and G2/M checkpoint as critical barriers for Hh-driven BCC. RESULTS AND DISCUSSION Epidermal inactivation of Sufu and Ptc1 leads to similar embryonic skin defects We previously reported that skin from epidermis-specific deletion of Ptc1, but not Sufu, embryos grafted onto the backs of nude mice developed BCC, suggesting that additional genetic alterations are required to transform Sufu-mutant cells into BCC.11 The function of Ptc1 in BCC formation has been extensively examined.6,24–27 To gain insight into the early events of malignancy, we compared the embryonic skin phenotypes of Sufu and Ptc1 mutants. Histological staining revealed strikingly similar epidermal defects between Sufu and Ptc1 mutants at E18.5, which include a reduction in the number of hair follicles (arrows in Figures 1a–c). Both Sufu and Ptc1 mutants displayed compromised epidermal differentiation (Supplementary Figures S1a–f) and ectopic expression of Hh target genes, Gli1 and Ptc1, in the interfollicular epidermis (arrows in Figures 1d–i). It has been suggested that a critical threshold of Hh pathway activity is required to induce BCC formation.28 To determine whether there is a difference in the extent of pathway activation in Sufu and Ptc1 mutant skin, we performed quantitative RT–PCR analysis on keratinocytes isolated from both mutants. Although differential expression of Gli2 and Gli3 proteins was found in Sufu and Ptc1 KO keratinocytes (Supplementary Figures S2a–c), Hh target genes, Gli1 and Ptc2, were upregulated to a similar extent in Sufu and Ptc1 KO keratinocytes compared with control keratinocytes (Supplementary Figure S2d). Furthermore, Sufu and Ptc1 mutant skin displayed a three-fold increase in actively cycling cells, as revealed by Ki-67 staining compared with the control (Figures 1j–l and p). These results demonstrate that despite the differential role of Sufu and Ptc1 in the postnatal skin, Sufu and Ptc1 have similar roles in embryonic skin development: they are required for the formation of hair follicles, and control proliferation and differentiation of the interfollicular epidermis. Loss of Sufu and Ptc1 results in aberrant cell cycle progression Sustained progression through the cell cycle is a hallmark of tumor cells. To determine what drives the increase in proliferation, we examined the expression of the key regulatory molecules involved in cell cycle entry, particularly D-type cyclins. Human BCC and mouse BCC models exhibit upregulation of cyclins D1 and D2, suggesting that expression of D-type cyclins contributes to tumor growth.28 Immunohistochemistry using an antibody that recognizes both cyclins D1 and D2 revealed ectopic expression in the interfollicular epidermis of Sufu and Ptc1 mutant skin compared with the control (Figures 1m–o and q). Consistent with this, western blot analysis revealed a two-fold increase in cyclins D1/2 protein level in Sufu and Ptc1 KO keratinocytes compared with the control (Figure 1r). These findings are consistent with previous reports demonstrating that Hh signaling promotes the expression of D-type cyclins,12,16 and suggest that the increased proliferation observed in Sufu and Ptc1 mutants is attributed to aberrant expression of D-type cyclins. Identification of biological pathways and processes discriminating Ptc1 from Sufu mutants Our analysis thus far has revealed biological processes that are similar between Ptc1 and Sufu mutants. To identify the molecular difference between Sufu and Ptc1 mutants, we performed microarray analysis on Sufu and Ptc1 KO keratinocytes coupled & 2014 Macmillan Publishers Limited

with Gene Set Enrichment Analysis, a computational method used to identify over-represented biological pathways and processes.29 Significant gene sets (FDR o0.05, Po0.01) were visualized as interaction networks with Cytoscape and Enrichment Map.30 From this stringent cutoff, we identified distinct and overlapping biological pathways that are characteristic of Sufu and Ptc1 KO keratinocytes. Overlapping gene sets between Sufu and Ptc1 mutants include those involved in sensory perception and neurotransmission (Supplementary Figure S3 and Supplementary Table 1), and their significance remains to be investigated. In this study, we focused on gene sets that are unique in Sufu KO keratinocytes compared with Ptc1 KO keratinocytes. We identified 13 gene sets that were significantly enriched in Sufu KO keratinocytes, whereby 11 of these are implicated in cell cycle control (Figure 2a). Specifically, Sufu KO keratinocytes were characterized by genes involved in chromosome maintenance (Zwilch, Sgol1 and Kif23), mitosis (CyclinB1, CenpE and Mad2) and DNA repair (Fancd2, Smc2 and Cercr2) (Figure 2b, Supplementary Table 2). DNA repair pathways are intimately associated with the cell cycle, as DNA lesions typically trigger cell cycle arrest. These results suggest that Sufu has a unique function in cell cycle regulation. We uncovered 35 gene sets that were significantly enriched in Ptc1 KO keratinocytes and over 50% of these are involved in transforming growth factor-b signaling, as well as extracellular matrix remodeling and collagen metabolic processes (Figure 2a). Genes associated with extracellular matrix destruction and cell migration (Mmp3, Mmp10, Slit3, Vcan and Adamts6) and genes involved in transforming growth factor-b pathway (Smad7, Tgfb, Tmp4 and Acvr1b) were enriched in Ptc1 KO keratinocytes (Figure 2b). We speculate that these biological events would favor invasion of tumor cells into surrounding stroma during BCC formation. The functional processes characterized by Ptc1 KO keratinocytes have some overlap with those found in previously published transcriptomes derived from human BCC. This includes genes associated with cell mobility, collagen metabolism/ anabolism and extracellular matrix processing.31–33 Although the transforming growth factor-b signaling pathway was not uncovered from human BCC microarray analysis, members of the transforming growth factor-b family were significantly upregulated at both the mRNA and protein levels in human skin tumors including BCC.34,35 In general, there is a low level of gene overlap between human BCC samples. The biological groups that reoccur in published human BCC studies are enriched in processes involved in enzymatic activity and structure/adhesion.36 Therefore, our comparative analysis between Ptc1 and Sufu KO keratinocytes here is informative, unveiling the major molecular pathways involved in BCC pathogenesis. Importantly, the unique cell cycle pathway signature in Sufu KO keratinocytes suggests that the difference in the tumorigenic potential of Sufu and Ptc1 mutants could be attributed to their distinct response(s) in cell cycle regulation. Activation of G2/M checkpoint in Sufu mutants To determine whether Sufu has a specific role in cell cycle regulation, we examined the distribution of cells in G0/G1, S and G2/M phases of the cell cycle in Sufu KO keratinocytes. Flow cytometric analysis of the DNA content of Sufu KO keratinocytes revealed a similar cell cycle profile as Ptc1 KO keratinocytes, both showing a significant enrichment in the G2/M phase (Figure 3a). Accumulation of cells in the G2/M phase may reflect increased cells at G2 phase (G2/M checkpoint) or M phase (mitotic cells). To distinguish between these possibilities, we evaluated the mitotic index in Sufu-mutant skin and keratinocytes, and found similar mitotic count as the control. In contrast, a significant increase in mitotic cells was found in Ptc1-mutant skin (Figures 3b–g and k). A higher proportion of mitotic cells found in Ptc1 mutants indicates Oncogene (2014) 2674 – 2680

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2676

hf

K5Cre;Sufu f/-

Control

K5Cre;Ptc1 f/-

K5Cre;Sufu f/-

Control

hf

hf

K5Cre;Ptc1 f/f

Gli1

Ptc1

600

*

*

700

Control K5Cre;Sufu f/K5Cre;Ptc1 f/f

500

Cyclin D cells/unit length

Ki-67 cells/unit length

700

400 300 200 100

600 500

300 200 100

f/f

f/-

1

u l

re C K5

ro nt co

Ki-67

actin

Control K5Cre;Sufu f/K5Cre;Ptc1 f/f

Fold change in cyclin D

2.5

cyclin D

;P

tc

uf ;S re C

cyclin D

K5Cre;Ptc1 f/f

hf

Control K5Cre;Sufu f/K5Cre;Ptc1 f/f

0

K5

l ro nt co

K5Cre;Sufu f/-

*

400

0

Control

*

* 2.0

*

1.5 1.0 0.5 0

hf

Figure 1. Skin-specific deletion of Sufu and Ptc1 exhibit similar phenotypes at E18.5. (a–c) H&E staining of Sufu and Ptc1 mutants reveals decreased number of hair follicles (arrows). (d–i) In situ hybridization of Gli1 and Ptc1 expression in Sufu and Ptc1 mutant skin, arrows denote ectopic expression. (j–l) Ki-67 (DAKO, Glostrup, Denmark) immunohistochemistry reveals increased proliferating cells in Sufu and Ptc1 mutants compared with the control. (m–o) Immunohistochemistry of cyclin D1/D2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in Sufu and Ptc1 mutant skin (see inset for higher magnification). (p–q) Quantification of ki-67 þ and cyclin D1/2 þ cells in the epidermis. Data are given as means ±s.e.m. (*Po0.05, one-way analysis of variance (ANOVA)). (r) Western blot analysis of cyclin D1/D2 and b-actin (Calbiochem, Danstadt, Germany) protein level in total cell extracts from Sufu and Ptc1 KO keratinocytes. Graphic representation of the mean cyclin D1/2 protein levels normalized to actin ±s.e.m. (n43 independent sets of mutant and control keratinocyte lysates) (*Po0.05, one-way ANOVA).

that loss of Ptc1 leads to increased proportion of cells at M phase, which is consistent with previous finding that Ptc1 inhibits the G2/M phase transition.6,37 To determine whether Sufu KO keratinocytes exhibit a G2/M checkpoint cell cycle arrest, we examined the expression and activity of a mitosis-promoting complex, cyclin B1–cyclin-dependent kinase 1. Immunohistochemistry revealed elevated cyclin B1-positive cells in Sufu- and Ptc1mutant skin compared with the control (Figures 3h–j and l). This may explain the enrichment of G2/M phase observed by flow cytometry (Figure 3a). Although the number of cyclin B1-positive cells increased in Ptc1 mutants, their level of expression was Oncogene (2014) 2674 – 2680

reduced compared with Sufu and control skin. In agreement with this finding, a modest, but significant, upregulation in the level of cyclin B1 protein was observed in Sufu KO keratinocytes, whereas Ptc1 KO keratinocytes did not exhibit a significant change when compared with the control (Figure 3m). Interestingly, Sufu and Ptc1 mutants displayed differential subcellular localization of cyclin B1 (Figure 3l). As nuclear localization of cyclin B1 is a feature of mitotic cells (reviewed in Takizawa et al.38), increased cytoplasmic cyclin B1 in Sufu mutants likely represents a block in G2. To substantiate this notion, we determined the activity of cyclin-dependent kinase 1, which is critical to drive entry into & 2014 Macmillan Publishers Limited

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2677

EXTRACELLULAR STRUCTURE ORGANIZATION

NUCLEAR DIVISION

HSA04510 FOCAL ADHESION

ORGANELLE FISSION

M PHASE CELL DIVISION CHROMOSOME "CELL CYCLE, MITOTIC"CHROMOSOMAL

EXTRACELLULAR MATRIX PART

EXTRACELLULAR REGION PART

TRANSMEMBRANE RECEPTOR PROTEIN SERINE/THREONINE KINASE ACTIVITY

TGF Signaling

Ptc1MT-1

Shared Genes

TUBE DEVELOPMENT LUNG DEVELOPMENT RESPIRATORY TUBE DEVELOPMENT

COLLAGEN METABOLIC PROCESS MULTICELLULAR ORGANISMAL MACROMOLECULE METABOLIC PROCESS

METALLOENDOPEPTIDASE ACTIVITY

HSA00280_VALINE_ LEUCINE_AND_ BUTANOATE_ ISOLEUCINE_ METABOLISM DEGRADATION INTEGRINS IN ANGIOGENESIS

VALINE_LEUCINE_ AND_ISOLEUCINE_ DEGRADATION

CELL REDOX HOMEOSTASIS PROTEIN COMPLEX BINDING

AMB2 INTEGRIN SIGNALING

Ptc1MT-3

Ptc1 mutants

SMAD BINDING

Ptc1MT-2

Sufu mutants

HEPARIN BINDING

TGFβ SIGNALING PATHWAY GROWTH FACTOR BINDING

TGFβ ACTIVITY

SufuMT-1

PART

Cell cycle control

COLLAGEN

HSA04512_ECM_ RECEPTOR_ INTERACTION

ACTIVATION OF THE PRE-REPLICATIVE COMPLEX

EXTRACELLULAR MATRIX

MULTICELLULAR ORGANISMAL METABOLIC COLLAGEN PROCESS FIBRIL MULTICELLULAR ORGANIZATION ORGANISMAL COLLAGEN INTEGRIN CELL PROTEINACEOUS CATABOLIC CATABOLIC SURFACE ECM PROCESS INTERACTIONS PROCESS

MITOSIS

EPIDERMIS MORPHOGENESIS

SufuMT-3

"CHROMOSOME, CENTROMERIC REGION"

M PHASE OF MITOTIC CELL CYCLE

EMC ORGANIZATION

SufuMT-2

SENSORY SENSORY PERCEPTION OF PERCEPTION OF CHEMICAL SMELL STIMULUS

ECM Remodeling

ECM STRUCTURAL CONSTITUENT

Cell cycle control Fancd2 Smc2 CyclinB1 Mad2 Zwilch Sgol1

TGFβ signaling Smad7 Tgfb1 Bmp4 Bmp6 Acvr1B ECM remodeling Mmp3 Mmp10 Slit3 Vcan Adamts6 Z-score -3.0

0.0

3.0

Figure 2. Gene Set Enrichment Analysis (GSEA) reveals distinct biological pathways and processes that define Sufu mutants from Ptc1 mutants. (a) GSEA illustrates that Sufu mutants (red) are enriched in biological pathways involved in cell cycle, whereas Ptc1 mutants (blue) are enriched with biological pathways associated with transforming growth factor-b (TGFb) signaling and extracellular matrix (ECM) remodeling (5% FDR, Po0.01). Cytoscape and Enrichment Map were used for visualization of the GSEA results. Enriched gene sets are represented by nodes, which are grouped and annotated based on gene similarity within each gene set. The size of each node is proportional to the total number of genes within each gene set. Thickness of the green line is proportional to the number of shared genes between gene sets. GSEA,29 as visualized in Cytoscape (version 2.7.0) and the Enrichment Map software30 were used to study pathway-level differences between Sufu and Ptc1 mutants. Gene sets were compiles from the National Cancer Institute (NCI), Kyoto Encyclopedia of Genes and Genomes (KEGG), Protein Families (PFAM), Biocarta and Gene Ontology (GO) pathway databases. (b) Heat map of the top 40 signature genes of cell cycle control and DNA repair, TGFb and ECM assembly identified by GSEA and Enrichment Map Analysis. Z-scores were generated for each gene by mean centering gene expression followed by division by the s.d. Gene expression values are shown as a range from less than to greater than three s.d.’s from the average gene expression level for a particular gene.

& 2014 Macmillan Publishers Limited

Oncogene (2014) 2674 – 2680

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2678 100

DNA content (PI)

Percentage of cells

250

30 20 10

G0/G1: 38.5% S: 33.6% G2/M: 24.4%

0

G0/G1: 35.7% S: 30.1% G2/M: 37.5%

G0/G1

G2/M

K5Cre;Ptc1f/f PH3 cells/ 40 microscopic views

cyclin B1

*

300 250 200 150 100 50 0

cyclin B1+cels/unit length

PH3 DAPI

*

S

*

% of PH3 postive cells

G0/G1: 32.8% S: 25.7% G2/M: 35.9%

PH3

*

*

70 60 50 40 30 20 10 0

50 *

40

Control K5Cre;Sufuf/K5Cre;Ptc1f/f

30 20 10 0

* Control f/K5Cre;Sufu K5Cre;Ptc1f/f *

cytoplasm

nuclear

2.5 2.0

*

cdk1-p

cyclin B1

cdk1

actin

actin

Control K5Cre;Sufuf/K5Cre;Ptc1f/f

1.5 1.0 0.5 0

p=0.09 Fold change in cdk1-p/cdk1

Fold change in cyclin B1

/f

co nt ro l K5 C re ;S uf co u f/nt ro l K5 C re ;P tc 1f

/f

hf

Control K5Cre;Sufuf/K5Cre;Ptc1f/f

DNA content (PI)

K5Cre;Sufuf/-

Control

0

** **

** **

40

co nt ro l K5 C re ;S uf co u f/nt ro l K5 C re ;P tc 1f

G0/G1: 31.9% S: 35.8% G2/M: 27.3%

20 200

250

200

150

100

0

50

0

40

150

20

100

40

50

60

0

60

Control f/f K5Cre;Ptc1

80

50

Control f/K5Cre;Sufu

80

Number of events (x1000)

Number of events (x1000)

100

3.5 3.0

*

Control K5Cre;Sufuf/K5Cre;Ptc1f/f

2.5 2.0 1.5 1.0 0.5 0

Figure 3. Loss of Sufu results in G2/M cell cycle checkpoint. (a) Representative histograms from the flow cytometry analysis are shown for each genotype. Flow cytometric analysis of DNA content reveals enrichment at G2/M phase of the cell cycle in Sufu and Ptc1 KO keratinocytes. Data were analyzed with FlowJo (Tree Star, Ashland, OR, USA) using the cell cycle platform (Watson–Pragmatic model). (b–d) Phospho-H3 (Cell signaling) staining (arrow heads) shows increased mitotic cells in Ptc1 mutant skin compared with both Sufu and control skin. (e–g) PH3 staining in Sufu and Ptc1 KO keratinocytes. (h–j) Cyclin B1 (Cell Signaling Technology, Beverly, MA, USA) immunohistochemistry reveals increased number of cyclin B1-positive cells in Sufu and Ptc1 mutant skin. In general, Ptc1 mutant skin exhibited weaker expression of cyclin B1 (arrows denote additional cells that were counted) compared with both Sufu mutant and control skin. See inset for higher magnification. Asterisks indicate nuclear staining. (k) Graph on left: quantification of PH3 þ cells in the epidermis. Data are given as means ±s.e.m. (*Po0.05, one-way ANOVA). Graph on right: quantification of the proportion of mitotic cells in keratinocytes, calculated by the number of PH3-positive cells/total number of cells. Data are given as means±s.e.m. of keratinocytes isolated from three independent biological samples (*Po0.05, one-way ANOVA). (l) Quantification of the proportion of cyclin B1 þ cells in the epidermis, based on cyclin B1 immunohistochemistry results. Data are given as means±s.e.m. (*Po0.05, one-way ANOVA). (m) Western blot analysis reveals an increase in cyclin B1 (Abcam, Cambridge, MA, USA) protein in total cell extracts from Sufu KO keratinocytes. Data are given as means±s.e.m. of n43 for control, n ¼ 3 for Sufu KO keratinocytes and n ¼ 2 for Ptc1 KO keratinocytes (*Po0.05, one-way ANOVA). (n) Western blot analysis shows an increase in inactive cyclindependent kinase 1 (cdk1) protein (Cell signaling) normalized to total cdk1 (Santa Cruz) in total cell extracts from Sufu KO keratinocytes. Graphic representation of cdk1-p/total cdk1, data are given as means±s.e.m. of n43 for control, n ¼ 4 for Sufu KO keratinocytes and n ¼ 3 for Ptc1 KO keratinocytes (*Po0.05, one-way ANOVA).

M phase and occurs through the dephosphorylation of Tyr15 and Thr14 (reviewed in Takizawa et al.38). Western blot analysis revealed an increase of Tyr15 phosphorylation in Sufu KO Oncogene (2014) 2674 – 2680

keratinocytes, indicating that the observed accumulation of cyclinB1 in Sufu KO keratinocytes likely represents an inactive cyclinB1–cyclin-dependent kinase 1 complex (Figure 3n). These & 2014 Macmillan Publishers Limited

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2679 findings reveal that loss of Sufu results in G2 cell cycle arrest, thereby restraining tumorigenesis and offers an explanation as to why inactivation of Sufu is insufficient to induce BCC formation. Uncoupling of DNA damage and p53 response in Sufu- and Ptc1-mutant skin Cell cycle arrest at G2 is typically elicited in response to DNA damage. Previous studies have demonstrated that oncogeneinduced aberrant cell-division cycles, such as overexpression of S-phase-promoting oncogenes, trigger cellular DNA replication stress and DNA damage.39,40 We found that Sufu- and Ptc1-mutant skin display expanded expression of PCNA, a marker most highly expressed during S-phase, compared with the control (Figures 4a–c). This increase of cell proliferation is associated with the appearance of DNA damage, as revealed by gH2AX staining, in Sufu and Ptc1 mutants (Figures 4d–f). No DNA lesions were detected in the control skin. DNA damage triggers a series of cellular cascades, leading to cell cycle arrest, DNA repair or apoptosis. These events require the tumor suppressor protein, p53, which turns over rapidly under normal conditions, but becomes stabilized and activated in response to DNA damage. Strikingly, p53 protein expression was undetectable in Sufu and Ptc1 mutants, despite the presence of DNA damage (Figures 4g–i). Indeed, p53 RNA expression was reduced by B20% in both Sufu and Ptc1 KO keratinocytes (Figure 4j). Consistent with this downregulation of p53 level, Sufu and Ptc1 KO keratinocytes K5Cre;Ptc1f/f

Control

H2AX

Relative fold change in p53

PCNA 1.4

Control

1.6

K5Cre;Sufuf/-

1.4

K5Cre;Ptc1f/f

Relative fold change

K5Cre;Sufuf/-

Control

displayed reduced expression of p53-responsive genes, p21 and mdm2 (Figure 4k). These observations provide in vivo evidence that Hh pathway activation suppresses p53 accumulation, as revealed previously in mouse embryonic fibroblasts.18 To exclude the possibility that the skin is a unique system, whereby response to DNA damage is independent of p53, Sufu and Ptc1 mutant embryos were exposed to ionizing radiation. Ionizing radiation induced the appearance of gH2AX-positive cells, accumulation of p53 protein, and p21 expression in Sufu- and Ptc1-mutant skin that is comparable to control (Supplementary Figure 4). While we cannot exclude the possibilities that cell cycle-driven and ionizing radiation-induced DNA damages elicit distinct (p53-dependent or -independent) responses, these data suggest that p53 can be induced by DNA damage in the skin. Taken together, these results indicate that both loss of Sufu and Ptc1 impair p53 response to cell cycle-driven DNA damage. In retinal neural progenitors, inactivation of Sufu leads to altered cell cycle kinetics and proliferation,41 suggesting that Sufu regulates the cell cycle. In this study, we demonstrate that loss of Sufu in keratinocytes results in G2/M cell cycle arrest. This is critical for tumor suppression, especially as Hh pathway activation abrogates p53 response. It remains unclear whether the cell cycle effect in Sufu mutants is Gli-dependent or -independent. For example, it is possible that differential expression of Gli proteins in Sufu and Ptc1 mutants (Supplementary Figure S2) accounts for the difference in tumor potential. However, based on our findings here, we favor the notion that cyclin B1 regulation by

1.2 1.0

*

0.8

*

0.6 0.4

K5Cre;Sufuf/K5Cre;Ptc1f/f

1.2 1.0 0.8

*

* *

0.6 0.4 0.2

0.2

0

0

p21

Mdm2

Hh pathway activation

Aberrant cell cycle

p53 DNA damage Sufu p53

G2/M arrest

Basal cell carcinoma

Figure 4. Hh signaling is associated with DNA damage and the suppression of p53. (a–c) Immunohistochemistry reveals expanded PCNA (Santa Cruz) expression in Sufu and Ptc1 mutant skin at E18.5. (d–f) gH2AX (Cell signaling) staining indicates the presence of DNA damage in Sufu and Ptc1 mutant skin at E18.5. (g–i) p53 (Leica Biosystems Newcastle Ltd, Newcastle Upon Tyne, UK) expression is undetectable in the skin of Sufu and Ptc1 mutant skin, whereas p53 is detectable in the p53-positive tissue—the thymus (see inset). (j–k) qPCR shows a significant decrease in p53 levels in Sufu and Ptc1 KO keratinocytes. p53-responsive genes p21 and Mdm2 are also reduced in the mutants compared with the control. SYBR Green gene expression assays were used to quantitative RT–PCR (Applied Biosystems, Carlsbad, CA, USA). Primer sequences used for qPCR: Gapdh_forwad 50 -TCGTCCCGTAGACAAAATGG-30 , Gapdh_reverse 50 -GAGGTCAATGAAGGGGTCGT-30 , Mdm2_forward 50 -CTCTGGACTCGGAAGAT TACAGCC-30 , Mdm2_reverse 50 -CCTGTCTGATAGACTGTGACCCG-30 , p21_forward 50 -CCATGTCCAATCCTGGTGATG-30 , p21_reverse 50 -CGAAGAGA CAACGGCACACTT-30 . (l) Model for BCC development. The Hh pathway activation triggers aberrant cell cycle progression (that is, increase D-type cyclin expression), which leads to replication stress and DNA damage. DNA damage triggers p53 accumulation and activation, as well as DNA damage response that results in G2/M cell cycle arrest. Impairment or escape from p53-mediated tumor surveillance system favors genetic instability and tumorigenesis. Therefore, the Hh pathway activation promotes tumorigenesis via two distinct mechanisms (1) suppression of p53 accumulation and (2) sustained Sufu-dependent G2/M progression. & 2014 Macmillan Publishers Limited

Oncogene (2014) 2674 – 2680

Evasion of p53 and G2/M checkpoints in Hh-driven BCC ZJ Li et al

2680 Ptc1 and unknown cell cycle regulators controlled by Sufu are the key mechanisms for this difference. Comparing Sufu mutants to other Hh-driven BCC models, such as skin-specific overexpression of Gli2 or oncogenic Smoothened, may be able to shed light into the potential mechanism. In conclusion, using genetic and biochemical approaches, we identified and characterized the molecular events involved in BCC pathogenesis using a comparative analysis between Ptc1 and Sufu mutants. Our data indicate that Hh pathway activation induces tumorigenesis by evasion of p53-mediated tumor-suppressive activity and G2/M cell cycle checkpoints (Figure 4l). We uncovered here a novel role for Sufu in cell cycle regulation, which may explain why inactivation of Sufu is not sufficient to induce BCC formation. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS We thank E Nieuwenhuis and W Nien for initial characterization of the Sufu and Ptc1 mutants, and T Satkunendran for technical help. This research is funded by the Canadian Cancer Society Research Institute (2011-700774) to CCH.

REFERENCES 1 Ribeiro GR, Francisco G, Teixeira LV, Romao-Correia RF, Sanches Jr. JA, Neto CF et al. Repetitive DNA alterations in human skin cancers. J Dermatol Sci 2004; 36: 79–86. 2 Sardi I, Piazzini M, Palleschi G, Pinzi C, Taddei I, Arrigucci S et al. Molecular detection of microsatellite instability in basal cell carcinoma. Oncol Rep 2000; 7: 1119–1122. 3 Reifenberger J, Wolter M, Knobbe CB, Kohler B, Schonicke A, Scharwachter C et al. Somatic mutations in the PTCH, SMOH, SUFUH and TP53 genes in sporadic basal cell carcinomas. Br J Dermatol 2005; 152: 43–51. 4 Hutchin ME, Kariapper MS, Grachtchouk M, Wang A, Wei L, Cummings D et al. Sustained Hedgehog signaling is required for basal cell carcinoma proliferation and survival: conditional skin tumorigenesis recapitulates the hair growth cycle. Genes Dev 2005; 19: 214–223. 5 Epstein EH.. Basal cell carcinomas: attack of the hedgehog. Nat Rev Cancer 2008; 8: 743–754. 6 Adolphe C, Hetherington R, Ellis T, Wainwright B. Patched1 functions as a gatekeeper by promoting cell cycle progression. Cancer Res 2006; 66: 2081–2088. 7 Kasper M, Jaks V, Are A, Bergstrom A, Schwager A, Barker N et al. Wounding enhances epidermal tumorigenesis by recruiting hair follicle keratinocytes. Proc Natl Acad Sci USA 2011; 108: 4099–4104. 8 Villani RM, Adolphe C, Palmer J, Waters MJ, Wainwright BJ. Patched1 inhibits epidermal progenitor cell expansion and basal cell carcinoma formation by limiting Igfbp2 activity. Cancer Prev Res (Phila) 2010; 3: 1222–1234. 9 Humke EW, Dorn KV, Milenkovic L, Scott MP, Rohatgi R. The output of Hedgehog signaling is controlled by the dynamic association between suppressor of Fused and the Gli proteins. Genes Dev 2010; 24: 670–682. 10 Tukachinsky H, Lopez LV, Salic A. A mechanism for vertebrate Hedgehog signaling: recruitment to cilia and dissociation of SuFu-Gli protein complexes. J Cell Biol 2010; 191: 415–428. 11 Li ZJ, Nieuwenhuis E, Nien W, Zhang X, Zhang J, Puviindran V et al. Kif7 regulates Gli2 through Sufu-dependent and -independent functions during skin development and tumorigenesis. Development 2012; 139: 4152–4161. 12 Mill P, Mo R, Fu H, Grachtchouk M, Kim PC, Dlugosz AA et al. Sonic hedgehogdependent activation of Gli2 is essential for embryonic hair follicle development. Genes Dev 2003; 17: 282–294. 13 Nilsson M, Unden AB, Krause D, Malmqwist U, Raza K, Zaphiropoulos PG et al. Induction of basal cell carcinomas and trichoepitheliomas in mice overexpressing GLI-1. Proc Natl Acad Sci USA 2000; 97: 3438–3443. 14 Grachtchouk M, Mo R, Yu S, Zhang X, Sasaki H, Hui CC et al. Basal cell carcinomas in mice overexpressing Gli2 in skin. Nat Genet 2000; 24: 216–217. 15 Mill P, Mo R, Hu MC, Dagnino L, Rosenblum ND, Hui CC. Shh controls epithelial proliferation via independent pathways that converge on N-Myc. Dev Cell 2005; 9: 293–303.

16 Kenney AM, Rowitch DH. Sonic hedgehog promotes G(1) cyclin expression and sustained cell cycle progression in mammalian neuronal precursors. Mol Cell Biol 2000; 20: 9055–9067. 17 Kenney AM, Cole MD, Rowitch DH. Nmyc upregulation by sonic hedgehog signaling promotes proliferation in developing cerebellar granule neuron precursors. Development 2003; 130: 15–28. 18 Abe Y, Oda-Sato E, Tobiume K, Kawauchi K, Taya Y, Okamoto K et al. Hedgehog signaling overrides p53-mediated tumor suppression by activating Mdm2. Proc Natl Acad Sci USA 2008; 105: 4838–4843. 19 Fan H, Khavari PA. Sonic hedgehog opposes epithelial cell cycle arrest. J Cell Biol 1999; 147: 71–76. 20 Cooper AF, Yu KP, Brueckner M, Brailey LL, Johnson L, McGrath JM et al. Cardiac and CNS defects in a mouse with targeted disruption of suppressor of fused. Development 2005; 132: 4407–4417. 21 Goodrich LV, Johnson RL, Milenkovic L, McMahon JA, Scott MP. Conservation of the hedgehog/patched signaling pathway from flies to mice: induction of a mouse patched gene by Hedgehog. Genes Dev 1996; 10: 301–312. 22 Svard J, Heby-Henricson K, Persson-Lek M, Rozell B, Lauth M, Bergstrom A et al. Genetic elimination of suppressor of fused reveals an essential repressor function in the mammalian Hedgehog signaling pathway. Dev Cell 2006; 10: 187–197. 23 Ng D, Stavrou T, Liu L, Taylor MD, Gold B, Dean M et al. Retrospective family study of childhood medulloblastoma. Am J Med Genet A 2005; 134: 399–403. 24 Aszterbaum M, Epstein J, Oro A, Douglas V, LeBoit PE, Scott MP et al. Ultraviolet and ionizing radiation enhance the growth of BCCs and trichoblastomas in patched heterozygous knockout mice. Nat Med 1999; 5: 1285–1291. 25 Mancuso M, Pazzaglia S, Tanori M, Hahn H, Merola P, Rebessi S et al. Basal cell carcinoma and its development: insights from radiation-induced tumors in Ptch1-deficient mice. Cancer Res 2004; 64: 934–941. 26 Nieuwenhuis E, Motoyama J, Barnfield PC, Yoshikawa Y, Zhang X, Mo R et al. Mice with a targeted mutation of patched2 are viable but develop alopecia and epidermal hyperplasia. Mol Cell Biol 2006; 26: 6609–6622. 27 Oro AE, Higgins K.. Hair cycle regulation of Hedgehog signal reception. Dev Biol 2003; 255: 238–248. 28 Grachtchouk V, Grachtchouk M, Lowe L, Johnson T, Wei L, Wang A et al. The magnitude of hedgehog signaling activity defines skin tumor phenotype. EMBO J 2003; 22: 2741–2751. 29 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 2005; 102: 15545–15550. 30 Merico D, Isserlin R, Stueker O, Emili A, Bader GD. Enrichment map: a networkbased method for gene-set enrichment visualization and interpretation. PLoS ONE 2010; 5: e13984. 31 Howell BG, Solish N, Lu C, Watanabe H, Mamelak AJ, Freed I et al. Microarray profiles of human basal cell carcinoma: insights into tumor growth and behavior. J Dermatol Sci 2005; 39: 39–51. 32 O’Driscoll L, McMorrow J, Doolan P, McKiernan E, Mehta JP, Ryan E et al. Investigation of the molecular profile of basal cell carcinoma using whole genome microarrays. Mol Cancer 2006; 5: 74. 33 Welss T, Papoutsaki M, Michel G, Reifenberger J, Chimenti S, Ruzicka T et al. Molecular basis of basal cell carcinoma: analysis of differential gene expression by differential display PCR and expression array. Int J Cancer 2003; 104: 66–72. 34 Gambichler T, Skrygan M, Kaczmarczyk JM, Hyun J, Tomi NS, Sommer A et al. Increased expression of TGF-beta/Smad proteins in basal cell carcinoma. Eur J Med Res 2007; 12: 509–514. 35 Lange D, Persson U, Wollina U, ten Dijke P, Castelli E, Heldin CH et al. Expression of TGF-beta related Smad proteins in human epithelial skin tumors. Int J Oncol 1999; 14: 1049–1056. 36 Van Haren R, Feldman D, Sinha AA. Systematic comparison of nonmelanoma skin cancer microarray datasets reveals lack of consensus genes. Br J Dermatol 2009; 161: 1278–1287. 37 Barnes EA, Kong M, Ollendorff V, Donoghue DJ. Patched1 interacts with cyclin B1 to regulate cell cycle progression. EMBO J 2001; 20: 2214–2223. 38 Takizawa CG, Morgan DO. Control of mitosis by changes in the subcellular location of cyclin-B1-Cdk1 and Cdc25C. Curr Opin Cell Biol 2000; 12: 658–665. 39 Gorgoulis VG, Vassiliou LV, Karakaidos P, Zacharatos P, Kotsinas A, Liloglou T et al. Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions. Nature 2005; 434: 907–913. 40 Bartkova J, Horejsi Z, Koed K, Kramer A, Tort F, Zieger K et al. DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature 2005; 434: 864–870. 41 Cwinn MA, Mazerolle C, McNeill B, Ringuette R, Thurig S, Hui CC et al. Suppressor of fused is required to maintain the multipotency of neural progenitor cells in the retina. J Neurosci 2011; 31: 5169–5180.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc) Oncogene (2014) 2674 – 2680

& 2014 Macmillan Publishers Limited