AP-2 factors act in concert with Notch to orchestrate terminal differentiation in skin epidermis

Abstract

The mechanisms by which mammalian epidermal stem cells cease to proliferate and embark upon terminal differentiation are still poorly understood. By conditionally ablating two highly expressed transcription factors, AP-2alpha and AP-2gamma, we unmasked functional redundancies and discovered an essential role for AP-2s in the process. In vivo and in vitro, AP-2 deficiency is accompanied by surprisingly minimal changes in basal gene expression but severely perturbed terminal differentiation and suppression of additional transcription factors and structural genes involved. In dissecting the underlying molecular pathways, we uncover parallel pathways involving AP-2 and Notch signaling, which converge to govern CCAAT/enhancer binding protein genes and orchestrate the transition from basal proliferation to suprabasal differentiation. Finally, we extend the striking similarities in compromising either Notch signaling or AP-2alpha/AP-2gamma in developing skin to that in postnatal skin, where all hair follicles and sebaceous gland differentiation are also repressed and overt signs of premalignant conversion emerge.

Figure 1.

Targeted ablation of AP-2α and AP-2γ in mouse skin. γ-cKO, α-cKO, and DcKO mice were compared with WT littermates. (A) Immunolocalization of AP-2α and AP-2γ in P0 skins. Color coding of markers is according to secondary antibodies used in detection; nuclei were counterstained with DAPI (blue). Solid white lines denote skin surface. Dotted lines denote dermoepidermal boundary. Note that DP (arrows) also express AP-2γ. Bars, 20 μm. (B) Immunoblot analyses with monospecific antibodies and real-time PCR on mRNAs were performed on P0 epidermis, enzymatically separated from the dermis/HFs. Molecular mass of proteins is shown. PCR primer sets were specific for the AP-2α and AP-2γ sequences targeted for deletion. DcKO real-time PCR values were consistently below the level of detection from three or more independent experiments. (C) P0 mutant mice and their WT littermates. (D and E) Histology and ultrastructure of P0 epidermis. Vertical bars in D represent epidermis thickness. Boxed areas in E are magnified in bottom panels. BL, basal layer; Sp, spinous layer; Gr, granular layer; SC, stratum corneum; der, dermis; KF, keratin filament; KG, keratohyalin granules. Bars, 10 μm. (F) Barrier assay, as determined by penetration of blue dye.

Figure 2.

Embryonic suppression of terminal differentiation in vivo and in vitro upon loss of AP-2s. (A) Histology and immunofluorescence of the epidermis from E17.5 γ-cKO, DcKO, and WT skins. Note parakeratosis in DcKO epidermis (arrows). Bars, 20 μm. (B and E) Real-time PCR on mRNAs from embryonic epidermis and cultured 1°MKs. DcKO and α-cKO levels are expressed as fold changes versus WT (set to 1). Error bars represent standard deviation for three or more independent experiments. Cultures were in 50 μM Ca²⁺ (differentiation-restricted) media. (C) α6 BrdU FACS was performed on cells from epidermis of E18.5 mice pulsed 4 h with BrdU before analyses. Asterisk denotes significant difference (P < 0.05) by Student's t test. (D) Anti-active caspase 3 staining and quantification. Total number of cells counted in E18.5 epidermis are shown along with number of cells scoring positive for Cas3. (F) At 0 h, 1.5 mM Ca²⁺ was added to induce terminal differentiation. mRNAs were isolated at indicated time points and real-time PCR was performed. Lor, loricrin; Fil, filaggrin.

Figure 3.

Ectopic suprabasal genes are expressed in the absence of AP-2α and AP-2γ. (A) Immunofluorescence of K6 and K8 in skins from E17.5 embryos. Dotted lines denote dermoepidermal boundary. Bars, 20 μm. (B) Real-time PCR of mRNAs from epidermis and 1°MK as indicated. DcKO levels are expressed as changes versus WT and error bars represent standard deviation for three or more independent experiments.

Figure 4.

C/EBP gene expression is regulated by AP-2α and AP-2γ. (A) Immunofluorescence with monospecific C/EBPα and C/EBPβ antibodies in E18.5 skin. Dotted lines denote dermoepidermal boundary. Bars, 20 μm. (B) Immunoblots of protein lysates from 1°MK cultured under 50-μM or 1.5-mM Ca²⁺ conditions. The C/EBPα antibody detects both Cα p42 and Cα p30 isoforms; the C/EBPβ antibody detects LAP* and LAP isoforms. β-Actin is used as a loading control. (C) Analyses of the effects of AP-2 ablation on C/EBP gene expression and promoter activity. Real-time PCRs with primers specific for C/EBPα and C/EBPβ were performed on mRNAs from in vivo epidermis and 1°MK. DcKO levels are expressed as changes versus WT and error bars represent standard deviation for three or more independent experiments. (bottom right) Luciferase assay of C/EBPα promoter activity in 1°MK in high Ca²⁺ media. 1.5 kb of C/EBPα promoter plus 5′-UTR sequences were used to drive firefly luciferase expression (pC/EBPα-pGL3; Tang et al., 1997). Fold change represents pC/EBPα-pGL3 firefly luciferase activity divided by basal level of pGL3 firefly luciferase activity, with cytomegalovirus Renilla luciferase activity as the standardized internal control for transfection efficiency. Three independent experiments performed in duplicate are represented. Error bars represent standard deviations. (D) Effects of ectopic expression of AP-2α, C/EBPα, or C/EBPβ in DKO and WT 1°MK. Where indicated (vector), cells were either transduced 36 h with an IRES-GFP empty vector (control) or an IRES-GFP retroviral expression vector encoding the transcription factor indicated. Transduced cells were purified by FACS, mRNAs were isolated, and real-time PCRs were performed with the primers indicated.

Figure 5.

Notch signaling is active but cannot induce terminal differentiation in the absence of AP-2α and AP-2γ. (A) Real-time PCR analyses of WT and DKO 1°MK transduced with either an IRES-GFP empty vector (control) or an IRES-GFP retroviral vector encoding the NICD. After 36 h, transduced cells were FACS enriched and K14, K1, and K10 mRNA levels were measured. Fold change upon NICD induction was compared with the control vector. (B) Same experiment as in A, but documenting the mRNA levels of Hes1, Hey1, and RBP/J. (C) Same experiment as in A, but showing that AP-2 mRNA levels are only modestly affected by activated Notch. (D) RBP/J^lox/lox 1°MK were cultured in duplicate in low Ca²⁺ medium, and then transduced with either an adenovirus expressing a cytomegalovirus-Cre recombinase (Cre) or cytomegalovirus-GFP (control). At t = 2 d, one set of transduced cells were switched to high Ca²⁺ to induce differentiation. FACS was performed at t = 4 d and mRNAs were subjected to real-time PCR analyses with the primers indicated. Error bars represent standard deviation for three or more independent experiments.

Figure 6.

Regulation of C/EBPα gene expression by canonical Notch signaling pathway. Real-time PCR (A) and immunoblot analyses (B) of C/EBPα gene expression in control vector or Cre-infected RBP/J^lox/lox (R KO) 1°MK cultured in low or high Ca²⁺ media. (C) Same experiment as in A, except NICD-induced C/EBPα gene expression is measured. All experiments were performed three or more times, error bars represent standard deviation, and quantifications and details were as described in Fig. 2. (D) Real-time PCR and immunoblot analyses of C/EBPα gene expression in E18.5 WT versus K14-Cre/RBP/J^lox/lox cKO (R cKO) epidermis in vivo. (E) C/EBPα immunofluorescence. Bars, 20 μm. (F) WT and DKO keratinocytes were infected with lentivirus harboring either an empty vector (control) or a vector expressing a shRNA targeting ∼60% RBP/J knockdown. Infected cells were selected with puromycin and real-time PCR analyses were performed on these cells cultured at high Ca²⁺ conditions.

Figure 7.

Loss of AP-2α and AP-2γ blocks rather than delays differentiation in postnatal skin: similarities to loss of Notch function. Skins from E18.5 WT and DcKO mice were engrafted onto Nude mice. (A) Gross appearance at indicated days after grafting. (B–G) Histology and immunofluorescence of frozen skin sections from 27-d grafts. Arrows in C denote K1 in a few degenerating HFs. Dotted lines denote dermoepidermal boundary. Bars, 40 μm.