Homeobox transcription factor DLX4 is not necessary for skin development and homeostasis

Abstract

Dlx4 is a member of a family of homeobox genes with homology to Drosophila distal-less (dll) gene. We show that Dlx4 expression pattern partially overlaps with its cis-linked gene Dlx3 during mouse development as well as in neonatal and adult skin. In mice, Dlx4 is expressed in the branchial arches, embryonic limbs, digits, nose, hair follicle and in the basal and suprabasal layers of mouse interfollicular epidermis in neonatal and adult skin. We show that inactivation of Dlx4 in mice did not result in any overtly gross pathology. Skin development, homeostasis and response to TPA treatment was similar in mice with loss of Dlx4 compared to wild type counterparts.

Graphical Abstract

Dlx4 gene belongs to a group of homeodomain transcription factors that are involved in patterning and differentiation during development. Herein, we report that DLX4 is expressed in epidermis and its function in skin is dispensable since ablation of the gene results in mice that are phenotypically indistinguishable from wildtype littermates.

BACKGROUND

The DLX family of homeobox genes is a group of transcription factors that control cellular patterning and differentiation during development [1–3]. In human and mouse, the DLX family is comprised of six members which have homology to Drosophila distal less (dll) gene [1–3]. The six paralog genes are organized as bigene clusters as Dlx1-Dlx2, Dlx3-Dlx4 and Dlx5-Dlx6 [1–3]. Sequence and expression comparisons show Dlx1-Dlx2 and Dlx5-Dlx6 as closely related clusters with Dlx3-Dlx4 being the most divergent pair [4]. Targeted inactivation of Dlx1, Dlx2, Dlx5 and Dlx6 demonstrated their role in craniofacial, limb and skeletal development [1, 5, 6]. While deletion of Dlx3 function resulted in embryonic lethality at E9.5 of development due to placental failure [7], epidermal-specific deletion of Dlx3 in mice resulted in skin, teeth and hair follicle abnormalities [8–13]. The expression of Dlx4 during development has been characterized, yet its function in early mouse development and adulthood has not been determined [4, 14]. Here we report that genetic ablation of Dlx4 does not overtly affect mouse development, nor does it alter skin formation and homeostasis.

QUESTIONS ADDRESSED

In this present study, we examined Dlx4 deletion in mouse development. DLX3, which is a cis-linked gene of Dlx4, is actively expressed in the hair follicle, the differentiating epidermal layers of skin and has an important regulatory role in hair development and keratinocyte differentiation and maintenance of skin barrier function [8–10]. Therefore, we specifically investigated the consequences of Dlx4 deletion in mouse skin.

EXPERIMENTAL DESIGN

Supporting Information (Appendix S1)

RESULTS

Loss of Dlx4 does not lead to major developmental defects

We generated a Dlx4 knockout mouse model (hereafter called Dlx4 KO) by recombining the β-galactosidase gene into the exon 2 of the Dlx4 locus (Fig. 1A). The excision of exon 2 was validated by PCR (Fig. 1B). In mice heterozygous for Dlx4, we detected β-galactosidase activity in branchial arches, limb bones, nose and digits (Fig. 1C), confirming previously reported sites of Dlx4 expression during development [3, 4, 14]. Dlx4 KO mice were healthy and did not exhibit any overt altered phenotype when compared to wild type littermates (Fig. S1A). In addition, skeletal staining at E18.5 did not reveal any defects in bone (Fig. S1B).

Figure 1. Dlx4 is dispensable for epidermal development and barrier formation.

(A) Schematic representation of targeting mouse Dlx4 locus by replacing exon 2 with LacZ. (B) Genotyping of littermate mice by genomic PCR. The lower and upper bands indicate the fragments corresponding to wild-type (WT) allele and LacZ (Dlx4 KO) allele respectively. (C) Whole mount beta galactosidase staining of WT and Dlx4 KO embryos. (D) DLX4 (red) immunostaining performed on P1 dorsal skin samples. Keratin 5 (K5, green) and Hoechst counterstaining (blue) was done to label the basal layer and cell nuclei, respectively. (E) DLX4 immunoblotting shows detection in epidermal basal and suprabasal compartments of skin. K5 detection confirms efficient separation of basal and suprabasal cells. Vinculin was used as a loading control. (F) mRNA expression level in the skin confirms the deletion of Dlx4. (G) Immunohistochemical analysis using anti-DLX4 antibody on WT and Dlx4 KO adult dorsal skin. (H) Immunoblotting of protein extracts from WT and Dlx4 KO skin using an anti-DLX4 antibody. (I) Dye exclusion assay (I) and performed in WT and Dlx4 KO embryos at embryonic days 16.5 (E.16.5), 17.5 (E17.5) and 18.5 (E18.5). (J) Cornified envelope preparation from WT and Dlx4 KO neonatal skins. (K) Immunofluorescence staining on P1 dorsal skin samples using antibodies for basal epidermal (K14) and differentiation markers (K10, Filaggrin). (L) Western blot analysis of epidermal marker expression of Dlx4 KO and heterozygous skin compared to skin from WT littermates. RPL11 was used as a loading control.

Dlx4 deletion does not alter epidermal differentiation or hair follicle development

While DLX3 expression is predominant in the suprabasal layers of the epidermis and in the hair matrix [8–11], DLX4 expression was observed in all layers of the epidermis as well as in hair matrix and dermal papilla (Fig. 1D). Western blot analysis also confirms the presence of DLX4 in both basal and suprabasal epidermal layers (Fig. 1E). Deletion of Dlx4 was confirmed both at the mRNA (Fig. 1F) and protein levels (Fig. 1G and H). Dye exclusion assays performed in WT and Dlx4 KO embryos at embryonic day 16.5, 17.5 and 18.5 (E16.5, E17.5 and 18.5) and cell envelope preparations from E18.5 skin showed no detectable defect in skin barrier formation in Dlx4 KO (Fig. 1I and J). Moreover, there was no noticeable expression changes in epidermal differentiation markers in Dlx4 KO skin (Fig. 1K and L). Dlx4 deletion did not affect the expression of its cis-linked paralog gene Dlx3 (Fig. S2A, S2B and 1L). Hair development and cycling throughout the first and second phase of anagen (P10 and P30) and telogen (P20 and P76) remained unchanged in Dlx4 KO mice when compared to wild type littermates (Fig. S3).

DLX4 is dispensible for keratinocyte proliferation and migration

DLX4 is overexpressed in a wide variety of cancer types such as leukemias and in lung, breast, ovarian and prostate cancers [15–19]. To determine if there is a correlation between ablation of DLX4 function and the proliferative capacity of keratinocytes, we evaluated proliferation and migration. Absence of DLX4 did not affect the proliferation of cultured keratinocytes (Fig. 2A) and no significant difference in Pcna expression levels though RNA sequencing (RNA-seq) was observed at P1 stage of WT and Dlx4 KO (Supplementary Figure 4A). However, a detectable increase in PCNA positive cells was seen in DLX4 KO epidermis and hair follicles (Supplementary Figure 4B). Migration of keratinocytes cultured in low and high calcium medium was not affected upon DLX4 deletion in presence or absence Mitomycin C treatment (Fig. 2B).

Figure 2. Keratinocyte proliferation, migration and skin homeostasis was unaffected in Dlx4 KO.

(A) Proliferation assay was performed on cultured keratinocyte using IncuCyte Zoom System. (B) Keratinocyte migration in low and high calcium medium (with and witout Mitomycin treatment) was assessed by scratch assay and relative wound density measurement using the IncuCyte Zoom System. (C) Visualization of RNA-seq coverage data confirming the deletion of Dlx4 exon 2 and exon 3 in KO skin. (D) Heatmap of mRNA expression levels in terms of log of RPKM from dorsal skin of P1 WT and Dlx4 KO mice (n=3). (E) Hematoxylin and eosin staining showing the epidermis of vehicle (acetone) and TPA-treated back skin of WT and Dlx4 KO mice 48 and 72 hrs after treatment. Insets on right panel shows a higher magnification view of epidermis and underlying dermis of TPA treated Dlx4 WT and KO. (F) Bar graphs represents the quantitative assessment of epidermal thickness.

Dlx4 deletion does not significantly alter skin transcriptome or response to TPA treatment

Based on the absence of any overt phenotypic changes in skin of Dlx4 KO mice, we performed RNA-seq analysis to identify at the molecular level, genes and pathways globally affected by Dlx4 deletion. RNA-seq reads mapped against the mouse genome confirmed total absence of second and third exons of Dlx4 in KO skin (Fig. 2C). No significant changes were detected at the transcriptome level in Dlx4 KO skin (no q-values below 0.05) (Fig. 2D).

We previously reported that epidermal Dlx3 deletion results in hyperproliferation which is exacerbated with TPA treatment [10]. We tested the effect of TPA treatment on DLX4 KO skin detected no difference between WT and Dlx4 KO as assessed by morphology and quantification of epidermal thickness (Fig. 2E and 2F).

CONCLUSIONS

In summary, our results indicate that while DLX4 expression is observed in both neonatal and adult murine skin, DLX4 function is dispensable for keratinocyte proliferation, differentiation and migration. Dlx4 KO mice were phenotypically indistinguishable from wild-type counterparts and skin transcriptome analysis did not determine any significant differentially expressed genes linked to the Dlx4 deletion. Indeed, there are mouse lines with knockouts for genes such as Calm4, periplakin, desmoyokin and TMEM45A which are highly expressed in skin and yet present no overt skin phenotype [20–23]. Lack of major phenotypic changes in Dlx4 KO mice suggests a non-essential function in skin development and homeostasis, although its role in more challenging conditions such as exposure to UV-B, injury or microbial infection remains to be determined.

However, DLX4 overexpression has been associated with aggressiveness of tumors. Therefore, overexpression of DLX4 in mouse will be important to study in future investigations [15, 16, 24, 25]. Recently, a mutation in DLX4 has been also shown to be associated with human orofacial clefting, with an identified single nucleotide deletion c.546_546delG in the DLX4 gene [26]. Mutations in the cis-linked gene DLX3, are associated with the tricho-dento-osseous (TDO) ectodermal dysplasia. [27]. The mutant TDO has a dominant negative function over the WT DLX3 [28]. Functionally, genetic mutations that lead to frame-shifts or altered aminoacid composition present a deleterious effect. Introducing the specific mutation associated with orofacial clefting in the DLX4 gene of mouse may give us a better understanding of DLX4 broader function during development.