Abstract
Background: The specific ablation of Trp53 gene in mouse epidermis leads to the spontaneous development of aggressive squamous cell carcinoma, a process that is accelerated by the subsequent loss of Rb gene. Materials and Methods: The possible mechanisms leading to spontaneous tumor formation in epidermis in the absence of Trp53 were studied focusing on hair cycle defects, inflammation and possible chromosomal instability (CIN). Results: Loss of p53 induces tumorigenesis primarily by mediating early CIN and, to a minor extent, nuclear factor κB activation. Notably, CIN occurs not only in p53-deficient skin, but also in epidermis lacking both Rb and Tp53 tumor suppressors, indicating a predominant role of this process in spontaneous tumorigenesis. Conclusion: These data identify CIN as a major mechanism in tumorigenesis originated by Trp53 loss in stratified epithelia and imply that therapies aimed to counterbalance CIN might be of relevance for the treatment of human cancer bearing impaired p53 functions.
The p53 tumor suppressor coordinates the cellular response to stress, including DNA damage, hypoxia and oncogenic stress through transcriptional mechanisms, resulting in cell cycle arrest, senescence, or apoptosis. Accordingly, p53 mutations are widely involved in human tumorigenesis (1). Moreover, altered p53 functions are also associated with poor prognosis and high metastatic potential in human tumors (1, 2).
We and others have previously described that the specific ablation of p53 gene in stratified epithelia (Tp53lox;K14cre hereafter Tp53ΔEC mice) leads to spontaneous squamous cell carcinoma (SCC) development (3, 4), in a process that is accelerated by the subsequent ablation of Rb gene (4). Detailed analysis indicates that such acceleration, characterized by earlier tumor development, is mediated by increased activation of Akt pathway, resulting in increased proliferation and augmented tumor angiogenesis (4). Nonetheless, the mechanisms leading to spontaneous tumor development by Tp53 loss in epidermis remain largely unknown. To gain some insight into these mechanisms, we recently performed functional genomic analysis of tumors arising in mice bearing the epidermal deletion of Trp53 or Rb and Trp53 (5). These studies revealed that these mouse tumors share relevant characteristics with multiple human malignancies, distinguished by poor prognosis, altered p53 status and high metastatic potential. Of note, these genomic alterations, which are independent of the pRb status, also revealed a primary involvement of cell cycle genes, in particular those implicated in mitosis, and a functional enrichment of stem cell-associated genes (5), in agreement with the hair follicle origin of the tumors (4).
Here we have performed a detailed analysis of the possible mechanisms leading to spontaneous tumor development in Tp53ΔEC and their possible metastatic characteristics.
Materials and Methods
Mice and genotyping. Mice of the different genotypes were kindly provided by Dr. A Berns (NKI, Amsterdam, the Netherlands) and they were enriched in FVB background by backcrossing (4). The different mice used in this study (RbΔEC (n=20), Tp53ΔEC (n=43) and RbΔEC;Tp53ΔEC (n=65)) were genotyped as previously described (4). Mice of all genotypes were monitored for tumor development over a period of 20 months and sacrificed for histopathological analyses when animals showed obvious tumors or morbidity. RbΔEC and control mice were subjected to a DMBA/TPA chemical carcinogenesis protocol as described elsewhere (6). The animal experiments were approved by the Animal Ethical Committee and conducted in compliance with Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT) Guidelines.
Histological procedures. For histological analysis, skin and tumoral samples were fixed in formalin and embedded in paraffin prior to sectioning, or were snap frozen in liquid nitrogen and embedded in OCT for cryosectioning. Sections of 5 μm were cut and stained with hematoxylin and eosin (H&E) for pathology assessment. Paraffin sections were deparaffinized. All sections were incubated with 5% horse serum for 30 minutes to block the Fc receptor in tissue, and then washed three times with sterile phosphate-buffered saline (PBS) (pH 7.5) prior to incubation with the appropriate primary antibodies diluted in PBS/bovine serum albumin (BSA). Antibodies were used as follows: in frozen sections 1/50 dilution of FITC-mouse anti-mouse CD45.2 (Ly 5.2) (BD Pharmingen, NJ, USA) for CD45 detection, fluorescein isothiocyanate (FITC)-rat anti-mouse CD11b (BD Pharmingen) for macrophage detection, FITC-rat anti-mouse Ly-6G and LY6C (BD Pharmingen) for granulocyte detection, and 1/1,000 dilution of anti-K5 (Covance, NJ, USA) for keratin 5 detection. In paraffin sections: 1/200 dilution of anti-γ-tubulin-centrosome marker (Abcam); 1/100 dilution of anti-α-tubulin for microtubule detection (Sigma, MO, USA); 1/50 dilution of P-ChK1 (S345) (Abcam) and P-Chk-2 (T68) (Abcam, Cambridge, UK). FITC- or TexasRed-conjugated secondary antibodies for immunofluorescence were purchased from Jackson ImmunoResearch (CA, USA) and used at 1/50 and 1/500, respectively. Diamidinophenylindole (DAPI) was used to counterstain the nuclei or chromosomes. Control slides were obtained by replacing primary antibodies with PBS (data not shown). The quantitative analyses of mitosis were performed using H&E-stained sections from at least five different mice/tumors scoring three to seven different sections. The total number of mitoses scored ranged from 150 to 500 for each data point. Double immunofluorescence stained sections against γ- and α-tubulin scoring at least 150 mitotic Figures for each data point. Data are shown as mean±SD.
Affymetrix mouse gene chip 430A analysis. Skin samples from newborns were preserved in RNAlater (Ambion, CA, USA) and disrupted and homogenized using Mixer Mill MM301 (Retsch, Haan, Germany). Total RNA was extracted and purified from 30 mg of skin using RNeasy Fibrous Tissue Mini kit (Qiagen, Venlo, the Netherlands) following the manufacturer's recommendations. The integrity of the RNA populations was tested in a Bioanalyzer (Agilent, CA, USA) showing an at least 1.4 28S to 18S ratio. Total RNA was hybridized at the Genomic Facility of the Centro de Investigación del Cáncer (Salamanca, Spain). We exported .cel files from Affymetrix GCOS software, and used the GEPAS Analysis Suite (7) to perform background subtraction with RMA (8), normalizing the chips using the quantile method (9) and log2 transformed and mean-centered the intensity values. ANOVA analysis (p<0.01), performed using MeV software (10), was used to select 1,435 genes differentially expressed among the quoted genotypes and control mice. The whole data sets are available at Gene Expression Omnibus database (GSE11990). The inflammation-related genes shown in Figure 2 were obtained by Gene Ontology using the DAVID Functional Annotation tool.
Western blot analysis. Whole skin extracts were ground with a mortar on liquid nitrogen, homogenized and lysed by freeze-thawing cycles in lysis buffer C (200 mM HEPES pH 7.9, 25% glycerol, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 μg/ml aprotinin, 1 μg/ml leupeptin, 1 mM PMSF, 20 mM NaF, 1 mM NaPPi, 1 mM Na3VO4, 2.5 mM DTT), and centrifuged to obtain supernatants containing total protein. 35μg protein per sample were resolved in SDS-PAGE gels and transferred to nitrocellulose membranes (Amersham, UK). Membranes were blocked with 5% non-fat milk diluted in TBS and incubated with the appropiate antibodies diluted in TBS-T 0.5% BSA. Antibodies were used as follows: anti-NFκB p65, anti-NFκB p50, anti-NFκB p52, anti-IKKα, anti-IKKγ all diluted 1/250 and purchased from Santa Cruz; anti-IκBα diluted 1/200 (Santa Cruz); anti-IKKβ diluted 1/1,000 (IMGenex, CA, USA); anti-actin diluted 1/100 (Santa Cruz). Secondary antibodies were purchased from Jackson ImmunoResearch and used at 1/5,000. Super Signal West Pico Chemiluminscence Substrate (Pierce, IL, USA) was used according to the manufacturer's recommendations to visualize the bands.
Results
The absence of Trp53 in stratified epithelia through cre-LoxP system (Tp53ΔEC mice) leads to spontaneous tumor formation in mice through as yet unknown mechanisms. To investigate the possible process involved in tumorigenesis, we focused on three previously reported activities of p53, which are also involved in skin tumorigenesis. Firstly, p53 is involved in hair cycle through catagen promotion (11); secondly, p53 represses inflammatory response by interfering with NFκB-dependent transcription (12), and thirdly, p53 prevents centrosome amplification and thus mitotic aberrations (13).
The possible involvement of p53 in tumorigenesis through interference with hair cycling would imply that processes leading to accelerated hair regrowth would also account for increased tumor development. To test this hypothesis, we performed repeated depilation and monitored tumor incidence in the treated areas. No increase in the number of tumors was recorded in mice (n=10) under repeated depilation after 4 months compared with untreated mice or with depilated controls (not shown), thus excluding hair cycle alterations as a main mechanism for tumor development in Tp53ΔEC mice.
To analyze the involvement of inflammatory processes in tumor development, we monitored the presence of inflammatory cells in non-lesional epidermis of control, Tp53ΔEC and Tp53ΔEC;RbΔEC adult mice, as these two mouse models display high incidence of spontaneous tumor development in which the presence of such cells is a common characteristic (4). We observed an increased number of lymphocytes (CD45-positive), macrophages (Mac-1-positive) and granulocytes (Gr-1-positive) in Tp53ΔEC mouse epidermis (Figure 1A′, B′, C′), compared to control epidermis (Figure 1A, B, C). Such increased numbers were not further augmented in Tp53ΔEC;RbΔEC skin (Figure 1A″, B″, C″) and only augmented to a moderate extent in epidermal tumors (Figure 1A″′, B″′, C″′). These data suggest that the absence of Trp53 is sufficient to elicit a proinflammatory response and recruit effector cells in epidermis prior to tumor development.
To further corroborate these results, we analyzed microarray data obtained from newborn mouse epidermis (4) in search for possible pro-inflammatory genes. This analysis (Figure 2A) indicated that there is a reduced number of genes that display increased expression in Tp53ΔEC exclusively, whereas the vast majority of them display a pattern characterized either by augmented expression in Tp53ΔEC;RbΔEC mouse epidermis, or by a moderate increase in Tp53ΔEC further increased in Tp53ΔEC;RbΔEC. These data indicate that the absence of Trp53 or, more importantly, the simultaneous absence of Rb and Trp53, promote expression changes in multiple inflammation-related genes. Finally, given the relevance of the NFκB pathway in regulating the expression of inflammatory genes, we analyzed the expression of proteins belonging to this pathway in mouse epidermis. We observed (Figure 2B) an increase in p65, p50 and IKKβ in Tp53ΔEC and Tp53ΔEC;RbΔEC samples, whereas no significant changes were observed in IKKα. Of note, the expression of p52 decreased in all transgenic samples, and a moderate increase in IKKγ and a moderate decrease in IκBα was only observed in Tp53ΔEC;RbΔEC mouse epidermis. Overall, these data indicate that the absence of p53 in epidermis triggers a pro-inflammatory response which is only partially dependent on NFκB activation.
Since mutations in p53 promote mitotic aberrations which may lead to spontaneous tumor development (14-16), we studied the presence of aberrant mitosis in Tp53ΔEC and Tp53ΔEC;RbΔEC mouse epidermis and tumors. Multiple mitotic aberrations were observed in epidermis (Figure 3A, B, E) and in epidermal tumors (Figure 3C, D, E), suggesting that deregulated mitotic progression could be responsible for spontaneous tumor development. We previously reported that tumors arising in Tp53ΔEC and Tp53ΔEC; RbΔEC mice display a clear hair follicle origin (4), since proliferation, and thus mitosis, is augmented in these areas compared with interfollicular epidermis (Figure 3F), this may help to explain our previous data. Moreover, proliferation in hair follicles is dramatically increased by the loss of Rb (Figure 3F), which might act synergistically with Trp53 loss to increase the number of mitotic aberration in Tp53ΔEC;RbΔEC hair follicles, thus providing further support for the accelerated tumor formation observed in Tp53ΔEC;RbΔEC mice (4). Finally, to further support the involvement of Trp53 loss in promoting mitotic aberrations in epidermis and thus favoring tumor development, we also monitored possible alterations in papillomas and SCCs arising upon chemical carcinogenesis protocols in RbΔEC epidermis (6). In this system, the development of papillomas is prevented by overactivation of p53-dependent pathways, leading to premature p53 loss of function and increased malignant conversion of papillomas in SCCs (6). Rb-deficient malignant papillomas, which have lost p53 functions (premalignant), displayed a significant increase of aberrant mitosis, similar to that observed in Rb-deficient SCCs and exceeding that shown by SCCs developed in control mice (Figure 3G).
One of the control mechanisms exerted by p53, affecting mitotic control, is the centrosome duplication (13, 17), and multiple multipolar mitoses were observed in our analysis (Figure 3B-D). Consequently, we monitored centrosome number in epidermis and tumors of Tp53ΔEC mice. Although the majority of mitoses in non-tumoral epidermis displayed a normal number of centrosomes (Figure 4A, C), we detected an increased number of centrosomes in both non-tumoral epidermis and epidermal tumors in Tp53ΔEC mice (Figure 4A´, B, C). These observations indicate that the altered mitoses in Tp53ΔEC mouse epidermis and tumors were mediated, at least in part, by aberrant centrosome duplication and positioning. Importantly, such altered centrosome number was associated with increased activation of mitotic checkpoints characterized by augmented expression of phosphorylated Chk1 and Chk2 (Figure 4D, D´, respectively), in agreement with our previous findings of increased expression of mitosis controlling genes in tumors arising in p53-deficient epidermis (5).
Collectively, the data presented here support the hypothesis that the spontaneous tumor development due to p53 absence is mediated by premature mitotic aberrations.
Discussion
The p53 gene is one of the most frequently mutated tumor suppressor genes in human malignancies, and its loss of function is associated with poor prognosis and high metastatic potential in multiple human tumors (1, 2). As a consequence, numerous efforts are directed to target the pathways affected by p53 loss as promising therapeutic alternatives (18). Such efforts require the use of well suited preclinical models to asses the efficiency of such therapies in an in vivo setting. We and others have previously reported the development of SCC in Tp53ΔEC mice (3, 4), and the collaboration of the subsequent ablation of Rb gene in the process (4). Moreover, functional genomic analysis of the tumors compared with control epidermis revealed that the loss of p53 is the most relevant characteristic for tumor development, and that the molecular changes allowed the identification and classification of human tumors harboring p53 mutations, poor prognosis and high metastatic behavior (5). Consequently, these mice represent an attractive alternative for testing molecularly targeted therapies. In spite of this, the molecular bases of tumor development in Tp53ΔEC mice are not well characterized. Here we focused on the three major mechanisms that might be relevant in the context of epidermis without external carcinogenic stimuli.
It is well assumed that most epidermal tumors arise in the hair follicle, probably in close relationship with epidermal stem cells (19, 20). Since p53 protein is suggested to act during catagen induction, the regression phase of the hair cycle (11), one may expect possible altered hair cycle and increased tumorigenesis in Tp53ΔEC mice upon repeated depilation. However, although spontaneous tumors arising in Tp53ΔEC mice display a clear hair follicle origin (4), we did not find tumor development in the treated areas, thus excluding the possibility that hair cycle alterations resulted in increased tumorigenesis. This can be explained if the functions of p53 can be carried by other proteins (21) and/or in the context of complete p53 loss, affecting epidermal and dermal cells (11), these functions can be discerned.
It has been reported that mutated p53 expression correlates positively with NFκB activity in cultured cancer cells and that mutant p53 can elevate NFκB activity via transactivation of the NFκB2 gene, in particular in the context of SCC development (12, 22-25). Our data add the functional characterization of proinflammatory cells prior to tumor development in mice bearing the epidermal specific deletion of Trp53 (Figure 1). Furthermore, in skin of newborns we found the expression of multiple genes involved in inflammatory response and deregulation of NFκB pathway proteins (Figure 2). These data highlight the possibility that inhibitors of inflammation, such as COX-2 inhibitors (26, 27) can be of use as anticancer therapies in mutant p53- bearing tumors.
Finally, we also analyzed possible premature chromosome instability in Tp53ΔEC mouse. Chromosomal instability is an early event in SCC development (28-30) and the functions of p53 mediating proper mitotic progression are well recognized (31-33). Therefore it is not surprising that the mutations in p53 promote mitotic aberrations leading to spontaneous tumor development (14, 15), which might be increased by Rb family inactivation (16). Overall, our findings are in agreement with these observations, as we detected frequent aberrant mitoses in Tp53ΔEC mouse tumors. Moreover, we also describe that such mitotic aberrations are frequent events prior to overt tumor development, thus reinforcing the oncogenic functions of chromosome instability. The observations that centrosome numbers are also altered prior to tumor formation also indicates that the functions of p53 controlling such process are essential for tumor suppressor activities. However, our data also seem to be in disagreement with the absence of centrosome aberrations in skin tumors promoted by Kras activation in Tp53ΔEC mice (14). Such apparent discrepancy can be explained by the use of an inducible cre system, the presence of Kras expression, or the induction of tumor formation by TPA treatment used by other authors (14). Indeed, the spontaneous tumor development between Tp53ΔEC and mice expressing the p53.R270H mutation is similar, and the differences are only appreciable upon UV carcinogenic stimuli (34).
Our data may also suggest that chromosome instability by complete loss of function, either by Trp53 loss or dominant functions, may represent a more dramatic effect than abrogating some specific p53 functions in oncogenesis. Indeed, the chromosomal instability observed in RbΔEC mice upon two-stage carcinogenesis leading to aggressive and poorly differentiated carcinomas (6, 35) is in contrast with the poorly aggressive and highly differentiated tumors arising in RbΔEC;p107-/- mice (36), in which the absence, or reduced expression, of p107 abrogates the proapoptotic functions of p53 without affecting cell cycle (36, 37). In support of this, a chromosomal instability signature represents a potential tool in determining malignancy and metastatic behavior over a broad range of human tumors (38), and the genomic analysis of Tp53ΔEC tumors, which display clear parallelism with chromosomal instability, is also able to predict malignant and metastatic behavior of several types of human cancer (5). Collectively, such observations indicate that Tp53ΔEC mice can be an excellent tool for the preclinical evaluation of drugs targeted against mitotic regulators.
Acknowledgements
Grant support: Ministerio de Ciencia e Innovación grant SAF2008-00121; Comunidad Autonoma de Madrid Oncocycle Program grant S2006/BIO-0232, Ministerio de Ciencia e Innovación grant PS-090100-2006-3, and Ministerio de Sanidad y Consumo grant ISCIII-RETICC RD06/0020 (J.M. Paramio). We thank Dr. A. Berns [Netherlands Cancer Institute (NKI), Amsterdam, the Netherlands] for providing different mutant mice, Jesús Martínez and the personnel of the animal facility of CIEMAT for the excellent care of the animals, and Pilar Hernández (CIEMAT) for the histological preparations.
- Received March 24, 2009.
- Revision received May 29, 2009.
- Accepted June 18, 2009.
- Copyright© 2009 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved