To investigate how oxidative stress pathways modulate αB-crystallin expression and/or function, we utilized sensitized zebrafish lines bearing a mutation within nrf2 that substantially reduces its activity (Mukaigasa et al., 2012). The nrf2 ^fh318 zebrafish line has a mutation in the basic region of the DNA-binding domain (exon 5), which leads to decreased transcriptional induction of Nrf2 target genes, which include antioxidant enzymes. We observed that the mRNA level of nrf2 downstream genes such as gstp1 and prdx1 were lower in the lens and heart tissues of nrf2 ^fh318/fh318 lines relative to WT fish (Supplementary Figure S1). These results are consistent with previous studies, although ours were carried out in the absence of oxidative stress (Mukaigasa et al., 2012). We also confirmed that the level of αBa-crystallin and αBb-crystallin transcripts was relatively higher in the lens than was in the heart and brain tissues, whereas the levels of nrf2 transcripts in the heart and brain were higher than that in the lens (Supplementary Figure S2).

To test if αB-crystallin expression is affected by Nrf2 deficiency, mRNA levels of αBa-crystallin and αBb-crystallin genes (hereafter, we use the nomenclature “cryaba” and “cryabb” for simplicity) were compared between WT and nrf2 ^fh318/fh318 embryos by the qRT-PCR analysis. Although the levels of cryaba transcript were not changed between WT and nrf2 ^fh318/fh318, cryabb transcription was highly upregulated in Nrf2-deficient embryos (Figure 1A). This modulation of cryabb transcripts appears tissue specific. cryaba transcripts were not substantially changed in the whole eyes, lens, heart, and brain tissues of nrf2-mutated zebrafish relative to WT (Figure 1B). By contrast, cryabb transcripts were upregulated strongly in the heart and brain tissues of Nrf2-deficient zebrafish (Figure 1C; Supplementary Figure S3). To test whether transcription of cryabb can be upregulated in response to increased oxidative stress, we measured the level of cryabb mRNA after treating WT embryos with 800 μM tert-Butyl hydroperoxide (tBHP) for 2 h at 4 days post fertilization (dpf). We observed an increase in the cryabb transcript (∼1.5 fold) with tBHP treatment (Supplementary Figure S4B), although to a lesser extent when compared to the reduced activity of Nrf2.

To investigate crosstalk between αB-crystallin and nrf2, we assessed the regulation of nrf2 and nrf2 downstream targets, such as antioxidant genes, in αB-crystallin knockout lines (cryaba ^−/−, cryabb ^−/−, cryaba ^−/−, and cryabb ^−/− double mutants). Although the level of nrf2 mRNA was not significantly changed in WT and αB-crystallin knockout lines, its targets, gpx1a and prdx1, were upregulated in cryabb ^−/− embryos (Supplementary Figure S5). Taken together, these results suggested a transcriptional link between nrf2 and cryabb but not cryaba. Further detailed experiments are required to examine the tissue-specific mechanisms underlying the upregulated antioxidant genes in the cryabb ^−/− lines.

In light of the aforementioned transcriptional relationship, we explored its consequences on the phenotypic manifestation of the loss of αB-crystallin function. Previously, we showed that knockout of either αB-crystallin causes lens defects in zebrafish embryos (Mishra et al., 2018; Zou et al., 2015). αB-crystallin mutant embryos, cryaba ^−/− and cryabb ^−/−, presented lens abnormalities at 4 dpf (Figure 2A). The phenotypic features characterizing these lines appeared as round puncta spread across the lens. These puncta lead to opacity and changes in light scattering. Consistent with our previous results, ∼50% and ∼30% of cryaba ^−/− and cryabb ^−/−embryos exhibited lens abnormalities, respectively (Supplementary Figure S6). In comparison, nrf2 ^fh318/fh318 showed a slightly increased percentage of lens defects (∼20%) relative to WT embryos (∼10%) (Figure 2B).

To investigate how Nrf2 deficiency affects lens development in the absence of αB-crystallin, we generated αB-crystallin/Nrf2 double knockout lines (cryaba ^+/−; nrf2 ^fh318/fh318, cryaba ^−/−; nrf2 ^fh318/fh318, cryabb ^−/−; nrf2 ^fh318/+, and cryabb ^−/−; nrf2 ^fh318/fh318). Then, we compared the penetrance of lens abnormalities in the progeny of WT, nrf2 ^fh318/fh318, cryaba ^−/− and cryaba ^−/−; nrf2 ^fh318/fh318 embryos. Strikingly, we found that the elevated level of lens defects in the cryaba ^−/− lines was suppressed in the cryaba ^−/− /nrf2 ^fh318/fh318 (Figure 3B, left panel). To confirm this result, we in-crossed cryaba ^+/−; nrf2 ^fh318/fh318 adult zebrafish and screened their embryos for lens defects in a genotype-blinded experimental protocol. After the lenses were imaged and sorted into normal and defective lens groups (Figure 2A), we categorized the cryaba genotype of individual embryos into three groups (cryaba ^+/+; nrf2 ^fh318/fh318, cryaba ^+/−; nrf2 ^fh318/fh318, and cryaba ^−/−; nrf2 ^fh318/fh318). The percentage of lens defects of cryaba ^−/−; nrf2 ^fh318/fh318 embryos is similar to the level of nrf2 ^fh318/fh318 (Figure 3B, right panel). The results are consistent with the conclusion that Nrf2 deficiency reduces the penetrance of the lens phenotype induced by αBa-crystallin loss of function. Moreover, among embryos with lens defects, the percentage of severe lens defects of cryaba ^−/− -mutated embryos was higher than that of cryaba ^−/−; nrf2 ^fh318/fh318, indicating that nrf2 deficiency alleviates the severity of lens defects (Figure 3B, bottom panel). By contrast, we did not observe large changes in the percentage of lens defects between cryabb ^−/− and cryabb ^−/−; nrf2 ^fh318/fh318 embryos, suggesting that the lens defects in cryabb KO lines cannot be rescued by abrogation of Nrf2 function (Figure 2C).

To identify the molecular pathway(s) mediating the suppression of lens defects in the cryaba mutant by Nrf2 deficiency, we performed a high-throughput transcriptome profiling on adult zebrafish lens tissues from the WT, cryaba ^−/−, nrf2 ^fh318/fh318, and cryaba ^−/−; nrf2 ^fh318/fh318 lines. Differentially expressed (DE) genes between the groups were determined using Illumina’s DRAGEN pipeline. Then, the DE genes were processed using the Ingenuity Pathway Analysis (IPA) to define the significantly regulated signaling pathways (Kramer et al., 2014). Interestingly, cholesterol biosynthesis was identified as the primary upregulated pathway in the lens of cryaba ^−/−; nrf2 ^fh318/fh318 when compared to that of WT or cryaba ^−/− (Figure 3A; Supplementary Figure S7B). mRNA of the gene-encoding enzymes involved in the superpathway of cholesterol biosynthesis—cyp51, dhcr24, ebp, lss, msmo1, sqlea, hmgcra, and hmgcs1—were upregulated in the lens of the cryaba ^−/−; nrf2 ^fh318/fh318 line when compared to that of the WT (Figures 3B, C).

Based on the RNA-seq data, we hypothesized a correlation between activated cholesterol biosynthesis and the alleviated lens defects in the cryaba ^−/−; nrf2 ^fh318/fh318 line. To test this hypothesis, we investigated whether the lens of cryaba ^−/−; nrf2 ^fh318/fh318 embryos exhibits lower tolerance to treatment with statins, which are competitive inhibitors of the HMG-CoA-R enzyme, an early rate-limiting step in cholesterol synthesis (Istvan and Deisenhofer, 2001). For this purpose, the cryaba ^−/−; nrf2 ^fh318/fh318 embryos were challenged with two different statins, atorvastatin and lovastatin, to inhibit the sterol biosynthetic pathway. cryaba ^−/−; nrf2 ^fh318/fh318 embryos, incubated with 5 μM atorvastatin from 1–4 dpf, showed increased lens abnormalities (Figure 4A). By contrast, vehicle-treated cryaba ^−/−; nrf2 ^fh318/fh318 (DMSO control) or WT embryos incubated with the same regimen of atorvastatin remained predominantly normal. In addition, we noticed that atorvastatin-exposed cryaba ^−/−; nrf2 ^fh318/fh318 embryos have significantly reduced lens area than the vehicle-treated control (Figure 4B). Similarly, cryaba ^−/−; nrf2 ^fh318/fh318 embryos displayed a higher percentage of lens defects than the WT in response to 4 μM lovastatin (Figure 4C). Together, these results suggest that upregulation of the cholesterol synthesis pathway in the lens is a plausible mechanism involved in the suppression of lens abnormalities in cryaba ^−/−; nrf2 ^fh318/fh318.

In addition to lens defects, the loss of αB-crystallin function is associated with a cardiac phenotype that presents as embryonic cardiac edema (Mishra et al., 2018). The penetrance of this phenotype increases in response to stress induced by exposure to external glucocorticoid receptor agonists such as dexamethasone (Dex), as was previously described by Mishra et al. (2018). Because it has been reported that Nrf2 has a protective role in cardiac cells under oxidative stress (Chen and Maltagliati, 2018), we examined whether stress-induced heart edema of αB-crystallin KO lines is modulated by Nrf2 deficiency. For this purpose, we compared the heart areas (see the Methods section) of zebrafish embryos derived from the loss-of-function lines of αB-crystallins, Nrf2, and crosses of cryaba ^−/−; nrf2 ^fh318/fh318 and cryabb ^−/−; nrf2 ^fh318/fh318. All single KO lines and nrf2 ^fh318/fh318, treated with 50 μM dexamethasone (Dex) from 1–4 dpf, developed pericardial edema manifested by increased heart area relative to the WT embryos (Figures 5A–C). Remarkably, we observed a large increase in the penetrance of the phenotype in the cryaba ^−/−; nrf2 ^fh318/fh318 line that was further accentuated in the presence of Dex (Figures 5B, D). Relative to cryaba ^−/−; nrf2 ^fh318/fh318, cardiac edema was blunted in cryabb ^−/−; nrf2 ^fh318/fh318 embryos treated with Dex. Furthermore, the vehicle-treated cryabb ^−/−; nrf2 ^fh318/fh318 group retained a WT-like distribution (Figure 5C).

Because the discovery of cryabb mRNA upregulation in Nrf2-deficient heart tissues (Figure 1C) suggests that cryabb is transcriptionally activated in response to oxidative stress, we further explored the level of cryabb transcript in the presence of Dex. The transcription of cryabb mRNA appears to be upregulated with Dex treatment in WT (∼2-fold) and cryaba KO (∼3-fold) at 4 dpf (Figure 5E), although the p-values for the two-way ANOVA were higher than 0.05. Furthermore, nrf2 ^fh318/fh318 and cryaba ^−/−; nrf2 ^fh318/fh318 embryos showed strongly upregulated cryabb mRNA regardless of the presence or absence of Dex (Figure 5E). Taken together, these results suggest that the two zebrafish αB-crystallin orthologues, cryaba and cryabb, have different regulatory mechanisms in the heart in response to oxidative stress. Our data suggest that cryabb functions as a stress–response gene regulated by both glucocorticoid stress and oxidative stress. Further detailed experiments are required to pinpoint the mechanism of the transcriptional control cryabb under different forms of stress.

To gain mechanistic insight into the origin of the heart edema phenotype, we performed transcriptome analysis of heart tissues from adult zebrafish comparing WT, cryaba ^−/−, nrf2 ^fh318/fh318, and cryaba ^−/−; nrf2 ^fh318/fh318. RNA-seq analysis identified DE genes in the heart tissue of Nrf2-deficient zebrafish (nrf2 ^fh318/fh318 and cryaba ^−/−; nrf2 ^fh318/fh318) that includes upregulation of cryabb transcripts, which is consistent with our data using the qRT-PCR analysis (Figure 1C, Supplementary Figure S8F). To derive a global understanding on how these genes affect heart health, the DE genes with a false discovery rate (FDR) cut off ≤ 0.05 between groups were used for further pathway analysis. The most significant gene ontology (GO) was calculated by using the WEB-based GEne SeT AnaLysis Toolkit (WebGestalt) (Wang et al., 2013). The top important GO terms between cryaba ^−/−; nrf2 ^fh318/fh318 and WT included the extracellular region, supermolecular fiber, and bicellular tight junction (Figure 6A). The extracellular region and tight junction pathways were also listed in the most enriched GO cluster between cryaba ^−/−; nrf2 ^fh318/fh318 versus cryaba ^−/− (Figure 6B). These findings suggest a significant alteration of genes related to extracellular interactions in the heart of cryaba ^−/−; nrf2 ^fh318/fh318 embryos.

To identify the extracellular components which are exclusively changed in the heart tissue of cryaba ^−/−; nrf2 ^fh318/fh318, the subset of genes in the GO term within the extracellular region from each comparison (cryaba ^−/−; nrf2 ^fh318/fh318 vs. WT, and cryaba ^−/−; nrf2 ^fh318/fh318 vs. cryaba ^−/−) was sorted by a Venn-diagram analysis (Supplementary Figure S9; Supplementary Tables S3, S4). The result identified cpa5, cstba, cpa2, anxa1c, hapln1a, ptx3a, rspo2, wnt7aa, and acana genes, which are closely involved in the degradation, remodeling, or inflammation of the extracellular matrix (ECM) (Figure 6C). In addition, the transcripts which encode components in the tight junction, such as cldn7b, cldn3, cldnb, oclna, cagnb, and marveld2b, showed higher expressions in the heart tissues of cryaba ^−/−; nrf2 ^fh318/fh318 than cryaba ^−/− (Figure 6D; Supplementary Figure S10). These results suggest that molecular changes of the ECM and tight junction pathways in the heart in cryaba ^−/−; nrf2 ^fh318/fh318 correlate with greater penetrance of the cardiac edema phenotype.

Finally, we performed Disease Ontology (DO) analysis with known disease markers to explore the potential pathological mechanism underlying enhanced heart edema in cryaba ^−/−; nrf2 ^fh318/fh318. For this purpose, the DE genes between cryaba ^−/−; nrf2 ^fh318/fh318 and WT in the heart tissues were analyzed using DOSE (Yu et al., 2015) to calculate the enrichment of DO terms. The results confirmed that the transcriptional changes in the heart of cryaba ^−/−; nrf2 ^fh318/fh318 are significantly associated with muscle diseases, specifically cardiomyopathy (Figures 6E, F). We also noticed that the human homologs of transcripts in the cardiomyopathy DO are functionally clustered via STRING, a database of protein–protein interactions (Szklarczyk et al., 2021) (Figure 6G). This result further supports the transcriptional changes in cryaba ^−/−; nrf2 ^fh318/fh318 being functionally associated with heart health.

Despite their ubiquitous expression, the physiological roles of sHSPs beyond chaperone activity continue to be enigmatic. While they have been implicated in numerous stress responses, there has been a lack of systematic investigation of how they are coupled to cellular protective pathways. To address this outstanding question, we have utilized zebrafish as a tractable model system to dissect the role of αB-crystallin both in the lens and other tissues. In a previous work, we demonstrated that αB-crystallins are critical for lens development and that the loss of function of αB-crystallins compromises the resistance of the heart to stress (Mishra et al., 2018; Zou et al., 2015). Here, we explored the role of the two αB-crystallins in response to oxidative stress signaling. The finding of tissue-specific transcriptional link between nrf2 and cryabb but not cryaba in zebrafish is novel and consistent with the recent findings that Nrf2 can function as a repressor (Brown et al., 2008; Liu et al., 2018). Equally novel is the identification of molecular pathways modulated in loss-of-function zebrafish lines. The results presented here suggest that cryabb plays a role in the stress response in the brain and heart tissues, but cryaba mainly has a function limited to the lens in zebrafish, most likely as a chaperone. Overall, our results set the stage for further studies on how nrf2 and cryabb are coupled in the zebrafish lens and heart. In both tissues, long-lived cells experience sustained oxidative loads that lead to a number of pathologies.

Human αB-crystallin is found in the mitochondria of both the lens and retinal cells (McGreal et al., 2012). Specifically, αB-crystallin interacts with cytochrome c both in vitro and in vivo, and its overexpression helps maintain mitochondrial membrane potential during oxidative stress (McGreal et al., 2012). Thus, exploring mitochondria changes in zebrafish αB-crystallin loss of function under stress of Nrf2 deficiency or other oxidative-challenged conditions will be critical.

A major finding of this work is that lens defects in the cryaba ^−/− line were rescued in the cryaba ^−/−; nrf2 ^fh318/fh318 line (Figure 2). Furthermore, RNA-seq analysis correlated this rescue with the upregulation of the cholesterol biosynthesis pathway (Figure 3). Following up on the hypothesis of a connection between cholesterol biosynthesis and alleviated lens defects in cryaba ^−/−; nrf2 ^fh318/fh318, we utilized statins and HMG-CoA reductase inhibitors to lower cholesterol synthesis. We determined that treatment with atorvastatin or lovastatin increased the penetrance of lens defect in cryaba ^−/−; nrf2 ^fh318/fh318.

The beneficial role of high cholesterol in the lens has been described and specifically linked to a reduction in oxygen transport across membranes to protect against oxidative damage (Dotson et al., 2017; Widomska et al., 2017; Widomska and Subczynski, 2019). Moreover, elevating cholesterol precursors, particularly lanosterol, has been studied as a means to reverse protein aggregation in cataract (Widomska et al., 2017; Kang et al., 2018; Widomska and Subczynski, 2019; Miyashita et al., 2022). In this context, a genetic mutation in lanosterol synthase (lss) is associated with cholesterol deficiency–associated cataracts (Mori et al., 2006). However, the precise molecular mechanism linking cholesterol biosynthesis and cataract remains largely unknown. Here, we provide an additional connection between cholesterol and lens integrity using high-throughput transcriptome profiling. Given that lens RNA-seq analysis revealed upregulation of cholesterol biosynthetic enzymes, we propose that lathosterol is likely the upregulated sterol in the lens, although the means by which Nrf2 deficiency upregulates cholesterol biosynthetic enzymes is still unclear. Thus, further experiments are required to delineate the intersection between the cholesterol synthesis pathway and Nrf2 in the lens.

Our results are consistent with previous reports that have uncovered Nrf2-dependent regulation of cholesterol biosynthesis in mouse liver and in cultured liver cells. A pharmacogenomics investigation of the Nrf2 activator 3H-1,2-dithiole-3-thione (D3T) in mice showed RNA enrichment of enzymes to be involved in cholesterol biosynthesis, whereas no effect on these enzymes was observed as a consequence of treatment with the Nrf2 activator 1-[2-cyano-3,12-dioxooleana-1,9 (11)-dien-28-oyl]-imidazole (CDDO-Im) (Wible et al., 2018). By contrast, a dietary supplement found to activate Nrf2 downregulated the cholesterol biosynthesis pathway (Hybertson et al., 2022). The contrasting results notwithstanding, these studies confirm the relationship between Nrf2 and the cholesterol biosynthesis pathway and point to the need for a more thorough understanding of how Nrf2 regulates this pathway.

Zebrafish is a popular model for studies on embryonic development, brain function, and disease progression. This is due to its many advantages, such as rapid development, easy-to-manipulate genetics, suitability for live imaging, and chemical screening. However, there are limitations in analyzing tissue-specific transcriptomics or proteomics at the embryonic stage. Even with the use of single-cell RNA sequencing technology, it is difficult to detect low-expressed transcripts. In this study, we conducted screening of embryo phenotypes, such as lens defects and heart edema. We also performed tissue-specific transcriptome analyses on the lens or heart tissues from adult zebrafish. However, phenotype screening was carried out on embryos, and pathway analysis was conducted in adult tissues. Thus, further experiments, such as histological analysis of lens or heart tissues from adult cryaba ^−/− zebrafish, are required to better understand the molecular and cellular pathways affected as a result of αB-crystallin loss of function in an oxidative stress context of nrf2 deficiency.

In contrast to the lens phenotype, the fraction of embryos displaying stress-induced heart edema in cryaba ^−/− was increased by the loss of Nrf2 function. More importantly, a higher penetrance of the heart defect phenotype was observed in cryaba ^−/−; nrf2 ^fh318/fh318 embryos, even without Dex treatment (Figure 5B). Furthermore, the search for disease-associated molecular pathways revealed cardiomyopathy as the top DO cluster in cryaba ^−/−; nrf2 ^fh318/fh318 adult zebrafish heart, suggesting a detrimental synergetic effect of the depletion of nrf2 and cryaba in heart tissue.

While more studies are required to mechanistically understand the link between these various pathways, our findings can be included in the extensive literature that implicates αB-crystallin in the oxidative balance of cardiomyocytes in mouse lines as well as in humans (Rajasekaran et al., 2007; Dimauro et al., 2018; Yin et al., 2019). First identified in investigations of the human cardiomyopathy mutant R120G of αB-crystallin (Fardeau et al., 1978; Vicart et al., 1998), a link between αB-crystallin and reductive stress has been elaborated in subsequent studies (Rajasekaran et al., 2007; Bauersachs and Widder, 2010; Xie et al., 2013; Morrow et al., 2015). Zebrafish is an ideal model to investigate this link further.

AB wild-type strain zebrafish (Danio rerio) were used. The embryos were obtained by natural spawning and raised at 28.5°C on a 14:10 h light/dark cycle in egg water 30 mg/L instant ocean in deionized water containing 0.003% PTU (w/v) to prevent pigment formation. Embryos were staged according to their ages (in dpf). The following mutant and transgenic fish lines were used: cryaba ^vu612 (cryαba ^−/−); cryabb ^vu613 (cryαbb ^−/−); nrf2 ^fh318. All animal procedures were approved by the Vanderbilt University Institutional Animal Care and Use Committee.

Zebrafish were euthanized, and the lens, heart, and brain tissues were dissected as described by Gupta and Mullins (2010). Tissues were immediately snap-frozen in liquid nitrogen, and RNA was extracted from all samples simultaneously using TRIzol (Invitrogen) and RNA Clean & Concentrator Kit (Zymo Research). A total of 500 μg of total RNA was then used as a template with the SuperScript III First-Strand Synthesis kit (Invitrogen) to produce cDNA. The specific targets were amplified by RT-PCR using oligonucleotides as given in Supplementary Tables S1, S2. β-actin was used as the internal control. The ANOVA and t-tests were performed to calculate the p-value to determine the significant difference between samples.

Total RNA from zebrafish tissues was isolated simultaneously using TRIzol (Invitrogen) and RNA Clean & Concentrator kit (Zymo Research). RNA-Seq libraries (n = 2) were processed at the Vanderbilt Technologies for Advanced Genomics (VANTAGE) core. Briefly, the samples were processed using the TruSeq Standard Sample Prep Kit (Illumina) to prepare cDNA libraries after Poly(A) selection. The libraries were sequenced on an Illumina NovaSeq 6000 to a depth of 50 million at 150 bp paired-end reads per library. For lens RNA-seq, FASTQ reads were aligned through Illumina’s DRAGEN RNA Seq pipeline. EdgeR (3.30.3) packages were used to measure differential gene expression with genes that achieved a count per million mapped reads (CPM). For heart RNA-seq, Vanderbilt Technologies for Advanced Genomics Analysis and Research Design (VANGARD) performed the DEG and further analyses.

For lens RNA seq, DEG (FDR cutoff ≤ 0.01) was analyzed through the use of IPA (QIAGEN, https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis). The GO clusters with the significant p-value were taken for further analysis. To use the human Disease Ontology database, the zebrafish genes were first matched to their human orthologues using Ensembl BioMart (www.ensembl.org/biomart). Then, the gene set of human orthologues was processed using Bioconductor DOSE packages (Yu et al., 2015) with a p-value cutoff of ≤ 0.05 to identify the corresponding DO terms.

Lenses of live embryos in 0.3× Danieau water with PTU/tricaine were analyzed by using bright field microscopy (Zeiss Axio Zoom V16) at 4 dpf. The percent of lens defects was scored as defined in our previous study (Mishra et al., 2018). Briefly, the phenotypic characteristic manifested as a spherical, shiny droplet spreading across the lens, which was classified as a defect. For heart imaging, the heart area of the live embryos at 4 dpf was analyzed by using bright field Axio Zoom microscopy. Quantification of the size of the heart area was performed by ImageJ (Duan et al., 2016; Incardona and Scholz, 2016).

One dpf embryo was manually dechorionated, and 10 embryos were placed in one well of a 24-well plate (polystyrene, tissue culture grade) with 1 mL of 0.3× Danieau water. Then, 50 μM dexamethasone (Sigma, D1756) diluted in 0.3× Danieau water was treated from 1 to 4 dpf to examine the cardiac phenotypes. For stain experiments, 2.5 or 5 μM atorvastatin (Santa Cruz Biotechnology, sc-337542A) was treated from 1 to 4 dpf to observe lens defects, and 4 μM lovastatin (Santa Cruz Biotechnology, sc-200850A) was treated for 16 h before examining lens abnormalities at 4 dpf.

Statistical analyses were carried out with GraphPad Prism software 9 (GraphPad) by means of Student’s t-test or ANOVA. Comparison between groups was performed with Bonferroni or Tukey test for one-way or two-way ANOVA. Statistical significance was defined as p < 0.05.