|Home | About | Journals | Submit | Contact Us | Français|
The human congenital syndromes ectrodactyly ectodermal dysplasia-cleft lip/palate syndrome, ankyloblepharon ectodermal dysplasia clefting, and split-hand/foot malformation are all characterized by ectodermal dysplasia, limb malformations, and cleft lip/palate. These phenotypic features are a result of an imbalance between the proliferation and differentiation of precursor cells during development of ectoderm-derived structures. Mutations in the p63 and interferon regulatory factor 6 (IRF6) genes have been found in human patients with these syndromes, consistent with phenotypes. Here, we used human and mouse primary keratinocytes and mouse models to investigate the role of p63 and IRF6 in proliferation and differentiation. We report that the ΔNp63 isoform of p63 activated transcription of IRF6, and this, in turn, induced proteasome-mediated ΔNp63 degradation. This feedback regulatory loop allowed keratinocytes to exit the cell cycle, thereby limiting their ability to proliferate. Importantly, mutations in either p63 or IRF6 resulted in disruption of this regulatory loop: p63 mutations causing ectodermal dysplasias were unable to activate IRF6 transcription, and mice with mutated or null p63 showed reduced Irf6 expression in their palate and ectoderm. These results identify what we believe to be a novel mechanism that regulates the proliferation-differentiation balance of keratinocytes essential for palate fusion and skin differentiation and links the pathogenesis of 2 genetically different groups of ectodermal dysplasia syndromes into a common molecular pathway.
The p53-related transcriptional activator p63 is expressed in basal cells of stratified epithelia, in myoepithelial cells of the breast and salivary glands, and in the proliferative compartment of gastric mucosa (1, 2). As a result of the alternative usage of 2 promoters and of complex alternative splicing, the p63 gene encodes 6 isoforms (1). Of these, the transactivation (TA) isoforms contain an N-terminal TA domain, which is absent in the ΔN isoforms (1). In ΔNp63 isoforms, an additional TA domain has been recognized in the C terminus, which is responsible for ΔN-specific transcriptional activities distinct from that of TA isoforms (3, 4). ΔNp63 isoforms are expressed in cells of stratified epithelia and contribute to the regulation of the proliferative potential of epithelial stem cells and to epidermal morphogenesis (5–7). The critical role for p63 in regulating epidermal morphogenesis is illustrated by the phenotype of p63–/– mice, which fail to develop an epidermis, other stratified epithelia, and epithelial appendages (2, 8). The single layer of epithelial cells covering p63–/– mice at birth fails to provide barrier function, resulting in early postnatal lethality due to severe dehydration.
Mutations in the p63 gene are found in a number of dominantly inherited human congenital disorders, including ectrodactyly ectodermal dysplasia-cleft lip/palate syndrome (EEC), ankyloblepharon ectodermal dysplasia clefting (AEC), and nonsyndromic split-hand/foot malformation (SHFM) (9, 10). These syndromes share several phenotypic features, consisting of ectodermal dysplasia, limb malformations, and cleft lip/palate, that can be related to alterations of a tightly controlled balance between proliferation and differentiation of precursor cells during the development of ectoderm-derived epithelia and organs (10). Since p63 is a transcription factor, the molecular basis of these defects most likely resides in the inability of mutated p63 proteins to properly activate/repress expression of target genes, which are beginning to be identified. Indeed, phenotypic defects found in p63-linked ectodermal dysplasias are common to other congenital syndromes caused by mutation in p63 target genes, such as the DLX genes, whose mutation/misregulation is associated with SHFMI (DLX5 and DLX6) (11) or with the trichodentoosseous syndrome (DLX3) (12), or Claudin-1, whose mutation leads to the NISCH syndrome (13). Interferon regulatory factor 6 (IRF6) is a member of a conserved family of IFN-dependent transcription factors known to regulate the proliferation-differentiation switch in epidermal cells and shown to be required for palate closure during embryonic development (14–16). Mutations in this gene cause the van der Woude and the popliteal pterygium syndromes (17), conditions resembling p63-linked ectodermal dysplasias in many aspects (10).
Here we show that IRF6 is a p63 transcriptional target that limits keratinocyte proliferation by inducing ΔNp63 proteasome-mediated degradation. Mutations in p63 or IRF6 disrupt this regulatory loop, altering the critical balance between differentiation and proliferation during development, leading to clinically evident defects.
In order to identify genes specifically regulated by TAp63 or ΔNp63, we performed transcriptional profiling of primary mouse keratinocytes, in which TAp63 or ΔNp63 were selectively depleted by siRNA (13) (Supplemental Figure 1, A and B; supplemental material available online with this article; doi: 10.1172/JCI40267DS1). In addition to known p63 targets (e.g., scotin, stratifin) (Supplemental Table 1), we identified Irf6 as a positively regulated target in ΔNp63-depleted cells but not in TAp63-depleted cells (Figure (Figure1,1, A and B).
Using ChIP assays, we determined that endogenous Irf6 is a direct p63 target in keratinocytes. We identified 2 p63-like consensus binding sites upstream of the transcription initiation site and in the first Irf6 intron (Figure (Figure1C);1C); both sites directly interact with p63 during calcium-induced differentiation of primary mouse keratinocytes (Figure (Figure1D).1D). Binding of p63 onto these Irf6 responsive elements (REs) precedes a marked increase in Irf6 mRNA and Irf6 protein levels (Figure (Figure1,1, E and F); moreover, exogenously expressed ΔNp63 increased Irf6 transcript levels (Figure (Figure1G)1G) and induced a reporter gene driven by an Irf6 promoter (Supplemental Figure 1C). In contrast, an Irf6 promoter, with the RE mutated so that it could not bind ΔNp63 in oligo pull-down experiments, was activated to a lesser degree by exogenously expressed ΔNp63 (Supplemental Figure 1, C and D).
To determine whether Irf6 expression is p63-dependent in vivo, we examined the distribution of Irf6 mRNA using in situ hybridization on sections from wild-type and p63–/– mice (8). As expected, the palatal and tooth epithelia of p63–/– embryos displayed reduced expression of Irf6 compared with the heterozygous or wild-type embryos. Expression of the mesenchymal marker Msx1 was not significantly changed (Figure (Figure2A),2A), demonstrating the correspondence of the anatomical structures and indicating that the initial steps of their morphogenesis were normally induced (18). Reduction of Irf6 expression was confirmed by real-time quantitative PCR (qPCR), performed on p63–/– embryonic ectoderm (Figure (Figure2B). 2B).
We also determined Irf6 mRNA expression levels in mouse embryos containing the knock-in p63 mutation R279H (found in EEC patients) at various developmental ages. Animals homozygous for this mutation were very similar to the null ones, in terms of skin, limb, and craniofacial defects, and displayed reduced Irf6 mRNA expression in the epidermis (Figure (Figure2B)2B) and in the tooth and palate epithelia (data not shown). Interestingly, animals heterozygous for this mutation show a low-penetrance, EEC-related phenotypic spectrum, including cleft palate and ectrodactyly, that more closely resembles the human EEC syndrome than that of p63-null mice (ref. 19 and A.A. Mills et al., unpublished observations). The skin of p63+/EEC embryos (Figure (Figure2C)2C) and newborns showed patches of hyperproliferating epidermis; in these patches, IHC staining revealed a marked reduction of Irf6 nuclear expression, as compared with the epidermis of wild-type mice (Figure (Figure2C,2C, left), and a marked increase in p63 nuclear expression (Figure (Figure2C,2C, right). Consistently, exogenous expression of ΔNp63s carrying either the R279H or the Q536L mutations (the latter associated to the AEC syndrome) in human keratinocytes resulted in a reduced IRF6 expression compared with wild-type ΔNp63 (Figure (Figure3A).3A). These findings were extended to human skin by examining IRF6 expression in biopsies from an AEC patient (carrying p63 I537T mutation) (13, 20). As shown in Figure Figure3B,3B, the AEC epidermis was parakeratotic and displayed strong reduction in IRF6 immunoreactivity as compared with normal skin, also confirmed at the mRNA level (Figure (Figure3C).3C). This observation correlated with abnormally high levels of p63, ectopically detected in the upper epidermal layers, including the flattened cells of intermediate zone between the spinous layer and the parakeratotic stratum corneum (Figure (Figure3B). 3B).
The developmental defects observed in Irf6 mutant mice are thought to result from the inability of Irf6 mutant cells to exit the cell cycle, leading to an undifferentiated hyperproliferative epidermis (14, 15). In addition, mice heterozygous for p63 and the Irf6 mutation R84C display cleft of the secondary palate, as the medial edge epithelial cells fail to undergo their normal differentiation program (21). In this regard, ΔNp63 is a key regulator of epithelial cell proliferative potential and is upregulated in the epidermis of Irf6 mutant mice (14, 15). We also noted a reciprocal relationship between IRF6 and p63 protein expression during calcium-induced differentiation of human primary keratinocytes (Figure (Figure1F)1F) and in immunofluorescence experiments performed on normal human epidermis (Supplemental Figure 2). Therefore, we reasoned that, following the initial activation of IRF6 expression by ΔNp63, IRF6 then promotes the downregulation of ΔNp63, and that this regulation might favor the exit of keratinocytes from the cell cycle. To test this hypothesis, we exogenously expressed IRF6 in normal keratinocytes and analyzed the endogenous expression of ΔNp63. The results indicate that IRF6-overexpressing cells display reduced levels of ΔNp63 (Figure (Figure4A4A and Table Table1). 1).
Accordingly, ΔNp63 mRNA and protein expression is markedly reduced during keratinocyte differentiation (1, 22). To evaluate the possible role of IRF6 in differentiation-induced ΔNp63 downregulation, we depleted IRF6 in human primary keratinocytes using siRNA. We observed that, in IRF6-depleted proliferating keratinocytes, the amount of ΔNp63α was similar to that in control cells, whereas differentiating keratinocytes failed to induce full ΔNp63 downregulation and also displayed higher levels of the proliferation marker PCNA (Figure (Figure4B).4B). IRF6 depletion caused inhibition of terminal differentiation, as observed by reduced levels of loricrin (LOR) mRNA in differentiating IRF6-depleted cells (Supplemental Figure 3, A and B).
Downregulation of ΔNp63 during keratinocyte differentiation is due, at least in part, to the expression of a specific miRNA (miR203) (23, 24). However, siRNA-mediated depletion of IRF6 had no significant effect on miR203 expression, indicating that IRF6 is unlikely to be involved in this mechanism (Supplemental Figure 3C). An alternative means to downregulate ΔNp63 expression is by proteasome-mediated degradation (25, 26), and indeed, we found that IRF6-induced ΔNp63 protein degradation was mediated by the proteasome, being blocked by treatment with the specific inhibitor MG132 (Figure (Figure4C),4C), which does not affect exogenous ΔNp63 mRNA level (Supplemental Figure 3D). Thus, IRF6 targets ΔNp63α and ΔNp63β isoforms for degradation but not ΔNp63γ or TAp63α isoforms (Supplemental Figure 4).
In addition to altered transcriptional properties, mutated ΔNp63 might also have different sensitivity to IRF6-dependent downregulation. We therefore tested the effect of IRF6 overexpression on the stability of the ΔNp63-R279H (EEC mutant) protein and found that this mutant is resistant to IRF6-mediated downregulation (Figure (Figure4D).4D). Conversely, IRF6 mutant R84C, a mutation causing van der Woude syndrome, was unable to induce ΔNp63 downregulation (Figure (Figure4E).4E). This observation correlates with increased levels of p63 in the perfusion medial edge epithelia and the midline epithelial seam of Irf6-R84C knock-in mice palates (21). Thus, disease-related p63- and IRF6-mutant proteins are unable to establish a correct biochemical regulation, and we can postulate that in epithelial cells of patients affected by these mutations, the amount of ΔNp63 remains abnormally high, consequent to alteration of this regulatory pathway. Failure of p63 to be downregulated at the appropriate time may result in increased proliferation of epidermal keratinocytes or cells of the embryonic palate.
To establish a role for IRF6 in epithelial cell proliferation, we determined the colony-forming efficiency of epithelial cells (TE13 cells) in which IRF6 was exogenously expressed. IRF6 overexpression resulted in a marked reduction in colony formation efficiency (Figure (Figure5A).5A). In the same assay, coexpression of normal IRF6 with the R279H mutated ΔNp63α, resistant to IRF6-dependent degradation, restored the colony-forming efficiency to a normal level, while wild-type ΔNp63α did not (Figure (Figure5A).5A). These results suggest that inhibition of epithelial cell growth requires downregulation of ΔNp63 protein, consequent to an increased IRF6 expression, and that disease-causing p63 mutations compromise this regulation.
To establish a role for IRF6 in the control of the proliferative potential of epidermal cells, we adopted the paradigm developed for therapeutic epidermal transplants on human epidermis, which estimates the self-renewal potential of epithelial cells via the generation of holoclones, meroclones, and paraclones in vitro (27, 28). We transduced primary human keratinocytes with retroviruses expressing IRF6 and observed that both the number and the size of clones generated by IRF6-expressing precursors were significantly lower than those obtained from control-infected cells (Figure 5B and Table 2). Thus, IRF6 expression reduces the proliferative potential of epidermal cells.
IRF6 is a transcription factor, mutated in a subset of developmental syndromes, characterized by cleft lip and ectodermal dysplasia (28). The first important finding reported here is that in keratinocytes IRF6 is a direct transcriptional target of ΔNp63 but not of TAp63. ΔNp63 maintains basal transcription levels of IRF6 via binding to distal p63 REs (21) and is promptly recruited to the proximal p63 element upon keratinocyte differentiation stimuli. This regulation most likely takes place in epithelial cells with high proliferative potential, as ΔNp63 is the predominant isoform detectable in this compartment. Importantly, mutations in p63, specifically those found in patients with the EEC and AEC syndromes, strongly reduce or abolish the ability of the ΔN protein to activate IRF6 transcription, and this correlates nicely with the reduced level of Irf6 protein expression found in the epidermis of p63R279H heterozygous animals and in the epidermis of the AEC patient we examined. Together, these observations suggest that misregulation of IRF6 expression contributes to development of the clinical phenotype of p63-related syndromes.
The second important finding is that, during epithelial cell differentiation, IRF6 promotes proteasome-dependent degradation of ΔNp63, thus establishing a cell-autonomous negative regulatory loop (Figure (Figure4C4C and Figure Figure5C).5C). During early differentiation, ΔNp63 promotes transcription of IRF6, and the IRF6 protein in turn promotes ΔNp63 degradation. Considering that ΔNp63 critically endows the epidermal cells with an elevated proliferation potential (7), this regulation may have the functional consequence of posing a limit to keratinocyte proliferation and allowing keratinocyte terminal differentiation and palate closure during embryonic development. Indeed, we show that increased IRF6 expression results in lower clonogenicity and reduced proliferation of cultured epithelial cells.
Importantly, the mutant IRF6 proteins found in patients with the van der Woude syndrome are unable to bind to their cognate REs (29) and cannot induce p63 degradation. Thus, in these patients, abnormally high levels of ΔNp63 protein are retained in progenitor cells, with pathological consequences. The R279H mutant p63 is also resistant to IRF6-mediated degradation, thus providing further evidence that the disruption or misfunctioning of the regulatory loop among these transcription factors, either resulting in hyperactivity or in hypoactivity, has pathological effects.
We propose that in normal epidermal cells the negative regulatory loop indicated by our model is initiated by ΔNp63 binding to the Irf6 promoter and is subsequently needed for the exit of the cells from the cell-cycle and for the initiation of a tissue-specific differentiation program. Alterations in this molecular pathway are likely to affect the execution of a normal differentiation program, essential for the correct maturation and stratification of the epidermis and for the morphogenesis of structures requiring epithelial-mesenchymal interaction, such as lip/palate closure and limb bud patterning and growth.
In summary, our data highlight the importance of a developmental feedback regulatory loop between p63 and IRF6 for the control of proliferation of epithelial cells and link two phenotypically similar, but genetically distinct, groups of ectodermal dysplasias into one common pathway. Future studies aimed at modulating this pathway for therapeutic purposes will require the identification of tissue-specific IRF6 targets and the full comprehension of the molecular mechanisms that tightly regulate p63 stability/degradation.
Mouse and human primary keratinocytes were cultured as described previously. Use of mouse and human keratinocytes was approved by the University of Torino and the University of Rome “Tor Vergata” Institutional Review Boards, respectively (30). TE13 cells (epithelial SCC cell line; gift of Pierre Hainaut, International Agency for Research on Cancer, Lyon, France) and H1299 cells were cultured in RPMI medium containing 10% FBS and antibiotics (100 U/ml penicillin and 10 μg/ml streptomycin) at 37°C. To measure clonogenic growth, TE13 cells transfected with Lipofectamine (Invitrogen) were selected in G418 (600 μg/ml) (Invitrogen) for 15 days, before fixation in 4% PFA and staining with Crystal Violet. Colonies with a diameter of more than 1 mm were scored using NIH Image software ( http://rsb.info.nih.gov/nih-image/). U2OS cells were cultured in DMEM containing 10% FBS and antibiotics (100 U/ml penicillin and 10 μg/ml streptomycin) at 37°C.
Expression plasmids encoding for Myc-tagged wild-type TA-ΔN-p63s and the ΔNp63α mutant R279H (EEC) have been reported previously (16). Human HA-tagged IRF6 expression plasmid was constructed by cloning the human IRF6 ORF into the pCDNA-HA vector. The R84C mutation was introduced in IRF6 by custom mutagenesis service (Biofab Research Inc.). Inserts were sequence verified.
A retroviral (pBABE-puro) expression vector with the HA-IRF6 ORF was derived by PCR amplification from the pCDNA plasmid. Luciferase reporter plasmid and retrovirus production are described in the Supplemental Methods.
For ChIP assays, mouse primary keratinocytes were used. ChIP assays were performed as described previously (20). The mouse monoclonal antibody anti-p63 (4A4) (Santa Cruz Biotechnology Inc.) or isotypic control antibodies (monoclonal anti-HA [F-7] antibody; Santa Cruz Biotechnology Inc.) were used. PCR reactions were performed for 25–35 cycles, with an annealing temperature of 55°C–57°C. Primers used in ChIP assay are indicated in the Supplemental Methods.
siRNA duplexes targeting ΔNp63, TAp63, and GFP, used for gene expression array (Figure (Figure1,1, A and B), were obtained from MWG-Biotech. Mouse primary keratinocytes, plated on collagen-coated 35-mm dishes, were transfected with 0.5 μg siRNA per dish using Lipofectamine 2000 (Invitrogen). For experiments described in Figure Figure3B3B and in Supplemental Figure 2, we used stealth siRNAs (Invitrogen). siRNA duplexes targeting human IRF6 (catalog no. 1299003; Invitrogen) were used to knockdown IRF6 mRNA; scrambled siRNA (Invitrogen) were used as control. Human primary keratinocytes were plated on collagen-coated 35-mm dishes and transfected with 50 nM siRNAs using Lipofectamine 2000.
Total RNA extracted from control and ΔNp63- and TAp63-depleted human primary keratinocytes was used to synthesize biotinylated complementary RNA (cRNA), using the GeneChip One-Cycle Target Labeling and Control Reagents kit (Affymetrix). Fifteen μg of fragmented cRNA were used for hybridization to GeneChip Mouse Genome 430 2.0 arrays (Affymetrix), for analysis of over 39,000 transcripts. Scanning and analysis were performed as described previously (28). Analysis criteria are described in the Supplemental Methods.
The following antibodies were used: monoclonal antibodies against HA (F-7) and Myc (9E10) epitopes, p63 (4A4), polyclonal antibodies against PCNA (PC10), and actin (C11) were purchased from Santa Cruz Biotechnology Inc.; anti-keratin (anti-K1) was purchased from Covance; mouse polyclonal anti-IRF6 from was purchased from Abnova Inc.; and goat polyclonal anti-IRF6 antibody (IMG3484) was purchased from Imgenex. For immunoblotting, cells were lysed in 50 mM Tris, pH 8.0, 120 mM NaCl, and 0.5% NP-40, and protein concentration was determined. Forty micrograms of total cell extract were separated by SDS-PAGE, blotted, incubated with the antibody, and developed with the ECL system (Amersham Pharmacia Biotech). The MG132 inhibitor was from Sigma-Aldrich.
For RT-qPCR, total RNA was extracted using TRIzol Reagent (Invitrogen). One μg of total RNA was reverse transcribed with GeneAmp RNA PCR (Applied Biosystems). For RT-qPCR in human cells, the RT reaction products were used for real-time qPCR amplification, which was performed with the MyiQ Single-Color Real-Time Detection System (Bio-Rad). The human acidic ribosomal protein P0 (hARP) gene was used for normalization. Total RNA was extracted from formalin-fixed, paraffin-embedded sections using the Optimum FFPE RNA Isolation Kit (Ambion), according to the manufacturer’s protocol. Primer sequences are available in the Supplemental Methods.
Four-mm punch biopsy specimens were taken from normal skin of healthy volunteers (n = 4). Declaration of Helsinki protocols were followed, and patients gave written approved consent before biopsy. Care and use of human tissues was approved by the University of Rome “Tor Vergata” Institutional Review Board. A patient affected by AEC (I537T mutation in p63) was subjected to skin biopsy on lesional skin, after obtaining informed consent from his parents. Immunofluorescence analysis was performed on 5 × 104 to 10 × 104 cells seeded on Chamber Slides (Nalge Nunc International), fixed with 4% PFA, and permeabilized with 0.2% Triton X-100. Immunohistochemistry and immunofluorescence were carried out by standard methods (Supplemental Methods).
The Brdm2 mouse strain, also designated p63– (8), that carries a null p63 allele was genotyped by RT-PCR on embryonic RNA, with primers annealing to the C terminus of the α isoform (Trp63) mRNA. In mice carrying the p63R279H mutation (an EEC mutation in humans) (19), a BsaHI restriction site was abrogated and was used to determine the genotype, following PCR amplification with primers flanking codon 279. The genotype was determined by PCR amplification with primers flanking codon 279, followed by digestion with BsaHI. In situ hybridization was performed as described previously (31) and as described in the Supplemental Methods.
Statistical evaluations were carried out using SigmaStat 2.03 (SPSS). For all tests, P values of less than 0.05 were considered significant.
We thank Antonio Ribuffo, the Master of Roman University School of Dermatology. We thank Pina Ricca for optimal technical assistance and Gioacchino Natoli, Giovanni Monteleone, and Maranke Koster for critically reading the manuscript and for helpful suggestions. This work was supported by grants from AIRC, Italian Ministry of Research (FIRB-IDEAS grant), Wyeth (Advances in Psoriasis Research grant), and European Union grants (FP6 “Active p53 Consortium”) to A. Costanzo, Telethon GGP05056 to L. Guerrini, and Telethon Career (Dulbecco Institute) S99003 to G. Merlo.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest doi:10.1172/JCI40267.
See the related Commentary p63 and IRF6: brothers in arms against cleft palate.