SP1‐mediated upregulation of LINGO‐1 promotes degeneration of retinal ganglion cells in optic nerve injury

Abstract Backgrounds Insults to the axons in the optic nerve head are the primary cause of loss of retinal ganglion cells (RGCs) in traumatic, ischemic nerve injury or degenerative ocular diseases. The central nervous system–specific leucine‐rich repeat protein, LINGO‐1, negatively regulates axon regeneration and neuronal survival after injury. However, the upstream molecular mechanisms that regulate LINGO‐1 signaling and contribute to LINGO‐1–mediated death of RGCs are unclear. Methods The expression of SP1 was profiled in optic nerve crush (ONC)–injured RGCs. LINGO‐1 level was examined after SP1 overexpression by qRT‐PCR. Luciferase assay was used to examine the binding of SP1 to the promoter regions of LINGO‐1. Primary RGCs from rat retina were isolated by immunopanning and RGCs apoptosis were determined by Tunnel. SP1 and LINGO‐1 expression was investigated using immunohistochemistry and Western bolting. Neuroprotection was assessed by RGC counts, RNFL thickness, and VEP tests after inhibition of SP1 shRNA. Results We demonstrate that SP1 was upregulated in ONC‐injured RGCs. SP1 was bound to the LINGO‐1 promoter, which led to increased expression of LINGO‐1. Treatment with recombinant Nogo‐66 or LINGO‐1 promoted apoptosis of RGCs cultured under serum‐deprivation conditions, while silencing of SP1 promoted the survival of RGCs. SP1 and LINGO‐1 colocalized and were upregulated in ONC‐injured retinas. Silencing of SP1 in vivo reduced LINGO‐1 expression and protected the structure of RGCs from ONC‐induced injury, but there was no sign of recovery in VEP. Conclusions Our findings imply that SP1 regulates LINGO‐1 expression in RGCs in the injured retina and provide insight into mechanisms underlying LINGO‐1–mediated RGC death in optic nerve injury.


| INTRODUC TI ON
Retinal ganglion cells (RGC), the projection neurons of the eye, bear the responsibility of propagating visual stimuli to the brain. Under many traumatic, ischemic nerve injury or degenerative ocular conditions, such as glaucoma, the dysfunction and/or loss of RGC is the primary determinant of visual field loss and are the measurable endpoints in current research into experimental therapies. 1,2 Evaluation of the molecular mechanism underlying RGC neuropathy is important as it may facilitate development of novel therapeutics that ameliorate glaucoma by promoting the survival and axonal regeneration of RGCs.
The lack of cellular and axonal regeneration in the event of neuronal injuries is due to myelin-associated inhibitory factors. [3][4][5] The leucine-rich repeat and immunoglobulin-like domain-containing protein 1, LINGO-1, is expressed by neurons and oligodendrocytes in the central nervous system (CNS) and is an essential component of the NgR/p75 or NgR/Troy signaling complex, which binds to myelin-associated inhibitory ligands. 6,7 LINGO-1 is reported to negatively regulate myelination and neurite extension and to mediate breakdown of the neuronal growth cone. 6,8,9 Importantly, its expression is elevated in patients with various degenerative diseases and those with CNS injuries, 7,10-12 suggesting the potential pathological role of LINGO-1 in CNS diseases.
Moreover, the LINGO-1 gene is related to risk for neuronal apoptosis in patients with neurodegenerative diseases, 13,14 implying that modulation of myelin inhibitor signaling may promote the survival of neurons and myelination of oligodendrocytes after injury. Inhibition of LINGO-1 has neuroprotective effects in models of several CNS diseases and injuries. In two previous studies, a LINGO-1 antagonist significantly increased oligodendrocyte and neural survival, and promoted axonal regeneration and functional recovery after spinal cord injury. 15,16 In a study that used a mouse model of Parkinson's disease, survival of dopaminergic neurons increased and behavioral abnormalities were reduced in LINGO-1-knockout mice compared to wild-type mice. 12 In addition, LINGO-1 antagonists have neuroprotective effects against injury-induced apoptosis in cultured neurons. 12,17,18 We previously reported that in an optic nerve crush (ONC) model, inhibition of LINGO-1 by RNA interference promoted regeneration of the optic nerve and the survival of RGCs. 19 Although inhibition of LINGO-1 promotes the survival of neurons and RGCs, the underlying mechanism is unclear.
Upon axonal injury, transcription factors in neurons are activated, resulting in a cascade of changes in the transcriptome and priming of the degeneration and regeneration pathways. 20 SP1, a multifunctional zinc finger transcription factor that binds to GC-rich motifs in DNA, is implicated in stress-related apoptosis of neurons and the pathogenesis of a variety of degenerative diseases. [21][22][23] However, little is known of the role of SP1 in the regulation of retinal neuropathy. Here, we demonstrate that SP1 regulates the expression of LINGO-1 and contributes to LINGO-1-mediated death of RGCs in the ONC-injured retina. These findings provide insight into the mechanism of the death of RGCs in patients with glaucoma and imply that SP1 and LINGO-1 may be potential therapeutic targets in neuroprotection strategies for glaucoma.

| Animals and ethics statement
A total of 52 male Sprague Dawley rats (weight, 180-200 g; age, 6-8 weeks) and 40 newborn rats (age, 3-5 days) were maintained in the Ophthalmic Animal Laboratory of Zhongshan Ophthalmic Center. All procedures involving animals were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All experimental procedures were approved by the institutional animal care and use committee of Zhongshan Ophthalmic Center (Permit SYXK 2018-025). All manipulations were performed with rats under general anesthesia with 2%-3% inhaled isoflurane, and the eyes of the rats were administered topical 0.5% Alcaine eye drops (Alcon) prior to surgery, experimentation, and electrophysiology examination.

| Optic nerve crush model
Optic nerve crush injury was performed as described previously. 19,24 In brief, after general anesthesia, a lateral canthotomy was performed on the temporal conjunctiva of the right eye of the rat using conjunctival scissors, the lateral rectus muscle was detached, and the optic nerve was exposed under a binocular surgical microscope.
A Dumont #5 clip (World Precision Instruments) was applied to the optic nerve 2 mm behind the posterior eye pole for 5 seconds to provide a consistent clamping force and so ensure the reproducibility of the injury. The left eyes of the rats underwent sham surgery, which entailed exposure of the optic nerve but no ONC injury.

| Constructs and dual-luciferase reporter assays
The LINGO-1 promoter fragment, comprising nucleotides −2104 to +121 bp of the LINGO-1 5′-flanking region relative to the transcription start site, was amplified by PCR (forward primer, and Renilla luciferase activity were measured using the Dual-Glo Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to that of Renilla luciferase and is presented as relative luciferase units.

| Primary culture of RGCs and survival assays
Primary RGCs were isolated and purified by immunopanning. 25 Briefly, the retinas of 3-day-old SD rats were triturated and digested

| Quantitative reverse transcription PCR
Total RNA was extracted from retina tissues using TRIzol reagent (Invitrogen). RT-qPCR was performed using PrimeScript

| Western blotting
Retinas were homogenized, and total protein was extracted using a Protein Extraction Kit (Beyotime Biotechnology). The total protein samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, electro-transferred to a nitrocellulose membrane, and exposed to anti-SP1 (1:500; Millipore) and anti-LINGO-1 (1:500; Upstate) antibodies. Next, the membrane was incubated with a horseradish peroxidase-conjugated secondary anti-rabbit antibody (CST); β-actin served as the loading control. The protein bands were detected by enhanced chemiluminescence (Pierce).

| Immunofluorescence assay
After cardiac perfusion with 0.9% saline, the eyes of the rats were collected, fixed in 4% paraformaldehyde for 12 hours, and cryoprotected in 30% sucrose for 12 hours at 4°C. The eyes were next embedded in optimal cutting temperature medium and sectioned (10 μm  and costained with DAPI. TUNEL-positive nuclei were quantified using ImageJ software. Finally, TUNEL-positive cells were enumerated in high-power fields of view of three wells per treatment, and the mean was calculated.

| Intravitreal injections
After general anesthesia and ocular surface anesthesia, intravitreal injections were performed 2 mm behind the limbus using a Hamilton micro-injector with a 30-gauge needle. Five microliters of AAV2-SP1 shRNA (1 × 10 12 GC/mL; GeneChem) was injected into the vitreous cavity of the rats in the experimental group at 14 days before ONC injury, avoiding damaging the lens and fundus hemorrhage and ensuring that the intraocular pressure did not increase markedly. The sequence of the SP1 shRNA was 5′-GCAACAUGGGAAUUAUGAATT-3′.

| Enumeration of RGCs in flat-mounted retinas
Rats were sacrificed and eyes were collected, fixed with 4% PFA for 2 hours. The intact retinas were separated, and five radial incisions were made to create a petal shape.

| Recording of the visual evoked potential
Before assessment of visual evoked potentials (VEPs), the rats were dark-adapted for > 2 hours. The rats were anesthetized by intraperitoneal injection of 10% chloral hydrate. VEPs were evaluated using a Roland RETI-scan system (Roland Consult) for a full-field flash stimulator. The stimulus intensity was 5 dB (9.49 cd × s/m 2 ), the stimulation frequency was 1.0 Hz, the passband was 0.5-50 Hz, and the stimulation frequency was 100. For quantitative analyses, the VEP system detection index was the N1 wave, P1 wave latency (ms), and N1-P1 wave amplitude (μv). A visual stimulus of 1 Hz white light (9.49 c × s/m 2 ) was generated by a full-field Ganzfeld stimulator under dark-adapted conditions. The amplitude of N1-P1 and the latency of the N1 and P1 peaks were measured using Roland software (Roland Consult). The amplitude of N1-P1 was determined as the interval from the trough of the first negative peak after light onset (N1) to the peak of the first positive wave (P1). The latency of the N1 and P1 waves was measured from light onset to the peak of N1 or P1.

| Optical coherence tomography imaging
To assess ONC-induced changes in peripapillary retinal nerve fiber layer thickness (RNFLT), we performed volume scans using a noninvasive high-resolution SD-optical coherence tomography (OCT) instrument (Spectralis HRA + OCT, Heidelberg Engineering) at baseline and 2, 7, 14, and 21 days after surgery as described previously. 26 The rats were anesthetized by intraperitoneal injection of 10% chloral hydrate. Next, they were placed on a freely rotating platform

| Statistics
All experiments were performed in at least triplicate biological repeats. Data are presented as means ± standard deviations (SDs).
Statistics was performed using the statistical package for the social sciences (SPSS) software. A P value <.05 was considered indicative of significance. Kolmogorov-Smirnov tests were used to assess data distribution for normality. The Student t test, one-way analysis of variance (ANOVA), two-way ANOVA, or repeated measure ANOVA was used to compare differences between groups.

| SP1 is upregulated in RGCs isolated from ONCinjured retina
To and AGCGGGCGG (−508 to −500 bp) ( Figure 1C). qRT-PCR analyses showed that the SP1 mRNA level was significantly increased in injured RGCs compared to control RGCs (7 days post-ONC) ( Figure 1D). Taken together, these data indicate that upregulation of SP1 modulates the expression of LINGO-1 at the transcriptional level.

| Regulation of LINGO-1 expression by SP1 and inhibition of SP1 attenuate LINGO-1-mediated death of RGCs in vitro
To

| LINGO-1 and SP1 colocalize in sham and ONCinjured retinas
To examine the association between LINGO-1 and SP1 in the

| Inhibition of SP1 enhanced the survival of RGCs after ONC in vivo
To examine the neuroprotective effects of inhibition of SP1, we As we reported previously, 19 there was considerable loss of RGCs in F I G U R E 2 Regulation of LINGO-1 by SP1 and inhibition of SP1 attenuated LINGO-1-mediated RGC death in vitro. A, RT-qPCR analysis showed that overexpression of SP1 increased LINGO-1 expression in HEK293 cells (n = 4, means ± SD compared to vector by Student's t test, ***P < .001). B, Luciferase reporter analysis showed that SP1 increased the promoter activity of pLINGO-1 (n = 3, means ± SD compared to vector by two-way ANOVA with Bonferroni test, *P < .05). C, Inhibition of SP1 promoted survival of RGCs in the presence of LINGO-1 in vitro.
Representative images of the effects of Nogo66 or LINGO-1 on serum-deprived RGCs with or without SP1-shRNA.

| Effect of inhibition of SP1 on the VEP after ONC
The preservation of RGC structure motivated us to evaluate the functional recovery of the visual circuits after injury to the optic nerve by assessing the VEP at 1-day pre-ONC (baseline) and at 7, 14, and 28 days post-ONC. Impairment of visual function after ONC was evidenced by reduced amplitude of N1-P1 waves and prolonged latency of N1 waves ( Figure 6A). However, no significant differences were observed in N1-P1 amplitude and N1 latency did not differ significantly between the ONC group and the SP1 transfection + ONC group at 7, 14, and 28 days ( Figure 6B). Therefore, inhibition of SP1 did not prevent or attenuate ONC-induced visual function impairment.  19 We then evaluated the protective effects of inhibition of SP1 on visual function. Surprisingly, inhibition of SP1 did not significantly modulate N1-P1 amplitude or N1 latency. The two major possible explanations for these conflicting findings are that inhibition of SP1 preserved the structure of RGCs but did not restore visual function. Effect of inhibition of SP1 on F-VEP in rats after optic nerve injury. A, Average N1-P1 amplitude at 1 d pre-ONC, and at 7, 14, and 28 d post-ONC. There were no significant differences between the ONC group and SP1 shRNA group at any time point (n = 6, means ± SD, by RM one-way ANOVA with the Greenhouse-Geisser correction and Tukey multiple comparison test, **P < .01). B, Average N1 latency component at 1 d pre-ONC, and at 7, 14, and 28 d post-ONC. There were no significant differences between the ONC group and SP1 shRNA group at any time point (n = 6, means ± SD, by RM one-way ANOVA with the Greenhouse-Geisser correction and Tukey multiple comparison test, *P < .05, **P < .01) formulate a neuroprotective strategy for glaucoma involving stimulation of the visual system.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.