Report

A Homolog of Drosophila grainy head Is Essential for Epidermal Integrity in Mice

See allHide authors and affiliations

Science  15 Apr 2005:
Vol. 308, Issue 5720, pp. 411-413
DOI: 10.1126/science.1107511

Abstract

The Drosophila cuticle is essential for maintaining the surface barrier defenses of the fly. Integral to cuticle resilience is the transcription factor grainy head, which regulates production of the enzyme required for covalent cross-linking of the cuticular structural components. We report that formation and maintenance of the epidermal barrier in mice are dependent on a mammalian homolog of grainy head, Grainy head-like 3. Mice lacking this factor display defective skin barrier function and deficient wound repair, accompanied by reduced expression of transglutaminase 1, the key enzyme involved in cross-linking the structural components of the superficial epidermis. These findings suggest that the functional mechanisms involving protein cross-linking that maintain the epidermal barrier and induce tissue repair are conserved across 700 million years of evolution.

Despite substantial structural differences, the surface epithelium of flies and that of mice exhibit notable functional parallels. Genetic studies in both organisms have identified highly conserved pathways regulating cell movement and polarity (1, 2), wound healing (3), and innate immunity (4). Less is known about the molecular and functional conservation of genes involved in assembly and maintenance of the impermeable surface barrier in the two organisms. In mammals, epidermal barrier function is largely provided by the stratum corneum (SC), a tough, insoluble layer composed of structural proteins and lipids covalently cross-linked by the enzyme transglutaminase 1 (TGase1) (5). In the fly, the epidermal epithelium deposits the embryonic cuticle, a rigid, waterproof layer essential for maintaining the structural integrity and microbial resistance of the organism (6). A limited number of genes have been linked to the regulation of cuticle formation in the fly embryo (7). One of these, grainy head (grh) (8), maintains the integrity of the cuticle through the regulation of Dopa decarboxylase (Ddc), which generates quinones that cross-link the cuticular proteins and chitin (9-11), similar to the function of TGase1 in mammals. We, and others, have recently identified homologs of grh that are highly conserved from Caenorhabditis elegans to human (12-15). We have demonstrated that mice lacking one of these genes, Grainy head-like 3 (Grhl3) (12), display neural tube defects (16). During these studies, we noted that expression of Grhl3 during mouse embryogenesis was largely restricted to the surface ectoderm, being expressed throughout this layer from embryonic day 12.5 (E12.5) onward (fig. S1), prompting us to examine the role of this factor in the maintenance of epidermal integrity.

We initially examined whether skin barrier function was perturbed in Grhl3-null mice. Changes in skin dye penetration reflect differences in the rates of embryonic acquisition of barrier function (17). Grhl3-/- and control embryos were exposed to 0.1% toluidine blue at various time points, and the ability of the embryos to prevent penetration of the dye was assessed (Fig. 1A). In agreement with previous studies, dye exclusion was not observed in either wild-type or Grhl3-null E16.5 embryos. Subsequently, the permeability barrier was established as a progressing front in the wild-type embryos between E17.5 and E18.5 (Fig. 1A, top row). In contrast, Grhl3-/- embryos demonstrated an inability to exclude dye, as shown by uniform blue staining at E17.5 and E18.5 (Fig. 1A, bottom row), reflecting their defective epidermal barrier.

Fig. 1.

Impairment of the epidermal barrier in Grhl3-null mice. (A) Skin permeability assay with 0.1% toluidine blue on Grhl3+/- intercross litters taken at the ages indicated. Wild type (+/+, top row); mutant (-/-, bottom row). (B) Transepidermal water loss over time on multiple equivalent skin sections from three separate E17.5 embryonic litters was analyzed. Wild type (+/+, n = 4); heterozygous (+/-, n = 10); mutant, (-/-, n = 5). Values are shown as the mean ± SD for each genotype at each time point. The differences between wild type and mutant were significant (P < 0.01, Student's t test) at each time point.

The skin barrier is critical for the key physiological function of fluid retention by the organism. Epidermal water loss can be measured by differential weight loss between wild-type and null mice under constant environmental conditions. Because the presence of thoracolumbosacral spina bifida in Grhl3-/- embryos may have confounded this total body weight analysis, we performed a transepidermal water loss (TEWL) assay on sections of skin from wild-type (+/+) and null (-/-) E18.5 embryos (Fig. 1B) (17). The cumulative TEWL from wild-type sections was less than 0.1 mg/mm2 over a 5-hour period. In contrast, water loss from the Grhl3-/- skin was 6.5 times this amount, with more than 0.5 mg/mm2 lost over the corresponding period. No statistical difference in water loss was observed between the wild-type and Grhl3+/- samples.

Because wound repair is another key function of the surface epithelium, we also examined this process in the Grhl3-null embryos. We harvested wild-type and Grhl3-/- embryos with their yolk sac intact at E12.5 (to assess embryonic wound healing) and E16.5 (to assess adult wound healing) (18). After wounding, the embryos were cultured for up to 24 hours in roller bottles and then analyzed by scanning electron microscopy (SEM) (Fig. 2A). Under these ex vivo conditions, the embryos are viable and continue to develop for up to 24 hours (19). In both E12.5 and E16.5 wild-type embryos, partial wound closure was observed after 6 hours, and this was largely complete after 24 hours. In contrast, the Grhl3-/- embryos of both gestations displayed minimal signs of healing at 24 hours.

Fig. 2.

Failed wound healing in Grhl3-null mice. SEM of a hindlimb amputation wound in wild-type (+/+) and mutant (-/-) embryos. The images are representative of the results obtained from six embryos in each group. E12.5 embryos (top row) were photographed 18 hours after amputation, and E16.5 embryos (bottom row) 24 hours after amputation. e, epidermis; cw, closed wound; ow, open wound. Scale bar, 86 μm at E12.5 and 170 μm at E16.5.

The molecular events underlying the defects observed in the Grhl3-null epidermis presumably reflected altered expression of target genes of this transcription factor. To date, none of these targets had been identified, and the consensus DNA binding sequence of this factor remained unknown. To further define the molecular mechanisms involved in Grhl3-dependent barrier formation and wound healing, we initially compared the phenotype, histology (fig. S2), and ultrastructural features (fig. S3) of the Grhl3-null mice with the features reported for mice deficient in other factors known to be critical for barrier formation. These included the fatty acid transport protein Fatp4 (20); Pig-a, which is responsible for glycosylphosphatidylinositol (GPI)-anchor synthesis (21); the transcription factors Arnt (22) and Klf4 (23); and the cornified envelope (CE) protein cross-linking enzyme, TGase1 (24, 25). Although all of these deficient mouse lines displayed some features that were shared with the Grhl3-null mice, there was a marked similarity between mice lacking Grhl3 and the animals deficient in either Klf4 or TGase1. Both lines displayed marked thickening of the SC with loss of corneocyte layering (fig. S2), loss of intact CEs, and defects in the SC lipid structures, with incomplete and irregularly deformed lamellae (fig. S3). We therefore examined the expression of Klf4 and TGase1 in wild-type and Grhl3-null epidermis by Northern analysis (Fig. 3A). The expression of TGase1 in the Grhl3-/- epidermis was reduced to less than one-fifth that of the wild-type control, but the levels of Klf4 RNA were identical in both samples. No change in expression between wild-type and Grhl3-/- epidermis was observed for the previously referred to “barrier” genes (26).

Fig. 3.

TGase1 is a putative target gene of Grhl3. (A) Northern analysis of TGase1 and Klf4 mRNA expression in skin from wild-type (+/+) and Grhl3-null (-/-) embryos.(B) Identification of a consensus sequence for recognition by Grhl3. The sequences from 49 clones obtained by CASTing (fig. S4, A and B) were aligned from positions -12 to +12. The consensus sequence was determined by the percent frequency of each nucleotide at each position. (C) DNA binding of Grhl3. Extract from the A431 cell line was studied with the Grhl3 consensus probe defined in (B). A 100-fold molar excess of unlabeled Grhl3 consensus probe or mutant probe or the two potential Grhl3 binding sites in the TGase1 promoter were added in the indicated lanes. The migration of the specific Grhl3/DNA complex is indicated with an arrow. The image shown in this figure has been cropped, and the full image is shown in fig. S4C.

To determine whether TGase1 was a potential Grhl3 target gene, we performed cyclic amplification and selection of targets (CASTing) (27) to define the DNA consensus binding site of Grhl3, using cellular extracts from an epidermoid carcinoma cell line (A431) transfected with a hemagglutinin epitope-tagged Grhl3-containing expression vector (28). After six cycles, the population of selected binding sites was cloned and individual isolates sequenced and aligned (Fig. 3B and fig. S4, A and B). The defined Grhl3 DNA binding consensus sequence matched the consensus sequence for Drosophila grh DNA binding, which we had previously identified by alignment of multiple grh-responsive gene-regulatory regions (12).

Recent studies have defined a region of the TGase1 promoter that is critical for its tissue-specific expression in epidermis and other epithelium (29). Scrutiny of this region (between -1.6 and -1.1 kb 5′ of the CAP site) identified a potential Grhl3 binding site, altered by the presence of one additional nucleotide. Further analysis of the upstream regions of the TGase1 gene identified a second potential Grhl3 binding site, 6 kb from the transcriptional start site. We examined these sites in electrophoretic mobility shift assays with the cellular extract from the A431 cells. As shown in Fig. 3C (and fig. S4C), a single specific protein/DNA complex was observed with the Grhl3 consensus probe defined in the CASTing experiment (lane 1). This complex was specifically competed with an unlabeled Grhl3 consensus probe (lane 2), but was not competed with the addition of an unlabeled mutant probe in which the two invariant nucleotides defined in the CASTing selection were altered (lane 3). Addition of either unlabeled TGase1 probe resulted in a reduction in Grhl3 binding to its consensus sequence (lanes 4 and 5). Taken together, these results suggest TGase1 as a putative target gene of Grhl3. The link between TGase1 and Grhl3 provides a mechanism that unifies the skin phenotypes of the Grhl3-deficient mice. TGase1-deficient mice also die early in neonatal life with markedly impaired barrier function and ultrastructural defects in their SC lipid lamellae (24, 25). Skin from these animals also displays marked delays in wound repair (30). However, differences also exist between these two deficient strains, particularly in the expression patterns of the CE precursor proteins and in the response to skin grafting (24), suggesting that the phenotype of the Grhl3-null mice may be influenced by the altered expression of additional target genes.

The abnormal barrier in the Grhl3-null mice appears comparable to the phenotype seen in Drosophila grh mutants, which exhibited fragile cuticles (7). Similarly, C. elegans embryos in which expression of a grainy head-like gene had been reduced by RNA interference failed to hatch and displayed a cuticle defect consistent with a loss of rigidity (14). Grh was initially identified as a factor involved in the transcriptional regulation of Ddc (9), which functions to pigment and harden the insect cuticle through the generation of quinones that cross-link the cuticular proteins (10). The identification of TGase1, the predominant enzyme involved in the generation of high molecular weight polymers of cross-linked CE proteins (30), as a putative target of Grhl3 indicates that the grh family is involved in the regulation of protein cross-linking in barrier formation across 700 million years of evolution. The grh mutant cuticular defects are unlikely to be caused solely by grh regulation of the Ddc gene, because null Ddc mutants do not express the above cuticular phenotype, and the hypopigmentation of cuticular structures in the grh mutants is not as severe as that in the null Ddc mutants. Similarly, the Grhl3-null epidermal defects suggest the presence of additional target genes other than TGase1. It is conceivable that some of these genes may be conserved from fly to human, a hypothesis strengthened by the demonstration of identical DNA binding consensus sequences for grh and Grhl3.

The recent identification of wound response enhancers in the Drosophila Ddc and pale (ple) genes that mediate protective functions of the epidermal wound response and require binding sites for grh, and grh genetic function, provides an additional compelling link between grh family members, cross-linking enzymes, and the integrity of the surface epithelium (31). Ple encodes tyrosine hydroxylase, which is also involved in the generation of quinones for protein/chitin cross-linking (10, 11). It is of interest that the regulation of cross-linking enzyme genes by grh-like factors has been conserved, even though the cross-linking genes themselves have diversified from fly to human. We found no change in Ddc gene expression in the Grhl3-null mice, consistent with the fact that this gene is not linked to protein cross-linking in the mammalian epidermis.

These studies identify an essential function for the grh family in maintenance of the integument barrier in diverse species. However, the grh gene is also critical for other aspects of epidermal and epithelial development in the fly, including cell polarity (32) and tubular morphogenesis (33). The complexity of gene function in the fly may also be seen in the mouse Grhl factors.

Supporting Online Material

www.sciencemag.org/cgi/content/full/308/5720/411/DC1

Materials and Methods

Figs. S1 to S4

References

References and Notes

View Abstract

Navigate This Article