Article Text
PDF
Abstract
Aim Retinopathy of prematurity (ROP) is a vasoproliferative eye disease in preterm infants. Based on its phenotypic similarities with familial exudative vitreo retinopathy (FEVR), the present study was conducted to screen the Norrin signalling pathway genes (already been implicated in FEVR) for understanding their involvement among Indian patients with ROP.
Methods The study cohort consisted of patients with ROP (n=246) and controls (n=300) that included full term (n=110) and preterm babies devoid of ROP (n=190). Screening of the NDP, FZD4, TSPAN12 genes were accomplished by resequencing the entire coding and untranslated regions (UTR). The genotype data of the patients with ROP were analysed in the background of their clinical manifestations and further analysed in conjunction with other available data on these genes worldwide.
Results Two novel variants in intron 1 (IVS1 +16A>G) and 3′UTR (c.5 22T>C) along with a previously reported change in the 5′UTR (c.395_409del14bp) were observed in the NDP gene in three patients with ROP. Screening of the FZD4 revealed four heterozygous variants, p.(Pro33Ser), p.(Pro168Ser), p.(Ile192Ile) and p.(Ile360Val), a compound heterozygous (p.(Pro33Ser)/p.(Pro168Ser)) and a 3′UTR (c*G>T) variants in the study cohort. Variants p.(Pro33Ser) and p.(Pro168Ser) were found to be significantly associated with ROP. A heterozygous variant p.(Leu119Arg) in TSPAN12 gene was observed in a patient with threshold ROP. However, a formal genotype–phenotype correlation could not be established due to the low frequencies of the variant alleles in these genes.
Conclusions This is a first study that revealed association of few variants in Norrin signalling genes among Indian patients with ROP that warrants further detailed investigation worldwide.
- ROP
- Premature births
- DNA
- genetics
Statistics from Altmetric.com
Introduction
Retinopathy of prematurity (ROP) is a proliferative retinal vascular disorder, characterised by abnormal retinal vascularisation in the presence of an incomplete retinal vascular development due to premature birth. It is a multifactorial disease contributed by various risk factors (gestational age (GA), birth weight (BW), early oxygen exposure and so on).1 2 ROP is a self-limiting disease, and only 15% of the preterm babies who develop ROP require treatment,3 while in others the disease regresses spontaneously. Severe ROP is mainly characterised by an abnormal blood vessel growth in the vitreous leading to vitreoretinal traction, retinal detachment and eventually blindness. Overall, the incidence of ROP in India varies from 30% to 48%.3 4 Hence, ROP remains a potential cause of avoidable childhood blindness in India.
Norrin β-catenin signalling plays an important role during the development of the fetal vasculature of the inner ear and retina.5 The genetic disruption of Norrin β-catenin signalling genes (NDP, FZD4, LRP5 and TSPAN12) has been implicated in many retinal vascular diseases like Coat’s disease, Norrie disease and familial exudative vitreoretinopathy (FEVR).5–9 The resemblance of clinical manifestations of ROP with FEVR further raised the possibility of involvement of Norrin β-catenin signalling genes in ROP pathogenesis. Few studies have looked for mutations in Norrin β-catenin signalling genes in ROP in different population worldwide.10–14 However, most of these studies targeted only one or two genes from the Norrin β-catenin signalling pathway and had quite conflicting results on the involvement of these genes in ROP pathogenesis.10 15 This warrant further investigations on the association of Norrin β-catenin signalling genes in ROP. Thus, the present study was conducted to screen NDP, FZD4 and TSPAN12 genes in patients with ROP from India in order to understand their relative contribution in ROP pathogenesis. We further tried to assess the genotype–phenotype correlation for the observed variants in order to understand their role in predicting the severity and visual outcomes of the disease.
Materials and methods
Enrolment of the study subjects
We adopted a case–control study design recruiting the preterm babies screened at a neonatal intensive care units and a tertiary eye care hospital in Hyderabad from the year 2008 to 2012. The study cohort included preterm babies with GA ≤35 weeks and/or BW ≤1700 g with ROP of any stage as cases and those with no ROP as controls. Additionally, ethnically matched full-term-born babies from the same geographical region with no history of any retinal disease were also recruited to serve as control. The medical and systemic history including race, GA, BW, early oxygen exposure, maternal health, other major complications, stage and zone of ROP, treatment details and pedigrees of the subjects recruited in the study were recorded on a predesigned proforma. Subjects with comorbid eye disease (secondary glaucoma, congenital cataract and so on) were excluded from the study. A prior informed consent was obtained from the parents of all the babies enrolled in this study. Diagnosis and severity of ROP was based on the amount of retinal vascularisation and the extent of each stage (I–V) of ROP as per the revised International Classification of Retinopathy of Prematurity16 guidelines using indirect ophthalmoscopy-based retinal drawings, done by fellowship trained ROP surgeons. This study was approved by the institutional review board of L.V. Prasad Eye Institute and adhered to tenets of the Declaration of Helsinki. Blood (0.3–1mL) samples were collected from the subjects in heparinised vacutainers (BD Biosciences, New Jersey, USA) by venipuncture for molecular genetic analysis.
Molecular genetic analysis
Isolation of genomic DNA from 0.3–1mL of peripheral blood of the enrolled subjects was done using MagNA Pure LC DNA isolation kit (Roche Applied Science, Indiana) on a MagNA Pure LC 2.0 System (Roche Applied Science). The DNA isolation from this machine is based on magnetic bead technology. DNA extraction protocol described in MagNA Pure LC DNA isolation kit manual was followed. The quality and quantity of DNA was measured using NanoVue plus (GE Life Sciences, New Jersey, USA). The integrity of DNA was further checked by running the DNA in 1% agarose gel. The primer sets used in this study for screening NDP, FZD4 and TSPAN12 gene were designed using primer designing tool (http://www.ncbi.nlm. nih.gov/tools/primerblast/), and the primers for 5′ untranslated region (UTR) and the coding regions of NDP genes were used as previously described17 (see online supplementary table 1). The specificity of primers was checked by primer blast. DNA of 50–100 ng was used for amplification of intron–exon boundaries, putative core promoter, 5′UTR region and the coding sequences of NDP, FZD4 and TSPAN12 genes using specific primers by PCR. The amplified PCR products were screened for detection of mutations/variants by sequencing on automated DNA sequencer (ABI 3130xl, Applied Biosystems (ABI), Foster City, California, USA). The observed variants in these genes were further validated by resequencing and further assessed for their association with disease susceptibility. Bioinformatics software like ClustalW Omega, SIFT,18 PolyPhen19 and Mutation Taster20 were used to predict the deleterious effect of the observed variants in these candidate genes.
Supplemental material
Statistical analysis
Allele frequencies of the polymorphisms identified in each of the genes screened were calculated by allele counting method. The risk conferred by the genetic polymorphisms towards ROP was assessed by calculating ORs and p value using χ2 test. The test of significance was set at p<0.05, and the CI was set at 95%. The allele frequencies of the observed variants in the present study were compared with the frequencies provided in Exome Aggregation Consortium (ExAC) database.21
Results
Of all the preterm babies screened for ROP (approx. 500 per year) by our team of ophthalmologists and nurses from the year 2008–2012, a total of 436 (including 246 ROP, 190 no ROP) were recruited in the study. Additionally, another 110 full-term normal controls who were recruited for a genetic study on FEVR were also screened for the observed variants.7 Demographics of the study participants are provided in online supplementary table 2. Babies with ROP had significantly lower mean BW (p=3.2×10−19) and mean GA (p=2.3×10−19) as compared with babies with no ROP. A total of 13 variants were observed in our ROP cohort after screening of entire coding and flanking region of NDP, FZD4 and TSPAN12 genes.
Supplemental material
The NDP gene screening in 78 ROP probands and 82 no ROP preterm controls revealed a 14 bp deletion in 5′UTR, an intronic variant and 3′UTR variant in three sporadic cases of ROP. The schematic locations of these variants are provided in figure 1. The 14 bp deletion (395_409del14bp) was detected in single patient and none of the controls, while the rest of the two variants were observed in both patients and controls (table 1).
Schematics diagram of (A) NDP, (B) FZD4 and (C) TSPAN12 showing localisation of observed variants within a gene. Variants identified in the present study in NDP, FZD4 and TSPAN12 genes are indicated by arrowhead (red arrow indicates position of amino acid change) in coding region and (black arrow indicates change at cDNA position. UTR, untranslated region.
Associations of variants/polymorphisms in the NDP, FZD4 and TSPAN12 gene with ROP
Screening of the exons and 3′UTR regions of FZD4 gene in 90 patients with ROP and 70 no ROP preterm controls revealed three non-synonymous c.97C>T (p.(Pro33Ser)), c.502C>T (p.(Pro168Ser)), c.1078A>G (p.(Ile360Val)) and a synonymous missense c.576 C>T (p.(Ile192Ile)) variants. Of the three observed non-synonymous variants, (p.(Pro33Ser) and p.(Ile360Val)) were detected only in patients while p.(Pro168Ser) was found in both patients and preterm controls (table 1). The schematic locations of these variants are provided in figure 1. The p.(Ile360Val) variant involving a highly conserved residue across the species was predicted to be pathogenic by PolyPhen-2 and Mutation Taster (figure 2). Likewise, both p.(Pro33Ser) (p=0.029) and p.(Pro168Ser) (p=0.003) were found to be significantly associated with ROP (table 1). While p.(Pro168Ser) was conserved across all the species and p.(Pro33Ser) was conserved in all species except opossum, chick and python (figure 2). The observed synonymous variant p.(Ile192Ile) though predicted to be disease causing by Mutation Taster was not found associated with ROP.
Multiple sequence alignment showing the conservation of wild-type residue with respect to P33S and I360V variants in FZD4 protein and L119R variant in TSPAN12 protein across different species: alignments were generated with the Clustal Omega software. The position of the variants is indicated by an arrow. Accession numbers (NCBI) and name of the species are mentioned at the left side of the picture. NCBI, National Center for Biotechnology Information.
The eight exons of TSPAN12 gene including intron–exon boundaries and 3′UTR regions were screened in 200 ROP probands and 147 no ROP preterm controls. Two coding region substitutions (p.(Leu119Arg) and p.(Pro255Pro)) and three UTR variants (c.*39C>T, c.*334A>T and c.*1243A>T) were observed in 65 patients with ROP. Schematic locations of these variants are provided in figure 1. On multiple sequence alignment, p.(Leu119Arg) in exon 5 of TSPAN12 gene was found to be an evolutionarily conserved residue in closely related species like human and macaque but not in mouse, dog, buffalo and so on (figure 2). This variant p.(Leu119Arg) was observed only in a proband and not in controls, while other synonymous variants p.(Pro255Pro) was present in both patients and controls with almost equal allele frequency. A single nucleotide polymorphism (SNP) (c.*39C>T) was observed at the 3′UTR, 39 base pair after the stop codon without any effect on splice site. None of these variants were found to be significantly associated with ROP (table 1).
Discussion
The overall incidence of ROP in southern India alone is almost 30%,3 which is very high, and this is expected to increase exponentially in the near future. Recently, it was observed that in India and in other middle-income countries,1 2 even the preterm babies with high BW and GA also tend to develop severe ROP. Varying incidence of ROP across different countries and the presentation of disease even in babies with relatively high GA and BW further suggested for the role of either genetics or environmental risk factors or their interactions in the disease progression.
Genetic variations in Norrin β-catenin signalling genes in ROP
The screening of candidate genes namely NDP, FZD4 and TSPAN12 led to the identification of six major coding region variants in 36 ROP probands. Of these, two of them, namely I360V in FZD4 and L119R in TSPAN12 gene (as reported in dbSNP databases22 with a minor allele frequency of 0.00007and 0.00005 in ExAC database21), were observed only in two ROP probands in our study cohort; therefore, they seem more likely to be the rare variants. Patients harbouring these variants had clinical features of advanced ROP like leukocoria, retinal haemorrhages and partial or total retinal detachments. Two polymorphisms—p.(Pro33Ser) and p.(Pro168Ser) in FZD4 gene—were found to be significantly associated with ROP (table 2).
Clinical features of ROP patients at presentation with variants in NDP, FZD4 and TSPAN12 genes
NDP genes variations in ROP
Mutations at the 5′ and 3′UTRs in NDP gene in patients with ROP have been suggested to alter the regulation of gene expression and messenger RNA stability thereby resulting in the reduction of Norrin β-catenin signalling and eventually poor vascular development.23 NDP gene mutations mostly have accounted for advanced ROP cases, and their frequencies varied across different population (table 3). In the present study, variants identified at the UTR region of NDP gene accounted for 2.7% (4/78) of the patients with ROP, which was comparable with previous studies (table 3).12 13 Furthermore, it was interesting to note that unlike FEVR, variants in NDP gene observed in ROP were exclusively restricted to the UTRs of the gene.7 In an earlier study, the whole gene screening of NDP by a direct sequencing method revealed a higher frequency of 597C>A variant in advanced ROP as compared with regressed ROP and no-ROP preterm controls from Kuwait14 (table 3). However, we did not observe any such common variant for the advanced ROP cases in our study. Likewise, the 14 bp deletion (395_409del14bp) predicted to be disease causing was observed in advanced ROP cases13; however, in the present study, the patient harbouring this change along with 3′UTR (c.*522T>C) had regressed ROP. Even in the previous studies on FEVR, similar 3′UTR variant was reported in mild cases7 thereby suggesting that 3′UTR change might be conferring protection to ROP progression. In a previous report, NDP was found to be one of the major genes involved in Indian FEVR cases7; however, it seemed to have a relatively minor role in ROP in the present study cohort.
Frequencies of NDP, FZD4 and TSPAN12 gene variants in patients with retinopathy of prematurity (ROP) in different ethnic groups
FZD4 gene variations in ROP
Several variations have been reported for severe ROP cases across different populations in FZD4, which is one of the key genes of Norrin β-catenin signalling pathway.10 15 24 25 Four heterozygous coding region variants including p.(Pro33Ser), p.(Pro168Ser), p.(Ile192Ile) and p.(Ile360Val) and a compound heterozygous change (p.(Pro33Ser)/p.(Pro168Ser)) were identified in 17.7% (16/90) of the patients with ROP in the present study. Both p.(Pro33Ser) and p.(Pro168Ser) were found to be significantly associated with ROP. Patients harbouring either of these variants exhibited classical clinical features of ROP and had poor outcome (table 2). Proline is known to exist in the turns of beta pleated sheets and provides rigidity, which is essential for proper conformation of protein. The in silico analysis predicted p.(Pro33Ser) to result in impaired stability of FZD4 protein causing its mistranslocation, which might lead to impaired Norrin β-catenin signalling further resulting in avascular retina.15 Interestingly, only these two variants from the present study were shared between patients from India and USA, while others were not found to be shared across different populations (table 3).
TSPAN12 gene variations in ROP
Several reports have demonstrated the important role of TSPAN12 gene mutations in FEVR pathogenesis. In the present study, we have identified p.(Leu119Arg) variants in only patients with ROP, while the other four reported SNPs c.*334A>T, c.765G>T, c.*39C>T and c.*1243A>T were observed in both patients with ROP and controls. None of the variants in TSPAN12 gene were pathogenic, and no significant differences in the frequency of these SNPs were observed among ROP patients and controls. Thus, the role of TSPAN12 gene in ROP needs to be investigated further across different study cohorts to establish its potential role in ROP.
Majority of the variations associated with severe ROP in NDP, FZD4 and TSPAN12 genes from the previous and present studies were either reported SNPs or non-pathogenic variations.10 15 24 Across all the studies published on Norrin β-catenin signalling genes in ROP, only very few pathogenic variants have been observed. Moreover, none of the pathogenic variants found in NDP, FZD4 and TSPAN12 genes for FEVR cases7 8 were detected in the patients with ROP, providing further evidence of locus heterogeneity across these two different phenotypes. Thus, we conclude that while patients with FEVR are born with specific genetic defects, ROP is much more a complex condition, where rare variants and SNPs might play an important role along with other prenatal and postnatal risk factors.
In conclusion, this study assessed the involvement of variants in NDP, TSPAN12 and FZD4 gene in the pathogenesis of ROP in Indian cohort. However, due to limited quantity of DNA concentration available for the experiments from a very low volume of blood from the preterm infants, different number of cases were screened for different genes. Additionally, for the above-mentioned reasons, some additional cases with genetic changes in the studied genes and other potential gene such as LRP5 might have been missed in our cohort. Despite of the challenges observed mainly in getting sufficient samples from preterm tiny infants, the results obtained in this study suggested that unlike FEVR, a Mendelian disorder, ROP seems to be a complex disease that is genetically more heterogeneous with multiple alleles of varying magnitudes of effect. Furthermore, screening of the rest of the Norrin β-catenin signalling pathway genes might expand the information on the involvement of this pathway in ROP pathogenesis. ROP, being a multifactorial disease, involvement of some putative genes apart from Norrin β-catenin signalling genes could not be ruled out. Exome sequencing or whole-genome sequencing could aid in identifying such novel genes in ROP.
Acknowledgments
The authors would like to thank Dr Subhabrata Chakrabarti for his kind help in data analysis and other valuable suggestions for the study.
References
Footnotes
Contributors Inderjeet Kaur and Subhadra Jalali conceived the idea and wrote the protocol; Inderjeet Kaur served as principal investigator; Subhadra Jalali, Divya Balakrishnan and Anjli Hussain were coinvestigators, performed clinical examinations and graded the fundus images and surgeries for the preterm; Sonika Rathi, Satish Patnaik and Ganeswara Rao Musada collected blood samples and family history of the probands; Sonika Rathi performed most of the molecular biology-based analysis of blood; Satish Patnaik and Ganeswara Rao Musada performed a part of molecular screening; Sonika Rathi and Inderjeet Kaur analysed the data and wrote the manuscript; and all authors revised the paper and approved the submitted version.
Funding This work was supported in parts by a grant from the Program Support grant by Department of Biotechnology (BT/01/COE/06/02/10 and BT/PR3992/MED/97/31/2011) and Champaulimaud Foundation grant Portugal to Inderjeet Kaur and Indian Council of Medical Research fellowship to Sonika Rathi.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study protocol was approved by the Institutional Review Board (LEC029149029) of the L.V. Prasad Eye Institute (LVPEI).
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Study data is available for research use from the corresponding author and principal investigator of the study upon request.
Linked Articles
- At a glance