Lhx2 is a direct NF-κB target gene that promotes primary hair follicle placode down-growth

Abstract

In the epidermis of mice lacking transcription factor nuclear factor-kappa B (NF-κB) activity, primary hair follicle (HF) pre-placode formation is initiated without progression to proper placodes. NF-κB modulates WNT and SHH signaling at early stages of HF development, but this does not fully account for the phenotypes observed upon NF-κB inhibition. To identify additional NF-κB target genes, we developed a novel method to isolate and transcriptionally profile primary HF placodes with active NF-κB signaling. In parallel, we compared gene expression at the same developmental stage in NF-κB-deficient embryos and controls. This uncovered novel NF-κB target genes with potential roles in priming HF placodes for down-growth. Importantly, we identify Lhx2 (encoding a LIM/homeobox transcription factor) as a direct NF-κB target gene, loss of which replicates a subset of phenotypes seen in NF-κB-deficient embryos. Lhx2 and Tgfb2 knockout embryos exhibit very similar abnormalities in HF development, including failure of the E-cadherin suppression required for follicle down-growth. We show that TGFβ2 signaling is impaired in NF-κB-deficient and Lhx2 knockout embryos and that exogenous TGFβ2 rescues the HF phenotypes in Lhx2 knockout skin explants, indicating that it operates downstream of LHX2. These findings identify a novel NF-κB/LHX2/TGFβ2 signaling axis that is crucial for primary HF morphogenesis, which may also function more broadly in development and disease.

Keywords: Cell migration; E-cadherin; EDA-A1; EDAR; Embryo; Hair follicle; LHX2; Mouse; NF-κB; Placode; Stem cell; TGFβ2.

Fig. 1.

NF-κB-dependent gene signature in primary HF placodes reveals a multifunctional role. (A) Generation of an NF-κB-responsive EGFP reporter mouse line (κ-EGFP). EGFP expression was observed in the developing HF and in blood vessels of the skin at E14.5. Inset shows high magnification of skin with EGFP-expressing placodes. (B) Venn diagram illustrating overlap of genes up- or downregulated in primary HF placodes with differentially expressed genes in ΔN versus control epidermal keratinocytes. For full list of microarray data, see Table S1. (C) Primary HF placode-specific gene signature obtained from microarray analysis of sorted EGFP-positive keratinocytes at E14.5. Genes mentioned in the text are underlined. Functional categories with representative examples of mRNAs upregulated ≥1.5× are listed. Fold differences between placodes and interfollicular epidermis (EGFP negative) are indicated in parentheses. Genes highlighted in red were downregulated and in blue upregulated in ΔN-positive epidermal keratinocytes. Genes in black were specifically enriched in HF placodes, but not regulated by NF-κB in a significant manner. Note that Fn1 (in light red) expression is enriched 5× in HF placodes, but is only weakly downregulated in ΔN versus controls (∼1.3×). (D) Quantitative real-time PCR (qRT-PCR) analysis using either RNA samples from epidermal keratinocytes of n=3 control or ΔN embryos at E14.5 (left graph), or from FACS-sorted EGFP-positive (HF) or EGFP-negative (IFE) keratinocytes of κ-EGFP embryos at E14.5 (right graph). Statistical analyses were performed using a two-tailed unpaired t-test. Data are presented as mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001. (E) In situ hybridization for the indicated mRNA probes using sagittal skin sections of n=3 control and ΔN embryos at E14.5. Arrowheads indicate expression in HF placodes and dermal papilla (Nrp2); dashed line delineates the dermal-epidermal boundary. Scale bars: 50 µm.

Fig. 2.

Lhx2 is a direct target gene of NF-κB during primary placode formation and acts in concert with NF-κB to regulate genes involved in cell migration. (A) qRT-PCR of Lhx2 mRNA expression from epidermal keratinocytes of n=3 control or ΔN embryos (left graph), or from EGFP-positive placode and EGFP-negative IFE keratinocytes (right graph) at E14.5. (B) NF-κB p65-specific ChIP assays using EGFP-positive placode and EGFP-negative IFE keratinocyte extracts from κ-EGFP embryos at E14.5, and Lhx2 and control Gapdh primers. (C) Immunostaining on serial sagittal back skin sections of n=3 κ-EGFP and ΔN mice at E14.5 using antibodies against EGFP, P-cadherin (cadherin 3), LHX2 and KRT14. Arrowheads indicate expression in HF placodes. Blue, nuclear DAPI staining. (D) Analysis of primary HF development in Lhx2-KO mice revealed a dramatic reduction of stage 1 (E14.5; left graph) and stage 2 (E15.5; middle graph) primary HF. Overall primary HF density in Lhx2-KO embryos at E14.5 and E15.5 was reduced by ∼30% (right graph). The graphs show quantification from multiple back skin sections of three biological replicates. (E) qRT-PCR for selected NF-κB target genes involved in cell migration using mRNA isolated from epidermal keratinocytes of either Lhx2-KO or control embryos at E14.5. (F) In situ hybridization for Nrp2 and Prokr2 mRNA on sagittal skin sections of control and Lhx2-KO embryos at E14.5. Arrowheads indicate mRNA expression in HF placodes and also in the dermal papilla for Nrp2; dashed line delineates dermal-epidermal boundaries. Scale bars: 50 µm. All statistical analyses (A,B,D,E) were performed using two-tailed unpaired t-test. Data are presented as mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001.

Fig. 3.

ΔN and Lhx2-KO mice exhibit keratinocyte migratory and proliferative defects in the down-growing primary HF placode. Immunostaining on sagittal sections of n=3 control, Lhx2-KO and ΔN embryos at E14.5. Cytoskeletal organization and dynamics (F-actin labeled by Phalloidin), and phosphorylation of FAK (pFAK) were used as markers for cell migration. As readout for proliferation Ki67 expression was used. Antibodies against P-cadherin and EDAR (both red) were used to differentiate HF placodes. Arrowheads indicate Ki67 expression in HF placodes; dashed line delineates dermal-epidermal boundaries. Blue, nuclear DAPI staining. Scale bars: 20 µm.

Fig. 4.

Lhx2-KO mice exhibit delayed primary HF development and, as in ΔN mice, impaired TGFβ signaling and lack of E-cadherin downregulation. (A) Immunostaining on sagittal sections of n=3 control, Lhx2-KO and ΔN embryos at E14.5 using antibodies against anti-pSMAD2 (green), TGFβ2 (red) and E-cadherin (green). Arrowheads indicate presence or absence of pSMAD2 and TGFβ2 expression in the dermal papilla and of E-cadherin expression the HF placode; dashed line delineates dermal-epidermal boundaries. Blue, nuclear DAPI staining. (B) In situ hybridization for Fn1 mRNA on sagittal sections of n=3 control, Lhx2-KO and ΔN embryos at E14.5. Arrowheads indicate Fn1 mRNA expression in the HF placode; dashed line delineates dermal-epidermal boundaries. (C,D) qRT-PCR for Tgfb2 (C) and Fn1 (D) using mRNA isolated from epidermal keratinocytes of n=3 ΔN, Lhx2-KO or control embryos at E14.5. Data are presented as mean±s.e.m. Scale bars: 20 µm in A; 50 µm in B.

Fig. 5.

Treatment of Lhx2-KO embryonic skin explants with TGFβ2 rescued proper primary HF development. (A) Embryonic skin explants of n=3 E14.5 control or Lhx2-KO mice were either left untreated (ctrl) or treated with recombinant TGFβ2 (100 ng/ml) for 24 h. Skin samples were stained with H&E (right), and follicles were quantified as percentage relative to the number of follicles per microscopic field in the control group (left). Statistical analyses were performed using two-tailed unpaired t-test. Data were pooled from three biological replicates and presented as mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001. Arrowheads indicate HF placodes. Scale bars: 100 µm. (B) Immunostaining on Lhx2-KO explants treated with recombinant TGFβ2 or left untreated (control) using antibodies against pFAK, EDAR, Ki67, KRT14, E-cadherin and pSMAD2. Arrowheads indicate pFAK and the presence or absence of E-cadherin expression in the HF placode border; dashed line delineates dermal-epidermal boundaries. Blue, nuclear DAPI staining. Scale bars: 20 µm (pFAK, Ki67, E-cadherin); 50 µm (pSMAD2).

Fig. 6.

Model for primary placode down-growth involving NF-κB signaling. We show here that LHX2 expression is directly regulated by EDA-A1/EDAR/NF-κB signaling, as NF-κB activity in primary placodes depends on EDA-A1/EDAR (Schmidt-Ullrich et al., 2006). By an as yet unknown mechanism LHX2 activates TGFβ2 signaling in primary HF placodes, which results in phosphorylation of FAK (pFAK), and, importantly, in downregulation of E-cadherin expression. Together, these results introduce a novel NF-κB/LHX2/TGFβ2 signaling axis that is required for establishing the proper conditions for placode down-growth, and might also be relevant for other epidermal-mesenchymal tissue interactions, for epithelial-mesenchymal transition or for tumor growth.