Loss of Rbfox1 Does Not Affect Survival of Retinal Ganglion Cells Injured by Optic Nerve Crush

Rbfox1 is a multifunctional RNA binding protein that regulates alternative splicing, transcription, mRNA stability and translation. Its roles in neurogenesis and neuronal functions are well established. Recent studies also implicate Rbfox1 in the regulation of gene networks that support cell survival during stress. We have earlier characterized the expression of Rbfox1 in amacrine and retinal ganglion cells (RGCs) and showed that deletion of Rbfox1 in adult animals results in depth perception deficiency. The current study investigates the effect of Rbfox1 downregulation on survival of RGCs injured by optic nerve crush (ONC). Seven days after ONC, animals sustained severe degeneration of RGC axons in the optic nerve and significant loss of RGC somas. Semi-quantitative grading of optic nerve damage in control + ONC, control + tamoxifen + ONC, and Rbfox1–/– + ONC groups ranged from 4.6 to 4.8 on a scale of 1 (normal; no degenerated axons were noted) to 5 (total degeneration; all axons showed degenerated organelles, axonal content, and myelin sheath), indicating a severe degeneration. Among these three ONC groups, no statistical significance was observed when any two groups were compared. The number of RGC somas were quantitatively analyzed in superior, inferior, nasal and temporal retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disc. The average RGC densities (cells/mm2) were: control 6,438 ± 1,203; control + ONC 2,779 ± 573; control + tamoxifen 6,163 ± 861; control + tamoxifen + ONC 2,573 ± 555; Rbfox1–/– 6,437 ± 893; and Rbfox1–/– + ONC 2,537 ± 526. The RGC loss in control + ONC, control + tamoxifen + ONC and Rbfox1–/– + ONC was 57% (P = 1.44954E-42), 58% (P = 1.37543E-57) and 61% (P = 5.552E-59) compared to RGC numbers in the relevant uninjured groups, respectively. No statistically significant difference was observed between any two groups of uninjured animals or between any two ONC groups. Our data indicate that Rbfox1-mediated pathways have no effect on survival of RGCs injured by ONC.

Rbfox1 is a multifunctional RNA binding protein that regulates alternative splicing, transcription, mRNA stability and translation. Its roles in neurogenesis and neuronal functions are well established. Recent studies also implicate Rbfox1 in the regulation of gene networks that support cell survival during stress. We have earlier characterized the expression of Rbfox1 in amacrine and retinal ganglion cells (RGCs) and showed that deletion of Rbfox1 in adult animals results in depth perception deficiency. The current study investigates the effect of Rbfox1 downregulation on survival of RGCs injured by optic nerve crush (ONC). Seven days after ONC, animals sustained severe degeneration of RGC axons in the optic nerve and significant loss of RGC somas. Semi-quantitative grading of optic nerve damage in control + ONC, control + tamoxifen + ONC, and Rbfox1 −/− + ONC groups ranged from 4.6 to 4.8 on a scale of 1 (normal; no degenerated axons were noted) to 5 (total degeneration; all axons showed degenerated organelles, axonal content, and myelin sheath), indicating a severe degeneration. Among these three ONC groups, no statistical significance was observed when any two groups were compared. The number of RGC somas were quantitatively analyzed in superior, inferior, nasal and temporal retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disc. The average RGC densities (cells/mm 2 ) were: control 6,438 ± 1,203; control + ONC 2,779 ± 573; control + tamoxifen 6,163 ± 861; control + tamoxifen + ONC 2,573 ± 555; Rbfox1 −/− 6,437 ± 893; and Rbfox1 −/− + ONC 2,537 ± 526. The RGC loss in control + ONC, control + tamoxifen + ONC and Rbfox1 −/− + ONC was 57% (P = 1.44954E-42), 58% (P = 1.37543E-57) and 61% (P = 5.552E-59) compared to RGC numbers in the relevant uninjured groups, respectively. No statistically significant difference was observed between any two groups of uninjured animals or between any two ONC groups. Our data indicate that Rbfox1-mediated pathways have no effect on survival of RGCs injured by ONC.

INTRODUCTION
The RNA binding protein, fox-1 (Rbfox) homolog family includes three evolutionarily conserved multifunctional proteins, Rbfox1, Rbfox2, and Rbfox3, that regulate RNA metabolism, including alternative splicing, transcription, mRNA stability and translation efficiency. The Rbfox proteins are master regulators of gene networks involved in both neurogenesis and mature neuronal functions. Disruption of Rbfox functions have been associated with several neurodevelopmental and neuropsychiatric disorders, including autism spectrum disorder (ASD), intellectual disability, epilepsy, ADHD, bipolar disorder, schizoaffective disorder and schizophrenia, sleep latency and heart disease (Bhalla et al., 2004;Martin et al., 2007;Sebat et al., 2007;Lal et al., 2013;Amin et al., 2016;McKean et al., 2016;Misra et al., 2020). The functional diversity of these proteins is supported by the expression of a number of isoforms for each family member (Kuroyanagi, 2009;Wei et al., 2016;Conboy, 2017).
Recently we have analyzed the expression of Rbfox1 in adult and differentiating mouse retinas and the effects of downregulation of this protein on visual function and the retinal transcriptome (Gu et al., 2018). We showed that in both developing and mature retinas, Rbfox1 expression is restricted to retinal ganglion cells (RGCs) and amacrine cells (ACs). RGCs are the projection neurons of the retina; they receive visual information from photoreceptors via bipolar (BP) and ACs, process this information and convey it via their axons in the optic nerve to postsynaptic targets in the brain. More than 46 brain regions have been identified in the mouse brain that receive image-forming and non-image visual information from different RGC types (Morin and Studholme, 2014). Functional and genetic classification of mouse RGCs identified at least 32 and 40 groups of RGCs, respectively (Baden et al., 2016;Rheaume et al., 2018). ACs are retinal interneurons that form synaptic connections with BPs and RGCs in the inner plexiform layer (IPL) and are involved in shaping spatial and temporal characteristics of RGC receptive fields (Diamond, 2017). In the mammalian retina, ACs similar to RGCs are represented by more than 30 morphological and functional subtypes. Downregulation of Rbfox1 in adult animals had no effect on retinal architecture or retinal cell morphology (Gu et al., 2018). However, Rbfox1 KO mice showed depth perception deficit, suggesting the involvement of Rbfox1 in the regulation of genes that support the functional integrity of the retino-geniculo-cortical pathway. Interestingly, deletion of Rbfox2 also resulted in depth perception abnormalities, with the normal gross retinal morphology (Gu et al., 2020). Furthermore, the deletion of the Rbfox3, which is normally expressed in most types of RGCs, some types of ACs and HCs, had no significant effect on retinal morphology, pupillary light response (PLR) and the optomotor response . Although, each Rbfox member has been shown to have specific roles in neuronal development, as well as in mature neuron functions (Lee et al., 2009(Lee et al., , 2016Gehman et al., 2011Gehman et al., , 2012Fogel et al., 2012;Hamada et al., 2015Hamada et al., , 2016Jacko et al., 2018;Vuong et al., 2018), the fact that these RNA binding proteins recognize the same (U)GCAUG element within their target genes and that many single KO models show no cellular phenotypes and have relatively modest change in the transcriptome than expected suggest possible redundancy in their function. Pan-neuronal Rbfox1/Rbfox2 double null mice (Rbfox1 loxP/loxP /Rbfox2 loxP/loxP /Nestin-Cre +/− ) on the other hand, exhibit a much more severe phenotype than either single KO and die prenatally (Gehman et al., 2012). This shows the importance of Rbfox1 and Rbfox2 in neurogenesis and neuronal function; even if there is some overlapping in the function of Rbfox proteins, these proteins cannot be considered fully redundant as Rbfox3, for instance, in Rbfox1/Rbfox2 double KO failed to substitute the missing family members.
One of the recently characterized functions of Rbfox1 is its involvement in post-transcriptional regulation of gene expression in response to stress. Using Drosophila oogenesis as an in vivo system for the stress response, it has been shown that Rbfox1 upregulation mediated by stress-responsive miR-980 promotes cell survival (Kucherenko and Shcherbata, 2018). Furthermore, Rbfox1 has been implicated in neuroprotective effect of miR-132 against amyloid β-peptide (Aβ) and glutamate excitotoxicity in Alzheimer's disease (El Fatimy et al., 2018). Based on these observations, we hypothesize that retinal cells that normally express Rbfox1, such as RGCs, will be more susceptible to the stress and damage in the Rbfox1 KO animals compared with the wild-type. The current study evaluates the effect of Rbfox1 downregulation on the survival of RGCs in response to optic nerve crush (ONC), an established procedure for acute injury of RGC axons, which leads to severe and specific degeneration of RGC axons and somata.

Generation of Rbfox1 KO Animals
The use of animals and all experimental procedures with animals were approved by the Animal Research Committee of the University of California at Los Angeles and were performed in compliance with the National Institutes of Health Guide for the Care and Use of Animals and the ARVO (The Association for Research in Vision and Ophthalmology) Statement for the Use of Animals in Ophthalmic and Vision Research. Animals were housed in a 12-h light-dark cycle with food and water available ad libitum. Rbfox1 KO animals were generated as described earlier (Gu et al., 2018). Briefly, homozygous transgenic mice with loxP sites flanking Rbfox1 gene exons 11-12 (Rbfox1 f l/fl ; kindly provided by Dr. Douglas Black, UCLA; Gehman et al., 2011) were crossed with Tg(UBC-Cre/ERT2)1Ejb mice (Jackson Laboratory, Bar Harbor, ME; Ruzankina et al., 2007) and the resulting heterozygous Rbfox1 f l/+; UBC-Cre +/− mice were crossed to Rbfox1 f l/fl mice to obtain homozygous Rbfox1 f l/fl /UBC-Cre +/− animals. The expression of tamoxifen-inducible Cre recombinase gene in UBC-Cre/ERT2 mice is controlled by the human ubiquitin C (UBC) promoter. Cre activity in homozygous Rbfox1 f l/fl /UBC-Cre +/− animals was induced with tamoxifen. Tamoxifen (Sigma, St. Louis, MO) was dissolved in corn oil to a final concentration of 50 mg/ml. Rbfox1 f l/fl /UBC-Cre +/− and age-matched heterozygous Rbfox1 f l/+ control mice, were

ONC and Optic Nerve Injury Grading
ONC was performed as described earlier (Wang et al., 2015). Briefly, a conjunctival incision was made on the temporal side of the globe and the optic nerve was exposed without damage to the optic nerve blood supply. Crush was applied approximately 2 mm behind the globe for 2 s with selfclosing forceps.
To quantify the axonal injury, an established method of grading optic nerve injury was adopted (Jia et al., 2000;Ishii et al., 2003). The tissues were dissected, fixed, processed, and embedded in acrylic resin. One-micrometer-thick sections of the proximal optic nerve were cut and stained with 1% toluidine blue. The samples were examined under microscope and the optic nerve injury was assessed in a masked fashion using a graded scale ranging from 1 (normal; no degenerated axons were noted) to 5 (total degeneration; all axons showed degenerated organelles, axonal content, and myelin sheath). In total, 6 animals per group were included in the assessment of axonal injury.

Cell Quantification
RGC quantification was performed on retinal flat mounts as described earlier. The retinas were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, incubated with 10% serum for 1 h to reduce non-specific staining and then with anti-Rbpms overnight at 4 • C. After washing, retinas were incubated with the corresponding secondary antibody overnight at 4 • C. The retinas were placed flat with the GCL facing upward. With several radial cuts, the retina was divided into four quadrants: superior, inferior, nasal and temporal and mounted flat on the glass slide with the GCL facing upward. Four sampling fields (0.24 × 0.24 mm each) were imaged at 0.5, 1, and 1.5 mm from the center of optic nerve disc in each retinal quadrant with a confocal laser scanning microscopy (Olympus FV3000, MA). Retinas from six animals per group were used in these experiments. Quantification Rbfox1-positive RGCs were counted in superior (S), inferior (I), nasal (N) and temporal (T) retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disk.
Frontiers in Neuroscience | www.frontiersin.org  Uninjured optic nerve in all three groups are normal. Optic nerves in the ONC groups sustained severe axonal degeneration, axonal swelling and degeneration of myelin sheath. Rectangular box is presented at higher magnification on the right side of each micrograph. (G) Semi-quantitative analysis of optic nerve injury after ONC. The grading of optic nerve in three uninjured groups (no ONC) were scored close to 1. The optic nerve injury grades for ONC groups was ranging from 4.6 to 4.8. No significant difference between these ONC groups was observed when any two groups were compared.
was performed in a masked manner. Data are presented as the mean ± SD.

Statistical Analysis
Data are presented as mean ± SD. An unpaired Student's t-test was used for quantitative analysis of axonal damage and RGC densities. P < 0.05 was considered statistically significant.

Rbfox1 Downregulation Has No Effect on Retinal Morphology in Uninjured or ONC Animals
The extent of Rbfox1 downregulation in Rbfox1 f l/fl /UBC-Cre +/− retinas and its effect on retinal morphology was evaluated immunohistochemically with antibodies against Rbfox1, Rbpms (RGC marker) and calbindin (AC marker). Age-matched heterozygous Rbfox1 f l/+ animals treated with corn oil (vehicle) or with tamoxifen were used as controls. As expected, Rbfox1 expression in mouse retinas was localized to RGCs and dACs in the ganglion cell layer (GCL) and innermost row of ACs in the inner nuclear layer (INL; Figures 1, 2). Colocalization of the Rbfox1-positive cells with Rbpms-positive cells in retinal sections of heterozygous Rbfox1 f l/+ control animals treated with corn oil or with tamoxifen is shown in Figures 1A,B, respectively. No difference between these two control groups was observed, indicating that tamoxifen has no adverse effect on the expression of Rbfox1, Rbpms and calbindin or on retinal morphology (Figures 1, 2). The downregulation of Rbfox1 in retinas of Rbfox1 f l/fl /UBC-Cre +/− animals, was more prominent in the GCL and particularly in RGCs. Very few Rbpms/Rbfox1positive were present (Figure 1C; see below for quantitative data). One week after ONC, the number of Rbpms-positive cells dropped significantly (Figures 1D-F) and almost no Rbfox1 staining was observed in the GCL (Figure 1F; see below for quantitative data). Co-immunostaining with Rbfox1 and calbindin showed significant overlap in the expression of these proteins in ACs localized in the GCL and INL (Figure 2). The vast majority of Rbfox1-positive/calbindin-negative cells in the GCL are RGCs. We also know from our earlier work that some subtypes of dACs (∼7% of dACs) do not express Rbfox1. Dramatic decrease in Rbfox1 expression in dACs in the GCL and ACs in the INL of Rbfox1 KO animals (Figures 2C,F) compared to control retinas (Figures 2A,B,D,E) was observed. For quantitative analysis of Rbfox1-positive RGCs in Rbfox1 −/− retinas vs. relevant controls, Rbfox1/Rbpms-immunostained cells were counted in superior, inferior, nasal and temporal retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic nerve disk in whole mounted retinas ( Table 1). The average densities (cells/mm 2 ) of Rbfox1-positive RGCs in uninjured animals were: control 6,188 ± 1,155; control + tamoxifen 5,732 ± 810; and Rbfox1 −/− 1,625 ± 524 (n = 6/group; Figures 3A-C,G). In animals with axonal injury, the average numbers of Rbfox1-positive RGCs were: control + ONC 1,016 ± 377; control + tamoxifen + ONC 1,037 ± 439; and Rbfox1 −/− + ONC 252 ± 173 (n = 6/group; Figures 3D-G). This translates to approximately 73 and 82% decrease in the number of Rbfox1-positive RGCs in Rbfox1 −/− retinas in uninjured and ONC retinas, respectively, compared to that of relevant control groups (Rbfox1 KO vs. control P = 3.53361E-52; Rbfox1 KO vs. control + tamoxifen P = 7.38334E-67; Rbfox1 KO + ONC vs. control + ONC P = 1.4715E-28; Rbfox1 KO + ONC vs. control + tamoxifen + ONC P = 8.01704E-25).

ONC-Induced RGC Degeneration in Rbfox1 KO Animals
To evaluate the effect of Rbfox1 deletion on the survival of injured RGCs, six groups of animals were used: control, control/tamoxifen, Rbfox1 −/− , control/ONC, control/tamoxifen/ONC and Rbfox1 −/− /ONC. ONC-induced damage was analyzed by grading the axonal degeneration in the RGCs were counted in superior (S), inferior (I), nasal (N), and temporal (T) retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disk.
Frontiers in Neuroscience | www.frontiersin.org FIGURE 5 | RGC quantification in uninjured and ONC animals. RGCs were counted in the superior, inferior, nasal and temporal retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disc. Representative whole mount retinal images of Rbpms labeled RGCs at 1 mm from the center of the optic nerve head from control (A), control + tamoxifen (B), Rbfox1 −/− (C), control + ONC (D), control + tamoxifen + ONC (E), and Rbfox1 −/− + ONC (F) animals. (G) RGC densities in the superior, inferior, nasal and temporal retinal quadrants in all three ONC-injured groups were significantly lower compared to that of corresponding uninjured animals. optic nerve and by counting RGC somata in the retina 7 days after injury. Axons in both uninjured control groups, as well as uninjured Rbfox1 KO animals, appeared to be normal and no noteworthy differences between these groups were observed (Figures 4A-C). Extensive optic nerve degeneration with degenerated myelin sheath, swollen axons and activated glial cells were noted in the optic nerves of all three groups of animals with ONC (Figures 4D-F). Semi-quantitative grading of optic nerve injury in three uninjured groups were scored close to 1 (normal, no degeneration), whereas the average injury grades for FIGURE 6 | Average RGC densities in Rbfox1 −/− and control animals. The average RGC densities (cells/mm 2 ) were: control 6,438 ± 1,203; control + ONC 2,779 ± 573; control + tamoxifen 6,163 ± 861; control + tamoxifen + ONC 2,573 ± 555; Rbfox1 −/− 6,437 ± 893; and Rbfox1 −/− + ONC 2,537 ± 526. The RGC loss in control + ONC, control + tamoxifen + ONC and Rbfox1 −/− + ONC was 57% (P = 1.44954E-42), 58% (P = 1.37543E-57) and 61% (P = 5.552E-59) compared to RGC numbers in the relevant uninjured groups, respectively. No statistically significant difference was observed between any two groups of uninjured animals or between any two ONC groups. groups with ONC ranged from 4.6 to 4.8 (severe degeneration; Figure 4G). Among these three ONC groups, no statistical significance was observed when any two groups were compared. Rbpms-labeled RGC somas were counted in superior, inferior, nasal and temporal retinal quadrants at 0.5, 1, and 1.5 mm from the center of the optic disc. The RGC densities in these locations for all six groups of animals are presented in Table 2 and Figure 5. The average RGC densities (cells/mm 2 ) were: control 6,438 ± 1,203; control + ONC 2,779 ± 573; control + tamoxifen 6,163 ± 861; control + tamoxifen + ONC 2,573 ± 555; Rbfox1 −/− 6,437 ± 893; and Rbfox1 −/− + ONC 2,537 ± 526 (n = 6/group; Figure 6). The RGC loss in control + ONC, control + tamoxifen + ONC and Rbfox1 −/− + ONC was 57% (P = 1.44954E-42), 58% (P = 1.37543E-57) and 61% (P = 5.552E-59) compared to RGC numbers in the relevant uninjured groups, respectively. No statistically significant difference was observed between any two groups of uninjured animals or between any two ONC groups.

DISCUSSION
Rbfox1 is a moonlighting protein, which regulates alternative splicing, transcription, mRNA stability and translation efficiency. Its role in regulation of gene networks associated with neurogenesis and neuronal function is well established. Recent studies implicate Rbfox1 in the stress-induced regulation of molecular pathways that promote cell survival. A mechanism by which Rbfox1 promotes cell survival during Drosophila oogenesis, which was used as an in vivo system for the stress response, involves its interaction with stress-dependent miRNA miR-980 (Kucherenko and Shcherbata, 2018). Stress-induced reduction of miR-980 expression led to an increased level of Rbfox1, extensive formation of ribonucleoprotein (RNP) granules, and resulted in higher cell viability. Rbfox1 was also implicated in the neuroprotective effect of miR-132 against amyloid β-peptide (Aβ) and glutamate excitotoxicity (El Fatimy et al., 2018). miR-132 has been associated with progression of both amyloid and Tau pathology in Alzheimer's disease. It is the most downregulated miRNA in Alzheimer's disease brain; the miR-132 downregulation precedes neuronal loss (Smith et al., 2015;Hernandez-Rapp et al., 2016;Salta et al., 2016;Pichler et al., 2017). Neuroprotective effects of miR-132 are mediated by direct regulation of the Tau modifiers including Rbfox1 (El Fatimy et al., 2018). Tau pathology has been also associated with dysfunction and degeneration of RGCs in a rat glaucoma model, 3xTg-AD mouse model of AD and m3R tau-Tg mice exhibiting Pick's Disease pathology (Chiasseu et al., 2016(Chiasseu et al., , 2017Ngolab et al., 2021). This suggests that the miR-132/Rbfox1-mediated mechanisms that promote neuronal survival in the brain of the AD mouse model may also be present in RGCs to protect these cells from stress-induced damage.
In our earlier work, we have characterized the expression of Rbfox1 in the retina and evaluated the effect of its downregulation on retinal integrity and visual function (Gu et al., 2018). We showed that Rbfox1 expression is restricted to RGCs and predominantly GABAergic ACs in the GCL and INL. Deletion of Rbfox1 in adult mouse retinas had no effect on retinal gross morphology but Rbfox1 −/− animals exhibited depth perception deficiencies. In this study, we evaluate the survival of RGCs 7 days after ONC in control and Rbfox1 −/− animals to determine whether or not Rbfox1 is involved in cell protection in response to stress. Downregulation of Rbfox1 in Rbfox1 −/− is tamoxifeninduced and, although, the Cre-loxP system and its induction by tamoxifen has been established as a reliable and safe method to delete target genes in the retina (Boneva et al., 2016), we have included in this study an additional tamoxifen-treated control group to rule out the potential cytotoxic effect of the drug itself. ONC injury resulted in virtually complete degeneration of RGC axons and more than a twofold decrease in the number of RGC somas. The average RGC densities in uninjured control, control + tamoxifen and Rbfox1 −/− groups were 6,438, 6,163, and 6,437 per mm 2 , respectively, which well correlate with reported RGC densities in mouse retinas adjacent to the optic disk (Dräger and Olsen, 1981). Average RGC densities as well as RGC densities in any of the four retinal quadrants were similar between Rbfox1 −/− and control groups indicating that downregulation of Rbfox1 in adult mouse retinas does not undermine RGC integrity. ONC-induced injury resulted in 57, 58, and 61% loss of RGC somas in control + ONC, control + tamoxifen + ONC and Rbfox1 −/− + ONC animals, respectively. No significant difference in RGC numbers was observed between Rbfox1 −/− and control groups. Based on these data, we can suggest that Rbfox1-mediated pathways have no effect on promoting RGC survival injured by ONC. Alternatively, since there is a functional redundancy of Rbfox proteins, the loss of Rbfox1 may be compensated by Rbfox2 or Rbfox3. We can also speculate that RGC protective effect in response to the injury is regulated by Rbfox2 or Rbfox3, and not by Rbfox1, as in the above mentioned models of cellular stress (El Fatimy et al., 2018;Kucherenko and Shcherbata, 2018). Rbfox3, for instance, has been shown to be downregulated in ONCinjured RGCs prior to neuronal degeneration (Tran et al., 2019), suggesting that the normal function of this protein is important for RGC survival. Therefore, evaluation of double or triple Rbfox knockouts may provide additional information about the role of Rbfox proteins in supporting the survival of RGCs under stressful conditions.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The animal study was reviewed and approved by The Animal Research Committee of the University of California at Los Angeles.