HNRNPU is essential for proper murine skin development

TO THE EDITOR

The Heterogeneous Nuclear Ribonucleoproteins (HNRNPs) are a large family of RNA-binding proteins that play crucial roles in various aspects of nucleic acid metabolism, including alternative splicing, mRNA stabilization, and transcriptional and translational regulation(Geuens et al., 2016). Our previous work has shown that HNRNP family members such as HNRNPK is necessary to prevent premature differentiation and sustain the proliferative capacity of human epidermal stem and progenitor cells(Li et al., 2019). A similar family member, HNRNPL, is essential for human basal layer epidermal stem and progenitor cells to express integrin and extracellular matrix genes to allow attachment to the underlying dermis(Li et al., 2021). While HNRNP family members have been shown to have profound impacts on cultured primary human keratinocytes and regenerated human skin as described above, it is unclear whether any of them are important for the development of the skin. Mutations in HNRNPU, the largest member of the HNRNP family, leads to neurodevelopmental disorders in humans(Poot, 2019). Brain specific deletion of Hnrnpu, showed that it is crucial for murine cerebral cortex development(Sapir et al., 2022). Since total body knockout (KO) of Hnrnpu is embryonic lethal, it is unclear if it has any role in the development of the skin(Roshon and Ruley, 2005).

Hnrnpu is one of the most highly expressed Hnrnp family of genes in the epidermis from P0 mice after Hnrnpk and Hnrnpl which are two family members we have already found to have a crucial role in the skin (Table S1). To investigate its role in the skin, we generated mice lacking Hnrnpu in the epidermis by crossing Hnrnpu^fl/fl mice with keratin 14 (Krt14)-Cre mice. Heterozygote Hnrnpu^wt/fl Krt14-Cre mice were used as controls. The vast majority of the Hnrnpu KO (Hnrnpu^fl/fl Krt14-Cre) mice died immediately after birth (only 2 out of 25 survived more than 5 days) and the birth rate was also significantly reduced compared to the expected level, possibly due to fetal resorption prior to birth (Figure S1a).

When comparing the skin conditions of embryos at days 16 and 18 (E16, E18), the skin of KO embryos on day E16 appeared more wet, transparent, and shiny compared to the aged matched control suggesting delayed barrier acquisition in the KO mice (Figure 1a). By E18, KO mice exhibited more visible subcutaneous blood vessels (white arrows), suggesting thinner skin. On the day of birth (Postnatal day 0: P0), the Hnrnpu KO skin appeared generally taut and erythematous, with visible subcutaneous blood vessels (Figure 1a).

Figure 1. Hnrnpu is required for the proper development of murine skin.

(a) Pictures of control (Hnrnpu^wt/fl Krt14-Cre) and Hnrnpu knockout (Hnrnpu^fl/fl Krt14-Cre) embryos at embryonic days (E16, E18), and post-natal day 0 (P0). White arrow denotes subcutaneous vessels. (b) Skin permeability assay using toluidine blue dye from E16, E18, and P0 pups. (c) Assessment of inside-to-outside barrier function with trans-epidermal water loss (TEWL) assay on P0 pups. (d) RT-QPCR of Hnrnpu mRNA expression from epidermis harvested from P0 pups. N is a minimum of at least 3 for the figure. All individual dots in bar graphs represent data from an individual mouse. Mean values are shown with error bars=SD (unpaired t-test).

The toluidine blue dye penetration was increased in the E16 Hnrnpu KO mice which suggests delayed development of the outside-to-inside barrier function (Figure 1b). By E18 and around the time of birth (P0), there was not as obvious of a difference between control and KO mice in their dye exclusion (Figure 1b). However, there were still large barrier defective lesions (white arrows) found throughout the animal possibly due to skin fragility (Figure 1b). Trans-epidermal water loss (TEWL) measurements of the KO skin (flank area) immediately after birth (P0) showed significantly higher levels of water loss, indicating a persistent abnormality in the inside-to-outside barrier function (Figure 1c). The barrier defective lesions and elevated water loss may explain the cause of the neonatal lethality in the knockout pups (Figure 1c, Figure S1a). Knockout of Hnrnpu was validated by the more than 90% reduction of Hnrnpu mRNA levels from the epidermis of P0 KO mice as well as the absence of Hnrnpu protein expression in the epidermis of P0 skin (Figure 1d and Figure S1b).

To determine the basis for the abnormal inside to outside barrier function, histology was performed on control and Hnrnpu KO dorsal skin at P0. The KO epidermis was significantly thinner and lacked epidermal appendages including hair follicles as compared to controls (Figure 2a, 2j). Under high magnification, the knockout epidermis was observed to not properly stratify, consisting of only 2 to 3 cell layers (Figure 2a: bottom panels). There was also epidermal fragility (red asterisks) with a lack of clear distinction between the epidermis and dermis (Figure 2a). The KO skin also lacked hair follicles with only rudimentary hair germ-like cell aggregations (red arrowheads) that appeared along the basal layer (Figure 2a). This suggests that there is delayed or failed development of the epidermis and its appendages. Supporting this, the mRNA expression of developmental regulators such as Zfp750, Grhl3, Trp63 and differentiation genes Krt1, Lor, and Flg were decreased in Hnrnpu KO epidermis (Figure 2b, 2d). P63 was expressed in the basal layer and the hair follicle of control mice but greatly diminished in the KO mice (Figure 2c, 2d). This diminished level of p63 was also correlated with decreased expression of terminal differentiation protein (Lor), diminished stratification, and decreased basal layer proliferation (Figure 2c-2f). The Hnrnpu KO mice also had diminished Krt15 expression in the hair follicle suggesting failure to properly develop hair follicle stem cells (Figure 2g, 2h). Notably, the epidermal separation observed in the KO mice occurred in the basal layer as evidenced by the split (white arrows) in the Krt14 positive cells which translated to skin that was highly susceptible to damage which tore upon handling (Figure 2i, Figure S1c). To determine the basis for the fragility in the epidermis, we analyzed the mRNA expression of basement membrane, extracellular matrix (ECM), and desmosomal genes. Most of the basement membrane/ECM genes were not altered between control and Hnrnpu KO epidermis (Figure S1d). However, the majority of the desmosomal genes which are critical for the intercellular connection between epidermal keratinocytes were downregulated in the KO epidermis (Figure S1e). Staining for desmoglein 2 (Dsg2) also showed diminished levels of this protein in the Hnrnpu KO epidermis (Figure S1f). In addition, Krt5 which when mutated in humans can cause epidermolysis bullosa simplex was downregulated on the protein and mRNA level in Hnrnpu deleted epidermis (Figure 2b, Figure S1g).

Figure 2. Hnrnpu knockout leads to epidermal thinning, skin fragility, dysregulation of differentiation, and loss of skin appendages.

(a) Hematoxylin and eosin staining of skin from control (Hnrnpu^wt/fl Krt14-Cre) and Hnrnpu KO (Hnrnpu^fl/fl Krt14-Cre) post-natal day 0 (P0) pups. Red asterisk denotes split in the epidermis. Red arrowheads denote rudimentary hair germ. Scale bar= 150 μm (top panels) and Scale bar= 50 μm (bottom panels). (b) RT-QPCR for the mRNA expression of epidermal differentiation regulators (Zfp750 and Grhl3), markers of the basal layer keratin 5 (Krt5), and markers of the differentiated layers of the epidermis Keratin 1 (Krt1), loricrin (Lor), filaggrin (Flg). RNA was isolated from the epidermis of the dorsal skin of P0 pups. (c) Immunofluorescence staining of control and Hnrnpu KO dorsal skin (P0) with antibodies against loricrin (differentiation protein: green), ΔNp63 (developmental transcription factor: red), and Hoechst 33342 staining of nuclei (blue). Scale bar= 150 μm. (d) mRNA expression of ΔNp63 from epidermis harvested from P0 pups. (e) Immunofluorescence staining of control and Hnrnpu KO dorsal skin (P0) with antibodies against Ki67 (proliferation marker: green), Krt14 (basal layer marker: red), and Hoechst 33342 staining of nuclei (blue). Scale bar=50 μm. (f) Quantitation of percent of epidermal cells that are proliferative (Ki67 positive). (g) Immunofluorescence staining of dorsal skin (P0) using antibodies against Krt15 (hair follicle stem cell marker: red) and Hoechst 33342 staining of nuclei (blue). Scale bar=50 μm. (h) mRNA expression of keratin 15 (Krt15) from the epidermis of P0 pups. (i) Immunofluorescence staining of dorsal skin (P0) using antibodies against Krt10 (spinous layer marker: green), Krt14 (basal layer marker: red), and Hoechst 33342 staining of nuclei (blue). White arrows denote areas of Krt14 positive epidermal cells that have split apart. Scale bar= 150 μm. (j) The thickness of the epidermis as measured by ImageJ in P0 pups. N is a minimum of at least 3 for the figure. All individual dots in bar graphs represent data from an individual mouse. Mean values are shown with error bars=SD (unpaired t-test). *p<0.05, **p<0.01.

To determine the global changes in gene expression upon Hnrnpu deletion, RNA sequencing (RNA-Seq) was performed on epidermis harvested from day 0 control and KO mice. Among the 1,804 upregulated genes, the gene ontology (GO) terms keratinocyte differentiation and epidermal cell differentiation were enriched in the Hnrnpu-deleted epidermis (Figure S2a, S2b, Table S2). While it may seem counterintuitive that the upregulated GO terms included epidermal and keratinocyte differentiation, a closer examination of the genes showed Sprrs (Sprr2f, Sprr3, Sprr2g, Sprr2b, Sprr1a, Sprr1b) and Krt16 being the genes upregulated (Figure S2b, S2d, Table S2). While these genes are classified with “epidermal/keratinocyte differentiation” GO terms, they have been shown to be associated with wound healing and disrupted barrier function (Segre, 2006) (Lessard et al., 2013). Specifically, the Sprrs have been shown to be upregulated in barrier deficient mice and human psoriatic plaques while Krt16 regulates innate immunity due to barrier breach (Segre, 2006) (Lessard et al., 2013). The impaired development of the skin in Hnrnpu knockout mice is likely due to the downregulation of developmental regulators and differentiation genes such as Grhl3, Zfp750, p63, Lor, and Flg, which impairs proper stratification and normal differentiation (Figure 2b-2d)(Sen et al., 2012). In particular, the skin of p63 knockout mice is similar to the skin of the Hnrnpu KO mice which includes loss of epidermal stratification, barrier disruption, as well as lack of epidermal appendage development (Supplementary reference 1). There is also decreased desmosomal gene expression as well as Krt5 which leads to epidermal fragility and further contributes to the wounding environment (Figure 2a-b, 2i, S1c-S1f).

The 1,592 downregulated genes were enriched for GO terms such as carboxylic acid transport, nuclear chromosome segregation, Wnt signaling pathway, and mitotic sister chromatid segregation (Figure S2a, S2c, Table S2). The downregulation of chromosome and chromatid segregation GO terms supports the decrease in proliferation seen in the epidermis of KO mice (Figure 2e-2f). Importantly, the Wnt signaling pathway (Tcf7l2, Tcf7l1, Wnt10b, Fzd2, Lef1, Wnt3) is downregulated which likely explains why the KO animals could not develop hair (Figure S2c, S2d, Table S2). These critical Wnt pathway genes have been previously described for the development, maintenance, and function of the hair follicle (Supplementary reference 2).

To summarize, HNRNPU is likely involved in coordinating the expression and processing of key genes that govern epidermal progenitor cell differentiation, proliferation, and epidermal structure. Hnrnpu may do this through transcriptional and post-transcriptional mechanisms (Geuens et al., 2016). On the genomic level, HNRNPU has been shown to directly regulate transcription as well as maintaining global chromatin structure (Fan et al., 2018). Post-transcriptionally, HNRNPU has been shown to regulate mRNA splicing, stability, and localization (Geuens et al., 2016). Thus, Hnrnpu regulation of key skin developmental genes likely influences the formation and maintenance of the epidermal barrier, as well as the development of skin appendages such as hair follicles in murine skin.

Supplementary Material

Supplementary Figures

Supplementary Figure 1. Hnrnpu knockout causes embryonic lethality, epidermal fragility, and downregulation of adhesion genes.

(a) The expected and actual birth rates of Hnrnpu KO (Hnrnpu^fl/fl Krt14-Cre) mice. Of the Hnrnpu KO mice that were born only two survived more than 5 days after birth. The other 23 Hnrnpu KO mice died less than a day after birth. (b) Immunofluorescence staining of control and Hnrnpu KO dorsal skin (P0) with antibodies against Hnrnpu (red) and Hoechst 33342 staining of nuclei (blue). Arrows and circles denote loss of Hnrnpu protein expression from the epidermis. Scale bar=50 μm. (c) Skin erosions (white arrows) in the Hnrnpu KO mice indicating their skin fragility. (d-e) RT-QPCR for the mRNA expression of basement membrane, extracellular matrix (ECM), and desmosomal structure genes. RNA was isolated from the epidermis of the dorsal skin of P0 pups. (f) Immunofluorescence staining of dorsal skin (P0) with antibodies against desmoglein 2 (Dsg2:red) and Hoechst 33342 staining of nuclei (blue). Scale bar=50 μm. (g) Immunofluorescence staining of dorsal skin (P0) with antibodies against Krt5 (basal layer marker: red) and Hoechst 33342 staining of nuclei (blue). Scale bar=50 μm. N is a minimum of at least 3 for the figure. All individual dots in bar graphs represent data from an individual mouse. Mean values are shown with error bars=SD (unpaired t-test). **p<0.01. NS=not significant.

Supplementary Figure 2. RNA sequencing analysis of Control and Hnrnpu KO epidermis (P0).

(a) Volcano plot of RNA-seq data showing the differentially expressed genes upon Hnrnpu knockout. 1,804 genes were upregulated while 1,592 genes were downregulated (≥2 fold change and FDR ≤ 0.05), n=3 in each group. RNA was harvested from the epidermis of control (Hnrnpu^wt/fl Krt14-Cre) and Hnrnpu KO (Hnrnpu^fl/fl Krt14-Cre) post-natal day 0 (P0) pups. (b) Top 5 gene ontology (GO) terms for the 1,804 genes upregulated in the Hnrnpu knockout epidermis. (c) Top 5 gene ontology (GO) terms for the 1,592 genes downregulated in the Hnrnpu knockout epidermis. (d) Heatmap for the genes found in the two GO terms that represent keratinocyte differentiation and the Wnt signaling pathway.

DATA AVAILABILITY

The RNA-Seq dataset generated from this study has been deposited in GEO with the accession number GSE273649 with the link:

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE273649