Keratin 76 is required for tight junction function and maintenance of the skin barrier

Abstract

Keratins are cytoskeletal intermediate filament proteins that are increasingly being recognised for their diverse cellular functions. Here we report the consequences of germ line inactivation of Keratin 76 (Krt76) in mice. Homozygous disruption of this epidermally expressed gene causes neonatal skin flaking, hyperpigmentation, inflammation, impaired wound healing, and death prior to 12 weeks of age. We show that this phenotype is associated with functionally defective tight junctions that are characterised by mislocalization of the integral protein CLDN1. We further demonstrate that KRT76 interacts with CLDN1 and propose that this interaction is necessary to correctly position CLDN1 in tight junctions. The mislocalization of CLDN1 has been associated in various dermopathies, including the inflammatory disease, psoriasis. These observations establish a previously unknown connection between the intermediate filament cytoskeleton network and tight junctions and showcase Krt76 null mice as a possible model to study aberrant tight junction driven skin diseases.

Figure 1. Krt76 gene trap disruption causes gross epidermal defects.

(A) Schematic showing Krt76 gene trap (knock-out first) targeting construct. (B) Whole mount LacZ staining of Krt76^tm1a/+ reporter mice, shows Krt76 expression in the dorsal and ventral snout and palate, eyelid, and vagina. (C) Mice homozygous for Krt76 gene trap disruption (Krt76^tm1a/tm1a) exhibit flaky skin following birth (see arrow-insert). Adult Krt76^tm1a/tm1a mice exhibit a scruffy coat and smaller body weight (n = 3 males, age 9 weeks, ***p<0.004) (D, E), as well as tail scaling (F). Krt76^tm1a/tm1a mice exhibit paw pad hyperpigmentation (G), concurring with regions of LacZ reporter expression (H). LacZ expression within paw pads is detected in exocrine glands (H′) and suprabasal epidermal layers (I). (J, J′) Haemotoxylin and Eosin (H&E) staining of paw pads from WT (J) and Krt76^tm1a/tm1a (J′) mice. Yellow arrowheads indicate abnormal dermal pigmentation. (K, L) Immunofluorescence analysis with indicated antibodies in wild type and Krt76^tm1a/tm1a mouse paw pad. Samples are counter stained with nuclear dye DAPI (4',6-diamidino-2-phenylindole). Coloured brackets indicate approximate distribution of FLG and KRT76 expression around the granular layer. (M) Western blot analysis of WT and Krt76^tm1a/tm1a dorsal skin and face skin extracts. (N) Immunofluorescence analysis with anti-KRT76 and anti-K14 antibodies in wild type mouse dorsal skin at E14.5, E18.5, P1 and adult time points and adult Krt76^tm1a/tm1a dorsal skin (N′). Asterisks indicate non-specific basal layer staining. (O) Krt76 mRNA qRT-PCR analysis of p3 dorsal skin relative to Gapdh. Scale bars represent 50 µm.

Figure 2. Krt76 is required for normal wound healing.

(A) Krt76^tm1a/tm1a gene trap mice show spontaneous wounds around the eyes and shoulders (sites of grooming). (B) H&E staining of dorsal skin from adult WT, early and late phenotype Krt76^tm1a/tm1a mice. Yellow arrowhead indicates abnormal pigmentation in the dermis and epidermis. (C) Immunofluorescence phospho-histone H3 (pHH3) analysis of wild type and Krt76^tm1a/tm1a mouse dorsal skin shows (D) increased proliferation/pHH3 positive cells (p = 0.005). E) Krt76^tm1a/tm1a mice die progressively from ∼2 weeks after birth with no animals surviving beyond 12 weeks of age. Treatment with Baytril reduces morbidity and mortality. (F) Immunofluorescence analysis with anti-KRT76 and anti-K14 antibodies in non-wounded and wounded wild type dorsal skin after 5 days. (G) LacZ staining of the wounded skin from Krt76^tm1a/+ reporter mice at 7 days post wounding. (H) Krt76, Krt6b and Krt16 mRNA qRT-PCR analysis of wounded skin relative to Gapdh, over 10 days. (I) Quantification of wound closure in wild type and Krt76^tm1a/tm1a mice over 10 days. *p<0.05, Error bars are S.E.M. Scale bars represent 50 µm.

Figure 3. Biochemical analysis of Krt76 disrupted skin.

(A-E, G–H) Immunofluorescence (or dye) analysis of wild type, early and late phenotype Krt76^tm1a/tm1a dorsal skin as indicated. (F) Isolated corneocytes from wild type and Krt76^tm1a/tm1a mouse dorsal skin exhibit a modest reduction in surface area (F′) (***p = 0.0004). I) Dye exclusion assay revealed no defects in outside to inside barrier function in Krt76^tm1a/tm1a mouse dorsal skin. Error bars  =  S.E.M. Scale bars represent 50 µm.

Figure 4. Histological and biochemical analysis of conditional Krt76 knockout skin.

(A) Exon structure and domain prediction of mouse Krt76 gene. Blue box represents insertion of β-galactosidase (β-gal/LacZ) cassette in the Krt76^tm1a reporter allele. Green circles and red triangles indicate frt and loxP sites. Validation of the mutant alleles was achieved using PCR amplification (see Protocol S1). (B) H&E staining of dorsal skin from 4OHT-treated control and Krt76^tm1d/tm1d mice. (C–J) Immunofluorescence analysis of 4OHT-treated control and Krt76^tm1d/tm1d mouse dorsal skin as indicated. Yellow arrowhead indicates absence of granular layer KRT76 staining. Asterisks indicate non-specific staining. Scale bars represent 50 µm.

Figure 5. Krt76 ^mutant mice show barrier function defects and KRT76 stabilises Claudin1 at tight junctions.

(A) Transepidermal water loss assay on P3 dorsal skin from wild type and Krt76^tm1a/tm1a mice. (B) P3 paw pad skin was dermally injected with a biotin tracer and diffusion through the epidermis assessed, with Filaggrin (FLG) and DAPI co-staining for tissue orientation. Yellow arrowhead shows diffusion in suprabasal keratinocytes into cornified layer. (C) Biotin tracer was assessed alongside TJ component, Claudin1 (CLDN1). Tracer exclusion indicated by flanking yellow arrowheads. (D) Immunofluorescence analysis of CLDN1 and Ecadherin (ECAD) distribution in wild type and Krt76^tm1a/tm1a mouse dorsal skin. (E) Image quantification at the cellular surface shows an inward shift and a decrease in intensity of CLDN1 not observed with ECAD. (F) Further quantification by image analysis of CLDN1 co-localisation at the cell surface with ECAD or DAPI in the nucleus. (G, H) Immunofluorescence analysis of CLDN1 localization in dorsal skin of wild-type and Krt76^tm1a/tm1a mice in early phenotype and biopsy wounded adult dorsal skin of wild-type and Krt76^tm1a/tm1a mice. (I) Dorsal skin fractionation assay showing localisation of different proteins to different fraction; relative lcoalisation of CLDN1 are indicated in (I′). (J, K) Immunofluorescence analysis of CLDN1 localization in adult dorsal skin and paw pads of 4OHT-treated conditional Krt76 knock-out mice and control sibling. Note paw pad phenotype from grooming transfer of tamoxifen. *p<0.05, **p<0.01. Error bars  =  S.E.M. Scale bars represent 50 µm.

Figure 6. KRT76 interacts with Claudin1.

(A) HIS-tagged KRT76 tail domain and HIS-tag alone where produced in E.coli, purified and immobilised on nickel-resin. Resin was then incubated with mouse paw pad lysates and the specific pull-down of CLDN1 with the KRT76- tail domain and not HIS-tag was shown by anti-Claudin1 WB. (B) Soluble extracts were prepared from A549 cells and anti-CLDN1 or non-immune IgG antibody immunoprecipitated. IP and lysate/input samples were then blotted for ZO-1, CLDN1 and KRT76. (C) A549 cells co-express CLDN1 and KRT76 and these colocalise in cytoplasmic punctate structures -see arrowheads.