Six-week-old female SJL/J mice were purchased from the Jackson Laboratory (Bar Harbor, ME) for induction of experimental optic neuritis. For isolation of primary retinal cells, 3–5-day-old C57/Bl6 mouse pups were removed from litters of timed-pregnant mice (Jackson). Treatment of animals was reviewed and approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania.

EAE was induced as previously described (Shindler et al., 2006). Briefly, eight-week-old female SJL/J mice were anesthetized with ketamine/xylazine and were injected subcutaneously at two sites on the back with 0.1 mL solution containing 0.5 mg/mL proteolipid protein peptide 139–151 (GenScript, Piscataway, NJ) emulsified in complete Freund adjuvant (CFA; Difco, Detroit, MI) containing 2.5 mg/mL Mycobacterium tuberculosis (Difco). Control mice were injected with equal volumes of PBS and CFA. All mice were injected with 200 ng pertussis toxin (List Biological, Campbell, CA) in 0.1 mL PBS intraperitoneally on day 0 (day of immunization) and again on day 2.

Cultures of RGC-5 cells (ATCC, Manassas, VA) were maintained in growth medium containing low-glucose Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin (Sigma-Aldrich, St. Louis, MO) in a humidified atmosphere of 95% air and 5% CO₂ at 37°C, as described by Krishnamoorthy et al. (2001). Cells were passaged every 3–4 days, with a doubling time of 18–20 h. During each experiment, cells were plated at a concentration of 1 × 105 cells/mL and after 16 h treated with various concentrations of RSV (Sigma-Aldrich) or SRTAW04 (Sirtris, a GSK company, Cambridge, MA) dissolved in DMSO with or without H₂O₂ for 24 h. To obtain staurosporine-differentiated RGC-5 cells, cells were exposed to 1 μM staurosporine (Cayman Chemical, Ann Arbor, MI) for 6 h, washed three times with PBS and then recovered in culture medium for 1 day. Control cells were treated with DMSO only. DMSO concentration in each well was below 0.5% v/v. For doxorubicin experiments, cells were plated at 1 × 105 cells/mL and after 16 h treated with 1 μM of doxorubicin (Sigma, St. Louis, MO) for another 24 h. For serum starvation experiments, media were replaced with serum free media and cells were grown for 24 h. Cell viability was measured by trypan blue staining and counted on a hemocytometer. Alternatively, cell viability was assessed using PrestoBlue™ Cell Viability Reagent (Invitrogen, Grand Island, NY). Twenty-two microliter was added to each well and incubated for 20 min, and fluorescence was read on a plate reader excited at 570 nm with non-conjugated light emissions collected at 610 nm.

The retina was removed from 3- to 5-day old C57/Bl6 mice and cells were dissociated as described previously for neonatal rat retina (Heng et al., 1999). Briefly, retinas were dissociated in solution containing 0.45 U papain (Worthington, Lakewood, NJ) for 30 min at 37°C. After digestion, the retinal tissue was rinsed twice with DMEM before being transferred into Neurobasal medium containing 1% BSA (Sigma) B27-supplement (1:50, Gibco) and penicillin/streptomycin (Gibco). Then, tissue was triturated and passed through a cell strainer before seeding of the isolated cells onto poly-D-lysine (0.1 mg/ml, molecular weight <300,000 Da, Sigma) and laminin (20 μg/ml, Sigma) coated chamber slides. To quantify RGC survival, chamber slides were washed twice with PBS and fixed in 4% paraformaldehyde for 15 min. Fixed cells were permeabilized using 0.3% Triton X-100 and blocked with 1% BSA in PBS for 1 h followed by incubation with anti Brn3a antibody (Santa Cruz biotechnology, Santa Cruz, CA) over night at 4°C. After washing with PBS, fluorescence-labeled secondary antibodies (Invitrogen) were applied for 1 h at room temperature (RT). Finally, cells were embedded in mounting medium containing DAPI (Vector labs, Burlingame, CA), pictures were taken using a fluorescence microscope (Eclipse E600; Nikon, Tokyo, Japan), at 10× magnification and the Brn3a specific RGCs were counted by a masked investigator.

MitoSOX Red (Invitrogen) mitochondrial superoxide indicator is a fluorogenic dye for selective detection of superoxide in the mitochondria of live cells. For RGC-5 cultures, MitoSOX reagent was diluted to a final concentration of 3 μM in warm DMEM 1% FBS and added to the cells. After 15 min incubation at 37°C, cells were fixed with 4% paraformaldehyde for 10 min, mounted onto glass slides with Mowiol mounting medium, and observed under a Eclipse E600 (Nikon, Tokyo, Japan) fluorescence microscope using excitation 510 nm/emission 580 nm. For optic nerves, mice were anesthetized with ketamine/xylazine and the optic nerves were removed, washed with PBS and incubated in 5 μM MitoSOX Red for 30 min at 37°C. After incubation, nerves were washed three times with PBS and mounted in OCT. Five micromole cross-sections were made, viewed by fluorescent microscopy, and photographed at 20× magnification. One photograph centered within each optic nerve cross section was taken by a blinded investigator using a standard exposure, and staining was quantified by calculating the optical density using Image J software (nih.gov).

Cell cultures were washed twice with PBS, trypsinized for 5 min, diluted with 10× volume of PBS and centrifuged at 2000g for 10 min. Pellets were lysed in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% desoxycholic acid, 0.1% SDS and 50 mM Tris, pH 8) to obtain total protein, and protein content was measured with the BCA protein assay kit (Thermo scientific, Rockford, IL). Proteins (30 μg) were separated by polyacrylamide gel electrophoresis using 10% SDS–polyacrylamide gels under reducing conditions, then transferred with semidry blotting (Bio-Rad, Hercules, CA) to Nitrocellulose High bound ECL membranes (GE Healthcare Biosciences, Pittsburgh, PA). Membranes were blocked with Odyssey Blocking Buffer (Licor Biotechnology, Lincoln, NE) for 1 h at RT. Rabbit polyclonal antibodies against SIRT1 (AbCam, Cambridge, MA) were used at a 1:1500 dilution overnight at 4°C. As secondary antibody, IRDye® 800CW Goat anti-rabbit IgG (Licor) was incubated at a 1:5000 dilution for 1 h at RT. Rabbit polyclonal SOD2 (GeneTex, Irvine, CA) and mouse monoclonal succinate dehydrogenase b (SDHb) (Abcam, Cambridge, MA) antibodies were used at 1:1000 dilutions. As secondary antibody, IRDye® 800CW goat anti-rabbit IgG and IRDye® 600 goat anti-mouse IgG were used at a dilution of 1:5000. Fluorescence was visualized using Odyssey infrared imaging system (Licor). For normalization of signals, blotted membranes were stained for β-actin (Sigma). The intensity of each band was determined using Image J software (nih.gov).

PGC-1α was immunoprecipitated from cell extract (500 μg of nuclear protein) using anti-PGC-1α antibody (Santa Cruz). Immunoprecipitated PGC-1α was electrophoresed in SDS-PAGE and immunoblotted with antibody specific for acetylated lysine (Cell Signaling Technology, Boston, MA) to detect acetylation, or antibody specific for PGC-1α (Novus Biologicals, Littleton, CO) to detect total PGC-1α levels.

JC-1 (Immunochemistry Technologies, Bloomington, MN) staining was performed according to manufacturer's instructions to assess mitochondrial membrane potential (Cossarizza et al., 1993). In healthy cells due to higher mitochondrial membrane potentials, JC-1 forms fluorescent “J aggregates” that appear red. However, JC-1 exists as a monomer at low mitochondrial membrane potential, as seen in apoptotic cells, and appears green. Thus, the emission of this cyanin dye can be used as a measure of mitochondrial membrane potential. A stock solution of JC-1 was prepared at 4 mg/mL in dimethylsulfoxide (DMSO). The stock JC-1 solution was added drop-wise, while vortexing, to control medium to a final concentration of 10 μg/mL. The diluted JC-1 solution was then passed through a 0.2-μM syringe filter (GeneMate, Kaysville, UT) and added to RGC-5 cultures for 15 min at 37°C. After incubation, staining solutions were decanted, each dish was washed three times with Ringer's buffer [in mM; 130 NaCl, 5 KCl, 2 CaCl₂.2H₂O, 1 MgSO₄, 8 NaOH, 1 NaPO₄, 5.5 D-glucose (pH 7.4)] at 37°C, and cells were left in fresh Ringer's buffer. Live cell images were acquired using the Argon laser on a fluorescent microscope (Eclipse E600; Nikon). JC-1 was excited at 488 nm and non-conjugated light emissions were collected at 530 nm (green) and conjugated at 590 nm (red). The level of JC-1 staining was quantified by photographing cultures with a standard exposure and calculating the optical density of staining using Image J software.

TMRM (Immunochemistry Technologies, Bloomington, MN) staining was performed according to manufacturer's instructions. Differentiated RGC-5 cells after treatment were incubated with 100 nM of TMRM for 15 min and washed twice with washing buffer. Live cell images were acquired using the fluorescent microscope (Eclipse E600; Nikon). The level of TMRM staining was quantified by photographing cultures with a standard exposure and calculating the optical density of staining using Image J software.

SIRT1-specific siRNA, and control siRNA were Silencer select predesigned siRNAs from Life Technologies (Invitrogen) and experiments were performed according to manufacturer's instructions. Briefly, mouse primary retinal cells or differentiated RGC-5 cells were treated with 25 nm of either SIRT1 specific siRNA or control siRNA using lipofectamine RNA iMAX transfection reagent. 24 h later, the cells were treated with H₂O₂ with or without drug treatment and cell survival was assessed 48 h after transfection.

Data are expressed as means ± SEM. For cell viability, SIRT1 activity, and western blot analysis, differences between culture groups were assessed using One-Way ANOVA followed by Student Neuman–Keuls post-hoc test. Statistical differences were considered significant at P < 0.05.

MitoSOX Red detection of superoxide within mitochondria was used to confirm prior studies suggesting a role of ROS accumulation in optic neuritis (Qi et al., 2007) in mice with EAE, a model of MS. EAE was induced in female SJL/J mice by immunization with proteolipid protein peptide, and mice were sacrificed 11 days later, when optic nerve inflammation is known to peak (Shindler et al., 2006, 2008). Ten optic nerves of 5 EAE mice and 5 control mouse optic nerves were isolated and incubated with MitoSOX Red. Fluorescent microscopy of cryosectioned EAE specimens revealed an increase in the superoxide anion compared to control optic nerves (Figure 1).

MitoSOX staining was used to determine whether cultured RGC-5 cells demonstrate similar superoxide accumulation in mitochondria in response to various stressors, as seen in RGCs axons in EAE optic neuritis. Cell viability was also measured. RGC-5 cells were plated and incubated for 16 h prior to being stressed by removal of serum, or by addition of doxorubicin or hydrogen peroxide. Serum starvation of the RGC-5 cells showed a significant decrease in cell viability by 16 h after serum removal, with an associated increase in MitoSOX Red staining (Figures 2A,B). Treatment with 1 μM doxorubicin induced a significant decrease in RGC-5 cells compared to control cultures, beginning within 6 h of incubation, also with a robust increase in the superoxide staining (Figures 2C,D), and similar RGC-5 cell loss and ROS accumulation occurred in cultures treated with 500 μM hydrogen peroxide (Figures 2E,F).

Because RGC-5 cells are a transformed, dividing cell line, we next pretreated RGC-5 cells with staurosporine (1 μM for 6 h), which drives neuronal differentiation with sprouting of neurites (Frassetto et al., 2006) (Figure 3A). Serum withdrawal and doxorubicin induce cell loss and ROS accumulation in differentiated RGC-5 cells (data not shown) as in undifferentiated cells (Figure 2). We also directly introduced oxidative stress by treating with hydrogen peroxide and found a dose dependent loss of differentiated RGC-5 cells (Figure 3B). We used hydrogen peroxide (500 μM) for subsequent experiments to directly drive oxidative stress.

Previous studies showed SIRT1 activators significantly attenuate RGC loss in experimental optic neuritis (Shindler et al., 2007, 2010; Fonseca-Kelly et al., 2012), although molecular mechanisms of effects on neuronal survival are not well understood. We investigated whether similar effects can be reproduced in neuronal cultures. RGC-5 cells were differentiated by 1 μM staurosporine for 6 h and then grown in media with 10% serum for 24 h. After 24 h, cells were observed for their morphological features. Their somas became rounder and neurites gradually increased, similar to prior studies (Frassetto et al., 2006). The cells contacted each other with multiple long neurites, features of neuronal differentiation (Figure 3A). Differentiated cells were then pretreated with SIRT1 activating compounds, 0.25 μM RSV or 5 μM SRTAW04, for 30 min before addition of 500 μM H₂O₂. Treatment with both SRTAW04 and RSV showed significant reduction of RGC-5 cell death compared to untreated cells, 24 h after H₂O₂ addition; treatment with a known SIRT1 inhibitor (Baur, 2010), nicotinamide (NAM; 200 μM), attenuated this protective effect whereas NAM treatment alone did not alter cell viability (Figure 3C). We also evaluated effects of higher RSV and SRTAW04 concentrations on cell survival, but found no increased benefit of RSV at concentrations higher than 0.25 μM, nor SRTAW04 higher than 5 μM (data not shown).

To investigate whether RSV and SRTAW04 attenuate superoxide accumulation produced by hydrogen peroxide treatment, RGC-5 cells were pretreated with 0.25 μM RSV or 5 μM SRTAW04, as used for the cell viability experiments. Cells were stained with MitoSOX Red 24 h after exposure to 500 μM H₂O₂, and experiments were repeated using both differentiated and non-differentiated RGC-5 cells. MitoSOX staining in differentiated cells showed that H₂O₂ treatment induces an increase in MitoSOX staining which is attenuated by both RSV and SRTAW04, and non-differentiated RGC-5 cells showed the same effect (Figure 4A). Quantification of the intensity of MitoSOX staining in differentiated RGC-5 cells shows that this effect is significant (Figure 4B). Untreated differentiated cells showed a small increase in MitoSOX Red staining compared to untreated non-differentiated cells, suggesting some effect from staurosporine treatment itself.

To determine whether H₂O₂ treatment results in loss of mitochondrial membrane potential, and how this is affected by SIRT1 activator treatment, JC-1 staining was used. Staurosporine-differentiated RGC-5 cells were treated for 24 h with 500 μM H₂O₂ with or without 0.25 μM RSV or 5 μM SRTAW04, and live cell fluorescent microscopy pictures were taken after JC-1 staining. Results reveal minimal remaining red fluorescent labeling of viable, high membrane potential mitochondria in H₂O₂-treated cells as compared with healthy control cells (Figure 5). Furthermore, the ratio of green to red fluorescent JC-1 staining showed a significant increase in H₂O₂-treated cells compared with control cells. The green to red ratio was significantly lower in H₂O₂-treated cells that were also treated with RSV or SRTAW04, as compared to H₂O₂-treated cells without SIRT1 activators. Results indicate that H₂O₂ treatment disrupts the mitochondrial membrane potential, and this is prevented by treatment with RSV and SRTAW04. While JC-1 stain can be used to mark mitochondrial membrane potential, cell membrane potential may also be detected. Therefore, additional mitochondrial membrane potential detection was performed by staining with TMRM. Results further suggest that H₂O₂ treatment disrupts the mitochondrial membrane potential, and treatment with RSV and SRTAW04 prevents this (Figure 5).

To confirm whether the ability of RSV and SRTAW04 to prevent loss of RGC-5 cells is dependent on SIRT1 activity, RGC-5 cells were transfected with SIRT1 siRNA or a control siRNA 24 h prior to addition of 500 μM H₂O₂ with or without RSV or SRTAW04. RGC-5 cell numbers were quantified 24 h later. SIRT1 siRNA blocked the ability of RSV and SRTAW04 to prevent RGC-5 cell loss (Figure 6A), whereas control siRNA did not (Figure 6B). SIRT1 protein expression was also measured by Western blotting of extracts from cultures treated with RSV. Results show H₂O₂ treatment leads to decreased SIRT1 expression, while RSV treatment significantly limits this decrease (Figure 6C).

RGC-5 cells differentiated with staurosporine were stressed with 500 μM H₂O₂ for 24 h, with or without addition of 0.25 μM RSV, then protein extracts were generated for Western blotting. SDH functions not only in mitochondrial energy generation, but also has a role in oxygen sensing (Baysal, 2006). Protein levels of SDH show a significant decrease during H₂O₂ treatment which is attenuated by RSV (Figure 7A). We also measured the protein expression levels of SOD2, a mitochondrial protein which binds to superoxide byproducts of oxidative phosphorylation and converts them to hydrogen peroxide and diatomic oxygen (Suski et al., 2012). Results show a significant decrease in levels of SOD2 during H₂O₂ treatment, and treatment with RSV significantly limits this decrease (Figure 7B).

The peroxisome proliferator activated receptor (PPAR) co-activator 1-α (PGC-1α) is a transcriptional co-activator identified as an upstream regulator of mitochondrial number and function (Russell et al., 2004) and is activated by SIRT1-mediated deacetylation (Lagouge et al., 2006). RGC-5 cell protein extracts were immunoprecipitated with PGC-1α antibody. Western blot of the precipitated protein was probed with either PGC-1α antibody, or antibody to acetylated lysine, and the ratio of acetylated to deacetylated PGC-1α was compared. Results show a significant increase in acetylated lysine antibody immunoreactivity following H₂O₂ treatment, compared to control cells, as well as compared to RSV and SRTAW04 treated RGC-5 cells (Figure 7C). Overall PGC-1α expression showed no significant change between culture conditions. When compared, the decreased ratio of acetylated lysine to PGC-1α immunoreactivity in RSV and SRTAW04 treated cells demonstrates SIRT1 activators prevent an increase in PGC-1α acetylation.

While RGC-5 cells differentiated with staurosporine exhibit some features of neurons, they are not a RGC specific cell line. Because optic neuritis and other optic neuropathies lead specifically to RGC loss, effects of RSV were further examined in primary retinal cell cultures, and RGC survival was assessed. Mixed dissociated retinal cells isolated from neonatal days 3 to 5 mice were incubated overnight, transfected for 24 h with SIRT1 specific or control siRNA, then treated for 24 h with 500 μM H₂O₂ with or without addition of RSV (3 μM). RGCs were identified within the mixed retinal cell cultures by immunohistochemical staining with antibodies to Brn3a (Figure 8A), and the number of RGCs was quantified by counting Brn3a positive cells by a blinded investigator. As seen in RGC-5 cells, H₂O₂ induced significant loss of primary RGCs, and RGC loss was reduced by treatment with RSV (Figure 8B). To determine whether the mechanism of this effect involves SIRT1 activity, primary RGC cultures were pretreated with SIRT1 siRNA. RSV neuroprotection of RGCs was blocked by SIRT1 siRNA but not by control siRNA (Figure 8B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.