Keratin 12 missense mutation induces the unfolded protein response and apoptosis in Meesmann epithelial corneal dystrophy

Abstract

Meesmann epithelial corneal dystrophy (MECD) is a rare autosomal dominant disorder caused by dominant-negative mutations within the KRT3 or KRT12 genes, which encode the cytoskeletal protein keratins K3 and K12, respectively. To investigate the pathomechanism of this disease, we generated and phenotypically characterized a novel knock-in humanized mouse model carrying the severe, MECD-associated, K12-Leu132Pro mutation. Although no overt changes in corneal opacity were detected by slit-lamp examination, the corneas of homozygous mutant mice exhibited histological and ultrastructural epithelial cell fragility phenotypes. An altered keratin expression profile was observed in the cornea of mutant mice, confirmed by western blot, RNA-seq and quantitative real-time polymerase chain reaction. Mass spectrometry (MS) and immunohistochemistry demonstrated a similarly altered keratin profile in corneal tissue from a K12-Leu132Pro MECD patient. The K12-Leu132Pro mutation results in cytoplasmic keratin aggregates. RNA-seq analysis revealed increased chaperone gene expression, and apoptotic unfolded protein response (UPR) markers, CHOP and Caspase 12, were also increased in the MECD mice. Corneal epithelial cell apoptosis was increased 17-fold in the mutant cornea, compared with the wild-type (P < 0.001). This elevation of UPR marker expression was also observed in the human MECD cornea. This is the first reporting of a mouse model for MECD that recapitulates the human disease and is a valuable resource in understanding the pathomechanism of the disease. Although the most severe phenotype is observed in the homozygous mice, this model will still provide a test-bed for therapies not only for corneal dystrophies but also for other keratinopathies caused by similar mutations.

Figure 1.

Generation of a humanized mutant K12 mouse model. The endogenous mouse Krt12 gene in its genomic context (A) was targeted for homologous recombination by a vector incorporating the human KRT12 gene (B). The targeting vector contained two homology arms (LHA and SHA) for targeted replacement of the endogenous mouse Krt12 with the humanized genomic region. The human allele includes all UTRs, exons and introns with additional coding for a FLAG-HA tag immediately prior to the stop codon and a single nucleotide change in exon 1 creating the L132P mutant (B). Additionally, an MTC is included within the 3′-UTR. FRT- and F3-flanked selection markers were included within introns 2 and 7 to assist in the selection of clones where the transgene had been correctly targeted (C). Transgene-positive mice were finally interbred with transgenic mice universally expressing Flp recombinase to excise the selection markers (D). Krt20 and LOC are neighbouring genes.

Figure 2.

Microscopic assessment of MECD mouse corneas. (A) Representative images of H&E-stained 8-week-old mice corneas. WT mice typically have a tightly packed corneal epithelium, with subtle small intracellular spaces occasionally visible in the heterozygous mice (white arrows). Homozygous mutant mice exhibited greater phenotypic changes, with cells less tightly packed and with large intracellular spaces leading to cytolysis (asterisks); delamination of the corneal epithelium at the stromal interface (black arrows) was observed in the homozygous mutant mice. Scale bar = 100 µm. (B) SEM micrographs of the corneal surface (n = 1 per genotype, both eyes) show squames flaking or beginning to flake off; however, no significant differences were observed on the corneal surface between genotypes (scale bar, 50 µm). (C) TEM micrographs of the basal layer cells of each genotype (25 000× magnification) show the stroma (black asterisks) and nuclei (white asterisks). WT cells appear to have a more distinct structure with tight cell–cell interfaces in comparison to homozygous mutant cells. Vacuoles and structures associated with mechanical weakness (white arrows) are observed in these cells. The corneas of heterozygous mutant more closely resembled those of the WT mice; however, some evidence of an increase in vacuolization was noted (white arrows).

Figure 3.

Verification of KRT12 transgene expression. (A) The KRT12 transgene has a C-terminal FLAG-HA epitope tag. The HA epitope is present only in the heterozygous and homozygous mutant mice (brown staining). Staining was observed in all epithelial layers moving centrally and inward from the limbal region (scale bar, 100 µm). (B) Corneal protein lysates from 7 to 9-week-old mice (n = 3 per genotype) were simultaneously probed for K12 and HA. Only mK12 was observed in corneas from WT mice (green lower band), whereas only the HA-tagged protein (green hK12 and red HA co-localize for a yellow merged image) was present in homozygous mutant mice. Both proteins were present in heterozygous mice. β-actin was used as a loading control.

Figure 4.

Differential expression of keratin mRNA in MECD mouse corneas. qRT-PCR validation of keratin mRNA levels obtained by RNA-seq. Fold change in expression for each keratin was calculated by normalization to that observed in the WT mouse cornea. Similar expression levels were observed for RNA-seq and qRT-PCR in all keratins analysed except KRT10 and KRT17, in which greater expression was found in heterozygous and homozygous mice by qRT-PCR, whereas RNA-seq deemed these keratins to be comparably expressed in all three genotypes.

Figure 5.

Differential expression of keratins in MECD mouse cornea. (A) Representative pictures of IHC staining (brown, nuclei blue) of the central corneas of 8-week-old mice are shown. K5 was present throughout all layers of the corneal epithelium in all genotypes. Expression of K14 was markedly increased in the homozygous animals. K6 was found in the outer layer of all corneas with staining significantly increased in the homozygous mutant; K16 was also strongly upregulated in homozygous mutant mice. Scale bar, 100 µm. (B) Immunoblotting of corneal lysates (n = 3 per genotype) for keratins K5, K14, K6, K16 and K13, normalized to β-actin. K5 levels were greatly reduced when WT and heterozygous mice and heterozygous and homozygous mutant animals are compared. K14 was strongly upregulated in homozygous mouse cornea, with a small change in heterozygous mice. K6 increased with increasing transgene expression when WT and homozygous mice are compared. Similar changes in expression pattern were observed with K16 and K13.

Figure 6.

Differential expression of keratins in the human MECD corneal epithelium. Histological assessment of the human MECD cornea demonstrated a loss in epithelial organization in comparison to the healthy unaffected cornea. Sections from both corneas were also assessed by IHC for the expressions of K12, K5, K14, K6A and K16 (green). Nuclei were counterstained with DAPI (blue). The MECD tissue appeared to have a decreased abundance of K12 protein in the epithelium, compared with the control. Expression of K5, K14, K6A and K16 was increased in the MECD cornea, compared with that found in the control cornea. Scale bar = 100 and 200 µm, left and right images, respectively.

Figure 7.

Assessment of induction of the UPR and apoptosis in MECD mouse corneas. (A) IHC analysis of UPR markers, Caspase 12 and CHOP, was performed on corneal sections from WT (+/+), heterozygous (+/−) and homozygous MECD mice (green). Nuclei were counterstained with DAPI (blue). Scale bar = 100 µm. Positive staining for both UPR markers was observed in the homozygous mouse corneal epithelium, whereas subtle staining of Caspase 12 was found in the corneal epithelium of heterozygous mice. (B) A TUNEL assay was performed to identify the presence of apoptotic cells within the corneas of these mice (green). The number of TUNEL-positive cells was counted to ascertain the apoptotic cells, whereas DAPI-stained cells were counted to gauge the total number of cells. A significant increase in the proportion of apoptotic cells was observed in homozygous mice when compared with WT, with the rate in heterozygous mice being comparable to WT. Scale bar = 100 µm.