Ectopic Rod Photoreceptor Development in Mice with Genetic Deficiency of WNT2B

Wnt/β-catenin signaling is essential for embryonic eye development in both the anterior eye and retina. WNT2B, a ligand and activator of the Wnt/β-catenin pathway, assists in the development of the lens and peripheral regions of the eye. In humans WNT2B mutations are associated with coloboma and WNT2B may also assist in retinal progenitor cell differentiation in chicken, yet the potential role of WNT2B in retinal neuronal development is understudied. This study explored the effects of WNT2B on retinal neuronal and vascular formation using systemic Wnt2b knockout (KO) mice generated by crossing Wnt2bflox/flox (fl/fl) mice with CMV-cre mice. Wnt2b KO eyes exhibited relatively normal anterior segments and retinal vasculature. Ectopic formation of rod photoreceptor cells in the subretinal space was observed in Wnt2b KO mice as early as one week postnatally and persisted through nine-month-old mice. Other retinal neuronal layers showed normal organization in both thickness and lamination, without detectable signs of retinal thinning. The presence of abnormal photoreceptor genesis was also observed in heterozygous Wnt2b mice, and occasionally in wild type mice with decreased Wnt2b expression levels. Expression of Wnt2b was found to be enriched in the retinal pigment epithelium compared with whole retina. Together these findings suggest that WNT2B is potentially involved in rod photoreceptor genesis during eye development; however, potential influence by a yet unknown genetic factor is also possible.


Introduction
Wnt/β-catenin signaling is a well-studied and well-conserved signaling pathway essential for embryonic eye development [1,2]. It regulates development in multiple cell types and regions of the eye, including the lens [3], retinal pigment epithelium (RPE) [4,5], and retinal vasculature [6], and is active throughout multiple developmental stages [1]. Wnt signaling acts within a large gene regulatory network in the embryonic eye, where it interacts with transcription factors and other signaling pathways, such as paired box 6 (Pax6), transforming growth factor beta (TGFβ), and bone morphogenetic proteins (BMP), to ensure proper tissue development [7][8][9].
WNT2B is a ligand and activator of the canonical Wnt/β-catenin pathway. Like other Wnt ligands, it binds to Wnt receptor complex Frizzled4/Lrp5/6 to subsequently stabilize β-catenin and allow for target gene transcription [10,11]. WNT2B has been shown to regulate the development of the peripheral regions of the eye [12,13] and to inhibit retinal neuronal differentiation in the embryonic chick eye [14]. It has also been reported that Wnt2b is highly expressed during BMP-induced RPE regeneration [7]. Humans with WNT2B mutations exhibit neonatal-onset chronic diarrhea [15], and a variety of ocular abnormalities, including iris colobomas and corneal defects, although it is unclear whether WNT2B is directly responsible for these ocular pathologies [16,17]. Overall WNT2B remains relatively understudied regarding retinal neuronal and vascular development.
In this study we investigated potential presence of ocular phenotypic abnormalities related to the previously described human ocular pathologies in mice with Wnt2b deficiency. We also explored the possible role of WNT2B in retinal neuronal and vascular development. We found that Wnt2b-deficient mice exhibit normal anterior segments and retinal vascular patterning. Yet a portion of Wnt2b-deficient mice displayed ectopic rod photoreceptor nuclei in the subretinal space, indicating abnormal rod maturation, differentiation, or migration. These findings suggest a potential involvement of WNT2B in photoreceptor genesis. Given that this phenotype was also occasionally observed in wild type (WT) mice as well at a lower frequency, potential influence or synergistic effects by another unknown gene cannot be excluded.

Generation of Wnt2b Knockout (KO) Mice
All animal studies were approved by the Boston Children's Hospital Animal Care and Use Committee (protocol #00001722, approved 27 June 2022). The studies also adhered to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Eye Dissection and Examination of Anterior Segments and Lens
Appearance of the eyelid and anterior segment of the eye (including cornea, iris, and lens) was examined both in live adult mice and in dissected eyes, after euthanasia with CO 2 inhalation followed by cervical dislocation. After eye enucleation, the lens was removed under a dissecting microscope by removing the anterior portion of the eye. Sizes of whole eyeballs and lens were measured using a ruler and analyzed in Adobe Photoshop.

Analysis of Retinal Vasculature in Retinal Flat Mounts
Staining of retinal vasculature was performed following protocols as described previously [20,21]. Adult mouse eyes (>4 weeks old) were enucleated and fixed in 4% paraformaldehyde for 1 h. Retinas were dissected in 1X phosphate-buffered saline (PBS) under a dissecting microscope by removing the anterior portion of the eye and peeling away the choroid/sclera complex. Retinas were first permeabilized in cold 70% ethanol on ice for 30 min, followed by permeabilization in PBS with 1% Triton X-100 (PBSTx) for 1 h. Retinas were then incubated in isolectin GS-IB 4 with Alexa Fluor 594 conjugate (Invitrogen, Waltham, MA, USA, Cat#: I21413) in 0.1% PBSTx overnight. Retinas were rinsed in PBS three times, then flattened onto slides by making 4 radial incisions with the photoreceptor side down. Slides were coverslipped using ProLong Glass Antifade Mountant (Invitrogen, Cat#: P36980) and imaged using a fluorescence microscope (AxioObserver.Z1 microscope; Carl Zeiss Microscopy, Oberkochen, Germany) to visualize vasculature in both superficial and deep layers.

Retinal Histology Characterization in Eye Cross Sections with H&E Staining
Retinal histology was examined in cross sections based on prior protocols [22]. After euthanasia, enucleated mouse eyes were embedded and frozen in Optimal Cutting Temperature compound (Sakura Finetek, Torrence, CA, USA, Cat#: 4583). Cross sections Cells 2023, 12, 1033 3 of 18 (12 µm) were cut using a cryostat (Leica Biosystems, Wetzlar, Germany) and placed on positively charged microscope slides (VWR, Radnor, PA, USA, Cat#: 16004-406). Sections were air-dried and briefly fixed in 4% paraformaldehyde for 15 min and stained with hematoxylin (Sigma-Aldrich, St. Louis, MO, USA, Cat#: HHS32-1L) for 3 min. Sections were then rinsed in ddH 2 O and developed in tap water for 5 min, followed by 10 dips in acid alcohol (0.5% hydrochloric acid in 70% ethanol) to remove excess stain. Sections were then stained with eosin Y solution (Sigma-Aldrich, St. Louis, MO, USA, Cat#: 1098441000) for 30 s and rinsed in ddH 2 O. The slides were dehydrated in a graded series of ethanol (50%, 70%, 95%, and 100%) and incubated with xylenes for at least 10 min. Slides were coverslipped using Permount Mounting Medium (Fisher Scientific, Waltham, MA, USA, Cat#: SP15-100) and air-dried before imaging.

Fundus and OCT Imaging
Adult mice (2-4 months old) were anesthetized with ketamine/xylazine mixture, and their pupils were dilated with topical administration of Cyclomydril drops (Alcon Laboratories, Fort Worth, TX, USA). Fundus and optical coherence tomography (OCT) images were taken using a Micron IV imaging system (Phoenix Research Lab, Pleasanton, CA, USA) as previously described [22,23].
The following primers and protocols were used for PCR genotyping.
Genotyping for Wnt2b: Wnt2b primers [18]: KO-1: GCCTCTCACACCAGCGTGTAAGAG; KO-4: GTAATTGAGTGGTCTCCACC; 2BNEO: ATCAGCAGCCTCTGTTCCACATAC. Wnt2b Reactions were initially incubated at 94 • C for 2 min. Reactions were then denatured at 92 • C for 15 s, annealed at 70 • C for 30 s, and elongated at 68 • C for 2.5 min. These steps were repeated for 9 cycles with the annealing temperature decreasing by one degree every cycle. Reactions then underwent 30 cycles at 92 • C for 15 s, 60 • C for 30 s, and 68 • C for 2.5 min, with a final incubation at 68 • C for 10 min.

RNA Isolation and Real-Time Quantitative PCR (RT-qPCR)
RT-qPCR experiments were carried out based on previous protocols [25]. Tissue samples from Wnt2b mice were first homogenized and lysed using QIAzol Lysis Reagent (Qiagen, Germantown, MD, USA, Cat#: 79306). Total RNA was isolated using Pure-Link RNA Mini Kit (Ambion, Austin, TX, USA, Cat#: 12183025) following manufacturer protocols. RNA concentration and quality was tested using a NanoDrop 8000 (Thermo Scientific, Waltham, MA, USA). RNA was further purified using DNase I treatment (Ther-moFisher, Waltham, MA, USA Cat#: EN0521) following manufacturer protocols. cDNA was synthesized using iScript Reverse Transcription Supermix (Bio-Rad, Hercules, CA, USA, Cat#: 1708841) following manufacturer protocols. RT-qPCR was performed using a CFX96 Real-Time System (Bio-Rad, Hercules, CA, USA) with SYBR Green qPCR Master Mix (ThermoFisher, Cat Waltham, MA, USA #: 4309155). The primer sequences of target genes are listed in Table 1. Copy number of each target gene cDNA was normalized to the housekeeping gene, Gapdh, using the comparative CT (∆∆CT) method.

Statistical Analysis
Quantitative data are presented as mean ± SD. p values listed were calculated using multiple Student's t tests for two groups of data. p values < 0.5 were considered statistically significant.

Normal Appearance of Anterior Segments of Wnt2b KO Mice
Genetic loss of Wnt2b in KO mice was validated with tail snip genotyping and RT-qPCR. We first examined the anterior segments of the Wnt2b KO mice to determine if any similar morphological abnormalities existed between this mouse strain and the previously reported human ocular phenotypes [16]. Grossly normal appearance of eyelids and cornea with clear lens was observed in both WT and Wnt2b KO mice, with comparable eye size and lens diameter, as well as relatively normal iris size and shape ( Figure 1A-E). These results suggest that Wnt2b KO mice did not appear to reproduce the human ocular phenotype with anterior segment defects associated with WNT2B mutation.

The Retinal Vasculature of Wnt2b KO Mice Appears Normal
WNT2B binds Frizzled4 (Fzd4) [13], and Fzd4 signaling is linked with familial exudative vitreoretinopathy in humans and in mice with incomplete development of retinal vasculature [26]. We therefore examined the retinal vasculature of Wnt2b KO mice. Using retinal flat mounts with isolectin staining, we found that WT, Wnt2b Het, and KO adult mice all showed grossly normal retinal vascular patterning in superficial and deep vessel layers in flat mounts ( Figure 2A-F), which is also confirmed in eye cross sections ( Figure 2G-I). Together these findings suggest that WNT2B is dispensable for retinal vascular development and not required for activating Frizzled4 signaling to exert its role in retinal angiogenesis.

Ectopic Formation of Rod Photoreceptor Clusters in the Subretinal Space in Adult Wnt2b Mutant Mice
Because of presence of abnormal DAPI-positive cell clusters in Wnt2b KO retinas ( Figure 2F), we next performed H&E staining on retinal cross sections to fully investigate retinal neuronal development and lamination. In adult mice (2-5 months old), we found clusters of ectopic cells in the subretinal space, near the photoreceptor inner and outer segments and RPE, in both Wnt2b Het and KO mice ( Figure 3A-C). Localization of these ectopic cells is random across the retina. All other retinal neuronal layers appeared normal in both thickness and organization ( Figure 3A-C). The Wnt2b fl/fl control mice, on the other hand, did not exhibit this phenotype ( Figure 3A).

The Retinal Vasculature of Wnt2b KO Mice Appears Normal
WNT2B binds Frizzled4 (Fzd4) [13], and Fzd4 signaling is linked with familial exudative vitreoretinopathy in humans and in mice with incomplete development of retinal vasculature [26]. We therefore examined the retinal vasculature of Wnt2b KO mice. Using retinal flat mounts with isolectin staining, we found that WT, Wnt2b Het, and KO adult mice all showed grossly normal retinal vascular patterning in superficial and deep vessel layers in flat mounts (Figure 2A-F), which is also confirmed in eye cross sections ( Figure  2G-I). Together these findings suggest that WNT2B is dispensable for retinal vascular development and not required for activating Frizzled4 signaling to exert its role in retinal angiogenesis. To identify the nature of these ectopic cells, we performed immunohistochemistry with photoreceptor cell markers. The ectopic cells in Wnt2b Het and KO retinas showed positive staining for rhodopsin ( Figure 3D-F). In addition, staining for Ribeye, a structural protein of ribbon synapse characteristic of photoreceptor synapses, was also positive around a group of ectopic cells that extended aberrantly from the normal outer nuclear layer in KO retinas ( Figure 3G-I). These findings indicate that these abnormal cells are displaced mature rod photoreceptor cells with photo-sensing pigments and ribbon synapse.

Ectopic Subretinal Cell Clusters Form during Wnt2b Mutant Mice Development
To elucidate the timing of the onset of this phenotype, we examined mice at developmental time points P7, P15, and P30. Ectopic cells were present in both Wnt2b Het and KO mice at all three of these time points ( Figure 4A-C), suggesting that the malformation likely occurs shortly after birth during the retinal progenitor cell differentiation process. At P7 and P15 only, we also observed pigment near the ectopic cells, primarily within the displaced inner and outer segments in KO retinas ( Figure 4C); however, these pigmented structures were not colocalized with cell nuclei marker DAPI ( Figure 4D), indicating an acellular nature or cell debris following loss of nuclei.

Ectopic Formation of Rod Photoreceptor Clusters in the Subretinal Space in Adult Wnt2b Mutant Mice
Because of presence of abnormal DAPI-positive cell clusters in Wnt2b KO retinas ( Figure 2F), we next performed H&E staining on retinal cross sections to fully investigate retinal neuronal development and lamination. In adult mice (2-5 months old), we found clusters of ectopic cells in the subretinal space, near the photoreceptor inner and outer segments and RPE, in both Wnt2b Het and KO mice ( Figure 3A-C). Localization of these ectopic cells is random across the retina. All other retinal neuronal layers appeared normal in both thickness and organization ( Figure 3A-C). The Wnt2b fl/fl control mice, on the other hand, did not exhibit this phenotype ( Figure 3A).
To identify the nature of these ectopic cells, we performed immunohistochemistry with photoreceptor cell markers. The ectopic cells in Wnt2b Het and KO retinas showed positive staining for rhodopsin ( Figure 3D-F). In addition, staining for Ribeye, a structural protein of ribbon synapse characteristic of photoreceptor synapses, was also positive around a group of ectopic cells that extended aberrantly from the normal outer nuclear In addition, the samples with displaced pigments (P15) did not show positive staining for RPE65, an enzyme and marker specific to RPE cells ( Figure 4E,F). This suggests these pigments are not functioning RPE cells.

Absence of Retinal Thinning and Degeneration in Aged Wnt2b KO Mice
To determine whether the ectopic cells may spontaneously regress over time and whether potential aging effects may alter or confound the observed phenotype, we evaluated 9-month-old KO retinas. Although the frequency of the phenotype at 9 month is lower (1/4 of Het and KO mice combined, Table 2) compared with younger ages in frequency analysis (Table 2), we did observe ectopic cells in a 9-month-old Wnt2b KO retina ( Figure 4G,H), confirming that this phenotype can persist well into aging. Except for disorganization of the photoreceptor layer due to the misplaced photoreceptor cells, we did not detect obvious signs of retinal thinning or other forms of age-related retinal degeneration in Wnt2b KO mice when compared to WT mice at 9 months old ( Figure 4G,H), suggesting that the phenotype is mostly a developmental defect that persists into adulthood yet does not appear to progress into retinal thinning and degeneration in later age. layer in KO retinas ( Figure 3G-I). These findings indicate that these abnormal cells are displaced mature rod photoreceptor cells with photo-sensing pigments and ribbon synapse.

Ectopic Subretinal Cell Clusters Form during Wnt2b Mutant Mice Development
To elucidate the timing of the onset of this phenotype, we examined mice at developmental time points P7, P15, and P30. Ectopic cells were present in both Wnt2b Het and KO mice at all three of these time points ( Figure 4A-C), suggesting that the malformation likely occurs shortly after birth during the retinal progenitor cell differentiation process. At P7 and P15 only, we also observed pigment near the ectopic cells, primarily within the displaced inner and outer segments in KO retinas ( Figure 4C); however, these pigmented structures were not colocalized with cell nuclei marker DAPI ( Figure 4D), indicating an acellular nature or cell debris following loss of nuclei.

Lack of Phenotype Detection in Fundus Imaging and OCT in Live Mice
In addition to morphological characterizations in dissected eyes, we performed fundus and optical coherence tomography (OCT) imaging on live Wnt2b mutants to determine if this phenotype could be observed on a macroscopic scale to facilitate live screening and monitoring. We did not observe any obvious abnormalities in the fundus images of Wnt2b Het or KO mice when compared to WT ( Figure 5A,B). We were also unable to detect any gross variations in retinal structure or lamination in the OCT images ( Figure 5C,D). In this technique, the light path is confined to the central retinal area and peripheral regions are not visible in the fundus imaging, potentially limiting our ability to evaluate the full retina through fundus imaging, even when the phenotype was confirmed in retinal sections ( Figure 5E,F). In addition, the samples with displaced pigments (P15) did not show positive staining for RPE65, an enzyme and marker specific to RPE cells ( Figure 4E,F). This suggests these pigments are not functioning RPE cells.

Absence of Retinal Thinning and Degeneration in Aged Wnt2b KO Mice
To determine whether the ectopic cells may spontaneously regress over time and whether potential aging effects may alter or confound the observed phenotype, we evaluated 9-month-old KO retinas. Although the frequency of the phenotype at 9 month is lower (1/4 of Het and KO mice combined, Table 2) compared with younger ages in

Occasional Presence of Phenotype in WT Mice
Despite our initial results suggesting global loss of Wnt2b may be the direct cause of this novel phenotype, we later observed the same phenotype in a small number of littermate WT control mice. Although the frequency was much lower in WT or fl/fl mice (17%) than in Het (53%) or KO (35%) mice (Table 2), we found similar clusters of displaced cells in WT retinal cross sections at P15, P30, and 4 months old (Table 2, Figure 6A-C). In P15

Occasional Presence of Phenotype in WT Mice
Despite our initial results suggesting global loss of Wnt2b may be the direct cause of this novel phenotype, we later observed the same phenotype in a small number of littermate WT control mice. Although the frequency was much lower in WT or fl/fl mice (17%) than in Het (53%) or KO (35%) mice (Table 2), we found similar clusters of displaced cells in WT retinal cross sections at P15, P30, and 4 months old (Table 2, Figure 6A-C). In P15 retinas, the ectopic displaced cells are also associated with pigment deposition ( Figure 6A). retinas, the ectopic displaced cells are also associated with pigment deposition ( Figure  6A). To confirm their genotype, we performed RT-qPCR on eyecup samples from mice, of which one WT had shown ectopic cells in cross sections. Our results were consistent with the previous genotype results from tail snips, with the abnormal phenotypic WT showing consistent Wnt2b expression at a lower level than, but statistically insignificant from, the WT without the phenotype ( Figure 6D). Absence of Wnt2b expression in the KO eye was also confirmed with RT-qPCR, as expected ( Figure 6D). This observation in the WT suggests two possibilities: WT eyes may express varying levels of Wnt2b, and those with lower levels may manifest a phenotype similar to that observed in Het and KO mice; or an unknown gene mutation may cause the displaced photoreceptor phenotype independently of or in combination with loss of Wnt2b.  Figure 4B,C). (D) qPCR of Wnt2b mRNA expression in WT1 (without phenotype), WT2 (with phenotype), and KO mouse eyes from age-matched adult mice (4 months old). Lack of Wnt2b expression is confirmed in KO, and the phenotypic WT2 mouse shows lower yet statistically insignificant levels of Wnt2b compared with WT1 without phenotype. Each sample contains a half eye cup with retinas after lens removal, with histological analysis of phenotype performed in the other half of the same eye for confirmation. For RT-qPCR, n = 3 technical repeats/group. Data presented as mean ± SD. * p < 0.05 KO compared to WT1. NS: not significant.  Figure 4B,C). (D) qPCR of Wnt2b mRNA expression in WT1 (without phenotype), WT2 (with phenotype), and KO mouse eyes from age-matched adult mice (4 months old). Lack of Wnt2b expression is confirmed in KO, and the phenotypic WT2 mouse shows lower yet statistically insignificant levels of Wnt2b compared with WT1 without phenotype. Each sample contains a half eye cup with retinas after lens removal, with histological analysis of phenotype performed in the other half of the same eye for confirmation. For RT-qPCR, n = 3 technical repeats/group. Data presented as mean ± SD. * p < 0.05 KO compared to WT1. NS: not significant.
To confirm their genotype, we performed RT-qPCR on eyecup samples from mice, of which one WT had shown ectopic cells in cross sections. Our results were consistent with the previous genotype results from tail snips, with the abnormal phenotypic WT showing consistent Wnt2b expression at a lower level than, but statistically insignificant from, the WT without the phenotype ( Figure 6D). Absence of Wnt2b expression in the KO eye was also confirmed with RT-qPCR, as expected ( Figure 6D). This observation in the WT suggests two possibilities: WT eyes may express varying levels of Wnt2b, and those with lower levels may manifest a phenotype similar to that observed in Het and KO mice; or an unknown gene mutation may cause the displaced photoreceptor phenotype independently of or in combination with loss of Wnt2b.

Expression of Wnt2b mRNA in the RPE and Retina
To understand the ocular cellular expression of Wnt2b, expression of Wnt2b mRNA in RPE and whole retina from adult eyes was compared ( Figure 7A), which showed enriched levels of Wnt2b in mouse RPE as compared to the retinas. Relative expression of Wnt2b and its potential Wnt receptors (FZD4, FZD7) and coreceptors (LRP5 and LRP6) in retinal cells were analyzed with published single-cell sequencing data from P14 mice [27] (Figure 7B). Within the retina, Wnt2b showed modest expression throughout retinal cells including photoreceptors (rods and cones), bipolar cells, and ganglion cells, with the exception of cholinergic amacrine cells, which have higher levels and more enrichment of Wnt2b than other cell types ( Figure 7B).

Expression of Wnt2b mRNA in the RPE and Retina
To understand the ocular cellular expression of Wnt2b, expression of Wnt2b mRNA in RPE and whole retina from adult eyes was compared ( Figure 7A), which showed enriched levels of Wnt2b in mouse RPE as compared to the retinas. Relative expression of Wnt2b and its potential Wnt receptors (FZD4, FZD7) and coreceptors (LRP5 and LRP6) in retinal cells were analyzed with published single-cell sequencing data from P14 mice [27] ( Figure 7B). Within the retina, Wnt2b showed modest expression throughout retinal cells including photoreceptors (rods and cones), bipolar cells, and ganglion cells, with the exception of cholinergic amacrine cells, which have higher levels and more enrichment of Wnt2b than other cell types ( Figure 7B).  [27]. Scaling is relative to each gene's expression across all cells of a cluster. Gene expression is scaled from zero to one (0.5 = the mean across all cells in the cluster file referenced for dot plotting). % expressing is the percent of cells that have one or more transcripts for the gene of interest. Dot size represents the percentage of cells expressing the specific gene, color indicates the relative expression level (red = higher expression).  [27]. Scaling is relative to each gene's expression across all cells of a cluster. Gene expression is scaled from zero to one (0.5 = the mean across all cells in the cluster file referenced for dot plotting). % expressing is the percent of cells that have one or more transcripts for the gene of interest. Dot size represents the percentage of cells expressing the specific gene, color indicates the relative expression level (red = higher expression).

Expression of Wnt Ligands and Frizzled (Fzd) Receptors Was Altered in Wnt2b KO Retinas
To assess whether other Wnt ligands and Fzd receptors are regulated in the absence of Wnt2b, we evaluated their mRNA expression levels in WT and Wnt2b KO retinas. Expression of Wnt2b was not detectable in KO retinas, as expected ( Figure 8A). Relative expressions of Wnt2, Wnt4, and Wnt5b were down regulated, whereas Wnt10a was modestly upregulated in Wnt2b KO retinas ( Figure 8A). Levels of Wnt3a, Wnt5a, and Wnt7a were comparable ( Figure 8B). For the receptors, we found that expression levels of Fzd2 were significantly upregulated and Fzd8 were downregulated in Wnt2b KO retinas ( Figure 8C), whereas Fzd1, Fzd3, Fzd4, Fzd6, and Fzd7 did not show significant change ( Figure 8D). These data suggest that altered expression of Wnt ligands and receptors may potentially compensate for the loss of function of WNT2B in its absence.
To assess whether other Wnt ligands and Fzd receptors are regulated in the absence of Wnt2b, we evaluated their mRNA expression levels in WT and Wnt2b KO retinas. Expression of Wnt2b was not detectable in KO retinas, as expected ( Figure 8A). Relative expressions of Wnt2, Wnt4, and Wnt5b were down regulated, whereas Wnt10a was modestly upregulated in Wnt2b KO retinas ( Figure 8A). Levels of Wnt3a, Wnt5a, and Wnt7a were comparable ( Figure 8B). For the receptors, we found that expression levels of Fzd2 were significantly upregulated and Fzd8 were downregulated in Wnt2b KO retinas ( Figure 8C), whereas Fzd1, Fzd3, Fzd4, Fzd6, and Fzd7 did not show significant change ( Figure 8D). These data suggest that altered expression of Wnt ligands and receptors may potentially compensate for the loss of function of WNT2B in its absence.
In addition, we also analyzed the expression of neurogenin1. Overexpression of neu-rogenin1 was found to cause ectopic retinas with multiple ectopic cell types and layers in the subretinal space [28][29][30], resembling somewhat the phenotype observed in Wnt2b KO eyes. We found that expression of neurogenin1 showed a trend of upregulation in Wnt2b KO retinas compared with WT, although it was statistically insignificant ( Figure 8E). This suggests a possibility that neurogenin1 dysregulation may potentially explain in part the phenotype in Wnt2b KO eyes.

Discussion
In this study, we examined mice with global genetic deficiency of Wnt2b to determine if they exhibit any ocular abnormalities and thus could be a useful mouse model to study WNT2B-related ocular birth defects found in humans [16,17]. In contrast to the findings in In addition, we also analyzed the expression of neurogenin1. Overexpression of neurogenin1 was found to cause ectopic retinas with multiple ectopic cell types and layers in the subretinal space [28][29][30], resembling somewhat the phenotype observed in Wnt2b KO eyes. We found that expression of neurogenin1 showed a trend of upregulation in Wnt2b KO retinas compared with WT, although it was statistically insignificant ( Figure 8E). This suggests a possibility that neurogenin1 dysregulation may potentially explain in part the phenotype in Wnt2b KO eyes.

Discussion
In this study, we examined mice with global genetic deficiency of Wnt2b to determine if they exhibit any ocular abnormalities and thus could be a useful mouse model to study WNT2B-related ocular birth defects found in humans [16,17]. In contrast to the findings in humans where anterior segment abnormalities were linked with WNT2B mutations, in Wnt2b KO mice we did not observe detectable gross abnormalities of the anterior eye, suggesting that Wnt2b is likely dispensable in anterior eye structure formation, and has a distinctly different role in mice versus humans. Because Fzd4, a major receptor for WNT2B, is important for retinal angiogenesis [26], we also examined formation of retinal vasculature in Wnt2b KO eyes. We found normal development of retinal vessels and patterning, suggesting that WNT2B is not requried for retinal blood vessel growth and formation in mice.
Importantly, we found abnormal retinal neuronal genesis with ectopic clusters of cell nuclei aggregated in the subretinal space. These cells appeared to be displaced aberrantly from the outer nuclear layer. We further confirmed these cells were of rod photoreceptor origin with ribbon synapse expressing both photoreceptor and synaptic markers. These abnormally displaced nuclei were present at P7 through adulthood to 9 months of age. During retinal development at P7 and P15, these nuclei were also accompanied by ectopic pigments of unknown origin, which disappeared at later time points. In addition, we found Wnt2b expression is higher in RPE than in the retina. Together, this observed phenotype of ectopic nuclei may reflect misguided photoreceptor differentiation and migration during development, potentially in the absence or knockdown of Wnt2b. This notion is consistent with the known role of β-catenin/Wnt signaling in controlling retinal progenitor cell fate and in photoreceptor development and regeneration. For example in chick eyes, Wnt2b expression was found in the RPE and the peripheral retina to inhibit differentiation of retinal neurons and regulate peripheral eye formation [12,13].
One interpretation is that disruption of WNT2B in mice primarily affects the RPE cells that typically guide rod photoreceptor development and lamination, causing this process to be disrupted and for additional, ectopic rod cells to develop and/or migrate incorrectly [31]. This notion is supported by our findings of enriched Wnt2b expression in the mouse RPE/choroid/sclera complex, which was significantly higher when compared to the retina (Figure 7). Expression levels of Wnt2b in many retinal cells are modest, except for in cholinergic amacrine cells (Figure 7). Receptors for Wnt2b, including frizzled4 (fzd4), fzd7 [32], and LRP5/6, are also expressed in the retina (Figure 8), and may potentially mediate the signaling of WNT2B. If WNT2B is at least partially responsible for the phenotype observed, WNT2B secretion from the RPE may thus guide retinal neuronal specification or differentiation through these receptors.
Alternatively, it may be possible that some RPE cells may transdifferentiate into ectopic retinal neuronal cells. Wnt signaling is known to be essential for early eye and RPE development, through interaction with other factors such as Mitf, Otx2, and BMP, and previous work shows that Wnt inactivation results in RPE transdifferentiation to neural retinal cells in embryonic mice [5,33]. Given that we observed disrupted lamination as early as P7, altered Wnt2b expression and potential dysregulation of other related developmental factors could result in RPE differentiation into ectopic rod photoreceptor cells. Particularly, Pax6 and Mitf are known to be connected to both Wnt signaling and RPE differentiation [34].
The fact that the observed ocular phenotype was also occasionally present in WT mice also suggests the alternative possibility of an unknown factor or mutation present in this strain that may either directly cause the observed phenotype independently of Wnt2b, or may interact with Wnt2b to precipitate the ocular phenotype. In this regard, one possible mutation we ruled out was the rd8 allele of the Crb1 gene. Rd8 mice show a somewhat similar disruption of retinal nuclear layers, accompanied by retinal degeneration and abnormal fundus imaging with spotting [19]. It was previously discovered that C57Bl/6N mice can harbor this mutation [19], which can confound interpretations of ocular phenotypes. The original Wnt2b fl/fl mice were of a C57Bl/6N background [18], and were crossed to CMV-Cre mice of C57Bl/6J background. We confirmed with genotyping that the Wnt2b KO mice lack the rd8 allele. Moreover, Wnt2b fl/fl mice did not show the abnormal ocular phenotype as observed in Wnt2b KO mice, suggesting that the observed phenotype is unrelated to rd8.
Another potential source of confounding genetic influence may come from the CMV-Cre breeder mice. The Wnt2b KO colony was not selectively bred to remove the Cre gene, hence the CMV-Cre gene was still present throughout the colony. Since all mice in this colony still contained CMV-Cre, the Cre recombinase gene itself is unlikely to be responsible the ocular phenotype. However, it is possible that an unknown gene originated from CMV-Cre strain may contribute to the phenotype independently or may interact with WNT2B to alter its expression or function and thereby leading to the observed ocular phenotype. Indeed, one out of six CMV-Cre eyes examined exhibited a similar phenotype of ectopic photoreceptors, suggesting the possibility of an unknown genetic factor originating from CMV-Cre strain background. To this end, it is worth noting that a separate strain of systemic Wnt2b KO mice was generated using the same Wnt2b fl/fl mice but bred with actin-Cre mice, another systemic Cre strain [18], where no gross eye abnormalities in external appearance in Wnt2b KO mice were reported. However, it is unclear in that study whether detailed examination of eye anatomy and retinal cellular structure was performed.
Although our observed ocular phenotype shows some similarities to those of certain known mutations, no previously discovered ocular phenotype truly matches what we have described in this study. Most observations of disrupted lamination and displaced photoreceptor nuclei in mice appear to result in significant retinal thinning and degeneration with age. The rd6 mutation of the Crb1 gene, for instance, causes ectopic, pigmented nuclei to develop in the subretinal space, but this mutation is also accompanied by severe retinal degeneration and spotting on fundus images [35]. Similarly, the microphthalmia mutation (loss of Mitf) in mice causes disrupted retinal neuronal patterning due to dysfunctional RPE differentiation, but is again accompanied by severe retinal degeneration as well as grossly small eye size [36], which were not observed in our mice. More similar to what we observed here was the phenotype seen in mice with expression of constitutively active RAC1 in rod photoreceptors [37]. These mice displayed abnormal retinal neuronal lamination accompanied by incorrect photoreceptor polarity, leading to improper migration during development. Although the histology at developmental time points closely matches our results, mice with RAC1 expression also showed significant photoreceptor loss by 3 months.
The closest phenotype reported thus far was seen in mice with overexpression of neurogenin1, a proneural gene. These mice developed an ectopic retina in the subretinal space with multiple cell types and layers [28][29][30]. Although the phenotype we observed was less severe and spontaneous, neurogenin1 might be a promising candidate in explaining the ectopic cells in our mice, which is consistent with our findings of a trend of neurogenin1 upregulation in Wnt2b KO retinas ( Figure 8). Interestingly, the Drosophila homolog of neurogenin, Tap, was also found to interact with Wnt protein Dishevelled to direct neuronal development [38]. Thus, it is possible that Wnt2b mutation might be linked with other neuronal guidance factors such as neurogenin1, which can cause a similar phenotype. A potential separate independent mutation of these other factors such as neurogenin1 also cannot be ruled out.
One significant technical aspect that limited our study is that we were only able to observe the ocular phenotype in retinal cross sections ex vivo, but not in live mice with fundus imaging or OCT. This made it technically challenging and impractical to screen a large number of mice to select those affected mice to further evaluate their visual function and other characteristics. Considering the microscopic nature and limited regions with the abnormal photoreceptors, it is possible that visual function might not be severely impaired in those affected mice.
To summarize, in this study we observed ectopic rod photoreceptor nuclei unaccompanied by retinal degeneration in a strain of systemic Wnt2b deficient mice. Given that this was also observed in a limited number of WT mice, we speculate that some other factor, possibly in conjunction with altered Wnt2b expression, is potentially responsible for this phenotype. More in-depth analysis will be needed to fully understand the nature of the genetic causes of the observed ocular phenotype. Nevertheless, the presence of ectopic photoreceptor genesis can make these mice a useful model to study photoreceptor differentiation and migration, with potential implications for photoreceptor regeneration as well.