Cell-Specific Expression of Human SIRT1 by Gene Therapy Reduces Retinal Ganglion Cell Loss Induced by Elevated Intraocular Pressure

Abstract

SIRT1 prevents retinal ganglion cell (RGC) loss in several acute and subacute optic neuropathy models following pharmacologic activation or genetic overexpression. We hypothesized that adeno-associated virus (AAV)-mediated overexpression of SIRT1 in RGCs in a chronic ocular hypertension model can reduce RGC loss, thereby preserving visual function by sustained therapeutic effect. A control vector AAV-eGFP and therapeutic vector AAV-SIRT1 were constructed and optimized for transduction efficiency. A magnetic microbead mouse model of ocular hypertension was optimized to induce a time-dependent and chronic loss of visual function and RGC degeneration. Mice received intravitreal injection of control or therapeutic AAV in which a codon-optimized human SIRT1 expression is driven by a RGC selective promoter. Intraocular pressure (IOP) was measured, and visual function was examined by optokinetic response (OKR) weekly for 49 days following microbead injection. Visual function, RGC survival, and axon numbers were compared among control and therapeutic AAV-treated animals. AAV-eGFP and AAV-SIRT1 showed transduction efficiency of ~ 40%. AAV-SIRT1 maintains the transduction of SIRT1 over time and is selectively expressed in RGCs. Intravitreal injections of AAV-SIRT1 in a glaucoma model preserved visual function, increased RGC survival, and reduced axonal degeneration compared with the control construct. Over-expression of SIRT1 through AAV-mediated gene transduction indicates a RGC-selective component of neuroprotection in multiple models of acute optic nerve degeneration. Results here show a neuroprotective effect of RGC-selective gene therapy in a chronic glaucoma model characterized by sustained elevation of IOP and subsequent RGC loss. Results suggest that this strategy may be an effective therapeutic approach for treating glaucoma, and warrants evaluation for the treatment of other chronic neurodegenerative diseases.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13311-023-01364-6.

Introduction

Glaucoma, a chronic progressive optic neuropathy, affects 70 million people worldwide with 10% of affected individuals rendered irreversibly blind [1, 2]. Currently, glaucoma is the world’s leading cause of irreversible blindness, partially due to its difficulty to diagnose until after significant progression has occurred [3, 4]. The disease’s pathology is characterized by slow degeneration of retinal ganglion cells (RGCs), resulting in debilitating loss of vision [5] over time. One factor studied and treated to prevent the progression of this disease is elevated intraocular pressure (IOP) [6]. Treatment of the disease is currently limited to methods that reduce IOP, such as arduous topical medications, laser trabeculoplasty, and surgery [7, 8]. These attempts can be helpful to slow the progression of the neurodegeneration, but do not address the cause of RGC death that occurs from glaucoma [9]. There is currently no treatment allocated to neuroprotection or aiding in the restoration of RGCs once the damage has occurred, severely limiting the ability to truly stop disease progression.

In many cases, glaucoma can be adequately controlled with medical treatment. Yet, despite availability of medications, 10% of patients still experience vision loss due to inadequate medical compliance, with rates of non-adherence to treatment ranging from 30 to 80% [10]. Furthermore, poor adherence has been correlated with worse glaucoma outcomes. There are also a large number of patients that still progress despite adequate IOP control, normal IOPs, and strict adherence to therapy [11]. There are likely many factors that contribute to the pathogenesis of this disease and subsequent loss of vision. However, RGC susceptibility to injury is a key interventional target in glaucoma therapy.

Herein, we investigated human SIRT1 (SIRT1) upregulation as a viable gene therapy approach for the treatment of chronic neurodegenerative diseases such as glaucoma. SIRT1, a sirtuin that encodes an NAD + -dependent deacetylase [12–14], has been shown to play a significant role in visual preservation during acute optic nerve injuries [15–18]. Molecularly, SIRT1 deacetylates and inactivates many key regulatory proteins that regulate inflammation, oxidative stress, and apoptosis, suggesting SIRT1 as a potential therapeutic target. Using pharmacologic and transgenic approaches, further studies demonstrated that upregulation of SIRT1 increased RGC survival in optic nerve crush (ONC) [15], inflammatory [19], and virus-induced optic neuropathies [20]. Furthermore, a recent study showed that upregulation of SIRT1 via AAV2-mediated gene transfer resulted in small but significant benefits in a model of optic nerve inflammation even though only 21% of RGCs were transduced with the AAV vector used as a delivery vehicle [21]. Subsequent studies found that increasing RGC transduction using a human, RGC-selective promoter and AAV7m8 capsid was able to reduce RGC loss induced by ONC [22]. However, the ability to treat chronic optic neuropathies like glaucoma remains to be examined.

In this study, we examined the effects of human SIRT1 overexpression in RGCs of mice with chronically elevated IOP induced by ferromagnetic microbead injection into the anterior chamber. We characterized the expression profile of an AAV7m8 vector that promotes RGC-selective expression of hSIRT1 in the mouse retina and examined its neuroprotective effects in the microbead occlusion-induced glaucoma model. Specifically, effects on the mitigation of RGC loss, axon death, and visual deterioration were examined.

Methods

Animals

C57B1/6 J mice comprised of both males and females were purchased from the Jackson Laboratory (Bar Harbor, ME). The animals were stored in the University of Pennsylvania’s vivarium. Treatment of the animals complied with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research, along with institutional and federal regulations designated by the Institutional Animal Care and Use Committee at the University of Pennsylvania. Mice were fed ad libitum and maintained on a 12-h/12-h light/dark cycle.

Optokinetic Response Recordings (OKRs)

Visual function was assessed by measuring the OKR using commercial software and apparatus (OptoMetry; CerebralMechanics, Inc., Medicine Hat, AB, Canada) as previously described [21]. OKR was determined as the highest spatial frequency where mice track a 100% contrast grating that is projected at different spatial frequencies. Briefly, mice were placed unrestrained on a platform in the center of four computer monitors with a video camera above the platform to capture the movement of the mouse. A rotating cylinder with vertical sine wave grating was computed and projected to the monitors. The sine wave grating, consisting of black and white bars at 100% contrast and 12 degrees/second, provides a virtual-reality environment to measure the spatial acuity of the left eye when it rotates clockwise and the right eye when it rotates counterclockwise. Thus, the asymmetry of the optokinetic motor reflex allows for testing of visual function for each eye separately. Testing was performed by a masked investigator and OKR responses were recorded for both eyes.

AAV Vector Design

Vectors were produced as described previously [4, 21] and further characterized. Briefly, human codon-optimized SIRT1 cDNA clones were ordered from GeneScript. SIRT1 cDNA sequences were amplified and cloned into an AAV expression plasmid using the In-Fusion HD commercial cloning kit (Clontech Laboratories, Mountain View, CA, USA). Transgene expression was promoted through the addition of a γ-synuclein promoter (SNCG). The selected cDNA sequences contained a C-terminal 3xFLAG epitope tag that terminates into a bovine growth hormone polyadenylation sequence. AAV-expression sequences were flanked by AAV7m8 inverted terminal repeats. A pro-viral plasmid containing identical cis regulatory elements drove the expression of enhanced green fluorescent protein (eGFP). AAV7m8-SNCG.SIRT1 and AAV7m8-SNCG.eGFP vectors were generated using previously described methods and purified with a CsCl gradient by the Center for Advanced Retinal and Ocular Therapeutics research core at the University of Pennsylvania [21, 22].

Intravitreal (IV) Injection

A cohort of 4-week-old mice was anesthetized using isoflurane inhalation. A 33 ½ gauge needle was used to create a small incision slightly posterior to the limbus. Following this, a 10-uL Hamilton syringe (701 RN; Hamilton Company, Reno, NV, USA) attached to a 33-gauge blunt-end needle was inserted into the vitreous cavity with the needle tip placed directly above the optic nerve head. A volume of 2 uL of AAV preparation containing 1 × 10¹⁰ vector genomes was injected into each eye. Both eyes of each mouse received either AAV7m8-SNCG.SIRT1 or AAV7m8-SNCG.eGFP allowing each eye to serve as an independent endpoint for raised IOP or no raised IOP following microbead injection into one eye only. This also was chosen to avoid transsynaptic contamination of vectors [23].

Microbead Preparation

Stock magnetic microbead (MB) solution (4.5 μm diameter, 4 × 10⁸ beads/mL; Dynabeads M-450 Epoxy, ThermoFisher Scientific, Cat# 14,011, Waltham, MA, USA) was vortexed to evenly distribute MBs in solution. Next, 1 mL of the stock MB solution was pipetted into 50 mL of 0.02 M sodium hydroxide (NaOH, MW 39.997 g/mol) in Tris 10 × buffer (MW 121.4 g/mol) and rotated for 24 h at RT to remove protective epoxy groups from the MBs. A magnet was attached to the bottom of the tube to collect the magnetic MBs. The tube was then oriented horizontally and rotated for an additional 4 h at RT. The supernatant of the MB solution was removed using a micropipette, while the magnetic MBs were resuspended in 50 mL of Tris 10 × buffer until evenly resuspended. MBs were washed three times with 5 mL of ultra-pure laboratory-grade water by vortexing for 2 min. A magnet was attached to the bottom of the tube to collect the MBs. MBs were further washed three times in a laminar flow hood with BSS by aspiration with a micropipette. MBs were resuspended in BSS to a final concentration of 1.6 × 10⁶ beads/μL to allow 2.4 × 10⁶ beads to be injected into the anterior chamber with a final volume of 1.5 μL. The MBs were stored at 4 °C until usage.

Magnetic Microbead Occlusion Model

A well-published magnetic MB mouse model of glaucoma was used to induce and sustain elevated intraocular pressure (IOP) [24, 25]. Briefly, all mice received an intracameral injection of 1.5 uL of magnetic MBs in either the right or left eye (selected randomly), with the fellow eye injected with sterile normal saline. The investigator injecting the MBs was masked to which AAV vector (eGFP or hSIRT1 expressing) had been injected in each eye. Mice were injected at day 0 and then again at day 24. IOPs were measured immediately post-injection and then weekly using Icare TONOLAB tonometry (Icare TONOVET, Vantaa, Finland). If the 1-week IOP measurement was not increased by 2 × of the baseline IOP, these eyes were removed from the experiment. Roughly 15% of mice did not have an associated and expected IOP spike. An average of 3 measurements per eye were used at each time point and at the final endpoint.

Retinal Immunohistological Analysis

Eyes were harvested from mice, placed in 4% paraformaldehyde (PFA) (Electron Microscopy Sciences, Cat# 15,710) for 1 h at room temperature, and washed in 1X PBS followed by dissection of retinas. Tissues were permeabilized with 2% Triton X-100 and blocked with 10% normal bovine serum. The blocked tissues were washed with 1X PBS and then incubated with rabbit anti-Brn3a antibody (Synaptic systems, Cat# 411 003) diluted 1:1000 overnight at 4 °C. For hSIRT1 staining, a goat anti-hSIRT1 antibody (ProSci, 42;428) was similarly incubated at 1:1000 dilution. For experiments including only Brn3a staining, retinas were washed and then incubated for one hour at room temperature in secondary antibody solution containing 2% Triton X-100, 10% normal bovine serum, and donkey anti-rabbit AlexaFluor 594 antibody diluted 1:1000. For dual staining of Brn3a and hSIRT1, retinas were washed and incubated with donkey anti-rabbit AlexaFluor 488 and donkey anti-goat AlexaFluor 594 (1:1000 dilutions each) for 1 h at room temperature. After washing, samples were flat-mounted onto glass slides with an aqueous mounting medium (SouthernBiotech, Birmingham, AL) containing 4′,6-diamidino-2-phenylindole (DAPI). RGCs were quantified as previously described [18, 22, 25]. Briefly, retinal micrographs were recorded at 40 × magnification in 12 standard fields (1/6, 3/6, and 5/6 of the retinal radius from the center of the retina in each quadrant). Total RGC counts from the 12 fields per retinal sample covered a total area of 0.407 mm²/retina. Brn3a-positive and hSIRT1-positive cells were counted by an investigator masked to the experimental conditions.

Optic Nerve Axon Staining and Quantification

Optic nerves were bisected at the level of the optic nerve head and fixed in 2.5% glutaraldehyde with 2.0% paraformaldehyde in a 1 M sodium cacodylate buffer at a pH of 7.4. RGC axons were stained and manually counted, adapted from previously described methods [22]. Briefly, optic nerves were further fixed in 2% osmium tetroxide and the tissue was dehydrated in increasing concentrations of ethanol. After the dehydration step, the tissue was embedded in epoxy resin Embed 812 (Electron Microscopy Science, Hatfield, PA), cut in 0.75 uM cross sections, and stained with 1% toluidine blue. Sections were obtained at approximately 1 mm from the proximal tip of the optic nerve head. The majority of each optic nerve section was captured with one centrally oriented image at 40 × magnification and axons were manually counted by a masked observer using ImageJ software. Three adjacent sections were counted for each optic nerve. Representative images were obtained with 60 × objective and then magnified four-fold to illustrate the axon density.

RT-PCR retinal tissues were dissected and then preserved in RNAlater™ Stabilization Solution (Cat#: AM7020, Invitrogen) at − 80 °C. Total RNA was isolated using PureLink^® RNA Mini Kit (Cat#: 12183025, Ambion by Life technologies). One microgram of total RNA, measured by NanoDrop 8000 Spectrophotometer (ThermoFisher, Waltham, MA) was converted into cDNA using a high-capacity cDNA reverse transcription kit with RNase inhibitor (Cat#: 4374966, Applied Biosystems). An equivalent amount of cDNA (10 ng based on 100% conversion from RNA) was used in each TaqMan duplex gene expression assay using human codon-optimized SIRT1 primers and probe set (manual design using IDT) and mouse Pou4f1 pre-designed TaqMan assay (Mm.PT.58.30469172, IDT). Relative fold change in mRNA expression level at different timepoints (1 week, 2 weeks, 4 weeks post intravitreal injection) was measured by real-time PCR via the 2^−ΔΔCT method.

Human codon-optimized SIRT1 assay:

• Probe: 5′-/56-FAM/TCCACACGA/ZEN/AGTGCCTCAAATCCT/3IABkFQ/-3′

• Primer 1: 5′- GACGAAGTGGATCTCCTGATTG-3′

• Primer 2: 5′- GCAGTGGTTCCCGGTTAATA-3′

Mouse Pou4f1 assay:

• Probe: 5′-/5SUN/ACCCTCCTC/ZEN/AGTAAGTGGCAGAGA/3IABkFQ/-3′

• Primer 1: 5′-CAACCAGAGACAGAAGCAGAA-3′

• Primer 2: 5′-ACAACTTCACGGTTCTCCAA-3′

Statistics

All data are represented as mean ± standard error of mean (SEM). Differences between treatment groups with respect to OKR responses, RGC quantification, and optic nerve histopathology were compared using either the Mann–Whitney U test, one-way ANOVA, or ANOVA of repeated measures followed by Tukey’s honest significant difference test, as indicated, using statistical software (GraphPad Prism 9.0; GraphPad Software, Inc., La Jolla, CA, USA). Differences were considered statistically significant at P < 0.05. Data meets the assumption of normal distribution of tests with variances between GFP and hSIRT1 groups.

Results

Design and Characterization of Reporter and Therapeutic Vectors

A reporter and therapeutic AAV vector (AAV7m8), expressing eGFP and hSIRT1, respectively (Figs. 1A and 2A), were characterized. The RGC-selective SNCG promoter drives the target gene expression in both vectors. Vector translational expression of eGFP was determined by in vivo IV injection of the reporter construct in 4–5-week-old male and female mice and subsequent visualization with color (Fig. 1B), red-free (Fig. 1C), and fluorescent (Fig. 1D) fundus photography. Cross-sections of retinas revealed no damage to retinal tissues at any layers (red filtered images) and showed cell-selective expression of eGFP in Brn3a stained RGCs (Fig. 1E and F). Quantification of eGFP-positive RGCs in retinal whole mounts demonstrated a 42% transduction efficiency (Fig. 1G).

Fig. 1.

Design and characterization of AAV7m8-SNCG-eGFP reporter vector. (A) Outline of proviral expression cassette for reporter/control target construct used in the study. (B–D) In vivo imaging of a representative retina post-intravitreal injection of 1 × 10.¹⁰ vg of AAV7m8-SNCG-eGFP showing fundus photography (B) of intact retina, absence of any autofluorescence confirming lack of inflammatory cells or tissue damage post intravitreal injection (C), and eGFP expression in retina from reporter vector (D). Ex vivo retinal cross-section demonstrating intact tissue and all layers of the retina (E) (red-Brn3a, green-transduced eGFP expression, blue-DAPI nuclear stain) and whole mount staining of a representative retina (F) demonstrate that vector-driven eGFP expression is highly selective for Brn3a-positive RGCs within the ganglion cell layer. Quantification of RGC transduction (G). Retinas isolated from both eyes of two mice transduced with vector (n = 4 eyes) show 42% RGC transduction efficiency (eGFP – green expressing cells as a percentage of Brn3a-red positive cells)

Fig. 2.

Design and characterization of AAV7m8-SNCG-hSIRT1 therapeutic vector. (A) Outline of proviral expression cassette for therapeutic SIRT1 target used in the study. (B) RT-qPCR analysis of relative quantities of human SIRT1 mRNA in mice after intravitreal injection (n = 6 retinas from 3 mice/group) compared with non-transduced wild-type mouse retinas shows high levels of expression by 2 weeks post-injection with continued increase at 4 weeks post-injection. (C) Immunofluorescent staining of human SIRT1 protein (red) and Brn3a expression (green) in a representative control/wild-type non-transduced retina demonstrates no expression of hSIRT1, whereas hSIRT1 staining in a transduced retina demonstrates clear expression of hSIRT1 within RGCs. Fluorescence microscopy images were taken at 10 × and 20 × magnification. (D) Quantification of RGC transduction of hSIRT1 in AAV-SNCG-hSIRT1 treated and untreated eyes. Retinas isolated from eyes transduced with vector (n = 5 eyes) shows 42% RGC transduction efficiency (hSIRT1-red expressing Brn3a-green positive cells). Untreated eyes (n = 5) had no detectable hSIRT1 immunostaining. The average number of RGCs was equivalent between untreated and AAV-SNCG-hSIRT1 treated eyes. Data represented as mean ± SEM. ****p < 0.0001 by ordinary one-way ANOVA with Tukey’s HSD post-test, n.s. non-significant

A therapeutic vector designed to express codon-optimized human SIRT (hSIRT1) (Fig. 2A) was intravitreally injected in 4–5-week-old male and female mice. Mice were subsequently sacrificed at week 1, week 2, and week 4 post-injection. Transduced retina displayed increasing mRNA expression in a time-dependent fashion demonstrating successful expression of the codon-optimized hSIRT1 transgene by week 2, as well as prolonged and increased expression at 4 weeks post-injection by qRT-PCR analysis (Fig. 2B). Retinas from mice with no vector injected did not display any expression of hSIRT1 protein, while retinas transduced with AAV-hSIRT1 demonstrated extensive hSIRT1 protein expression with immunofluorescent antibodies specific for human-SIRT1 (Fig. 2C and D). Quantification of transduced RGCs demonstrated a 42% transduction efficiency (Fig. 2D) and there was no RGC toxicity.

In Vivo Injection of Magnetic Microbeads in Mice Induces Elevated IOP with Loss of Visual Response and Reduction of RGC Number

C57Bl/6J mice received injections of either magnetic MBs or BSS in opposite eyes selected at random as experimental and control samples, respectively. Examination of eyes injected with nothing, BSS, and MBs in vivo and post-mortem demonstrated intact visual axis as well as deposition of MBs into the anterior angle of the eye (MB-injected). IOP was measured once a week for 7 weeks after intracameral injection of BSS or MBs, with significant and sustained elevation in IOPs being demonstrated in MB-injected eyes compared to BSS-treated eyes. This IOP elevation was apparent starting 1 week following injection, persisting until the end of the 7-week observation period (Fig. 3A). The visual function of the treated mice was measured by OKR weekly, demonstrating a significant decline in OKR scores in the MB-injected eyes as compared to BSS treated eyes (Fig. 3B). There was no difference in average IOP levels or OKR responses between male and female mice injected with MBs (data not shown); thus, subsequent histologic analyses were performed using eyes from mice of both genders. Additional sets of mice were euthanized at either 1, 6, or 8 weeks after MB injection (N = 4 mice/group), and RGC counts were tabulated across 8 weeks (Fig. 3C, D). Statistically significant declines in RGCs were noted at 6 and 8 weeks in retinas from MB-injected eyes as compared to BSS-treated eyes (BSS-treated = 4142 ± 327 RGCs/mm²; MB-treated (1 week) = 3458 ± 118; p = 0.0571; MB-treated (6 weeks) 2319 ± 94; p = 0.0286; MB-treated (8 weeks) = 2160 ± 186; p = 0.0230; Mann–Whitney U test.

Fig. 3.

Characterization of the magnetic MB mouse model, and effect of gene transfer on aqueous humor physiology and visual acuity. C57Bl6/J (WT) mice were injected with MBs in one eye and BSS in the contralateral eye. (A) Intraocular pressure measurements in WT animals show a significant difference between BSS (N = 10) treated eyes as compared to MB (N = 12) treated eyes. (B) OKR analysis of mice after intracameral injection of BSS and MBs suggests MB-induced chronic elevation of IOP results in decreased visual function over time. (C, D) Quantitative (C) and representative (D) RGC counts by nuclear Brn3a (red) staining shows a decrease in RGC number over time. Fluorescence microscopy images were taken at 40 × magnification. There is a significant decrease in RGC counts at 6 weeks and 8 weeks post IOP elevation. (E) C57Bl6/J mice were injected with AAV-eGFP in one eye and AAV-SIRT1 in the contralateral eye at 4 weeks of age. At 8 weeks of age, disease was induced by intracameral injection of magnetic MBs in one eye and mice were monitored until termination at 8 weeks. IOP measurements obtained weekly (n = 10 eyes per condition) show a statistical difference between BSS (gray) treated eyes as compared to MB (red) treated eyes, and no difference between MB injected eyes treated with AAV-eGFP as compared to eyes treated with AAV-SIRT1 constructs. (F) Visual function was measured weekly OKR analysis of mice after both AAV-eGFP and AAV-SIRT1 injection and subsequent intracameral injection of BSS and MBs shows that chronic elevation of IOP results in decreased visual function over time (n = 10 eyes per condition) in mice that received control vector AAV-eGFP, whereas eyes that received AAV-SIRT1 showed significant preservation of OKR responses. Data represented as mean ± SEM. *p < 0.05, **p < 0.01 by a Mann–Whitney U test of RGC counts, and ANOVA of repeated measures of OKR scores and IOP

hSIRT1 Expression Demonstrates Preservation of Vision and RGC Protection in Mice with Elevated IOP

C57Bl/6J mice treated with AAV-eGFP vector in one eye and the experimental AAV-hSIRT1 construct in the contralateral eye at 4 weeks of age were injected with either MB or BSS 4 weeks later, and IOP was measured weekly across a period of 7 weeks (Fig. 3E). The IOPs of MB-treated eyes were significantly higher, as compared to those of the BSS-treated eyes. IOP increased 1 week after MB injection and remained elevated throughout the time period of the experiment. There was no significant difference between IOPs in animals treated with AAV-eGFP as compared to AAV-hSIRT1, as MB injection induced elevated IOP in eyes treated with both vectors (Fig. 3E). Visual function was assessed by OKR (Fig. 3F). Within the MB-injected groups, there was significant preservation of visual function in AAV-hSIRT1-treated mice as compared to AAV-eGFP-treated mice starting at the 35-day time point (Fig. 3F) as well as over time assessed by ANOVA of repeated measures. Mice were euthanized 8 weeks after MB injection, and quantification of RGC numbers (Fig. 4A, B) showed no effect of vector treatment on RGC numbers in BSS-injected mice; whereas loss of RGCs induced by MB injection was significantly attenuated by treatment with AAV-SIRT1 (BSS AAV-eGFP = 3766 ± 163 RGCs/mm²; BSS AAV-hSIRT1 = 3538 ± 128; MB AAV-eGFP = 1931 ± 120, MB AAV-hSIRT1 = 2705 ± 96). Axon counts at the proximal end of the optic nerve head were quantified, demonstrating a significant difference between the MB, AAV-eGFP group and the MB, AAV-hSIRT1 group (Fig. 4C, D) (MB, AAV-eGFP 1.8 × 10⁴ ± 1533, MB-AAV-hSIRT1 2.6 × 10⁴ ± 1052, *p = 0.011; one-way ANOVA followed by Tukey’s honest significant difference test). Lastly, there was no significant difference in RGC axon densities between eyes treated with BSS, AAV-eGFP and BSS, AAV-hSIRT1.

Fig. 4.

Effect of gene transfer on RGC soma and axons. Mice shown in Fig. 3E and F were euthanized 8 weeks after MB disease induction and RGC numbers were quantified (A) in retinal whole mounts (N = 5 per group) stained for Brn3a from mice treated with AAV-eGFP (gray bars) and AAV-SIRT1 (red bars) with normal and elevated IOP. There is a significant decrease in RGC counts in mice with sustained IOP without AAV-SIRT1 treatment (gray bars) compared with those that received AAV-SIRT1 (red bars) (A). Photographs of Brn3a-positive RGCs in representative retinas illustrate that MB injection induces a decrease in RGC numbers only when IOP is successfully elevated, and AAV-SIRT1 attenuates this RGC loss. Fluorescence microscopy photographs were taken at 40 × magnification. (B) Axon counts are shown in optic nerves from sections taken 1 mm behind the optic nerve head in mice treated with AAV-eGFP (gray bars) or AAV-SIRT1 (red bars). There is a significant decrease in axon density in BSS-treated eyes as compared to that of MB-induced ocular hypertension eyes, with significant preservation of axon numbers in MB-induced ocular hypertension eyes of mice treated with AAV-SIRT1 as shown in counts (C) and representative photographs (D). Data represented as mean ± SEM. *p < 0.05, **p < 0.01 by ordinary one-way ANOVA with Tukey’s HSD post-test

Discussion

Glaucoma is the most common neurodegenerative disease that significantly affects vision [26]. The loss of vision is intricately connected to elevated IOP, but this factor is not necessarily the cause [9]. To date, the only therapies that can be offered to patients presenting with this diagnosis are means to lower the IOP. This modification slows the progression of the disease in roughly 60–70% of patients; however, a population of patients will progress despite physiologically low IOPs [9]. This suggests a need to better understand the mechanism of RGC death, but also to develop and test new therapies that can preserve visual function and extend RGC survival to prevent unnecessary vision loss that is independent of IOP levels, as was demonstrated in the current experiments.

There are many murine models of glaucoma currently used in the literature separated into genetic models or those with a disruption in normal aqueous humor physiology [27]. While genetic models are excellent at teasing out specific signaling and genetic mutations that contribute to the disease, we chose an inducible and controllable model for a more generalizable disease model in which to examine neuro-protective effects of SIRT1 signaling. In previous studies, our construct has shown promise reducing the visual decline and RGC degeneration following optic nerve crush, an acute, harsh model of RGC injury, but it was not known whether a similar treatment could prevent RGC loss in the presence of a sustained insult such as persistent elevated IOP. Thus, the MB model was used and shown here to induce chronic IOP elevation with decreased visual function and time-dependent loss of RGCs, similar to prior studies [25].

Our data showed that an RGC-selective AAV (AAV7m8) vector expressing eGFP under the control of the SNCG promoter was able to transduce ~ 40% of RGCs, while maintaining a healthy retina. Similarly, AAV7m8 carrying a SNCG promoter-driven therapeutic transgene was able to drive expression of the human SIRT1 gene sustainably for 4-week post intravitreal injection, transducing ~ 40% of RGCs. This transgene overexpression was not toxic to RGCs as RGC number, quantified by Brn3a, were equivalent in AAV-hSIRT1 treated and untreated eyes, similar to prior studies [28, 29]. While prior studies showed similar expression levels of this eGFP-expressing vector [22], hSIRT1 was not previously studied due to the lack of human specific antibodies that are now available and used in the current study. Results showed that hSIRT1 expression is expressed as efficiently as the eGFP reporter. Furthermore, our MB model shows similar degrees of IOP elevation and RGC loss as previous studies did [24, 25, 30]. We observed some disparity in the extent of IOP increases among animals of the same groups. Such variability, also reported in other studies by other investigators, can be attributed to the heterogeneity of MB diameter and/or physiological differences among injected mice [24, 25, 30].

In our study, RGC survival was quantified through the determination of Brn3a⁺ cell density in the retina, and functional integrity was determined by OKR measurement. We recognize that several markers of RGCs in the retina have been used in the literature including Brn3a; this was selected based on the laboratory’s expertise in the use of this marker. There is some evidence of Brn3a downregulation with elevated IOP, along with demyelination effects on the axons of RGCs [31, 32]. However, the literature fails to provide sufficient evidence against the effectiveness of Brn3a as a marker for RGCs. In the case of the present study, we attribute Brn3a downregulation to be a result of RGC death from elevated IOP within the eye.

The AAV-SIRT1 construct was protective for visual function, RGC somatic degeneration, and axonopathy observed in the MB-induced glaucoma model. This neuroprotective effect seems to be independent of IOP levels since IOP remained elevated in AAV-SIRT1-treated mice. Thus, this SIRT1-based therapy may be effective in patients that fail to respond to IOP-lowering treatments. While the mechanisms underlying the therapeutic effects of AAV-SIRT1 have not been explored in this study. Further experimentation in this direction could yield new insights into the survival pathways mediated by SIRT1-based gene therapy. However, findings introduce a novel strategy of RGC-specific protection using targeted gene therapy treatment. Another limitation of the current study is that gene therapy was administered prior to inducing IOP elevation. While treatment would not be given until after disease onset in human patients, due to the slow progressive nature of glaucoma, gene therapy would still likely be an option prior to the degeneration of most RGCs. Thus, we believe current results support further exploration of this strategy as a treatment for glaucoma, and future studies will include treatment initiated at later time points as well.

In human disease, vision loss and loss of visual acuity is the undesirable and irreversible outcome of glaucoma [27]. This clinical manifestation is associated with degeneration of RGCs with subsequent loss of RGC soma, dendritic loss, and finally axonal loss and optic nerve atrophy [33]. Because of this clinical finding, our study focused on three main endpoints to test for therapeutic efficacy: vision loss, as evidenced by OKR decline; RGC degeneration, by RGC reduction in retinal whole mounts; and axon loss, by reduction of axon counts 1 mm behind the optic nerve head. Our results demonstrate that prolonged IOP elevation results in vision loss, RGC loss, and axonopathy, which clinically mirrors human disease. We also show that treatment of mice with AAV-eGFP or AAV-SIRT1 has no effect on IOP or visual function in wild type mice. In these experiments, there was no difference in the average IOP between mice treated with MBs that received AAV-eGFP as compared to those that received AAV-SIRT1. Each of the groups treated with MBs had a decline in OKR responses. However, treatment with therapeutic vector, AAV-SIRT1, by 35 days, was able to preserve 55% of the visual function that would otherwise be lost as a result of chronically elevated IOP. This corresponded to a significant preservation of Brn3a positive RGCs as well as preservation of total axons compared with animals injected with AAV-eGFP. Thus, by enhancing SIRT1 signaling, we are able to reduce, in part, the neurodegenerative effects of IOP induced glaucomatous optic neuropathy.

The degree of RGC survival induced by this gene therapy is notable, especially given that under half (about 40%) of Brn3a-positive cells were transduced with the vector. Of note, Brn3a identifies only up to 90% of the total RGCs, while 10% or more are missed [34], thus there could be small differences in the transduction rate not accounted for in the current study. Nonetheless, the significant degree of improved RGC survival and visual function suggest that perhaps many uncounted RGCs are also transduced, or it is possible that overexpression of SIRT1 in a subset of RGCs may lead to secreted factors that promote RGC survival even in non-transduced cells.

The use of gene therapy to treat glaucoma in an IOP-independent fashion represents a novel treatment strategy for a common and debilitating disease. To date, most gene therapies for the treatment of retinal disease have been to replace known genetic mutations that contribute to the pathology. We have shown that novel enhancement of the SIRT1 pathway, in an RGC-selective fashion, can rescue the neurodegenerative effects of IOP-dependent damage without modulating the IOP itself. This finding suggests a potential broad strategy to target neuronal loss induced by multiple different insults besides strictly genetic disease. Results suggest future studies may be warranted to examine the potential of gene therapy to overexpress SIRT1 in specific neuron populations in other neuronal degenerations inside and outside of the eye.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This study is supported by National Institutes of Health Grants (EY019014, EY301163), RWJ-Harold Amos Faculty Development Award, Linda Pechenik Montague Investigator Award, Foundation Fighting Blindness, Research to Prevent Blindness, Paul and Evanina Mackall Foundation Trust, Center for Advanced Retinal and Ocular Therapeutics, and the F. M. Kirby Foundation.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

AGR and KSS hold intellectual property relevant to this study, and receive research funding from Gyroscope Therapeutics which was not used in this study. The other authors certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.