Pannexin 1a regulates visual system development and ocular integrity in zebrafish

TL;DR

This study shows that Pannexin 1a is crucial for eye development and function in zebrafish, as its absence leads to vision problems, eye structure issues, and retinal thinning.

Contribution

The study identifies Pannexin 1a as a key regulator of ocular integrity and refractive development in zebrafish.

Findings

Panx1a deficiency causes impaired visuomotor behavior and reduced contrast sensitivity in zebrafish larvae.

Adult zebrafish lacking Panx1a show axial myopia and progressive lens disruption.

Retinal thinning and ganglion cell loss are observed in Panx1a−/− zebrafish.

Abstract

Pannexin 1a (Panx1a), a large-pore ATP release channel broadly expressed in the vertebrate eye, has an unclear role in ocular development. Using a zebrafish Panx1a knockout, we observed progressive ocular phenotypes, including impaired visuomotor behavior, axial myopia, cataract-like lens abnormalities, lens defects, and retinal thinning. Optokinetic response (OKR) assays revealed markedly reduced responsiveness in Panx1a−/− larvae, with response rates declining from 91.7% to 50% at 20% contrast and from 100% to 41.7% at 100% contrast, along with prolonged saccade intervals (2-fold at 20% and 4-fold at 100% contrast). High-resolution optical coherence tomography (OCT) in adults showed increased normalized axial length (7–8% across ages) and reduced lens-to-axial length ratios (0.58 vs. 0.63 at 8 months). Histology revealed age-dependent lens epithelial disruption and increased light scatter. Retinal analyses demonstrated thinning of the whole retina (10%) and ganglion cell layer (33%), with occasional vascular tortuosity. Together, these findings identify Panx1a as a key regulator of ocular integrity and refractive development in zebrafish.

•Optimized whole-eye and retinal OCT enables detailed analysis of adult zebrafish eye structure•Panx1a loss impairs larval visuomotor behavior and reduces contrast sensitivity•Whole-eye OCT reveals axial myopia and age-progressive lens disruption in Panx1a−/− adults•Panx1a deficiency causes retinal thinning, ganglion cell loss, and tortuous blood vessels

Optimized whole-eye and retinal OCT enables detailed analysis of adult zebrafish eye structure

Panx1a loss impairs larval visuomotor behavior and reduces contrast sensitivity

Whole-eye OCT reveals axial myopia and age-progressive lens disruption in Panx1a−/− adults

Panx1a deficiency causes retinal thinning, ganglion cell loss, and tortuous blood vessels

Molecular neuroscience; Developmental neuroscience; Sensory neuroscience

Figures

Panx1 ablation affects responsiveness to moving gratings(A) Moving gratings are presented to immobilized 6-day post-fertilization larvae of both genotypes. The eye movements were video recorded and analyzed offline using Stytra (Vilim Stih et al., PLOS Computational Biology, 2019).(B) Traces of angular horizontal eye movements of both genotypes at 20% and 100% light. *n* = 3 larvae (C) Positive responses (%) of larvae stimulated by moving gratings at 20% and 100% light. *n* = 12 larvae.(D) Time between saccades is identified as the time between the maximum amplitude of a saccade and the onset

Ocular geometry measurements of 8-month-old and 1-year-old zebrafish(A) shows an OCT 3-D volume in different cuts.(B and C) display B-mode and en-face images with the measured parameters (scale bars, 500 μm).(D–K) present estimation plots for normalized axial length (axial length/body length), RRE, Lens diameter, and proportion of lens diameter to normalized axial length for each group, respectively. (*N* = 30/genome/age). As it is shown, in the 8-month-old and 1-year-old group, the normalized axial length of the eye is significantly elongated, and RRE has a significant negative shift (*p*-val

Panx1a mutant lenses have abnormal lens texture and focusing power(A and B) Macroscopic comparison of Panx1a^+/+^ and Panx1a^−/−^ eyes under the same lighting conditions.(C) shows the light focus pattern through a control Panx1a^+/+^ lens with the black arrow highlighting the 140 μm lens working distance.(D) Illustrates the laser light passage through a defective Panx1a^−/−^ lens, where no distinct focal point is observed.(E) demonstrates how the MATLAB code selects a region of interest (ROI) from the lens and calculates the full width at half maximum (FWHM) and mean intensity to analyze lens

Retinal immunohistochemistry and OCT image of adult zebrafishPanels (A) and (B) present the expression and localization of the Panx1a (in green) antibody throughout the retinal layer. The Panx1a is localized in the horizontal cell layer (HC) and in the ON/OFF ganglion cell layer, as it is shown in panel (a1). The Panx1a-KO served as the negative control (scale bars, 25 μm). Panel (C) is a B-mode OCT image of the TL zebrafish retina obtained using our modified 880 nm OCT system and zebrafish retina probe, and Panel (D) is a B-mode retina OCT of Panx1a−/−. The red square in Panel (C) indicates t

Graphical summary of the study and tested hypothesisInitial OKR analysis in larvae revealed ocular motor deficiencies in Panx1a^−/−^ zebrafish, leading to the hypothesis that Panx1a deletion causes structural and/or functional deficits in the eye. Subsequent analyses in adult control and mutant zebrafish confirmed that Panx1a mutation affects emmetropization, induces RGC loss, and produces cataract-like changes/structural defects in the lens, thereby verifying the hypothesis and validating the understanding that Pannexin1a regulates visual system development and ocular integrity in zebrafish.

Full text

Introduction

Visual impairments such as myopia, cataracts, and glaucoma are major global health challenges, affecting hundreds of millions of people and substantially diminishing both quality of life and economic productivity.1 These disorders are complex, multifactorial, and often progressive, necessitating deeper insights into their underlying genetic, molecular, and cellular mechanisms. To this end, animal models have been instrumental in elucidating the pathophysiology of ocular diseases, providing tractable platforms to dissect gene function and test therapeutic strategies. For example, zebrafish models have advanced clinical health outcomes by enabling the discovery of disease-causing genes such as SLC25A38 in sideroblastic anemia2 and CCBE1 in Hennekam syndrome,3 while also facilitating drug development efforts, including Prostaglandin E2 (ProHema) for hematopoietic stem cell transplantation4 and Leflunomide for melanoma treatment.5

Among available vertebrate models, the zebrafish (Danio rerio) has emerged as a powerful system for studying ocular development and disease. Zebrafish exhibit key anatomical and functional similarities to the human eye, including a layered retina, cone-dominant photoreception, and diurnal vision cycles.6^,^7 Their external fertilization, optical transparency, and rapid development allow high-resolution in vivo imaging and large-scale genetic screening, making them particularly suitable for studying refractive errors, lens biology, and retinal degeneration.8^,^9

Recent genome-wide association studies and functional genomics have implicated numerous genes in eye development and disease, yet many remain incompletely characterized. One such candidate is Pannexin1 (Panx1), a large-pore ATP-permeable channel protein known for its roles in purinergic signaling, mechanotransduction, and cell-cell communication. Panx1 is expressed in a variety of tissues, including the retina, lens, and cornea, where it participates in responses to membrane stretch, extracellular ATP, and calcium flux.10^,^11 Its relevance to ocular physiology is underscored by studies in mice and humans suggesting that Panx1 contributes to intraocular pressure regulation, neuronal survival, and inflammation, all of which are processes relevant to glaucoma and other degenerative eye conditions.12^,^13^,^14^,^15

In zebrafish, the gene family includes two Panx1 paralogs, Panx1a and Panx1b, arising from a teleost-specific genome duplication event.10^,^16 While Panx1a is predominantly expressed in the horizontal cells of the outer retina, Panx1b is enriched in the inner retinal layers, including ganglion and amacrine cells, indicating potentially non-redundant functions.16 Functional studies have begun to elucidate these roles; for instance, Panx1a knockout zebrafish exhibit visual behavior deficits, including impaired optokinetic response (OKR) and abnormal dopaminergic signaling.17 These findings suggest that Panx1a plays a key role in retinal processing and sensorimotor integration.

However, the potential role of Panx1a in lens development, refractive tuning, and epithelial homeostasis has not yet been explored in zebrafish. This represents a significant knowledge gap, as pannexins are broadly expressed in vertebrate ocular tissues and are key regulators of extracellular ATP signaling and ionic fluxes. Evidence from rodent, porcine, and other mammalian lenses demonstrates that Panx1 functions as an ATP-release channel that modulates purinergic signaling, Ca^2+^ wave propagation, and epithelial stress responses. In these systems, Panx1 activity regulates Na^+^/K^+^-ATPase function and intracellular Ca^2+^ dynamics, processes essential for maintaining lens transparency, ionic balance, and volume homeostasis.18 Panx1 activation has also been linked to oxidative stress pathways and redox-sensitive ATP release, both of which are particularly relevant to the metabolically active lens epithelium.19 Furthermore, given its structural and functional parallels with connexins, proteins indispensable for lens fiber differentiation, gap-junctional communication, and refractive development, Panx1a may plausibly contribute to axial length regulation and emmetropization through analogous regulatory mechanisms.20^,^21^,^22^,^23^,^24^,^25 Collectively, although no studies have previously examined Panx1a function in the zebrafish lens, mammalian evidence strongly supports a model in which Panx1a could influence lens ionic homeostasis, ATP-dependent signaling, and optical integrity.

Human data also support the clinical relevance of this channel. A germline mutation in PANX1 was recently associated with a multisystemic disorder characterized by intellectual disability, sensorineural hearing loss, and night blindness, suggesting a broader role in sensory system development.26 Yet, ocular-specific phenotypes in such cases remain poorly defined, and functional modeling in vertebrates is lacking.

In this study, we investigate the role of Panx1a in ocular development using a knockout zebrafish model across four developmental stages, assessing visual behavior, ocular biometry, lens transparency, and retinal morphology. Building on our previous findings of altered visual-motor behavior and dopaminergic signaling in Panx1a^−/−^ larvae,27^,^28^,^29^,^30^,^31 we carried out optokinetic response assays and found significant abnormalities in the mutants during larval development. Because OKR performance is dependent on the integrity of image-forming structures such as the retina and lens, and impaired OKR can be indicative of structural visual deficits rather than only primary reflex pathway dysfunction,32^,^33 we sought to further investigate the basis of this impairment. We therefore refined our hypothesis to distinguish between two possible mechanisms: (1) a functional disturbance of visuomotor reflex circuitry, or (2) structural abnormalities in ocular tissues responsible for forming a clear image. Given the nature of the OKR deficits, we hypothesized that Panx1a deficiency primarily leads to structural defects in ocular components rather than a generalized functional visual reflex disorder. To test this, we performed high-resolution whole-eye and retinal 3D OCT imaging and complementary histological analyses to characterize the adult ocular phenotypes. Our findings reveal that Panx1a deletion is associated with a distinct ocular phenotype, including progressive myopia, lens structural defects and cataract-like changes, and retinal thinning, highlighting Panx1a as a critical regulator of ocular structure and function. This work expands our understanding of pannexin channel biology in the eye and establishes the Panx1a knockout zebrafish as a valuable disease model for investigating the impact of Panx1 mutations in humans, providing a foundation for future studies into ATP signaling and mechano-transduction in refractive development and visual pathology.

Results

Pannexin 1a−/− larvae exhibit ocular motor deficiency

To determine whether Panx1a contributes to visual-motor function during early development, we assessed the optokinetic response (OKR) in Panx1a^−/−^ and Panx1a^+^/^+^ zebrafish larvae using contrast-dependent visual stimuli. Here, we evaluated the optokinetic response (OKR) in six-day post-fertilization larvae. Eye movements were recorded under clockwise-moving gratings at 20% and 100% contrast (Figure 1A). Horizontal saccadic movements (>15°) were extracted from traces (Figure 1B). Panx1a^−/−^ larvae exhibited reduced responsiveness to moving gratings compared to Panx1a^+/+^ larvae (Figure 1C). At 20% contrast, 91.7 ± 1.3% of Panx1a^+/+^ larvae responded, while the percentage responding was significantly lowered to 50 ± 15.1% in Panx1a^−/−^ larvae (p = 0.027). Similar trend was observed at 100% contrast, with 100% of Panx1a^+/+^ larvae, and 41.7 ± 14.9% of Panx1a^−/−^ larvae, responding to the stimuli (p = 0.0024).Figure 1. Panx1 ablation affects responsiveness to moving gratings(A) Moving gratings are presented to immobilized 6-day post-fertilization larvae of both genotypes. The eye movements were video recorded and analyzed offline using Stytra (Vilim Stih et al., PLOS Computational Biology, 2019).(B) Traces of angular horizontal eye movements of both genotypes at 20% and 100% light. n = 3 larvae (C) Positive responses (%) of larvae stimulated by moving gratings at 20% and 100% light. n = 12 larvae.(D) Time between saccades is identified as the time between the maximum amplitude of a saccade and the onset of a new saccade.(E) Saccade amplitudes are identified as the maximum angle (in degrees) reached when eyes jerked in response to the stimulus presented at 3.6°/sec.(F) Saccade duration is calculated as the time between the onset and the peak of the eye movement (ms).(G, H) Immunohistochemistry-based detection of Panx1a in the brain and retina using the anti-Panx1a antibody from.17 Arrows in (G) indicate immunoreactive arborization fields in a cross-section of a whole head (6dpf). Arrows in (H) indicate the optic nerve (ON), inner plexiform layer (IPL), retinal ganglion cell layer (RGCL), and lens epithelium (LE) in the retina. The number of animals used for analysis, Panx1a^+/+^ (20%) n = 8, Panx1a^+/+^ (100%) n = 3, Panx1a^−/−^ (20%) n = 12 and Panx1a^−/−^ (100%) n = 5. Statistical analysis: Welch’s t test; p-value significance ∗ <0.05 ∗∗ <0.01 ∗∗∗ <0.005 ∗∗∗∗ <0.001. Scale bars, (G) (150 μm), (H) (50 μm).

Saccade intervals in Panx1a^+/+^ larvae were significantly reduced with increased contrast (20%: 13.6 ± 1.3s vs. 100%: 4.0 ± 1.5s, p < 0.0001) (Figure 1D). An increase in the contrast from 20% to 100% had no significant effect on the saccade intervals of Panx1a^−/−^ larvae. However, Panx1a^−/−^ larvae showed prolonged saccade intervals at both contrasts (20%: 24.1 ± 3.4s, 100%: 16.3 ± 1.8s), significantly differing from controls at 20% (p = 0.012) and 100% contrast (p < 0.0001).

Saccade amplitude decreased with higher contrast in Panx1a^+/+^ larvae (20%: 23.3 ± 0.7°, 100%: 20.3 ± 0.9°, p = 0.0187) (Figure 1E). However, Panx1a^−/−^ larvae showed no significant amplitude change (20%: 21.4 ± 1.0°, 100%: 20.8 ± 0.6°). Similarly, saccade duration remained unchanged across all groups (Panx1a^+/+^ 20%: 129.1 ± 5.8 ms, 100%: 109.6 ± 10.0 ms; Panx1a^−/−^ 20%: 137.5 ± 5.6 ms, 100%: 125 ± 4.1 ms) (Figure 1F).

The immunohistochemistry image highlights Panx1a expression (green) in the inner plexiform layer (IPL) of the retina, arborization fields in the midbrain (including the optic tectum), and cells surrounding the lens (Figures 1G and H). In larval zebrafish, the localization of Panx1a supported its roles in both retinal signal processing and sensorimotor integration, which are essential for OKR function. The presence of Panx1a in cells around the lens raised the possibility that it may also contribute to lens maintenance or transparency.

To further investigate the extent of Panx1a′s influence on the visual system, we prioritized the question of whether structural and synaptic integrity deficits can be observed in the adult retina. Given its expression in the lens, the lens morphology and transparency were also investigated to determine whether Panx1a deletion predisposes the zebrafish to cataract-like phenotypes or other refractive abnormalities during adult developmental stages, further contributing to visual dysfunction.

Pannexin 1a deletion induces myopic ocular alterations revealed by volumetric whole-eye optical coherence tomography imaging

To assess the refractive status of the mutant eyes, we measured geometrical parameters using optical coherence tomography (OCT), a non-invasive imaging technique that has become a gold standard analytical tool in the field of ophthalmology.27^,^31^,^34 Ophthalmic OCT studies in zebrafish frequently use shorter central wavelengths (800–1050 nm) to obtain high-resolution retinal images. However, these wavelengths experience greater light attenuation in larger adult eyes, which can limit penetration depth and reduce the visibility of deeper ocular structures. Although adult zebrafish imaging has been successfully performed using commercial systems such as those described by Collery et al., the image quality at shorter wavelengths often does not fully resolve whole-eye morphology in older fish. To overcome these limitations, we optimized a custom-developed 1310 nm Spectral-Domain OCT (SD-OCT) system to enable high-resolution whole-eye 3-D imaging of adult zebrafish (Figure 2A). By raster scanning the laser beam across the eye, we acquired high-resolution volumetric images which could be sliced to obtain either B-mode (aka. sagittal or coronal planes), or en-face images (aka. transverse plane). The central B-mode images of the volumetric datasets were then used for extracting geometrical measures, as schematically shown in the OCT B-mode and en-face images of Figures 2B and 2C. To avoid developmental biases, we normalized the measured axial length of each eye to the respective fish body length, measured from the top of the head to the beginning of the fin. We also calculated the Relative Refractive Error (RRE) for each eye, which is a unitless measure, with values less than zero indicating myopia and positive values indicating hyperopia. The formula commonly used for calculating RRE of zebrafish is RRE = 1 − retinal radius/F; where F = idealized focal lengtℎ = lens radius ∗ 2.324.35^,^36^,^37^,^38 Measurements were conducted for both Panx1a^+/+^ and Panx1a^−/−^ groups at three developmental stages of 8 months, 1 year, and 1.5 years of age, with n = 30 eyes per age group per genotype. Figures 2D and 2H depict the normalized axial lengths measured from OCT images of 8-month-old and 1-year-old Panx1a^+/+^ and Panx1a^−/−^zebrafish. Results from the 1.5-year-old Panx1a^−/−^ group are not included in Figures 3D and 3H because of the presence of severe lens defects in these eyes (discussed in the next section), which significantly altered the eye structures and rendered measurements of the geometrical parameters inaccurate. Our results indicate a significant elongation in the normalized axial length of the eye in mutants at 8-month-old (Panx1a^−/−^: 48.79 ± 2.24 μm/mm, Panx1a^+/+^: 45.34 ± 3.59 μm/mm, p < 0.0001). Similar pattern was observed in the 1-year-old mutants (Panx1a^−/−^: 48.73 ± 1.38 μm/mm, Panx1a^+/+^: 47.57 ± 1.58 μm/mm, p = 0.0012). Additionally, the relative refractive error (RRE) in the 8-month-old and 1-year-old mutant groups showed a significant negative shift compared to age-matched controls (8-month-old, Panx1a^−/−^: −0.05 ± 0.04 μm/μm, Panx1a^+/+^: 0.04 ± 0.05 μm/μm, p < 0.0001) (1-year-old, Panx1a^−/−^: −0.05 ± 0.03 μm/μm, Panx1a^+/+^: 0.02 ± 0.03 μm/μm, p < 0.0001) (Figures 2E and 2I). Both the elongated axial length and the negative shift in RRE indicated a myopic shift in the mutants’ eyes. While the difference of the normalized lens diameter of the two genotypes was not statistically significant at 8 months (p > 0.05), the normalized lens diameter was significantly reduced in mutants at the 1-year-old developmental stage (Panx1a^−/−^: 28.02 ± 1.28 μm/mm, Panx1a^+/+^: 29.42 ± 1.20 μm/mm, p < 0.0001). Accordingly, the ratio of lens diameter to axial length of the eye was significantly reduced in the 8-month-old and 1-year-old Panx1a^−/−^ mutants (8-month-old, Panx1a^−/−^: 0.58 ± 0.017 μm/μm, Panx1a^+/+^: 0.63 ± 0.024 μm/μm, p < 0.0001) (1-year-old, Panx1a^−/−^: 0.58 ± 0.014 μm/μm, Panx1a^+/+^: 0.61 ± 0.016 μm/μm, p < 0.0001).Figure 2. Ocular geometry measurements of 8-month-old and 1-year-old zebrafish(A) shows an OCT 3-D volume in different cuts.(B and C) display B-mode and en-face images with the measured parameters (scale bars, 500 μm).(D–K) present estimation plots for normalized axial length (axial length/body length), RRE, Lens diameter, and proportion of lens diameter to normalized axial length for each group, respectively. (N = 30/genome/age). As it is shown, in the 8-month-old and 1-year-old group, the normalized axial length of the eye is significantly elongated, and RRE has a significant negative shift (p-value<0.001).(L–O) shows the development of normalized axial length, RRE, lens diameter, and proportion of lens diameter to normalized axial length in Panx1a^+/+^ versus Panx1a^−/−^ group in both developmental stages. Statistical analysis: Welch’s t test; Significance ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.Figure 3. Progressive lens defects in the Panx1a mutant(A) Presenting an OCT image of a Panx1a^+/+^ healthy eye.(B) Depicts a specific type of lens defect in the eye of an 8-month-old Panx1a^−/−^ fish (scale bars, 500 μm).(C) Illustrates a different type of lens defect in the eye of a 1-year-old Panx1a^−/−^ fish.(D) shows another variant of lens defect in the eye of a 1.5-year-old Panx1a^−/−^ fish.(E) Presents an OCT image indicating complete aphakia, observed in a fish from the 1.5-year group. Blue rectangles mark the same area of the lens in all the B-modes. Yellow squares emphasize the same retina section in all the B-modes, and in the mutants, we can observe the retina bumps.Yellow arrows show the lens rings, which are different in mutant eyes (B-E) compared to healthy eyes (A).(F–I) Represent the selected lens area from the whole-eye images above each, respectively. The green arrow is pointing to the highly scattering areas.(J) Displays a pie chart of Panx1a^−/−^ eyes with lens defects across different age groups, indicating an age-related increase in the occurrence of this phenotype. Panel (K) shows a 4× histological image of a 1.5-year-old Panx1a^+/+^ eye, highlighting the normal lens structure, while panel (M) presents a corresponding 20× magnified view of the same area, displaying the clear lens interior and organized, square-shaped epithelial cells. In contrast, panel (L) shows a 4× histological image of a Panx1a^−/−^ eye with a defective lens, and panel (N) provides a 20× magnified view of the same region, revealing disorganized cellular structures within the lens, including migrated epithelial-like cells with visible nuclei.

To study the variation of geometrical parameters with age, we compared the normalized axial length, RRE, normalized lens diameter, and ratio of lens diameter to axial length for 8-month-old and 1-year-old groups (Figures 2L–2O). Figure 2L shows significant growth in the normalized axial length of the eye from 8 months to 1 year of age (p < 0.01) in the Panx1a^+/+^ group (as expected), but such growth is not observed in the Panx1a^−/−^ group (p > 0.05). The relative refractive error (RRE) followed a similar pattern with a significant difference between the two developmental stages in the Panx1a^+/+^ group (p < 0.05), but not in the Panx1a^−/−^ group (p > 0.05).

Similarly, the growth of normalized lens diameter from 8 months old to 1 year old was different between Panx1a^+/+^ and Panx1a^−/−^ (Figure 2N). The normalized lens diameter was significantly increased from 8 months old to 1 year old in the Panx1a^+/+^ group (p < 0.001), while no significant growth was detected in the Panx1a^−/−^ group between the 8-month-old and 1-year-old groups. As the axial length in the mutants does not increase at the same rate as in the Panx1a^+/+^ group and their lens diameter did not grow with age in the Panx1a^−/−^ group, the ratio of lens diameter to axial length stays the same from 8 months to 1 year old (Figure 2O). However, in the Panx1a^+/+^ group, the proportion of lens diameter to axial length is significantly decreased from 8-month-old to 1-year-old (p < 0.05). (Table S1).

Pannexin 1a deletion induces age-progressive lens epithelial disorganization and cellular intrusion in zebrafish

Another significant finding from our whole-eye OCT images is the presence of lens epithelial defects in the Panx1a^−/−^ mutants. Figure 3A shows a control Panx1a^+/+^ lens with epithelial layers and the capsule appearing round and intact (as evident in the zoomed view in panel F). In contrast, OCT images from 8-month-old, 1-year-old, and 1.5-year-old mutants show representative epithelial defects highlighted by the blue rectangles in Figures 3B–3D, and in zoomed views in panels G-I. The observed epithelial defects appear as saturated regions (green arrows in panels G-I) in the OCT images and frequently with rupture-like presentations. Additionally, we identified three cases of complete aphakia (absence of the lens) in the 1-year-old (n = 1) and 1.5-year-old (n = 2) mutant groups (Figure 3E). Qualitatively, we observed an increase in severity and quantity of defects with age in the Panx1a^−/−^ groups, many of which manifested as complete holes in the lens epithelium surrounded by highly scattering areas accompanied by the disorganization of the lens layers. Quantitative analysis revealed an age-related increase in the number of lens defects, with 13% of cases (3/23) in the 8-month-old group, 26% (6/23) of cases in the 1-year-old group, and 61% (14/23) of cases in the 1.5-year-old group. (Figure 3J). Additional observations in the whole-eye images include apparent "bumps" in the Panx1a^−/−^ retina, highlighted with yellow squares in panels (B–D), whereas no such bumps are observed in the Panx1a^+/+^ retina shown in panel A. These “bumps” are not true structural protrusions of the retina, but rather optical artifacts caused by the altered material properties of the mutant lens and discontinuity of refractive index through which the light passes before it gets to the retina. Upon closer inspection, the lens epithelium rings (yellow arrows in Figures 3A–3D) exactly above these perceived retina bumps in the mutant fish appear thicker and more light scattering compared to controls.

Histological H&E staining on representative defective eyes provided a more comprehensive understanding of the nature of defects visualized in OCT images (Figures 3K–3N). Panel (K) (and the zoomed view in panel (M)) depicts the typical histological structure of a healthy Panx1a^+/+^ lens, characterized by square-shaped lens epithelial cells lining the anterior surface and an otherwise clear lens interior with no cellular content or scattering structures. In contrast, a defective Panx1a^−/−^ lens, shown in panel (L) and the magnified panel (N), displays abnormal cellular features within the lens. These include cells with visible nuclei (arrow in panel N) that resemble differentiated, migrated epithelial cells, some showing signs of lumen formation. These abnormal cell types correspond to the regions of high scattering observed in the OCT images of Panx1a^−/−^ lenses. Notably, such scattering patterns and intra-lens cellular structures are absent in the Panx1a^+/+^ controls. These correlations between OCT and histological findings reveal the histological origins of the lens defects observed in the OCT images of Panx1a^−/−^ groups. Collectively, our results demonstrate the presence of several types of lens defects in Panx1a^−/−^, which increased in severity and quantity with age.

Pannexin 1a deletion induces abnormal lens texture and impaired focusing power in zebrafish

The qualitative macroscopic inspection of zebrafish eyes suggested increased opacity in the lenses of Panx1a^−/−^ compared to those of Panx1a^+/+^ (Figures 4A vs. 4B), suggesting possible textural alterations in the mutants. To investigate further, we performed texture analysis of OCT datasets from a subset of Panx1a^−/−^ lenses that did not show overt structural abnormalities as described in the previous section. To assess the functional capacity of these morphologically subtle lenses, we also surgically removed the lenses from the eyes and evaluated their focusing power in a controlled environment using a custom-designed experimental setup (see STAR methods section for details). Using this setup, we studied the lens’ ability to focus green collimated light for both Panx1a^+/+^ and Panx1a^−/−^ zebrafish (Figure 4C vs. 4D). Results show that Panx1a^+/+^ lenses produced a sharp and distinct focal point with a working distance of 140 μm (Figure 4C), indicating the ability to effectively focus light on the retina. In contrast, Panx1a^−/−^ lenses, while narrowing the laser beam, failed to focus the light into a sharp focal point (Figure 4D), indicating a disruption in lens focusing ability. These results suggest that Panx1a^−/−^ lenses are functionally compromised, likely leading to impaired visual performance, and motivated us to pursue quantitative analysis using OCT.Figure 4. Panx1a mutant lenses have abnormal lens texture and focusing power(A and B) Macroscopic comparison of Panx1a^+/+^ and Panx1a^−/−^ eyes under the same lighting conditions.(C) shows the light focus pattern through a control Panx1a^+/+^ lens with the black arrow highlighting the 140 μm lens working distance.(D) Illustrates the laser light passage through a defective Panx1a^−/−^ lens, where no distinct focal point is observed.(E) demonstrates how the MATLAB code selects a region of interest (ROI) from the lens and calculates the full width at half maximum (FWHM) and mean intensity to analyze lens texture (scale bars, 500 μm).(F and G) Histograms of representative healthy Panx1a^+/+^ lens and a mutant lens.(H–K) Estimation plots representing the FWHM and mean intensity in both groups, revealing that both FWHM and mean intensity were significantly higher (p < 0.05) in the mutant group at 8-month-old and 1-year-old, indicating notable differences in the texture of healthy and Panx1a^−/−^ lenses, such as cataracts.(L and M) Histology images of Panx1a^+/+^ and Panx1a^−/−^ lens with different textures. Statistical analysis: Welch’s t test; Significance ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Scale bars, 500 μm.

To quantitatively assess structural abnormalities in Panx1a^−/−^ lenses, we performed OCT texture analysis on defined regions of interest (ROIs, e.g., yellow rectangle in Figure 4E) across multiple age groups (8 months, 1 year, and 1.5 years), with 30 lenses analyzed per genotype per age group. Histograms of pixel intensity within each ROI were generated, and two metrics were extracted: mean intensity and full width at half maximum (FWHM) of distributions. Mean intensity value serves as an indicator of the average scattering of light by lens tissue. FWHM values, on the other hand, serve as an indicator of textural heterogeneity; higher values reflect increased disturbance and non-homogeneity, whereas lower values indicate more uniform, organized tissue microstructure.

Qualitative analysis of the extracted texture data revealed that Panx1a^+/+^ lenses exhibited narrow histograms with low mean intensities and FWHM values, indicating minimal and uniform optical scattering as expected from a homogenous internal lens microstructure. In contrast, Panx1a^−/−^ lenses showed significantly broader histograms with elevated mean intensities compared to those of Panx1a^+/+^ lenses, suggesting more scattering and heterogeneous microstructures in the mutants (Figures 4F vs. 4G). Quantitative analysis (Figures 4H–4K) confirmed significantly larger mean intensities for Panx1a^−/−^ groups compared to the Panx1a^+/+^ groups at both 8-month (p < 0.0001) and 1-year (p < 0.001) developmental stages (93.85 ± 7.31 vs. 83.03 ± 7.82 and 98.15 ± 7.4 vs. 90.1 ± 4.08, respectively). Similarly, FWHM values were significantly different at both 8 months (Panx1a^−/−^: 29.1 ± 6.51, Panx1a^+/+^: 22.19 ± 3.56, p < 0.0001) and 1 year (Panx1a^−/−^: 34.7 ± 7.5, Panx1a^+/+^: 26.2 ± 4.9, p < 0.0001).

To shed light on the origins of the observed OCT textural differences, we performed histological analysis using H&E staining. In Panx1a^+/+^ lenses, we observed well-organized concentric rings extending from the core to the epithelium, with no nucleated cells present within the lens body (Figure 4L). In contrast, Panx1a^−/−^ lenses exhibited disrupted ring organization. That is, while the organized rings were detectable in the lens core, they abruptly ceased toward the periphery, which also stained lighter and lacked the structural definition observed in controls (Figure 4M). These histological disruptions are consistent with the increased textural heterogeneity and optical scattering identified in the OCT analyses.

Pannexin 1a deletion induces retinal structural deficits and ganglion cell layer loss revealed by multimodal imaging

Previous studies have correlated Panx1 with glaucoma, characterized by retinal thinning and retinal ganglion cell (RGC) loss.13^,^14 Given these prior reports, we performed immunohistochemistry using adult Panx1a^+/+^ and mutant zebrafish retina to identify differences in the localization of the Panx1a protein (green expressions in Figures 5A and 5B). In Panx1a^+/+^, Panx1a was localized in the horizontal cell layer (HC) and the ON/OFF ganglion cell layer. Conversely, in Panx1a knockout (Panel B) Panx1a protein was absent from the retina, particularly within the ganglion cells of the Ganglion Cell Layer (GCL). In addition to the loss of Panx1a protein, qualitatively, the mutant inner nuclear layer exhibited a noticeably sparser distribution of nuclei, aligning with the retinal thinning observed in our quantitative OCT measurements.Figure 5. Retinal immunohistochemistry and OCT image of adult zebrafishPanels (A) and (B) present the expression and localization of the Panx1a (in green) antibody throughout the retinal layer. The Panx1a is localized in the horizontal cell layer (HC) and in the ON/OFF ganglion cell layer, as it is shown in panel (a1). The Panx1a-KO served as the negative control (scale bars, 25 μm). Panel (C) is a B-mode OCT image of the TL zebrafish retina obtained using our modified 880 nm OCT system and zebrafish retina probe, and Panel (D) is a B-mode retina OCT of Panx1a−/−. The red square in Panel (C) indicates the region selected by the custom MATLAB code for the analysis and measurement of each layer's thickness. Panel (E) shows the fitted plot of the retinal layers. The green line shows the thickness of the GCL layer, and the yellow line presents the whole retina thickness. Panels (F and G) present estimation plots of whole-retina thickness and ganglion cell layer (GCL) thickness in the 8-month-old group, demonstrating a significant decrease in both measurements (n = 30 for Panx1a^+/+^ and KO, Statistical analysis: Welch’s t test; p < 0.05). Panel (H) is showcasing an en-face OCT image of Panx1a^+/+^ retina. Panel (I) is an en-face OCT image of Panx1a^−/−^ retinal vessels, with the red rectangle highlighting the "tortuous vessel" phenotype observed in a few Panx1a^−/−^ eyes. Statistical analysis: Welch’s t test; Significance ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Scale bars, 500 μm.

To determine whether the absence of Panx1a is associated with anatomical remodeling at the tissue level, we optimized our 880 nm OCT system equipped with a zebrafish retinal probe lens to investigate in vivo retinal structure and geometries (Figure 5C). Acquired volumetric datasets were analyzed to extract the thicknesses of the retina and the GCL layer (Figure 5D). The retina of 8-month-old Panx1a^+/+^ and Panx1a^−/−^ zebrafish (N = 30 per genotype) showed changes in both the retina and GCL layer thicknesses. In the Panx1a^−/−^ mutants the retina and the GCL layer thickness were significantly reduced when compared to Panx1a^+/+^ controls (Retinal thickness, Panx1a^−/−^: 56.3 ± 5.7 pixels, Panx1a^+/+^: 62.2 ± 3.6 pixels, p < 0.0001) (GCL layer thickness, Panx1a^−/−^: 11.6 ± 4.07 pixels, Panx1a^+/+^: 17.4 ± 4.08 pixels, p < 0.0001) (Figures 5E and 5F). We also analyzed the volumetric images in the en-face orientation, which indicated the presence of tortuous blood vessels in a few of the mutants (N = 3), all of which had defective lenses in the same eye (Figure 5H). We did not find such a pattern in the retina en face images of any of the Panx1a^+/+^ group fish (Figure 5G).

Discussion

This study presented the first comprehensive analysis of the structural and functional consequences of Pannexin1a (Panx1a) deletion in the adult zebrafish eye. Our findings established Panx1a as a critical regulator of ocular development and function, affecting multiple anatomical compartments of the eye, including the retina, lens, and axial dimensions. The phenotypic spectrum we observed, ranging from visual behavior deficits to myopia, lens structural defects and opacification, and retinal thinning, highlighted a pleiotropic role for Panx1a in maintaining ocular integrity (Figure 6).Figure 6. Graphical summary of the study and tested hypothesisInitial OKR analysis in larvae revealed ocular motor deficiencies in Panx1a^−/−^ zebrafish, leading to the hypothesis that Panx1a deletion causes structural and/or functional deficits in the eye. Subsequent analyses in adult control and mutant zebrafish confirmed that Panx1a mutation affects emmetropization, induces RGC loss, and produces cataract-like changes/structural defects in the lens, thereby verifying the hypothesis and validating the understanding that Pannexin1a regulates visual system development and ocular integrity in zebrafish.

Visual behavior and sensorimotor dysfunction in pannexin 1a-Deficient zebrafish

Our previous work demonstrated that Panx1a^−/−^ larvae exhibit altered visual-motor behavior and dopaminergic signaling, as assessed by visual motor response (VMR) analysis.17 In zebrafish, VMR and optokinetic response (OKR) assays probe distinct, yet complementary, aspects of visual processing. VMR measures gross locomotor responses to abrupt changes in illumination, making it a robust and high-throughput approach for detecting broad visual deficits.39^,^40 In contrast, OKR evaluates reflexive eye movements in response to moving visual stimuli, enabling the assessment of image-forming ocular structures and pathways and fine-scale functional impairments. Together, these assays provide a comprehensive framework for investigating visual system development and function.33^,^41^,^42

In zebrafish, the OKR is mediated by a well-defined visuomotor pathway that begins with retinal image detection and projects to the pretectal nuclei and tegmental regions of the midbrain before descending to the oculomotor nuclei to drive compensatory eye movements. The pretectum, particularly the nucleus of the optic tract (NOT), plays a central role in motion detection and direction selectivity, while downstream premotor neurons in the tegmentum transform retinal slip signals into motor commands for the extraocular muscles.43^,^44 Electrophysiological studies have demonstrated that deficits in any component of this circuit, including reduced retinal contrast sensitivity, altered pretectal activation, or disrupted tegmental firing, can impair OKR gain and saccade timing.45

Using OKR assays, we observed that Panx1a^−/−^ larvae exhibited significant deficits in visual behavior. These included prolonged saccade intervals and reduced contrast sensitivity, suggesting impairments in both photoreceptor function and central sensorimotor processing. Previous studies have localized Panx1a expression to the inner plexiform layer of the retina and visual midbrain regions in zebrafish, implicating its involvement in retinal signal integration and dopaminergic neuromodulation.10^,^17 The localization of Panx1a to the inner plexiform layer and midbrain regions suggests a role in retinal signal transduction and central visual pathway integration.12^,^16 These observations are consistent with a broader role for pannexin channels in modulating visual-motor coordination via ATP release and purinergic signaling pathways. Furthermore, our behavioral data parallel those from other Panx1-deficient models, including rodents, in which altered visual reflexes and electroretinographic responses have been reported.13 This supports the evolutionary conservation of Panx1-mediated neurosensory regulation across vertebrate species. Taken together, the OKR anomalies observed at the larval stage, along with histological evidence of abundant Panx1a expression in the zebrafish eye, led to the hypothesis that Panx1a deletion causes defects specifically affecting image-forming ocular structures. To test this, control and mutant zebrafish at three adult developmental stages (8, 12, and 18 months) were examined using whole-eye OCT, high-resolution retinal OCT, histology, and immunohistochemistry. These analyses confirmed that Panx1a deletion is associated with systemic ocular structural and functional deficiencies.

Axial myopia and refractive shift

Our quantitative biometric analyses revealed a significant elongation of axial length and a corresponding negative shift in relative refractive error (RRE) in Panx1a^−/−^ zebrafish, particularly from 8 months to 1 year of age. These changes are indicative of progressive axial myopia, a condition associated with excessive eye growth and image defocus on the retina.46 The reduction in the lens-to-axial length ratio in mutants further supports this refractive imbalance, aligning with emmetropization theories where a mismatch between ocular components leads to refractive error.47

While the molecular underpinnings of axial elongation remain incompletely understood, previous work has implicated purinergic and dopaminergic signaling in regulating eye growth.48^,^49 While the molecular mechanisms underlying axial myopia remain incompletely defined, prior studies have implicated dopaminergic and purinergic signaling in the regulation of ocular growth.46^,^48 Given that Panx1 channels are known mediators of ATP release, their absence may disrupt extracellular signaling environments critical for controlling scleral remodeling and choroidal feedback loops involved in eye growth regulation. The deletion may alter signaling cues essential for ocular growth regulation. However, such mechanistic connections warrant further targeted molecular and functional validation in future studies. Although the refractive error in panx1a mutants was statistically significant, the magnitude of the myopic shift was relatively modest. Importantly, the axial elongation and resulting refractive phenotype in panx1a mutants may be partially attributable to lens abnormalities. The pronounced epithelial disruption, internal defects, and increased light scatter seen in the mutant lenses are expected to degrade image quality at the retina. Because retinal image blur is a well-established driver of experience-dependent axial elongation and myopia in fish and other vertebrates, the optical degradation caused by Panx1a^−/−^ lenses may actively contribute to the axial elongation observed in these mutants. In this framework, the modest but significant myopic shift may reflect a downstream consequence of insufficiently sharp retinal images rather than an isolated defect in axial growth regulation. Future studies that experimentally separate lens-induced optical blur from intrinsic axial-growth mechanisms will be essential for determining the extent to which lens pathology drives the refractive changes in Panx1a^−/−^ eyes.

Lens opacification and structural abnormalities

Combined OCT imaging and histological evaluation of Panx1a^−/−^ zebrafish revealed marked disruptions in lens structure, including epithelial abnormalities and increased light scattering. These changes were progressive with age and, in advanced cases, resulted in aphakia. Quantitative texture analysis of OCT images indicated elevated pixel intensity and broader full-width at half maximum (FWHM) values in Panx1a^−/−^ lenses, parameters commonly associated with cataract formation.

These observations are consistent with prior evidence implicating Pannexin 1 in maintaining lens homeostasis. Specifically, Panx1 has been linked to the regulation of epithelial Na^+^/K^+^-ATPase activity and calcium signaling processes that are critical for lens transparency and the organization of fiber cells.18^,^24 In line with this, similar disruptions in lens circulation and clarity have been documented in models of connexin deficiency, suggesting that Panx1 and connexins may participate in overlapping or complementary regulatory pathways.20^,^50

Our OCT analysis confirmed that Panx1a^−/−^ lenses were not only reduced in size but also exhibited distinct internal abnormalities, including localized regions of increased scatter, focal opacification, and discrete voids or lumen-like structures. These features increased in prevalence and severity with age. Correspondingly, histological sections revealed the presence of nucleated cells beneath the lens epithelium in regions where mature, enucleated fiber morphology would normally predominate. While these findings are consistent with an aberration in epithelial organization or differentiation, we cannot determine the precise cellular identity or origin of these structures based solely on OCT and routine histology. Without transmission electron microscopy or molecular markers, interpretations involving epithelial migration or failed fiber maturation remain hypotheses. Future studies employing TEM or targeted immunostaining will be essential to resolve the underlying cellular pathology associated with Panx1a deficiency. Although the mechanisms underlying these structural changes remain to be fully clarified, one possibility is that they represent an early morphological change associated with tissue remodeling. Prior work has implicated Panx1 in regulating cellular processes related to vascular development, including ATP-mediated signaling and modulation of VEGF pathways. While angiogenesis is not a typical feature of lens tissue, the presence of organized luminal structures in the mutant lenses may warrant further investigation into whether similar signaling pathways are aberrantly activated in this context.11^,^51^,^52

Retinal thinning and ganglion cell loss

In adult Panx1a^−/−^ zebrafish, retinal OCT imaging revealed significant thinning of the retinal ganglion cell layer (GCL) and overall retinal thickness, hallmarks of the retinal degeneration often associated with glaucoma.14 Immunohistochemistry confirmed the absence of Panx1a in mutant GCL layers. These structural alterations mirror prior findings in Panx1-deficient rodent models, where GCL dysfunction was observed.13 Given that Panx1 channels respond to mechanical stress and modulate inflammatory signaling, their absence could impair pressure homeostasis and increase susceptibility to retinal ganglion cell loss under stress conditions.11^,^15

Additionally, the observation of vascular tortuosity in some Panx1a^−/−^ retinas raises the possibility of vascular remodeling secondary to lens and retinal degeneration. While the genetic basis of vessel tortuosity includes contributions from angiogenic regulators such as VEGF, its relevance in this context requires further molecular and histological validation.53

Panx1 channels have been proposed to act as mechanosensitive conduits for ATP release, modulating intraocular pressure and glial signaling in response to biomechanical stress.11^,^15 Their deletion may impair pressure adaptation mechanisms, contributing to retinal ganglion cell loss, which manifested in our OCT images as a reduction in the GCL layer thickness. Furthermore, sporadic vascular tortuosity observed in mutant retinas suggests possible remodeling of the ocular vasculature, which could be secondary to hypoxic or inflammatory responses, though this remains speculative without further molecular analysis.53

Broader implications

Our findings establish Panx1a as a critical regulator of ocular structure and function, highlighting its potential relevance to the pathogenesis of multiple human eye diseases. The phenotypes observed in zebrafish, myopia, lens opacification, and ganglion cell loss, are key hallmark features of human refractive errors, cataract, and glaucoma, respectively. Together, these disorders account for most of the global visual impairment, underscoring the translational potential of these findings.

In mammalian systems, Panx1 is highly expressed in the retina, lens epithelium, and ciliary body, where it contributes to aqueous humor dynamics and cellular stress responses. Mouse models have demonstrated that Panx1 dysregulation sensitizes RGCs to oxidative stress, alters visual processing, and modulates immune signaling pathways.13^,^14 The similarity of these phenotypes to those described here in zebrafish supports the conserved role of pannexins in ocular homeostasis.

Limitations of the study

Although our analyses across multiple developmental stages demonstrated that ocular and retinal defects in panx1a knockout zebrafish progressively increased in both frequency and severity with age, several limitations should be considered. First, while the use of wild-type littermates as controls minimizes confounding from genetic background, these comparisons cannot fully exclude indirect effects, such as the activation of compensatory pathways or developmental alterations in non-ocular tissues that may secondarily influence eye morphology.

Second, although the temporal analyses presented here support a critical role for panx1a in both the development and long-term maintenance of ocular integrity, we did not perform functional rescue experiments to directly test causality. The generation of transgenic zebrafish lines expressing panx1a under the control of lens- or retina-specific promoters, followed by the longitudinal assessment of phenotypic rescue, would provide the most definitive evidence.

Third, we did not explicitly examine whether males and females exhibit differential susceptibility to the ocular phenotypes associated with panx1a loss. Subtle sex-specific differences in growth rate, lens aging, or retinal maintenance may contribute to variability in our measurements and should be evaluated in future work through sex-stratified analyses.

Fourth, although all fish were reared under standardized photoperiod and light-intensity conditions, environmental lighting is known to influence ocular growth, refractive development, and retinal physiology in zebrafish. Because we did not test how variations in light intensity or spectral composition interact with panx1a loss, the extent to which our findings generalize to other lighting environments remains limited. Future studies incorporating controlled manipulation of light conditions will help clarify how the photic environment interacts with Panx1a function during eye development.

Finally, our study focused primarily on structural endpoints obtained through ocular imaging. While these provide robust evidence of morphological disruption, they do not fully capture the functional consequences of panx1a loss. Incorporating complementary functional assays, such as electrophysiological recordings or intraocular pressure measurements, would provide a more comprehensive assessment of visual capacity and retinal physiology.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Nima Tabatabaei (nimatab@yorku.ca).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All data are available from Nima Tabatabaei upon reasonable request and signing of institutional materials transfer agreements (MTAs).
• •All original code has been deposited at https://doi.org/10.7910/DVN/ADFBXZ and is publicly available as of the date of publication.
• •Any additional information required to reanalyze the data reported in this article is available from Nima Tabatabaei upon request.

Acknowledgments

We gratefully acknowledge Petia Stefanova of the Histology Laboratory at Sunnybrook Hospital for her expert assistance with histological analyses. We also thank Christiane Zoidl for her support, and Janet Fleites-Medina and Veronica Scavo of the York University Zebrafish Vivarium for their excellent care and husbandry of the zebrafish.

Funding: This study was funded by the 10.13039/501100000038Natural Sciences and Engineering Research Council of Canada (NSERC). Discovery grants RGPIN-2022-04605 (NT) and RGPIN-2019-06378 (GRZ).

Author contributions

Conceptualization: S.S., G.R.Z., and N.T.; methodology: S.S., S.H.T., G.R.Z., and N.T.; investigation: S.S., S.H.T., G.R.Z., and N.T.; visualization: S.S., S.H.T., G.S.O.Z., G.R.Z., and N.T.; supervision: G.R.Z. and N.T.; writing – original draft: S.S., G.R.Z., and N.T.; writing – review and editing: S.S., G.S.O.Z., G.R.Z., and N.T.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

An artificial intelligence tool was utilized during the preparation of this article to support clarity and readability. Specifically, ChatGPT version 4.5 (OpenAI, San Francisco, CA) was employed for grammar correction and writing edits. After using this tool or service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

STAR★Methods

Key resources table

REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesanti-Panx1a (rabbit polyclonal)Custom-madehttps://www.davids-bio.de/ZDB-ATB-100609-1Prochnow et al. 2009Safarian et al. 2020Vroman et al. 2014Kurtenbach et al. 2013Secondary antibody Alexa Fluor 488 conjugateThermo Fisher ScientificCat # A-11094; RRID:AB_221544Biological samplesTuebingen Long Fin (TL)in-house colonyZDB-GENO-990623-2ZFIN line: yku1 (Panx1a-/-)generated in-houseZDB-ALT-210414-16Safarian et al. 2020Deposited dataTexture analysis codeCustom MATLAB codehttps://doi.org/10.7910/DVN/ADFBXZRetina layer detection codeCustom MATLAB codehttps://doi.org/10.7910/DVN/ADFBXZExperimental models: organisms/strainsTuebingen Long Fin (TL)in-house colonyZDB-GENO-990623-2ZFIN line: yku1 (Panx1a-/-)Generated in-houseZDB-ALT-210414-16Safarian et al. 2020Software and algorithmsStytraOpen source, Github(Vilim Stih et al., PLos Computational Biology, 2019https://github.com/portugueslab/stytraGraphPad PrismCommercial, GraphPadhttps://www.graphpad.comMojo Hand softwareCommercialhttp://talendesign.orgImageJ softwareCommercialhttps://imagej.netZEISS ZEN softwareCommercialhttps://www.zeiss.com

Experimental model and study participant details

Zebrafish (Danio rerio) of the Tupfel Long Fin (TL) strain were used to generate the panx1a^−/−^ mutant line. The fish were maintained in a recirculation system (Aquaneering Inc., San Diego, CA) at 28 C on a 14-hour light/10-hour dark cycle. All experiments and procedures involving animals were conducted at York University’s zebrafish vivarium, in accordance with the guidelines of the Canadian Council for Animal Care (CCAC) and were approved by the Animal Care Committee (ACC) under protocol GZ#2014-19 (R3). The number of experiments, including those involving zebrafish larvae, was minimized to the necessary amount. In this study we used adult zebrafish in 4 developmental stages of larvae, 8-month-old, 1-year-old and 1.5-year-old. We used zebrafish larvae for OKR study and for all other tests we humanly used adult zebrafish. Additionally, all other experiments were approved by the York University Biosafety Committee (YUBC) under permit #04-11.

Method details

Generation of Panx1a-/- line

Potential TALEN target sites on the panx1a gene (NM_200916.1) were identified using the Mojo Hand software (http://talendesign.org). Sequence-specific TALEN constructs were assembled using Golden Gate cloning methods. The Panx1a^-/-^ zebrafish line used in this study was generated previously by our group using transcription activator-like effector nuclease (TALEN) technology, as detailed in Safarian et al. published paper.17

Briefly, for generation of the knockout zebrafish, adult founder (F0) zebrafish were anesthetized in pH-buffered 0.2 mg/ml ethyl-3-aminobenzoate methane sulfonate solution (MS-222; Sigma-Aldrich), and a 2 mm segment of the caudal fin was excised for genomic DNA extraction and screening of indel mutations in the panx1a locus. Mutant F0 carriers were subsequently out-crossed with wild-type TL zebrafish to generate F1 offspring, which were genotyped by PCR and AfeI restriction digestion to confirm germline transmission. Verified heterozygous (panx1a^+/-^) F1 fish were in-crossed to establish homozygous F2 mutants, and all experiments in the present work were conducted on progenies of the homozygous F3 generation.

Optical coherence tomography (OCT) imaging

Two OCT imaging systems were developed and optimized for whole-eye imaging (Figure S1) and high-resolution retinal imaging (Figure S2) of zebrafish, respectively.

Whole-eye imaging: Each sample was placed on a XYZ translation stage (Thorlabs MT3-Z8) for imaging. All whole-eye images were collected using our custom-developed spectral-domain optical coherence tomography (SD-OCT) system. The system was built in Michelson configuration, employing a super luminescent laser diode centered at 1,310 nm (± 75 nm at 10 dB; Exalos, Switzerland) and a 2048-pixel line scan camera spectrometer with a maximum acquisition rate of 147 kHz (Wasatch Photonics; United States of America). The details of the system can be found in Figure S1. The axial and lateral resolutions of the system in tissue were measured as 8.5 μm and 10 μm, respectively.

Before imaging age-matched adult zebrafish were humanely euthanized using MS222 and immediately imaged. Zebrafish (n = 15/genotype/age) were placed in a silicon mold to orient the eye toward OCT system’s objective lens. To minimize specular reflections from sample surface, a thin layer of matching index liquid (∼80 μm, n=1.33) was placed over the eye before imaging. We collected 10 volume scans of 2.5x5mm^2^. All volumes were then computationally averaged to yield one despeckled volume from each sample. The despeckled OCT volumes were interrogated in ImageJ software which enabled measurement of various geometrical parameters through a standard calibration process. Because the axial distances measured by OCT represent optical path lengths (OPL) rather than true geometric distances, we corrected for the refractive index of each constituent ocular medium. Specifically, we divided the OPL by the refractive index n of the medium to obtain the true geometric length:

$$[eqn]$$

For the zebrafish lens, we used the peak refractive index value of 1.5 as reported by Wang et al.54 for adult zebrafish lenses. A uniform refractive index was assumed for the aqueous/vitreous compartments (n =1.33), consistent with previous biometry studies in zebrafish.55 After index-correction, we calculated normalized lens-to-axial length ratios and refractive error metrics. In addition, we developed a MATLAB code to analyze our OCT images, specifically for texture analysis of the lens using whole-eye images. A fixed-size square region of interest (ROI) was placed at the geometric center of the lens in each B-scan. The lens core was used as a landmark to ensure consistent sampling across lenses of different sizes and to avoid peripheral artifacts or localized defects. The code selects the same region of interest from the images and calculates the mean intensity and full width at half maximum (FWHM) of histogram of the ROI. This analysis was performed on both 8-month-old and 1-year-old groups (N=30 eyes per genotype per age group).

Retinal imaging: Each sample was placed on an XYZ translation stage (Thorlabs MT3-Z8) for imaging. All retinal images were collected using a modified Lumedica SD-OCT system (Figure S2), employing a super luminescent laser diode centered at 880nm (± 90 nm at 10 dB; Exalos, Switzerland) and a 2048-pixel line scan camera spectrometer with a maximum acquisition rate of 80 kHz. The OCT system was outfitted with a custom retina probe specialized for imaging zebrafish retina. Similar to whole eye imaging, zebrafish (n = 15/genotype) were placed in a silicon mold to orient the eye toward OCT system’s objective lens. To minimize specular reflections from sample surface, a thin layer of matching index liquid (∼80 μm, n=1.33) was placed over the eye before imaging. Despeckling was carried out by averaging 10 volumes from each sample.

For measuring the whole retina and retinal ganglion cell (RGC) layer thickness, we used another custom MATLAB code. This code selects the same square region from the retina in each image. By summing up the values in each row and plotting the results, the code generates a plot with peaks and valleys representing different layers of the retina. Total retinal thickness was calculated from the distance between the first and last major peaks, while RGC layer thickness was defined as the full width at half maximum of the first peak. As our aim was to compare relative differences between genotypes rather than absolute biometric values, this approach ensured robust internal consistency.

Histology

After capturing the OCT images, 3 mutant eyes with proper defects and 2 Panx1a^+/+^ eyes were chosen for histology study. The eyes were surgically removed from the eye cap in a way that optic nerve is visible. Then submerged in 4% PFA overnight and then transferred to ethanol solution for 5 hours. The sample then is positioned in paraffin block so sagittal cuts can be made to be comparable with our B-mode OCT images. The sample then went under histological processing. Each histological slide contained one 5-μm-thick hematoxylin and eosin–stained tissue samples sectioned at an interval of 20 microns per slide to catch the lens defect at its best. Ultimately, we received 50 slides from each eye and imaged them using Nikon Eclipse Ti Inverted Fluorescence microscope with white light at 20x,10x and 4x lens with Tucsen FL20 camera.

Setup for studying the zebrafish lens

We designed a set up using a Nikon Eclipse Ti Inverted microscope, Tucsen FL20 camera, and a 526nm collimated green laser source to study focusing dynamics of light by zebrafish lens (Figure S3). From the selected eyes, the lens was surgically removed without any harm to the lens (Video S1). Next, we place the lens in a glass chamber filled with a dilute water-lipid emulsion to enable visualization of green light propagation in chamber and focus by zebrafish lens.

Video S1. Zebrafish Lens ExtractionProcedure for zebrafish lens extraction without damage to the retina or lens.

Immunohistochemistry

The adult zebrafish was humanly euthanized using MS-222 solution (0.2%, Sigma-Aldrich). Tissues were fixated with 4% Paraformaldehyde (PFA) in 1xPBS for 5 days at 4° fridge, followed by an overnight fixation with 15% sucrose, and another overnight fixation with 30% sucrose the following day. An OCT medium (obtained from VWR clear Frozen Section Compound, cat#95057-838) was used for freezing the tissues at room temperature (RT). A Cryotome FSE machine (serial#CC0220C0904) acquired from ThermoSCIENTIFIC was used to cut 15μ tissue slices. To eliminate unwanted debris, the tissues were placed in 1xPBS medium (obtained from 100ml of 10xPBS, 900ml of distilled water, and 2ml of 0.1% Tween®20). The tissues were washed 3x for 5minutes for a total of 15minutes with 1xPBS. The tissues were then placed in a moist chamber, and an antibody dilution blocker was added to each slice. Tissues were left at RT for an hour. This was followed by another 3x wash with 1xPBS and then placed in the wet chamber and the primary antibody was applied onto the tissues and left at 4° fridge for 2 hours. Afterwards, another 3x wash with 1xPBS and then the secondary antibody was applied. Slides were kept in the dark at RT for an hour. This was followed by the last 3x wash with 1xPBS and one wash with distilled water to wash out the ions. 10μl DAPI is applied onto the tissues to stain the nuclei and microscopic cover glass (18x18mm) was placed onto the tissue slides. The LSM 700 laser scanning microscope was used to visualize the staining and ZEISS ZEN software was used to capture the high-resolution images (2048x2048 pixel).

OKR

A custom-made OKR setup was built with a 55mm/F2.0 59873 lens (acquired from EDMUND OPTICS) to investigate the 6dpf larvae’s eye movement. The videos were captured at 40fps and recorded at 60-second intervals. A constant speed of 0.01 degree/second was utilized to evoke an OKR response. The OKR response was captured at spatial frequency 0.2 cycle/degree and the larvae were exposed to two contrasts (20% and 100%). The Python code for the moving grating stimulus was created using PsychoPy (https://www.psychopy.org/). For OKR, the 6dpf larvae were used for the experiment, and in each run, one larva was placed in a plate and was immobilized with 6% methylcellulose under the OKR camera. A sample size of 12 was used for each condition. The videos acquired from the OKR experiment were analyzed using Stytra (acquired from https://portugueslab.com/stytra/) to acquire an Excel file as an output and investigate the precision and efficacy of the larvae’s eye movement patterns and visual tracking behavior.

The larva’s eye movement patterns were analyzed using a custom Python script to detect saccades. The script removed eye movement under a threshold of 15° and isolated saccades greater than 15° (aka. positive responses). The rightward-moving grating stimulus was analyzed. The parameters saccade amplitude, duration, responsiveness to the stimulus, and time between saccades were analyzed at the highest (100%) and lowest (20%) contrast settings in both Panx1a^+/+^ and Panx1a^-/-^. The custom Python script will be shared upon reasonable request. Raw traces were smoothed using GraphPad Prism. Significance was determined using Welch’s t-test.

Quantification and statistical analysis

For all measurements, data visualization was performed using violin plots, followed by normality tests to assess the distribution of the data. For parameters exhibiting a normal Gaussian distribution, a two-tailed, two-sample Student's t-test (Welch t-test) was conducted to compare the Panx1a^+/+^ and Panx1a^-/-^ groups. For parameters that did not follow a normal distribution, non-parametric tests such as the Mann-Whitney test were used. Each group per age and genotype included a sample size of 30 eyes from 15 different fish. Complete results of the statistical analyses of geometrical eye parameters extracted from OCT tomograms are depicted in Table S1. For the OKR analysis a sample size of 12 larvae was used. Statistical significance was determined with p-values less than 0.05 and significance is defined as ∗ p<0.05, ∗∗ p<0.01, ∗∗∗ p<0.001, ∗∗∗∗ p<0.0001. All statistical analyses and plots were performed using GraphPad Prism 10.