Therapeutic effect of a histone demethylase inhibitor in Parkinson’s disease

Selective iron suppression effect of GSK-J4 on DA neurons

To examine the effects of GSK-J4 on the potentially toxic intracellular labile iron pool (LIP) in cells, a calcein-AM assay was performed. Since the fluorescence of calcein is partially quenched upon binding to iron, compounds that displace intracellular iron from its complex with calcein would result in an increased fluorescence signal. As shown in Fig. 1a, b, SH-SY5Y cells treated with the iron supplement FAC (50 μg/ml) for 12 h displayed a lower calcein fluorescence level than the control group (P < 0.05). However, calcein fluorescence increased with the addition of 0.5 μM GSK-J4 (P < 0.001, compared with control group), suggesting a decrease in free labile iron in the SH-SY5Y cells with the treatment of GSK-J4. Moreover, the addition of FAC (50 μg/ml) inhibited the GSK-J4-mediated increase of calcein fluorescence (P < 0.01, compared with GSK-J4 group), further indicating that the elevated calcein signal by GSK-J4 was caused by the iron suppressing effect of GSK-J4.

We further characterized the effect of GSK-J4 on cellular iron homeostasis by examining the expressions of the iron importer protein TfR1 and the iron storage protein ferritin light chain (FtL). As shown in Fig. 1c, d in SH-SY5Y cells treated with the iron chelator deferoxamine (DFO, 50 μM), TfR1 synthesis was increased compared to untreated control cells (P < 0.01), whereas synthesis of FtL is strongly inhibited (P < 0.01). Interestingly, similar to DFO, only a trace amount of GSK-J4 (0.5 μM) treatment caused a significant increase of TfR1 (P < 0.001) and a decrease of FtL (P < 0.01). These findings are consistent with that under the condition of free labile iron depletion, IRP1 stabilizes TfR1 mRNA to prolong its half-life, while inhibiting the translation of FtL mRNA^23,24. These effects enhance iron uptake into cells and stops it from being sequestered away for storage. However, the effects of GSK-J4 treatment on FtL and TfR1 were not found in HEK293 (Fig. 1e, f) and HepG2 cell line (Fig. 1g, h). In contrast, the classical iron chelator DFO (50 μM) caused a significant increase of TfR1 (P < 0.01, compared with each control group) and a significant decrease of FtL (P < 0.01, compared with each control group), in all these cell lines (Fig. 1c–h). In addition, the iron suppression effect was also found in the MES23.5 cell line (Supplementary Fig. 1a, b). Since SH-SY5Y and MES23.5 both are widely used DA neuronal cell lines, these results indicate that GSK-J4 has a selective iron suppressing effect on DA neurons.

GSK-J4 displays a potent neuroprotective effect in 6-OHDA-induced PD model in vitro

Considering the cell-type specificity of the iron-suppressing effect of GSK-J4, we hypothesize that it is beneficial in PD. To investigate the effect of GSK-J4 on a PD model in vitro, SH-SY5Y cells were treated with the neurotoxin 6-hydroxydopamine (6-OHDA). As shown in Fig. 2a, 6-OHDA induced a dose-dependent shrinkage, detachment, and cell death in SH-SY5Y cells in 24 h. Compared with 6-OHDA-treated cells, pre-treatment with GSK-J4 (0.5 μM) clearly protected cells against 6-OHDA toxicity, even at a high concentration of 6-OHDA (40 μM). The protective effect of GSK-J4 was also quantitatively assessed by the MTT assay (Fig. 2b). For example, upon 40 μM of 6-OHDA treatment for 24 h, 0.5 μM of GSK-J4 significantly increased cell viability from 30.5 ± 10.9 to 86.6 ± 8.4% (P < 0.001). On the other hand, treatment with GSK-J4 alone (0.5 μM) did not have a significant effect on morphology and cell viability. These data suggest that GSK-J4 showed a powerful neuroprotection in the 6-OHDA-treated SH-SY5Y cells. We also examined the expressions of two important markers of cell apoptosis, namely cleaved Caspase-3 and Bcl-2 by Western blot. 6-OHDA treatment increased the level of cleaved caspase-3 to 160% (Fig. 2c, d) and decreased the level of Bcl-2 to 51% when compared with the control group, while pretreatment with GSK-J4 significantly prevented these changes (Fig. 2c, d). These data suggest that GSK-J4 decreased the 6-OHDA-induced apoptosis in SH-SY5Y cells.

GSK-J4 is a highly cell permeable ethyl ester derivative of GSK-J1¹⁹ (Fig. 2e). To compare the neuroprotective effects of these two compounds, SH-SY5Y cells were pretreated with GSK-J1 (0–10 μM) 1 h before the 6-OHDA (30 μM, 12 h) treatment, and MTT assay was performed. Consistently, GSK-J1 prevented 6-OHDA-induced cell death (P < 0.01, comparing the 6-OHDA group with 6-OHDA+GSK-J1 group) (Fig. 2f), but a concentration of 10 μM GSK-J1, that is, almost 40 times that of GSK-J4 at 0.25 μM, was required in order to display the same neuroprotective effect. The difference is therefore attributable to the high cell membrane permeability of GSK-J4 compared with GSK-J1.

Neuroprotective effect of GSK-J4 in PD models in vivo

To explore the neuroprotective and therapeutic effects of GSK-J4 in in vivo PD model, SD rats were mico-injected with 6-OHDA (8 μg/2 μl) at the medial forebrain bundle (MFB) in the right side of the brain. GSK-J4 (1 μg/1 μl) was injected into MFB one day before the 6-OHDA administration (Fig. 3a). Three weeks after injection, the relevant behavior tests were performed before sacrifice of the animals for histological and biochemical assessments. In the cylinder test, unilateral 6-OHDA injection decreased the use of the contralateral forelimb to about 20% while GSK-J4 treatment partially but significantly rescued this motor defect, in which the use of contralateral forelimb was almost 40% (Fig. 3b). In the apomorphine-induced rotation test, GSK-J4-treated rat exhibited far fewer number of contralateral rotation (15 rpm) in 5 min when compared with the unilateral 6-OHDA-treated rat, which was about 36 rpm in the same period of time (Fig. 3c). These results indicate that GSK-J4 can prevent the 6-OHDA-induced motor defects in rats. At the same time, based on tyrosine hydroxylase (TH) staining, it was found that 6-OHDA injection caused a significant decrease in the number of TH-positive neurons in SN, which was about 25% of the saline-treated side. With GSK-J4 treatment, 50% of TH-positive neurons survived after 6-OHDA injection indicating a neuroprotective effect of GSK-J4 (Fig. 3d, e).

To establish the link between the beneficial effect of GSK-J4 and iron level in the PD animal model, we first asked whether iron was altered in the unilateral 6-OHDA-induced rat model. Iron staining (Fig. 3f, g) showed that the number of iron-positive cells significantly increased in the 6-OHDA-lesioned SN, indicating increased iron content after 6-OHDA treatment. However, the treatment of GSK-J4 significantly inhibited this increase of iron, suggesting that GSK-J4 can reduce iron accumulation in response to 6-OHDA treatment.

Anti-oxidative stress is involved in the neuroprotection of GSK-J4

Iron-induced oxidative stress as a result of iron accumulation is regarded as an important cause of neurodegeneration in many experimental PD models. Given the iron suppression effect of GSK-J4 that we observed, we investigated whether anti-oxidative stress plays a role in the neuroprotective effect of GSK-J4 in PD model.

We determined the levels of different oxidative stress markers in the SN of 6-OHDA-treated rats in the absence and presence of GSK-J4 pre-administration. It was found that GSK-J4 could significantly suppress the 6-OHDA-induced increase of reactive oxygen species (ROS) production, as measured by DCF fluorescence, from 176 ± 7.9% to 131 ± 6.3% (Fig. 4a), the increase of MDA level from 0.79 ± 0.10 to 0.52 ± 0.12 nmol (Fig. 4b), the increase of protein carbonyl level from 5.84 ± 0.54 to 3.52 ± 0.25 nmol (Fig. 4c) and the decrease of total GSH level from 10.27 ± 1.05 to 15.27 ± 1.08 μg/mg (Fig. 4d). These results indicate that GSK-J4 acts as a strong antioxidant in the 6-OHDA-induced PD model.

To further confirm the antioxidant nature of GSK-J4, we tested the effect of this compound on SH-SY5Y cells treated with H₂O₂. We found that GSK-J4 at 0.25 μM prevented H₂O₂-induced changes in cell death reflected by changes in morphology (Fig. 4e). In both 150 and 300 μM concentrations of H₂O₂, pretreatment with GSK-J4 maintained a high level of cell viability, confirmed with MTT assay (Fig. 4f).

GSK-J4 prevents abnormal histone methylation in 6-OHDA model

As a selective inhibitor of histone demethylases including KDM5 and KDM6, GSK-J4 could increase the levels of H3K4me3 and H3K27me3, respectively^19,20. Therefore, apart from the effect on iron level and oxidative stress, we asked whether the effects of GSK-J4 on H3K27me3 and H3K4me3 contribute to its therapeutic action in PD animals. As shown in Fig. 5a, b, SH-SY5Y cells treated with GSK-J4 (0.5 μM) for 24 h exhibited an increase in the level of H3K27me3 (1.72-fold of control group), and an increase in the level of H3K4me3 (1.62-fold of control group) in SH-SY5Y cells. Moreover, there were significant reductions in H3K27me3 (P < 0.01) and H3K4me3 (P < 0.05) under the 6-OHDA (25 μM) treatment when compared with the control group but pretreatment with GSK-J4 significantly attentuated this decrease (P < 0.05, when compared with the 6-OHDA-lesioned group, Fig. 5a, b). These data reveal abnormal levels of H3K4me3 and H3K27me3 in PD model that could be rectified by GSK-J4.

H3K4me3 contributes to the neuroprotection of GSK-J4 treatment on the 6-OHDA-induced PD model

To further explore and distinguish the roles of H3K27me3 and H3K4me3 in the protective effect of GSK-J4 on the 6-OHDA-induced PD model, two sets of experiments were performed. First, SH-SY5Y cells were treated with GSK-J4 (0.5 μM) followed by treatment with a highly selective EZH2 methyltransferase inhibitor (GSK126, 1 μM) that could potently inhibit H3K27me3^25,26. As shown in Fig. 5c, in 6-OHDA-treated cells, the increase of H3K27me3 induced by GSK-J4 was inhibited by GSK126 (P < 0.001, when comparing the 6-OHDA+GSK-J4 group with 6-OHDA+GSK-J4+GSK-126). However, MTT assay showed that there was no significant difference in the cell viabilities of these two groups (Fig. 5d). Next, we performed similar experiments to determine the role of H3K4me3. SH-SY5Y cells were treated with GSK-J4 (0.5 μM), followed by the treatment with 3-deazaneplanocin A (DZNep, 1 μM), which was reported to inhibit H3K4me3 levels²⁷. As shown in Fig. 5e, H3K4me3 was inhibited in the presence of DZNep (P < 0.01, when comparing the 6-OHDA+GSK-J4 group with 6-OHDA+GSK-J4+DZNep). In addition, MTT assay showed that DZNep partially antagonized the neuroprotective effect of GSK-J4 (P < 0.05, when comparing the 6-OHDA+GSK-J4 group with 6-OHDA+GSK-J4+DZNep) (Fig. 5f). These data suggest that the increase of H3K4me3, but not that of H3K27me3, contributes to the neuroprotective effect of GSK-J4 against 6-OHDA treatment.

H3K4me3 plays a role in iron metabolism under GSK-J4 treatment

Since GSK-J4 was found to selectively regulate iron metabolism in SH-SY5Y cells, while histone modifications is often locus type and cell type specific^28–30. To fully elucidate the therapeutic mechanism of action of GSK-J4, we explored the relationship between H3K4me3 and iron homeostasis. First, in SH-SY5Y cells, Western blot analysis revealed that, when compared with the control cells, GSK-J4 treatment (0.5 μM, 24 h) significantly increased the levels of iron uptake proteins TfR1 (P < 0.01), the iron export protein Fpn1 (P < 0.01), and decreased the iron storage protein FtL (P < 0.01) (Fig. 6a, b). Together, these data are consistent with that increased iron export by Fpn1 is the primary cause of iron-suppressing effect of GSK-J4. Indeed, iron storage protein FtL was largely increased (P < 0.01) by 6-OHDA treatment when compared with control cells, while a marked decrease of FtL was found in the 6-OHDA plus GSK-J4 treatment group (P < 0.01, when compared with 6-OHDA-treated cells). Moreover, GSK-J4 could significantly prevent the 6-OHDA-induced decrease of Fpn1 and TfR1. In another set of experiments, it was found that the increase of Fpn1 could be inhibited by the addition of DZNep (decreasing H3K4me3) (Fig. 6c, d) while not by the addition of GSK126 (decreasing H3K27me3) (Fig. 6e, f). Consistent with the effect on iron metabolism, the effects of GSK-J4 on H3K4me3 and H3K27me3 were only found in SH-SY5Y cell line but not in HEK293 or HepG2 cell lines (Supplementary Fig. 2a, b). Furthermore, real-time PCR experiments showed that GSK-J4 could also upregulate the expression of Fpn1 at the mRNA level (Fig. 6g). These results together implicate that H3K4me3 plays a role in the iron metabolism under GSK-J4 treatment.

Upregulation of H3K4me3 increases iron exporter Fpn1 and exhibits neuroprotection

To further probe the role of H3K4me3 in cell-specific iron regulation of GSK-J4, SH-SY5Y cells were pretreated with CPI-455, a specific inhibitor of KDM5, which could increase the global level of H3K4me3³¹. We found that CPI-455 (1 and 5 μM) caused a significant increase (over 2-fold) in the global level of H3K4me3 (P < 0.01, when compared with the control group, Fig. 7a, b). Under this condition, CPI-455 treatment also resulted in an increase in the iron exporter Fpn1. In addition, CPI-455 treatment improved cell viability in response to 6-OHDA exposure (Fig. 7c).

KDM5C is the main histone demethylase that can specifically alter H3K4 methylation in several cell types^31,32. To further probe the role of H3K4me3 in regulating iron metabolism, the effect of AAV-KDM5C-silence virus was tested. As shown in Fig. 7d, KDM5C silencing strongly decreased KDM5C levels to 17% of control based on the mRNA level and to 16% of control based on the protein level (Fig. 7e, f).

Remarkably, significant increases of Fpn1 and H3K4me3 were observed under this condition (Fig. 7e, f). Also, KDM5C silencing improved the viability of cells under 6-OHDA treatment (Fig. 7g). Thus, KDM5C-mediated upregulation of H3K4me3 is involved in the increase of iron exporter (Fpn1) underlying the neuroprotective effect of GSK-J4.

Drugs and chemicals

The following chemical agents were used in the experiment: GSK-J4 (S7070, Selleck Chemicals), GSK-J1 (S7581, Selleck Chemicals), 6-OHDA (H116, SIGMA), H₂O₂ (H1009, SIGMA), CPI-455 (S8287, Selleck Chemicals), GSK126 (15415, Cayman Chemical), DZNep HCl (S7120, Selleck Chemicals).

Animals

Sprague-Dawley (SD) rats weighing 320–350 g were used in this study. The animals were housed in standard cages with ad libitum access to both food and water. All animals were housed at a temperature of 25 °C under a 12-h light-on–light-off schedule. After that, all animals were randomly divided into different groups. All animal procedures were performed in accordance with the protocol approved by the Chinese University of Hong Kong Animal Ethics and Experimentation Committee.

Cell culture

Cell cultures were prepared using the methods described previously⁴⁹. In brief, undifferentiated SH-SY5Y cells and MES23.5 cells (kindly provided by Prof. Wei-dong Le, Shanghai Institutes for Biological Sciences, CAS) were grown in a 1:1 mixture of Dulbecco’s modified Eagle’s medium (DMEM) (10313021, Gibco, USA) and F-12 nutrient mixture (Ham12) (11765062, Gibco, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (10082147, Gibco, USA). In addition, the human hepatocellular carcinoma (HepG2) cell line and HEK293 cells (human embryo kidney cells), purchased from the American ATCC Cell Line Center, were maintained as an adherent cell line in DMEM supplemented with 10% FBS, 2 mmol/l l-glutamine (Gibco, USA), and 1× nonessential amino acids (Gibco, USA). Cells were passaged as needed using 0.5% trypsin–EDTA (Gibco, USA). All culture mediums contained 100 units/ml of penicillin (Gibco, USA) and 100 mg/ml of streptomycin (Gibco, USA). All cells were cultured at 37 °C in a 5% CO₂ and 95% air-humidified atmosphere.

MTT assay

To study the effect of GSK-J4 on neuroprotection, MTT assays were performed to assess cell viability as previously described⁴⁹. In brief, SH-SY5Y cells were cultured in the 96-well plates at a density of 1 × 10⁵ cells per well at 37 °C. Two days later, cells were placed in a fresh medium with different treatments. After a particular treatment, cell viability was assessed by the MTT assay (11465007001 ROCHE, Cell Proliferation Kit I (MTT)). First, the supernatant in each culture well was removed and the wells were washed with PBS twice. 20 μl MTT solution (5 mg/ml in PBS) was added to each well (containing 100 μl culture medium). After incubating for 4 h, the medium was removed carefully. Then, 150 μl DMSO was added to each well and shaken for 10 min to allow the crystals to be fully melted. Finally, the absorbance intensity was measured at 490 nm with a microplate reader (Bio-RAD 680, USA) and together with a reference wavelength of 620 nm. Control wells (cells with the same concentration of drug in medium) were set at the same time. The cell viability (%) was expressed as a percentage relative to the control group.

PD animal models

Behavioral test

Immunostaining

Immunostaining was carried out as reported before with some modifications⁵³. Briefly, the rats having received different treatments were deeply anesthetized with 4% chloral hydrate (0.4 mg/g, i.p.) followed by perfusion with normal saline through the left ventricle. The brains were removed and post-fixed in 4% paraformaldehyde (PFA) followed by cryoprotection with serial sucrose solutions. After that, brain tissues were immersed in cryo-embedding media (OCT) and brain sections (30 μm) that contained the SN (AP: −4.80 mm to AP: −6.12 mm from bregma) were prepared by a cryotome. Every fourth section was stained for TH and analyzed. The brain slices were then incubated in blocking solution followed by overnight incubation at 4 °C with primary antibodies such as anti-TH (1:1000, Millipore). After washing with PBS, the slides were incubated with Alex 546 or FITC-conjugated secondary antibody for 1 h at room temperature. The nuclei were counterstained with DAPI. The slides with immunofluorescence staining were visualized with a Nikon C-1 confocal laser scanning microscope (Nikon). The number of TH-positive cells as well as the areas of SN (in mm²) were analyzed by using ImageJ software (from National Institutes of Health and available at http://rsb.info.nih.gov/ij/).

Western blotting

Brain tissues or cells were washed and homogenized in RIPA lysis buffer and then sonicated. Aliquots of the extract containing about 30 μg of protein were loaded and ran on a single track of 10–12% SDS–PAGE under reducing conditions and subsequently transferred to a pure nitrocellulose membrane (Bio-Rad). The blots were blocked and incubated with primary antibodies anti-H3K4me3 (#9751, Cell Signaling Technology, 1:1000), anti-H3K27me3 (07-449, Merck Millipore, 1:1000), anti-KDM5C/SMCX/Jarid1C (ab190180, ABCAM 1:1000), anti-Ferritin-L (ab69090, ABCAM, 1:1000), anti-transferrin receptor 1 (13-6800, Thermo Fisher Scientific, 1:1000), anti-ferroportin/SLC40A1 antibody (NBP1-21502, Novus Biologicals, 1:1000), anti-Bcl2 (ab692, ABCAM, 1:1000), and anti-cleaved caspase-3 (#9664, Cell Signaling Technology, 1:1000) overnight at 4 °C. After the incubation, the blots were washed and then incubated with goat anti-rabbit or anti-mouse IRDye 800 CW secondary antibodies (1:5000, Li-Cor) for 1 h at room temperature. The intensities of specific bands were detected and analyzed by Odyssey infrared imaging system (Li-Cor). To ensure even loading of the samples, the same membrane is probed with rabbit anti-β-actin polyclonal antibody at a 1:5000 dilution.

Real-time PCR

Total RNA was isolated from cultured cells using Trizol^® Reagent (Invitrogen Corp., Carlsbad, CA) following the manufacturer’s protocol. RNA concentrations were evaluated utilizing spectrophotometer and the purity of the RNA was assessed by calculating the ratio of A260–A280 nm signals. RNA was reversely transcribed using Superscript II Reverse Transcriptase (Invitrogen Corp., Carlsbad, CA) and Oligo(dT) Primer (Invitrogen Corp., Carlsbad, CA) following the manufacturer’s instruction. The cDNA for primer design for RT-PCR was previously reported. Primer sequences were as follows:

KDM5C forward primer, 5′-GAGGTGACCCTGGATGAGAA-3′, and KDM5C reverse primer, 5′-CAGGAGCTGAGGTCTGAAC-3′;

FPN1 forward primer, 5′-CCA CCT GTG CCT CCC AGA T, and FPN1 reverse primer, 5′-CCC ATG CCA GCC AAA AAT AC-3′.

cDNA amplification and detection were fulfilled using MyiQ Real-time PCR Detection System (Bio-Rad Laboratories, CA) and iQ SYBR Green Supermix (Bio-Rad Laboratories, CA). The initial denaturation was 95 °C for 5 min, then 95 °C for 10 s followed by 60 °C for 45 s, which cycled for 40 times.

The β-actin cDNA (3′-primer, 5′-CTCTCAGCTGTGGTGGTGAA-3′; 5′-primer, 5′-GTCGTACCACTGGCATTGTG-3′) was simultaneously amplified as the internal control. Relative quantification exploited the comparative Cx method. The mRNA level of KDM5C of each sample was normalized to that of the level of β-actin and control group samples were used as calibrator. Relative mRNA level was calculated by the 2^−ΔΔCT method.

DAB-enhanced Perls’ iron staining

DAB-enhanced Perls’ iron staining was performed as described previously⁵⁴. Briefly, the slides were fixed with 4% PFA for 10 min. After three washes with PBS, the slides were incubated in Perls’ solution containing 8% potassium ferrocyanide and an equal volume of 1.2 mmol/l hydrochloride acid solution for 16 h at 4 °C and then washed with deionized water three times, followed by dehydration through 95% alcohol and mounting with xylene. Afterwards, the sections were washed and incubated in a solution of DAB to enhance the signals. Iron staining was observed under a Nikon Eclipse TE2000-U microscope (Nikon, UK).

Calcein-AM assay

The ferrous iron measurement on cell was performed using calcein-AM method as previously described, with some modifications⁵⁵. Briefly, SH-SY5Y cells were grown in DMEM supplemented with 10% heat-inactivated fetal bovine serum, and cultured at 37 °C under humidified 5% CO₂ atmosphere. Cells were treated with ferric ammonium citrate (FAC) or GSK-J4 or GSK-J4 plus FAC for 12 h in serum-free media. Then they were washed with PBS twice, followed by the addition of calcein-AM (50 nM), and incubated for 30 min at 37 °C. After that, fluorescence was measured using a microplate fluorometer (Thermo Fisher Scientific Inc).

ROS, protein carbonylation determination

ROS were measured by the cell-permeant 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) assay as previously described with minor modifications⁵⁶. Briefly, brain, cell, or protein extraction were incubated with 5 μM H₂DCFDA in dark at 37 °C for 30 min and then washed three times with PBS. DCF fluorescence was observed under Nikon C-1 confocal laser scanning microscope and quantified using a fluorescence microplate reader (LS55 fluorescence Spectrometer, PerkinElmer) with an excitation/emission wavelength of 485/515 nm in 96-well fluorescent plates. Protein carbonyls were measured by using an OxyBlotTM protein oxidation detection kit (Chemicon), in which carbonyl groups in the protein side chains are derivatized to 2,4-dinitrophenylhydrazone (DNP) by reaction with 2,4-dinitrophenylhydrazine (DNPH).

MDA level

Lipid peroxidation was determined in total brain lysates using the N-methyl-2-phenylindole based LPO-586™ lipid peroxidation kit (Oxis International, OR, USA)⁵⁷. This assay measures the level of MDA. Standard curves of MDA were established using 1,1,3,3 tetramethocypropane. To minimize non-specific oxidation during sample preparation, 5 mm butylated hydroxytoluene (BHT dissolved in acetonitrile) was added to the extraction buffer. The assay was performed in triplicates according to manufacturer’s recommendation using 4 mg of total brain lysates per reaction, and the results were calculated as picomoles of MDA per mg of protein. Briefly, 200 µl of brain lysates (4 mg) was added to 10 µl of 0.5 M BHT and 650 µl of 0.1 mM N-methyl-2-phenylindole, followed by gentle mixing. For MDA measurement, 150 µl of reagent grade HCl (∼36%) was added. The sample was mixed and incubated for 1 h at 45 °C. After incubation, the sample was centrifuged, and the supernatant was extracted and measured at 586 nm.

GSH level

Total reduced GSH was measured as described previously⁵⁷. Briefly, the acid-soluble fraction was obtained from cell lysates and tissue homogenates by adding perchloric acid to a final concentration of 3%, followed by centrifugation at 14,000 × g for 10 min. The acid-soluble fraction was neutralized to pH 7 with 0.5 M KOH/50 mM Tris. After the removal of precipitate (potassium perchlorate) by a second centrifugation, 50 μl aliquots of sample was combined with 100 μl of reaction mixture consisting of 2.5 ml of 1 mM 5,5′-dithio-bis-(2-nitrobenzoic acid) (DTNB), 2.5 ml of 5 mM NADPH, 2.5 ml of phosphate buffer solution (100 mM NaPO4, pH 7.5, 1 mM EDTA), and glutathione reductase (5 U/ml final volume). GSH-mediated reduction of DTNB was measured at 412 nm at 30 s intervals over 30 min. GSH content was normalized to protein.

Construction of AAV virus

The coding silence sequence of KDM5C will be first ligated into a shuttle plasmid pAAV with EGFP-tag (Addgene). Insertion of the coding sequence was then verified by sequencing. The positive recombinant plasmid was co-transfected with pAAV-RC, pAAV-helper (Addgene) into HEK293 cells. AAV was harvested 3 days later and designated as “AAV-KDM5C”. The titers of the virus were measured according to standard procedure.

Statistical analysis

All data were presented as mean ± standard error of the mean (SEM). Graphpad Prism was used for statistical analysis. Unpaired Student’s t test or one-way ANOVA with a post hoc Tukey’s test was performed, as appropriate, to determine significant differences between groups. A probability value of P < 0.05 was taken to be statistically significant. All the experiments consisted of at least three replicates.