Ethics statement

This study was approved by the National Institute of Biomedical Genomics Institutional Animal Ethics Committee (NIBMG-IAEC) for all zebrafish related experiments under the certification number NIBMGZF/002/2022-23. NIBMG Zebrafish Facility is registered with the Committee for Control and Supervision of Experiments on Animals in India with a registration number 2194/GO/ReBi/S/22/CPCSEA. No sample from human participants were used in this study.

Conditional analysis of CNTNAP5 locus

We selected 13 Single Nucleotide Polymorphisms (SNPs) for analysis, each characterized by p-values, odds ratios (OR), and 95% confidence intervals (CI) from our previous GWAS on PACG (S1 Table). To determine if the association signal for glaucomatous neurodegeneration previously identified through GWAS points to CNTNAP5 as the likely causal gene within the chromosome 2 locus, we performed conditional and joint analysis using summary statistics of SNP genotype-phenotype associations across the region. This was carried out using version 1.93 of genome-wide complex trait analysis (GCTA) software[14]. LD-independent genome-wide significant SNPs within ±500kb window flanking CNTNAP5 transcription start site were selected from the broader chromosome 2 locus showing evidence of association in the primary GWAS discovery analysis. The GCTA stepwise model selection (—cojo-slct algorithm) was applied conditioning on the top sentinel SNP across iterations to refine association signals, pinpoint candidate independent causal variants, and infer local genes mapping to said variants statistically accounting for most of the GWAS signal.

Analysis of Cis-Regulatory Elements (CRE) annotations at CNTNAP5 locus

To investigate the non-coding regulatory landscape surrounding prioritized candidate gene CNTNAP5 for follow-up functional study, we searched for annotated CREs within the broader ±500 kb flanking region using the SCREEN DNA Elements Explorer platform [15]. We specifically queried genomic coordinates chr2:125,083,095–125,084,384 spanning the CNTNAP5 locus for reported CREs characterized by empirical DNaseI, CTCF, H3K4me3, and H3K27ac signals. Data on genomic location, biochemical epigenetic readouts, normal cell/tissue source identifiers and other annotation metadata were systematically compiled for downstream analysis in R software (R, version 4.1.2) environment as described in Chakraborty et al, 2023 [16].

Analysis of CNTNAP5 SNPs in DeepSEA database and RegulomeDB

We selected 13 SNPs for analysis, each characterized by p-values, odds ratios (OR), and 95% confidence intervals (CI) from our previous GWAS on PACG. These SNPs were prioritized based on their significant associations with the risk of PACG.We input genomic coordinates and reference/alternate alleles for the 13 CNTNAP5 intronic and 3’ UTR SNPs into the DeepSEA web server (http://deepsea.princeton.edu). Regulatory impacts of SNPs disrupting or creating binding sites and composite functional significance scores were compiled from DeepSEA multi-task predictions exploring the known and predicted regulatory potential of PACG-associated non-coding variants annotated to the CNTNAP5 gene provides additional support linking the genetic association signals observed through orthogonal bioinformatics evidence [17]. To assess potential genetic regulation of CNTNAP5, we examined 13 identified SNPs in RegulomeDB. This database integrates various regulatory information including expression quantitative trait loci (eQTLs) and chromatin immunoprecipitation (ChIP-seq) data to predict regulatory potential of variants. SNPs were input and the resulting RegulomeDB scores were recorded, with lower scores indicating increased evidence for regulatory function.

Expression quantitative trait loci (eQTL) Analysis

Further to elucidate the potential impact of 13 selected SNPs within the CNTNAP5 gene on its expression levels in the retina, eQTL analysis was conducted to assess the association between genetic variants (SNPs) and gene expression levels in retinal samples. The average expression of the CNTNAP5 gene, measured in Counts Per Million (CPM), was derived from transcriptomic datasets of Age-related Macular Degeneration (AMD) retinal samples. The analysis utilized QTLtools, which computes nominal p-values to identify potential eQTLs. The identified eQTLs were integrated with methylation QTLs (mQTLs) to better understand the regulatory impact of these SNPs on CNTNAP5 expression [18].

Analysis of ocular tissue expression patterns of CNTNAP5

To evaluate the expression patterns of CNTNAP5 across ocular tissue types, we obtained RNA sequencing datasets from the eyeIntegration v2.12 (https://eyeintegration.nei.nih.gov/) and human eye transcriptomic atlas(https://www.eye-transcriptome.com/) encompassing tissues from the cornea, iris, lens, ciliary body, retinal pigment epithelium, choroid, central and peripheral retina.

Exploring chromatin topology at CNTNAP5 locus using 3D, 3DIV and HUGIn genome browser

To investigate potential long-range chromatin interactions connecting distal regulatory elements to prioritized candidate gene CNTNAP5, we utilized the 3D Genome Browser (http://3dgenome.org). We searched 3D Genome Browser encompassing the full CNTNAP5 gene body along with ±500kb flanking regions. Interactive conformation plots centred on the CNTNAP5 Transcription Start Site (TSS) were generated to identify significant long-range intra-TAD chromatin loop anchors that may harbour distal regulatory elements. By utilizing overlays with tissue-specific epigenomic annotations, this investigation of 3D genomic architecture maps surrounding CNTNAP5 facilitates the identification of potential non-coding SNPs identified by GWAS signals that may have an impact on CNTNAP5 expression [19]. Additionally, we queried the same 13 SNPs of CNTNAP5 in the 3DIV database [20] to retrieve a list of statistically significant distal chromatin interaction partners and their genomic annotations. The distance normalized interaction frequencies and bias removed interaction frequencies were noted for each SNP. Further, we utilized the HUGIn browser [21] to visualize and compare the observed versus expected read counts between the CNTNAP5 gene locus and 13 SNPs of interest with distal chromatin regions in specifically in in the dorsolateral prefrontal cortex (DPLC) and neuronal progenitor cells (NPC). The raw observed and expected read counts were extracted for each SNP for descriptive statistical analysis. By leveraging these publicly available Hi-C data resources, we obtained processed long-range chromatin interaction data between our loci of interest and distal regulatory elements across a human neural cell and tissue types. The normalized interaction frequencies and raw read counts were utilized to characterize and compare the chromatin architecture surrounding the CNTNAP5 and its 13 SNPs under study.

Gene Co-expression analysis

To identify genes co-expressed with CNTNAP5, we utilized two online databases—STRING (https://string-db.org/) and GeneMANIA (https://genemania.org/). In STRING, we input CNTNAP5 and selected Homo sapiens as the organism. The analysis was run with a high confidence level (0.700). In GeneMANIA, we similarly input CNTNAP5 with Homo sapiens selected. The co-expression analysis was run with automatic weighting. The resulting list of co-expressed genes from both databases were compiled.

Pathway enrichment analysis

The combined list of CNTNAP5 co-expressed genes was used as a input into Enrichr[22] for pathway enrichment analysis. The analysis was run against the KEGG and Reactome pathway databases with the default settings. Significantly enriched pathways were identified using an adjusted p-value cutoff of 0.05.

Analysis of CNTNAP5 variant-phenotype associations using HumanBase

To examine CNTNAP5, the gene was searched in the HumanBase database(https://hb.flatironinstitute.org/). The algorithm returns confidence scores predicting the likelihood of association between the queried gene and various diseases. These confidence scores are derived from the tissue-specific gene networks and analysis of known disease ontology terms.

Analysis of CNTNAP5 variant-phenotype associations using PheWAS

We queried the database of Phenotype-Wide Association Studies (PheWAS) catalog (phewascatalog.org) [23]. This aggregator tool combines published genotype-phenotype association statistics across a diversity of disease states profiled in large genomic datasets like the UK Biobank. We searched for all SNPs and small insertions/deletions mapped to within the genomic co-ordinates of CNTNAP5 on chromosome 2 (chr2:125,083,095–125,084,384). Statistical association data linked to clinical diagnosis billing codes (ICD-10 classifications) were compiled across available PheWAS studies for variants catalogued within this CNTNAP5 locus. Significantly associated phenotypes and disorders (at FDR or Bonferroni adjusted p-values < 0.05 threshold) were systematically tabulated and categorized by broad human diseases.

Dual luciferase reporter assay

To determine our PACG associated prioritized SNP rs2553628 located downstream of the CNTNAP5 locus exerts regulatory effects on enhancer activity; we carried out dual luciferase reporter assay. A 245 bp region flanking the rs2553628 variant was PCR amplified from PACG patient genomic DNA and directionally cloned into the multiple cloning site upstream of the firefly luciferase cassette using KpnI and XhoI restriction enzymes (NEB) in the pGL4.20 luciferase construct driven by an SV40 promoter (Promega). Patient and control subjects’ DNA were used to generate the risk and reference allelic reporter constructs. HEK293 cells were cultured in DMEM (Gibco) supplemented with 10% FBS (Gibco) and seeded into 24-well plates at a density of 5x10⁴ cells per well. Cells were co-transfected after 24 hours with 50 ng of the CNTNAP5 luciferase reporter constructs along with 10 ng of pRL-TK Renilla luciferase internal control using FuGENE HD transfection reagent (Promega). At 48 hours post-transfection, cell lysates were collected and Firefly and Renilla luciferase signals measured using the Dual-Luciferase Reporter Assay System (Promega) as per manufacturer’s protocol on Tecan infinite 200 PRO (Tecan). Normalized promoter activity was calculated as the ratio of Firefly to Renilla luminescence. Differences in normalized luciferase ratios between SNP alleles across three biological replicates were statistically compared by Student’s t-test.

Zebrafish husbandry and maintenance

Zebrafish (Danio rerio) were maintained in recirculating Tecniplast benchtop rack systems under standard laboratory conditions on a 14-hour light/10-hour dark cycle. Adult fish were set up for natural spawning to obtain synchronized embryo clutches for experiments. Usually, zebrafish (Tubingen) begin breeding at first light (7 AM). After feeding at 6:30 PM on the preceding day, pairwise breeding was established on a sloped breeding tank filled with zebrafish water. In order to maintain a 2:1 female to male ratio, four females and two males were housed in the breeding tank. When the light came on at 7AM in the morning, they often breed the next day. Approximately 500 eggs were found at the base of the breeding tank. After utilizing a strainer to gather the eggs, the strainer was rinsed with 1X E3 media (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4, and 0.01% methylene blue) before the eggs were put into the Petri dish (with an average of 25 embryos/Petri dish) and kept in the incubator at 28.4°C. All experiments in this study were conducted in compliance with institutional animal protocols under Committee for Control and Supervision of Experiments on Animal (CPCSEA) approval (No. NIBMGZF/002/2022-23).

Microinjection of morpholino

Translation blocker (TB) morpholinos and 5 base mismatch morpholino against cntnap5a and cntnap5b were purchased from Gene Tools, LLC. The cntnap5a and cntnap5b genes in zebrafish are orthologs of the human CNTNAP5 gene, sharing significant sequence and functional similarity. Mismatched controls (MM) were designed with morpholino sequences that differ by five base pairs from the cntnap5-targeting morpholinos. These controls were used to ensure that any observed phenotypic differences, such as retinal disorganization, were specifically due to cntnap5 knockdown rather than non-specific effects of the morpholino (MO) injection. MO and MM sequences are listed in S2 Table. Morpholino stocks were diluted in 10 mM KCl and 25% Phenol Red and each morpholino (2 ng) were mixed in a volume of 2 nL was microinjected injected into single cell-stage embryos using calibrated injection volumes [24]. S1 Fig represent the key steps of the experimental process including morpholino injection at the single cell stage, embryo development, eye measurements at 96 hpf, tissue processing, and subsequent histological and behavioural analyses.

Eye morphological assessment

At 96 hpf (hour post fertilization), larval zebrafish were anesthetized in tricaine solution and lateral images captured under a stereomicroscope (LEICA M205 FA) for morphological measurements. Eye diameter, defined as the axial length between the nasal and temporal periphery across the centre, was quantified for each larva using LAS 4V.13 software (LEICA). A 2-sample t-test was performed to compare the gross anatomical eye size both vertical and horizontal diameter of morphants and mismatch control in zebrafish to evaluate changes in eye size and morphology resulting from cntnap5 knockdown.

Tissue processing and cryosectioning

96 hpf zebrafish were anesthetized in tricaine methane sulfonate and subsequently euthanized following approved animal care protocols (Ferguson et al, 2019) [25]. Zebrafish were fixed overnight at 4°C in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS). After fixation, samples were washed in PBS 2–3 times to remove residual PFA. Tissues were then immersed in 30% sucrose in PBS overnight at 4°C for cryoprotection. Subsequently, zebrafish were oriented in OCT compound in peel-away cryomolds and quickly frozen using dry ice. Cryosections were cut at 15 μm thickness using a cryostat (LEICA CM 1860 UV) and collected directly on charged slides (Autofrost, Cancer Diagnostics). Sections were adhered to the slides by drying overnight at room temperature and stored at -80°C until use. Prior to immunostaining, slides were thawed at room temperature for 30–60 minutes and dried completely before beginning staining procedures.

Eosin and hematoxylin staining

The sections on glass slides were stained as follows: PBS for 12 mins, Mayer’s hematoxylin (SRL) in aqueous solution for 15 sec, water for 15 sec, eosin (SRL) in ethanolic solution for 20 sec, followed by dehydration with 100% ethanol for 5 sec twice, and finally mounted in DPX (Sigma) with cover slip. The tissue sections were observed under the light microscope (EVOS XL core Invitrogen). Detailed retinal analysis was performed on a representative subset of histological sections from 6–8 larvae per injection batch for each group (control and morphant).

Immunofluorescence staining and confocal microscopy

Immunofluorescence staining on cryosections was performed following previously published protocols [25]. Briefly, Cryosections were thawed and dried at room temperature for 30 minutes prior to immunostaining. Tissues were rehydrated with PBS and blocked for 1 hour at room temperature with 5% normal goat serum (Gibco) in PBS containing 0.3% Triton X-100. Sections were incubated with primary antibodies overnight at 4°C. The following primary antibodies and dilutions were used: mouse anti-acetylated tubulin (1:1000, Sigma Aldrich), rabbit anti-cntnap5 (1:200, Abgenex). After incubation, slides were washed 3 times with PBS and incubated with isotype-specific anti-rabbit secondary antibody labelled with Alexa Fluor 488 (1:500, Abcam) and anti-mouse secondary antibody labelled with Alexa Fluor 568 (1:500, Abcam) for 2 hours at room temperature protected from light. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (1 μg/mL, Invitrogen) for 5 min at room temperature in dark. Finally, sections were washed with 1X PBS and mounted with ProLong Gold Antifade Mountant (Invitrogen). Images were acquired using confocal microscopy (Nikon Ti2 Eclipse). Image analysis was conducted using Nikon NIS-Elements software and ROI intensity was compared by Student t-test using GraphPad Prism software.

Generation of customized cntnap5 antibody for gene knockdown validation

The sequence of cntnap5 Protein (XP_009294486.1 5 isoform X2, Danio rerio) was analysed by Hydroplotter software from protein lounge (proteinlounge.com) to select the suitable epitopes for antibody generation. The selection is based on analysing the amino acid sequences in Kyte-Doolittle and Hopp-Woods plots and selecting a suitable hydrophilic region also Kolaskar Tongaonkar for antigenicity. Based on the above analysis we have selected RSERNVREASLQVDQLPLR (571-589aa) for antibody development. (Antigenicity Index: 1.028, Hydrophobicity Index: -1.084, Hydrophilicity Index: 1.268)

A cysteine has been added in the N-terminus of the peptide for conjugation to carrier molecule. After synthesis of the peptide, it is conjugated to KLH (Keyhole lymphatic hemocyanin). Then it was immunized to the two healthy New Zealand white rabbits with freund’s complete adjuvant (CFA) followed by freund’s incomplete adjuvant (IFA). After four boosters the first immune sera were collected. The initial screening of the developed antisera was carried out by coating ELISA plate with the unconjugated peptide (200ng/well) (S2A Fig).

Based on the above information, Rabbit B 3rd bleed was selected for further peptide specific antibody purification by conjugating the free peptide with Sepharose 4B matrix. The purified antibody was again validated by employing indirect ELISA (S2B Fig).

Whole mount immunofluorescence

At 32hpf, embryos were dechorionated and transferred in 4% 1x PBS. Samples were permeabilized in ice-cold 100% methanol drop wise at -20°C for 2 hour and then washed with 1x PDT. Non-specific binding was blocked with 5% goat serum, 1% bovine serum albumin in PBS-Tween for 1 hr at room temperature. Embryos were incubated overnight at 4°C with rabbit anti-active caspase-3 primary antibody (1:500; Cell Signalling Technology) to label apoptotic cells. On the following day, embryos were washed and incubated with Alexa Fluor 568 goat anti-rabbit IgG secondary antibody (1:500; Molecular Probes) for 2 hrs at room temperature. Samples were washed with 1x PDT prior to imaging. 4% methylcellulose was used for whole mount imaging under the confocal microscope [26].

Neurotrace staining

Cryosections ranging from 12–15 μm thickness were collected on charged slides. Zebrafish eye tissue sections were stained with NeuroTrace 435/455 Blue Fluorescent Nissl Stain (1:50, Thermo Fisher) for 30 mins at room temperature as per manufacturer’s instructions. Sections were then washed 3 times in PBS and mounted with ProLong Gold Antifade (Invitrogen). Images were captured using confocal microscope (Nikon Ti2 Eclipse). Image intensity was measured using Nikon NIS-Elements software. The whole eye was selected as region of interest (ROI) and the fluorescence intensity within the designated ROIs was quantified. In order to account for non-specific fluorescence, background correction was done when recording mean intensity values. ROI intensity was compared by Student’s 2-sample t-test using GraphPad Prism software.

Zebrafish protein isolation

Zebrafish embryos were collected for protein isolation at 96 hpf after morpholino injection. Embryos were washed with PBS three times to remove any E3 media remaining. Whole cell lysis buffer was added to the embryos (HEPES (20mM) pH 7.6, 100mM NaCl, 1.5mM MgCl₂, 1mM DTT, 0.1% Triton X-100, 20% Glycerol and freshly added 10ul 0.1 M PMSF, 5ul Protease inhibitor cocktail per mL of buffer and homogenized using cordless mortar and pestle (Sigma-Aldrich) keeping on ice. After incubation in ice for 30 minutes the samples were centrifuged at 13,000 rpm for 20 minutes at 4°C. The supernatant was collected and used for western blotting [27].

Western blotting

Protein concentration of the prepared zebrafish protein lysates was estimated using Bradford reagent (Bio-Rad). Equal amounts of protein (30ug) were resolved by 10% sodium dodecyl sulphate–polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose membrane. Membranes were blocked with 5% non-fat milk in TBST (Tris-buffered saline-Tween 20) for 1 hour and incubated with primary antibodies overnight at 4°C. The following primary and secondary antibodies were used as per the mentioned dilutions: anti Acetyl-α-Tubulin Antibody (Sigma Aldrich), anti GAPDH (1:1000, CST), Goat Anti-Mouse IgG H&L (1:10,000, Abcam), Goat Anti-Rabbit IgG H&L (HRP) (1:10,000, Abcam). The blots were visualized using SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermofisher) in ChemiDoc XRS+ System (Bio-Rad). Densitometric quantification was measured in unsaturated images using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Spontaneous movement analysis

Larval zebrafish at 96 hpf were arrayed into 6-well transparent microtiter plate (1 larvae per well and each time 3 well for mismatch controls and 3 well for morphants) containing 200 μL E3 medium in each well. Plates were loaded into the Zantiks MVP system for locomotor tracking. The system was calibrated to maintain constant 28°C temperature and stabilize lighting for 30 minutes prior to recording. Following a 60-minute habituation period under darkness, the integrated software was programmed for automated alternating 10-minute epochs of light on and light off with complete darkness over a total 60-minute period. The movement tracks and total distance travelled (mm) per well were quantified for each light/dark interval period by the Zantiks tracking software. Statistical analysis was conducted using GraphPad Prism software (v9,). A nonparametric Mann-Whitney test was performed to compare the groups in light and dark separately.

In silico and in vitro characterization of CNTNAP5 genic region

Conditional analysis of thirteen genome-wide significant SNPs from our previous study clustered within a ~500kb region of CNTNAP5 on chromosome 2 showed maintenance of statistically significant genomic association with PACG. This indicates CNTNAP5 is likely a candidate gene for the glaucomatous trait (Fig 1A). Next, bokeh plot of linkage disequilibrium (LD) revealed that our earlier associated 13 variants of CNTNAP5, clustered into two LD block (Fig 1B). To contextualize the 13 PACG-associated intronic and downstream CNTNAP5 variants, uncovered from conditional analysis in terms of their putative regulatory functions, we intersected their locations with candidate cis-regulatory elements (CREs) annotated in the region. Plotting the CNTNAP5 SNPs relative to adjacent CRE genomic positions revealed two distinct spatial clusters exhibiting differences in predicted regulatory element abundance. One cluster (cluster I, Blue Square) containing rs17011381, rs2901264, rs2115890, rs733112, rs1430263, rs17011394 and rs780010) reside in a genomically sparse region largely devoid of annotated CREs. In contrast, the cluster II (Red square) encompassing rs17724018, rs2553625, rs17011420, rs2553628 and rs17011429 directly overlap with a dense hotspot of CREs enriched in putative active chromatin marks (Fig 1C). This differential CRE density profile provides preliminary evidence suggesting the latter SNP cluster has a higher probability of harbouring non-coding variants interfering with functionally relevant gene regulatory elements that may contribute to a breakdown in CNTNAP5 expression. Additionally, in DeepSEA database, we found same set (cluster II) of SNP are having significant enrichment scores on histone modifications demarcating putative active enhancer elements within introns of CNTNAP5 (Fig 1D). The RegulomeDB score of these 13 SNPs are enlisted in the S3 Table. Further the eQTL analysis of these SNPs within the CNTNAP5 gene revealed varying degrees of association between genetic variants and gene expression in the retina. Among the significant findings, rs2553628 (p = 0.00529), rs17011420 (p = 0.00365) and rs17011429 (p = 0.00522, r² = 0.01930) showed a notable positive association with CNTNAP5 expression, as indicated by their positive beta coefficients of 0.15117, 0.17926 and 0.31897, respectively. This suggests that the presence of these variants may enhance CNTNAP5 expression levels. Additionally, rs17011381 (p = 0.0093) also displayed a significant positive association with a beta coefficient of 0.37025. Other variants such as rs2553625 (p = 0.04423) demonstrated more moderate effects. Conversely, SNPs like rs2901264 and rs2115890 had higher p-values (0.25371 and 0.10104, respectively), suggesting weaker associations with CNTNAP5 expression. (S4 Table). The analysis of regulatory features for SNPs rs2553628 and rs17011429 revealed that both variants have high gene regulatory potential and both are residing in a strong linkage disequilibrium (LD). Examination of the HaploReg database showed that rs2553628 alters the motif for transcription factor TCF4 and having a CTCF binding site, while rs17011429 alters motifs for GR and Hsf. TCF4 is reportedly associated with neurological disorders [28]. Furthermore, rs2553628 demonstrated significantly altered motif activity between the reference and alternative alleles (Fig 2A). To functionally check the enhancer activity, in the dual luciferase assay, the G allele of rs2553628 (from the cluster II) was showing the significantly higher luciferase activity than the A allele (P value = 0.0034) (Figs 2B and S3). On further exploration on genomic architectural aspects surrounding CNTNAP5, we visualized Hi-C chromosomal conformation contact matrices encompassing the gene locus using the 3D Genome Browser. Notably, CNTNAP5 resides within a topologically associating domain (TAD) suggesting capacity for long-range spatial contacts between the gene body and more distal enhancer elements (S4A Fig). The contact map, as well as the corresponding arc also suggests distal as well as proximal interactions in the regions surrounding the prioritized genic region in the dorsolateral prefrontal cortex (S4B Fig). Moreover, the HUGIn analysis of the genomic region of CNTNAP5 reveals high proximal as well as distal interactions in the surrounding area as the observed counts were more than the expected counts especially outside the gene. The 4C plot from the database indicates the presence of frequently interacting regions (FIREs) near the CNTNAP5 locus in the dorsolateral prefrontal cortex (DPLC), alongside peaks for H3K27ac, H3K4me1, and H3K4me3 in both DPLC and neuronal progenitor cells (NPC) (Fig 2C–2E and S5 Table). The Human Protein Atlas expression dataset showed that the expression of CNTNAP5 is restricted only in retinal and neural tissues (S5 Fig). Analysis of human eye transcriptomic atlas datasets revealed marked enrichment of CNTNAP5 expression within central retinal tissues amongst profiled ocular cell types (S6 Fig). The neural pathways were significantly enriched after performing pathways enrichment analysis of the co-expressed genes of CNTNAP5 (S7 and S8 Figs). Analysis of the CNTNAP5 gene in the HumanBase database revealed several disease associations. The strongest association was with autism spectrum disorder, with a confidence score of 0.54. A weaker association was also found with epilepsy (confidence score 0.11). Most notably, HumanBase analysis also predicted an association between CNTNAP5 and retinal disease, with a confidence score of 0.11. To visualize these results, the confidence scores for CNTNAP5 across various neurological and retinal diseases were shown (S6 Table). Together, with the genetic and genomic architectural evidence, the expression patterns further strengthen the candidacy of CNTNAP5 in contributing to retinal phenotype outcomes in PACG. Additionally, mining of the phenotype-wide association study (PheWAS) catalog revealed significant genetic associations between variants in the CNTNAP5 locus and several neurological disorders including epilepsy, schizophrenia and Parkinson’s disease (S9 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

A. Regional Association Plot indicating the statistical strength of association among associated genomic region of CNTNAP5 gene with PACG after performing conditional analysis B. Bokeh plot shows the pairwise linkage disequilibrium between the 13 variants around the region of CNTNAP5 in two LD blocks C. CREs and SNPs plotted along with their genomic and epigenomic features of 13 SNPs of CNTNAP5. Based on the features, blue and red boxes are shown that roughly highlight SNP and CRE clusters. The elements and variants belonging to the red box show better likelihood for having functional relevance. D. The Whisker plot shows the DeepSEA gene regulatory score (P-value) of 13 SNPs of CNTNAP5. The plot includes a Bonferroni-corrected significance threshold line at 0.0038, indicating the adjusted significance level after correcting for multiple testing. SNPs below this threshold are considered statistically significant. Based on the features, blue (Cluster I) and pink (Cluster II) boxes are shown, representing different DeepSEA gene regulatory scores (P-values). The elements and variants belonging to the pink square have a better likelihood of having gene regulation potential.

https://doi.org/10.1371/journal.pgen.1011502.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

A. Variant detailed view of rs2553658 and rs17011429 in HaploReg. The regulatory motif TCF4 altered in reference and alternative allele of rs2553658. B. The data was drawn for a total of 3 biological replicates considering triplicate technical replicates for each biological experiments representing three different plasmid preparations (pGL3 SV40 minimal promoter vector) transfected in to HEK293T cells grown in 24-well plates. The height of the bar shows the mean value of all the 3 experiments, while error bars are depicting SE. The 4C plot of genomic region of CNTNAP5 (chr2:125,083,095–125,084,384) showing the FIREs, ChIP-Seq data and Hi-C counts in DPLC (C) and NPC (D) from HUGIn. The ratio of observed read counts to expected read counts is higher in NPC and is also statistically more significant than in DPLC (E).

https://doi.org/10.1371/journal.pgen.1011502.g002

cntnap5 morpholino-injected embryos show ocular phenotypic alterations in zebrafish

Following extensive genetic fine-mapping data integration indicating CNTNAP5 as a putative contributor in glaucomatous neurodegeneration, we next sought direct functional validation in vivo using antisense knockdown approaches in zebrafish larvae. Quantitative gross anatomical phenotyping revealed significantly reduced diameters along both the nasal-temporal axis (horizontal) and superior-inferior axis (vertical) of eyes measured from cntnap5 a and b translation-blocking double morpholino (dMO) injected embryos (n = 72, biological replicates = 6) at 72 hours post fertilization compared to 5 base pair mismatch (MM) morpholino controls (n = 72, biological replicates = 6) (P<0.001) (Fig 3A and 3B). There was no significant death observed in both control and morphant groups. The injection rate, death rate and other anatomical parameter was compared in both groups (S10 Fig). In deep phenotyping, histological examination showed noticeably diminished retinal nuclear layers with gaps and discontinuities in laminar organization in haematoxylin & eosin-stained thin sections from cntnap5 deficient larvae (Fig 3C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

A. Morpholino knockdown of cntnap5a and 5b, left cartoon is created with BioRender.com licensed version. Zebrafish were microinjected with a cntnap5 translation blocking morpholinos. Images taken 96 hpf, eye of a mismatch morphant (left) and an eye cntnap5 morpholino injected zebrafish (right) showing smaller anatomical eye than mismatch control. B. Quantitative gross anatomical diameters along both the nasal-temporal axis (horizontal) and dorsal-ventral axis (vertical) of eyes measured from cntnap5 translation-blocking morpholino injected embryos and mismatch morpholino controls zebrafish at 96 hours post fertilization stage. bars = mean ± SD, ns not significant, ***p < 0.005. C. Histological analysis of zebrafish eye microinjected with mismatch morpholino (left) a cntnap5 translation blocking morpholino (right) using H&E stain at 96 hpf stage showing disrupted retinal layers in cntnap5 morphant. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; PR, photoreceptor layer; RPE, retinal pigment epithelium; CH indicates choroid. D. Representative IF image of tissues stained with anti-cntnap5 showing cntnap5 expression in mismatch morphant (upper left; lower Transmitted Light Differential Interference Contrast (TD) image) but no expression in a cntnap5 translation blocking morphants (lower left: right: TD image).

https://doi.org/10.1371/journal.pgen.1011502.g003

CNTNAP5 antibody mediated validation of gene knockdown

In conjunction with the gross morphological and histological analyses demonstrating perturbation of eye and retinal development upon repression of endogenous cntnap5 translation, we also aimed to directly confirm successful protein knockdown in our zebrafish model using a target-specific customized antibody. As expected, dMOs exhibited a near complete absence of detectable anti-cntnap5 immunofluorescence throughout the retinal layers of eye. In contrast, the mismatch controls show cntnap5 antibody immunostaining in their whole retinal layers, P value <0.0005 (Figs 3D and S11). Thus, our data clearly suggest that cntnap5 deficiency disrupts retinal development in larval zebrafish.

cntnap5 morpholino-injected embryo eye showed disturbed retinal cytoarchitecture

To investigate the retinal tissue architecture after deep phenotyping using H&E staining, the morphant retina showed profoundly increased anti-acetylated tubulin immunofluorescence signal intensity and disoriented retinal tissue organization throughout the retinal ganglion cell layer compared to mismatch morphant retina; P-value = 0.0108 (S12 Fig). To validate this, we co-stained with anti-cntnap5 antibody along with DAPI to stain the nuclei (Fig 4A). Similarly, western blot quantification of whole embryo lysates further corroborated a significant ~1.5-fold upregulation of acetylated alpha tubulin protein levels in cntnap5 deficient embryos relative to mismatch controls; P value = 0.031 (Figs 4B and S13).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4.

A. Cntnap5 and Acetylated Tubulin Expression in Zebrafish Eye Tissue (96 hpf): Representative confocal images of cntnap5 and acetylated tubulin expression of eye tissues from mismatch control fish (upper) and cntnap5 morphant (lower) zebrafish at 96 hpf (blue- DAPI, green- cntnap5, red- acetylated tubulin, and merged). B. Acetylated Alpha Tubulin Levels (Whole Embryo) (72 hpf): Western blot image of acetylated tubulin expression of whole zebrafish tissue lysate (72 hpf). MO: cntnap5 morpholino injected fish MM: mismatch control morpholino injected fish. A two-sided t-test. bars = mean ± SD, ns not significant, *p < 0.05.

https://doi.org/10.1371/journal.pgen.1011502.g004

cntnap5 morpholino-injected zebrafish embryos show visual neurodegeneration and greater apoptosis

During our investigation, we employed the NeuroTrace stain to examine zebrafish retinal neurons. Notably, our findings revealed a significant reduction in NeuroTrace stain mirroring less Nissil granule intensity of retinal neurons within the dMOs compared to the MMs (P<0.005) (Fig 5A and 5B). This observed decrease in NeuroTrace stain intensity strongly suggests the presence of neurodegeneration in the cntnap5 morphant zebrafish. Finally, we performed whole mount immunofluroscence to observe the apoptosis by cleaved caspase 3 as an executioner caspase at a system level. Whole mount cleaved caspase 3 immunofluorescence staining showed dMOs had significantly more cleaved caspase 3 staining compared to mismatch controls (P<0.005). This indicates greater apoptosis in the dMOs versus controls (Fig 5C and 5D). Together, these results demonstrate increased apoptosis in vivo in dMOs, supporting the importance of cntnap5 for cell survival.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5.

A. Representative eye tissue images of Nissl granules expression of zebrafish embryos (96 hpf) injected with a mismatch control (left) and cntnap5 morpholino (right). Corresponding cartoons are created using BioRender.com licensed version. B. Comparative analysis of eye mean intensity of eye for both groups. bars = mean ± SE, ns not significant, ***p < 0.005. C. Representative whole mount confocal images of cleaved expression of zebrafish embryos (32 hpf) injected with cntnap5 morpholino (upper) or mismatch control (lower) D. Comparative analysis of mean intensity of eye for the both groups, bars = mean ± SE, ns not significant, ***p < 0.005.

https://doi.org/10.1371/journal.pgen.1011502.g005

Spontaneous locomotory movement in cntnap5 knockdown fish in light and dark phase

We analyzed the effects of cntnap5 knockdown on visually guided locomotor activity under intermittent light/dark conditions using the Zantiks system. During 10-minute light periods, cntnap5 morphants (dMO) showed significantly less average swimming distance compared to mismatch control larvae (p<0.005), reflecting impaired significant locomotory movement in bright light phase. However, distance moved in dark conditions were not significantly different between groups (p = 0.23) (Fig 6A and S1 Video). In summary, cntnap5 morphant zebrafish spontaneous movement was impaired in light phase compared to mismatch controls indicating towards their vision loss (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Locomotory movement analysis of zebrafish.

A. Total distance moved (mm) of 96 hpf cntnap5 morpholino injected zebrafish (blue circles) and 96 hpf cntnap5 mismatch control (red circles) zebrafish for the light (left image) and light off (right image) periods. Left image, (light on) p  =  <0.0001, two-sided Mann–Whitney test, n  =  24 cntnap5 morpholino injected zebrafish and n = 24 cntnap5 mismatch control, Right image (light off), p  =  0.5767, bars = mean ± SEM ns not significant, ****p < 0.0001. B. Average distance moved (mm) of 96 hpf cntnap5 morpholino injected zebrafish (blue circles) and 5dpf cntnap5 mismatch control (red circles) zebrafish exposed to four cycles of 10 min light on and 10 min light off after 60-minutes habituation period. Each point represents 1 min, n  =  24 cntnap5 mismatch control zebrafish and n  =  24 cntnap5 morpholino injected zebrafish.

https://doi.org/10.1371/journal.pgen.1011502.g006

Broader implications for glaucoma genetics

Our findings shed light on the genetic landscape of PACG, particularly the role of CNTNAP5 as a contributor to disease susceptibility. The genetic fine-mapping of the CNTNAP5 locus identified several intronic and downstream variants that are significantly associated with PACG. These variants overlap with CREs, which are regions that regulate gene expression by modulating the interactions between promoters and enhancers. The presence of these regulatory variants suggests that changes in CNTNAP5 expression levels could influence susceptibility to PACG, potentially altering the structural integrity of the retina.

The results of this eQTL analysis provide insights into how specific SNPs within the CNTNAP5 gene might influence its expression levels in retinal tissues, potentially impacting retinal structure and function. The significant associations identified for rs2553628, rs17011420 and rs17011429 highlight their potential regulatory roles, as these variants appear to increase CNTNAP5 expression. This upregulation could be critical in understanding the broader role of CNTNAP5 in retinal development and neurodegeneration.

The cross-referencing with the PheWAS catalogue revealed that CNTNAP5 polymorphisms are not only associated with glaucoma but also with several other neurodevelopmental and neurological disorders, such as schizophrenia, bipolar disorder, epilepsy, and Alzheimer’s disease [29–32]. This shared genetic basis indicates that CNTNAP5 may be involved in a range of neural processes that influence both central and peripheral nervous systems, including retinal neurons. The expression of CNTNAP5 in retinal cells, as observed in ocular tissue expression database, suggests that it may contribute to retinal development and maintenance, potentially making the retina more resilient to stressors like elevated intraocular pressure (IOP).

The role of CNTNAP5 in regulating the interactions between promoters and enhancers suggests that disruptions in these regulatory networks could alter the expression patterns of genes involved in retinal integrity. Understanding the mechanisms through which CNTNAP5 operate at the regulatory level could provide new insights into the genetic underpinnings of glaucomatous neurodegeneration and related conditions.

Molecular Genetic Underpinnings of the Disease

Our study emphasizes the role of CNTNAP5 in the regulation of retinal structure through its interaction with chromatin architecture and gene regulation. The Hi-C chromatin conformation data revealed that CNTNAP5 is positioned within a topological associating domain (TAD), facilitating interactions between its promoter region and distal regulatory elements[33]. Additionally, the local density and activity of cis-regulatory elements (CREs) such as enhancers within a TAD can further modulate expression [34]. These interactions likely influence the expression of CNTNAP5, impacting its role in retinal development. The presence of PACG-associated variants within regulatory regions suggests that genetic variations may alter the structural organization of the retina by modifying CNTNAP5 expression.

The dual luciferase reporter assays provided evidence of allele-specific enhancer activity for rs2553628, one of the variants within the CNTNAP5 locus. The PACG risk allele rs2553628 showed increased enhancer activity, suggesting that this allele may promote higher CNTNAP5 expression in certain contexts. Additionally, rs17011429, which has strong linkage disequilibrium with rs2553628, was identified as a co-eQTL associated with schizophrenia [29].This allele-specific regulation provides a potential mechanism by which genetic variations at the CNTNAP5 locus could contribute to the development of retinal degeneration, by influencing CNTNAP5 levels during critical stages of retinal development.

In vivo analysis to support the hypothesized mechanisms of action

Using antisense knockdown experiments in zebrafish, we demonstrated cntnap5 knockdown results in ocular phenotypes including reduced eye size, disrupted retinal architecture with discontinuous nuclear layers as observed with reduced NeuroTrace intensity and increased acetylated tubulin levels. In contrast to the eye enlargement seen in human congenital glaucoma due to elevated intraocular pressure, the reduced vertical and horizontal eye diameters observed in cntnap5-morphant zebrafish likely reflect disruptions in retinal and ocular development due to CNTNAP5 knockdown. Moreover, increased acetylated tubulin levels following cntnap5 knockdown indicate alterations in microtubule dynamics in cntnap5 morphant retinas possibly leading to cytoskeletal defects and disturbed cytoarchitecture. Acetylation of tubulin is a post-translational modification that alters microtubule stability and intracellular trafficking [35]. Earlier reports suggested that tubulin acetylation restricts axon over branching by reducing microtubule plus-end dynamics in nerves [36]. Defects in microtubule-based transport could disrupt inner retinal laminar organization which manifest as gaps and discontinuous histology. Dysfunctional intracellular trafficking resulting from tubulin hyperacetylation provides a plausible mechanism contributing to the neurodegenerative effects of cntnap5 knockdown. Axonal transport defects are linked to several neurodegenerative diseases [36,37]. Additionally, while the morphant eyes exhibited disorganization across all retinal layers, which contrasts with the selective retinal ganglion cell (RGC) loss typical in human glaucoma, this broader disorganization likely reflects the developmental role of CNTNAP5 in maintaining retinal structure and cell integrity.

Our findings reveal a significant reduction in NeuroTrace stain intensity in the retinal neurons of cntnap5 morphant zebrafish compared to mismatch control zebrafish. The decreased NeuroTrace stain intensity indicates loss of Nissl substance and suggests the presence of neurodegeneration in the morphant zebrafish retina. The loss of Nissl staining we observed in all the neuro-retinal layers of cntnap5 morphants reflects chromatolysis, or disintegration of Nissl bodies, which is a characteristic feature of neurodegeneration [38]. This indicates that CNTNAP5 is essential for maintaining the structural integrity of retinal layers.

Our findings of increased cleaved caspase 3 immunofluorescence in the wholemount cntnap5-morphant zebrafish provides evidence that cntnap5 deficiency leads to greater apoptosis during development. The caspase family of cysteine proteases play a central role in programmed cell death pathways [39]. Executioner caspases including caspase 3 are activated by initiator caspases in response to pro-apoptotic signals, ultimately leading to widespread proteolysis and cell death [40]. Increased cleaved caspase-3 staining in the cntnap5 morphant retina indicates heightened apoptosis, suggesting cntnap5 is essential for retinal cell survival. This deficiency likely makes retinal neurons more vulnerable to apoptosis, leading to thinner retinal layers. Similar findings in glaucoma link disrupted pro-survival signalling to neurodegeneration through caspase-mediated cell death. Our data support a link between cntnap5 loss and increased retinal apoptosis, connecting this novel glaucoma risk gene to the apoptotic death of retinal neurons [40–43].

CNTNAP5 belongs to the neurexin superfamily, a group of proteins known for their roles in synaptic organization, neural signalling, and cell adhesion [29]. Within the retina, these functions are essential for the proper arrangement of neurons and the maintenance of retinal structure. The role of CNTNAP5 in axon guidance and synaptic stability likely contributes to its ability to maintain the alignment and structure of retinal cellular architecture, preventing structural abnormalities that could predispose the retina to further damage. Supporting these roles, our findings clearly show that knockdown of cntnap5 in zebrafish larvae result in widespread retinal disorganization and overall disruption of cellular architecture of the retina.

Finally, our analysis of visually guided behaviour provides functional evidence that cntnap5 knockdown impairs vision in larval zebrafish. Analysis of locomotor activity clearly showed significantly decreased movement during intermittent light periods in cntnap5 morphants compared to controls. However, swimming distance in dark period was equivalent between groups. This selective reduction in light phase locomotion indicates the cntnap5 deficient larvae have impaired visual function despite intact neuromuscular capacity. Similar light/dark dependent behavioural defects have been reported in zebrafish models of inherited blindness [44]. Visually-guided locomotion requires integration and transmission of visual cues to motor circuits [45]. Decreased photoperiod movement in cntnap5 morphants reflects an inability to respond appropriately to light onset due to deficiencies in retinal processing. Loss of cntnap5 expression likely impairs retinal circuit development and function, manifesting as vision-dependent locomotor deficits. Further electrophysiological analysis of visual responses and circuit connectivity in cntnap5 morphants will provide greater mechanistic detail into how cntnap5 loss impacts retinal signalling pathways leading to functional vision loss. However, our locomotor activity data clearly demonstrates cntnap5 plays a critical role in the proper development of visual behaviour and function in zebrafish larvae. Our findings are consistent with the morphological and histological retinal abnormalities resulting from cntnap5 knockdown underlying vision loss.