Graphical abstract

   graphic file with name fx1.jpg
   [37]Open in a new tab

Highlights

     * •
       Combined omics analysis and animal experiments
     * •
       Omics analysis indicated occurrence of ferroptosis in epilepsy
       patients
     * •
       Animal experiments confirmed the hypothesis of omics analysis
     * •
       The key role of HIF-1α/HO-1 pathway in epilepsy
     __________________________________________________________________

   Health sciences; Molecular biology; Neuroscience; Cell biology;
   Metabolomics

Introduction

   Epilepsy is a common chronic disease of the central nervous system
   (CNS) characterized by repeated seizures caused by abnormal electrical
   activity in the brain. Despite long-term efforts in the study of
   epilepsy, the exact pathogenic mechanism is still not fully
   elucidated.[38]^1^,[39]^2 Epileptic seizures can cause neuronal damage,
   ion pathway dysfunction, mossy fiber budding, gliosis, synaptic
   plasticity, and inflammatory response, especially hippocampal neuron
   damage. In turn, hippocampal neuron damage can further promote
   seizures. Therefore, signaling pathways involved in hippocampal neuron
   damage are considered potential targets for alleviating seizures or
   delaying the development of epilepsy.[40]^3 Iron is an important trace
   element that is essential for maintaining neuronal function and
   metabolic processes in the CNS.[41]^4 However, when iron, especially
   ferrous iron, accumulates excessively in the CNS, it can lead to
   oxidative stress and cell death. Recent studies have shown that iron
   accumulation is strongly associated with the occurrence and development
   of a variety of neurological disorders, including epilepsy.[42]^5

   Metabolomics, which focuses on small molecule metabolites (<1000 Da)
   present in biological systems, is a powerful tool used to
   comprehensively evaluate the dynamic changes of metabolites after
   endogenous or exogenous interference. Metabolomics analysis can be used
   to understand the metabolic characteristics of an organism in
   physiological or pathological states, as well as the interaction and
   regulatory network between metabolites.[43]^6 Metabolomics can provide
   comprehensive metabolic information and provide important support for
   understanding the physiological status, disease mechanism and drug
   response of organisms, and is widely used in medicine, medicine and
   agriculture.[44]^7^,[45]^8 However, metabolomics approaches still have
   limitations. They only focus on the terminal variation caused by
   interferences, but ignore the endogenous mechanisms of metabolite
   activity.[46]^9 DisGeNET database is a comprehensive gene-disease
   association (GDA) database that provides extensive information on
   variants and genes associated with human disease. Recently, DisGeNET
   database has been widely applied to investigate the potential target
   genes of various diseases.[47]^10 Therefore, the integration of
   metabolomics and DisGeNET database can help overcome the limitations
   due to the lack of experimental basis for the former and the lack of
   molecular mechanism explanations for the latter, and will contribute to
   a better understanding of the pathogenesis of epilepsy. In this study,
   an untargeted metabolomics approach was employed to comprehensively and
   unbiasedly detect differential metabolites in the peripheral blood of
   epilepsy patients compared to the normal population. The results
   revealed elevated levels of free fatty acids, decreased levels of
   glutamine, increased levels of bilirubin, and enhanced iron metabolism
   in the peripheral blood of epilepsy patients. Further functional
   enrichment analysis of the differential metabolites highlighted the
   significant activation of hypoxia-inducible factor (HIF) -1 pathway in
   the pathological state of epilepsy. Additionally, we also used the
   DisGeNET database to collect target genes related to human epilepsy for
   functional enrichment analysis, and the analysis results further
   support that HIF-1 is an important disease target for human epilepsy.

   HIF-1 is a crucial transcription factor involved in regulating cellular
   adaptation to low oxygen environments. Comprising two subunits, HIF-1α
   and HIF-1β, HIF-1α serves as the primary functional subunit.[48]^11
   HIF-1α orchestrates the transcriptional activation of over 60 genes,
   including heme oxygenase (HO)-1.[49]^12 HO-1 is a stress-inducing
   enzyme that metabolizes heme into biliverdin (rapidly converted to
   bilirubin), carbon monoxide, and ferrous iron (which induces an
   increase in ferritin).[50]^12 Recent research has highlighted the
   significant role of HIF-1α/HO-1-mediated ferroptosis in mouse diabetic
   nephropathy[51]^13 and infertility.[52]^14 However, the interaction
   between HIF-1α and HO-1, along with their neurotoxic mechanisms
   associated with hippocampal neuronal ferroptosis in epilepsy, remains
   unclear. Therefore, we propose a scientific hypothesis that the
   HIF-1α/HO-1 pathway promotes the occurrence and development of epilepsy
   by mediating hippocampal neuronal ferroptosis.

   This study combines metabolomics with human disease public databases to
   uncover the potential role of the HIF-1α/HO-1 pathway in hippocampal
   neuronal ferroptosis during epilepsy. Subsequently, we will employ
   various experimental approaches in vivo, including molecular biology,
   cell biology, and behavioral studies, to validate the findings from the
   metabolomics research and further elucidate the critical role of the
   HIF-1α/HO-1 pathway in the pathogenesis of hippocampal neuronal
   ferroptosis-related epilepsy. This research holds promise for enhancing
   our understanding of the mechanisms underlying epilepsy and providing a
   theoretical basis for future development of novel therapeutic
   strategies and drugs.

Results

Sample characteristic

   The study consisted of a total of eighty-two subjects: sixty-two
   epilepsy patients categorized into infrequent seizure (IS) (n = 31) and
   frequent seizures (FS) (n = 31) based on the frequency of their
   seizures. Additionally, 20 healthy volunteers served as NC (n = 20).
   There were no significant differences in age and sex among the three
   groups. There were no statistically significant differences in age of
   first onset or seizure type between the two groups. ([53]Table 1).

Table 1.

   Demographic and clinical characteristics of enrolled subjects
   Demographic and clinical characteristics NC
   （n = 20） IS
   （n = 31） FS
   （n = 31） p-value
   Age (mean ± SD) 28.05 ± 4.17 31.39 ± 13.77 35.10 ± 18.06 0.218[54]^a
   Male/Female Ratio 10/10 18/13 18/13 0.819[55]^b
   Age at onset (mean ± SD) 24.55 ± 16.13 25.97 ± 16.84 0.736[56]^c
   Seizure type (n)
    Focal 11 6 0.376[57]^b
    Generalized 14 18 0.376[58]^b
    Mixed 6 6 0.376[59]^b
   [60]Open in a new tab

   NC, Normal control; IS, Infrequent seizure; FS, Frequent seizures.

   n, number of samples; SD, Standard deviation.
   ^a

   p value calculated using ANOVA test.
   ^b

   p value calculated using chi-square test.
   ^c

   p value calculated using T-test.

Untargeted metabolomic differential metabolite analysis

   Firstly, we constructed an OPLS-DA model based on the untargeted
   metabolomics results of epilepsy patients and control subjects
   ([61]Figure 1A). The results revealed significant inter-group
   differences and good intra-group reproducibility in the differential
   metabolites. The differentially expressed metabolites, identified based
   on the criteria of VIP score > 1.0 and p < 0.05, have the potential to
   serve as biomarkers for diagnosing epilepsy. A total of 153
   significantly altered metabolites were identified in the positive ion
   mode, and 76 differential metabolites were identified in the negative
   ion mode when comparing the NC, IS, and FS group ([62]Table S1). We can
   find from the HMDB classification ring maps that the differential
   metabolites of epilepsy patients are mostly concentrated in lipids and
   small lipid molecules (30.68%), organic heterocyclic compounds
   (17.05%), organic acids and derivatives (11.36%) in positive ion mode.
   While in the negative ion mode, differential metabolites were mainly
   concentrated in organic heterocyclic compounds (22.86%), organic acids
   and derivatives (20%), lipids and small lipid molecules (20%)
   ([63]Figure 1B). In terms of fatty acids metabolism, compared to the NC
   group, the epilepsy groups showed elevated levels of palmitic acid,
   oleic acid, linoleic acid, and myristic acid. Furthermore, the levels
   of these metabolites increased with the frequency of epileptic
   seizures. It is worth noting that the difference in palmitic acid
   levels between the FS group and the IS group was statistically
   significant (p < 0.05) ([64]Figure 1C). Regarding amino acid
   metabolism, glutamine and arginine levels in peripheral blood of
   epileptic patients were lower than those of NC group. As the frequency
   of seizures increased, glutamine and arginine levels further decreased.
   Similarly, there was a statistically significant difference in
   glutamine levels between the FS and IS groups (p < 0.05)
   ([65]Figure 1D). Additionally, the epilepsy groups showed elevated
   levels of bilirubin compared to the control group. Moreover, as the
   frequency of epileptic seizures increased, there was a trend of further
   elevation in bilirubin levels ([66]Figure 1E). HO-1 can metabolize heme
   into bilirubin, carbon monoxide, and ferrous iron. Heme iron represents
   the main source of iron in the body, accounting for approximately 67%
   of total body iron.[67]^15 Enhanced heme metabolism can lead to changes
   in iron content in the body. Therefore, the alterations in peripheral
   blood metabolites in epilepsy patients suggest the presence of abnormal
   iron metabolism and oxidative stress, indicating a potential role of
   ferroptosis in the development of epilepsy.

Figure 1.

   [68]Figure 1
   [69]Open in a new tab

   Untargeted metabolomic differential metabolites analysis among NC, IS,
   and FS groups

   (A) OPLS-DA score in positive and negative ion mode.

   (B) HMDB classification ring maps in positive and negative ion mode.

   (C‒E) Relative abundance of several representative differential
   metabolites. One-way ANOVA was used to compare three groups, followed
   by conducting pairwise comparisons between groups using Welch’s T-test.
   The data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01,
   ∗∗∗p < 0.001, and ns for no significance.

KEGG pathway enrichment analysis of differential metabolites

   We performed Kyoto encyclopedia of genes and genomes (KEGG) pathway
   enrichment analysis on the differential metabolites. In the positive
   ion mode, no significantly enriched pathways were identified among the
   differential metabolites. However, in the negative ion mode, the HIF-1
   signaling pathway, fatty acid biosynthesis signaling pathway, and
   unsaturated fatty acid biosynthesis signaling pathway were
   significantly enriched (Adjusted p < 0.05) ([70]Figure 2). This is
   consistent with previous knowledge, as ferroptosis involves excessive
   reactive oxygen species (ROS) production and accumulation through
   iron-catalyzed lipid peroxidation, leading to oxidative cell death.
   Furthermore, previous studies have indicated that the increase in
   unsaturated fatty acids or free fatty acids provides substrates for
   iron-mediated cell death. Therefore, our functional enrichment results
   align with existing understanding in the field.

Figure 2.

   [71]Figure 2
   [72]Open in a new tab

   KEGG pathway enrichment analysis of differential metabolites in
   positive and negative ion mode

Screening and functional enrichment analysis of epilepsy-associated genes
based on DisGeNET database

   DisGeNET database is a comprehensive GDA database that provides a
   wealth of information on variations and genes associated with human
   diseases. Using “epilepsy” as a keyword, we collected 1215 disease
   target genes related to epilepsy from the DisGeNET database
   ([73]Table S2). Additionally, we constructed a protein-protein
   interaction (PPI) network using the String database and visualized the
   PPI network using Cytoscape 3.7.2 software. By applying selection
   criteria (degree cutoff = 2, node score cutoff = 0.2, K-core = 2, Max.
   Depth = 100), we identified a core network and 73 hub target genes
   ([74]Figure 3A). Next, we conducted functional and pathway enrichment
   analysis for the 73 hub target genes associated with epilepsy. The Gene
   Ontology (GO) analysis yielded a total of 540 biological processes (BP)
   terms, 62 cellular components (CC) terms, and 68 molecular functions
   (MF) terms ([75]Table S3). The GO enrichment results were visualized as
   a histogram based on the p value (TOP30). In terms of BP, the
   predominant themes encompassed positive regulation of gene expression,
   aging, positive regulation of apoptosis, inflammatory response, and NO
   synthesis. CC analysis primarily highlighted extracellular regions,
   macromolecular complexes, cytoplasm, plasma membrane, and transcription
   factor complexes. MF analysis mainly focused on cytokine binding,
   enzyme binding, transcription factor binding, protein binding, and
   protein kinase binding ([76]Figure 3B). Furthermore, the KEGG analysis
   identified 153 significantly enriched signaling pathways
   ([77]Table S4). Based on the p value (TOP30), we visualized The KEGG
   enrichment results using a bubble plot. Notably, the HIF-1 signaling
   pathway also exhibited significant enrichment (p < 0.0001)
   ([78]Figure 3C). Key molecules enriched in the HIF-1 pathway include
   HIF-1α, NOS3, STAT3, IGF1, EGFR, MTOR, RELA, INS, VEGFA, IL6, IFNG,
   AKT1, MAPK1, GAPDH, and TLR4. These findings further support the notion
   that the HIF-1α signaling pathway may play a crucial role in
   ferroptosis and the development of epilepsy.

Figure 3.

   [79]Figure 3
   [80]Open in a new tab

   Screening and functional enrichment analysis of epilepsy-associated
   genes based on DisGeNET database

   (A) PPI network for epilepsy-associated genes.

   (B) GO analysis.

   (C) KEGG analysis.

Validation of hippocampal neurons ferroptosis in PTZ kindling epilepsy mice
models

   We successfully established the PTZ kindling epilepsy mice models
   according to the previous stable model establishment method. The flow
   schematic depiction of study design is showed in [81]Figure 4A.
   Compared with CON group, the epileptic seizure score of mice in the PTZ
   group was significantly higher, in addition, the incidence and duration
   of epilepsy of mice in the PTZ group met the criteria for successful
   modeling ([82]Figure 4B, [83]Table 2). We observed the structure of
   CA1, CA3, and DG regions in the hippocampus of PTZ and CON group by HE
   staining. In the control group, the neurons in CA1, CA3, and DG regions
   of the hippocampus were arranged neatly, the structure was complete,
   the chromatin distribution was uniform, the nucleolus was clear, and no
   obvious neuron loss was observed. However, in the PTZ group, the
   neurons in CA1 and CA3 regions of the hippocampus were disordered, some
   neurons were missing, the number was reduced, the shape of neurons was
   abnormal, the regular round and oval gradually shrank and the color was
   deepened, the nucleus was shrunk, the large nucleolus disappeared, and
   the neurons were vacuolized. In addition, the damage in DG region was
   not obvious ([84]Figure 4C). We evaluated the subcellular structure of
   hippocampal neurons in two groups of mice by TEM. The results showed
   that the nucleus size of hippocampal neurons in both groups was normal
   and there was no chromatin condensation. However, compared with the CON
   group, the mitochondria of hippocampal neurons in the PTZ group became
   smaller, the membrane density increased, and the mitochondrial ridge
   decreased or disappeared, which was consistent with the mitochondrial
   change characteristic of ferroptosis reported in the literature
   ([85]Figure 4D). In contrast to the control group, the mice in PTZ
   group displayed modified mitochondrial structure, showing significant
   reductions in both mitochondrial length and area ([86]Figure 4E). This
   suggests an abnormality in mitochondrial dynamics. Then, we detected
   the expression of ferroptosis marker gene prostaglandin-endoperoxide
   synthase 2/cyclooxygenase 2 (PTGS2/COX2) in the two groups of mice by
   RT-qPCR and western blot. We found that at both transcription and
   translation levels, PTGS2/COX-2 expression were significantly elevated
   in mice in the PTZ group ([87]Figures 4F and 4G). In addition, Fe^2+
   accumulation in the hippocampal tissues of two groups of mice was also
   detected by iron content detection kit. We found that compared with the
   CON group, the relative Fe^2+ content in the hippocampus of mice in the
   PTZ group was significantly increased ([88]Figure 4H). As for ferritin
   (FE), because FE has heavy chain (FTH-1) and light chain (FTL), FTH1
   mainly exists in neurons, while FTL is mainly expressed in
   microglia.[89]^16 Therefore, we detected the expression levels of FTH-1
   and transferrin (TRF) in the hippocampus of two groups of mice.
   According to the results of RT-qPCR, transcription levels of FTH1 and
   TRF in the hippocampus of mice in the PTZ group were significantly
   increased compared with the CON group ([90]Figure 4I). ELISA results
   showed that the contents of FE and TRF in the hippocampus of mice in
   the PTZ injection group were significantly increased ([91]Figure 4J).
   Western blot results also showed that the expressions of FTH-1 and TRF
   proteins in the hippocampus of mice in the PTZ group were significantly
   increased ([92]Figure 4K).

Figure 4.

   [93]Figure 4
   [94]Open in a new tab

   Validation of hippocampal neurons ferroptosis in PTZ kindling epilepsy
   mice models

   (A) Flow schematic depiction of study design.

   (B) Statistical chart of seizure score changes after each PTZ injection
   in epileptic model group.

   (C) HE staining of CA1, CA3, and DG regions in hippocampus of PTZ and
   CON group.

   (D) TEM results of hippocampus in PTZ and CON group with 5.0 μm,
   2.0 μm, and 500 nm scale.

   (E) The quantification of mitochondrial length and area in hippocampus.

   (F and G) RT-qPCR and western blot results for PTGS2/COX-2 in
   hippocampus.

   (H) Relative Fe2+ accumulation in hippocampus.

   (I‒K) RT-qPCR, ELISA and western blot results for FTH-1 and TRF in
   hippocampus. Unpaired two-tailed Student’s t-test was used to compare
   two groups. The data are presented as mean ± SEM. ∗p < 0.05,
   ∗∗p < 0.01, and ∗∗∗p < 0.001.

Table 2.

   Behavioral statistical results of epileptic seizure in different
   treatment groups
   Group N Percentage of kindled
   (%) Latency time of kindled
   (d) Last seizure scale ≥4
     __________________________________________________________________

   Latency time (s) Duration time (s)
   CON 30 0 – – –
   PTZ 30 83.3 16.48 ± 2.19 208.36 ± 15.78 62.96 ± 6.08
   [95]Open in a new tab

Validation of excessive oxidative stress of hippocampus in PTZ kindling
epilepsy mice models

   Aiming to observe the antioxidant ability of hippocampal tissue after
   epilepsy, we firstly measured the activities of SOD, CAT, and GSH-Px.
   Compared with CON group, the activities of these three antioxidant
   enzymes were significantly decreased in PTZ group ([96]Figure 5A).
   GPX4, which has a broader substrate preference, is the only enzyme
   described that can directly reduce complex phospholipid hydroperoxides.
   Therefore, we measured GPX4 expression levels in hippocampal tissues of
   two groups of mice from transcriptional and protein levels,
   respectively. The results showed that GPX4 mRNA and protein expression
   in hippocampus of mice in PTZ group were significantly lower than those
   in CON group ([97]Figures 5B and 5C). We also used ROS frozen section
   staining method to detect ROS content in hippocampal CA1, CA3, and DG
   regions of mice in CON group and PTZ group, respectively. The results
   showed that ROS content in hippocampal CA1, CA3, and DG regions of mice
   in PTZ group was significantly increased compared with the CON group
   ([98]Figure 5D). MDA is a natural product of lipid oxidation, and the
   content of MDA reflects the level of lipid oxidation. Therefore, the
   determination of MDA is widely used as an index to react lipid
   oxidation. We found that compared with the CON group, MDA content in
   the hippocampus of mice in the PTZ group was significantly increased
   ([99]Figure 5E), indicating lipid peroxidation in the hippocampus of
   mice in the PTZ group. In conclusion, the mice in PTZ group showed
   typical excessive oxidative stress and ferroptosis in the hippocampus.

Figure 5.

   [100]Figure 5
   [101]Open in a new tab

   Validation of excessive oxidative stress of hippocampus in PTZ kindling
   epilepsy mice models

   (A) Relative activity of antioxidant enzymes in hippocampus, including
   SOD, CAT, and GSH-Px.

   (B and C) western blot and RT-qPCR results for GPX4 in hippocampus.

   (D) ROS staining of CA1, CA3, and DG regions in hippocampus.

   (E) Relative MDA content hippocampus. Unpaired two-tailed Student’s
   t-test was used to compare two groups. The data are presented as mean ±
   SEM. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001.

Validation of the activation of HIF-1α/HO-1 pathway in hippocampus in PTZ
kindling epilepsy mice models

   To verify that HIF-1α/HO-1 pathway is activated in epilepsy, we
   performed RT-qPCR and western blot assays to detect hippocampal tissues
   of PTZ and CON mice at transcriptional and translational levels.
   RT-qPCR results showed the mRNA levels of HIF-1α and HO-1 in the
   hippocampus of mice in the PTZ group were significantly increased
   compared with the CON group ([102]Figure 6A). Western blot results also
   showed that the protein expressions of HIF-1α and HO-1 in the
   hippocampus of mice in the PTZ group were significantly increased
   ([103]Figure 6B). These results suggest that HIF-1α/HO-1 pathway is
   significantly activated in the hippocampus of PTZ kindling epileptic
   mice.

Figure 6.

   [104]Figure 6
   [105]Open in a new tab

   Validation of the activation of HIF-1α/HO-1 pathway in hippocampus in
   PTZ kindling epilepsy mice models

   (A and B) RT-qPCR and western blot results for HIF-1α and HO-1 in
   hippocampus. Unpaired two-tailed Student’s t-test was used to compare
   two groups. The data are presented as mean ± SEM. ∗p < 0.05,
   ∗∗p < 0.01, and ∗∗∗p < 0.001.

Discussion

   Epilepsy is a chronic brain disease caused by highly synchronized
   abnormal discharge of brain neurons, which has the characteristics of
   recurrence, seizure, transitory, and stereotyped.[106]^17 As estimated,
   the current prevalence of epilepsy has achieved 1% in the general
   population, with 80% of people with epilepsy living in low- and
   middle-income countries, causing a substantial financial
   burden.[107]^18 The pathogenesis of epilepsy remains elusive, with two
   plausible mechanisms currently being postulated: the ion channel theory
   and the abnormal network theory. According to the ion channel theory,
   epilepsy arises from genetic abnormalities that result in dysregulation
   of excitatory or inhibitory neurotransmitters, leading to aberrant
   opening of ion channels and subsequent neuronal discharge, culminating
   in epileptic seizures. While the abnormal network theory proposes that
   epilepsy originates from neuronal death, wherein the remaining neurons,
   along with newly generated neurons and proliferating glial cells, form
   novel networks. When these networks facilitate the formation and
   propagation of epilepsy, recurrent seizure activity is triggered,
   perpetuating the cycle of neuronal death and establishing a malignant
   loop.[108]^19^,[109]^20^,[110]^21 Ion channels exert rapid effects and
   are the primary targets of the antiepileptic drugs commonly used in
   clinical practice. Although these medications effectively control
   seizures, their long-term use may lead to adverse reactions such as
   increased blood drug concentrations, drug resistance, and neurologic
   and psychiatric impairments. Unfortunately, there is currently no ideal
   pharmacological intervention that regulates neuronal death to improve
   epilepsy. Therefore, further exploration of the role and molecular
   biology mechanisms of neuronal damage in epilepsy holds promise for
   identifying novel drug targets that could revolutionize the treatment
   of epilepsy.[111]^18

   Untargeted metabolomics, which involves comprehensive analysis of
   metabolic products in an organism under specific conditions, holds
   significant importance in unraveling the pathogenesis of epilepsy and
   identifying novel therapeutic targets.[112]^22 Through conducting
   untargeted metabolomics studies in patients with epilepsy, we can
   uncover metabolic abnormalities associated with the condition, thereby
   deepening our understanding of the pathogenesis of epilepsy. In this
   study, we performed comprehensive metabolomic analysis on peripheral
   blood samples collected from both epilepsy patients and healthy
   individuals, aiming to provide insights into the identification of
   molecular and biological targets for epilepsy treatment. Blood samples
   are widely used in clinical research on various diseases, including
   epilepsy. They are often employed as a substitute for brain tissue due
   to their ease of accessibility. Furthermore, blood samples can
   effectively reflect the overall physiological changes occurring before
   and after the onset of the disease.[113]^22^,[114]^23 Our research
   findings demonstrated that compared to the NC group, there is an
   elevated expression of palmitic acid, oleic acid, linoleic acid, and
   myristic acid in the epilepsy group. Palmitic acid has been shown to
   play a role in inflammation. A study by Wen et al.[115]^24 indicated
   that palmitic acid can activate NOD-like receptor thermal protein
   domain associated protein 3, which in turn enhances the generation of
   ROS in macrophages, thereby attenuating the adenosine
   5‘-monophosphate-activated protein kinase (AMPK) signal. The AMPK
   signal serves as a negative regulator of ROS generation and
   inflammation, thus further increasing ROS production, ultimately
   resulting in a vicious cycle. Additionally, palmitic acid can also
   induce ferroptosis in osteoblasts through the activation of the
   methyltransferase like 3/apoptosis signal-regulating kinase 1-protein
   38 signaling pathway.[116]^25 The relationship between oleic acid and
   oxidative stress has been extensively studied. Utilizing oleic acid to
   induce the generation of ROS in HepG2 liver cancer cells is a
   well-established oxidative stress model.[117]^26 Furthermore, research
   has indicated that oleic acid can promote ROS production in mouse
   dermal fibroblasts, leading to lipid peroxidation.[118]^27 High
   concentrations of linoleic acid (500 μg/L) exhibit significant
   inhibitory effects on algal growth, and studies have shown that the
   application of high concentrations of linoleic acid to algae results in
   a significant increase in ROS, potentially promoting
   ferroptosis.[119]^28 Moreover, research has demonstrated a positive
   correlation between the content of myristic acid and ROS levels in
   tumor cells.[120]^29 Comparing epilepsy patients to the healthy
   individuals, there are significant differences in amino acid metabolism
   as well. Our study reveals a notable decrease in the levels of
   glutamine and arginine in the epilepsy patients. Looking at the body as
   a whole, the synthesis of arginine mainly occurs in the epithelial
   cells of the small intestine. These cells utilize glutamine and
   glutamate to produce citrulline, which is then converted to arginine
   with the assistance of renal tubular cells in the kidneys. Therefore,
   the decrease in arginine can be partly attributed to the decline in
   glutamine. Glutamine participates in the synthesis of glutathione,
   enhancing the body’s antioxidant capacity. It serves as an important
   reducer or antioxidant. Under normal conditions, more than 60% of the
   free amino acids in the human body appear in the form of glutamine, and
   the body can produce excess glutamine to meet its needs. However, in
   many disease states, there is a significant decrease in glutamine,
   leading to a weakening of the body’s reducing capacity.[121]^30 We also
   observed an elevation in bilirubin levels in epilepsy patients.
   Bilirubin is generated through the breakdown of heme by HO-1,
   accompanied by the production of carbon monoxide and ferrous iron. Heme
   iron serves as the primary source of iron in the body, accounting for
   approximately 67% of total body iron. Therefore, an enhanced metabolism
   of heme, as indicated by increased bilirubin, naturally leads to
   changes in iron levels within the body.[122]^15 Our subsequent in vivo
   experiments in PTZ kindling epileptic mice further demonstrated a
   significant increase in HO-1 expression and elevated Fe2+ levels in the
   hippocampal tissue. The upregulation of HO-1 leads to Fe^2+ overload,
   intensifies the Fenton reaction, accumulates ROS, and ultimately
   contributes to the occurrence of ferroptosis.[123]^31^,[124]^32 Studies
   have reported that FE and TRF are also regulated by HO-1. The
   upregulation of HO-1 leads to increased levels of ferritin and
   transferrin, further enhancing iron accumulation.[125]^33 This finding
   has been confirmed in our subsequent animal experiments.

   Cell death of hippocampal neurons plays a crucial role in the
   pathogenesis of epilepsy. Currently, research on hippocampal neuron
   cell death and its association with epilepsy mainly focuses on two
   aspects: apoptosis and necrotic cell death. In epilepsy, apoptosis may
   be triggered by abnormal neuronal electrical activity.[126]^34 Studies
   have found a significant increase in the extent of neuronal apoptosis
   during epileptic seizures.[127]^35^,[128]^36 Necrotic cell death is an
   irregular form of cell death that is often associated with factors such
   as acute ischemia, hypoxia, and inflammatory responses.[129]^37 In
   epilepsy, ischemia and hypoxia are common pathological and
   physiological alterations, which can contribute to neurodegeneration
   and necrotic cell death of neurons.[130]^38 Ferroptosis, initially
   described by Dixon et al. in 2012, is an iron-dependent regulated cell
   death pathway that primarily relies on iron-mediated oxidative damage
   and subsequent membrane disruption, particularly mitochondrial membrane
   damage.[131]^39 Ferroptosis of hippocampal neurons plays a crucial role
   in the pathogenesis of epilepsy; however, the precise molecular
   mechanisms underlying this process require further
   investigation.[132]^40 The essence of ferroptosis is the iron-catalyzed
   accumulation of lipid ROS, which reacts with membrane polyunsaturated
   fatty acids, resulting in the accumulation of lipid peroxides and
   subsequent oxidative cell death.[133]^40 Ferroptosis differs
   morphologically and biologically from other types of cell death.
   Morphologically, during ferroptosis, mitochondria appear smaller with
   increased membrane density, reduced or vanished cristae, and swelling
   or even rupture of the outer mitochondrial membrane, while the cell
   nucleus maintains a normal size without chromatin condensation.[134]^41
   Biologically, ferroptosis is characterized by iron accumulation and
   redox imbalance resulting from multiple signaling pathways, ultimately
   leading to lipid peroxidation.[135]^42 A clinical study conducted by
   Ikeda et al.[136]^43 revealed that compared to the healthy control
   group, patients with epilepsy had significantly higher transferrin
   saturation in their peripheral blood, suggesting that iron overload in
   the body may be an important risk factor for epilepsy. Another clinical
   study in pediatric epilepsy patients found elevated markers of
   ferroptosis in their peripheral blood, including increased levels of
   the lipid peroxidation byproduct 4-hydroxy-2-nonenal, partial
   inactivation of GPX4, and a significant decrease in GSH levels.[137]^44
   Our animal experimental research further supports these findings by
   demonstrating significant ferroptosis and oxidative stress in
   hippocampal neurons of the epilepsy animal model, thereby promoting the
   development of epilepsy. These findings may provide novel targets and
   strategies for the treatment of epilepsy.

   HIF-1, a transcription factor belonging to the HIF-1 family, consists
   of two subunits, namely HIF-1α and HIF-1β. HIF-1α serves as the primary
   subunit responsible for functional modulation. HIF-1α exhibits the
   capacity to regulate the transcriptional activation of more than 60
   genes. Its principal role lies in gene expression regulation under
   hypoxic conditions, facilitating cellular adaptation to oxygen
   deprivation. The stability and activity of HIF-1α are modulated by
   various factors, encompassing oxygen levels, intracellular iron
   content, and activation of other signaling pathways.[138]^11 In recent
   years, it has been discovered that in addition to its role in hypoxia
   adaptation, HIF-1 can also contribute to iron-mediated cell death in
   certain disease states, such as diabetic nephropathy and malignant
   mesothelioma.[139]^13^,[140]^45 HO-1, one of the target genes regulated
   by HIF-1α, is a stress-inducible enzyme. Studies conducted by Feng
   et al.[141]^13 have demonstrated that HIF-1α/HO-1-mediated ferroptosis
   plays a significant role in the pathogenesis of diabetic nephropathy.
   Furthermore, Wu et al.[142]^14 have discovered that the environmental
   pollutant di-2-ethylhexyl phthalate can induce ferroptosis in mouse
   testes and lead to infertility through the mediation of the HIF-1α/HO-1
   signaling pathway. In addition, HIF-1α/HO-1 pathway was also confirmed
   to promote ferroptosis in the lungs induced by polystyrene
   nano-plastics.[143]^46 Hippocampal ferroptosis is a mechanism that
   triggers neuronal damage and epileptic seizures. However, whether the
   HIF-1α/HO-1 signaling pathway can mediate hippocampal ferroptosis and
   contribute to the occurrence and development of epilepsy remains
   unclear. This study aims to investigate whether the HIF-1α/HO-1 pathway
   can promote epilepsy by mediating hippocampal ferroptosis. Firstly, the
   plasma untargeted metabolomic analysis in epileptic patients suggests
   the occurrence of ferroptosis and activation of the HIF-1α/HO-1 pathway
   following epileptic seizure. Additionally, analysis of epilepsy disease
   target genes based on the DisGeNET human disease database further
   supports HIF-1α as an important disease target gene in human epilepsy.
   In the subsequent animal experimental validation, our results
   demonstrate a significant increase in the expression of HIF-1α and HO-1
   in PTZ kindling chronic mice elilepsy models. These findings indicate a
   potential promoting role of HIF-1α/HO-1 in epileptic seizures. Further
   investigation reveals that the activation of the HIF-1α/HO-1 pathway
   may positive correlated with abnormal iron metabolism and increased
   accumulation of Fe^2+ within hippocampal neurons. Excessive
   accumulation of Fe^2+ has been shown to induce oxidative stress and
   cellular damage, thereby triggering ferroptosis in hippocampal neurons.
   Furthermore, the activation of the HIF-1α/HO-1 pathway may reduce the
   activity of antioxidant enzymes in hippocampal tissue, further
   exacerbating oxidative damage in hippocampal neurons and the
   progression of epilepsy. In conclusion, our research findings indicated
   that the HIF-1α/HO-1 pathway might promote the occurrence and
   development of epilepsy by mediating hippocampal ferroptosis. This
   discovery provides novel targets and strategies for the treatment of
   epilepsy. Inhibiting the activation of the HIF-1α/HO-1 pathway and
   regulating iron metabolism may offer potential approaches for epilepsy
   treatment. Additionally, enhancing the activity of antioxidant enzymes
   in hippocampal tissue and boosting the antioxidative ability of
   hippocampal neurons may have a positive impact on epilepsy therapy.

Limitations of the study

   However, there are several limitations in this study. First of all, the
   study predominantly establishes correlations between the HIF-1α/HO-1
   pathway activation, ferroptosis markers, and epilepsy. However,
   establishing a causal relationship requires further mechanistic
   studies, such as targeted genetic or pharmacological interventions to
   modulate the pathway and observe resultant changes in ferroptosis and
   epileptic phenotypes. Moreover, more direct and comprehensive brain
   tissue analysis in human subjects would strengthen the link between the
   HIF-1α/HO-1 pathway and epileptogenesis. Thirdly, dihydroethidium used
   in the present study is superoxide-anion specific dye, it cannot able
   to detect other kinds of ROS, and we intend to include assays that can
   specifically detect hydrogen peroxide, hydroxyl radicals, and other ROS
   in the future work.

STAR★Methods

Key resources table

   REAGENT or RESOURCE SOURCE IDENTIFIER
   Antibodies
     __________________________________________________________________

   Rabbit anti-HO-1 monoclonal antibody Abcam Cat# AB68477;RRID:
   [144]AB_11156457
   Rabbit polyclonal anti-HIF-1α antibody Abmart Cat# P50517R1
   Rabbit anti-GPX4 monoclonal antibody Abmart Cat# T56959F
   Rabbit anti-Transferrin monoclonal antibody Abmart Cat# T40111F
   Rabbit anti-Ferritin monoclonal antibody Abmart Cat# T55648F
   Rabbit anti-Cox2 monoclonal antibody Abmart Cat# T58852F
   Mouse anti-β-actin antibody Sigma-Aldrich Cat# A1978; RRID:
   [145]AB_476692
     __________________________________________________________________

   Chemicals, peptides, and recombinant proteins
     __________________________________________________________________

   PTZ Sigma-Aldrich P6500
   TranScript One-Step gDNA Removal and cDNA Synthesis SuperMix TransGen
   Biotech AT-311-03
   SYBR Premix Ex Taq™ II Takara RR820A
   Iron Assay Kit Sigma-Aldrich MAK025
   SOD Assay kit Beyotime S0101S
   CAT Assay kit Beyotime S0051
   GPx Assay kit Beyotime S0056
   Ferritin ELISA Assay Kit Lai Er Bio-Tech LE-M2088
   Transferrin ELISA Assay Kit Lai Er Bio-Tech LE-M1662
   DHE Sigma-Aldrich D7008
   MDA Assay kit Beyotime S0131S
     __________________________________________________________________

   Experimental models: Organisms/strains
     __________________________________________________________________

   Male C57BL/6 mice Yisi Laboratory Animal Technology Co., LTD NA
     __________________________________________________________________

   Software and algorithms
     __________________________________________________________________

   Prism 8.0 GraphPad [146]https://www.graphpad.com/
   Adobe Illustrator CC 2018 Adobe
   [147]https://www.adobe.com/products/illustrator.html
   SPSS statistics IBM RRID:[148]SCR_002865
   Microsoft Excel Microsoft RRID:[149]SCR_016137
   ZEN 2012 ZEISS [150]https://www.zeiss.com.cn/
   Image J NIH [151]https://imagej.net
   R version 3.4.1 Vienna [152]https://www.r-project.org/
   [153]Open in a new tab

Resource availability

Lead contact

   Further information and requests for resources and reagents should be
   directed to and will be fulfilled by the lead contact Songyan Liu
   ([154]liu_sy@jlu.edu.cn).

Materials availability

   This study did not generate new unique reagents.

Experimental model and study participant details

Ethics statement

   The study was conducted in accordance with the ethical guidelines of
   the Declaration of Helsinki. The research protocol was approved by the
   Ethics Committee of China-Japan Union Hospital of Jilin University
   (Approval Number: 2023022701). Written informed consent was obtained
   from each patient's next of kin/participant. For animal experiments,
   the protocols were approved by the Institutional Animal Care and Use
   Committee of Northeast Normal University in China (Approval Number: AP
   20221008).

For human studies

   Sixty-two epilepsy patients from China-Japan Union Hospital of Jilin
   University, (Changchun, China) who met the inclusion and exclusion
   criteria were enrolled in the present study. The entirety of the
   enrolled participants belonged to the of Han Chinese ethnicity.
   Furthermore, [155]Table 1 displayed the subjects' age and gender
   details. The Institutional Review Committee of China-Japan Union
   Hospital of Jilin University approved the research protocol, The study
   was conducted in accordance with the ethical guidelines of the
   Declaration of Helsinki. Written informed consent was obtained from
   each patient's next of kin/participant.

For animal studies

   Eight-week-old male C57BL/6 mice were obtained from Yisi Laboratory
   Animal Technology Co., LTD (Changchun, China). Mice were provided ad
   libitum access to food and water. The mice were housed under a 12-hour
   light/dark cycle at 25 ± 2°C with a relative humidity of 50 ± 5%.

Method details

Experimental group and plasma samples collection

   The inclusion criteria of the present study included: 1. Patients
   diagnosed with various types of epilepsy according to the 2017
   International League against Epilepsy Diagnostic Criteria; 2. Age ≥16
   years and follow-up time ≥3 months; 3. Good mental state and ability to
   cooperate well with examination and treatment; 4. Informed consent of
   the patient or their guardian. The exclusion criteria as follows: 1.
   Febrile convulsion, syncope, hysteria and other non-epileptic diseases;
   2. Age <16 years, loss of follow-up, or follow-up time <3 months; 3.
   Serious systemic diseases including liver, kidney, and heart
   dysfunction, thyroid diseases, and blood system diseases; 4.
   Alcoholism, family history of mental illness, or inability to cooperate
   with the examination. Sixty-two epilepsy patients who met the inclusion
   and exclusion criteria were divided into the infrequent seizures group
   (IS, n=31) and the frequent seizures group (FS, n=31). Based on our
   clinical experience, we defined more than three seizures per month as
   frequent seizures and vice versa as infrequent seizures. Additionally,
   20 non-epileptic adults (healthy volunteers) who were recruited during
   the same period and matched for sex and age were selected as the normal
   control group (NC, n=20). To avoid the effect of diurnal variation,
   blood samples were collected in predetermined daytime hours between 9
   a.m. and 10:00 a.m. 5 mL of venous blood was collected from all
   subjects using a heparin anticoagulant tube after fasting on the
   predetermined time. The collected blood was temporarily stored at 4°C,
   and within 1 hour, it was sent to the specimen library of the
   China-Japan Union Hospital of Jilin University for plasma preparation
   and frozen at −80°C.

Plasma untargeted metabolomics analysis

Plasma samples preparation for metabolomics

   The plasma samples were prepared by thoroughly mixing 100 μL of the
   sample with 400 μL of cold methanol acetonitrile (v/v, 1:1) using
   vortexing. The mixture was sonicated for 1 h in ice baths, incubated at
   -20°C for 1 h, and centrifuged at 4°C for 20 min at a speed of 14,000
   g. The resulting supernatants were harvested and dried under vacuum for
   UHPLC-MS/MS analysis.

UHPLC-MS/MS analysis

   Metabolomics profiling was performed using a UPLC-ESI-Q-Orbitrap-MS
   system consisting of a UHPLC (Shimadzu Nexera X2 LC-30AD, Shimadzu,
   Japan) coupled with a Q-Exactive Plus mass spectrometer (Thermo
   Scientific, San Jose, USA). For hydrophilic interaction liquid
   chromatography separation, the samples were analyzed using a
   2.1 mm × 100 mm ACQUIY UPLC BEH Amide 1.7 μm column (Waters, Ireland).
   The flow rate was set at 0.5 mL/min, and the mobile phase consisted of
   two components: A) 25 mM ammonium acetate and 25 mM ammonium hydroxide
   in water, and B) 100% acetonitrile (ACN). The gradient elution started
   at 95% B for 1 min, followed by a linear decrease to 65% B over 7 min,
   then to 35% B in 2 min and maintained for 1 min, and finally increased
   to 95% B in 0.5 min, with a 2 min re-equilibration period. Both
   electrospray ionization (ESI) in positive ion mode and negative ion
   mode were used for MS data acquisition. The HESI source conditions were
   set as follows: Spray Voltage: 3.8 kV (+) and 3.2 kV (-); Capillary
   Temperature: 320 (±); Sheath Gas: 30 (±); Aux Gas: 5 (±); Probe Heater
   Temp: 350 (±); S-Lens RF Level: 50. In MS-only acquisition mode, the
   instrument was set to acquire data in the m/z range of 80-1200 Da. Full
   MS scans were performed at a resolution of 70,000 at m/z 200, and MS/MS
   scans were performed at a resolution of 17,500 at m/z 200. The maximum
   injection time was set to 100 ms for MS and 50 ms for MS/MS. The MS2
   isolation window was set to 2 m/z, and the normalized collision energy
   (stepped) was set to 27, 29, and 32 for fragmentation. Quality control
   (QC) samples were prepared by pooling representative aliquots from all
   samples and used for data normalization. Blank samples (75% ACN in
   water) and QC samples were injected every sixth sample during
   acquisition.

Data preprocessing and filtering

   The raw MS data from the NC, IS, and FS groups were processed using
   MS-DIAL for peak alignment, retention time correction, and peak area
   extraction. Metabolites were identified based on accurate mass (mass
   tolerance < 0.01 Da) and MS/MS data (mass tolerance < 0.02 Da), which
   were matched against the Human Metabolome Database (HMDB), MassBank,
   and other public databases, as well as a self-built metabolite standard
   library from Bioprofile Biotechnology (Shanghai, China). Only variables
   with more than 50% non-zero measurement values in at least one group
   were retained in the extracted-ion features.

Multivariate statistical analysis

   We utilized R version 3.4.1 (R Foundation for Statistical Computing,
   Vienna, Austria) for all multivariate data analyses and modeling. The
   data were mean-centered using Pareto scaling. We constructed models
   using Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA).
   Consequently, a score plot was generated to observe the separation
   among the NC, IS, and FS groups. The variable importance in projection
   (VIP) scores established by OPLS-DA were utilized to select the
   metabolites that exhibited the highest discriminatory power within each
   group. A VIP score greater than 1 was defined as the threshold for the
   current analysis. We conducted a one-way ANOVA test analysis to compare
   the abundance of metabolites among the three groups. A significance
   level of P-value < 0.05 was applied in the OPLS-DA analysis to identify
   differentially abundant metabolites. Subsequent to that, we generated
   HMDB classification ring maps using these differentially abundant
   metabolites as the basis. Additionally, we employed Welch's t-test
   analysis to compare the abundance of metabolites between the two
   groups.

Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis

   To identify disrupted biological pathways, we conducted KEGG pathway
   analysis using the KEGG database ([156]http://www.kegg.jp) with the
   differential metabolite data from the NC, IS, and FS groups. KEGG
   enrichment analyses were conducted using the Fisher's exact test, and
   multiple testing correction was applied using the FDR method. The
   enriched KEGG pathways showed nominal statistical significance at a
   p-value < 0.05 threshold.

Screening and functional enrichment analysis of epilepsy-associated genes
based on DisGeNET database

     * 1.
       Retrieve epilepsy-associated gene targets

   Enter "epilepsy" in the DisGeNET database
   ([157]https://www.disgenet.org/) and download the summary files of
   gene-disease associations.
     * 2.
       Protein-Protein Interaction (PPI) network analysis

   Epilepsy-associated target genes obtained from the DisGeNET database
   were imported into the String database
   ([158]https://cn.string-db.org/). The network was constructed by
   selecting "Multiple proteins" and "Homo sapiens," resulting in a
   network diagram consisting of 1122 nodes and 22654 edges. Utilizing the
   MCODE plugin (Cytoscape 3.7.2 plugin), the following settings were
   applied: "In whole network," "degree cutoff = 2," "Haircut" with "node
   score cutoff = 0.2, K-core = 2, Max. Depth = 100." Finally, the gene
   targets ranked first in terms of "K-core" were selected as the hub
   targets for epilepsy.
     * 3.
       Functional enrichment and pathway analysis

   The key target genes of epilepsy were initially uploaded into the
   online software, David ([159]https://david.ncifcrf.gov/), with the
   settings "H. sapiens" and "P < 0.05." Subsequently, Gene Ontology (GO)
   enrichment analysis was performed by selecting GO Biological Processes
   (BP), GO Molecular Functions (MF), and GO Cellular Components (CC). The
   GO enrichment results were visualized as a histogram based on the
   P-value (TOP30) for analysis. In addition, KEGG enrichment analysis was
   performed by selecting KEGG Pathway. The enrichment results of KEGG
   pathways were visualized as a bubble map based on the P-value (TOP30)
   for analysis.

Establishments of pentylenetetrazol (PTZ) kindling epilepsy models

   The mice were divided into two groups: (1) PTZ group: C57BL/6 mice were
   intraperitoneally injected with PTZ (P6500, Sigma-Aldrich, USA) at a
   dose of 35 mg/kg once every other day for a total of eleven
   injections.[160]^47 Mice showing more than three consecutive stage 4
   seizures were considered kindled. (2) Control group: Control C57BL/6
   mice were simultaneously injected with saline. Seizure behavior was
   video-monitored for 30 minutes after each injection. Seizure intensity
   was assessed using the modified Racing scale as follows: Stage 0, no
   response; Stage 1, ear and facial twitching; Stage 2, convulsive
   twitching axially through the body; Stage 3, myoclonic jerks and
   rearing; Stage 4, wild running and jumping; Stage 5, generalized
   tonic-clonic seizures; and Stage 6, death.[161]^48^,[162]^49

Hematoxylin-eosin (HE) staining

   For HE staining, mice were perfused with cold PBS followed by 4%
   paraformaldehyde (PFA) through transcardial perfusion. The mice brains
   were post-fixed overnight at 4°C and subsequently transferred to a 30%
   sucrose solution. Next, the brains were embedded in paraffin.
   Subsequently, tissue blocks were sectioned into 6 μm slices and stained
   with hematoxylin and eosin. Representative images were captured using
   an Olympus FSX100 microscope.

Transmission electron microscope (TEM)

   For TEM assay, mice were transcardially perfused with heparinized
   normal saline, followed by perfusion with a mixture of 2.5%
   glutaraldehyde and 1% PFA in 0.1 M phosphate buffer (PB, pH 7.4).
   Next,the hippocampus was dissected and subjected to TEM
   analysis.[163]^50 Briefly, the hippocampus was dissected from the
   entire brain. Next, the dissected hippocampal tissue samples were diced
   and fixed in 2.5% glutaraldehyde. Then, the hippocampus was
   transversely sectioned at 70 μm using a vibratome. The sections were
   then treated with 0.5% OsO4 (in 0.1 M PB) for 1 hour, dehydrated in a
   series of graded alcohols, flat-embedded in EMbed 812 (Electron
   Microscopy Sciences, Hatfield, PA), and cured for 48 hours at 60°C.
   Ultrathin sections of 60 nm were cut and mounted on Formvar-coated
   single-slot grids. The sections were stained with uranyl acetate and
   lead citrate, and examined under an electron microscope (HT-7800,
   Hitachi, Tokyo, Japan). As for quantification, the Image-Pro Plus 6.0
   software was used to select 5 mitochondria under each field of view and
   measure their length (μm) and area (μm^2) with a 2.5k 5μm scale as the
   standard. 5 fields of view are selected for each sample.

Real-time quantitative PCR (RT-qPCR)

   RT-qPCR assays were performed as a previously described
   protocol.[164]^51^,[165]^52 Specifically, total RNA was extracted from
   mouse hippocampus using TRIzol reagent (257401, Invitrogen, USA).
   Complementary DNA (cDNA) was synthesized using the TranScript One-Step
   gDNA Removal and cDNA Synthesis SuperMix (AT-311-03, TransGen Biotech,
   Beijing, China) according to the manufacturer’s protocol. RT-qPCR was
   performed using SYBR Premix Ex Taq™ II (RR820A, Takara) using the ABI
   StepOnePlus Real-Time PCR System. Each sample was analyzed in
   triplicate, and the mRNA levels were normalized to the β-actin mRNA
   level using the 2^−ΔΔCT method. Melting curves were automatically
   recorded and analyzed for each reaction. The primer sequences are
   listed in the below table.

   Primer sequences information
           Gene                  Primer sequences
   β-actin -F           5′- GGT GAA GGT CGG TGT GAA CG -3′
   Mouse-β-actin -R     5′- CTC GCT CCT GGA AGA TGG TG -3′
   Mouse-HIF-1α-F       5′- GAATGAAGTGCACCCTAACAAG -3′
   Mouse-HIF-1α-R       5′- GAGGAATGGGTTCACAAATCAG -3′
   Mouse-HO-1 -F        5′- CCGCTACCTGGGTGACCTCTC -3′
   Mouse-HO-1 -R        5′- GACGAAGTGACGCCATCTGTGAG -3′
   Mouse-GPX4-F         5′- CATGCCCGATATGCTGAGTGTGG -3′
   Mouse-GPX4-R         5′- TAGCACGGCAGGTCCTTCTCTATC -3′
   Mouse-FTH1-F         5′- TGCCATCAACCGCCAGATCAAC -3′
   Mouse-FTH1-R         5′- ATTCAGCCCGCTCTCCCAGTC -3′
   Mouse-Transferrin -F 5′- GGA CGC CAT GAC TTT GGA TG -3′
   Mouse-Transferrin -R 5′- GCC ATG ACA GGC ACT AGA CC -3′
   Mouse-PTGS2 -F       5′- TTCCAATCCATGTCAAAACCGT -3′
   Mouse-PTGS2 -R       5′- AGTCCGGGTACAGTCACACTT -3′
   [166]Open in a new tab

Protein extraction and western blotting

   Mouse hippocampal tissues were subjected to protein extraction using a
   modified RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM sodium chloride,
   1% NP-40, 0.25% sodium deoxycholate, and proteinase inhibitors).
   Protein concentrations were measured using a BCA Protein Assay Kit
   (P0012, Beyotime, Shanghai, China). Western blotting was conducted
   following a previously described protocol.[167]^51 Specifically, the
   protein samples were separated by SDS-PAGE and subsequently transferred
   onto PVDF membranes (Millipore). The membranes were then incubated
   overnight at 4°C with a specific primary antibody, followed by
   incubation with the corresponding HRP-conjugated secondary antibody.
   The chemiluminescent signals were visualized using ECL Prime Western
   Blot Detection reagent (GE Healthcare). The ECL signal was detected
   using a Tanon-5200 system (Tiangen (Beijing) Biotech Co., Ltd.) at the
   appropriate exposure time points. Band intensities were quantified
   using ImageJ software (NIH), and the values were normalized to the
   expression level of β-actin in each sample. The following antibodies
   were used in the study: Rabbit anti-Heme Oxygenase 1 monoclonal
   antibody (Abcam, AB68477, 1:10000); Rabbit polyclonal anti-HIF-1α
   antibody (Abmart, P50517R1, 1:1000); Rabbit anti-GPX4 monoclonal
   antibody (Abmart, T56959F, 1:1000); Rabbit anti-Transferrin monoclonal
   antibody (Abmart, T40111F, 1:1000); Rabbit anti-Ferritin monoclonal
   antibody (Abmart, T55648F, 1:1000); Rabbit anti-Cox2 monoclonal
   antibody (Abmart, T58852F, 1:1000); Mouse anti-β-actin antibody
   (Sigma-Aldrich, A1978, 1:5000); HRP-conjugated secondary antibodies
   against mouse or rabbit were obtained from Invitrogen (Carlsbad, USA).

ELISA

   The levels of Ferritin and Transferrin[168]^2 proteins in mouse
   hippocampus lysates were quantified using ELISA Assay Kits (LE-M2088
   and LE-M1662, Lai Er Bio-Tech, Anhui, China) following the
   manufacturer's instructions. The OD values were measured at 450 nm
   using a Milton Roy Spectronic 3000 Array spectrophotometer. The total
   protein concentration in each sample was determined using the Bradford
   method.

Ferrous iron (Fe^2+) assay

   The relative concentration of Fe^2+ in the mouse hippocampus tissue
   lysates was determined using an Iron Assay Kit (MAK025, Sigma-Aldrich,
   USA) following the manufacturer's instructions. OD values were measured
   at 593 nm using a Milton Roy Spectronic 3000 Array spectrophotometer. T
   The total protein content in each sample was determined using the
   Bradford method.

Antioxidant enzyme activity assay

   The relative activity of Superoxide Dismutase (SOD), catalase (CAT),
   and Glutathione Peroxidase (GPx) in the mouse hippocampus tissues was
   examined using commercial test kits (S0101S, S0051, and S0056,
   Beyotime, Shanghai, China) following the manufacturer's protocols. The
   absorbance was measured at 450 nm for SOD activity, 340 nm for GPx
   activity, and 520 nm for CAT activity using a Milton Roy Spectronic
   3000 Array spectrophotometer. The total protein content in each sample
   was measured using the Bradford method.

Reactive oxygen species (ROS) staining

   For ROS staining, mice were sacrificed, and the whole brain was
   immediately dissected, snap-frozen, and stored at -80°C. The brain
   samples were embedded in optimal cutting temperature compound (G6059,
   Wuhan servicebio technology CO., LTD, China) and frozen. Frozen
   sections of 8 μm thickness were prepared using a cryostat microtome
   (CRYOSTAR NX50; Thermo Fisher Technology (China) Co., LTD). Then, the
   brain slices were washed with distilled water, and were incubated in
   dihydroethidium (DHE) staining solution (D7008, Sigma-Aldrich, USA) for
   30 min. Subsequently, the slices were immersed in nucleus staining
   solution (G1012, Wuhan servicebio technology CO., LTD, China) for
   10 min, followed by washing with PBS (pH 7.4). Finally, the slices were
   used for measuring ROS content after being washed with PBS (pH 7.4)
   again. The ROS-stained sections were observed using a Zeiss LSM 780
   confocal microscope with ZEN 2012 software. After imaging, only the
   brightness, contrast, and color balance were adjusted. The ROS
   fluorescence intensity in different regions of the hippocampus was
   measured using ImageJ software (Image J, NIH, Bethesda, MD, USA).

Lipid peroxidation assay

   The relative concentration of malondialdehyde (MDA) in the mouse
   hippocampus lysates was determined using a Lipid Peroxidation MDA Assay
   Kit (S0131S, Beyotime, Shanghai, China) following the manufacturer's
   instructions. OD values were measured at 532 nm using a Milton Roy
   Spectronic 3000 Array spectrophotometer. The total protein content in
   each sample was measured using the Bradford method.

Quantification and statistical analysis

   The statistical analysis was performed using SPSS 22.0 software. After
   confirming normal distribution, unpaired two-tailed Student's t-test
   was used to compare two groups (P > 0.05). One-way ANOVA was used to
   compare three or more groups, followed by conducting pairwise
   comparisons between groups using Welch’s T-test. The data are presented
   as mean ± standard error of the mean (SEM). P values ≤ 0.05 were
   considered statistically significant. Graphs were generated using
   GraphPad Prism 8 for Windows or Microsoft Excel for Windows. ∗P<0.05,
   ∗∗P<0.01, ∗∗∗P<0.001, and ns for no significance.

Acknowledgments