Is Keratoconus a Heritable or Genetic Disease?

There are numerous studies which support a role of heredity in the development of keratoconus. There is a strong familial predisposition in keratoconus development. A positive family history is reported by 6Ð10% of patients [3Ð5] or even as high as 23.5% in some populations [6]. The estimated prevalence of keratoconus in Þrstdegree relatives is 3.34% or 15Ð67 times higher than general population prevalence of 0.23Ð0.05% [7]. In most published studies, the inheritance pattern of keratoconus is autosomal dominant with incomplete penetrance or variable expressivity [4, 8Ð12]. Low expressivity forms of keratoconus, referred to as subclinical or Ôforme frusteÕ keratoconus, can be detected using corneal topography in the relatives of keratoconus patients [13, 14]. Studies in consanguineous populations strongly suggest the existence of recessive forms of keratoconus [15, 16]. Additionally, in a genetic modelling study in a multi-ethnicity population, a major recessive genetic defect was the most parsimonious genetic model [7]. X-linked inheritance has been reported rarely [17]. The role of heredity in disease development can be implied from twin studies, with a higher concordance rate between monozygotic versus dizygotic twins and non-twins, supportive of a genetic aetiology rather than environmental effects. Ideally, the zygosity should be conÞrmed using genetic typing. Most studies employing corneal topography support the concept of greater concordance between monozygotic twins and hence the role of heredity in keratoconus development [18Ð20]. Keratoconus commonly presents as an isolated sporadic condition but can be associated with a variety of single-gene disorders and chromosomal aneuploidies [3]. The increased prevalence of keratoconus in trisomy 21, 0.5Ð15% or 10Ð300 times the normal population prevalence, has implicated chromosome 21 as a positional candidate for the causative gene [21, 22].

Two approaches have been used to determine the genetic basis of keratoconus: candidate gene sequencing and genetic mapping. Candidate genes are identiÞed based on functional or biological information which makes them plausible agents in the disease pathogenesis or with genetic mapping and linkage analysis also known as positional cloning. Genetic mapping is a powerful technique as no assumptions are made about the causative gene and therefore genes of unknown function or deemed unlikely to be related to disease pathophysiology can be identiÞed.

Mutational Screening of Candidate Genes in Keratoconus

Visual System Homeobox Gene 1 (VSX1)

HŽon et al. [23] used linkage analysis to map a major gene for posterior polymorphous corneal dystrophy-1 (PPCD1) to chromosome 20p11-q11 and subsequently identiÞed mutations in the visual system homeobox gene 1 (VSX1) in PPCD1 and keratoconus. HŽon et al. [24] inferred a role for VSX1 in keratoconus pathogenesis as earlier case reports had documented the co-existence of PPCD and keratoconus. Following this original publication, there has been debate in the literature about the

signiÞcance of VSX1 mutations in keratoconus. Some authors have questioned the role of VSX1 in keratoconus because of a failure to detect VSX1 mutations in some keratoconus cohorts [25Ð27], coupled with the fact that the VSX1 knockout mouse did not have a corneal phenotype on histological analysis [28] and the initial report of VSX1 expression in the normal adult human cornea [29] could not be replicated in subsequent studies on normal and keratoconic corneas [24, 30, 31]. However, VSX1 expression has been detected in the murine cornea [32, 33] and was recently reported in the human neonatal cornea [34]. VSX1 has a role in corneal wound healing participating in the differentiation of corneal keratocytes into myoÞbroblasts [35] which may be relevant to the pathogenesis of keratoconus.

VSX1 belongs to the Ôpaired-likeÕ subfamily of homeodomain (HD) proteins. The homeodomain in this family is related to the homeodomain (HD) of the Drosophila ÔpairedÕ protein. VSX1 also contains a highly conserved CVC domain, essential for transcriptional regulation, which takes its name from the genes it was originally identiÞed in: mouse Chx10, goldÞsh Vsx1, and Caenorhabditis elegans Ceh-10 genes [29]. VSX1 is expressed in embryonic craniofacial, adult retina and the cornea. The Ôpaired-likeÕ homeodomain proteins have been implicated in craniofacial and ocular development [28, 29, 32, 36]. Previously, VSX1 was reported as a Þve-exon gene but two novel exons downstream of the original VSX1 gene sequence have been identiÞed. VSX1 is now known to consist of seven exons with a complex splicing pattern producing a total of six transcripts [34].

In order to determine pathogenicity in any gene-sequencing study, researchers must determine whether the detected sequence variants are deleterious and disease-causing, that is a mutation. Pathological sequence variants or mutations are not seen in ethnically matched controls (and if present, the variant is a polymorphism) and if the condition is familial, segregation of the variant should be seen within the family (affected family members carry the mutation and unaffected have a normal sequence). Pathogenicity can also be predicted based on amino acid conservation, biochemistry and structure using a range of computational tools. The ultimate test is to demonstrate a functional and pathological effect from the sequence variant in disease pathogenesis. HŽon et al. [24] reported a functional impact of the Arg166Trp VSX1 mutation associated with keratoconus on homeodomain binding.

A number of studies [24Ð27, 37Ð44] have reported the Þndings from VSX1 sequencing in keratoconus and identiÞed sequence variants as shown in Table 3.1. Some of these sequence variants were classiÞed as pathogenic mutations, but the data from subsequent studies or a re-assessment of the original data, shows some of these variants are polymorphisms. To date, there have been 12 mutational studies of VSX1 in keratoconus involving greater than 1,000 patients. A critical appraisal of all reported VSX1 mutations in keratoconus leaves six sequence variants which are pathological mutations given the available data (see Table 3.1): Leu17Pro, Asn151Ser, Gly160Asp, Gly160Val, Arg166Trp and Gln175His. Across all published studies, mutations in VSX1 are present in approximately 2Ð3% of keratoconus patients. VSX1 represents the only major genetic defect identiÞed to date causing keratoconus. The most commonly reported mutation was Gly160Val, although the

data was skewed by one study in which Gly160Val mutations were seen in 13/249 of sporadic Korean keratoconic patients [42].

From the available data, the pathogenicity and hence classiÞcation of Gly160Asp and Asp144Glu can be debated. Gly160Asp was seen in four unrelated keratoconus patients from Northern Europe and Italy in two independent studies and therefore accounts for 17.4% of VSX1 mutations in keratoconus. HŽon et al. [24] initially reported the segregation of Gly160Asp with posterior polymorphous corneal dystrophy and failed to detect this sequence variant in 277 control individuals, although did report that the glycine residue at position 160 was not highly conserved across species. Biscegali et al. [37] detected Gly160Asp in two families with keratoconus. In one family, segregation was not present as Gly160Asp was detected in one family member with keratoconus, but was only seen in 2/4 family members with topographically suspected keratoconus. In the second family, Gly160Asp was present as a compound heterozygote with Leu17Pro in a patient with keratoconus. The Leu17Pro VSX1 sequence variant was present in two other Italian families in this study. Segregation was demonstrated in one family; however, the patient from the second family was essentially a sporadic case. Leu17Pro was not seen in 125 control individuals and from the evidence appears pathogenic in its own right. Therefore, the role of Gly160Asp in the compound heterozygote state (Gly160Asp/ Leu17Pro) is difÞcult to ascertain, especially as two other family members had Leu17Pro alone and had evidence of topographically suspected keratoconus. Dash et al. [41] detected Gly160Asp in two sporadic keratoconic patients from Northern Europe which was not seen in 100 control individuals. This author reported that the Gly160 residue is not well conserved across species and was predicted to be a benign variant using bioinformatics modelling. Consequently, the pathogenicity of Gly160Asp can be debated. Gly160Asp is unlikely to represent a rare polymorphism as the glycine 160 amino acid is invariant in 433 controls (combined data across all keratoconus studies). Including the data from HŽon et al. [24] on this variant in PPCD, the glycine 160 residue is invariant in 710 controls. Gly160Asp may represent a genetic modiÞer as compound heterozygotes, as Gly160Asp in association with Leu17Pro [37] and Pro247Arg [24] had clinically more severe corneal phenotypes, suggesting it has an additive effect on keratoconus pathogenesis.

Asp144Glu was reported as a mutation in keratoconus in one study [37], but has been seen in controls [39, 40] and did not segregate in familial keratoconus in two studies [38, 41]. HŽon et al. [24] detected Asp144Glu in two family members affected with keratoconus and PPCD and in 1/90 glaucoma patients. The data in the literature suggests Asp144Glu is a polymorphism. Eran et al. [40] proposed Asp144Glu may confer a degree of susceptibility to keratoconus as the residue is highly conserved across species. Additionally, the author reported a higher frequency of the Asp144Glu substitution in keratoconus patients versus the control population. Similarly, Mok et al. [42] reported a higher frequency of an intronic variant (IVS1Ð11*a) in keratoconus patients versus controls increasing the risk of keratoconus. Some VSX1 sequence variants (Gly160Asp; Asp144Glu; IVS1Ð11*a) may not be directly pathogenic but confer a susceptibility to disease development.