Background Knobloch syndrome (KS) is a developmental disorder characterised by occipital skull defect, high myopia, and vitreo-retinal degeneration. Although genetic heterogeneity has been suspected, COL18A1 is the only known KS disease gene to date.
Objective To identify a novel genetic cause of KS in a cohort of Saudi KS patients enrolled in this study.
Methods When COL18A1 mutation was excluded, autozygosity mapping was combined with exome sequencing.
Results In one patient with first cousin parents, COL18A1 was excluded by both linkage and direct sequencing. By filtering variants generated on exome sequencing using runs of autozygosity in this simplex case, the study identified ADAMTS18 as the only gene carrying a homozygous protein altering mutation. It was also shown that Adamts18 is expressed in the lens and retina in the developing murine eye.
Conclusion The power of combining exome and autozygome analysis in the study of genetics of autosomal recessive disorders, even in simplex cases, has been demonstrated.
- cutis aplasia
- lens subluxation
- vitreoretinal degeneration
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- cutis aplasia
- lens subluxation
- vitreoretinal degeneration
Knobloch syndrome (KS, OMIM 267750) is an autosomal recessive developmental disorder of the eye and the occipital region of the skull.1 Typical eye abnormalities include high myopia, cataract, dislocated lens, vitreoretinal degeneration, and retinal detachment. The occipital skull defect typically presents as occipital encephalocoele, although milder forms such as isolated bony defect without brain protrusion or even occult cutis aplasia are known to occur. Although the eye abnormalities clearly reflect defective embryonic eye development, the occipital defect is a subject of debate in terms of whether it represents a neural tube closure defect or a bona fide bony defect of the skull.2 Additional clinical findings such as renal anomalies, lung hypoplasia and central nervous system malformation have also been reported, but only occasionally.2–5
Positional mapping led to the successful identification of COL18A1 as the KS disease gene and 17 mutations have been reported to date.2 4–9 Remarkably, virtually all mutations are truncating which raises interesting questions about what the phenotype of missense mutations might be. Despite the clear exclusion of COL18A1 in some KS patients as evidence of locus heterogeneity, no other disease genes or even loci have been identified; a second locus of KS was later retracted as erroneous.7 10 11 Identification of additional KS disease genes has the potential to reveal novel molecular mechanisms that control eye development and to determine the nature of the characteristic occipital defects.
Based on our experience with locus heterogeneity of several autosomal recessive diseases in our highly consanguineous population, we sought to identify novel genetic causes of KS by studying a local cohort of phenotypically well characterised KS patients. After excluding COL18A1 in one patient, we performed exome sequencing followed by filtration of the resulting variants based on her pattern of autozygosity. Using this approach, we identified a novel mutation in ADAMTS18 which we show to have a strong cytoplasmic expression in the developing lenticular and retinal tissues but not in the neural tube. Our data suggest that ADAMTS18 is the latest ADAMTS family member to be linked to a human eye disease.
Subjects and methods
KS was diagnosed based on the dual presence of one or both typical eye findings (lens subluxation and vitreoretinal degeneration) and an occipital defect of any degree (most severe form being encephalocoele and mildest form being cutis aplasia). All patients and, when available, parents and unaffected siblings were enrolled using an institutional review board approved written informed consent (KFSHRC RAC#207023). Blood was collected in EDTA tubes for DNA extraction.
All patients had their COL18A1 sequenced using specially designed primers followed by bidirectional sequencing (primer sequences and PCR conditions are available upon request). The coding region and flanking intronic sequences of ADAMTS18 were sequenced in patient F-6, and the mutation-containing fragment was also amplified in her relatives to check segregation and in 386 Saudi controls to rule out the possibility that it represents a rare polymorphism.
Autozygome and exome analysis
For one patient (F-6) with no identifiable COL18A1 mutation, we proceeded with autozygome analysis as described before.12 13 Briefly, genotyping was performed using Axiom platform (Affymetrix) and homozygosity mapping was run on autoSNPs to determine the extent and distribution of runs of homozygosity, with the assumption that these are autozygous (homozygous by descent) since parents were first cousins. In addition, we performed exome sequencing on patient F-6. Briefly, a DNA sample was randomly fragmented and adapters were ligated to both ends of the resulting fragments. This was followed by size selection, amplification, purification, hybridisation to the SureSelect Biotinylated RNA Library for enrichment and washing. The captured library was then loaded on HiSeq 2000 platform and run to get at least 30-fold coverage. The software SOAPsnp was used to assemble the consensus sequence and call genotypes in target regions (http://soap.genomics.org.cn/soapsnp.html).
Whole mount in situ hybridisation (WISH)
Whole mount RNA in situ hybridisation for Adamts18 was performed on E12.5 mouse embryos. Adamts18 probe corresponded to the area spanning c.3831-4777 (ENSMUST00000093113). SP6 and T7 tagged primers were used for generating, respectively, sense and antisense digoxygenin labelled RNA probes using the MaxiScript Kit (Ambion, Austin, Texas, USA) and Roche's DIG RNA Labelling Mix (Indianapolis, Indiana, USA). Embryos were permeabilised with proteinase K (10 μg/ml) at 37°C for 5 min, and in situ hybridisation was performed with the InsituProVSi (Intavis AG, Koeln, Germany) in accordance with a manufacturer recommended protocol.
Timed pregnant mice were set up to get E14.5 mouse embryos which were fixed in 4% paraformaldehyde (PFA) overnight followed by paraffin embedding and sectioning. The sections were blocked with blocking solution (1% goat serum and 0.1% triton in phosphate buffered saline (PBS)) for 2 h followed by one wash in PBS for 5 min followed by overnight incubation in a 1:50 dilution of anti-ADAMTS18 (Abgent, San Diego, CA, USA). This was followed by two washes in PBS, 5 min each, and 2 h incubation with rhodamine goat anti-rabbit antibody at 1:500 dilution. Finally, sections were washed with PBS and stained with DAPI-containing mounting medium (Vector Laboratories, Burlingame, CA, USA) and viewed on Eclipse 90i microscope (Nikon Instruments, Melville, NY, USA). Images were taken using the Cytovision V4.02 software (Applied Imaging Corp, San Jose, CA, USA). As a negative control, peptide competition assay was used where anti-ADAMTS18 antibody was preincubated with the peptide antigen at a 1:2 ratio (w/w) for 1 h at room temperature, before staining.
Clinical phenotype and COL18A1 analysis
In total, 13 patients representing six families were recruited for this study. All were Saudi in origin and consanguineous. Four families were multiplex and two had one affected member each. With the exception of patient F-6, all other patients harboured truncating homozygous mutations in COL18A1 (figure S1). Table 1 summarises the clinical and molecular features of all families but we will highlight the clinical phenotype of the only patient (F-6) in whom COL18A1 was excluded (see below). She is a 7-year-old girl, the third of four children born to healthy first cousin parents. Delivery was at full term and spontaneous vaginal. Birth weight was 2700 g. Occipital meningocoele was noted immediately after birth but brain CT was normal otherwise. She had successful surgical correction of the defect. Her cognitive and motor development has been normal. She presented at 18 months of age because parents were concerned about poor vision. Diagnosis of ectopia lentis, cataract and myopia was made and she was followed regularly in clinic. Over time, she developed progressive retinal degeneration and serous retinal detachment (figure 1).
ADAMTS18 defines a novel KS locus
Patient F-6 lacked any pathogenic COL18A1 mutation at the genomic level. She was therefore considered a good candidate for mapping a novel KS locus. Lack of runs of homozygosity (ROH) overlapping with COL18A1 further ruled out this locus. Her ROH analysis is summarised in figure 2 and table 2. In total, this patient had 242 Mb of autozygous DNA which equates to an inbreeding coefficient of 0.078. This is higher than 0.0625 expected for a product of a first cousin union but reflects background inbreeding, a very common observation in highly consanguineous populations. We used this autozygome data to filter the exome results we obtained for patient F-6 as follows. Exome analysis revealed the presence of 58 764 variants. By excluding previously reported variants, the list was narrowed to 10 110 variants. By limiting the analysis to variants that reside within ROH, the number was reduced to 238. By only focusing on non-synonymous, splicing, frameshift insertions and deletions, and nonsense variants, we decreased the number to 15 variants. Further filtration of the 15 variants by examining other Saudi exomes that we have generated, and by only considering the genes that are expressed in the eye, left us with ADAMTS18 as the only candidate with its non-synonymous homozygous missense mutation c.536C>T (p.Ser179Leu). Both parents were heterozygous for the change as was one unaffected sibling. The remaining two unaffected siblings were homozygous for the wild-type allele. This mutation was not seen in 386 Saudi control samples or in the Exome Variant Server (http://snp.gs.washington.edu/EVS/), is highly conserved across species, and is predicted to be disease causing with a high probability of 0.9048 according to MutationTaster (http://www.mutationtaster.org) (figure 3). PolyPhen predicted the variant to be possibly damaging with a score of 0.515.
In order to examine a potential role for ADAMTS18 in development, we studied the expression pattern of its murine orthologue in the developing mouse. A strong domain of expression was noted in the eye of E12.5 embryos, most notably in the lens, but no comparable strong expression was observed in the cranial part of the neural tube (figure 4). In E14.5, strong cytoplasmic expression was noted in the lens and retina which was abrogated by peptide competition (figure 4).
We have previously suggested the value of combining autozygome analysis with next generation sequencing and have recently demonstrated the utility of this approach in identifying the causative mutation in autosomal recessive disorders.12 Patients with autosomal recessive disease and consanguineous parents almost always have their causative mutation in an ROH. Thus, by focusing the search on these regions, one can filter out a significant percentage of variants generated by exome sequencing. In the setting of first cousin parents, this percentage is around 94%. Indeed, we show in this study the power of this approach by demonstrating that a single patient (patient F-6) was sufficient to uncover a novel genetic cause of KS.
ADAMTS (A Disintegrin-like And Metalloproteinase with ThromboSpondin), like ADAMS (A Disintegrin And Metalloproteinase), is a family of metalloproteinases but is distinct from ADAMS by the additional presence of thrombospondin motifs in the C-terminus and lack of transmembrane domain.14 So far, 19 ADAMTS members have been identified with highly variable tissue expression profile and a diverse portfolio of physiological functions. Unlike other metalloproteinases, ADAMTS members demonstrate a narrow substrate specificity due to the various exosites located in the C-terminal regions of the enzymes, which influence protein recognition and matrix localisation.15 ADAMTS-1 and -8 are inhibitors of angiogenesis, ADAMTS-2 and -3 function as N-propeptidases to process procollagens, and ADAMTS-4 and 5 are involved in cartilage remodelling in response to inflammation.16–18 Interestingly, mutations in ADAMTS10 and ADAMTS17 are known to cause a syndrome known as Weill–Merchasani syndrome which displays some overlapping eye phenotype with KS in the form of subluxated lens and cataract.19 Similar to ADAMTS18, the molecular pathogenesis of ADAMTS-10 and -17 linked eye disease remain to be elucidated. However, it seems likely that these proteins influence signalling through their action on matrix biology.20 The presumed loss of function that results from our ADAMTS18 mutation, therefore, could perturb this signalling although we acknowledge that experimental evidence will be required. We do show, however, strong expression of Adamts18 in the lens and retina during development which may be compatible with the proposed role in modifying the signalling environment required for the proper formation of these two eye tissues that are consistently malformed in KS patients.
In conclusion, we propose ADAMTS18 as a novel KS gene. Future research should focus on both refining the role of ADAMTS18 in eye development and identifying the other KS disease gene(s).
We thank the patients and their families for their enthusiastic participation. We also thank the Genomic and Sequencing Core Facilities for their technical help.
Funding This study was funded in part by KACST grants 08-MED497-20 and 09-MED941-20 (FSA) and a Dubai Harvard Foundation for Medical Research Collaborative Grant (FSA).
Competing interests None.
Patient consent Obtained.
Ethics approval This study was approved by KFSHRC IRB.
Provenance and peer review Not commissioned; externally peer reviewed.