ΔNp63, a Target of DEC1 and Histone Deacetylase 2, Modulates the Efficacy of Histone Deacetylase Inhibitors in Growth Suppression and Keratinocyte Differentiation*

The p63 gene, a member of the p53 family, is expressed as TA and ΔN isoforms. ΔNp63 is the predominant isoform expressed in cells of epithelial origin and frequently overexpressed in cancers. However, what regulates p63 expression is uncertain. Here, we showed that ΔNp63 is regulated by the transcription factor DEC1, a p53 family target. We also showed that the ability of DEC1 to regulate ΔNp63 is enhanced by histone deacetylase (HDAC) inhibitors or knockdown of histone deacetylase 2 (HDAC2). Consistent with this, we found that DEC1 and HDAC2 physically interact and knockdown of HDAC2 leads to increased binding of DEC1 to the ΔNp63 promoter. Interestingly, we found that growth suppression induced by HDAC inhibitors is attenuated by ectopic expression of DEC1 in a ΔNp63-dependent manner. In addition, we showed that ectopic expression of DEC1 inhibits, whereas knockdown of DEC1 promotes, keratinocyte differentiation via modulating ΔNp63 expression. Finally, we showed that DEC1 cooperates with HDAC inhibitors to further decrease keratinocyte differentiation. Together, we conclude that ΔNp63 is a novel target of DEC1 and HDAC2 and modulates the efficacy of HDAC inhibitors in growth suppression and keratinocyte differentiation.

The p63 gene, a member of the p53 family, is expressed as TA and ⌬N isoforms. ⌬Np63 is the predominant isoform expressed in cells of epithelial origin and frequently overexpressed in cancers. However, what regulates p63 expression is uncertain. Here, we showed that ⌬Np63 is regulated by the transcription factor DEC1, a p53 family target. We also showed that the ability of DEC1 to regulate ⌬Np63 is enhanced by histone deacetylase (HDAC) inhibitors or knockdown of histone deacetylase 2 (HDAC2). Consistent with this, we found that DEC1 and HDAC2 physically interact and knockdown of HDAC2 leads to increased binding of DEC1 to the ⌬Np63 promoter. Interestingly, we found that growth suppression induced by HDAC inhibitors is attenuated by ectopic expression of DEC1 in a ⌬Np63-dependent manner. In addition, we showed that ectopic expression of DEC1 inhibits, whereas knockdown of DEC1 promotes, keratinocyte differentiation via modulating ⌬Np63 expression. Finally, we showed that DEC1 cooperates with HDAC inhibitors to further decrease keratinocyte differentiation. Together, we conclude that ⌬Np63 is a novel target of DEC1 and HDAC2 and modulates the efficacy of HDAC inhibitors in growth suppression and keratinocyte differentiation.
p63 shares a similarity with p53 and p73 in sequence-specific DNA binding, activation, and tetramerization domains (1). Like p53, p63 is a transcription factor and capable of inducing a diverse array of target genes responsible for p63 function in inducing cell cycle arrest, differentiation, and apoptosis, and in maintaining the proliferative potential of epithelial stem cells (2)(3)(4)(5)(6)(7). Due to the presence of the upstream P1 promoter and the P2 promoter in intron 3, the p63 gene is expressed as TA and ⌬N isoforms (1). TAp63 contains the N-terminal transactivation domain (TA) conserved among the p53 family (1). ⌬Np63 contains 13 unique residues, which together with the prolinerich domain constitutes a novel activation domain (3,8). In addition, RNA splicing leads to production of at least five alternatively spliced isoforms, ␣, ␤, ␥, ⑀, and ␦ (9).
It has been shown that mice deficient in p63 are defective in epidermal development, including severe limb malformation and an almost complete absence of mammary glands, skin, teeth, and hair (7,10,11). Another study (12) showed that mice heterozygous for p63 are prone to develop an increased tumor burden and metastasis rate, which is compounded in mice harboring heterozygous alleles of p53 and/or p73. Consistent with this, evidence showed that p63, especially ⌬Np63, is found to be amplified and/or overexpressed in multiple tumors and correlated with tumor progression and poor prognosis (13). In addition, overexpression of ⌬Np63 promotes cell proliferation in vitro and tumor growth in vivo (14). These results suggest that TAp63 and ⌬Np63 have opposing functions in tumor suppression. However, somatic mutations of p63 in tumors are rarely detected (15). Thus, the control of p63 activity is likely at the level of p63 expression.
DEC1, 2 also named Stra13 (stimulated with retinoic acid 13) in mouse and sharp2 (enhancer of split and hairy related protein 2) in rat, along with DEC2, belongs to a subfamily of basic helix-loop-helix transcription factors (16). In addition to its role in circadian rhythm, DEC1 plays a role in cell cycle arrest, apoptosis, and differentiation in response to stress signals (16). DEC1 expression is dynamically regulated during mammary gland development (17). In addition, DEC1 expression is found to be increased during progression from normal to in situ in invasive breast carcinomas and in a variety of other cancers, including brain tumors, melanoma, colorectal adenocarcinoma, leukemia, and lymphoma (18,19). However, whether dysregulated DEC1 plays a role in tumorigenesis is not clear.
In this study, we found that ⌬Np63 is a novel target of DEC1 and histone deacetylase 2 (HDAC2). We found that ⌬Np63 is transcriptionally induced by DEC1. In addition, we identified a proximal Sp1 site in the p63 P2 promoter that is responsive to DEC1. Moreover, we found that HDAC2 attenuates the ability of DEC1 to regulate ⌬Np63 expression via physical interaction. Furthermore, we showed that knockdown of HDAC2 increases the binding of DEC1 to the P2 promoter. Finally, we showed that DEC1 attenuates growth suppression induced by HDAC inhibitors and keratinocyte differentation via modulating ⌬Np63 expression. Taken together, we provide evidence that ⌬Np63 is a mediator of DEC1 and HDAC inhibitors in cell survival and differentiation.
Antibodies-Anti-DEC1 is a generous gift from Dr. Yan (University of Rhode Island) (20). Antibodies against p63 (4A4), HDAC8, and IVL were purchased from Santa Cruz Biotechnology. Antibodies against HDAC1, HDAC2, and HDAC3 were purchased from Upstate. Anti-HA epitope was purchased from Covance. Anti-actin, mouse IgG, and rabbit IgG were purchased from Sigma.
Colony Formation Assay-Cells were seeded at 4,000 per well in a 6-well plate with or without induction of DEC1 or DEC1 siRNA, and then treated with trichostatin A (TSA) or suberoylanilide hydroxamic acid (SAHA) for 24 h. For controls, cells were seeded at 1,000 per well and mock treated with dimethyl sulfoxide. The assay was performed as described (26).
Luciferase Assay-The dual luciferase assay was performed in triplicate according to the manufacturer's instructions (Promega). Briefly, 0.25 g of a luciferase reporter, 0.25 g of empty pcDNA4 or pcDNA4 that expresses DEC1 or DEC1-M, and 3 ng of an internal control Renilla luciferase assay vector pRL-CMV (Promega) were transfected into MCF7 cells by Express-Fect reagent according to the manufacturer's instructions (Denville). Cells were seeded at 4 ϫ 10 4 per well in 24-well plates 24 h before transfection. 24 h post-transfection, luciferase activity was measured with the dual luciferase kit and Turner Designs luminometer. The fold-change in relative luciferase activity is a product of the luciferase activity induced by DEC1 or DEC1-M, divided by that induced by an empty pcDNA4 vector.
Cornified Cell Envelope Assay-Cornified cell envelopes (CCEs) were analyzed as described previously (25,28). Briefly, cells were trysinized and resuspended in 1.1 ml of PBS containing 2 mM EDTA, and 100-l aliquotes were counted for the total number of cells. The remaining cells were centrifuged and resuspended in dissociation buffer (2% SDS, 20 mM dithiothreitol, 5 mM EDTA, and 0.1 M Tris Base, pH 8.5), and boiled for 5 min. Detergent-insoluble CCEs were cooled, centrifuged, resuspended in 100 l of PBS, and counted in a hemacytometer. The data were presented as total CCEs/total cells ϫ100.

RESULTS
⌬Np63 Expression Is Regulated by DEC1-Our previous study indicated that DEC1 is regulated by the p53 family and mediates DNA damage-induced cellular senescence (21). DEC1 also plays a role in cell cycle arrest, apoptosis, and differentiation in response to various stress signals (16). To determine how these biological activities are regulated by DEC1, cDNA microarray analysis was performed in MCF7 cells uninduced or induced to express DEC1 by tetracycline (26). We found that p63 mRNA was elevated in cells overexpressing DEC1. To confirm this, we found that the level of the ⌬Np63␣ protein in MCF7 cells was increased upon induction of wildtype DEC1 but not DNA-binding mutant DEC1-R58P (Fig. 1A, compare lanes 1, 3, and 5 with lanes 2, 4, and 6, respectively). Conversely, we found that upon knockdown of DEC1, the level of ⌬Np63␣ protein was decreased in MCF7 cells (Fig. 1B, compare lanes 1 and 3 with lanes 2 and 4, respectively). To rule out potential cell type-specific effects, DEC1 was transiently knocked down by DEC1 siRNA in MCF10A cells. Consistent with the above observation, the levels of ⌬Np63␣ and ⌬Np63␤ were decreased in DEC1-KD MCF10A cells (Fig. 1C). Next, we examined whether an increase in the p63 protein correlates with an increase in the p63 transcript, RT-PCR was performed with primers specifically for amplifying the ⌬Np63 isoform. We found that upon induction of DEC1, the level of ⌬Np63 mRNA was increased (Fig. 1D). As a control, DEC2, a target of DEC1, was repressed by ectopic expression of DEC1 (Fig. 1D), consistent with the previous report (30). DEC1 functions as a transcription repressor by binding to canonical E-boxes (30,31). Recently, DEC1 was found to induce survivin expression by binding to the Sp1 sites in the survivin promoter (22). Thus, we searched for potential E-boxes and Sp1 sites in the genomic locus of the p63 gene and found several potential DEC1 responsive elements within the proximal P2 promoter region ( Fig. 2A). Next, a luciferase reporter under control of the p63 P2 promoter (nucleotides Ϫ1977 to ϩ74), which contains three E-boxes and two Sp1 sites, was constructed and designated p63-P2-1977. We showed that the luciferase activity for p63-P2-1977 was increased by DEC1 (Fig.  2B). As a control, the survivin promoter was activated, whereas the DEC2 promoter was inhibited, by DEC1 (Fig. 2B). However, DEC1-M, a mutant that lacks residues 53-65 in the DNA binding domain (22), was inert (Fig. 2B). To examine which element in the P2 promoter is responsive to DEC1, a luciferase reporter under control of a truncated P2 promoter, which lacks the distal Sp1 site and all of the E-boxes, was generated (nucleotides Ϫ314 to ϩ74) and designated p63-P2-314 (Fig. 2C, left panel). The luciferase assay showed that DEC1 was still able to activate this reporter (Fig. 2C, right panel). To further test this, we generated a mutant p63-P2-314 reporter in which the proximal Sp1 site was mutated, designated p63-P2-314M (Fig. 2C, left panel, mutated nucleotides in boldface). We showed that DEC1 was incapable of increasing the luciferase activity for p63-P2-314M (Fig. 2C, right panel). This suggests that the proximal Sp1 site is likely to be the DEC1 responsive element. Next, a ChIP assay was performed to measure the binding of DEC1 to the p63 promoter in MCF7 cells uninduced or induced to express HAtagged DEC1 (26). The binding of DEC1 to survivin, DEC2, or GAPDH promoter was determined as a positive or nonspecific binding control (Fig. 2D, left panels). We found that DEC1 bound to the promoters of the p63, survivin, and DEC2 genes (Fig. 2D, right panel). However, the GAPDH promoter was not recognized by DEC1 (Fig. 2D, right panel). Taken together, it is likely that ⌬Np63 is a direct target of DEC1.
The Activity of DEC1 to Regulate ⌬Np63 Expression Is Modulated by HDAC2-HDACs and histone acetyltransferases coordinately regulate gene expression through histone acetylation and promoter access by transcription factors, including DEC1 (32)(33)(34)(35). To further test this, we examined whether TSA, a pan-inhibitor of class I, II, and IV HDACs, has an effect on DEC1 to regulate ⌬Np63. We found that DEC1 induction of ⌬Np63␣ was further enhanced in MCF7 cells treated with TSA (Fig. 3A, compare lane 2 with lane 4). We also found that upon treatment with TSA, the level of endogenous DEC1 was slightly increased with concomitant induction of ⌬Np63␣ ( Class I HDACs are primarily localized in the nucleus, whereas class II HDACs are localized in both the nucleus and cytoplasm (36). Because DEC1 is mainly expressed in the nucleus, it is likely that one or more HDACs localized in nucleus (HDAC1, -2, -3, or -8) are involved in regulating ⌬Np63 expression. To test this, we found that the level of ⌬Np63␣ protein was markedly increased by knockdown of HDAC2 but only moderately by knockdown of HDAC1, -3, or -8 (Fig. 3D, compare lane 1 with lanes 2-5, respectively). To test this further, ⌬Np63 expression was examined in MCF7 cells in which HDAC2 was transiently knocked down along with overexpression or knockdown of DEC1. We found that the ability of DEC1 to induce ⌬Np63␣ was further enhanced by knockdown of HDAC2 (Fig. 3E, compare lane 2 with lane 4). In addition, the increased expression of ⌬Np63␣ by knockdown of HDAC2 was abrogated by knockdown of DEC1 (Fig. 3F, compare lane 3 with  lane 4). Thus, like the effect of HDAC inhibitors on p63 expression, the effect of HDAC2 on p63 expression is also DEC1-dependent.
DEC1 forms a complex with HDAC1 in vitro (35), but their interaction has not been defined in vivo. In addition, whether DEC1 interacts with HDAC2 is not clear. To test this, endogenous DEC1 and HDAC2 were immunoprecipitated with anti-DEC1 and anti-HDAC2, respectively. Western blot analysis showed that endogenous HDAC2 was detected in the DEC1 immunocomplex (Fig. 4A, left panel). Conversely, endogenous DEC1 was detected in the HDAC2 immunocomplex (Fig. 4A, right panel). We also showed that HDAC2 associated with ectopically expressed HA-tagged DEC1 in anti-HA immunocomplex (Fig. 4B, left panel). Similarly, HA-DEC1 was detected in the HDAC2 immunocomplex (Fig. 4B, right panel). In addition, HDAC1 was also detected in these complexes (Fig. 4B). Next, to determine how HDAC2 modulates DEC1 activity, the binding of endogenous DEC1 to the P2 promoter was examined and found to be highly increased upon knockdown of HDAC2 (Fig. 4C).
Growth Suppression by HDAC Inhibitors Is Attenuated by DEC1 in a ⌬Np63-dependent Manner-Accumulating evidence demonstrated that HDACs play a role in cancer pathogenesis. Thus, small molecules targeting HDACs have been developed for cancer therapy in the past decade (36 -38). TSA and SAHA, a hydroxamate class of HDAC inhibitors, are able to induce cell cycle arrest, cell death, and differentiation (37)(38)(39)(40)(41).
To determine whether ⌬Np63 and DEC1 play a role in growth suppression by HDAC inhibitors, the colony formation assay was performed in MCF7 cells in which DEC1 can be inducibly knocked down. To rule out a potential effect of doxycycline, M7-TR clone 7, the parental tet-on inducible cell line, was used as a control. We showed that MCF7 cell proliferation was decreased by TSA (Fig. 5A, top panel, compare right column with left column), but not doxycycline (Fig. 5A, top panel, left  column). Nevertheless, the ratio of colonies formed in cells treated with doxycycline versus that in control cells remained Luciferase activity under control of the p63, DEC2, and survivin promoters was measured in the presence of wild-type or mutant DEC1. The luciferase assay was performed as described under "Experimental Procedures." C, left panel, schematic presentation of luciferase reporters (p63-P2-314 and p63-P2-314M) under control of the p63 P2 promoter in which a Sp1 site remains intact or is mutated (boldface, mutated nucleotides in the Sp1 site). Right panel, the proximal Sp1 site is essential for DEC1 to activate the p63 P2 promoter. The luciferase assay was performed as in B. D, left panel, schematic presentation of the ⌬Np63, survivin, DEC2, and GAPDH promoters with the locations of potential DEC1-REs and PCR primers for ChIP assays. Right panel, DEC1 binds to the P2 promoter in vivo. MCF7 cells uninduced or induced to express HA-tagged DEC1 were cross-linked with formaldehyde followed by sonication. Chromatin was immunoprecipitated (IP) with anti-HA or a control IgG. The binding of DEC1 to ⌬Np63, survivin, DEC2, and GAPDH promoters was quantified by PCR. unchanged regardless of TSA treatment (Fig. 5A, bottom  panel). However, although DEC1-KD alone had little if any effect (Fig. 5B, left column, top and bottom panels), DEC1-KD markedly sensitized MCF7 cells to TSA (Fig. 5B, right column,  top and bottom panels). Conversely, whereas ectopic expression of DEC1 had a weak effect on colony formation (Fig. 5C, left  column, top and bottom panels), the ability of both TSA and SAHA to suppress colony formation was markedly inhibited by ectopic expression of DEC1 (Fig. 5C, middle and right columns).
Previous reports showed that ⌬Np63 is associated with cell survival in multiple cell lines (42)(43)(44). To test the role of ⌬Np63 in DEC1-and/or HDAC inhibitor-mediated growth suppression, we generated MCF7 cell lines in which DEC1 is inducibly expressed and ⌬Np63 is stably knocked down. Western blot analysis showed that comparable levels of DEC1 protein were induced, whereas ⌬Np63␣ was undetectable in p63-KD clones 3 and 17 compared with that in DEC1-producing clone 16 (Fig.  5D). M7-sip63-DEC1 clone 3 was then used to examine DEC1 activity in the absence of p63. We found that TSA and SAHA were capable of inhibiting cell proliferation regardless of p63-KD (Fig. 5E, top panel, compare middle and right columns with left column). However, the increased survival by ectopic expression of DEC1 detected in MCF7 cells treated with TSA or SAHA (Fig. 5C) was abrogated by knockdown of ⌬Np63 (Fig.  5E). Similarly, quantification of the colonies showed that the fold-increase of colonies by DEC1 overexpression (Fig. 5C, bot-FIGURE 3. The activity of DEC1 to regulate ⌬Np63 is modulated by HDAC2. A, DEC1 induction of ⌬Np63 is enhanced in cells treated with TSA. Western blots were prepared with extracts from MCF7 cells uninduced (Ϫ) or induced (ϩ) to express DEC1 for 12 h, and then untreated (Ϫ) or treated (ϩ) with TSA (10 ng/ml) for 24 h. B, DEC1 is required for TSA to induce ⌬Np63. Western blots were prepared with extracts from MCF7 cells uninduced (Ϫ) or induced (ϩ) to express DEC1 siRNA for 3 days, and then untreated (Ϫ) or treated (ϩ) with TSA (10 ng/ml) for 24 h. C, the experiment was performed as in B except that MCF10A cells, in which DEC1 was transiently knocked down, were used. D, ⌬Np63 is induced upon knockdown of HDAC2. Western blots were prepared with extracts from MCF7 cells transfected with scrambled siRNA or siRNA targeting HDAC1, -2, -3, or -8 for 3 days. E, knockdown of HDAC2 enhances DEC1 induction of ⌬Np63. Western blots were prepared with extracts from MCF7 cells transfected with scrambled siRNA or siRNA against HDAC2 for 3 days, and then uninduced (Ϫ) or induced (ϩ) to express DEC1 for 24 h. F, the induction of ⌬Np63 by knockdown of HDAC2 is DEC1-dependent. Western blots were prepared with extracts from MCF7 cells uninduced (Ϫ) or induced (ϩ) to express siRNA against DEC1, followed by transfection with scrambled siRNA or siRNA against HDAC2 for 3 days. Left panel, extracts from MCF7 cells were immunoprecipitated (IP) with anti-DEC1 or a control IgG, which were then used to detect HDAC2 and DEC1 along with whole cell lysates as Input control. Right panel, extracts from MCF7 cells were immunoprecipitated with anti-HDAC2 or a control IgG, which were then used to detect DEC1 and HDAC2 along with whole cell lysates as Input control. B, left panel, the experiment was performed as in A, left panel, except that anti-HA was used to immunoprecipitate HA-tagged DEC1 in DEC1-producing MCF7 cells. Right panel, the experiment was performed as in A, right panel. A mixture of anti-HDAC1 and anti-HDAC2 was used to measure the levels of HDAC1 and HDAC2, which were co-immunoprecipitated with DEC1 or immunoprecipitated with anti-HDAC2. Anti-DEC1 was used to detect the levels of DEC1, which was immunoprecipitated with anti-DEC1 or co-immunoprecipitated with anti-HDAC2. C, knockdown of HDAC2 increases the binding of DEC1 to the P2 promoter. MCF7 cells, which were transfected with scrambled siRNA or siRNA against HDAC2 for 3 days, were cross-linked with formaldehyde followed by sonication. Chromatin was immunoprecipitated with anti-DEC1 or a control IgG. The promoter region from nucleotides Ϫ112 to ϩ77 was amplified by PCR with primers as described in the legend to Fig. 2D. APRIL 8, 2011 • VOLUME 286 • NUMBER 14 JOURNAL OF BIOLOGICAL CHEMISTRY 12037 tom panel) was significantly diminished by knockdown of ⌬Np63 (Fig. 5E, bottom panel). DEC1 Inhibits Keratinocyte Differentiation-It is well characterized that ⌬Np63 plays a role in maintaining the proliferative potential of basal cells and down-regulation of ⌬Np63 is associated with keratinocyte differentiation (1,45,46). Thus, we examined whether modulation of ⌬Np63 expression by DEC1 is involved in keratinocyte differentiation (47). We found that consistent with the above observations, transient expression of DEC1 increased, whereas transient knockdown of DEC1 decreased, the levels of ⌬Np63␣ and ⌬Np63␤ in HaCaT cells (Fig. 6A). Calcium influx is a physiological trigger for keratinocyte differentiation and eventual formation of insoluble cornified cell envelope (48). To examine the role of DEC1 in such a process, parental HaCaT and DEC1-producing HaCaT cells were untreated or treated with calcium (1.5 mM) for 3 days. We showed that upon treatment with calcium, ⌬Np63␣ was decreased along with a concomitant increase of involucrin, a differentiation marker (Fig. 6, B, compare lane 1 with lane 2,  and C, compare lane 1 with lane 3), consistent with previous reports (29, 45,46). Interestingly, we found that upon induction of DEC1, the level of ⌬Np63␣ was increased along with decreased expression of involucrin regardless of treatment with calcium (Fig. 6C, compare lanes 2 and 4 with 1 and 3, respectively). In addition, ectopic expression of DEC1 decreased, whereas knockdown of DEC1 increased cornified cell envelope formation (Fig. 6, D and E). Consistent with the effect on cornified cell envelope formation, knockdown of DEC1 decreased ⌬Np63 expression but increased involucrin expression induced by treatment of calcium (Fig. 6F). Because the ability of DEC1 to regulate ⌬Np63 expression is modulated by HDAC inhibitors, we examined whether HDAC inhibitors also affect keratinocyte  ; n ϭ 3). B, top panel, colony formation assay was performed with MCF7 cells uninduced or induced to express siRNA against DEC1 for 3 days along with mock treatment or treatment with TSA (80 ng/ml) for 24 h, and then maintained for 14 days. Bottom panel, the ratio of colonies formed in DEC1-KD cells versus that in control cells without or with TSA treatment (mean Ϯ S.D.; n ϭ 3). C, top panel, colony formation assay was performed with MCF7 cells uninduced or induced to express DEC1 for 24 h along with mock treatment or treatment with TSA (80 ng/ml) or SAHA (10 M) for 24 h, and then maintained for 12 days. Bottom panel, the ratio of colonies formed in DEC1 overexpression cells versus that in control cells without or with TSA or SAHA treatments (mean Ϯ S.D.; n ϭ 3). D, generation of MCF7 cell lines that DEC1 is inducibly expressed and ⌬Np63 is stably knocked down. Western blots were prepared with extracts from ⌬Np63-KD MCF7 cells uninduced (Ϫ) or induced (ϩ) to express DEC1 for 12 h. E, top panel, colony formation assay was performed with ⌬Np63-KD MCF7 cells uninduced (Ϫ) or induced (ϩ) to express DEC1 for 24 h along with mock treatment or treatment with TSA (80 ng/ml) or SAHA (10 M) for 24 h, and then maintained for 12 days. Bottom panel, the ratio of colonies formed in ⌬Np63-KD cells overexpressing DEC1 versus that in control ⌬Np63-KD cells without or with TSA or SAHA treatments. differentiation. We found that upon treatment of TSA, spontaneous cornified cell envelope formation was inhibited (Fig. 6G), along with decreased expression of involucrin (Fig. 6H). Finally, we found that DEC1 cooperated with TSA treatment to further inhibit cornified cell envelope formation (Fig. 6I).

DISCUSSION
The ⌬N isoform of p63 is the one predominantly expressed in epithelial cells and elevated expression of ⌬Np63 is correlated with tumor progression and poor prognosis in multiple cancers of epithelium origin (13). Consistent with this, overexpression of ⌬Np63 promotes cell proliferation in vitro and tumor growth in vivo (14). However, because p63 is rarely mutated in tumors, uncovering the mechanism by which p63 is regulated would provide an insight into how to target tumors with dysregulated p63 expression. Here, we found that ⌬Np63 is transcriptionally regulated by DEC1, a direct target of the p53 family (21). Interestingly, we found that HDAC2 physically interacts with DEC1 and attenuates the ability of DEC1 to regulate ⌬Np63 expression via repressing DEC1 to bind to the p63 P2 promoter. Finally, we showed that DEC1 attenuates growth suppression induced by HDAC inhibitors and cooperates with HDAC inhibitors to suppress keratinocyte differentiation via modulating ⌬Np63 expression. We conclude that ⌬Np63 is a mediator of DEC1 and HDAC inhibitors in cell survival and differentiation. Thus, modulation of ⌬Np63 expression by HDAC inhibitors represents a previously unrecognized mechanism by which HDAC inhibitors regulate gene expression. Furthermore, our results point to a possibility that ⌬Np63 inhibits the efficacy of chemotherapies with HDAC inhibitor(s) for tumors with amplified ⌬Np63 and/or DEC1. Thus, selective knockdown of the pro-survival targets of HDAC inhibitors, such as ⌬Np63 and DEC1, would enhance the efficacy of HDAC inhibitors-based chemotherapies.
DEC1 is classified as a transcription repressor by directly binding to class B E-boxes (30,31). However, DEC1 is capable of activating survivin gene expression via binding to the Sp1 sites (22). Here, we showed that a proximal Sp1 site in the P2 promoter is essential for DEC1 to induce ⌬Np63 expression (Fig.  2). In addition, HDAC2 forms a complex with DEC1 and attenuates the binding of DEC1 to the P2 promoter (Fig. 4). It has been shown that TSA is capable of increasing the level of DEC1 mRNA (35). However, we found that the level of endogenous DEC1 protein is slightly increased upon treatment with TSA ( Fig. 3B) or knockdown of HDAC2 (Fig. 3F). These suggest that other than increased expression of DEC1, the increased binding of DEC1 to the P2 promoter may play an essential role in upregulating p63 expression. Thus, we conclude that under normal conditions, DEC1 binds to and initiates ⌬Np63 transcription from the P2 promoter. However, due to the interaction with HDAC2, DEC1 activity is attenuated. Once the activity of HDAC2 is inhibited via HDAC2 siRNA or HDAC inhibitors, the binding of DEC1 to the P2 promoter is markedly increased, leading to enhanced ⌬Np63 gene expression. Interestingly, HDAC1 is also present in the same complex (Fig. 4B), but has much less effect on DEC1 to regulate ⌬Np63 expression (Fig.  3D). These data indicate that DEC1 can function as both transcriptional activator and repressor in regulating various genes by utilizing different responsive elements and associating with different transcriptional co-factors.
Previous studies showed that DEC1 is frequently overexpressed in tumors (19,49) and in some cases along with overexpression of ⌬Np63 (13). We found that knockdown of DEC1 decreased ⌬Np63␣ expression in MCF7 and MCF10A cells (Fig. 1). Thus, one mechanism by which ⌬Np63 expression is elevated may be due to overexpression of DEC1 in tumors. It has been shown that DEC1 is a downstream target of TSA (35). Here, we found that induction of ⌬Np63 by TSA is dependent on DEC1 (Fig. 3). In addition, we found that the sensitivity of MCF7 cells to TSA and SAHA is enhanced by DEC1-KD, but decreased by DEC1 overexpression (Fig. 5, A-C). Moreover, knockdown of ⌬Np63 abrogates the prosurvival effect of DEC1 in MCF7 cells treated with TSA or SAHA (Fig. 5E). Taken  2) or cells transfected with scrambled siRNA or siRNA against DEC1 for 3 days (lanes 3 and 4). DEC1, p63, involucrin, or GAPDH were quantified by their respective antibodies. B, DEC1 expression is repressed by calcium. Western blots were prepared with extracts from HaCaT cells untreated (Ϫ) or treated (ϩ) with calcium (1.5 mM) for 3 days. C, involucrin expression is repressed by DEC1. Western blots were prepared with extracts from HaCaT cells uninduced (Ϫ) or induced (ϩ) to express DEC1 for 24 h, and then mock-treated (Ϫ) or treated (ϩ) with calcium (1.5 mM) for 3 days. D, ectopic expression of DEC1 represses spontaneous formation of CCEs. 70 -80% subconfluent HaCaT cells were cultured with defined keratinocyte-SFM medium for 24 h, then uninduced or induced to express DEC1 for 24 h, followed by treatment of 1.5 mM calcium for 12 days. CCEs were analyzed as described under "Experimental Procedures," and presented as percentage of total CCEs to total cells (mean Ϯ S.D.; n ϭ 3). E, knockdown of DEC1 enhances CCEs formation. HaCaT cells were transfected with scrambled siRNA or siRNA against DEC1 for 3 days, followed by treatment of calcium (1.5 mM) for 9 days. F, the level of DEC1, ⌬Np63, involucrin, and GAPDH was measured in HaCaT cells used in E. G, TSA inhibits CCEs formation. HaCaT cells were untreated or treated with TSA (5 ng/ml) for 24 h, followed by treatment of 1.5 mM calcium for 11 days. H, the level of ⌬Np63, involucrin, and actin was measured in HaCaT cells used in G. I, suppression of keratinocyte differentiation by DEC1 is further enhanced by treatment with TSA. HaCaT cells (HaCaT-DEC1 clone 5) were uninduced (Ϫ) or induced (ϩ) to express DEC1 for 12 h along with mock treatment or treatment with TSA (5 ng/ml) for 24 h, followed by treatment of 1.5 mM calcium for 11 days.
together, our data indicated that ⌬Np63 is a novel target of DEC1 and mediates DEC1 function in cell survival in response to HDAC inhibitors.
Reports showed that HDAC inhibitors are able to alter the acetylation status of p53, and thus regulate the transcriptional activity of p53 toward its targets involved in cell cycle arrest and apoptosis, such as p21 and Bax (41,50). Interestingly, HDAC inhibitors also induce p21 expression in a p53-independent manner (51). In addition, a recent study showed that despite of induction of p53 upon TSA treatment, up-regulation of p73 contributes to TSA-induced Bax and apoptosis in A2780 ovarian cancer cells (52). Here, we found that like in MCF7 and MCF10A cells containing wide-type p53, DEC1 overexpression or TSA treatment increases ⌬Np63 expression in HaCaT cells containing a mutant p53 (Fig. 6, A and H). Furthermore, it has been shown that TSA up-regulates DEC1 (35). Therefore, although DEC1 is a target of p53 in response to DNA damage, up-regulation of DEC1 and ⌬Np63 upon treatment with TSA may be independent of p53. Because DEC1 is a gene responsive to a broad spectrum of extracellular stimuli, including DNA damage, hypoxia, serum starvation, cytokines, cAMP, insulin, retinoic acid, and light (16), the potential correlation of DEC1 with p53 family members in pro-or anti-tumor activities upon other stress signals needs to be further investigated.
DEC1 is found to mediate chondrocyte differentiation induced by transforming growth factor ␤ in ATDC5 cells (34) and neuronal differentiation induced by retinoic acid in P19 cells (53). However, DEC1 inhibits adipogenesis in L1 cells in response to hypoxia (54). This suggests that DEC1 either promotes or inhibits differentiation in a cell type-dependent manner. Previous studies also showed that ⌬Np63 is a molecular switch between proliferation and differentiation in epidermal keratinocytes, and elevated ⌬Np63 expression suppresses differentiation (1,45,46). Here, we showed that DEC1 is involved in calcium-induced keratinocyte differentiation in HaCaT cells. We also showed that DEC1 induction of ⌬Np63 is correlated with decreased cornified cell envelope formation and decreased expression of involucrin, a differentiation marker (Fig. 6, B and C). In contrast, decreased expression of ⌬Np63␣ by DEC1-KD is correlated with increased cornified cell envelope formation and increased involucrin expression (Fig. 6, D and E). Moreover, we showed that treatment with TSA inhibits cornified cell envelope formation (Fig. 6, G and H) and DEC1 cooperates with TSA to further suppress cornified cell envelope formation (Fig.  6I). These data suggest that suppression of keratinocyte differentiation via ⌬Np63 is at least in part responsible for poor differentiation of tumors with amplified DEC1 and ⌬Np63.