Defective glucocorticoid receptor signaling and keratinocyte-autonomous defects contribute to skin phenotype of mouse embryos lacking the Hsp90 co-chaperone p23

Abstract

p23 is a small acidic protein with intrinsic molecular chaperone activity. It is best known as a co-chaperone of the major cytosolic molecular chaperone Hsp90. p23 binds the N-terminus of Hsp90 and stabilizes the ATP-bound and N-terminally closed Hsp90 dimer. It is in this configuration that many Hsp90 clients are most stably bound. Considering the important role of p23 in the Hsp90 cycle, it came as a surprise that it is not absolutely essential for viability in the budding yeast or for mouse development. Mice without p23 develop quite normally until birth and then all die perinatally because of immature lungs. The only other apparent phenotype of late stage embryos and newborns is a skin defect, which we have further characterized here. We found that skin differentiation is impaired, and that both apoptosis and cell proliferation are augmented in the absence of p23; the consequences are a severe thinning of the stratum corneum and reduced numbers of hair follicles. The altered differentiation, spontaneous apoptosis and proliferation are all mimicked by isolated primary keratinocytes indicating that they do require p23 functions in a cell-autonomous fashion. Since the phenotype of p23-null embryos is strikingly similar to that of embryos lacking the glucocorticoid receptor, a paradigmatic Hsp90-p23 client protein, we investigated glucocorticoid signaling. We discovered that it is impaired in vivo and for some aspects in isolated keratinocytes. Our results suggest that part of the phenotype of p23-null embryos can be explained by an impact on this particular Hsp90 client, but do not exclude that p23 by itself or in association with Hsp90 affects skin development and homeostasis through yet other pathways.

Fig 1. Lack of p23 results in decreased epidermal thickness, delayed differentiation and increased rates of proliferation and apoptosis in E18.5 skin.

(a) Representative sections stained with H&E showing epidermal skin (delimited by red line) thinning with almost absent stratum corneum in embryonic p23^-/- skin sections along with lower numbers of hair follicles. E, epidermis; HF, hair follicle; SC, stratum corneum. (b) Quantitation of epidermal thickness and (c) number of hair follicle per 100 μm basal membrane (BM) in p23^-/- and wild-type E18.5 skin sections; significant decrease of epidermal thickness (*** p<0.001) and hair follicle numbers (** p<0.01) in p23^-/- skin (n: WT = 4, p23^-/- = 3). (d) Immunohistochemical staining for the differentiation markers loricrin, filaggrin and involucrin shows decrease in staining intensity in p23^-/- E18.5 skin sections. (e) TUNEL assay indicates significantly higher apoptotic rate in p23^-/- E18.5 skin sections compared to wild-type (*** p<0.001; n: WT = 4, p23^-/- = 4). (f) BrdU staining of nuclei indicates increased proliferation in p23^-/- skin sections (** p<0.01; n: WT = 7, p23^-/- = 7). Black bars in panels a and d indicate 50 μm.

Fig 2. Expression of p23, GR, Fkbp51 and other GR target genes in E18.5 skin.

(a) Immunohistochemical staining for p23, GR, and Fkbp51 in E18.5 skin sections. Note that GR staining was detected almost exclusively in basal keratinocytes, with very few positive cells in the suprabasal layers, and that GR is also present in the dermal compartment. Fkb51 levels were markedly lower in the epidermis of p23^-/- E18.5 embryos, while expression of GR was slightly lower in some but not all p23^-/- skin sections compared to wild-type. For each protein and genotype, sections of two different embryos are shown. White bars indicate 50 μm. (b) Representative immunoblot (top panel) of extracts from 4 WT and 3 p23^-/- embryos shows reduced GR protein levels in E18.5 epidermis in some but not all p23^-/- E18.5 epidermal samples compared to wild-type. The quantitation of two immunoblots with a total of 6 WT and 7 p23^-/- samples is shown as a bar graph. A trend to a reduction of GR levels is suggested by the statistical analysis (p = 0.0566), whereas Fkbp51 levels are undoubtedly decreased in p23^-/- E18.5 epidermis (*** p<0.001). Immunoblotting for p23 was used to confirm the genotype and GAPDH served as loading control. (c) Relative mRNA levels of several other GR target genes in p23^-/- vs. wild-type E18.5 epidermis, quantified by qPCR. Results are representative of three independent experiments, each with the 6 WT and 7 p23^-/- samples.

Fig 3. Isolation of primary MPKs.

(a) Total number of MPKs that can be obtained from the epidermis of individual p23^-/- E18.5 embryos compared to wild-type is significantly lower (* p<0.05; n: WT = 24, p23^-/- = 26). (b) Representative micrographs showing differences in cell confluency after seeding MPKs; bars = 50 μm. (c) Increased growth rate of p23^-/- MPKs; * p value at 120 h <0.05 (n: WT = 3, p23^-/- = 3). (d, e) Increased apoptosis of p23^-/- compared to wild-type MPKs without (d; **** p<0.0001; n: WT = 10, p23^-/- = 6) and following UV treatment (e; **** p<0.0001; n: WT = 6, p23^-/- = 11).

Fig 4. Differentiation of p23^-/- MPKs in vitro is impaired.

(a) Undifferentiated wild-type and p23^-/- MPKs from E18.5 embryos were induced to differentiate with 1.2 mM Ca²⁺ for 24 h and 48 h. No obvious morphological differences were noted by phase contrast microscopy (bars = 50 μm). (b) Immunofluorescence staining for epithelial markers demonstrates impaired differentiation of p23^-/- MPKs. Note that only a subset of cells are expected to express K-10 at the relatively early 24 h time point. K-10, keratin-10.

Fig 5. Dex differentially affects p23^-/- versus wild-type MPKs.

(a) Nuclear localization of GR is impaired in p23^-/- MPKs. Immunofluorescence micrographs of cells treated with 100 nM Dex for 3 h and stained with an antibody to GR (bars = 50 μm). (b) Opposite effects of 1 μM Dex on the proliferation of wild-type (* p<0.05) and p23^-/- (* p<0.05) MPKs 120 h after seeding (n: WT = 3, p23^-/- = 3).

Fig 6. GR-mediated effects on MPK differentiation are not affected by loss of p23.

(a) Phase contrast micrographs of MPKs induced to differentiate with 1 μM Dex for 48 h. (b) Immunofluorescent staining for the differentiation markers keratin-10 (K-10), E-cadherin and involucrin in wild-type and p23^-/- MPKs treated with 1 μM Dex for 48 h (bars = 50 μm).