Diagrams of the 6 pedigrees (A-F) that were investigated. Asterisk indicates examined and genotyped individuals. Unaffected status in some individuals who were not examined is based on information from other family members, and mild keratoconus therefore cannot be excluded. Perpendicular lines with a double horizontal line indicate consanguinity; diagonal lines with question mark indicate twins with unknown zygosity.
Plots of best logarithm of odds (LOD) scores obtained for a region over the position of the markers on the chromosome. A, Linkage between rs1074501 and rs755212 located at 14q24.3. B, Linkage around rs717822 at 6q26.
Liskova P, Hysi PG, Waseem N, Ebenezer ND, Bhattacharya SS, Tuft SJ. Evidence for Keratoconus Susceptibility Locus on Chromosome 14A Genome-wide Linkage Screen Using Single-Nucleotide Polymorphism Markers. Arch Ophthalmol. 2010;128(9):1191-1195. doi:10.1001/archophthalmol.2010.200
JANEY L.WIGGSMD, PhD
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
To search for genetic factors that could increase susceptibility to keratoconus.
A single-nucleotide polymorphism chip method was used to generate whole-genome data in a multiethnic panel of 6 families with 3 to 5 members affected by keratoconus. Candidate gene screening was performed by direct sequencing.
Linkage analysis results were strongest for a locus on chromosome 14q24.3. This region contains a relatively small number of genes of potential interest, including VSX2, a homeobox gene known to be involved in eye development and implicated in a spectrum of ocular disorders. However, sequencing the coding region of VSX2 did not reveal a sequence variant segregating with disease in any of the families described.
To our knowledge, this is the first report of linkage for keratoconus to 14q24.3 and the region is likely to harbor important inheritable genetic factors that may affect susceptibility to keratoconus.
Further genetic research is needed to identify the genes responsible for keratoconus. This knowledge will aid in understanding the molecular pathophysiology of this condition and may lead to improved treatment strategies.
Keratoconus is a noninflammatory disorder in which there is thinning and ectasia of the cornea. The estimated prevalence varies from 29 to 86 per 100 0001- 3; however, some studies suggest that the condition may be underreported.4 The onset of disease is typically after puberty, with subsequent progression at a variable rate over the following decades. Visual acuity is initially reduced by irregular corneal astigmatism but scarring can also develop. There are characteristic clinical signs in moderate and advanced disease,5 but despite the introduction of accurate computer-assisted videokeratography, the identification of mild keratoconus remains a challenge.6,7
Although extensive research has been undertaken, the etiology of keratoconus is still unknown. Some families have multiple affected members and genetic factors are thought to play a major role.1,5,8 The frequency of familial disease was reported to be between 6% and 8%5,8 but can be as high as 23.5% in some populations.9 Environmental factors such as repeated corneal trauma from eye rubbing or contact lens wear have also been associated with the development of keratoconus.5 However, in the majority of apparently sporadic cases, the relative contribution of genomic vs environmental factors toward the phenotype is unknown.8 Previous studies have shown that keratoconus is a genetically heterogeneous disorder and several potential loci have been identified. However, no susceptibility genes have been reported following linkage or association studies of nonsyndromic patients.10- 20
Because familial aggregation is present in a significant proportion of keratoconus cases,21 a family-based linkage study could identify potential genes and their sequence variants that contribute to disease susceptibility. Single-nucleotide polymorphism genomic microarrays (SNP chips) are a new development in the field. Because of their high-density marker coverage compared with the microsatellite panel, they can give a more accurate localization of the disease locus. We have therefore investigated genetic factors influencing keratoconus by genotyping 10 805 single-nucleotide polymorphism markers in 35 individuals from 6 families of different ethnic backgrounds.
The study was approved by an appropriate institutional review board and all participants signed informed consent. Patients with a family history of keratoconus were identified during routine examination at Moorfields Eye Hospital, London, England. Family members were then invited to participate in the study.
Affection status for the participating subjects was determined by clinical examination that included best-corrected visual acuity, slitlamp examination, videokeratography, and pachymetry using an Orbscan II operating with software version 3.12 (Bausch & Lomb Incorporated, Rochester, New York), which was performed in all eyes that had not been grafted. A diagnosis of keratoconus was considered to be clinically confirmed if 1 or more characteristic clinical signs on slitlamp biomicroscopy (central corneal thinning, corneal ectasia, Fleischer ring, Vogt striae, anterior stromal scar of the central cornea) were observed in at least 1 eye. In cases that had corneal surgery, there was clear documentation in the clinical record that there was keratoconus before the procedure. In the absence of slitlamp changes, the criteria used for keratoconus suspect were those described by Lim et al22: an area of relative inferior, central, or superior corneal steepening combined with an oblique cylinder of more than 1.5 diopters (D) on videokeratography and/or steep keratometric values more than 47 D and/or central corneal thickness less than 500 μm. The status of unaffected family members was confirmed using the same examination methods.
Venous blood samples (10 mL) were taken from available affected and unaffected family members. DNA was extracted using the Nucleon III BACC3 genomic DNA extraction kit (GE Healthcare, Chalfont St. Giles, England), according to the manufacturer's instructions.
Thirty-five subjects from 6 families from different ethnic and racial backgrounds were genotyped using the GeneChip Mapping 10k Array (Affymetrix, Santa Clara, California).
Pedigrees were checked for mendelian consistencies using PEDSTATS.23 Genotypes whose presence in related individuals would have required improbable recombination were identified using Merlin24 and these genotypes were then removed from the analysis. Linkage was assessed using LAMP,25,26 which uses a maximum likelihood model to generate genetic parameters that best suit the pedigrees and the known general epidemiologic parameters of the phenotype. Linkage results obtained through LAMP were validated using Merlin. For each locus, a number of parameters were permuted and different models of inheritance were considered until the parameters that fit the observed family structure best were applied for the analysis. We assumed a population prevalence for keratoconus in this multiethnic sample of 0.001.
The transcription factor visual system homeobox 2 gene (VSX2; OMIM *605020, also designated as CHX10 or HOX10), known to be involved in eye development and implicated in a spectrum of ocular disorders,27- 30 was selected as the best positional candidate gene and its coding region was sequenced using 5 primer pairs. Amplification was carried out in a 25-μL mixture containing 12.5 μL of ReddyMix PCR Master Mix (ABgene, Epsom, England), with 50 pmol of gene-specific primers and approximately 50 ng of genomic DNA. Primer sequences and PCR conditions are available on request. The PCR amplicons were purified and sequenced on both strands as previously described.31 Nucleotide sequences were compared with the published reference sequence (NCBI Entrez Gene NM_182894). Initially, 1 affected individual from each of the 6 families was screened. Where sequence alterations were identified, other available affected and unaffected family members were sequenced to test for segregation of these changes with keratoconus.
The 6 genotyped pedigrees with 3 to 5 members affected by keratoconus are shown in Figure 1 while their main characteristics are summarized in Table 1. All 21 individuals indicated as being affected had clinical signs of keratoconus in at least 1 eye evident at slitlamp examination. Where image capture was not precluded by extensive scarring, keratoconus was also confirmed by videokeratography and pachymetry. No individuals fulfilled the criteria of keratoconus suspect.
The Affymetrix GeneChip Mapping 10k Array has a total 10 805 markers. Of these, 1033 markers were excluded because of either excessive Hardy-Weinberg disequilibrium (P < .05) among founders, an inheritance pattern incompatible with mendelian laws (>1 per marker), or low genotyping from the chip output (<70% of all genotypes across the panel of participants). For the remaining 9772 markers, the average genotyping rate was 98.71%. The average distance between 2 consecutive markers was 0.35 cM (median, 0.13 cM/marker).
Using the criteria of Lander and Kruglyak32 on genetic linkage interpretation, we observed 1 suggestive and 1 significant logarithm of odds (LOD) score for 2 genomic regions (Figure 2 and Table 2).
The strongest linkage centered between marker rs1074501 and rs755212 located at 14q24.3 for which identical LOD scores of 3.58 were obtained assuming the same dominant model of inheritance (Table 2). Given that these 2 markers are only 2.8 Mb from each other, it is likely that they indicate the same linked gene. A susceptibility allele present at this locus in linkage with rs1074501 with a minor allele frequency predicted to be around 0.0001 would be responsible for about 20.77% of the familial keratoconus cases. In the interval between rs1074501 and rs755212, the power to detect linkage was insufficient because of low heterozygosity. The second locus on chromosome 6 also showed LOD scores more than 3. However, its contribution to the disease in our sample (1.5%) was remarkably small (Table 2).
To test for consistency of results, and to check for other potentially hidden signal peaks across the whole genome, we used Merlin to reanalyze the data under parametric models assuming additive and dominant models similar to that observed for the chromosome 14 locus. These analyses replicated previous findings identifying the same linkage peaks with the same markers and did not find additional linkage in any region beyond the one reported herein on chromosome 14. Finally, we checked other genomic regions where statistically significant linkage has been previously reported. We did not observe a maximum LOD score more than 2 in any of the loci known to be associated with keratoconus.10,12- 20
There are a limited number of transcripts coded for in the interval between rs1074501 and rs755212, 47 of which are annotated in the Ensembl genome browser (www.ensembl.org) as known protein coding genes. Sequencing the coding region of VSX2, the best candidate gene within the mapped region, did not reveal a sequence alteration segregating with disease in any of the families described.
A number of chromosomal regions have been linked to keratoconus using microsatellite genotyping.10,12- 17,19,20 The various loci reported may be the result of genetic differences between the populations. In this study, similar to the study by Burdon et al,18 we used a high-density whole-genome chip with the median distance between single-nucleotide polymorphism markers of 105 kb and report a potential linkage of keratoconus to a genomic region on 14q24.3. To the best of our knowledge, this is the first report of keratoconus linkage to this locus. Although there is uncertainty about the true prevalence of keratoconus, when we replicated our linkage under several scenarios differing only in disease prevalence all of these analyses produced identical results. In our panel, the strongest effect was detected on a locus whose likely mode of inheritance with respect to the phenotype was autosomal dominant. However, family F (Figure 1) shows an apparent recessive mode of transmission and therefore the genetic cause in this family is expected to be different than in the other pedigrees.
The locus reported herein on chromosome 14 contains a number of potential candidate genes. VSX2 maps to the middle of this interval. It is a paralog of VSX1, which is the only gene so far associated with keratoconus.11,33VSX2 is a member of the homeobox family with an essential role in retinal progenitor proliferation and bipolar cell differentiation.27 This gene is thought to be particularly important for eye development. Patients with VSX2 mutations have anophthalmia or microphthalmia associated with congenital cataract and iris abnormalities.28,29 Six individuals from 2 consanguineous families also had cloudy corneas with microphthalmia.30 Although we did not observe any sequence alterations in the VSX2 gene segregating with disease, we cannot fully exclude VSX2 as a candidate gene in these families because other mutational events, such as deletions, copy number variations, or mutations outside of the coding region, would not be detected by the methods described and may contribute to the phenotype.
The region of association on chromosome 14 was identified using a combination of genome-wide linkage analysis and a dominant mode of inheritance that best fit our expectations for a keratoconus-linked locus. A linkage to chromosome 14 with a maximum LOD score of 2.91 at 14q11.2 has previously been reported from an affected sibling pair association study of white families with keratoconus.17 Further linkage on chromosome 14 around q11.2 was reported by Bisceglia et al.19 Although these additional loci on chromosome 14 do not appear to be in the immediate vicinity of our locus, linkage is a low-resolution technique for mapping and in some cases a common locus of origin may be suspected even if there are nonoverlapping peaks. It is also possible that the linkage comes from 2 separate loci both having some effect on the development of disease with several adjacent genes regulating the same pathways. The other genetic locus we were able to link with keratoconus was on chromosome 6. Within the limits of the analytical power that our current sample size confers, we were unable to confirm a link between this locus and keratoconus under any other (eg, recessive) mode of inheritance.
There are a number of potential limitations to our study. First, the size of the sample is modest and any generalization of these results to larger populations requires caution. Second, maximizing analytical power by refining parameters to detect linkage proportionally increases the risk of overfitting. We accepted this trade-off reluctantly but judged that it could potentially contribute additional useful information about the disease. Third, our study assumes a certain degree of genetic homogeneity. This assumption may not be valid given the multiethnic nature of our panel. Fourth, replication of linkage results is essential to validate these findings. Finally, we concentrated our efforts in individuals with familial disease who should have a stronger genetic component than patients with apparently sporadic keratoconus. As a consequence, these results might have overestimated the magnitude of the effect of these loci in the general population with disease, where the majority of cases are sporadic, or they may even only be relevant to the population with familial disease. Despite the need for caution before linkage results are validated, our study identified a locus of potential interest on chromosome 14 that may help guide further studies on keratoconus.
Correspondence: Petra Liskova, PhD, Division of Genetics, Institute of Ophthalmology, University College London, 11-43 Bath St, London EC1V 9EL, England (firstname.lastname@example.org).
Submitted for Publication: June 3, 2009; final revision received March 11, 2010; accepted March 14, 2010.
Author Contributions: Drs Liskova and Hysi contributed equally to this work. Drs Liskova and Hysi had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This work was supported by the Special Trustees of Moorfields Eye Hospital and NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital London. Dr Liskova was also supported by the Czech Ministry of Education, Youth, and Sports research project 0021620806/20610011.
Additional Contributions: Quincy Prescott, MSc (Institute of Ophthalmology, University College London), provided technical support for the project and Kerra Pearce (UCL Genomics) assisted with sample processing. We thank Alison J. Hardcastle (Institute of Ophthalmology, University College London) for valuable comments on the study design.
This article was corrected online for typographical errors on 9/12/2010.