A Novel Human Congenital Cataract Mutation in EPHA2 Kinase Domain (p.G668D) Alters Receptor Stability and Function

Abstract

Purpose: To identify the genetic defect in a four-generation Chinese family that causes autosomal dominant congenital posterior subcapsular cataracts, and to understand how this EPHA2 kinase domain mutation affects EPHA2 activity.

Methods: Variants in 54 cataract-associated genes were screened by targeted next generation sequencing (NGS) and then validated by Sanger sequencing. EPHA2 wild-type cDNA was synthesized in vitro, and EPHA2 p.G668D mutant was constructed by PCR site-directed mutagenesis. Western blotting and fluorescence microscopy were used to analyze the expression level of protein and its subcellular localization, respectively. A wound-healing assay was performed to analyze changes to cell migration.

Results: A novel heterozygous missense mutation was identified in the kinase domain of the EPHA2 gene (c.2003G>A, p.G668D). This is the third congenital cataract mutation being reported in this domain. Functional study revealed that the kinase domain mutation (p.G668D) decreased EphA2 protein level (P = 0.036) via a proteasome-dependent pathway, altered its subcellular localization of the EphA2 from cell-cell contacts to a diffuse perimembranous distribution, and changed the distribution of β-catenin as well. The expression of mutant EphA2 significantly promoted the migration of human lens epithelial cells (P = 0.002).

Conclusions: Our study presented the evidence for a novel EPHA2 kinase domain mutation that causes congenital posterior subcapsular cataracts. The first functional study on an EPHA2 kinase domain mutation that causes a congenital cataract revealed that the G668D mutation destabilized the receptor, changed its subcellular localization, and altered the activation of EphA2 with its ligand ephrin. The mutant EphA2 resulted in a reduced inhibition of cell migration. As a consequence, the c.G668D mutation promoted cell migration and caused the formation of cataracts.

Cataract is the leading cause of blindness globally,¹ and is believed to cause 33% of the total visual impairment and 51% of blindness worldwide.² Congenital cataract is particularly serious as it has the potential to inhibit visual development, causing nystagmus, strabismus, amblyopia, or even permanent vision loss.³ In some cases, cataract may be inherited in a Mendelian fashion (∼1/10,000 births), either as an isolated phenotype or in association with other ocular and/or systemic abnormalities.^4,5 The inheritance pattern of cataract is most commonly autosomal dominant, but also can be seen as autosomal recessive, or X-linked.⁶ According to Online Mendelian Inheritance in Man, at least 42 genes have been identified for inherited forms of isolated or primary cataract with minimal additional ocular signs (e.g., microcornea). Identified genes mainly encode cytoplasmic crystallins, membrane proteins, cytoskeletal proteins, and DNA/RNA-binding proteins.⁷ 

The EPHA2 gene encodes one of the ephrin receptors that comprise the largest family of tyrosine kinase receptors.⁸ Structurally, it is a single-pass transmembrane glycoprotein that has multiple functional domains including an extracellular ligand-binding domain, a tyrosine kinase (TK) domain, and a sterile-α-motif (SAM) domain.⁹ EPHA2 is observed to play a crucial role in embryonic development,¹⁰ and is highly expressed in many solid tumors.¹¹ It is also expressed in both humans and mouse lenses,^12–14 and is believed to play an essential function in lens cell migration and lens fiber alignment.^14–17 Shiels and colleagues¹⁸ first linked a mutation in EPHA2 (p.G948W) to affected members of a four-generation pedigree with congenital posterior subcapsular cataract. According to the Cat-Map database (Jan 2019 version), 20 EPHA2 mutations have been identified to cause different forms of congenital cataract.⁴ Twelve of them are autosomal dominant, five are autosomal recessive, and three are sporadic. By using target region capture sequencing, we identified a novel EPHA2 gene mutation (c.2003G>A, p.G668D) in the kinase domain that results in congenital posterior subcapsular cataract. The mutation causes increased EphA2 protein degradation in a proteasome-dependent pathway, and also alters the subcellular distribution of the protein. These alterations result in an increased cell migration activity. To our knowledge, this is the first functional study of an EPHA2 kinase domain mutation that can cause congenital cataract. 

Methods

Patient Recruitment

A four-generation Han Chinese family from Zhejiang Province with autosomal dominant congenital cataracts (ADCCs) was recruited in the Eye Center, the Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China. All available individuals had comprehensive physical and ophthalmic examinations. A control group represented 100 unrelated healthy subjects with the same ethnic background. The project adhered to the guidelines of the Declaration of Helsinki and was approved by the Medical Ethics Committees of the Second Affiliated Hospital, College of Medicine, Zhejiang University (Hangzhou, China). Appropriate informed consent was obtained from each participant. 

Sample Collection and DNA Extraction

Peripheral blood was collected by venipuncture in EDTA-coated Vacutainer tubes. Genomic DNA of subjects was isolated from 2 mL peripheral blood samples using QIAamp DNA Blood kits (Qiagen, Hilden, Germany). The purity and quantity of DNA samples were then measured by the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, USA). 

Target Region Capture and Next Generation Sequencing (NGS)

Genomic DNA of the proband was sheared by the CovarisTM system. The sample was then prepared by using a Truseq DNA Sample Preparation Kit (Illumina, Inc., San Diego, CA, USA). The coding exons, flanking regions and promotor regions of 54 genes related to inherited cataracts (Supplementary Table S1) were selected and captured using a SureSelect Target Enrichment Kit (Agilent Technologies, Inc, Santa Clara, CA, USA) as previously described.¹⁹ The enrichment libraries were sequenced on an Illumina HiSeq2000 Sequencer (Illumina, Inc.). The average depth of sequencing was 500-fold. 

Bioinformatics Analysis

The low-quality reads and adaptor sequences were filtered out with the FASTX program. The Picard program was used to remove the PCR duplicates. After high-quality reads were retrieved, clean data were aligned using the BWA program according to human genome parameters (hg19). Subsequently, we determined single nucleotide polymorphisms (SNPs) using the SOAPsnp program, realigned the reads with BWA, and detected the deletions or insertion (InDels) with GATK software. After SNPs were identified, we used ANNOVAR to do annotation and classification. Finally, all nonsynonymous variants were evaluated by three algorithms, SIFT (http://sift.jcvi.org/), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), and Mutation Taster (http://www.mutationtaster.org/). 

Variant Identification and Validation

DNA samples of the proband and other family members were subjected to further Sanger sequencing, to confirm segregation of the potential pathogenic variants detected by the NGS. PCR was performed in a 20 μL reaction system using the primer pairs below: 

•  
  For the EPHA2 mutation:
•  
  Forward primer 5′-TGCCTGCTCGTAGGCAGCTT-3′
•  
  Reverse primer 5′-CTCTGCTGTGCTGCCTTGGG-3′
•  
  For the SIL1 mutation:
•  
  Forward primer 5′-CCCTCTAGGCGGATGATGTT-3′
•  
  Reverse primer 5′-GCATGCTGAAGACATCCTCG-3′

Sequenced PCR products were analyzed using SnapGene Viewer software (Version 4.3; GSL Biotech LLC, Chicago, IL, USA) and compared with sequences from the NCBI human genome database. 

Cell Culture

The Hek293T cell line was purchased from Genechem (Shanghai, China), cultured and maintained in Dulbecco's Modified Eagle's Medium–high glucose (Corning, Inc., Corning, NY, USA) supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA) and 1% penicillin-streptomycin mixture (Thermo Fisher Scientific, Inc., Waltham, MA, USA) at 37°C in the presence of 5% CO₂. 

Lentivirus Vector Construction, Lentivirus Infection and Screening

The full-length human EPHA2 (NM_004431.4) fragment was synthesized by GeneChem Co., Ltd. (Shanghai, China). The mutant EPHA2^G668D plasmid was constructed by using the following primers: 

•  
  Forward primer 5′-GGCATCATGGACCAGTTCAGCCACCACAACATC-3′
•  
  Reverse primer 5′-CTGAACTGGTCCATGATGCCGGCCTCGCCG-3′

After the DNA sequences of both EPHA2^WT and EPHA2^G668D were confirmed, the fragments were ligated into the GV358 vector (Ubi-MCS-3FLAG-SV40-EGFP-IRES-puromycin; GeneChem Co., Ltd.), and transformed into the competent Escherichia coli. Positive clones were confirmed by PCR and Sanger sequencing. The transfer and packaging plasmids (20 μg of GV358, 15 μg of pHelper 1.0, and 10 μg of pHelper 2.0) were mixed. Hek293T cells cultured on 6-cm plates were then transfected with 5 μg of DNA plasmid using Lipofectamine 2000 regent (Invitrogen, Carlsbad, CA, USA). The medium was replaced, and the cells were washed with PBS after 6 hours. Forty-eight hours after transfection, cell supernatants were collected by centrifugation at 10,000g for 10 minutes at room temperature and stored at −80 °C. Hek293T cells were infected with a double volume of recombinant viruses, and the cells were selected with 1 μg/mL puromycin for 2 weeks according to the manufacturer's protocol. Antibiotic-resistant colonies were then expanded for further analysis. Hek293T cell lines that stably expressed EphA2-Flag were confirmed by Western blotting. 

Western Blot Analysis and Inhibition of Protein Degradation

Protein Level in Transient Transfected Hek293T

The wild-type EPHA2 was synthesized by GeneChem Co., and pEGFP-N1-EPHA2^WT plasmids were constructed to create EPHA2- EGFP fusion proteins. The expression vector for G668D mutant EPHA2 (pEGFP-N1-EPHA2^G668D) was constructed using site-directed mutagenesis with the following primers: 

•  
  Forward primer 5′-GGCATCATGGACCAGTTCAGCCACCACAACATC-3′
•  
  Reverse primer 5′-CTGAACTGGTCCATGATGCCGGCCTCGCCG-3′

The mutation was confirmed by DNA sequencing. Hek293T cells at 80% confluence were transfected with 1 μg of pEGFP-N1-EPHA2^WT or pEGFP-N1-EPHA2^G668D plasmids DNA separately. Transfection was performed by using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Twenty-four hours after transfection, cells were treated with the proteasome inhibitor MG-132 (CalbioChem, San Diego, CA, USA) at 30 μM and a corresponding amount of dimethyl sulfoxide (DMSO) for 12 hours. 

Protein Level in Stable Hek293T Cell Lines

Hek293T cell lines stably expressing wild-type (WT) and G668D EphA2-Flag at 80% confluence were treated with either proteasome inhibitor MG-132 (10 μM or 30 μM), or DMSO (10 μM) for 12 hours. 

Hek293T cells were harvested and lysed with cell lysis buffer. Total protein was extracted and separated on 10% SDS-PAGE gels, and transferred to PVDF membranes separately. The membrane was then incubated with the following antibodies: anti-Actin (Cell Signaling, Danvers, MA, USA), anti-GFP (green fluorescent protein) (ProteinTech, Rosemont, IL, USA), anti-Flag-Tag antibody (Sangon Biotech, Shanghai, China), and anti-GAPDH antibody (Cell Signaling). Fluorescent signals were visualized using chemiluminescent detection. Protein band intensities were quantified using NIH ImageJ software (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA). 

Fluorescence Microscopy Analysis

Stably expressing Flag-EPHA2 Hek293T cells in 24-well plates were grown to 80% confluence. The cells were washed with PBS, fixed with 4% paraformaldehyde for 15 minutes at room temperature, and permeabilized in 0.5% Triton X-100 in PBS buffer for 15 minutes. Cells were then incubated overnight with the anti-Flag-Tag mouse monoclonal antibody (BBI Life Science, Shanghai, China), anti-Calnexin rabbit polyclonal antibody (Proteintech), anti-Giantin rabbit polyclonal antibody (Proteintech), and anti-beta-Catenin rabbit polyclonal antibody (Proteintech). After incubation with the anti-mouse Alexa Fluor 555 labeled secondary antibody (Cell Signaling) and anti-rabbit Alexa Fluor 647 labeled secondary antibody (Invitrogen) for 2 hours; the nuclei were then labeled with 4′,6-diamidino-2-phenylindole (0.5 mg/mL; Sigma-Aldrich, St. Louis, MO, USA). Images were captured using a Leica TCS SP8 confocal microscope (Leica, Wetzlar, Germany), merged and labeled by using NIH ImageJ software. Staining was repeated at least three times, and representative results are shown. 

Wound-Healing Assay

A wound-healing assay was performed following a previously described protocol.²⁰ HLE B3 cells were cultured on six-well plates. Wounds were introduced by scraping the confluent cell cultures with a micropipette tip, after being transfected with GV358 vector, GV358- EPHA2^WT plasmid, or GV358-EPHA2^G668D plasmid. Debris was removed and edge of the scratch was smoothed by washing the cells once with media. Images were captured using an Olympus (Tokyo, Japan) IX71 microscope at 0, 12, 24, 36, and 48 hours after scraping. The transfection efficiency was monitored by observing GFP signal from the GV358 vector. The relative migration distance of cells into the wound was determined using Adobe (San Jose, CA, USA) Photoshop CC software. 

Statistical Analysis

All the values in this article were presented as the standard deviation (±SD) of the mean from at least three independent experiments. Statistical analyses were performed using IBM SPSS Statistics software (Version 25; IBM, Armonk, NY, USA). A two-tailed t-test was used to analyze statistical significance between two groups. A P value less than 0.05 was considered statistically significant. 

Results

Clinical Features

Clinical histories were obtained from 14 members (Fig. 1A, six affected individuals and eight unaffected individuals) of a four-generation Chinese family affected with ADCC; all the affected members had a history of congenital cataracts. Blood samples from six available individuals in this family were collected. The proband was diagnosed with congenital cataracts at age of 18 months. His slit lamp examination showed bilateral posterior subcapsular cataracts with right posterior lenticonus (Fig. 1B). He underwent bilateral cataract surgery at age 2. His father's medical record also documented posterior polar cataract. Other affected individuals all had bilateral congenital cataract after birth, and underwent cataract surgeries in infancy or early childhood. 

Figure 1

 

The clinical features of the family. (A) Pedigree of the family. Solid boxes and circles indicate affected males and females. The Proband (IV:2) is marked with a black arrow. Plus (+) signs indicate the c.2003G>A mutation in EPHA2 gene, and minus (−) signs indicate a normal EPHA2 genotype on one allele; thus, heterozygotes are depicted as +/−, and WT as −/−. (B) An anterior segment photograph of the proband age 2 illustrated a posterior subcapsular cataract (opacity marked with a white arrow).

Target Region Capture and NGS

By using the custom-made capture panel described in the Methods, 94.7% of designed target regions were covered in the proband. Two rare mutations, in SIL1 (NM_022464) c.1093C>T (p.R365C) and EPHA2 (NM_004431) c.2003G>A (p.G668D), were identified by NGS. 

Bioinformatics Analysis

The SIL1 c.1093C>T mutation was predicted to be “probably damaging” by PolyPhen-2 with a score of 0.937, “deleterious” by SIFT with a score of 0.02, and “disease causing” by MutationTaster with a score of 0.884. SIL1 mutations have been linked to Marinesco-Sjögren Syndrome defined by early onset of cataracts, myopathy, and ataxia.^21,22 

The EPHA2 gene c.2003G>A mutation was predicted to be “probably damaging” by PolyPhen-2 with a score of 0.999, “deleterious” by SIFT with a score of 0, and “disease causing” by MutationTaster with a score of 0.999. The connection between EPHA2 mutations and cataracts has been well established. Multiple sequence alignments of the EPHA2 gene showed that glycine at position 668 of EphA2 is highly conserved among various species (Fig. 2A). 

Figure 2

 

The causative mutation in the family. (A) Multiple species alignments of the region of the human EPHA2 protein including residue 668 showed conservation of the mutated residue across species. (B) Forward Sanger sequencing showed the EPHA2 mutation c.2003G>A (p.G668D), marked with an arrow, to be present in the proband.

Variant Identification and Validation

Sanger sequencing of both mutations was done. The results indicated only subject III:5 and IV:2 carried SIL1 c.1093C>T mutation. This mutation thus did not co-segregate with the disease, and was assumed to be an SNP. The EPHA2 c.2003G>A mutation was confirmed to co-segregate with congenital cataracts in available patients by Sanger sequencing (Fig. 2B). The status of EPHA2 c.2003G>A mutation in available participants in this family is marked in Figure 1A. 

Western Blot Analysis and Inhibition of Protein Degradation

In Hek293T cells that were transiently transfected with pEGFP-N1-EPHA2 plasmids, Western blot indicated that EphA2^G668D-GFP mutant protein level was significantly reduced compared with EphA2^WT-GFP (Figs. 3A, B; DMSO group, P < 0.001). However, after 12 hours of treatment with proteasome inhibitor MG-132, the level of mutant EphA2^G668D-GFP protein recovered significantly (Figs. 3A, 3B; P < 0.001). The difference between EphA2^WT-GFP protein level and EphA2^G668D-GFP protein level was not as significant as that in the DMSO group (Figs. 4A, 4B; MG-132 group, P = 0.039). 

Figure 3

 

Protein level of EphA2^WT and EphA^G668D in Hek293T cells. (A) EphA2-GFP protein level in transient transfected Hek293T cells. The EPHA2 c.2003G>A mutation significantly reduced protein expression level of EphA2-GFP. This decrease can be partially rescued by 12-hour treatment with proteasome inhibitor MG-132. (B) The relative EphA2-GFP amount compared to β-actin. The error bars shown for DMSO showed a significant difference, whereas those for MG-132 did not. (C) EphA2-Flag protein level in stable Hek293T cells. Compared with the EphA2^WT-Flag, EphA2^G668D-Flag level significantly decreased (P = 0.036). No difference was detected between EphA2^WT-Flag and EphA2^G668D-Flag in both 10 μM and 30 μM MG-132 treatment groups (P > 0.05). (D) The relative EphA2-Flag amount compared with GADPH. Each bar represents the quantification (mean ± SD) of Western blots from three independent experiments. ***P < 0.001; *P < 0.05; and ns, no statistically significant difference between the two groups. Values of P < 0.05 were considered to be statistically significant.

Figure 4

 

(A) Colocalization of EphA2 with the Golgi apparatus marker Giantin. A few cells expressing G668D mutant EphA2 showed some colocalization with the Golgi apparatus (white arrows). (B) Colocalization of EphA2 with the ER marker Calnexin. A few cells expressing G668D mutant EphA2 showed little colocalization with the ER (white arrows). Representative images are shown from three independent experiments. Scale bar: 10 μm.

In stable cell lines, we also detected a significant EphA2-Flag protein level decline in the G668D mutant (Figs. 3C, 3D; P = 0.036), although not as drastic as in the transiently transfected cells. This difference can also be recovered by the treatment of both 10 mM and 30 mM MG-132 for 12 hours (Figs. 3C, 3D; WT and G668D groups, P > 0.05). 

Our data indicate that the EPHA2 c.G668D mutation in the kinase domain may enhance proteasome-mediated EphA2 protein degradation, and results in a relatively low EphA2 receptor level. 

Immunofluorescence Analysis

To investigate the subcellular localization of both WT and G668D mutant EphA2 receptors in the cell, stable Hek293T cell lines were used expressing the EPHA2 gene from integrated lentiviral vectors. Anti-Giantin rabbit polyclonal antibody was used to label the Golgi apparatus, and anti-Calnexin rabbit polyclonal antibody was used to label endoplasmic reticulum (ER). 

Comparing cells that expressed WT EphA2, the red fluorescence signal, which indicated the EphA2-Flag protein, was relatively weaker in cells that expressed G668D mutant protein. Both WT and G668D proteins were predominantly seen in the cell periphery. However, WT EphA2 had distinct membrane expression, especially at the cell-cell border. Unlike the WT EphA2, the G668D mutant had a more diffuse perimembranous distribution pattern (Figs. 4, 5; EphA2-Flag group). By observing 331 cells in the WT group and 211 cells in the G668D group from five separate fields, we found that the 24 cells (7.25%) that express EphA2^WT had a perimembranous distribution, whereas the 106 cells (54.07%) that express EphA2^G668D showed a perimembranous distribution. Few cells in the G668D group showed some colocalizations of EphA2 protein with the Golgi apparatus and ER (Figs. 4A, 4B, white arrows), whereas cells in the WT group did not. 

Figure 5

 

Colocalization of the EphA2 with the β-catenin. WT EphA2 colocalized with β-catenin precisely at cell-cell contacts (white arrows). The cells expressing G668D mutant EphA2 showed a noticeably diffuse pattern of β-catenin distribution. Some β-catenin appeared in the cytoplasm where was no EphA2 signal (cyan arrows). Representative images are shown from three independent experiments. Scale bar: 10 μm.

Given that previous studies have shown a close relationship between EphA2 and β-catenin, we conducted a double labeling immunofluorescence experiment of EphA2 and β-catenin in stable Hek293T cell lines. In the cells that expressed WT protein, β-catenin was colocalized with the EphA2 receptor almost exclusively at cell-cell contacts. In contrast, noticeably diffused distribution of β-catenin appeared in the cells that expressed G668D protein, affecting approximately 60% of total cell populations. G668D mutant protein still showed some perimembranous colocalization with β-catenin, although not precisely colocalized. Some β-catenin was sparsely distributed in the cytoplasm where there was no EphA2 signal (Fig. 5). 

Wound-Healing Assay

To understand how the p.G668D mutation affected EphA2 function, we investigated its effect on cell migration using a wound-healing assay, a common assay of cell migration. Compared with control, HLE B3 cells transfected with GV358-EPHA2^WT plasmid showed no difference in migration after 24 hours and 48 hours (P = 0.443 and P = 0.474, respectively). HLEB3 cells expressing EPHA2^G668D migrated remarkably faster than WT, occupying 38.33% ± 2.39% (P = 0.004) of the cell-free area after 24 hours and up to 78.55% ± 2.32% (P = 0.002) after 48 hours (Figs. 6A, 6B). These observations suggested that the EPHA2 p.G668D mutation promotes cell migration. Dysregulated cell migration might then be the primary cause of cataract formation. 

Figure 6

 

EPHA2 gene p.G668D mutation promotes cell migration. (A) A scratch wound was made with a micropipette tip, and the wound was monitored 0, 12, 24, 36, and 48 hours after scraping. (B) Quantification of the effect of EPHA2 gene p.G668D mutation on HLE B3 migration. Statistical differences were calculated by two-tailed t-tests. **P < 0.01; ns, not significant. A P value less than 0.05 was considered statistically significant.

Discussion

EphA2 Receptor Function in Animal Lens

Cooper et al.²³ first demonstrated that ephrin-A5 interacts with the EphA2 receptor to regulate the adherens junction complex by enhancing recruitment of beta-catenin to N-cadherin, suggested that ephrin receptors and their ligands are critical regulators of lens development and maintenance. Jun et al.¹⁴ then reported that homozygous deletion of EPHA2 in two independent strains of mice resulted in progressive cortical cataract, and concluded that EPHA2 helps maintain lens clarity with age. The lens fulcrum is a distinct region at the lens equator where the apical tips of elongating epithelial cells constrict to form an anchor point before fiber cell elongation in the lens. In the normal lens, hexagonal equatorial epithelial cells are packed into organized meridional rows at the lens fulcrum. Cheng et al.¹⁵ found that in the EPHA2−/− mouse lens, equatorial cells failed to form precisely aligned meridional rows, and the lens fulcrum was disrupted. Further study showed that, EphA2 and ephrin-A5 did not form a lens receptor-ligand pair, and EphA2 had other binding partners in the lens to help align differentiating equatorial epithelial cells.²⁴ In contrast, Zhou and Shiels¹⁷ did not find cataracts, but only optical degradation, in Epha2-null lenses generated by homologous recombination on the B6J background. All these data suggest that the EphA2 receptor plays an important role in lens development and lens fiber alignment, yet the exact function of EphA2 in the lens is still not well-understood. 

EPHA2 Mutations in Human Congenital Cataract

Shiels and colleagues¹⁸ first reported that EPHA2 mutation can cause human cataract. To date, 21 mutations in EPHA2 have been reported to cause congenital cataract (Table), which are dispersed in different domains. 

Table

 

Summary of the Known EPHA2 Gene Mutations That Cause Congenital Cataract

 

Fewer mutations were identified (6/21) in the extracellular region that result in congenital cataract. Five of them were located in the fibronectin III regions. Among them, two were autosomal recessive, one sporadic, one autosomal dominant, and one unknown. Interestingly, there was no mutation located in the ligand-binding domain that could cause congenital cataract so far. In general, congenital cataract causative mutations were more often seen in the intracellular region (15 of 21), and were most commonly inherited in an autosomal dominant pattern (11 of 15). This information suggests that intracellular region is crucial in maintaining the stability of the lens. One-third of the mutations (7 of 21) were clustered in the SAM domain, and they were believed to destabilize the receptor and cause the loss of cell migration activity.²⁵ Other mutations located in the kinase domain (3 of 21) and PDZ-binding motif (2 of 21) suggest that these two domains also play important roles in stabilizing the lens. However, the functional changes caused by mutations in these two domains remain unknown. 

EPHA2 Kinase Domain and Cataract

The novel EPHA2 missense mutation presented in this study, c.2003G>A (p.G668D) specifically affects the potential kinase domain composed of amino acids 613 to 871. The kinase domain is a structurally conserved domain containing the catalytic function of protein kinases. A consequence of many kinase domain mutations, in other types of ephrin receptors, is a reduction in cell surface localization, which suggests that the mutations cause misfolding and/or alter receptor trafficking.²⁶ 

Our Western blot results suggested a decline in mutant protein level. Also, the mutant proteins were likely degraded in a proteasome-dependent pathway, as treatment with the proteasome inhibitor MG-132 enhanced EphA2 protein levels. From all the evidence mentioned above, a EphA2 loss-of-function resulted from a reduced G668D mutant protein level was first considered. However, Western blot results from stable Hek293T cells suggested that the mutation decreased protein level by 30%. Given this is a heterozygous mutation, the EphA2 protein level would be even higher in lens tissue in vivo. Considering Epha2-null lenses in a B6J background mouse did not result in cataracts but only lenses with degraded optical quality¹⁷ at birth, it is unlikely that the serious phenotype observed in the family described here resulted only from an approximate 15% reduction of the EphA2 level. Previous studies also pointed out that the level of EphA2 kinase activity not only relied on the amount of protein, but also largely depended on the phosphorylation level of tyrosine. Singh et al.²⁷ suggested some EPHA2 mutations changed the Ser-897 and Tyr-772 phosphorylation levels of TK domain, and resulted in a stimulation effect of cell migration. Thus, we sought other reasons to explain the decrease of function of EphA2 we observed in this study. 

G668D mutation discovered in this study was the third kinase domain mutation linked to congenital cataract. The first reported mutation, c.2353G>A (p.A785T), caused autosomal recessive congenital cataracts in a Pakistani family.²⁸ The second TK domain mutation, c.2007G>T (p. Q669H), reported recently, caused autosomal dominant infantile cataract in a Saudi Arabian family.²⁹ According to the multiple species alignment (Fig. 2A), amino acids in the kinase domain that were affected in ADCC (668D and 669Q) were both highly conserved across different species, suggesting that this region of the EPHA2 gene plays an important role in lens development. However, no functional study on the TK domain has ever been conducted before mutations that relate to congenital cataract. Only one TK domain mutation EPHA2 c.R721Q, which associated with age-related cataract, was found to increase the basal activation level of EphA2 kinase.¹⁴ 

Despite limited information on EPHA2 kinase domain mutations related to cataracts, somatic kinase domain mutations that are seen in solid tumor have been extensively studied. Chavent et al.³⁰ found that the interaction of the EphA2 receptor TK domain with the cell membrane was largely affected by the interaction between the basic residues on the surface of kinase domain with anionic lipids at the cytoplasmic surface of the cell membrane. Mutations in the kinase domain changed the interaction between EphA2 with lipids containing membrane, and thus changed the autophosphorylation level of the kinase. The glycine at position 668 is one of the nonpolar amino acids that tends to cluster its side chains together in the interiors of proteins. In contrast, aspartate is a negatively charged amino acid that is very polar and is almost always found on the outsides of proteins. This significant difference between the two amino acids could be the underlying reason for the pathogenicity of the G668D mutation, consistent with the suggestion of Chavent et al.³⁰ that the change of the EphA2 kinase domain surface charge altered the autophosphorylation level of the EphA2 kinase. If we take a step back to look at the EPHA2 R721Q and A785T mutations in the kinase domain, which were also found to cause congenital cataracts, we will find that they also result in a dramatic change in amino acid polarity. On the other hand, another known SNP (rs1479127438) resulting in a G668A mutant protein, does not result in a drastic change in amino acid polarity, as both glycine and alanine are nonpolar aliphatic residues. Thus, G668A mutation did not cause a pathologic phenotype. 

EPHA2 and Beta-Catenin

This study also detected the disruption of β-catenin distribution in the stable cells that express G668D mutant protein. Beta-catenin serves a key role in epithelial-mesenchymal transition (EMT), and EMT of lens epithelial cells contributes to the regeneration of crystallin-expressing lenticular fibers and folding of lens epithelial cells on the posterior capsule.³¹ A disruption of β-catenin distribution can thus potentially harm the lens differentiation process. 

Cooper et al.²³ indicated that EphA2 may serve as a membrane anchor for β-catenin, and activation of EphA2 promotes recruitment of β-catenin to N-cadherin. Peng et al.³² also found that appropriate EphA2 function promotes the stabilization of β-catenin. Cheng and Gong³³ indicated that noticeably diffused distribution of β-catenin proteins appeared in EphA (−/−) fiber cells. In the present study, our data suggested that β-catenin in the cells that express mutant EphA2 presented a diffused distribution. This finding is similar to Cheng and Gong's observation in EphA (−/−) mouse lens.³³ This provided a piece of evidence to support that G668D mutant results in an alteration of EphA2 function. 

EPHA2 and Cell Migration

The EphA2 receptor is a transmembrane protein that spans the cell membrane. This feature allows it to interact with surface-associated ligands, and to trigger downstream signal pathways. EPHA2 was found to be highly expressed in tumors and played a crucial role in regulating the cell migration,²⁶ as reviewed by Dunne et al.³⁴ Some studies suggested that EphA2 was a poor prognostic marker in cancer and might promote cell migration and tumor metastasis. 

Park et al.²⁵ found that WT EphA2 receptor promoted the migration of mouse lens epithelial αTN4–1 cells, whereas mutants exhibit significantly reduced migration activity. Considering the information that links EphA2 to cell migration, we evaluated the functional effects of the EPHA2 kinase domain mutation using the wound-healing assay. Similar to the result of Park et al.,²⁵ EPHA2 gene overexpression in our study tended to favor cell migration, although a statistical significance was not achieved in the present study (P > 0.05). However, rather than an inhibition of cell migration caused by SAM domain mutations, c.G668D mutation in the kinase domain promoted cell migration in HLE B3 cells. 

EphA2 receptors plays a complex role in cell migration. A previous study revealed that overexpression of the EphA2 receptor could either promote cell migration or inhibit migration. Miao et al.³⁵ addressed that a possible cause of this apparent paradox is diametrically opposite roles of EphA2 in regulating cell migration and invasion. They found that the activation of EphA2 with its ligand ephrin-A1 inhibited migration of glioma cells, whereas EphA2 overexpression promoted migration in a ligand-independent manner.³⁵ Cheng et al.²⁴ showed that expression of ephrin-As (A1, A3, A4, and A5) and ephrin-Bs (B1, B2 and B3) were both detected in lens epithelial cells. However, the binding partner of EphA2 in the lens is still unknown.²⁴ To simulate the ephrin expression status in vivo, we studied a human lens epithelial cell line in our wound-healing assay. Our result revealed that G668D promoted cell migration in HLE B3 cells; the mechanism of this effect remains unknown. An increased level of ligand-independent EphA2 activation can be one explanation. 

Another more reliable explanation for an increased cell migration caused by G668D would be a decreased EphA2 inhibition on cell migration due to less interaction between EphA2 receptor with its ligands. Zelinski et al.³⁶ revealed that although breast carcinoma cells overexpressed the EphA2 receptor, the protein was diffusely distributed throughout the cytoplasm. This reduction of protein cell surface localization prevents EphA2 from interacting with its ligands, and facilitates metastasis. Jun et al.¹⁴ also observed cytosolic retention of a EPHA2 kinase domain mutation (c.R721Q). A reduction in cell surface localization could be a common consequence for the kinase domain mutations. The mutant EphA2^G668D protein also had a more diffuse perimembranous distribution pattern, which might have affected its interaction with a ephrin ligand.³² Future investigations are needed to draw more definitive conclusions on how mutations in the kinase domain affect cell migration. 

To conclude, our study reported a novel EPHA2 gene mutation (c.2003G>A, p.G668D) in the kinase domain that results in a congenital posterior subcapsular cataract in a four-generation Chinese family. Functional studies revealed that the consequence of the G668D mutation is an altered EphA2 receptor function, supported by the β-catenin diffuse distribution in cell lines that express the G668D mutant, which was similar to the β-catenin distribution in the lenses of EPHA2 double knockout mice. The altered EphA2 function is believed to be caused by three mechanisms. First, the mutation destabilized the EphA2 protein in a proteasome-dependent pathway, which was supported by both Western blotting results. Moreover, the change of a nonpolar glycine at position 668 by a negatively charged polar aspartate altered the interactions between EphA2 and its binding partners. This was supported by the fact that other cataract-causing mutations in the kinase domain also resulted in similar consequences. Finally, the G668D mutation changed the EphA2 subcellular localization from a precise cell-cell border pattern to a diffuse perimembranous distribution, and this prevented EphA2 from interacting with its ligand properly. All three conditions may decrease the activation of EphA2 with its ligand ephrin, which is believed to result in the inhibition of cell migration. Thus, the G668D mutation promoted cell migration and caused the formation of cataract. 

Acknowledgments

The authors thank the family for participating in this study. The authors also express sincere gratitude to Ian M. MacDonald, MD CM, for his help with developing and editing this manuscript. 

Supported by Program of National Natural Science Foundation (No. 81570822, No. 81870641, and No.81700816), Zhejiang Key Laboratory, Fund of China (No2011E10006), and Medical science and technology project of Zhejiang Province (2017209519). 

Disclosure: Y. Zhai, None; S. Zhu, None; J. Li, None; K. Yao, None