Slug expression in mouse skin and skin tumors is not regulated by p53.

TO THE EDITOR Slug (Snai2), a highly conserved transcriptional repressor important in epithelial-mesenchymal transformation (EMT) in development, has also been implicated in EMT-like processes during tumor progression (Hemavathy et al., 2000). Slug-null mice chronically exposed to UVR develop a lower skin tumor burden than wild-type mice, with fewer aggressive spindle cell tumors (Newkirk et al., 2007). Slug expression is transiently induced in keratinocytes by UVR and Slug expression is persistently elevated in UVR-induced skin tumors (Newkirk et al., 2007; Kusewitt et al., 2009). Slug expression is regulated both at the transcriptional and post-translational levels, and there are conflicting reports about the role of the p53 tumor suppressor in this regulation, as discussed below. In the present studies, we investigated the relationship between Slug and p53 expression in UVR-exposed skin and UVR-induced skin tumors by immunohistochemistry using a specific anti-Slug antibody (Cell Signaling, Danvers, MA), the CM5 antibody (Leica Microsystems, Buffalo Grove, IL) that recognizes all forms of p53, and the PAb240 antibody (Novus, Littleton, CO) that recognizes p53 in the unfolded conformation assumed by many mutant forms. Animal studies were performed in accordance with all applicable state and national animal welfare guidelines. Several studies have suggested that wild-type p53 suppresses Slug expression and enhances Slug degradation (Wang et al., 2009; Rinon et al., 2011; Zhang et al., 2011); however, we demonstrated that induction of wild-type p53 by UVR in keratinocytes did not prevent concurrent Slug induction. In the epidermis of wild-type 129 mice exposed 1 time to 4,800 J m−2 UVR and harvested 24 hours after exposure, both Slug and p53 were expressed predominantly in basal keratinocytes, with some nuclei clearly expressing high levels of both p53 and Slug (Figure 1). Slug has been reported to be positively regulated by p53 in hematopoietic cells in response to ionizing radiation (Wu et al., 2005). To determine whether Slug is a p53 target gene, we quantified Slug expression in p53-null mice (Donehower, 1996) following a single exposure to 2,400 J m−2 UVR. We observed no significant difference in the number of Slug-positive nuclei in the unexposed skin of p53-null FVB mice compared to wild-type FVB skin (Figure 2a). Moreover, in both wild-type and p53-null mice, UVR exposure significantly (P0.05) enhanced Slug expression, and there was no significant difference between Slug induction in wild-type versus p53-null epidermis (Figure 2a). Thus, wild-type p53 was not required for Slug induction by UVR. A number of reports indicate that wild-type p53 suppresses while mutant p53 enhances Slug expression (Wang et al., 2009; Roger et al., 2010; Rinon et al., 2011; Zhang et al., 2011). We therefore examined the relationship between Slug and p53 expression in skin lesions expressing mutant p53. After repeated UVR exposures, hairless mice develop focal aggregates of p53-positive keratinocytes; many of these foci harbor p53 mutations and some represent precursors of squamous cell carcinomas (Rebel et al., 2001; Rebel et al., 2005; de Gruijl and Rebel, 2008). Virtually all UVR-induced tumors in the skin of SKH-1 hairless mice express 1 or more mutant p53 alleles (Melnikova et al., 2005). We examined the skin of five SKH-1 hairless mice exposed three times weekly to UVR for a total of 10 weeks to induce the formation of preneoplastic p53-positive foci. At the time of killing, 20 weeks after the last UVR exposure, UVR-exposed skin without gross abnormalities was examined. Of the 30 foci identified using the CM5 antibody, only 14 of these were also immunoreactive with the PAb240 antibody. Examination of skin from similarly treated mice at earlier time points revealed that CM5-positive foci appeared before Pab240-positive foci were detectable, suggesting that some CM5-positive foci expressed stabilized wild type rather than mutant p53. We also examined p53 expression in a total of 32 squamous cell carcinomas and spindle cell tumors induced by UVR in SKH-1 mice. Four of the squamous cell carcinomas did not express immunohistochemically detectable p53; the p53 gene was likely deleted in these tumors. Of the remaining 23 squamous cell carcinomas, 14 expressed p53 detectable using both Pab240 and CM5 antibodies. Another nine squamous cell carcinomas expressed p53 detectable only with the CM5 antibody. All five spindle cell tumors examined had adjacent or intermingled neoplastic epithelial components, and all spindle cell components were immunoreactive with both p53 antibodies. Slug expression in these UVR-induced preneoplastic foci and skin tumors was compared to p53 expression (Figure 2b). Both p53-negative epidermis and p53-positive preneoplastic foci in UVR-exposed skin had significantly fewer Slug-positive nuclei than p53-positive epithelial tumors (P0.05). There was no significant difference between non-tumor skin, whatever the p53 status, and p53-negative epithelial tumors or p53-positive spindle cell tumors. Spindle cell tumors had significantly lower Slug expression than CM5-postive epithelial tumors (P0.05), but did not differ significantly from PAb240-positive epithelial tumors. P53-positive epithelial tumors expressed higher levels of Slug than p53-positive preneoplastic foci and there was reduced expression of Slug in p53-positive spindle cell tumors. Thus, it appeared that p53 mutation was not the critical factor driving Slug expression during skin tumor progression. Our results failed to demonstrate a relationship between Slug and p53 expression in acutely UVR-exposed skin, UVR-induced preneoplastic foci, or UVR-induced skin tumors. Activation of wild-type p53 did not prevent Slug induction, p53 was not required for UVR induction of Slug, and mutant p53 expression did not enhance Slug expression. Indeed, in tumors, expression of mutant p53 detectable with pAB240 or complete absence of p53 expression was actually associated with somewhat decreased Slug expression. Our findings are consistent with a recent report showing no correlation between p53 and Slug immunoreactivity in human squamous cell carcinomas (Chen et al., 2013) and suggest that the relationship between Slug expression and p53 status in vivo requires additional investigation. Slug interactions with the p53 protein are believed to be mediated via Mdm2, with formation of a wild-type p53–Mdm2–Slug complex leading to degradation of Slug (Wang et al., 2009). However, the relationship between Slug and Mdm2 expression is also unresolved. Some studies have shown that decreased Mdm2 expression is correlated with increased Slug expression (Kim et al., 2010; Wu et al., 2013). On the other hand, another study indicates that Mdm2 actually enhances Slug mRNA expression, although Mdm2 increases Slug protein levels only in the absence of wild-type p53 (Jung et al., 2013). Studies in Mdm2 knockout mice similar to those performed in the present study in p53 knockout mice will be required to determine the interrelationship of Mdm2 and Slug in vivo. The authors state no conflict of interest. This work was supported by National Institutes of Health grants R21AR054361 (DFK), R01 ES015587 (DGJ), R01 CA133629 (TMO), R01 CA109204 (TMO), P30 CA16672, and P30 ES07784. We thank the MD Anderson Comprehensive Cancer Center Support Grant Mutant Mouse Pathology Service and the Science Park Center for Research in Environmental Disease Histology Core Facility for their excellent technical support for these studies.