Abstract
Mitochondrial dysfunction is an early event in the pathogenesis of neurologic disorders and aging. Sirtuin 3 (SIRT3) regulates mitochondrial function in response to the cellular environment through the reversible deacetylation of proteins involved in metabolism and reactive oxygen species detoxification. As the primary mitochondrial deacetylase, germline, or peripheral tissue-specific deletion of SIRT3 produces mitochondrial hyperacetylation and the accelerated development of age-related diseases. Given the unique metabolic demands of neurons, the role of SIRT3 in the brain is only beginning to emerge. Using mass spectrometry-based acetylomics, high-resolution respirometry, video-EEG, and cognition testing, we report targeted deletion of SIRT3 from select neurons in the cortex and hippocampus produces altered neuronal excitability and metabolic dysfunction in female mice. Targeted deletion of SIRT3 from neuronal helix-loop-helix 1 (NEX)-expressing neurons resulted in mitochondrial hyperacetylation, female-specific superoxide dismutase-2 (SOD2) modification, increased steady-state superoxide levels, metabolic reprogramming, altered neuronal excitability, and working spatial memory deficits. Inducible neuronal deletion of SIRT3 likewise produced female-specific deficits in spatial working memory. Together, the data demonstrate that deletion of SIRT3 from forebrain neurons selectively predisposes female mice to deficits in mitochondrial and cognitive function.
SIGNIFICANCE STATEMENT Mitochondrial SIRT3 is an enzyme shown to regulate energy metabolism and antioxidant function, by direct deacetylation of proteins. In this study, we show that neuronal SIRT3 deficiency renders female mice selectively vulnerable to impairment in redox and metabolic function, spatial memory, and neuronal excitability. The observed sex-specific effects on cognition and neuronal excitability in female SIRT3-deficient mice suggest that mitochondrial dysfunction may be one factor underlying comorbid neuronal diseases, such as Alzheimer's disease and epilepsy. Furthermore, the data suggest that SIRT3 dysfunction may predispose females to age-related metabolic and cognitive impairment.
Introduction
Aging is an inevitable process that is modulated by various internal and external factors, including biological sex (Austad, 2011). Sirtuin 3 (SIRT3) is gaining attention as one such modulator of cellular aging and has been shown to regulate lifespan in model organisms, such as yeast and worm (Kaeberlein et al., 1999). In humans, SIRT3 is associated with increased longevity in men and increased hormone receptor-positive breast cancer in women, suggesting a role for SIRT3 in human aging that may differ according to biological sex (Rose et al., 2003; Bellizzi et al., 2005, 2007). Curiously, despite its importance in regulating global mitochondrial function, germline deletion of sirt3 produces no overt phenotype, although its deficiency is associated with the accelerated development of age-related diseases, such as cancer, cardiovascular disease, and metabolic diseases (Hirschey et al., 2011). Additionally, SIRT3 levels are altered in postmortem tissue obtained from patients with Alzheimer's disease (AD) and Parkinson's disease and decrease as a function of age, suggesting a potential role for SIRT3 in neurologic disease (Flurkey et al., 2007; Joseph et al., 2012; Weir et al., 2012; Shi et al., 2017; Yin et al., 2018; Pukhalskaia et al., 2020).
Reversible mitochondrial protein acetylation is a dynamic post-translational modification (PTM) that regulates mitochondrial function at multiple levels. Acetylation is a largely inhibitory PTM that is particularly enriched in mitochondria (Kim et al., 2006; Verdin and Ott, 2015). The high concentration of acetyl-CoA and basic pH within mitochondria favors an environment for nonenzymatic acetylation of susceptible proteins (Garland et al., 1965; Wagner et al., 2017). SIRT3 is a member of the highly conserved sirtuin superfamily (SIRT1-7) and one of three sirtuins localized to the mitochondria where it acts as the primary deacetylase (Onyango et al., 2002; Lombard et al., 2007). Its deacetylase activity is dependent on NAD+, linking SIRT3 function to the metabolic status of the cell (Finnin et al., 2001). With a broad range of target proteins involved in antioxidant response, oxidative phosphorylation (OXPHOS), the tricarboxylic acid (TCA) cycle, and fatty acid oxidation, SIRT3 is therefore postulated to be a critical, redox-sensitive interface between cellular environment and metabolism (Kumar and Lombard, 2015; Yang et al., 2016). In postmitotic cells, such as neurons, SIRT3 may be particularly important as mitochondrial acetylation can accumulate in an age-dependent manner potentially contributing to mitochondrial dysfunction.
Recent studies suggest that SIRT3 expression and function are tissue-dependent with brain exhibiting unique patterns of mitochondrial protein acetylation compared with liver mitochondria (Jin et al., 2009; Dittenhafer-Reed et al., 2015). This supports the role of SIRT3 as a key cellular rheostat linking metabolic processes to the energy demands of the cell. Whereas SIRT3 function in the liver has been intensively studied, its importance in the brain is only beginning to emerge. Moreover, its function in neurons which comprise a fraction of all cell types in brain but consume the majority of ATP is not well characterized (Herculano-Houzel, 2014). In the current study, we sought to determine the specific role of neuronal SIRT3 by selectively deleting SIRT3 in principal neurons within the hippocampus and neocortex. Deletion of SIRT3 in these highly metabolically active cells resulted in hyperacetylation of SOD2, increased cortical superoxide (O2−.) levels, metabolic reprogramming, altered neuronal excitability, and spatial memory deficits in 12-month-old female, but not 12-month-old male, mice. These findings reveal a novel and unique vulnerability of female mice to the loss of neuronal SIRT3 and provide a potential mechanism by which sex differences can emerge in the context of aging and neurodegenerative disease.
Materials and Methods
Reagents
All reagents were obtained from Sigma Aldrich or Fisher Scientific unless otherwise noted.
Mice
Animal studies were conducted in accordance with the National Institute of Health's Guide for the care and use of laboratory animals (Publication no. 80-23). All procedures were approved by the Institute Animal Care and Use Committee of the University of Colorado Anschutz Medical Campus, which is fully accredited by the American Association for the Accreditation of Laboratory Animal Care. Sirt3 p lox C57BL/6CrSlc mice were obtained through a material transfer agreement from the University of California–San Francisco (E.V. laboratory). These mice were crossed with NEX-Cre C57BL/6 (Goebbels et al., 2006) or the inducible Camk2cre/ERT 2 mice from The Jackson Laboratory (Madisen et al., 2010) to generate the heterozygous F1 generation. The offspring were intercrossed, and genotyping was performed on the F2 progeny from ear snips collected at the time of weaning. To identify the p lox copy number, PCR was performed (ABI TaqMan probe Mm01275637_g1). Individuals carrying the NEX-Cre insert were identified as described by Goebbels et al. (2006). All mice, including breeders, were housed in ventilated caging on a 14/10 light/dark cycle with ad libitum access to standard rodent chow and filtered water. Mice (6 or 12 months old, as specified) were killed by continuous inhalation of CO2 followed by decapitation.
Tamoxifen injection
Tamoxifen (20 mg/ml) was prepared fresh, in the dark, by first dissolving in a small amount of 95% EtOH followed by corn oil. Ten- to 12-month-old male and female Camk2cre/ERT2 mice (cnSIRT3+/+, cnSIRT3−/−) were injected intraperitoneally with 100 mg/kg tamoxifen or vehicle daily for 5 d. Two weeks following the last injection, we verified SIRT3 knockdown in a small subset of mice by RT-PCR and identified a significant reduction in SIRT3 expression in tamoxifen-treated cnSIRT3−/− mice compared with vehicle-treated cnSIRT3−/− and tamoxifen-treated cnSIRT3+/+ (mean fold difference ± SEM: 0.55 ± 0.12, p = 0.0008, n = 6-8/group unpaired t test and 0.37 ± 0.07, p = 0.0004, n = 6/group unpaired t test, respectively). Multiple separate cohorts of mice were treated with tamoxifen or vehicle to generate sufficient numbers for behavioral testing (described below). During the tamoxifen treatment period and the following 2 weeks, all mice were monitored for health effects by daily visual inspection and weight checks. Any mouse found to have lost >10% of its body weight was immediately killed. At the conclusion of behavior testing (described below), ∼4 weeks after the initiation of tamoxifen treatment, mice were killed, and cortices collected for verification of sirt3 knockdown by RT-PCR. Mice were excluded from behavior analysis if the sirt3 RQ value was >0.7. This resulted in the following group numbers across three cohorts: male vehicle-treated cnSIRT3+/+ mice = 10, male vehicle-treated cnSIRT3−/− mice = 11, male tamoxifen-treated cnSIRT3+/+ mice = 9, male tamoxifen-treated cnSIRT3−/− mice = 12, female vehicle-treated cnSIRT3+/+ mice = 12, female vehicle-treated cnSIRT3−/− mice = 12, female tamoxifen-treated cnSIRT3+/+ mice = 10; female tamoxifen-treated cnSIRT3−/− mice = 6. Of these eight groups, only 2 (female tamoxifen-treated cnSIRT3−/− mice and male tamoxifen-treated cnSIRT3−/− mice) showed significantly decreased sirt3 expression, as expected. Thus, the remaining six groups are controls.
Estrus cycle monitoring
Estrus cycle stage was monitored daily and/or at the time of tissue collection by vaginal lavage as previously described (Au-McLean et al., 2012). Briefly, vaginal cells were collected in ddH2O, slide prepped, and visualized using cresyl violet stain. Cycle stage was determined as estrus when cells were comprised of mostly cornified squamous epithelial cells, proestrus when nucleated epithelial cells were predominant, and diestrus was determined when mainly leukocytes were present. For the mitochondrial protein acetylation experiment, estrus cycle in 6-month-old mice was balanced across groups or specifically stratified for diestrus. Indeed, by stratifying female samples for diestrus or using sample sizes large enough to encompass all estrus phases, it is highly unlikely that the observed sex differences could be directly attributed to differences in circulating hormones. Moreover, SIRT3 was selectively deleted in a subset of neurons likely precluding any potential alterations to total cortical estrogen/hormone signaling.
Primary cortical cultures
Mixed cortical cell cultures were established from male and female nSIRT3−/− and nSIRT3+/+ postnatal day 1-3 mice. Tail snips from tattooed littermates were collected and genotyped within 1 d of birth. Pups were killed and cortices dissected and placed in sterile HBSS + 10 mm HEPES. Tissue was minced finely with micro scissors, then trypsinized in a 1% solution of trypsin/HBSS for 25 min at 37°. Cells were rinsed with HBSS and triturated in a 1% solution of MEM and DNase (Sigma) until homogeneous. Cells were pelleted by centrifugation at 800 rpm for 10 min, followed by resuspension in NeuroCult plating medium (STEMCELL Technologies). Cells were then plated in PDL-treated 96 well XF96 cell culture microplates (cell density: 60,000/well; Agilent Technologies) or 48 well multielectrode array (MEA) cell culture plates (cell density: 50,000/well; Axion BioSystems) and maintained at 37°C, 5% CO2/95% room air. After 5 DIV, one-half of the plating medium was replaced with BrainPhys neuronal medium (STEMCELL Technologies). Downstream applications are described below.
Immunocytochemistry
To validate the maturity of synapses, we performed immunocytochemistry on DIV 9 mixed cortical cells. Cells were fixed in 4% PFA and subsequently blocked for 1 h at room temperature (0.5% BSA, 2% NGS, 0.005% Tween-20, 0.05% Triton X-100). Primary antibody was applied overnight at 4°C (guinea pig anti-NeuN, Millipore N90, 1:100; rabbit anti-Synapsin I, Abcam ab254349, 1:100). Following rinses with PBS (3 × 5 min) and secondary antibody incubation at room temperature for 1 h (goat anti-guinea pig AlexaFlour 488, Invitrogen A11073, 1:100; goat anti-rabbit ATTO 647, Rockland 611-156-122, 1:100), cells were rinsed with PBS (3 × 5 min) and mounted with Prolong Diamond Antifade Mountant (Fisher Scientific). Images were captured on an Olympus FV-1000 confocal microscope with a 60× objective (Shinjuku).
MEA recordings
MEA recordings (Axion BioSystems) were performed as previously described (McConnell et al., 2012). Briefly, spontaneous network activity was measured in DIV 9 cortical cultures grown in MEA plates equipped with 16 embedded nanoporous platinum electrodes (I30 μm diameter, 200 μm center-to-center spacing; Axion BioSystems) at a density of 50,000 cells/well. Data acquisition for all 768 channels was performed by Axion's Integrated Studio (AxIS 2.1.1) software, at a sampling rate of 12.5 kHz/channel and 1200 × gain. Culture viability was evaluated by monitoring spontaneous network activity at DIV 5-8. Synapse maturity at DIV 9 is illustrated in Extended Data Figure 5-1. Wells showing no spontaneous activity were excluded from analysis. Spontaneous network activity was measured as total number of spikes (variable threshold spike threshold detector set at 8× SDs of the rms noise of each channel, bandpass filter: high pass 300 Hz, low pass 5000 Hz). Studies were performed at 37°C.
Bioenergetic profiling in male and female, nSIRT3−/− and nSIRT3+/+ cultures
Mitochondrial respiration and extracellular glycolytic parameters were investigated in primary mixed neuronal glial cultures at DIV 8 using a Seahorse XFe96 Analyzer (Agilent BioSciences) following standard protocols (Pelletier et al., 2014). On the day of the experiment, cell culture media was exchanged for glucose and serum-free XF assay media and plates (60,000 cells/well) were transferred to a 37°C, CO2-free incubator for 1 h before measurement. Real-time extracellular acidification rate was measured 3 times under baseline conditions and following the acute injection of glucose (16 mm), oligomycin (1 μm), and 2-deoxyglucose (50 mm). Basal oxygen consumption rates were obtained before glucose injection. Values were normalized to protein amount as determined by Bradford Assay (Fisher Scientific).
Mitochondrial isolation
Crude mitochondria were obtained from the forebrain (cortex and hippocampus) of adult mice as previously described (Liang and Patel, 2006). Briefly, ice-cold brain homogenates were centrifuged for 10 min at 3000 × g (4°C), and the resulting supernatant was centrifuged for another 10 min at 13,000 × g (4°C). Mitochondrial pellets were stored at −80°C for future use.
Immunoblotting
To assess protein expression, cortical tissue or mitochondria from 6- and 12-month-old-mice were lysed in SDS buffer and protein yield was determined by Bradford Assay (Fisher Scientific). Samples were diluted in PBS, combined with Laemmli sample buffer (2.5% β mercaptoethanol [1:1], Bio-Rad), heated to 95°C for 10 min, and placed on ice. Criterion stain-free gels (4%-20%) were then loaded with 25 µg protein and electrophoresed for 1 h at 200 V in 1× Tris/glycine/SDS buffer (Bio-Rad). Protein was transferred to PVDF membranes on a Trans-blot turbo transfer system following the manufacturer's instructions (Bio-Rad) and then blocked for 1 h in 5% milk and TBST (0.1% Tween). Membrane was incubated with primary antibody overnight at 4°C (SOD2 ab13534, Abcam, 1:200; SOD2 K68 ab137037, Abcam, 1:200; SOD2 K122 ab214675, Abcam, 1:200; acetyl-lysine 9814, Cell Signaling, 1:1000; acetyl-lysine ICP0380, Immunechem Pharmaceuticals, 1:200; SIRT3 5490, Cell Signaling, 1:200) followed by secondary antibody (HRP-conjugated goat α rabbit, Invitrogen 31460) at 1:10,000 concentration for 1 h. Selection of primary antibodies was based on peer-reviewed reports that illustrate validation using in vitro acetylation treatment (Assiri et al., 2017; Gano et al., 2018). The membrane was developed with SuperSignal West Pico PLUS Chemiluminescent substrate (Fisher Scientific) and imaged for both total protein and HRP signal on a Bio-Rad ChemiDoc MP imaging system. The HRP signal was quantified using ImageLab 6.0.1 software (Bio-Rad) and normalized to the total protein signal.
Mass spectrometry-based acetylomics
Cortical mitochondria were isolated from 12-month-old male and female nSIRT3−/− and nSIRT3+/+ mice (6/group = 24 total mice) as described above for acetylomic analysis. To have enough protein for analysis, mitochondria were pooled from 2 mice of the same sex and genotype to form one sample. Analysis was then performed on three samples/group. Sample preparation, data collection, accurate mass and retention time (AMRT) library generation, and data analysis were performed as previously described (Ali et al., 2019). A total of 3.2 mg from each pooled sample was spiked with 75 ng of acetylated whole protein BSA and trypsin-digested overnight, acidified with TFA, purified using Sep-Pak C18 Classic Cartridges (Waters, #WAT051910), frozen at −80°C for 4 h, and lyophilized for 48 h. Samples were incubated 2 h at 4°C with immunoaffinity beads conjugated to acetyl lysine antibody (PTMScan Acetyl-Lysine Motif [Ac-K] Immunoaffinity Beads (#13416, Cell Signaling). After incubation, the supernatants were collected and the remaining beads were washed with IAP buffer twice (PTM Scan IAP Buffer 9993, Cell Signaling) and with Burdick and Jackson LC-MS grade water 3 times (Honeywell). Peptides were eluted with 0.15% TFA twice, pooled, cleaned by Pierce C18 Spin Tips (Fisher Scientific, #84850), evaporated to dryness, and frozen at −80°C. Dried samples were resuspended in 15 fmol/μl of PROCAL Retention Time Standard Kit peptides (JPT Peptide Technologies, #RTK-1 10 pmol) in 3% ACN, 0.1% formic acid in water for LC-MS and LC-MS/MS analysis.
Enriched acetyl lysine peptides were loaded to a 2 cm PepMap 100, nanoviper trapping column and resolved online using a 0.075 × 250 mm, 2.0 μm Acclaim PepMap RSLC reverse-phase nano column (Fisher Scientific) by 1290 Infinity II LC system with a nanoadapter (Agilent Technologies). Mobile phases consisted of water + 0.1% formic acid (A) and 90% aqueous acetonitrile + 0.1% formic acid (B). Samples were loaded onto the trapping column at 3.2 μl/min for 3.5 min before separation via flow rate of 330 nl/min and a gradient of 3%-7% B over 2 min, 7%-35% B over 43 min, and 35%-70% B over 2 min for a 47 min gradient at 42°C. The column wash was performed at 70% B for 5 min. Data were acquired using a 6550 Q-TOF with a nano source (Agilent Technologies) operated in MS-Only mode for peptide quantification for every sample or pooled samples of nSIRT3−/− and nSIRT3+/+ female and male groups. Analysis was performed using intensity-dependent CID MS/MS for peptide identifications. Capillary voltage, drying gas flow, and drying gas temperature were 1300 V, 11.0 L/min, and 175°C, respectively. MS/MS data were collected using positive ion polarity at mass ranges 260-1700 m/z and a scan rate of 6 spectra/s for MS scans and mass ranges 50-1700 m/z at a scan rate of 3 spectra/s for MS/MS scans. All charge states, excluding singly charged species, were included for MS/MS acquisition, and preference was given for charge States 2 and 3. MS-only analysis was performed at a fixed scan rate of 1.5 spectra/s.
We used SpectrumMill software (Agilent Technologies) to extract, search, and review peptide identities. The SwissProt Mus Musculus database allowing up to four missed tryptic cleavages and fixed carbamidomethyl (C) and variable deamidated (NQ), oxidation (M), and acetyl (K) modifications. Monoisotopic mass tolerance allowed was ±20.0 ppm and the MS/MS tolerance was ±50 ppm. Acetylated peptides having a minimum peptide score of 8 and scored peak intensity of 50% were included in building the AMRT library.
LC-MS quantitative data were extracted and aligned with Profinder V.B.10.00 software (Agilent Technologies). Retention times, neutral masses, and chemical formulas identified in the AMRT library were used for batched feature extractions; for extraction, an ion count threshold of 2 or more ions and 8000 counts along with score threshold of 50. Scoring includes MS quality, isotope abundance, and isotope spacing that match to a targeted chemical formula in a given retention time window in the AMRT library. Charge States 2-6 were included with H+ adducts. Mass window alignment and retention time tolerances were set to 10 ppm and 0.8 m, respectively. The extraction and alignment data were quantified using Mass Profiler Professional version 14.9 (Agilent Technologies). Acetylated BSA peptides were used to assess the variance in sample preparation between samples. Overall, internal standards used in our approach accounted for acetyl peptide enrichment, sample injection, and chromatographic retention. Based on rigorous PROCAL and acetyl-BSA standardization, acetyl peptide identifications acquired from 0-33 min were included from both sexes for bioinformatics analysis.
Statistical analysis included peptides found in 100% of 1 of 4 conditions for group-to-group comparisons. Volcano plots of acetyl peptides were generated using a moderated t test for peptides with a fold change ≤2.0 or ≥2.0 and a Benjamini-Hochberg multiple-testing correction p value < 0.05. A two-way ANOVA between genotypes and sex was performed to determine the overall effects of genetic differences versus sex differences at various lysine acetylation sites.
Superoxide measurement
The O2−. sensitive, 2-hydroxyethidium (2-OH-E+) was measured by an HPLC coupled with an electrochemical detector following the method previously described with minor modifications (Zielonka et al., 2006; Maghzal and Stocker, 2007). Twelve-month-old-mice were injected with 50 mg/kg (s.c.) dihydroethidium (DHE, Sigma Aldrich #37291) and killed 1 h after injection. Brain samples were sonicated in ice-cold 0.1N PCA (0.1 g/ml, w/v) and centrifuged at 16,000 × g at 4°C for 10 min. An aliquot of 20 µl was injected into the HPLC-EC system (ESA Biosciences), which consists of two pumps (model 582), a temperature controlled autosampler (model 542), and an 8-channel electrochemical CoulArray (model 5600) detector set to the potentials 0, 200, 280, 365, 400, 450, 500, and 600 mV versus the palladium reference electrode. The mobile phase contained 50 mm PB, pH 2.6, 2% acetonitrile. HE, 2-OH-E+, and E+ were separated on a Phenomenex Polar-RP 80 A column (4 μm, 250 × 4.6 mm; Phenomenex). The flow rate was 0.6 ml/min, and the pressure was ∼78 bar.
EEG analysis
Twelve-month-old-mice were anesthetized (isoflurane) and implanted with bilateral dural stainless-steel electrodes (Pinnacle Technology; AP: −0.5 mm from bregma, ML: ±2.0 mm from midline). Two additional electrodes positioned near the lambdoid suture on left and right hemispheres served as a reference and ground electrode. Electrodes were secured to the skull with dental acrylic. Mice received buprenorphine once daily for 2 d. After a 5 d recovery period, synchronized time-locked video and EEG signals (Pinnacle Technology) were recorded continuously for 14 d. The EEG signals were digitized at 2000 Hz with acquisition filters set at 0.5 Hz (low) and 500 Hz (high) and stored on a hard disk for offline analysis. EEG records were searched for seizure-related parameters, including interictal spikes (ISs) by an investigator blinded to animal genotype and sex. An IS was noted when a high amplitude (at least 3× baseline) sharp wave was observed. IS quantitation and power spectral plots of EEG signal were acquired using Sirenia Seizure Pro software (Pinnacle Technology).
High-resolution respirometry
To assess Complex I (CI) function, maximum mitochondrial respiratory capacity was measured pairwise in freshly isolated cortical mitochondria from 12-month-old sex- and genotype-matched mice. Samples were compared by high-resolution respirometry (Oxygraph-2k: O2k, Oroboros Instruments) in the presence of CI and Complex II (CII) substrates according to established protocols with minor modification (Votion et al., 2012). Briefly, mitochondria were isolated as described above in mitochondrial isolation buffer A (320 mm sucrose, 10 mm Tris-Cl, 1 mm K+ EDTA, 2.5 g/l BSA, pH 7.4). Mitochondria were resuspended in respiration buffer (MiR05) and loaded into a previously oxygenated (200 mm catalase and 200 mm H2O2), oxygraph chamber at 37°C. After the addition of mitochondria, the following compounds were added to the oxygraph: 10 mm glutamate, 2 mm malate, 5 mm pyruvate, 5 mm ADP, 10 μm cytochrome C, 10 mm succinate, 0.5 μm FCCP, 0.5 μm rotenone, and 2.5 μm antimycin A. After each injection of substrate, uncoupler, or inhibitor (SUIT-011D024), a 5-point averaging window was used to determine the oxygen consumption from the change in concentration of oxygen within the chamber.
Enzyme activity assays
SOD2 enzyme activity was measured by colorimetric detection of formazan dye following production of O2−. via xanthine oxidase in 12-month-old-mice (Sun et al., 1988). Mitochondria were isolated from cortical tissue by homogenizing in cold buffer containing 20 mm HEPES, pH 7.2, 1 mm EGTA, 210 mm mannitol, and 70 mm sucrose. Samples were centrifuged first at 1500 × g followed by 10,000 × g. Residual CuZn-SOD and EC-SOD activity was deactivated via incubation with 3 mm potassium cyanide for 30 min. SOD2 enzyme activity was then assessed with a Cayman Chemical superoxide dismutase assay kit, and 460 nm absorbance was measured on a BioTek Synergy four plate reader (Agilent Technologies). Protein normalization was performed via Coomassie Blue assay (Fisher Scientific).
NADH:ubiquinone oxidoreductase (CI) activity was measured in crude mitochondria isolated from 12-month-old-mice as previously described (Birch-Machin and Turnbull, 2001). Crude mitochondria were lysed by freeze-thawing in hypotonic buffer (25 mm KH2PO4, 5 mm MgCl2, pH 7.4). The reaction was initiated by the addition of 50 µg mitochondria to the assay buffer (hypotonic buffer with 65 μm ubiquinone, 130 μm NADH, 2 µg/ml antimycin A, and 2.5 mg/ml defatted BSA). The oxidation of NADH by CI was monitored spectrophotometrically at 340 nm for 2 min at 30°C. Activity was further monitored for an additional 2 min following the addition of rotenone (2 µg/ml). The difference between the rate of oxidation before and after the addition of rotenone was used to calculate CI activity.
Aconitase and fumarase activities were measured as described previously (Patel et al., 1996). Briefly, mitochondria were isolated from cortical tissue and homogenized in 70 mm sucrose, 210 mm mannitol, 10 mm Tris HCl, 1 mm EDTA, pH 7.4. Mitochondria were isolated from the homogenate following two differential centrifugation steps (3000 × g followed by 13,000 × g) and snap frozen in liquid nitrogen; 50 μl of the mitochondria was combined with either 100 μl 200 mm isocitrate (aconitase assay) or 50 mm malate (fumarase assay) and placed in a cuvette containing 850 μl reaction buffer (aconitase assay: 50 μl mm Tris HCl and 600 μm MnCl2; and fumarase assay: 30 mm KH2PO4 100 μm EDTA, pH 7.4). The rate of the conversion of cis-aconitate to isocitrate (aconitase assay) or fumarate to L-malate (fumarase assay) was measured spectrophotometrically by detecting the increased rate of absorbance at 240 nm. Protein normalization was performed via Coomassie Blue assay (Fisher Scientific).
Cognitive testing
Multiple cohorts of mice underwent Y-maze testing for spatial working memory at 6 or 12 months of age. Mice were acclimated to conditions approximating that of the Y-maze over several days. Specifically, mice were brought to the dedicated behavioral testing room for 2 h before returning to the housing room. The next day, mice performed an open field task as an initial assessment of locomotion and exploratory drive. The following day, test order was randomized using a random number generator (www.random.org) and mice were allowed to explore the Y-maze apparatus (arm length 30 cm) for 8 min. A successful alternation was noted on three consecutive arm entries (i.e., A-B-C), while an unsuccessful alternation was recorded when the mouse returned to the most recently explored arm (i.e., A-B-A). Total arm entries and sequence of arm entries were recorded, and percent alternation was calculated as follows: [(the number of actual alternations ÷ the maximum possible alternations) × 100].
Statistical analysis
The distribution of data before analyses was evaluated with the Shapiro–Wilk test for normal distribution. Data meeting this criterion were compared with Student's t test and evaluated for effects of genotype and sex using two-way ANOVA. Significant main effects and interactions (p < 0.05) were probed using Sidak's multiple comparison test. All data are presented as mean ± SEM unless otherwise noted. Analyses were performed using Prism 9.3.1 software (GraphPad Software).
Results
Targeted deletion of neuronal-SIRT3 results in hyperacetylation of mitochondrial proteins
To investigate the role of SIRT3 in a clinically relevant subset of neurons, sirt3 floxed mice were crossed with the constitutive Cre line under the control of NEX, a basic helix-loop-helix protein (Lombard et al., 2007) (Fig. 1A). This Cre-lox approach has been previously shown to restrict target gene deletion to principal excitatory neurons of the forebrain in primarily cortical and hippocampal areas (Goebbels et al., 2006). Offspring were born in a normal Mendelian ratio and displayed no obvious physical phenotype. We verified the presence of the sirt3 gene in liver in both nSIRT3+/+ and nSIRT3−/− mice and an ∼50% reduction of brain SIRT3 expression in nSIRT3−/− mice (Fig. 1B), consistent with neuronal-specific KO of SIRT3. To ensure that deletion of SIRT3 from just a subset of neurons was sufficient to increase cortical mitochondrial acetylation, a direct consequence of its enzymatic function, we performed mass spectrometry-based acetylomics in 12-month-old male and female mice (schematic represented in Fig. 1C). Using this approach in isolated cortical mitochondria, we identified 205 significantly hyperacetylated peptides corresponding to 88 different proteins in nSIRT3−/− mice across gender (Fig. 1D; Extended Data Table 1-1). Further bioinformatics analysis identified pathologically relevant pathways enriched in biological processes (Fig. 1E), molecular functions (Fig. 1F), and KEGG pathways (Fig. 1G), all associated with the 88 proteins found to have significant changes in acetylation by at least a 1.2-fold increase. Most of the identified proteins were involved in metabolic processes, including the TCA cycle and oxidative processes (Fig. 1E–G; Table 1; Extended Data Table 1-1).
Identity of hyperacetylated proteins involved in TCA cycle and mitochondrial electron transport in nSIRT3−/− mice (n = 6 nSIRT3−/− mice, n = 6 nSIRT3+/+ mice)
Targeted deletion of neuronal SIRT3 results in hyperacetylation of mitochondrial proteins. A, Schematic represents Cre-loxP strategy for targeting Nex to induce forebrain-specific neuronal SIRT3 deletion. B, Sirt3 gene expression in liver and cortex (n = 5 or 6/group). **p < 0.01 (unpaired t test). C, Mass spectrometry-based acetylomic workflow. D, Volcano plot of significantly hyperacetylated (red) and unaltered (black) proteins in nSIRT3−/− cortical mitochondria from 12-month-old-mice (n = 6/group; Benjamini-Hochberg–corrected p < 0.05). Bioinformatic analysis of significantly increased acetyl peptides using DAVID (version 6.8) identified enriched pathways related to (E) biological processes, (F) molecular functions, and (G) KEGG pathways (fold change > 1.2). Proteins identified to have significant changes in acetylation levels are shown in Extended Data Table 1-1.
Table 1-1
Proteins identified to have significant changes in acetylation levels. These proteins exhibit at least a 1.2-fold increase in the level of acetylation and are mostly involved in metabolic processes. Download Table 1-1, DOCX file.
Neuronal-SIRT3 deficiency predisposes to mitochondrial hyperacetylation but not enzymatic dysfunction
We sought to characterize any potential sex differences in these mice by directly comparing acetylation levels with SDS page and immunoblotting. We first verified that mitochondrial SIRT3 protein levels were equivalent between the sexes and its deficiency in 12-month-old nSIRT3−/− mice (Fig. 2A). Using acetyl-lysine antibodies, we found global mitochondrial acetylation to be significantly increased in both male and female nSIRT3−/− mice at 6 months (Fig. 2B) and 12 months of age (Fig. 2C). We rationalized that if circulating hormones such as estrogen or progesterone were sufficient to increase brain mitochondrial acetylation, this would be reflected in the hormonal fluctuations associated with estrus cycle. To determine whether hormonal fluctuations associated with estrus cycle would increase brain mitochondrial acetylation, we assessed mitochondrial pan-acetylation in female mice at 6 months of age staged in proestrus, estrus, or diestrus. Regardless of estrus stage, nSIRT3−/− mice exhibited significantly increased acetylation relative to nSIRT3+/+ mice (Fig. 2D).
Neuronal SIRT3 deficiency predisposes to mitochondrial hyperacetylation but not enzymatic dysfunction. Western blot and band densitometry analysis of (A) SIRT3 in cortical mitochondrial isolates (n = 6/group), total mitochondrial protein acetylation levels in cortical mitochondrial isolates in (B) 6-month-old and (C) 12-month-old-mice (n = 6/group), and of total mitochondrial protein acetylation levels across estrus cycle in 6-month-old female mice and male mice (n = 3/group) (D). Protein expression and protein acetylation levels were normalized to total protein load. Enzymatic activity in 12-month-old mice (E) CI activity (n = 4-6/group), (F) aconitase activity (n = 11 or 12/group), and (G) fumarase activity (n = 11 or 12/group). Total mitochondrial acetylation levels and enzymatic activity were compared by two-way ANOVA followed by Sidak's multiple comparison test. Post hoc significance: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Error bars indicate mean ± SEM.
As a major target of SIRT3-mediated deacetylation, multiple subunits of CI were hyperacetylated in nSIRT3−/− mice (Table 1). Among these were core subunits comprising the minimal assembly required for catalysis, including the 75 kDa subunit (Ndufs1) which was acetylated at two lysine sites, K68 and K122 (Table 1). We sought to identify the enzymatic effect of these findings by examining CI activity in 12-month-old mice. Interestingly, CI activity was unaffected in nSIRT3−/− mice relative to nSIRT3+/+ (Fig. 2E). Moreover, hyperacetylation at multiple sites of mitochondrial aconitase (Table 1; Fig. 2F) and fumarase (Fig. 2G) was not sufficient to alter the activities of these TCA cycle enzymes. Together, this suggests that deletion of SIRT3 is sufficient to induce cortical mitochondrial protein hyperacetylation; however, this hyperacetylation does not appear to influence downstream enzymatic function of specific target proteins.
Neuronal-SIRT3 deficiency results in sex-specific alterations to SOD2
SOD2, a critical target of SIRT3-mediated deacetylation, is the primary enzyme capable of detoxifying mitochondrial O2−., and inhibition of its activity by hyperacetylation leads to mitochondrial oxidative stress (Qiu et al., 2010; Tao et al., 2010; Chen et al., 2011; Dikalova et al., 2017). Given that global mitochondrial protein acetylation was significantly increased in nSIRT3−/− mice relative to nSIRT3+/+ mice, we sought to determine whether SOD2 key functional lysine residues, such as K68 and K122, exhibited a similar acetylation pattern. Despite no changes in the protein levels of SOD2 (Fig. 3A), acetylation at K68 was significantly increased in both male and female nSIRT3−/− mice relative to nSIRT3+/+ at 12 months of age (Fig. 3B). Interestingly, acetylation at K122 was only significantly elevated in female nSIRT3−/− mice (Fig. 3C) at 12 months of age.
Neuronal SIRT3 deficiency results in sex-specific alterations to SOD2. Western blot and band densitometry analysis of (A) SOD2 in 12-month-old-mice (n = 6/group), (B) acetylated K68 of SOD2 in 12-month-old-mice (n = 6/group), and (C) acetylated K122 of SOD2 in 12-month-old-mice (n = 6/group). Protein expression and protein acetylation levels were normalized to total protein load. D, SOD2 activity in 12-month-old-mice (n = 7-9/group). E, HPLC-based detection of 2-OH-E+, which is generated specifically by the superoxide-derived oxidation of DHE in 12-month-old-mice (n = 8 or 9/group). Data analyzed by two-way ANOVA followed by Sidak's multiple comparison test. Total mitochondrial acetylation levels and enzymatic activity were compared by two-way ANOVA followed by Sidak's multiple comparison test. Post hoc significance: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Error bars indicate mean ± SEM.
To determine whether differences in SOD2 K122 hyperacetylation conferred alterations to its functional capacity, we examined SOD2 activity in crude mitochondria from 12-month-old-mice. Compared with their WT controls, female nSIRT3−/− mice exhibited significantly decreased SOD2 activity (Fig. 3D). To examine whether decreased SOD2 activity was sufficient to increase steady-state O2−. levels, we assessed O2−. levels in cortical tissue using HPLC-based detection of O2−. sensitive 2-OH-E+. 2-OH-E+ is formed from the specific oxidation of DHE by O2−. (Zhao et al., 2003, 2005; Fink et al., 2004). After peripheral injection of DHE, we detected significantly increased levels of 2-OH-E+ in female nSIRT3−/− mice compared with female nSIRT+/+ (Fig. 3E). Together, these data suggest that neuronal-specific SIRT3 deletion results in sex-specific hyperacetylation to SOD2 key functional lysine residue K122 and result in female-specific elevated cortical O2−. levels and decreased SOD2 activity.
Neuron-specific SIRT3 deletion results in sex-specific effects on cellular metabolism
Through deacetylation of enzymes associated with various metabolic pathways, SIRT3 is capable of affecting global shifts in cellular metabolism (Finley and Haigis, 2012). We therefore sought to investigate cellular bioenergetics using a combination of approaches. First, we determined the consequences of neuron-specific deletion of SIRT3 in primary cultured cells grown in culture (DIV 9). Real-time bio-energetic profiling of extracellular acidification rate and oxygen consumption rate allowed for measurement of glycolysis and mitochondrial respiration, respectively. Extracellular acidification rate was significantly elevated by genotype and sex such that cultures obtained from female nSIRT3−/− mice exhibited increased baseline glycolysis (Fig. 4A,B) and glycolytic capacity (Fig. 4A,C) compared with all other groups. Baseline oxygen consumption rate was not different between genotypes (mean % of KO male ± SEM; 89.23 ± 7.647 for WT male, 112.3 ± 6.415 for WT female, 108.9 ± 6.415; n = 4-9/group).
Neuron-specific SIRT3 deletion results in sex-specific effects on cellular metabolism. A, Representative image of in vitro extracellular acidification rate. B, Quantification of baseline glycolysis and (C) glycolytic reserve. n = 9-15/group (biological replicates), 10 technical replicates, across 9 total experiments. Data compared by two-way ANOVA followed by Sidak's multiple comparison test. Mitochondrial oxygen consumption compared by high resolution respirometry in the presence of CI and CII substrates and inhibitors in 12-month-old (D) female and (E) male mice (n = 6-10). *p < 0.05 (multiple t test). Error bars indicate mean ± SEM.
To further assess mitochondrial bioenergetics, maximum mitochondrial respiratory capacity was measured pairwise in freshly isolated cortical mitochondria from 12-month-old sex-matched, nSIRT3+/+ and nSIRT3−/− mice. Samples were compared by high-resolution respirometry (Oroboros, O2K) in the presence of CI and CII substrates and inhibitors. Relative to female nSIRT3+/+ mice, female nSIRT3−/− mice exhibited significantly decreased CI-linked respiration, significantly reduced CI+CII-linked respiration, and a trend toward decreased ET capacity associated with CI+CII respiration (Fig. 4D). ET capacity in the presence of CI inhibitor, rotenone, was not significantly different, suggesting that these differences were driven by CI. Conversely, maximum respiratory capacity was not significantly altered in male nSIRT3−/− mice (Fig. 4E). CI enzymatic activity was not significantly different between male and female nSIRT3−/− mice (Fig. 2E), suggesting that decreased CI activity does not account for differences in CI oxygen consumption. Together, these data suggest that neuronal SIRT3 deficiency increases in vitro glycolytic parameters without overt changes to OXPHOS and induces sex-dependent mitochondrial dysfunction.
Neuron-specific SIRT3 deficiency results in female-specific altered neuronal excitability and working spatial memory impairment
Proper neuronal activity depends on tightly regulated metabolic control. To determine whether neuronal-specific SIRT3 deficiency produced functional deficits, we first investigated network excitability by observing baseline neuronal spiking events using an MEA in male and female, nSIRT3+/+ and nSIRT3−/− mixed neuronal-glial cultures (DIV 9). The total number of spikes recorded over a 30 min period was significantly elevated in female nSIRT3−/− cultures compared with all other groups (Fig. 5A), suggesting a propensity toward neuronal hyperexcitability. Maturity of the cultures was verified by coexpression of Synapsin I and NeuN (Extended Data Fig. 5-1). To determine whether this altered neuronal excitability occurred in adult mice, 12-month-old nSIRT3+/+ and nSIRT3−/− mice were implanted with bilateral EEG electrodes located over the motor/somatosensory cortex. Although no spontaneous seizures were detected over the 2 week observation period, sharp waves known as ISs appeared consistently in EEG records obtained from nSIRT3−/− mice on one or both cortical channels (at least 2 spikes appeared in records obtained from 40% of male nSIRT3+/+ mice, 16.6% of female nSIRT3+/+ mice, 60% of male nSIRT3−/− mice, and 66.7% of female nSIRT3−/− mice). Neuronal spikes are generated by the sudden, synchronous discharge of groups of neurons and are considered a biomarker of hyperexcitability and epilepsy (for review, see Staley and Dudek, 2006). The total number of sharp discharges at least 3× baseline on one or both cortical channels was highly variable within groups but corroborated our in vitro findings of increased excitability in nSIRT3−/− mice with a tendency for increased spiking in female nSIRT3−/− mice (Fig. 5B). Representative traces of interictal spiking and power spectral plot of the corresponding 30 s data segment in 12-month-old female nSIRT3−/− and nSIRT3+/+ are illustrated in Figure 5C. Together, these data demonstrate that neuron-specific SIRT3 deficiency results in altered neuronal excitability in female mice.
Neuronal deletion of SIRT3 results in female-specific altered neuronal excitability. A, Baseline neuronal spiking in an in vitro MEA. Synapse maturity at DIV 9 was verified by coexpression of Synapsin I and NeuN (Extended Data Fig. 5-1). B, In vivo EEG assessment of spikes over a 2 week period in 12-month-old-mice (n = 5-12/group). C, Representative trace of interictal spiking in 12-month-old female nSIRT3−/− and nSIRT3+/+ mice illustrated in the 30 s time series (left column) and power spectral plots of the corresponding data segments (right column) for both hemispheres. *p < 0.05 (two-way ANOVA followed by Sidak's multiple comparison test). Error bars indicate mean ± SEM.
Figure 5-1
Fluorescent micrograph of mixed cortical cells showing Synapsin I expression. These cortical cells were harvested from postnatal day 1-3 mice and matured in culture 9 days (DIV 9). Neurons (green) and synapses (purple) are identified with NeuN and Synapsin I antibodies, respectively. Scale bar, 10 µm. Download Figure 5-1, TIF file.
To probe cognitive function, short-term working spatial memory was assessed using the spontaneous alternation version of the Y-maze task. Female nSIRT3−/− mice exhibited significantly reduced correct spontaneous alternations at 6 months (Fig. 6A), 12 months (Fig. 6B), and in inducible SIRT3 KO (Camk2cre/ERT2) mice (Fig. 6C). These deficits were not accompanied by differences in locomotion, as measured by total arm entries (Fig. 6D,E), indicating that neuronal SIRT3 deficiency is capable of inducing working spatial memory impairment in female mice, specifically.
Neuronal deletion of SIRT3 results in female-specific working spatial memory impairment. Y-maze assessment of spatial working memory in mice at (A) 6 months of age (n = 11-19/group, 2 cohorts), (B) 12 months of age (n = 11-17/group, 3 cohorts), and (C) on inducible neuronal SIRT3 deletion (n = 6-12/group, 4 cohorts). Locomotion assessment, measured by total arm entries in mice at (D) 12 months of age (n = 11-17/group) and (E) on inducible neuronal SIRT3 deletion (n = 6-12/group, 4 cohorts). *p < 0.05 (two-way ANOVA followed by Sidak's multiple comparison test). Error bars indicate mean ± SEM.
Discussion
Increasing evidence suggests that SIRT3 functions in a complex and context-dependent manner to influence health and disease. Outright metabolic dysfunction in germline SIRT3−/− mice occurs only after an additional insult, such as fasting or high-fat diet, leading to the general consensus that SIRT3 functions as a mitochondrial stress sensor (Ahn et al., 2008; Sundaresan et al., 2009; Hirschey et al., 2010; Qiu et al., 2010; Someya et al., 2010; Hirschey et al., 2011; Fernandez-Marcos et al., 2012; Tyagi et al., 2018). Thus, it appears that SIRT3 dysfunction per se is not sufficient to drive pathology perhaps because of compensatory effects. Importantly, most studies in SIRT3 KO mice have reported data comprised mostly or entirely of male mice. Here, we demonstrate that targeted SIRT3 deletion in select neurons of the neocortex and hippocampus, preferentially predisposes female mice to differential protein acetylation, increased cortical O2−. levels, altered neuronal excitability, and working spatial memory deficits. This would suggest that SIRT3 deficiency and resultant SOD2 at lysine site K122 hyperacetylation exerts sexually dimorphic effects on metabolic function. Moreover, data from this study suggest that neuronal deletion of SIRT3 is sufficient to drive dysfunction in females but not in males. Importantly, the data presented here implicate a novel mechanism by which sex differences can arise in aging and neurologic disease.
Mitochondria have evolved to respond to changes in energy demand. Reversible acetylation/deacetylation of proteins provides one such mechanism to rapidly alter mitochondrial function in response to redox state and/or energy demand. In line with its role as the primary mitochondrial deacetylase, germline and cell type-specific deletion of SIRT3 in mice consistently increases mitochondrial protein acetylation (Lombard et al., 2007; Hirschey et al., 2010, 2011; Fernandez-Marcos et al., 2012; Dittenhafer-Reed et al., 2015; Cheng et al., 2016; Tyagi et al., 2018; Nandi et al., 2019). Likewise, our neuron-specific SIRT3 deletion resulted in significant cortical mitochondrial acetylation albeit with fewer identified acetylated sites than previous studies, likely owing to the presence of functional SIRT3 in glia, oligodendrocytes, and other neuronal populations in our study. Like other acetylomic analyses, the majority of the hyperacetylated proteins identified were involved in core metabolic processes, such as the TCA cycle and OXPHOS (Dittenhafer-Reed et al., 2015). Protein immunoblotting revealed significantly increased mitochondrial acetylation in male and female nSIRT3−/− mice, suggesting a unique and persistent effect of reduced SIRT3 activity.
To what degree protein hyperacetylation affects protein activity is largely context-dependent (Fernandez-Marcos et al., 2012). For example, acetylation at one key lysine residue may or may not have greater implications for target protein activity than acetylation at multiple ancillary sites. Here, we identified direct sex-dependent alterations to SOD2 at key functional lysine site K122 (Qiu et al., 2010; Tao et al., 2010; Chen et al., 2011). In addition, we illustrate a female-dependent decrease in SOD2 enzymatic activity, suggesting that acetylation at K122, at least in neurons, is sufficient to affect SOD2 activity in female mice. Similarly, we identified hyperacetylation of known SIRT3 targets, aconitase and CI subunits. Examination of the activity of these proteins in nSIRT3−/− mice revealed that hyperacetylation did not alter CI, aconitase, and fumarase activity, regardless of sex. Hyperacetylation of mitochondrial proteins, under most circumstances, leads to inhibition of target protein activity (Schlicker et al., 2008). Aconitase is one protein whose activity has been shown to increase on hyperacetylation (Fernandes et al., 2015). Conversely, the iron-sulfur core of aconitase is sensitive to specific oxidation by O2−., which is associated with decreased enzymatic activity (Gardner et al., 1995; Patel et al., 1996). Increased O2−. levels in female nSIRT3−/− mice could offset any alterations to aconitase activity induced by hyperacetylation. Thus, here we show that, despite differences in protein acetylation between the sexes, the indirect effects of these major SIRT3 target protein activities are unaffected by biological sex.
CI and SOD2 are robust mitochondrial enzymes, and slight inhibition of activity is not necessarily sufficient to diminish function. CI provides ∼40% of the proton-motive force required for cellular respiration and is also the primary site of O2−. production (Giachin et al., 2016). As a potential dual threat in the highly oxidant-vulnerable brain, CI function is tightly regulated in an activity-dependent manner. Nonsynaptic mitochondria can withstand upwards of 70% depletion of CI activity before a decrease in oxygen consumption (OC) is observed, whereas inhibition of CI activity by as little as 10% is sufficient to impair OC in synaptic mitochondria (Davey et al., 1997, 1998). Here, despite our reported absence of change in CI activity in both male and female nSIRT3−/− mice, only female nSIRT3−/− mitochondria exhibited significantly impaired OC and elevated O2−. levels. Electron flow through CII stimulated by succinate addition was unable to correct deficient OXPHOS in female nSIRT3−/− mice leading to a trend toward overall decreased maximum respiratory capacity. CI is a well-known source of intracellular O2−.. Deficits in CI could contribute additional electrons to the formation of O2−. by electron leak through CI and may be sufficient to create a vicious cycle of O2−. driven O2−. production in female nSIRT3−/− mice. The observed increase in glycolysis in primary cultures from female nSIRT3−/− mice may be indicative of shunting to bypass CI dysfunction and/or to minimize ROS production. On the other hand, this could provide further evidence of aberrant glucose metabolism, which may be attributed to the indirect effects of mitochondrial protein acetylation. Nonetheless, this demonstrates that neuronal-specific SIRT3 deletion induces sexually divergent effects on mitochondrial function, rendering females more vulnerable to metabolic dysfunction and oxidative stress.
Mitochondrial dysfunction and oxidative stress have been implicated as key factors underlying cognitive aging and neurodegenerative disease (Kapogiannis and Mattson, 2011; Johri and Beal, 2012; Joseph et al., 2012; Giachin et al., 2016; Shi et al., 2017). Yet, the role of sirtuins in cognitive aging and cognition more generally is only beginning to emerge (Satoh et al., 2017; Diaz-Perdigon et al., 2020). We provide evidence that female nSIRT3−/− mice exhibit deficits in spatial working memory. Spatial working memory is the ability to keep information active for recall over short periods and is mediated primarily by cortical and hippocampal areas of the brain, regions that exhibited decreased SIRT3 expression in our model (van Asselen et al., 2006). Notably, impairment of working memory is highly affected by the aging process and one of earliest predictors of AD (Dobbs and Rule, 1989; Baddeley et al., 1991; Kirova et al., 2015; Cansino et al., 2020). We observed sex-specific deficits in female nSIRT3−/− mice regardless of age. Moreover, we were able to replicate this sex-specific deficit in spatial working memory on inducible deletion of SIRT3 in adult mice. This suggests that SIRT3 deficiency selectively predisposes female mice to deficits in working memory. In this regard, our data are consistent with the literature in that male nSIRT3−/− mice are relatively unremarkable under normal circumstances. Thus, we are the first to report deficits in memory function as a specific result of neuronal-SIRT3 deletion. Further, these effects were present despite functional SIRT3 in other neuronal and non-neuronal cells in the brain, implicating a particular dependency of forebrain neurons on functional SIRT3. Cortical and hippocampal neurons generally exhibit increased activity and metabolism relative to other brain regions and therefore may be vulnerable to not only AD-related pathology but to the loss of SIRT3 (Buckner et al., 2009).
Further, we show that both in vitro and in vivo female nSIRT3−/− mice exhibit increased baseline neuronal excitability relative to all other groups, a finding that adds to the growing literature suggesting a role for sirtuins in neuronal hyperexcitability and seizures (Wang et al., 2016; Hall et al., 2017). We have previously demonstrated SIRT3 dysfunction in a male rat model of temporal lobe epilepsy (Gano et al., 2018). Neuronal-specific deletion of the SIRT3 downstream effector, SOD2, is sufficient to produce a catastrophic epilepsy (Fulton et al., 2021). Additionally, SIRT3 haploinsufficiency when combined with an AD mouse model exacerbates neuronal spiking and seizure incidence in a GABA-dependent manner (Cheng et al., 2020). Although spontaneous seizures were not detected in the 2 week observation period, interictal spiking was observed. Increased neuronal excitability puts an additional metabolic stress on cellular respiration and likely contributes to the observed female-specific O2−. levels. Increased O2−. can cause further oxidative damage to core ETC components, thus engaging female nSIRT3−/− mice in a feedforward cycle that may lead to synaptic dysfunction.
While neuronal spiking was reported in primary cultures derived from nSIRT3+/+ and nSIRT3−/− mice using MEA recordings and further confirmed in vivo using video-EEG, a limitation of this study is that additional experiments, such as seizure threshold tests to measure hyperexcitability, would be needed to validate a hyperexcitability phenotype. In addition, IS analysis would have been further refined by the inclusion of hippocampal and electromyogram recordings (Gureviciene et al., 2019).
In conclusion, we demonstrate that neuronal-specific deletion of SIRT3 is sufficient to drive sex differences in brain metabolism, altered neuronal excitability, and cognitive function. This work provides a model by which female-specific deficits resulting from mitochondrial dysfunction might arise in aging and neurologic disease.
Footnotes
This work was supported by the following grants: National Institute of Health R37NS039587 (to M.P.), National Institute of Health RO1NS039587S1 (to M.P. and J.N.P.-S.), National Institute of Health R01NS086423 (to M.P.), American Epilepsy Society Postdoctoral Fellowship (to L.B.G.), National Research Service Award NRSA F32NS090808 (to L.B.G.), and National Institute of Health T32 AG000279 (to J.N.P.-S.). Behavioral and EEG studies were performed in conjunction with the Behavior and In Vivo Neurophysiology Core, a member of the NeuroTechnology Center at the University of Colorado-School of Medicine, funded in part by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health Award P30NS048154. Assistance with genomic analysis was provided by the University of Colorado-School of Medicine Bioinformatics and Biostatistics Shared Resource Core supported by the National Institutes of Health P30CA046934. The Graduate Experience for Multicultural Students Program was supported by National Institutes of Health Grant 1R25HL 103286-08. Imaging was performed in the Advanced Light Microscopy Core Facility of the NeuroTechnology Center at the University of Colorado Anschutz Medical Campus, which is supported in part by Rocky Mountain Neurological Disorders Core Grant P30 NS048154 and by Diabetes Research Center Grant P30 DK116073. We thank Robert Valdez and Timothy Corrigan for technical assistance with EEG implantation; Katrina Diener for her technical support with acetylomics and Elizabeth Kaye (University of Colorado Anschutz Medical Campus Advanced Light Microscopy Core) for confocal imaging assistance.
The authors declare no competing financial interests.
- Correspondence should be addressed to Manisha Patel at Manisha.patel{at}cuanschutz.edu
