Hdac1 and Hdac2 act redundantly to control p63 and p53 functions in epidermal progenitor cells

Abstract

Epidermal and hair follicle development from surface ectodermal progenitor cells requires coordinated changes in gene expression. Histone deacetylases alter gene expression programs through modification of chromatin and transcription factors. We find that deletion of ectodermal Hdac1 and Hdac2 results in dramatic failure of hair follicle specification and epidermal proliferation and stratification, phenocopying loss of the key ectodermal transcription factor p63. Although expression of p63 and its positively regulated basal cell targets is maintained in Hdac1/2-deficient ectoderm, targets of p63-mediated repression, including p21, 14-3-3σ, and p16/INK4a, are ectopically expressed, and HDACs bind and are active at their promoter regions in normal undifferentiated keratinocytes. Mutant embryos display increased levels of acetylated p53, which opposes p63 functions, and p53 is required for HDAC inhibitor-mediated p21 expression in keratinocytes. Our data identify critical requirements for HDAC1/2 in epidermal development and indicate that HDAC1/2 directly mediate repressive functions of p63 and suppress p53 activity.

Figure 1. HDAC1 and HDAC2 display overlapping, dynamic expression in developing epidermis and hair follicles

Immunohistochemistry (brown) for HDAC1 and HDAC2 at the stages indicated. (A, B) HDAC1 and HDAC2 are expressed in all cells of the E13 single layer epidermis. (C, D) HDAC1 and HDAC2 are ubiquitously expressed in E14.5 epidermis and hair follicle placodes (arrows). (E, F) HDAC1 and HDAC2 are expressed most strongly at the leading edge of E16.5 hair follicles (black arrows) and in differentiating epidermal cells (blue arrows). (G–J) At E18.5, HDAC1 and HDAC2 are expressed strongly in differentiating epidermal cells (blue arrows). Dashed black lines indicate dermal-epidermal junctions. Scale bar in (J) applies to all images.

Figure 2. Embryos lacking epidermal HDAC1 and HDAC2 display striking defects in epidermal and ectodermal appendage development

(A–F) Immunofluorescence (green) reveals expression of HDAC1 and HDAC2 in the epidermis and developing hair follicles of control E14.5 and E16.5 Hdac1^fl/fl Hdac2^fl/fl littermate skin (A, C, E) and their absence in KRT14-Cre Hdac1^fl/fl Hdac2^fl/fl (DcKO) epidermis (B, D, F) (arrows). (G, H) Immunofluorescence (green) reveals increased levels of H3K9Ac in DcKO (H) compared with control (G) epidermis (arrows). (I, J) Open eyes, thin smooth skin, and failure of digit septation in a DcKO embryo at E18.5 (J) compared with a control KRT14-Cre Hdac^fl/+ Hdac2^fl/fl littermate (I) (arrows). (K–X) Hematoxylin and eosin staining reveals defects in stratification and ectodermal appendage development in DcKO compared with control littermate embryos. (K, L) absence of hair follicle placodes in E14.5 DcKO epidermis (L) compared with control skin (K, arrow). (M–P) DcKO epidermis remains as a single cell layer at E16.5 (M, N) and E18.5 (O, P) instead of stratifying, and lacks hair follicles. (Q–T) Abnormal DcKO dental lamina morphology at E16.5 (Q, R, arrows) and degeneration of incisor (black arrows) and molar (blue arrows) dental structures at E18.5 (S, T). (U, V) Failure of filiform papilla morphogenesis in E18.5 DcKO tongue surface ectoderm (arrows). (W, X) Sections through E18.5 hindlimbs reveal digit formation and failure of epidermal differentiation between the digits in the DcKO (arrows). Nuclei in (A–H) are DAPI-stained (blue). Dashed lines mark dermal-epidermal borders. Size bar in (F) applies to (A–F); size bar in (H) applies to (G, H); size bar in (P) applies to (K–P); size bars in (R, T, V, X) also apply to their respective littermate controls. See also Figure S1 and Figure S2.

Figure 3. Loss of epidermal Hdac1/2 causes failure of epidermal stratification, absence of hair follicle specification, and progressive defects in proliferation and cell survival

(A) Immunofluorescence (red) with antibodies to the basal layer marker KRT14, the suprabasal marker KRT10 and the granular layer marker loricrin, reveals expression of KRT14, but lack of KRT10 and loricrin in DcKO epidermis at E14.5 E18.5. The weak signal in E16.5 DcKO skin incubated with anti-KRT10 is due to background fluorescence in the periderm. (B) Whole mount in-situ hybridization (blue-purple) for Ctnnb1 reveals formation of hair follicle pre-placodes in E14.0 control littermate skin (arrows), and their absence in DcKO skin. (C) Immunohistochemistry for β-catenin (brown) and immunofluorescence for LEF1 (green) reveal upregulated expression in developing hair follicle placodes (black or white arrows), and lower or absent expression in inter-placodal epidermis (blue or yellow arrows) in controls, and uniform low level expression in DcKO epidermis (black or white arrows). (D) Immunofluorescence for phospho-histone H3 (PH3) (green) reveals decreased proliferation in DcKO epidermis at E14.5 and almost absent proliferation at E16.5 (arrows). E16.5 samples were co-stained with anti-KRT14 (red) to mark basal cells. (E) TUNEL staining (green) reveals increased cell death in DcKO embryos at E16.5 (arrows) but not at E14.5 compared to littermate controls. (F) Quantification of the percentage of PH3 positive basal layer cells reveals a statistically significant decrease in proliferation in DcKO compared with control embryos at E14.5, and a further decrease in the relative rate of proliferation at E16.5. 20 fields of view were counted for each genotype. (G) Quantification of the number of TUNEL positive cells per field of view at 40× magnification reveals a statistically significant increase in the number of apoptotic basal cells in DcKO embryos at E16.5, but not at E14.5. 20 fields of view were counted for each genotype. Statistical data are represented as mean +/− SEM. Dashed lines indicate dermal-epidermal junctions. NS, not statistically significant. See also Figure S3.

Figure 4. Hdac1 and Hdac2 are required for repression of negatively regulated ΔNP63 target genes

(A) Immunofluorescence with pan-p63 antibody (green) reveals expression in the basal layer of littermate controls and in the single layered DcKO epidermis at E14.5, E16.5 and E18.5 (arrows). (B-E) Expression of ΔNP63 (B, C, arrows) and undetectable expression of TAP63 (D, E) in both control (B, D) and DcKO (C, E) epidermis at E14.5. (F–K) Positively regulated p63 targets FASN (F, G) (green), CDC25c (H, I) (brown), and claudin1 (J, K) (green) and are similarly expressed in single-layered DcKO (G, I, K) and basal control (F, H, J) epidermis (arrows). (L–O) Immunofluorescence for 14-3-3σ at E14.5 (L, M) (green signal) and E16.5 (N, O) (red signal) reveals its expression in control suprabasal cells (N, yellow arrow) and absence in control basal cells (N, white arrow). In DcKO skin, 14-3-3σ is expressed in the single-layered epithelium at both E14.5 (M) and E16.5 (O) (yellow arrow). (P–S) p21 immunostaining (brown) is slightly more intense in DcKO (Q, S) compared with control basal epidermis (P, R) at both E14.5 (P, Q) and E16.5 (R, S) (arrows). (T–W) p16/INK4a (brown immunostaining) is dramatically upregulated in DcKO (U, W) compared with control (T, V) epidermis at both E14.5 (T, U) and E16.5 (V, W) (arrows). Dashed lines indicate dermal-epidermal boundaries. (X) CHiP assays with extracts of untreated HEKs and HEKs treated with TSA for 6 hours or 12 hours, and antibodies to HDAC1, HDAC2, H3K9Ac, or rabbit IgG negative control, as indicated. Primers amplifying sequences that encompass GAPDH negative control promoter regions or p63 binding sites in the promoter regions of the 14-3-3σ, p21, and p16/INK4a genes were used for PCR as indicated. All panels in (A) were photographed at the same magnification; scale bar in (E) applies to (B–E); scale bars in (G, I, K, M, O) also apply to their respective littermate controls; scale bar in (W) applies to (P–W). See also Figure S4 and Supplemental Table.

Figure 5. HDAC1/2 suppress hyper-acetylation of p53 in embryonic epidermis

(A–D) PCNA (green) is expressed in control basal cell and some suprabasal cell nuclei at E14.5 (A) and E16.5 (C, arrows). PCNA is expressed in DcKO epidermis at E14.5 (B), but is downregulated by E16.5 (D). (E–H) GADD45A (green) is expressed in control suprabasal but not basal cells at E14.5 (E) and E16.5 (G). In single-layered DcKO epidermis, GADD45A is weakly expressed at E14.5 (F) and gives a strong signal at E16.5 (H, arrows). (I–P) p53-AcK379 (I–L) and p53-AcK386 (M–P) (green) are undetectable in control basal nuclei at E14.5 (I, M) or E16.5 (K, O), and are weakly detected in a subset of differentiating nuclei at E16.5 (K, O). DcKO epidermis displays strong staining for acetylated p53-AcK379 (J, L) and p53-AcK386 (N, P) at both E14.5 (J, N) and E16.5 (L, P) (arrows). Dashed lines indicate dermal-epidermal junctions. Size bars represent 25μm.

Figure 6. p53 is required for elevated p21 expression in response to HDAC inhibition in keratinocytes

(A–F) TSA treatment of HEKs causes decreased proliferation, and increased expression of acetylated p53 and p21. HEKs treated with 500nM TSA in DMSO (D–F) or DMSO alone (A–C) were subjected to immunofluorescence (green) using antibodies to PH3 (A, D), p53-AcK379 (B, E) or p21 (C, F). (G) The TSA-induced decrease in proliferation requires p53 and p21. HEKs were transduced with retroviruses carrying control knockdown (KD), p53 KD or p21 KD shRNAmir constructs as indicated, and treated with 50nm TSA in DMSO (light grey bars) or DMSO only (dark grey bars). The percentage of PH3-positive nuclei is indicated for each condition. 10 fields of view were counted per sample at 40x magnification and the experiment was performed in duplicate. (H–N) p53 is required for the TSA-induced increase in p21 expression. HEKs were transduced with retroviruses carrying control (H, K), p53 (I, M) or p21 (J, L) knockdown shRNAmirs and treated with 50nM TSA. Immunofluorescence with antibodies to total p53 (H–J) (red) or p21 (K–M) (green), reveals efficient knockdown of p53 or p21 with the corresponding shRNAmir (H, I, K, L), and decreased p21 expression in p53 knockdown compared with control knockdown cells (K, M). Knockdown of p21 does not affect expression of p53 (H, J). (N) Quantification of p21 mRNA levels by real-time PCR reveals a statistically significant increase in p21 mRNA expression in TSA-treated control knockdown cells compared with DMSO-treated control knockdown cells (dark grey bars) (p<0.05), and no TSA-induced increase in p21 expression in either p21 (light grey bars) or p53 (medium grey bars) knockdown cells. (O) Model for interactions of HDAC1/2, ΔNp63 and p53 in embryonic epidermis. ΔNp63 and HDAC1/2 cooperate to directly repress expression of p16/INK4a and may also directly contribute to repression of p21. HDAC1/2 independently deacetylate p53, suppressing its ability to induce apoptosis and activate target genes including p21. Data in (G, N) are represented as mean +/− SEM; asterisks indicate statistical significance; NS, not statistically significant. See also Figure S5.