Mice Deficient in TAZ (Wwtr1) Demonstrate Clinical Features of Late-Onset Fuchs' Endothelial Corneal Dystrophy

Abstract

Purpose: We sought to define the role of Wwtr1 in murine ocular structure and function and determine the role of mechanotransduction in Fuchs' endothelial corneal dystrophy (FECD), with emphasis on interactions between corneal endothelial cells (CEnCs) and Descemet's membrane (DM).

Methods: A Wwtr1 deficient mouse colony was established, and advanced ocular imaging, atomic force microscope (AFM), and histology/immunofluorescence were performed. Corneal endothelial wound healing was assessed using cryoinjury and phototherapeutic keratectomy in Wwtr1 deficient mice. Expression of WWTR1/TAZ was determined in the corneal endothelium from normal and FECD-affected patients; WWTR1 was screened for coding sequence variants in this FECD cohort.

Results: Mice deficient in Wwtr1 had reduced CEnC density, abnormal CEnC morphology, softer DM, and thinner corneas versus wildtype controls by 2 months of age. Additionally, CEnCs had altered expression and localization of Na/K-ATPase and ZO-1. Further, Wwtr1 deficient mice had impaired CEnC wound healing. The WWTR1 transcript was highly expressed in healthy human CEnCs comparable to other genes implicated in FECD pathogenesis. Although WWTR1 mRNA expression was comparable between healthy and FECD-affected patients, WWTR1/TAZ protein concentrations were higher and localized to the nucleus surrounding guttae. No genetic associations were found in WWTR1 and FECD in a patient cohort compared to controls.

Conclusions: There are common phenotypic abnormalities seen between Wwtr1 deficient and FECD-affected patients, suggesting that Wwtr1 deficient mice could function as a murine model of late-onset FECD. Despite the lack of a genetic association between FECD and WWTR1, aberrant WWTR1/TAZ protein subcellular localization and degradation may play critical roles in the pathogenesis of FECD.

Figure 1.

Wwtr1 deficient mice exhibit reduced corneal endothelial cell density (ECD) and abnormal morphology when compared with WT controls. (A) In vivo confocal microscopy revealed reduced ECD in both Wwtr1^+/− and Wwtr1^−/− mice at 2 months compared with WT mice. Despite age-related corneal endothelial cell (CEnC) loss, Wwtr1 deficient mice continued to have further declines in ECD at 6 and 12 months of age. Additionally, CEnCs from Wwtr1 deficient mice demonstrated abnormal morphology, lacking the normal hexagonal shape and even distribution, and altered reflectivity when compared with WT mice. (B) Each dot represents the mean ECD from three representative regions in a single eye (n ≥ 9 eyes in each group), horizonal line represents mean of the group and error bars reflect standard deviation. A 2-way ANOVA was performed to detect statistical differences, *, **, ***, **** represent P < 0.05, P < 0.01, P < 0.001, P < 0.0001, respectively. Scale bar equivalent to 50 µm.

Figure 2.

Wwtr1 deficient mice demonstrate abnormal tissue biomechanics of Descemet's membrane and irregular protein localization in corneal endothelial cells (CEnCs) compared with WT controls. (A) Atomic force microscopy (AFM) performed on Descemet's membrane revealed decreased elastic modulus (softer) in Wwtr1 deficient mice, with Wwtr1^−/− mice having the lowest measurement, followed by Wwtr1^+/− mice and finally WT mice (stiffest). Each dot represents a single eye (n ≥ 9 eyes in each group with each dot representing an average elastic modulus over five locations and five force curves were generated for each location), horizonal line represents mean of the group and error bars reflect standard deviation. A 2-way ANOVA was performed to detect statistical differences, *, **, ***, **** represent P < 0.05, P < 0.01, P < 0.001, P < 0.0001, respectively. (B, C) Representative CEnC flatmount preparations from WT, Wwtr1^+/−, and Wwtr1^−/− mice at 2, 6, and 12 months of age, staining for Na/K-ATPase (B, green) and ZO-1 (C, red), and nuclei (DAPI, blue). Both WT and Wwtr1^+/− demonstrated discrete localization of Na/K-ATPase and ZO-1 to the cell periphery, whereas CEnCs from Wwtr1^−/− mice had diffuse and lobulated staining for Na/K-ATPase and disrupted to absent ZO-1 staining. Scale bar equivalent to 20 µm.

Figure 3.

Wwtr1 deficient mice demonstrated impaired corneal endothelial cell (CEnC) regeneration after corneal cryoinjury. A cryoinjury wound was created with a 2 mm diameter steel probe immersed in liquid nitrogen for 3 minutes (–196°C) and subsequently applied to the cornea for 10 seconds in WT (n = 8), Wwtr1^+/− (n = 8), and Wwrt1^−/− (n = 5) mice. On day 2, animals were euthanized, and the right eyes were stained with Alizarin red to calculate total denuded area (A) and the left eyes were stained with EdU to assess cell proliferation (B). (A) On day 2, there was a significantly larger denuded area in Wwtr1^+/− mice compared with WT mice. Additionally, Wwtr1^−/− mice had a trend towards larger denuded areas compared with WT mice. (B) There were no differences in EdU staining, indicating that the proliferative capacity of the CEnCs were equal across groups. Both Alizarin red and EdU staining were analyzed using Kruskal-Wallis tests, ** P < 0.01. Alizarin red scale bars equivalent to 500 µm and EdU staining scale bars equivalent to 20 µm (inset).

Figure 4.

Wwtr1 deficient mice demonstrated impaired corneal endothelial wound healing after phototherapeutic keratectomy (PTK). An endothelial wound was created with an excimer laser (20 µm depth) in 6-month-old WT (n = 6) and Wwtr1^+/− (n = 6) mice. Optical coherence tomography (OCT) (A, B), slit lamp biomicroscopy (C), and in vivo confocal microscopy (IVCM) (D) were performed on post-injury days 1, 3, 5, 7, 10, and 14. Central corneal thickness (CCT) measured with OCT did not show statistical difference between the two genotypes (B), however, CCT reached its maximal thickness on post-injury days (PIDs) 7 to 10 and corneal edema was more marked on PID 10 in Wwtr1^+/− mice (C). IVCM revealed the regenerating course of injured endothelial cells throughout the study period (D). Following euthanasia on PID 14, endothelial wholemounts were stained with alizarin red (E), and endothelial cell density (ECD) was measured. ECD was significantly reduced in Wwtr1^+/− mice versus WT, suggesting impaired endothelial healing in Wwtr1 deficient mice. CCT was analyzed using a 2-way ANOVA, ECD was analyzed using a Mann-Whitney U test, * P < 0.05. OCT scale bars equivalent to 250 µm, IVCM and Alizarin red microscopy scale bars equivalent to 50 µm.

Figure 5.

WWTR1/TAZ expression levels are similar to that of other genes implicated in FECD. (A) RNA was extracted from healthy human endothelial tissue from eye bank donors without the TCF4 repeat and expression values were extracted for WWTR1 and other genes known to play a role in FECD pathogenesis. Box and whisker plot with each dot representing a single patient specimen. (B) RNA was extracted from healthy human endothelial tissue from eye bank donors without the TCF4 repeat (control), presymptomatic with the repeat expansion (presymptomatic), patients with late-stage FECD without the repeat (FECD_NR), and patients with late stage FECD with the repeat (FECD). RNA sequencing was performed and WWTR1 expression was compared across groups. Each dot represents a single patient specimen, horizonal line represents mean of the group, and error bars reflect standard deviation. Data were analyzed using a Kruskal-Wallis test (P > 0.05).

Figure 6.

TAZ protein expression is elevated in FECD surgical explants compared to non-FECD healthy controls. (A) Microfluidic detection of TAZ protein in FECD surgical explants (N = 16, in pools of 4) had 92% higher expression of TAZ protein than non-FECD, healthy donor samples (N = 12, in pools of 4), where each samples’ TAZ expression was normalized to total protein in that capillary (** P = 0.0046). (B) Localization of TAZ protein (green) in FECD surgical samples is higher in the nucleus in cells adjacent to guttae compared to cells not adjacent to guttae. As cells lose confluency, or are growing on top of developing guttae, the nuclear translocation of TAZ increases. The expression of TAZ protein is higher in immunohistochemical images of FECD samples compared to non-FECD donor samples. Images are representative samples of seven FECD and four donor controls used in the study. Nuclear counter stain = DAPI (blue). Scale bars are 50 µm.