A mouse model of brittle cornea syndrome caused by mutation in Zfp469

Abstract

Brittle cornea syndrome (BCS) is a rare recessive condition characterised by extreme thinning of the cornea and sclera. BCS results from loss-of-function mutations in the poorly understood genes ZNF469 or PRDM5. In order to determine the function of ZNF469 and to elucidate pathogenic mechanisms, we used genome editing to recapitulate a human ZNF469 BCS mutation in the orthologous mouse gene Zfp469. Ophthalmic phenotyping showed that homozygous Zfp469 mutation causes significant central and peripheral corneal thinning arising from reduced stromal thickness. Expression of key components of the corneal stroma in primary keratocytes from Zfp469BCS/BCS mice is affected, including decreased Col1a1 and Col1a2 expression. This alters the collagen type I/collagen type V ratio and results in collagen fibrils with smaller diameter and increased fibril density in homozygous mutant corneas, correlating with decreased biomechanical strength in the cornea. Cell-derived matrices generated by primary keratocytes show reduced deposition of collagen type I, offering an in vitro model for stromal dysfunction. Work remains to determine whether modulating ZNF469 activity will have therapeutic benefit in BCS or in conditions such as keratoconus in which the cornea thins progressively. This article has an associated First Person interview with the first author of the paper.

Keywords: Brittle cornea; Collagen; Keratocyte; ZNF469; Zfp469.

Fig. 1.

Pathogenic mutations in ZNF469 are distributed throughout the protein and result in premature termination codons or disruption of conserved residues in C2H2 zinc-finger domains. (A) Schematic representation of the protein structure of ZNF469, showing the position of previously reported pathogenic variants in relation to eight C2H2 zinc-finger domains and regions of compositional bias. Nonsense or frameshift variants in C2H2 ZF domains resulting in a premature stop are indicated by black arrows; missense mutations are indicated by red arrows. (B) Sequences of the eight C2H2 zinc-finger (ZF) domains in the mouse (Zfp469, NP_001354553.1) and human (ZNF469, NP_001354553.1) proteins were aligned using Clustal Omega. Fully conserved residues are indicated by ‘*’; ‘:’ indicates conservation between groups of amino acids with strongly similar properties; and ‘.’ indicates conservation between groups of amino acids with weakly similar properties. Residues shown in red are the paired cysteines (C) and histidines (H), which bind the zinc ion. Residues in yellow are structurally important hydrophobic amino acids.

Fig. 2.

Genome editing Zfp469 to create a mouse model of brittle cornea syndrome (BCS). (A) CRISPR/Cas9 genome editing was used to insert an in-frame V5 tag and premature stop codon into Zfp469 at p.Gly634, the conserved residue equivalent to the human mutation p.Gly677*. Sequencing chromatograms from a homozygous Zfp469^BCS/BCS mouse compared to a wild-type littermate indicating the position and sequence of the insertion. The V5 tag contained a spontaneously arising A>T mutation, resulting in the amino acid change Asn (N)>Ile (I) at position 5 in the V5 tag. (B) A representative image of two female Zfp469^BCS/BCS and two female Zfp469^+/+ littermate mice at 3 months of age. Zfp469^BCS/BCS mice showed no gross phenotypic abnormalities but were smaller than Zfp469^+/+ sex-matched littermates. (C) Bodyweight of Zfp469^BCS/BSC was significantly reduced relative to Zfp469^+/+ age- and sex-matched animals, as determined by one-way ANOVA with Dunnett's multiple comparison test (*P<0.05, **P<0.01). Data are mean±s.d. (D) qPCR analysis revealed no significant difference in the relative expression of Zfp469 in keratocytes isolated from the corneas of four wild-type, two heterozygous and four homozygous mice aged 3 months. Data are mean±s.d. with the average of assays performed in duplicate from each sample shown by data points.

Fig. 3.

Corneal thinning in Zfp469^BCS/BCS mice as a result of decreased stromal thickness. (A) Corneal opacity was not observed in 3-month-old Zfp469^BCS/BCS mice examined by slit lamp. (B) AS-OCT of 3-month-old Zfp469^BCS/BCS mice demonstrated extreme thinning of the central and peripheral cornea compared to Zfp469^+/+ and Zfp469^+/BCS age- and sex-matched mice. Scale bars: 200 µm. (C) CCT was decreased in Zfp469^BCS/BCS mice (n=4 for each genotype and sex; P=0.0025 in males, P<0.0001 in females, P<0.0001 for sexes combined; one-way ANOVA with Tukey's multiple comparison test). **P<0.01, ***P<0.001. Data are mean±s.d., with individual measurements from each eye shown by data points. (D) H&E staining of corneal sections from 3-month-old mice revealed visibly thinner corneal stroma in Zfp469^BCS/BCS mice.

Fig. 4.

Corneal thinning in Zfp469^BCS mice is apparent during corneal development and is not progressive. (A) Extreme thinning of the cornea in 1-month-old Zfp469^BCS/BCS mice compared to Zfp469^+/+ and Zfp469^+/BCS age- and sex-matched mice was observed by AS-OCT. (B) At 1 month of age, CCT was significantly decreased in Zfp469^BCS/BCS mice (P=0.0016 in males, P=0.0043 in females, P<0.0001 for sexes combined, one-way ANOVA with Tukey's multiple comparison test). N=3 for each genotype in males; n=3 for Zfp469^+/+ and Zfp469^BCS/BCS, n=4 for Zfp469^+/BCS in females. Data are mean±s.d., with individual measurements from each eye shown by data points. (C) Unilateral corneal opacity observed in one eye from a Zfp469^BCS/BCS female at 1 month of age resembled corneal edema (white arrowhead). Green arrows indicate the direction and position of the OCT scans. (D) CCT at 6 months of age remained similar to CCT measured at 1 and 3 months of age, with significant central and peripheral thinning apparent only in Zfp469^BCS/BCS mice. (E) CCT was significantly decreased in Zfp469^BCS/BCS mice at 6 months of age (P=0.0020 in males, P=0.0010 in females, P<0.0001 for sexes combined, one-way ANOVA with Tukey's multiple comparison test). N=4 for each genotype in males; n=4 for Zfp469^+/+, n=6 for Zfp469^+/BCS, n=5 for Zfp469^BCS/BCS in females. Data are mean±s.d., with individual measurements from each eye shown by data points. (F) Corneal distortion was seen by AS-OCT after the application of eye drops in Zfp469^BCS/BCS eyes, but not in heterozygous or wild-type sex-matched littermates, suggesting a loss of biomechanical strength in the mutant corneas. (G) Globular protrusion of the eye, resembling keratoglobus, was observed after dilation of the pupil in Zfp469^BCS/BCS female mice, but not in male mice or heterozygous or wild-type sex-matched controls. (H) ACD in the eyes of Zfp469^BCS/BCS female mice, measured from AS-OCT images, was significantly increased compared to heterozygous or wild-type sex-matched controls after dilation of the pupil. P=0.0047, Welch's ANOVA test). n=4 for Zfp469^+/+, n=3 for Zfp469^+/BCS, n=4 for Zfp469^BCS/BCS in males; n=4 for Zfp469^+/+, n=6 for Zfp469^+/BCS, n=5 for Zfp469^BCS/BCS in females. Box plot shows all data, with individual measurements from each eye shown by data points. Whiskers show minimum to maximum values. *P<0.05, **P<0.01, ***P<0.001.

Fig. 5.

Collagen type I is less abundant and affects fibril architecture in the stroma of Zfp469^BCS/BCS corneas. (A) Collagen type I and type V localise to the corneal stroma in 6-month-old mice, with decreased stromal area apparent in Zfp469^BCS/BCS compared to wild-type and heterozygous age- and sex-matched samples subjected to immunofluorescence microscopy. Nuclei are stained blue with DAPI, and collagen type I and type V are stained green. (B) Western blot showing that Col1a1 is decreased in Zfp469^BCS/BCS corneas at 2 months of age. Fibronectin-1 was used as a loading control. (C) Representative TEM of the cornea anterior stroma showed regular fibril organisation and packing of lamellae in mice at 9 months of age. (D) Colour-coded images of collagen fibrils in the images shown in C, showing that fibrils in wild-type corneal stroma are mostly green (40-49 nm), whereas homozygous mutant fibrils in the anterior stroma are mostly <29 nm (red). Fibrils with a diameter of 30-39 nm are coded yellow and those with a diameter of 50-59 nm are coded blue. (E) Fibril diameter was significantly decreased in the anterior stroma of Zfp469^BCS/BCS mice (P<0.0001, unpaired two-tailed Student's t-test). The range of diameters in 1059 measurements made in sections from three wild-type mice was 27.93-99.05 nm, and in 912 measurements made in sections from Zfp469^BCS/BCS mice the range was 15.79-45.70 nm. (F) The smaller fibrils in Zfp469^BCS/BCS stroma were more tightly packed, as shown by significantly increased fibril density in the corneal stroma of Zfp469^BCS/BCS mice, measured in 23 sections from three wild-type mice, and 22 sections from Zfp469^BCS/BCS mice (P<0.0001, unpaired two-tailed Student's t-test). ***P<0.001. Error bars are mean±s.d. Scale bars: 100 µm (A); 200 nm (C,D).

Fig. 6.

Mutation in Zfp469 affects the expression of ECM genes by corneal stroma fibroblasts in culture. (A) Relative expression of Zfp469 in primary keratocytes cultures, determined by qRT-PCR, was not significantly different between genotypes. There was an allele dosage effect on the expression of Zfp469 transcript containing the V5 tag and premature termination codon (V5_PTC), with Zfp469^+/BCS expressing half as much of this transcript as Zfp469^BCS/BCS keratocytes lines (n=5 for Zfp469^+/+, n=4 for Zfp469^+/BCS, n=3 for Zfp469^BCS/BCS; one-way ANOVA with Dunnett's multiple comparison test). Data are mean±s.d., with the average of assays performed in triplicate from each sample shown by data points. (B) Primary keratocytes and MEFs were seeded at known density at passage three, and the doubling time was determined by cell counting after 96 h in culture. No significant difference in doubling time was observed, although Zfp469^BCS/BCS keratocytes showed the greatest variability and longest average doubling time. Data are mean±s.e.m., with the doubling time of individual cell lines shown by data points. (C) Relative adhesion of Zfp469^BCS/BCS cells to fibronectin and fibrinogen-coated wells after 45 min compared to litter-mate wild-type keratocytes was significantly decreased [wild-type adhesion normalised to 1, average adhesion of Zfp469^BCS/BCS cells to fibronectin was 0.755±0.049 (mean±s.d.) (P=0.0133, Welch's t-test), and average adhesion of Zfp469^BCS/BCS cells to fibrinogen was 0.766±0.087 (mean±s.d.), P=0.0433, Welch's t-test]. Data are presented as relative adhesion (wild-type absorbance 560 nm/homozygous absorbance 560 nm)±s.d. Three cell lines for each genotype were used. (D) Relative expression of key genes in the ECM and in BCS was assessed by qRT-PCR in primary keratocytes. Col1a1, Col1a2 and Kera showed a significant Zfp469^BCS allele dosage effect on gene expression. Hprt and Gapdh were used to normalise the amount of mRNA, and data were analysed using the 2^−ΔΔCT method for relative quantitation. Fold change is relative to Zfp469^+/+. Data are mean±s.d., with the average of assays performed in triplicate from each sample shown by data points (n=5 for Zfp469^+/+, n=4 for Zfp469^+/BCS, n=3 for Zfp469^BCS/BCS; one-way ANOVA with Dunnett's multiple comparison test). ns, not significant.

Fig. 7.

Secretion and deposition of collagen type I by primary keratocytes is impaired when Zfp469 is mutated. (A) The concentration of proCol1a1 in serum-free medium after 7 days in culture, as measured by ELISA, showed that secretion of proCol1a1 into culture medium by confluent Zfp469^BCS/BCS primary keratocytes was significantly impaired compared to wild-type cells. Data are mean±s.d. with the proCol1a1 concentration of individual cell lines shown by data points (n=4 for Zfp469^+/+, n=4 for Zfp469^BCS/BCS; unpaired two-tailed Student's t-test). (B) Representative western blot of the serum-free conditioned medium samples used in ELISA that were probed with an anti-Telo Col1a1 antibody that detects prepro, pro and mature Col1a1 shows decreased Col1a1 in the medium of Zfp469^BCS/BCS. (C) CDMs generated by primary keratocytes were decellularised and then denatured and reduced prior to western blotting. The CDM of Zfp469^BCS/BCS contained less Telo Col1a1 than wild-type CDM. Fibronectin was used as a loading control. (D) Quantification of Col1a1 signal normalised to fibronectin for CDMs generated from three Zfp469^+/+ and three Zfp469^BCS/BCS by densitometry shows a significant decrease in the abundance of Telo Col1a1 in mutant cell lines (n=3 per group, P=0.0097; unpaired two-tailed Student's t-test). (E) Representative immunofluorescence images of CDMs generated by primary keratocytes after 20 days exposure to ascorbic acid-containing medium, and stained for collagen type I (green), collagen type V (green) and fibronectin (red). Scale bars: 20 µm. (F) Immunofluorescence staining intensity for collagen type I and collagen type V was normalised to fibronectin signal for CDMs from three Zfp469^+/+ and two Zfp469^BCS/BCS cell lines, using five non-overlapping images from each sample. Data are mean±s.d. with the average normalised collagen staining intensity of individual cell lines shown by data points. Deposition of collagen type V did not differ between genotypes but collagen type I deposition was decreased in Zfp469^BCS/BCS CDMs.