Novel p63 target genes involved in paracrine signaling and keratinocyte differentiation

Abstract

The transcription factor p63 is required for proper epidermal barrier formation and maintenance. Herein, we used chromatin immunoprecipitation coupled with DNA sequencing to identify novel p63 target genes involved in normal human epidermal keratinocyte (NHEKs) growth and differentiation. We identified over 2000 genomic sites bound by p63, of which 82 were also transcriptionally regulated by p63 in NHEKs. Through the discovery of interleukin-1-α as a p63 target gene, we identified that p63 is a regulator of epithelial-mesenchymal crosstalk. Further, three-dimensional organotypic co-cultures revealed TCF7L1, another novel p63 target gene, as a regulator of epidermal proliferation and differentiation, providing a mechanism by which p63 maintains the proliferative potential of basal epidermal cells. The discovery of new target genes links p63 to diverse signaling pathways required for epidermal development, including regulation of paracrine signaling to proliferative potential. Further mechanistic insight into p63 regulation of epidermal cell growth and differentiation is provided by the identification of a number of novel p63 target genes in this study.

Keywords: FGF10; GM-CSF; KGF; cytokines; p53; p73.

Figure 1

Identification of novel direct p63 target genes. (a) Rapidly growing cultures of NHEKs were analyzed for p63 isoforms by comparison with protein markers generated by transfection of the indicated expression vectors in H1299 cells. Total protein was isolated and expression of p63 isoforms was analyzed by western blot. (b) NHEKs were chemically crosslinked and quantitative chromatin immunoprecipitation used to determine the relative binding of p63 to a negative control region (con) and to three previously published p63 binding sites located in the p21, MDM2, and Jagged-1. (c) p63-bound genes in NHEKs were overlayed with p63 target genes identified in ME180 cells. Bootstrap distribution statistical analysis was performed to compare the significance of gene overlap between the two datasets. (d) NHEKs were transduced with a retrovirus-expressing shRNAs specific to GFP (con) or the DNA-binding domain of p63 (sip63). Cells were puromycin selected for 48 h and harvested for western analysis of p63 and actin. (e) RNA was isolated and submitted, in duplicate, for Affymetrix microarray analysis to identify genes with expression patterns dependent on the presence of p63. Data shown are representative of at least three independent experiments. p63-bound genes identified in the ChIP dataset (left circle) were overlayed with genes exhibiting p63-dependent expression in microarray analyses (right circle). Eighty-two genes, as annotated by the GeneSpring 7 software, were present in both datasets and were chosen for further biological analyses

Figure 2

Validation of p63-bound response elements located in/near target genes. Rapidly growing NHEKs were chemically crosslinked, snap frozen, and submitted for FactorPath query analysis. Specific p63 binding sites were selected from the regions identified by the initial ChIP experiment. A majority of the sites queried were enriched by p63 immunoprecipitation, with statistically significant hits marked with an asterisk (^*P<0.05; ^**P<0.01). Data presented were generated from three independent experiments. Please note that for BDKRB2, we identified two independent p63 binding sites, denoted (1) and (2)

Figure 3

Validation of genes displaying a p63-dependent expression pattern. (a) NHEKs were transduced with adenovirus-expressing p63DBD siRNA, or shuttle virus as a control. Protein lysates were harvested 48 h later and western blot analysis performed for p63 and actin. (b) Rapidly growing NHEKs were treated as in (a) and RNA was isolated 48 h after infection and target genes analyzed by quantitative real-time PCR. Data presented are representative of three independent experiments. (c) Three TAp63-specific siRNAs were co-transfected with TAp63γ into H1299 cells. The expression of TAp63 was analyzed as in (a). (d) TAp63 siRNAs were co-transfected with TAp73β. p73-specific siRNAs were used as a positive control for knockdown. Expression of TAp73 and actin were analyzed as in (a). (e) TAp63 siRNAs were transfected into NHEKs and levels of ΔNp63α, p53, and actin were analyzed by western blot. (f) NHEKs were transfected with TAp63-specific siRNAs and qPCR was used to analyze endogenous levels of ΔNp63 and TAp63 transcript levels 48 h after transfection. (g) NHEKs were transfected with NS, or TAp63-specific siRNAs as in (b), and RNA was analyzed by quantitative real-time PCR to observe transcript levels of a subset of novel direct p63 target genes. Each of the three siTAp63 oligos were transfected into independent cultures of NHEKs, and data presented next to the original microarray data (gray bar) represent the mean of all three TAp63 siRNA experiments (black bar)

Figure 4

p63 regulates paracrine pathways between keratinocytes and dermal fibroblasts. (a) Model of paracrine signaling between the epidermis and the dermis. (b) NHEKs were transiently transfected with control (NS) or p63-specific siRNA, harvested 72 h later, and p63 and actin protein levels analyzed by western blot. (c) Conditioned medium from transfected NHEKs was collected 72 h after transfection and secreted IL1A concentration was determined by ELISA. (d) NHEK-conditioned medium (cm) was obtained by growing NHEKs for 48 h, filtered, and given as the only media to NHDFs for 24 h. Unconditioned NHEK medium (ucm) was used as a control. NHDFs were harvested 24 h after the addition of medium and harvested for RNA. Quantitative real-time PCR was used to quantify the expression of the fibroblast-specific growth factors KGF and FGF10 in response to NHEK-conditioned medium, relative to an unconditioned control. (e) NHDFs were treated as in (d) and GM-CSF levels were analyzed and graphed independently owing to the robust induction observed following cm addition. (f) NHEKs were transfected with NS or IL1A-specific siRNA and IL1A depletion was confirmed by qPCR 48 h after transfection. (g) NHEKs were transfected with p63 or IL1A siRNA alone or in combination, as well as non-silencing siRNA (NS) as a control. At 48 h after transfection, the media were replaced, conditioned medium was generated, and NHDFs were treated as in (d). KGF expression was measured by qPCR 24 h after the addition of conditioned medium. (h) NHDFs treated as in (g) and qPCR was used to analyze the expression of GM-CSF or FGF10 (i). (j) NHEKs were transduced with shuttle (sicon) or p63 siRNA- (sip63) expressing adenovirus. Cells were harvested, RNA isolated, and quantitative real-time PCR performed to analyze the expression of p63 and KGFR. All quantitative PCR data were normalized to actin levels and error bars represent standard deviation of at least three independent experiments. ^*P<0.5

Figure 5

TCF7L1 is a negative regulator of keratinocyte differentiation. (a) At confluency, rapidly growing monolayers of NHEKs were treated with 1.4 mM calcium to induce differentiation. Cultures were harvested at the indicated time points, RNA isolated, and qRT-PCR performed to determine the expression of p63 and TCF7L1 during differentiation. Expression levels were normalized to actin levels and presented relative to cultures at day 0. Error bars represent standard deviation of three replicate experiments. (b) NHEKs were transduced with control or ectopic TCF7L1-expressing lentivirus and induced to differentiate 48 h after infection. Cultures were collected at indicated time points and RNA was isolated for qPCR analysis of differentiation genes. Error bars represent standard deviation of three independent experiments. (c) NHEKs were treated as in (b), protein harvested at indicated time points, and levels of TCF7L1, p63, involucrin, loricrin, and actin analyzed by western blot

Figure 6

Loss of TCF7L1 results in early differentiation of NHEKs in OTC. (a) NHEKs were transfected with siRNAs targeting TCF7L1 or non-silencing oligos (NS) as a control. At 48 h after transfection, RNA was isolated and qRT-PCR performed to assess the expression of TCF7L1. (b) Immunofluorescent analysis of control and siTCF7L1 OTCs were performed with antibodies specific for filaggrin, involucrin, loricrin, keratin 2, and E-cadherin. Dotted line represents the epidermis–dermis interface. For proliferation analysis, OTCs were grown in medium containing BrdU for 16 h before harvesting in order to label proliferative cells. Immunofluorescence with a BrdU-specific antibody was performed to identify actively proliferating cells in the epidermal layers of OTCs. (c) BrdU-positive cells from (b) were quantified and graphed. Data shown are representative of three independent experiments