Fig 1.
Immunohistochemical localization of Rbfox1 expression in mouse retinal sections.
Rbfox1 immunoreactivity is present in the ganglion cell layer (GCL) and inner nuclear layer (INL) of the retina. Rbfox1-positive cells in the INL, which contains cell bodies of horizontal cells (HCs), bipolar cells and amacrine cells (ACs), as well as Muller glia cells, were primarily localized proximal to the inner plexiform layer (IPL). A. Rbfox1 was colocalized with Rbpms-positive RGCs. White arrows point at several Rbfox1/Rbpms-positive RGCs. B. Calbindin-positive displaced ACs (dACs; blue arrows) and ACs with cell bodies localized at the margin with IPL (white arrows) were also immunoreactive for Rbfox1. C. Colocalization of Rbfox1 and Rbfox2 showed significant overlap in expression of these genes within GCL. In the INL, as stated above, Rbfox1 is expressed in ACs adjacent to the IPL, whereas the expression of Rbfox2 is more widely distributed among ACs. White arrows point at Rbfox2-positive/Rbfox1-negative ACs and dACs in the INL and GCL, respectively. Blue arrows indicate Rbfox2-positive HCs. D and E. Colocalization of Rbfox1 and Rbfox2 with cells that are immunoreactive for calbindin generated against C-terminal peptide. These anti-calbindin antibodies show strong immunoreactivity in HCs. Arrows point at HCs in the INL that are immunoreactive for calbindin and Rbfox2. ONL, outer nuclear layer, OPL; outer plexiform layer, DAPI; 4',6-diamidino-2-phenylindole.
Fig 2.
Immunohistochemical colocalization of Rbfox1 positive cells with GABAergic and cholinergic amacrine cells.
A. In the inner nuclear layer (INL) many Rbfox1 immunoreactive cells were colocalized with GABAergic ACs. Blue arrows point at several Rbfox1/GABA-positive cells in the INL. White arrows point at GABAergic cells in the INL and GCL that were not immunostained for Rbfox1. In the ganglion cell layer (GCL), a vast majority of GABAergic dACs were Rbfox1-immunopositive. B. All ChAT-immunoreactive starburst ACs (SACs) in the INL (type a) and in the GCL (type b) were Rbfox1-positive. White and blue arrows point at several type a and type b SACs, respectively that are colocalized with Rbfox1-positive cells. ONL, outer nuclear layer, OPL, outer plexiform layer; IPL, the inner plexiform layer; DAPI (4',6-diamidino-2-phenylindole).
Fig 3.
Quantitative analysis of Rbfox1-positive cells in whole mount adult mouse retina.
A. All Rbpms-positive RGCs were also stained with Rbfox1 antibody. There are some Rbpms stained cells that appeared to be negative for Rbfox1 (indicated by arrowheads), but at higher magnification these cells have faint Rbfox1 staining. B. Approximately 94% of calbindin-positive cells were also positive for Rbfox1. C. Colocalization of Rbfox1 and Rbfox2 positive cells revealed approximately 6% of cells with Rbfox2 expression were negative for Rbfox1. Arrowheads in (B) and (C) point at cells immunoreactive for calbindin or Rbfox2, respectively but negative for Rbfox1.
Fig 4.
Rbfox1 expression pattern at embryonic day 12 (E12) of mouse ocular development.
A. A distinctive Rbfox1 staining at E12 is clearly present in the surface ectoderm, lens and retina. B. Rbfox1 expression in retinal cells appears to be cytoplasmic, whereas, Rbfox2 expression is predominantly nuclear. The boxed area in is enlarged below. These cells are most likely RGCs, since at that developmental stage RGCs are the main type of differentiated retinal cells. L, lens; LC, lens capsule; R, retina; SE, surface ectoderm.
Fig 5.
Dynamics of subcellular localization of Rbfox1 during early postnatal development of the retina.
A. At P0, along with cells with cytoplasmic Rbfox1 expression, there are cells in the GCL with pronounced nuclear staining (pointed by arrows). Most of Rbfox1 and Rbpms expression is colocalized indicating that these cells are RGCs. B and C. At P5, Rbfox1 expression switches to be predominantly nuclear especially in RGCs and dACs. More cells with nuclear expression of Rbfox1 also appear in the INL. Arrows point at several cells in the GCL and INL with Rbfox1 nuclear staining. D. A significant overlap exists between Rbfox1 and Rbfox2 expression at P5, though in both, the GCL and INL, there are cells with specific expression of one or the other paralog. Rbfox1 staining in the IPL (B, C and D) is non-specific since it was also observed in a control experiment with only secondary antibodies. Blue and white arrows point at several Rbfox1-positive/Rbfox2-negative and Rbfox2-positive/Rbfox1-negative cells, respectively.
Fig 6.
Rbfox1 distribution in the lens and cornea.
A. Rbfox1 staining in the lens observed at E12 (Fig 3A) remained unchanged at P10, P12, P14, P15 and in adult animals. It was predominantly localized in the lens capsule (non-specific), although relatively faint expression can be clearly seen in the cytoplasm of lens epithelial cells from P10 throughout adulthood. Rbfox2 expression is localized to lens epithelial cells (appears to be nuclear) and fiber cells. B. In the cornea, strong Rbfox1 staining is observed in the stroma (non-specific) and endothelial cells. Less pronounced expression is present in epithelial cells. The pattern of this staining in the cornea did not undergo detectable change from P10 into adult stage. Rbfox2 expression was observed in corneal epithelial and endothelial cells, as well as keratocytes in the stroma. Strong Rbfox1 staining in lens capsule and corneal stroma is non-specific since it is present in negative control immunohistochemistry without Rbfox1 primary antibodies.
Fig 7.
Rbfox1 KO animals have normal retinal morphology.
A and B. The Tg(UBC-cre/ERT2)1Ejb transgenic mouse was used to generate Rbfox1 KO animals. The images (modified from www.informatics.jax.org/recombinase/specificity?id=MGI:3707333&system=sensory+organs) show beta-galactosidase expression in ocular tissues of an adult Tg(UBC-cre/ERT2)1Ejb transgenic mouse. The rationale for using this transgene is based on a strong Cre expression in RGCs, optic nerve, in the majority of dACs, and certain types of ACs adjacent to the IPL, as well as in the cornea. These locations are most relevant to study the effect of Rbfox1 downregulation on visual function. C-E. Colocalization of Rbfox1 with Rbpms (C), calbindin (D) and Rbfox2 (E) in retinas of wild-type and Rbfox1 KO animals. Very few Rbfox1-positive cells in the GCL were detected. Less dramatic (compared to the GCL), but evident downregulation of Rbfox1 is also observed among ACs in the INL. D. Calbindin immunoreactivity is decreased in Rbfox1 KO retinas; almost no calbindin staining is observed in the INL and very few calbindin-positive cells are present in the GCL. E. Rbfox2 expression in the GCL cells appears to be increased. F. Light microscopic examination of epoxy resin-embedded specimens also failed to identify any change in retinal morphology two months after Rbfox1 downregulation. G. Expression of glial fibrillary acidic protein (GFAP) in Rbfox1 KO retinas. GFAP staining was used to evaluate potential “stress” in the retina associated with Rbfox1 downregulation. However, no significant difference in the levels of GFAP immunoreactivity between control and Rbfox1 KO retinas was observed. OS, photoreceptor outer segments; IS, photoreceptor inner segments; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.
Fig 8.
RGC quantification in retinas of Rbfox1 KO animals.
The possible effect of Rbfox1 downregulation on RGC survival was evaluated two months after last administration of tamoxifen. RGCs were immunolabeled with antibodies against Rbpms and counted in inferior, superior, temporal and nasal retinal quadrants. The average densities of RGCs in all four retinal quadrants of Rbfox1 KO (n = 3) and control (n = 3) animals were similar: in the superior quadrant 2913 ± 338.14 cells/mm2 in control vs 3139 ± 364.21 cells/mm2 in Rbfox1 KO, P = 0.376; in the inferior quadrant 2750 ± 502.14 cells/mm2 in control vs 2921 ± 341.66 cells/mm2 in Rbfox1 KO, P = 0.526; in the nasal quadrant 3057 ± 238.79 cells/mm2 in control vs 3037 ± 339.50 cells/mm2 in Rbfox1 KO, P = 0.898; and in the temporal retina 2884 ± 310.62 cells/mm2 in control vs 2743 ± 387.54 cells/mm2 in Rbfox1 KO, P = 0.461. Data are presented as the mean ± SEM.
Fig 9.
Visual cliff test reveals depth perception impairment in Rbfox1 KO mice.
Rbfox1 KO and control mice were subjected to two modifications of the test: A. the first test determines the time the animal spends in deep versus shallow side of the chamber; B. in the second test animals were placed on a pedestal between deep and shallow sides and animal’s preference to step down on the perceived deep or shallow side was recorded. The test is designed to identify visual dysfunction that can alter animal’s avoidance of the deep side of the chamber. C. Rbfox1 KO mice spent more time on the deep side than on the shallow side. The overall mean (+/-SD) time spent on the deep side was 179.7 +/- 58.8 seconds for all animals in the Rbfox1 KO group and 42.3 +/- 28.8 seconds for all animals in the control group, respectively. There was a statistically significant mean difference in time spent on the deep side between two groups: 137.4 seconds (95% CI = 82.8–192.0 seconds; p = 0.002). D. Rbfox1 KO mice were also less selective when choosing shallow vs deep side to step down. Control mice have clear preference for the shallow side of the box and avoid the deep side (C and D).
Fig 10.
Identification of Rbfox1-regulated genes in the retina.
A. Heatmap representing the top differentially expressed genes in Rbfox1 KO vs control retinas after RNA sequencing. Rows are transcripts and each column is an experimental sample compared to three control samples. Red and green in the heat map indicate up- and down-regulation, respectively. B. Real-time PCR quantification of several differentially regulated genes identified by RNA-seq. Downregulation of prostaglandin D2 synthase (Ptgds), vesicle-associated membrane protein 1 (Vamp1), RPE-retinal G protein receptor (Rgr), calbindin 1 (Calb1) and upregulation of transient receptor potential cation channel, subfamily V, member 1 (Trpv1), estrogen receptor 1 (Esr), O-acyltransferase like protein (Oacyl) and synaptonemal complex central element protein 1 (Syce1) in the retinas of Rbfox1 KO animals observed by quantitative real-time PCR are in agreement with the results of RNA-seq. The data are presented as the mean ± SEM. C. Immunohistochemical analysis of Vamp1 and Vamp2 in mouse retinal sections. Vamp1 is localized to a very small population of RGCs, whereas Vamp2 is widely expressed in the inner and outer plexiform layers. Vamp1 and Vamp2 were noticeably downregulated in retinas of Rbfox1 KO animals, which is consistent with RNA-seq and real-time PCR data. ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.
Table 1.
Differentially regulated genes in Rbfox1 null mouse retinas that have been reported to be have neuronal function.
Table 2.
Association of differentially regulated genes in retinas of Rbfox KO animals with neurological diseases.
Table 3.
Primers used for quantitative real-time PCR.