Figures
Abstract
Posterior polymorphous corneal dystrophy 1 (PPCD1) is a genetic disorder that affects corneal endothelial cell function and leads to loss of visual acuity. PPCD1 has been linked to a locus on chromosome 20 in multiple families; however, Sanger sequencing of protein-coding genes in the consensus region failed to identify any causative missense mutations. In this study, custom capture probes were utilized for targeted next-generation sequencing of the linked region in a previously reported family with PPCD1. Variants were detected through two bioinformatics pipelines and filtered according to multiple criteria. Additionally, a high-resolution microarray was used to detect copy number variations. No non-synonymous variants in the protein-coding region of annotated genes were identified. However, 12 single nucleotide variants in 10 genes, and 9 indels in 7 genes met the filtering criteria and were considered candidate variants for PPCD1. Eleven single nucleotide variants were confirmed by Sanger sequencing, including 2 synonymous variants and 9 non-coding variants, in 9 genes. One microdeletion was detected in an intron of OVOL2 by microarray but was subsequently not identified by PCR. Using a comprehensive next-generation sequencing approach, a total of 16 genes containing single nucleotide variants or indels that segregated with the affected phenotype in an affected family previously mapped to the PPCD1 locus were identified. Screening of these candidate genes in other families previously mapped to the PPCD1 locus will likely result in the identification of the genetic basis of PPCD1.
Citation: Le DJ, Chung D-WD, Frausto RF, Kim MJ, Aldave AJ (2016) Identification of Potentially Pathogenic Variants in the Posterior Polymorphous Corneal Dystrophy 1 Locus. PLoS ONE 11(6): e0158467. https://doi.org/10.1371/journal.pone.0158467
Editor: Alvaro Galli, CNR, ITALY
Received: October 31, 2015; Accepted: June 16, 2016; Published: June 29, 2016
Copyright: © 2016 Le et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All array comparative genomic hybridization files are available from the GEO DataSets database (accession number GSE72617). All next generation sequencing files will be available after the manuscript is accepted.
Funding: The following are the sources of funding for this work: National Eye Institute (https://nei.nih.gov/), R01EY022082 (AJA), core grant P30EY000331; Research to Prevent Blindness (https://www.rpbusa.org/rpb/), unrestricted grant (AJA). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The corneal dystrophies are a heterogeneous group of genetic disorders that are associated with bilateral, progressive loss of visual acuity due to changes in the cornea [1]. Four corneal dystrophies, posterior polymorphous corneal dystrophy (PPCD), Fuchs endothelial corneal dystrophy, congenital hereditary endothelial dystrophy, and X-linked endothelial corneal dystrophy, affect the corneal endothelium and are collectively known as the endothelial corneal dystrophies.
PPCD is characterized by bands, vesicles, and gray opacities at the level of the corneal endothelium and is associated with corneal steepening [2]. Extracorneal manifestations such as glaucoma, keratoconus, and Alport syndrome are also associated with PPCD. At the cellular level, the hexagonal corneal endothelial cells exhibit changes in cellular morphology and size [3]. Additionally, the affected corneal endothelial cells exhibit epithelial cell-like characteristics such as stratification, desmosomes, and microvilli [4]. In some cases, these abnormal cells can affect the iridocorneal angle and trabecular meshwork, leading to glaucoma. PPCD also exhibits genetic locus heterogeneity with linkage demonstrated to two different genomic loci: PPCD1 (MIM ID #122000), associated with an unknown variant on chromosome 20p11.2-q11.2, and PPCD3 (MIM ID #609141), associated with truncating mutations in the zinc finger E box-binding homeobox 1 gene (ZEB1) gene on chromosome 10p11.22.
Multiple groups have reported PPCD1 families linked to a common region on chromosome 20, but the genetic basis for PPCD1 is still unknown (Fig 1). The first report for PPCD1 described genetic linkage to a locus on chromosome 20 between STS markers D20S98 and D20S108 [5]. Since then, multiple groups have reported other PPCD1 families, all showing linkage within the initial interval reported by Heon et al. [6–9]. All together, these studies suggest that the genetic basis for PPCD1 is found within the common support interval between D20S182 and D20S139 (approximately 3.6 cM or 1.8 Mb), which contains 32 genes according to the NCBI Annotation Release 105.
Relative position of previously reported intervals associated with PPCD1 are on the left of the ideogram. Relative position of the interval enriched for NGS in this study is on the right of the ideogram. Ideogram and genomic coordinates are based on the hg19 reference build. *The interval reported by Hosseini et al. refines the original interval reported by Heon et al.
Despite these reports that map PPCD1 to a common locus on chromosome 20, screening of the coding regions of multiple candidate genes within the initial PPCD1 interval and the PPCD1 common support interval have failed to identify a causal variant [6–8, 10–14]. Since screening of the exon and exon-intron boundaries of the genes in the common support interval, and other genes outside of the common region, have failed to identify the causal variant for PPCD1, we used a targeted next-generation sequencing (NGS) approach to screen variants within the linked region bordered by flanking markers D20S182-D20S195 in a previously reported PPCD1 family [9]. We previously published a limited study showing the utility of targeted NGS for PPCD1 but herein describe a substantially more comprehensive and robust NGS approach for the identification of the genetic basis of PPCD1 [13].
Materials and Methods
This study followed the Declaration of Helsinki and was approved by the Institutional Review Board at the University of California at Los Angeles (UCLA IRB# 94-07-243-(14-33A), 02-10-092-(4,11)). Written consent was obtained from all subjects in this study.
Subject selection and DNA collection
A total of 29 members from an affected family, previously mapped to the PPCD1 locus on chromosome 20, were enrolled in this study [9]. Clinical characterization was previously described. Genomic DNA was purified from peripheral blood leukocytes using the FlexiGene DNA Isolation Kit (Qiagen, Valencia, CA) or extracted from buccal epithelial cells using the Oragene Saliva Collection Kit (DNA Genotek, Ottawa, Canada) according to the manufacturer’s instructions.
Library preparation and next-generation sequencing
DNA from four affected and four unaffected members of the family previously mapped to the chromosome 20 locus were prepared for next-generation high-throughput sequencing at the UCLA Clinical Microarray Core. A DNA library was prepared using the Seqcap EZ Choice XL kit (Roche Diagnositics, Indianapolis, IN). In order to completely encompass the PPCD1 common support interval, a custom-designed oligonucleotide probe array (Roche NimbleGen Iceland LLc., Reykjavik, Iceland) was used to enrich for the previously linked chromosome 20 region in this family and an additional 500 Kbp 5’ of the linked interval (hg19: 17.3 Mbp– 31.8 Mbp). Highly repetitive regions such as the centromere were excluded from enrichment. High-throughput sequencing was performed on the Illumina HiSeq2000 platform (Illumina, Inc., San Diego, CA).
Variant calling bioinformatics pipelines
FASTQ files were downloaded from the UCLA Clinical Microarray Core and processed with two independent bioinformatics pipelines.
Burrows Wheeler Aligner/GATK HaplotypeCaller pipeline (BWA/GATK).
All reference files for this pipeline were obtained from the Broad Institute’s reference file directory (https://www.broadinstitute.org/gatk/download/) associated with UCSC’s hg19 human reference genome. Files were then processed and analyzed according to recommendations from the Genome Analysis Toolkit (GATK) Best Practices Pipeline [15–17]. FASTQ files were first aligned to the hg19 reference genome with the Burrows Wheeler Aligner (BWA, http://bio-bwa.sourceforge.net/) for paired-end reads [18]. After alignment, Picard Tools was used to convert files from SAM to BAM, sort by genomic coordinates, and mark optical duplicates (http://broadinstitute.github.io/picard/). Files were then realigned to known indels using GATK. Variant calling for single-nucleotide variants (SNVs) and insertions/deletion (indels) was then conducted with GATK’s HaplotypeCaller algorithm.
Variant filtering and annotation
Variant output files from both bioinformatics pipelines were analyzed using Partek® Genomics Suite® platform. The following criteria were utilized to filter candidate variants: (1) within the enriched interval (chr20: 17,316,434 bp– 26,319,280 bp and 29,420,138 bp– 31,826,081 bp); (2) read depth of ≥ 5; (3) quality score ≥ 20; (4) heterozygous, (allele frequency from 0.4–0.7); (5) not present in unrelated, unaffected controls; (6) present in affected family members and absent in unaffected family members; and (7) rare or novel (minor allele frequency (MAF) ≤ 0.05). MAFs were assigned using the Database of Single Nucleotide Polymorphisms (National Center for Biotechnology Information, National Library of Medicine; Bethesda, MD [dbSNP Build ID: 138]), and novel variants were defined as lacking a reference SNP cluster ID. After filtering, SNVs were then annotated with the Ensembl (v75) and RefSeq (compiled 02-02-2015 by Partek Inc.) annotation databases. SNVs were categorized as missense, nonsense, synonymous, 5’ or 3’ splice site, 5’ or 3’ untranslated region (UTR), promoter, intronic, or non-coding RNA (ncRNA). Indels were annotated with only the Ensembl (v75) database using Ensembl’s Variant Effect Predictor tool (VEP, http://grch37.ensembl.org/info/docs/tools/vep/index.html).
SNV validation and screening of additional family members
Sanger sequencing was performed to validate the filtered NGS-detected SNVs. Validated SNVs located in protein-coding genes were screened for in an additional 19 family members (seven affected and 12 unaffected). Novel SNVs in protein-coding genes were screened for in 100 ethnically matched controls. Sequencing primers are given in S1 Table.
In silico analysis of filtered promoter region SNVs and synonymous substitutions identified in protein-coding genes
Variants identified within the promoter regions of protein-coding genes were analyzed using JASPAR (http://jaspar.genereg.net/) in order to detect possible changes within transcription factor binding sites [19]. A threshold cutoff score of ≥ 75% was used. In addition, cryptic splice site prediction for synonymous SNVs was performed using MutPred Splice (http://mutdb.org/mutpredsplice/submit.htm) and NetGene2 (http://www.cbs.dtu.dk/services/NetGene2) [20, 21].
Determination of corneal endothelial expression of protein-coding genes in which validated, filtered, SNVs were identified
Transcript levels of OVOL2, CCM2L and THBD in the corneal endothelium were previously determined by RNA-seq, while the level of the encoded proteins was determined by fluorescence-immunohistochemistry (F-IHC) [22]. A cadaveric donor cornea from an unaffected individual and two corneas from individuals with PPCD without a ZEB1 mutation (non-PPCD3) obtained at time of surgery were fixed in 10% Tris-buffered formalin and subsequently paraffin embedded. Immunodetection was performed using a standard immunohistochemistry protocol with antibodies directed against OVOL2, CCM2L and THBD (S2 Table). In brief, sections were deparaffinized in Histo-Clear (National Diagnostics, Atlanta, GA) and rehydrated through a series of alcohols (100%, 95% and 80%) and water. Antigen retrieval was performed using proteinase-K digestion at 37°C for 15 minutes and sections were washed in PBST (PBS and 0.5% Tween 20). Non-specific epitope blocking was achieved by a 1 hour incubation with PBST supplemented with 1% bovine serum albumin and 10% normal serum. The sections were subsequently incubated overnight with each primary antibody diluted 1:500 (OVOL2) or 1:100 (CCM2L and THBD) in blocking buffer, followed by three washes in PBST. Incubation with a secondary antibody, Alexa Fluor 594 (Life Technologies, Carlsbad, CA), diluted 1:500 in blocking buffer was performed. After washing three times with PBST and one time with PBS, sections were mounted with Vectashield aqueous mounting medium containing 4′,6-diamidino-2-phenylindole (Vector Laboratories Inc., Burlingame, CA). To account for non-specific fluorescence, a control was performed using only the secondary antibody. Images were obtained using a fluorescence confocal microscope. Quantification of the fluorescence signal corresponding to each protein was performed using the Volocity 3D Image Analysis Software (PerkinElmer, Waltham, MA). Final fluorescence quantities were determined by subtracting non-specific fluorescence values, which were obtained by measuring the fluorescence in a field devoid of tissue in each image from the fluorescence in the endothelium of the secondary-only control.
Copy number variant analysis using high-resolution array comparative genomic hybridization
Copy number variation (CNV) analysis was performed using genomic DNA samples from the aforementioned four affected and four unaffected individuals that underwent NGS. The genomic DNA samples were submitted to the UCLA Clinical Microarray Core for array comparative genome hybridization (aCGH) using a custom Agilent 8x60K array (Agilent Technologies, Inc., Santa Clara, CA). Interrogation of a 16.7 Mbp region (hg19: 17.3 Mbp– 34.0 Mbp) encompassing the linked PPCD1 locus and approximately 2.7 Mbps (0.54 Mbp 5’ and 2.17 Mbp 3’) of sequence flanking the PPCD1 locus was performed using 52,828 oligonucleotide probes. This design resulted in a median probe spacing of 159 bp. Data analysis was performed using the Agilent CytoGenomics 3.0 software. The raw data files are available from the GEO DataSets database (accession number GSE72617; National Center for Biotechnology Information [NCBI], Bethesda, MD, USA). CNV validation was performed using agarose gel electrophoresis of PCR products amplified with primers flanking the putative CNV (S1 Table).
Results
Sequencing reads align to the enriched region
After performing alignment algorithms using two bioinformatics pipelines, NGS reads from one representative sample that underwent NGS were confirmed to align to the enriched region on chromosome 20 (Fig 2). Oligonucleotide probes for regions of low complexity (e.g., centromere) were avoided and therefore lack reads.
Histogram depicts the number of reads aligning to the PPCD1 candidate region on chromosome 20 for a representative individual (hg19 reference sequence; histogram produced using Partek® Genomics Suite®).
Comparison of two bioinformatics pipelines for the detection of SNVs
After variant filtering, a comparison of SNVs detected by the BT2/SAM and BWA/GATK pipelines was conducted to determine concordance between the two bioinformatics pipelines. The BT2/SAM pipeline detected a total of 839 SNVs and the BWA/GATK pipeline detected a total of 885 SNVs. Of these, 820 SNVs were concordant between both pipelines, while 19 SNVs were detected by only BT2/SAM and 65 SNVs were detected by only BWA/GATK.
Comparison of two gene annotation databases for the classification of identified SNVs
After performing variant filtering, a comparison of annotated SNVs was conducted to determine concordance between annotation databases for each bioinformatics pipeline. For the BT2/SAM pipeline, 386 SNVs were annotated by the RefSeq database, 441 SNVs were annotated by the Ensembl database, and 384 annotated variants were concordant between both annotation databases (Fig 3A). For the BWA/GATK pipeline, 409 SNVs were annotated by the RefSeq database, 469 SNVs were annotated by the Ensembl database, and 407 SNVs were concordant between both annotation databases (Fig 3B).
Comparison of functional annotations between the RefSeq (RS-Feb15) and the Ensembl (Enbl75) databases for SNVs detected by both bioinformatics pipelines. (A) Number of annotated SNVs detected by the BowTie2-SAMtools pipeline. (B) Number of annotated SNVs detected by the BWA-GATK HaplotypeCaller pipeline.
SNV and indel variant analysis of the PPCD1 interval
SNV analysis.
The majority of detected SNVs were intergenic SNVs (not annotated) or annotated as intronic or ncRNA SNVs (data not shown). A total of 12 SNVs located in 10 genes that were neither intronic nor ncRNA passed our filtering criteria (Table 1). Three of the 10 genes were protein-coding: OVOL2, THBD and CCM2L. Two of the 12 SNVs were coding region variants, resulting in synonymous substitutions in OVOL2 and CCM2L. Nine of the 12 SNVs were located in the promoter regions of eight genes, including two variants in the promoter region of OVOL2. The remaining SNV was located in the 3’ untranslated region (UTR) of THBD. Of the two SNVs that were novel, only one, NM_021220:c.-307T>C in OVOL2, affected a protein-coding transcript.
Indel analysis.
Since indel realignment is not available in Partek® Flow®, only indels detected by the BWA/GATK pipeline were analyzed to determine segregation with the affected phenotype. Due to difficulties in indel annotation in Partek® Genomics Suite®, Ensembl’s VEP tool was utilized to annotate indels. A total of 168 indels segregated with the affected status, with 159 indels annotated as intergenic or intronic. The remaining nine indels were located in seven genes, all of which were protein-coding (Table 2). Validation of these nine indels was not performed since the majority of the indels are located in regions of low complexity. Five of the nine indels were insertions located in intron 1 of CRNKL1 and the promoter of C20orf26 while the other four indels were located in the non-coding regions of five different genes. One of these four indels, which was mapped to the intronic region of HCK, was also mapped to the exonic region of a non-protein-coding transcript for HCK within the Ensembl database.
SNV validation and screening of additional family members
Of the 12 SNVs that survived the filtering criteria, 11 SNVs were confirmed by Sanger sequencing to be present in the four affected individuals who underwent NGS (the SNV in FAM182A (n.-452C>T) was not detected) and each was confirmed to be absent in the four unaffected individuals who underwent NGS (S1 Fig). Four of the 12 SNVs (OVOL2 c.327C>A; OVOL2 c.-307T>C; THBD c.351A>G; CCM2L c.1107G>A) were located in protein-coding genes. Nineteen additional family members (7 affected and 12 unaffected) who did not undergo NGS were screened for each of these four SNVs. Three of the four SNVs (OVOL2 c.327C>A, OVOL2 c.-307T>C, and THBD c.351A>G) segregated with the affected status of the additional family members. Although present in all affected family members, CCM2L c.1107G>A was also identified in one unaffected family member. OVOL2 c.-307T>C, the only novel SNV found within a protein-coding gene, was not found in 100 controls.
In silico analysis of promoter region SNVs and synonymous substitutions identified in protein-coding genes
Of the variants that passed the filtering criteria, OVOL2 c.-307T>C was the only variant found within the promoter region of a protein-coding gene. Thus, this variant was analyzed to determine whether it could cause any changes to the transcription factor binding sites within the OVOL2 promoter. According to JASPER, OVOL2 c.-307T>C is predicted to cause the formation of an additional FOXO3 enhancer transcription factor binding site (with a relative score of 77.8%) to the OVOL2 promoter.
Of the other variants that passed the filtering criteria, OVOL2 c.327C>A and CCM2L c.1107G>A were annotated as synonymous substitutions. Thus these variants were analyzed with MutPred Splice and NetGene2 to determine whether a cryptic splice site would be created. MutPred Splice predicted that the synonymous substitution in exon 3 of OVOL2 (c.327C>A) was a splice neutral variant with a score of 0.17. Additionally, NetGene2 did not predict that this synonymous substitution in OVOL2 would alter splicing, predicting the wild type splice acceptor and splice donor sites that flank exon 3 with confidence values of 1.00. MutPred Splice predicted that the synonymous substitution in exon 7 of CCM2L (c.1107G>C) was also a splice neutral variant with a score of 0.09 while NetGene2 predicted the creation of a splice acceptor site with a confidence of 0.19. The wild type splice acceptor site flanking exon 7 of CCM2L was predicted with a confidence value of 0.33, and the wild type splice donor site was predicted with a confidence value of 0.95.
Expression of OVOL2, CCM2L and THBD in PPCD corneal endothelium
A recent study using RNA-seq to profile the ex vivo human corneal endothelial cell transcriptome demonstrated transcript levels for OVOL2 (0.03 RPKM) and CCM2L (0.00 RPKM) at levels significantly below the background cutoff of 1 RPKM, while THBD (50.87 RPKM) was significantly above this cutoff [22]. Expression of the proteins encoded by OVOL2, CCM2L and THBD was investigated using F-IHC (Fig 4). In agreement with the transcript levels determined by RNA-seq, THBD was detected in the normal donor endothelium while OVOL2 and CCM2L were not. Similarly, one of the PPCD corneas did not show expression of OVOL2 or CCM2L, while the second PPCD cornea demonstrated increased OVOL2 (2.4 fluorescence units per pixel (FU/px)) and CCM2L (3.9 FU/px) expression compared with the normal donor cornea. THBD was detected in both normal donor (2.6 FU/px) and PPCD corneas (14.5 and 30.1 FU/px) with a marked increase in both PPCD corneas compared with normal donor cornea.
H&E: Hematoxylin and eosin stain (row 1). Primary antibodies directed against the proteins encoded by OVOL2 (row 2), CCM2L (row 3) and THBD (row 4) were used to detect protein expression in the corneal endothelium of a normal donor (column 1) and two PPCD corneas (columns 2 and 3). A secondary antibody conjugated to a fluorescent moiety (Alexa Fluor 594, red) was used to visualize the localization of the primary antibodies. The sections were counterstained with DAPI, which stained the nuclei blue. Numbers located at lower right corner of each panel represent the quantification of the fluorescent signal in fluorescence units per pixel (FU/px), corrected for autofluorescence.
Copy number variant analysis of the PPCD1 interval
CNV analysis was performed to identify a potentially pathogenic microdeletion or microinsertion in the PPCD1 interval. DNA samples from the four affected and the four unaffected individuals for whom NGS was performed were subjected to high-resolution aCGH. Fourteen CNVs (7 gains and 7 losses) were identified in at least one individual and ranged in size from 121 to 413,066 base pairs. A single CNV (318 bp loss) within an intron of OVOL2 was identified in all the affected individuals but none of the unaffected individuals. However, subsequent validation by PCR did not identify the 318 bp loss in the region predicted by aCGH (data not shown).
Discussion
In this study, we present, to the best of our knowledge, the first screening of the PPCD1 interval to identify potentially pathogenic coding and non-coding SNVs, indels and CNVs. NGS was used to sequence the entire linked region, which includes the PPCD1 common support interval, excluding regions of low complexity such as the centromere. With adequate coverage of the sequenced region, two independent bioinformatics pipelines were utilized for alignment and variant calling, and two independent annotation databases were utilized to annotate SNVs. Indels were analyzed and annotated using only the BWA/GATK pipeline and the Ensembl annotation database. After performing genetic filtering, a total of 11 validated candidate SNVs in nine genes and nine non-validated indels in seven genes were identified. Additionally, aCGH was used to interrogate CNVs in the linked region, although the single CNV identified within the PPCD1 interval proved not to be present after validation was performed. The most plausible explanation for the identification of a CNV that is not subsequently observed by PCR is a false positive result associated with the phenomenon of competitive hybridization observed with high-density array designs. In this case, the observation of the 318 bp loss only in the affected individuals may be due to the presence of a smaller genetic variation (i.e., SNV or small indel) in these individuals that results in an alteration in the competitive hybridization of the three involved probes.
No non-synonymous coding variants were detected in any of the annotated genes within the linked region, consistent with previous reports that failed to identify non-synonymous coding region variants in multiple genes within the PPCD1 locus. However, the recent association of another corneal dystrophy with a synonymous substitution in COL17A1 that creates a cryptic splice donor site, resulting in the loss of 18 amino acids, highlights the potential pathogenicity of synonymous substitutions [23, 24]. Therefore, we performed an in silico analysis of the synonymous substitutions identified in OVOL2 and CCM2L. The synonymous substitutions in OVOL2 was not predicted to create a cryptic splice site while one bioinformatics tool predicted that the synonymous substitution in CCM2L creates a relatively weak cryptic splice acceptor site in comparison to the wild type splice acceptor site. Thus, the exclusion of potentially pathogenic coding region SNVs and CNVs in the PPCD1 interval indicates that the genetic variant responsible for PPCD1 is likely in a non-coding region. Of particular interest, the novel SNV that we report in the promoter region of OVOL2 (c.-307T>C) results in the formation of a binding site motif for the transcription factor FOXO3, a promoter of gene transcription [25]. Given the pathogenic role of ZEB1 in PPCD3, and the fact that OVOL proteins are involved in the suppression of ZEB1 transcription, OVOL2 is a functional as well as positional candidate gene for PPCD1 [26–29]. To this end, the observation that OVOL2 protein was elevated in the corneal endothelium of one of the corneas from two individuals with PPCD provides some evidence to support this hypothesis. Although these corneas were obtained from individuals in whom Sanger sequencing of the exons in the ZEB1 gene did not reveal a pathogenic mutation, it may not necessarily be assumed that both of these individuals have PPCD1 as neither are members of a family linked to the PPCD1 interval. Thus, the failure to detect OVOL2 expression in the corneal endothelium of one individual does not exclude the possibility that OVOL2 may be ectopically expressed in the corneal endothelium in individuals with PPCD1.
Although intronic, ncRNA, and intergenic variants were not analyzed in this study, these variants may have potentially pathogenic effects. In particular, intronic variants may also lead to formation of cryptic splice sites and aberrant alternative splicing. Additional regulatory elements may also exist in the intronic and intergenic regions, which may cause abnormal expression of protein-coding genes. Additionally, studies show that ncRNA such as microRNA and long intergenic non-coding RNA, have potentially important functional roles in regulating gene expression [30]. Since the causative variant or gene has yet to be identified, we advocate identification and screening of non-coding variants of the PPCD1 locus and suggest that variants affecting the expression of OVOL2 may be causative of PPCD1.
Supporting Information
S1 Table. Primers for validation of detected variants.
https://doi.org/10.1371/journal.pone.0158467.s001
(DOCX)
S2 Table. Antibodies used for fluorescence immunohistochemistry.
https://doi.org/10.1371/journal.pone.0158467.s002
(DOCX)
S1 Fig. Validation of the 12 filtered SNVs detected by NGS. WT: wild-type sequence. MU: mutant sequence.
https://doi.org/10.1371/journal.pone.0158467.s003
(TIF)
Author Contributions
Conceived and designed the experiments: DJL DDC RFF AJA. Performed the experiments: DJL DDC RFF MJK AJA. Analyzed the data: DJL DDC RFF MJK AJA. Contributed reagents/materials/analysis tools: DJL AJA. Wrote the paper: DJL DDC RFF MJK AJA.
References
- 1. Weiss JS, Moller HU, Aldave AJ, Seitz B, Bredrup C, Kivela T, et al. IC3D classification of corneal dystrophies—edition 2. Cornea. 2015;34(2):117–59. pmid:25564336.
- 2. Aldave AJ, Ann LB, Frausto RF, Nguyen CK, Yu F, Raber IM. Classification of posterior polymorphous corneal dystrophy as a corneal ectatic disorder following confirmation of associated significant corneal steepening. JAMA ophthalmology. 2013;131(12):1583–90. pmid:24113819; PubMed Central PMCID: PMC3888803.
- 3. Laganowski HC, Sherrard ES, Muir MG. The posterior corneal surface in posterior polymorphous dystrophy: a specular microscopical study. Cornea. 1991;10(3):224–32. pmid:2055029.
- 4. Krachmer JH. Posterior polymorphous corneal dystrophy: a disease characterized by epithelial-like endothelial cells which influence management and prognosis. Transactions of the American Ophthalmological Society. 1985;83:413–75. pmid:3914130; PubMed Central PMCID: PMC1298709.
- 5. Heon E, Mathers WD, Alward WL, Weisenthal RW, Sunden SL, Fishbaugh JA, et al. Linkage of posterior polymorphous corneal dystrophy to 20q11. Human molecular genetics. 1995;4(3):485–8. pmid:7795607.
- 6. Hosseini SM, Herd S, Vincent AL, Heon E. Genetic analysis of chromosome 20-related posterior polymorphous corneal dystrophy: genetic heterogeneity and exclusion of three candidate genes. Molecular vision. 2008;14:71–80. pmid:18253095; PubMed Central PMCID: PMC2267740.
- 7. Gwilliam R, Liskova P, Filipec M, Kmoch S, Jirsova K, Huckle EJ, et al. Posterior polymorphous corneal dystrophy in Czech families maps to chromosome 20 and excludes the VSX1 gene. Investigative ophthalmology & visual science. 2005;46(12):4480–4. pmid:16303937.
- 8. Liskova P, Gwilliam R, Filipec M, Jirsova K, Reinstein Merjava S, Deloukas P, et al. High prevalence of posterior polymorphous corneal dystrophy in the Czech Republic; linkage disequilibrium mapping and dating an ancestral mutation. PloS one. 2012;7(9):e45495. pmid:23049806; PubMed Central PMCID: PMC3458081.
- 9. Yellore VS, Papp JC, Sobel E, Khan MA, Rayner SA, Farber DB, et al. Replication and refinement of linkage of posterior polymorphous corneal dystrophy to the posterior polymorphous corneal dystrophy 1 locus on chromosome 20. Genetics in medicine: official journal of the American College of Medical Genetics. 2007;9(4):228–34. pmid:17438387.
- 10. Aldave AJ, Yellore VS, Principe AH, Abedi G, Merrill K, Chalukya M, et al. Candidate gene screening for posterior polymorphous dystrophy. Cornea. 2005;24(2):151–5. pmid:15725882.
- 11. Aldave AJ, Yellore VS, Vo RC, Kamal KM, Rayner SA, Plaisier CL, et al. Exclusion of positional candidate gene coding region mutations in the common posterior polymorphous corneal dystrophy 1 candidate gene interval. Cornea. 2009;28(7):801–7. pmid:19574904; PubMed Central PMCID: PMC2714875.
- 12. Heon E, Greenberg A, Kopp KK, Rootman D, Vincent AL, Billingsley G, et al. VSX1: a gene for posterior polymorphous dystrophy and keratoconus. Human molecular genetics. 2002;11(9):1029–36. pmid:11978762.
- 13. Lai IN, Yellore VS, Rayner SA, D'Silva NC, Nguyen CK, Aldave AJ. The utility of next-generation sequencing in the evaluation of the posterior polymorphous corneal dystrophy 1 locus. Molecular vision. 2010;16:2829–38. pmid:21203404; PubMed Central PMCID: PMC3012649.
- 14. Yellore VS, Rayner SA, Emmert-Buck L, Tabin GC, Raber I, Hannush SB, et al. No pathogenic mutations identified in the COL8A2 gene or four positional candidate genes in patients with posterior polymorphous corneal dystrophy. Investigative ophthalmology & visual science. 2005;46(5):1599–603. pmid:15851557.
- 15. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature genetics. 2011;43(5):491–8. pmid:21478889; PubMed Central PMCID: PMC3083463.
- 16. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome research. 2010;20(9):1297–303. pmid:20644199; PubMed Central PMCID: PMC2928508.
- 17. Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, Del Angel G, Levy-Moonshine A, et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Current protocols in bioinformatics / editoral board, Baxevanis Andreas D [et al]. 2013;11(1110):11 0 1–0 33. PubMed PMID: 25431634; PubMed Central PMCID: PMC4243306.
- 18. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25(14):1754–60. pmid:19451168; PubMed Central PMCID: PMC2705234.
- 19. Mathelier A, Zhao X, Zhang AW, Parcy F, Worsley-Hunt R, Arenillas DJ, et al. JASPAR 2014: an extensively expanded and updated open-access database of transcription factor binding profiles. Nucleic acids research. 2014;42(Database issue):D142–7. pmid:24194598; PubMed Central PMCID: PMC3965086.
- 20. Hebsgaard SM, Korning PG, Tolstrup N, Engelbrecht J, Rouze P, Brunak S. Splice site prediction in Arabidopsis thaliana pre-mRNA by combining local and global sequence information. Nucleic acids research. 1996;24(17):3439–52. pmid:8811101; PubMed Central PMCID: PMC146109.
- 21. Mort M, Sterne-Weiler T, Li B, Ball EV, Cooper DN, Radivojac P, et al. MutPred Splice: machine learning-based prediction of exonic variants that disrupt splicing. Genome biology. 2014;15(1):R19. pmid:24451234; PubMed Central PMCID: PMC4054890.
- 22. Frausto RF, Le DJ, Aldave AJ. Transcriptomic Analysis of Cultured Corneal Endothelial Cells as a Validation for Their Use in Cell-Replacement Therapy. Cell Transplant. 2015. Epub Sep 2. pmid:26337789.
- 23. Jonsson F, Bystrom B, Davidson AE, Backman LJ, Kellgren TG, Tuft SJ, et al. Mutations in collagen, type XVII, alpha 1 (COL17A1) cause epithelial recurrent erosion dystrophy (ERED). Human mutation. 2015;36(4):463–73. pmid:25676728.
- 24. Oliver VF, van Bysterveldt KA, Cadzow M, Steger B, Romano V, Markie D, et al. A COL17A1 Splice-Altering Mutation Is Prevalent in Inherited Recurrent Corneal Erosions. Ophthalmology. 2016. pmid:26786512.
- 25. Paik JH, Kollipara R, Chu G, Ji H, Xiao Y, Ding Z, et al. FoxOs are lineage-restricted redundant tumor suppressors and regulate endothelial cell homeostasis. Cell. 2007;128(2):309–23. pmid:17254969; PubMed Central PMCID: PMC1855089.
- 26. Jia D, Jolly MK, Boareto M, Parsana P, Mooney SM, Pienta KJ, et al. OVOL guides the epithelial-hybrid-mesenchymal transition. Oncotarget. 2015;6(17):15436–48. pmid:25944618.
- 27. Roca H, Hernandez J, Weidner S, McEachin RC, Fuller D, Sud S, et al. Transcription factors OVOL1 and OVOL2 induce the mesenchymal to epithelial transition in human cancer. PloS one. 2013;8(10):e76773. pmid:24124593; PubMed Central PMCID: PMC3790720.
- 28. Lee B, Villarreal-Ponce A, Fallahi M, Ovadia J, Sun P, Yu QC, et al. Transcriptional mechanisms link epithelial plasticity to adhesion and differentiation of epidermal progenitor cells. Developmental cell. 2014;29(1):47–58. pmid:24735878; PubMed Central PMCID: PMC4153751.
- 29. Watanabe K, Villarreal-Ponce A, Sun P, Salmans ML, Fallahi M, Andersen B, et al. Mammary morphogenesis and regeneration require the inhibition of EMT at terminal end buds by Ovol2 transcriptional repressor. Developmental cell. 2014;29(1):59–74. pmid:24735879; PubMed Central PMCID: PMC4062651.
- 30. Cech TR, Steitz JA. The noncoding RNA revolution-trashing old rules to forge new ones. Cell. 2014;157(1):77–94. pmid:24679528.