Dysregulation of Grainyhead-like 3 expression causes widespread developmental defects

Abstract

Background: The gene encoding the transcription factor, Grainyhead-like 3 (Grhl3), plays critical roles in mammalian development and homeostasis. Grhl3-null embryos exhibit thoraco-lumbo-sacral spina bifida and soft-tissue syndactyly. Additional studies reveal that these embryos also exhibit an epidermal proliferation/differentiation imbalance. This manifests as skin barrier defects resulting in peri-natal lethality and defective wound repair. Despite these extensive analyses of Grhl3 loss-of-function models, the consequences of gain-of-function of this gene have been difficult to achieve.

Results: In this study, we generated a novel mouse model that expresses Grhl3 from a transgene integrated in the Rosa26 locus on an endogenous Grhl3-null background. Expression of the transgene rescues both the neurulation and skin barrier defects of the knockout mice, allowing survival into adulthood. Despite this, the mice are not normal, exhibiting a range of phenotypes attributable to dysregulated Grhl3 expression. In mice homozygous for the transgene, we observe a severe Shaker-Waltzer phenotype associated with hearing impairment. Micro-CT scanning of the inner ear revealed profound structural alterations underlying these phenotypes. In addition, these mice exhibit other developmental anomalies including hair loss, digit defects, and epidermal dysmorphogenesis.

Conclusion: Taken together, these findings indicate that diverse developmental processes display low tolerance to dysregulation of Grhl3.

Keywords: Grhl3 overexpression; Shaker-Waltzer phenotype; inner ear malformation; neural tube defects; skin barrier defects.

FIGURE 1

Generation of Grhl3 ^Cre ;Rosa26 ^{Grhl3 cDNA} transgenic mouse model. (A) Gene‐targeting strategy for making Rosa26 ^{Grhl3 cDNA} allele. The loxP‐STOP‐loxP‐Flag‐Grhl3 transgene was targeted into the RMCE‐compatible Rosa26 locus of murine G4 ROSALUC embryonic stem cells. (B) With the presence of the Cre recombinase, the expression of the transgene will be constantly activated, which is driven by the Rosa26 promoter, after the Cre‐mediated excision of the loxP‐flanked transcriptional stop sequence (loxP‐STOP‐loxP). (C) G1 Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/+} mice is generated by crossing G0 breeding pairs, Grhl3 ^Cre/+ and Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} mice. Intercrossing of G1 Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/+} mice generated offspring with nine different genotypes. (D) PCR genotyping of Grhl3 ^Cre allele. WT, product of wild‐type Grhl3 allele; Cre, product of targeted Grhl3 allele. (E) PCR genotyping of Rosa26 ^{Grhl3 cDNA} allele. WT, product of wild‐type Rosa26 allele; Tg, product of targeted Rosa26 allele. (F) Representative PCR products of floxed allele and delta allele of Grhl3 transgene on Rosa26 locus

FIGURE 2

Misexpression of Grhl3 in Grhl3 ^Cre ;Rosa26 ^{Grhl3 cDNA} transgenic mouse model. Q‐RT‐PCR on E18.5 epidermis showing abundance of total Grhl3 mRNA (A), endogenous Grhl3 mRNA (B) and transgene Grhl3 mRNA (C) varies across genotypes. (D) Western blot analysis of GRHL3 and GFP protein expression in E18.5 wild‐type, constitutive Grhl3‐null and transgenic mice. (E) Q‐RT‐PCR on E18.5 epidermis showing comparable expression levels of Tgm1 in wild‐type and transgenic mice. Bar graph presented as a mean ± standard error of mean (SEM). A one‐way ANOVA test following by a Dunnett's multiple comparison test between wild‐type and other genotypes were used for data analysis. An additional Mann‐Whitney test was used to compare the total Grhl3 mRNA level (A) between E18.5 wild‐type and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} epidermis. *P value <.05, **P‐value <.01, ***P value <.001, ****P value <.00001. n.s., not significant

FIGURE 3

Restoration of Grhl3 expression rescues neurulation and skin barrier defects of Grhl3‐null mice while misexpression of Grhl3 expression causes digit defects. (A‐D) Skin barrier assay on E18.5 wild‐type and transgenic mice. Wild‐type embryos (N = 4) showed a fully acquired skin barrier (A). Penetration of toluidine blue into skin indicated impaired skin barrier in Grhl3 ^Cre/Cre ;Rosa26 ^+/+ embryos (N = 4). Grhl3 ^Cre/Cre ;Rosa26 ^+/+ embryos also showed syndactyly, curly tail and spina bifida (B). Restoration of Grhl3 expression rescued skin barrier defect and spina bifida in Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/+} embryos (N = 4) (C). Over‐expression of Grhl3 in Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} embryos (N = 4) fully rescued Ghrl3‐null phenotypes (D). (E) Gross appearance of digits from wild‐type (N = 4), Grhl3 ^Cre/Cre ;Rosa26 ^+/+ (N = 4) and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 4) embryos at E18.5. (F) Skeletal preparations of wild‐type (N = 4), Grhl3 ^Cre/Cre ;Rosa26 ^+/+ (N = 4) and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 2) embryos at E18.5. Grhl3 ^Cre/Cre ;Rosa26 ^+/+ embryos exhibited soft‐tissue syndactyly. Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} embryos exhibited synpolydactyly and incompletely formed digits. The dotted line outlines the soft‐tissue syndactyly. (G) Penetrance of digit defects among Grhl3 ^Cre/Cre ;Rosa26 ^+/+, Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/+} and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} embryos at E18.5. Arrows, curly tail; black arrowheads, spina bifida; red arrowheads, digits

FIGURE 4

Restoration of Grhl3 expression partially rescues epidermis abnormalities in embryos. Histological and immunostaining analysis of wild‐type and transgenic E18.5 epidermis. N = 4. Arrowheads, the extra tissue structure residing superiorly to the epidermis; asterisks, compacted stratum corneum layer between the extra tissue structure and the stratum granulosum layer; arrows, compacted stratum corneum layer. SC, stratum corneum; SG, stratum granulosum; SP, stratum spinosum; SB, stratum basale; E, epidermis; D, dermis

FIGURE 5

Misexpression of Grhl3 expression leads to basal keratinocyte hyperproliferation. PCNA and Ki67 immunostaining analyses and quantifications of PCNA‐ and Ki67‐positive cells in wild‐type and transgenic E18.5 epidermis. N = 4. Arrowheads, the extra tissue structure residing superiorly to the epidermis; asterisks, compacted stratum corneum layer between the extra tissue structure and the stratum granulosum layer; arrows, compacted stratum corneum layer. SC, stratum corneum; SG, stratum granulosum; SP, stratum spinosum; SB, stratum basale; E, epidermis; D, dermis. Bar graph presented as a mean ± standard error of mean (SEM). A one‐way ANOVA test following by a Dunnett's multiple comparison test between wild‐type and other genotypes were used for data analysis. ***P value <.001, ****P value <.00001. n.s., not significant

FIGURE 6

Overexpression of Grhl3 perturbs epidermal homeostasis in adult. (A) Adult Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} mice displayed large scale of alopecia in the dorsal region (N = 4). (B) Hematoxylin and eosin staining of adult wild‐type, Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} hair follicles (N = 4). (C) Histological and immunostaining analysis of wild‐type, Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} adult epidermis (N = 4). (D) Quantifications of PCNA‐ and Ki67‐positive cells in wild‐type, Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} adult epidermis. Arrowheads, clefts between the inner root sheath and the outer root sheath. SC, stratum corneum; SG, stratum granulosum; SP, stratum spinosum; SB, stratum basale; E, epidermis; D, dermis. Bar graph presented as a mean ± standard error of mean (SEM). A one‐way ANOVA test following by a Dunnett's multiple comparison test between wild‐type and other genotypes were used for data analysis. *P value <.05, ***P value <.001, ****P value <.00001

FIGURE 7

Misexpression of Grhl3 does not compromise the expression of hair shaft anchorage genes Dsg1a and Grhl1. Q‐RT‐PCR on E18.5 epidermis showing the mRNA abundance of two hair shaft anchorage genes Dsg1a (A) and Grhl1 (B) in wild‐type and transgenic mice. Bar graph presented as a mean ± standard error of mean (SEM). A one‐way ANOVA test following by a Dunnett's multiple comparison test between wild‐type and other genotypes were used for data analysis. n.s., not significant

FIGURE 8

Inner ear malformation and hearing impairment in Grhl3 transgenic mice. (A) Quantification of the number of times that transgenic mice circled per minute (N = 3‐4 mice per genotype). (B) 3D reconstructions of the inner ear X‐ray micro‐computed tomography scanning. (C) Average ABR thresholds of wild‐type (N = 10), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 7) and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 3) mice when respond to a click containing mixed frequency noise. (D) Average ABR thresholds of wild‐type (N = 9), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 9) and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 3) mice when respond to pure tone stimuli at 4, 8, 16, and 32 kHz. Black arrowheads, lateral semi‐circular canal; green arrowheads, posterior semi‐circular canal; red arrowheads, anterior semi‐circular canal; purple arrowheads, ampullae; asterisks, artifacts of the scanning and reconstruction. V, vestibular apparatus; C, cochlea. Bar graph and scatter plots presented as a mean ± standard error of mean (SEM). A one‐way ANOVA test following by a Dunnett's multiple comparison test between wild‐type and other genotypes were used for data analysis. *P value <.05, **P value <.01, ****P value <.00001. n.s., not significant

FIGURE 9

Misexpression of Grhl3 does not impair the orientation of cochlea hair cell stereocilia and loss of Grhl3 expression does not cause inner ear malformation. (A) Phalloidin staining of cochlea hair cells from the inner ear of postnatal wild‐type and transgenic mice. N = 3. Scale bar = 10 μm. (B) Skeletal preparations of heads from E18.5 wild‐type and Grhl3 ^Cre/Cre ;Rosa26 ^+/+ embryos. N = 4. C, cochlea. Black arrowheads, lateral semi‐circular canal; green arrowheads, posterior semi‐circular canal; red arrowheads, anterior semi‐circular canal

FIGURE 10

Misexpression of Grhl3‐induced hearing impairment is not a neurological condition. Representative click‐evoked ABR waveforms from wild‐type (N = 10), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/+} (N = 7), Grhl3 ^Cre/+;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 7) and Grhl3 ^Cre/Cre ;Rosa26 ^{Grhl3 cDNA/Grhl3 cDNA} (N = 3) adult mice. I–VII represent the presences of ABR peaks, with each peak corresponds to one action potential as signal transmits through the vestibulocochlear nerve and the brainstem