Cell culture

Mouse hippocampal neuronal cells (HT22) were provided by Procell (Wuhan, China). Cells were incubated at 37°C in an incubator with 5% CO₂ using Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), and low-glucose DMEM was used for comparison.

Cells were treated with Mg²⁺, erastin, RSL3, and Fer-1, as well as the inhibitor of apoptosis (z-VAD-fmk), pyroptosis (PBzyme) and necroptosis (Nec-1s).

RT-qPCR

Cells were lysed using TRIzol buffer (R0016; Beyotime, China) and the total RNA was collected. RNA concentration and purity were detected using a UV spectrophotometer (Bio-Rad, USA). Then RNA was reverse transcribed into cDNA using a reverse transcription kit (RR037B; Takara, Japan). Relative mRNA expression was determined using a SYBR Premix Ex Taq kit (DRR041A; Takara, Japan) and calculated using the 2^–ΔΔCq method. The sequences of the primers used in PCR were as follows: histone deacetylase 4 (HDAC4), F: 5′-CTGCAAGTGGCCCCTACAG-3′ and R: 5′-CTGCTCATGTTGACGCTGGA-3′; and glyceraldehyde-3-phosphate dehydrogenase, F: 5′-AGGTCGGTGTGAACGGATTTG-3′ and R: 5′-GGGGTCGTTGATGGCAACA-3′.

Western blot analysis

Proteins were extracted from HT22 cells. Following concentration and denaturation, the proteins were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes, which were blocked with 5% skimmed milk. Then membranes were incubated with primary antibodies, such as anti-HDAC4 (ab240643; 1 : 1000, Abcam, UK), anti-acetyl H3K9 (ab190479; 1 : 1000, Abcam, UK), and anti-transferrin receptor (TFRC) (ab214039; 1 : 1000, Abcam, UK), followed by secondary antibody (ab205718; 1 : 10000, Abcam, UK). Finally, the immunoreactive bands were captured using an enhanced chemiluminescence reagent. β-actin (ab227387; 1 : 10000, Abcam, UK) served as a loading control.

Co-immunoprecipitation (Co-IP) assay

Cells were lysed with radio-immunoprecipitation assay lysis buffer (20-188; Millipore, USA). Afterwards, the protein lysates were collected and incubated with Protein A/G magnetic beads containing antibodies against HDAC4 (ab240643; 1 : 30, Abcam, UK) and TFRC (ab214039; 1: 30, Abcam, UK). Subsequently, the beads were boiled and subjected to western blot assay.

Determination of malondialdehyde (MDA)

The MDA levels were detected in cell lysates and measured using an MDA Kit (MAK568; Sigma-Aldrich, USA). Briefly, cells were lysed and centrifuged. Then the lysates were supplemented with thiobarbituric acid. Finally, MDA levels were determined at a wavelength of 532 nm.

Determination of intracellular iron levels

The intracellular iron levels were determined using a Colorimetric Iron Assay Kit (ab83366; Abcam, UK).

Glutathione (GSH) assay

GSH levels were determined using a GSH Assay Kit (ab112132; Abcam, UK). Briefly, cells were plated in a 24-well plate. After supplementation with magnesium ion (Mg²⁺)-free solution, low-glucose medium and erastin, cells were lysed with GSH assay buffer. GSH was determined using a kinetic assay and calculated at a wavelength of 412 nm.

Cell Counting Kit-8 (CCK-8) assay

After transfection, cells were collected and plated in a 24-well plate. After being cultured for 48 h, the cells were treated with CCK-8 solution (96992-100TESTS-F; Sigma-Aldrich, Germany) and mixed for another 90 min at 37°C. Subsequently, cell viability was quantified by measuring absorbance at 450 nm using a microplate reader.

Propidium iodide (PI) staining

After fixation in 4% paraformaldehyde and permeabilization with 0.2% Triton X-100, cells were blocked with 5% bovine serum. Then cells were stained with PI. The image was taken using a microscope (M205 FA; Leica, Germany).

Animal experiment

Male Sprague Dawley (SD) rats (80–100 g) were purchased from Changsheng Biology (Liaoning, China). Rats had free access to food and water. This study was approved by the Animal Care Board of The First Affiliated Hospital of Harbin Medical University.

Rats were randomly divided into three groups: the sham group, the kainite acid (KA) group, and the KA + KD group. Rats in the KA group were hypodermically injected with 10 mg/kg KA. Rats in the KA + KD group were hypodermically injected with 10 mg/kg KA and administered KD formula (76% fat, 16% protein, 3% carbohydrate, and 5% dietary fibre in kcal) (Kuibuqianli Biology, Xuancheng, China). Rats in the sham group were hypodermically injected with saline.

Immunofluorescence assay

After anaesthesia, the rats were sacrificed and the brain tissues were collected. The slides were fixed in paraffin, deparaffinized and immersed in ethylenediaminetetra-acetic acid buffer. After blocking with 1% bovine serum albumin regents, the sections were incubated with primary antibodies against NeuN (ab177487; 1 : 100, Abcam, UK) and HDAC4 (ab313474; 1 : 200, Abcam, UK), and then with secondary antibody (ab150077; 1 : 200, Abcam, UK). Nuclei was counterstained with 4’,6-diamidino-2-phenylindole. Finally, the images were captured using a microscope (M205 FA; Leica, Germany).

Statistical analysis

Data were presented as mean ± standard deviation. Two-group comparison was analysed using Student’s t test. Multi-group comparison was analysed using analysis of variance. P < 0.05 was deemed as statistically significant.

Epilepsy-induced ferroptosis-related neuronal loss

Ferroptosis-mediated neuron loss is involved in the progression of epilepsy. To further confirm this, neurons were treated with a ferroptosis inducer (erastin) as well as an inhibitor (Fer-1) after exposure to Mg²⁺-free solution. We found that erastin treatment significantly enhanced the effects of Mg²⁺ treatment and suppressed cell viability of neurons, suggesting that the progression of epilepsy may induce neuronal ferroptosis. However, epilepsy can induce other forms of cell death, such as apoptosis, pyroptosis and necroptosis. To further confirm this, cells were treated with inhibitors of apoptosis (z-VAD-fmk), pyroptosis (PBzyme) and necroptosis (Nec-1s). As shown in Figure 1 B, the decrease in the cell viability induced by Mg²⁺ treatment was significantly alleviated by Fer-1, which showed no significant alteration by the inhibitors of apoptosis, pyroptosis, and necroptosis. Mg²⁺ treatment significantly increased the release of MDA and ferrous iron (Fe²⁺), and decreased GSH (Figures 1 C–E), which was significantly enhanced by erastin, but reversed by Fer-1. Moreover, the increase in the percentages of PI positive cells was significantly enhanced by erastin (Figures 1 F, G), but reversed by Fer-1. These findings confirmed that the progression of epilepsy is accompanied by the ferroptosis of neurons.

Figure 1

Epilepsy-induced ferroptosis-related neuronal loss. A, B – Cell viability was detected using CCK-8 assay. C – Release of MDA was determined using MDA assay. D – Release of ferrous iron was detected using ELISA assay. E – Release of GSH was detected using GSH assay. F, G – Death of neurons was determined using PI staining

MDA – malondialdehyde (MDA), GSH – glutathione, CCK-8 – Cell Counting Kit-8, PI – propidium iodide. **P < 0.01, ***p < 0.001, ##p < 0.01, ###p < 0.001.

KD suppresses the ferroptosis of neurons induced by epilepsy in vitro

KD is an effective strategy for epilepsy. To conform this, HT22 cells were treated with KD after Mg²⁺ treatment. We found that the cell viability was significantly decreased by Mg²⁺ treatment, which was reversed by KD treatment. KD treatment also significantly antagonized the effects of Mg²⁺ treatment and decreased the release of MDA and ferrous iron (Fe²⁺), and increased GSH (Figures 2 B–D). KD treatment significantly reduced the percentage of PI-positive cells induced by Mg²⁺ treatment (Figures 2 E, F). These findings suggested that KD treatment alleviates neuronal ferroptosis in epilepsy.

Figure 2

KD suppresses ferroptosis of neurons induced by epilepsy in vitro. A – Cell viability was detected using CCK-8 assay. B – Release of MDA was determined using MDA assay. C – Release of ferrous iron was detected using ELISA assay. D – Release of GSH was detected using GSH assay. E, F – Death of neurons was determined using PI staining

KD – ketogenic diet, MDA – malondialdehyde (MDA), GSH – glutathione, CCK-8 – Cell Counting Kit-8, PI – propidium iodide. ***P < 0.001, ##p < 0.01, ###p < 0.001.

HDAC4 is downregulated in epilepsy

GDS954 was used to analyse genes differentially expressed in epilepsy. We found that HDAC4 was significantly decreased (data not presented here). HDAC4 participates in the progression of epilepsy. Then we determined HDAC4 expression in epilepsy in vitro models. We found that HDAC4 mRNA expression was significantly reduced after Mg²⁺ treatment (Figure 3 A). This was paralleled by the results from western blot assay. As shown in Figures 3 B, C, Mg²⁺ treatment significantly suppressed HDAC4 protein expression.

Figure 3

HDAC4 is downregulated in epilepsy. A – HDAC4 mRNA expression was detected using RT-qPCR. B, C – HDAC4 protein expression was detected using western blot

HDAC4 – histone deacetylase 4. ***P < 0.001.

HDAC4 deficiency promotes the ferroptosis of neurons

To further confirm the roles of HDAC4 in epilepsy, neurons were treated with the HDAC4 specific inhibitor LMK235. As shown in Figure 4 A, KD treatment-mediated restoration of neuronal viability was significantly alleviated by LMK235. LMK235 also significantly promoted the release of MDA and Fe²⁺ (Figures 4 B, C), whereas it decreased GSH (Figure 4 D). Moreover, LMK235 treatment significantly antagonized the effects of KD treatment and increased the percentages of PI positive cells (Figures 4 E, F). These findings suggested that KD inhibits the ferroptosis of neurons in epilepsy via upregulating HDAC4.

Figure 4

HDAC4 deficiency promotes ferroptosis of neurons. A – Cell viability was detected using CCK-8 assay. B – Release of MDA was determined using MDA assay. C – Release of ferrous iron was detected using ELISA assay. D – Release of GSH was detected using GSH assay. E, F – Death of neurons was determined using PI staining

KD – ketogenic diet, MDA – malondialdehyde, GSH – glutathione, CCK-8 – cell counting Kit-8, PI – propidium iodide, HDAC4 – histone deacetylase 4. **P < 0.01, ***p < 0.001, ##p < 0.01, ###p < 0.001, &&p < 0.01.

HDAC4 promotes the histone deacetylation of TFRC

Ferroptosis is characterized by iron overload-induced lipid peroxidation [17]. Therefore, we hypothesized that iron metabolism may be involved in the progression of epilepsy. TFRC, as an iron uptaker, promotes the ferroptosis of neurons during the pathogenesis of epilepsy. We found that the protein expression as well as the global acetylation of TFRC was significantly increased after Mg²⁺ treatment (Figures 5 A, B). HDAC4, as a member of the histone deacetylases, regulates gene expression via exerting histone deacetylase activity. Therefore, we hypothesized that KD may regulate ferroptosis via regulating the HDAC4/TFRC pathway. Endogenous IP assay showed that TFRC can interact with HDAC4 (Figure 5 C). The results of IP experiments showed that overexpression of HDAC4 could significantly reduce the acetylation of TFRC (Figure 5 D).

Figure 5

HDAC4 promotes histone deacetylation of TFRC. A, B – Protein expression was detected using western blot. C, D – Interaction between HDAC4 and TFRC was detected using Co-IP assay

HDAC4 – histone deacetylase 4, TFRC – transferrin receptor.

KD alleviates epilepsy progression in vivo

To further confirm the effects of KD on epilepsy, rats were administered KA and/or KD. We found that KD treatment significantly reduced the total number of seizures induced by KA (Figure 6 A). Moreover, KD treatment also significantly decreased the mean number of seizures/week from week 2 (Figure 6 B). We further found that KD treatment significantly improved motor skill learning (Figure 6 C) as well as hippocampal associative memory (Figure 6 D).

Figure 6

Alleviation of epilepsy progression in vivo. A – Total number of seizures after KA and/or KD treatment. B – Mean seizures per week after KA and/or KD treatment. C – Motor learning skill test after KA and/or KD treatment. D – Contextual fear conditioning testing after KA and/or KD treatment

KA – kainite acid, KD – ketogenic diet. **P < 0.01, ***p < 0.001, ##p < 0.01, ###p < 0.001.

As shown in Figures 7 A–C, KA treatment significantly increased the release of MDA and Fe²⁺, whereas it decreased GSH; however, this was alleviated by KD administration. KA injection significantly promoted the loss of neurons, which was alleviated by KD administration (Figure 7 D). Moreover, we also found that KD administration reversed the effects of KA treatment and increased the percentages of HDAC4⁺NeuN⁺ cells (Figure 7 E). This further confirmed that KD alleviated epilepsy-mediated ferroptosis of neurons via upregulating HDAC4.

Figure 7

KD inhibits ferroptosis in vivo. A – Release of MDA was determined using MDA assay. B – Release of ferrous iron was detected using ELISA assay. C – Release of GSH was detected using GSH assay. D – NeuN expression in brain tissues was determined using immunofluorescence. E – NeuN and HDAC4 expression in brain tissues was determined using immunofluorescence

KA – kainite acid, KD – ketogenic diet, MDA – malondialdehyde (MDA), GSH – glutathione, HDAC4 – histone deacetylase 4. ***P < 0.001, ##p < 0.01.