c‐MAF Transcriptionally Activates Slc40a1 to Repress Ferroptosis in Sepsis‐Associated Encephalopathy

ABSTRACT

Background

Ferroptosis, an iron‐dependent programmed cell death driven by lipid peroxidation, has emerged as a potential contributor to sepsis‐associated encephalopathy (SAE). However, the relationship between ferroptosis and cognitive deficits following sepsis needs to be further elucidated.

Methods

Transcriptome sequencing was employed to identify solute carrier family 40 member 1 (Slc40a1) as a candidate ferroptosis‐related gene in the hippocampus of septic mice. The SAE mouse model was established via cecal ligation and perforation (CLP) after treatment with recombinant adeno‐associated virus 9 (AAV9)‐CaMKII to knock down or overexpress musculoaponeurotic fibrosarcoma (c‐Maf) or Slc40a1. We assessed cognitive performance, Nissl staining, and ferroptosis‐associated parameters. Dual‐luciferase reporter gene assays and chromatin immunoprecipitation assays were performed to illuminate the mechanism by which c‐MAF transcriptionally activates Slc40a1.

Results

Hippocampal neurons of mice subjected to CLP showed downregulation of Slc40a1. Neuron‐specific knockdown of Slc40a1 or c‐Maf deteriorated sepsis‐induced cognitive impairment, oxidative stress, and ferroptosis. Conversely, overexpression of Slc40a1 or c‐Maf attenuated acute mortality and cognitive impairment following CLP, hampered lipid peroxidation and iron deposition, and enhanced antioxidant capacity. Moreover, Slc40a1 silencing neutralized the anti‐ferroptotic property of c‐Maf in SAE. Mechanistically, c‐MAF was found to directly bind to the Slc40a1 promoter and facilitate its transcription.

Conclusions

Our findings suggest that c‐MAF/Slc40a1 may represent a promising prevention target for SAE.

c‐MAF and its downstream target Slc40a1 were identified as key regulators of ferroptosis in septic mouse hippocampi. Neuron‐specific knockdown of either gene aggravated CLP‐induced cognitive deficits and ferroptosis, whereas their overexpression exerted protective effects. c‐MAF directly activates Slc40a1 transcription, indicating the c‐MAF/Slc40a1 axis as a potential therapeutic target for SAE.

1. Introduction

Sepsis is an intricate condition characterized by fatal multiple organ dysfunction arising from an imbalanced host immune response to infection [1]. Sepsis‐associated encephalopathy (SAE) occurs in approximately 50% of septic patients during intensive care unit (ICU) hospitalization, and a portion of survivors experience enduring cognitive deficits post‐discharge [2]. Emerging evidence implicates that various elements, such as microglial hyperactivation, blood–brain barrier (BBB) compromise, ischemic or hemorrhagic cerebral insults, and neurotransmitter dysregulation, have been proposed as key contributors to SAE [3]. However, the exact molecular mechanisms underlying SAE remain poorly understood, underscoring the imperative to delineate its pathogenesis for developing targeted therapeutic interventions aimed at decreasing the occurrence of SAE.

Ferroptosis, an iron‐dependent non‐apoptotic cell death paradigm, is pathologically characterized by distinctive mitochondrial ultrastructural alterations, including increased bilayer membrane density, cristae diminution or disappearance, and outer membrane rupture [4]. Iron overload and dysregulated lipid peroxidation are two cardinal features of ferroptosis. Cells employ both enzymatic and non‐enzymatic defense mechanisms to counteract cytotoxic membrane lipid peroxidation [5]. The solute carrier family 40 member 1 (SLC40A1) gene encodes the cellular iron exporter ferroportin (FPN) that facilitates the efflux of divalent iron from intracellular compartments to maintain iron homeostasis [6]. Loss of SLC40A1 initiates a pathological cascade wherein iron overload synergizes with lipid peroxidation to trigger ferroptosis [7]. Accumulating evidence indicates that Slc40a1 deficiency exacerbates renal tubular ferroptosis and alveolar epithelial injury during sepsis [8, 9]. However, its role in the pathogenesis of SAE remains unclear.

As a transcriptional activator or repressor, musculoaponeurotic fibrosarcoma (MAF, also known as c‐MAF) encompasses the C‐terminal basic region‐leucine zipper (bZIP) domain and plays a pivotal role in neural development and synaptic plasticity [10]. c‐MAF interacts with the Maf response element (MARE) site of the Il10 promoter to exert anti‐inflammatory effects during sepsis‐induced acute lung injury [11, 12, 13]. It also indirectly regulates both glutathione synthesis and mitochondrial function by transcriptionally repressing PHB1 [14]. Elevated intracellular iron levels induce c‐MAF expression in regulatory T cells (Tregs) [15], indicating its potential involvement in ferroptosis during SAE.

In this study, we sought to determine whether c‐MAF protects against SAE through dampening ferroptosis. Moreover, we investigated the regulatory interactions between c‐MAF and Slc40a1 in the progression of SAE.

2. Materials and Methods

2.1. Animals

Specific‐pathogen‐free (SPF) male C57BL/6J aged 8 weeks and weighing 18–22 g were sourced from Hunan SJA Laboratory Animal Co. Ltd. (Changsha, China). Animals were housed under pathogen‐free conditions with a 12‐h light/dark cycle at 22°C–24°C and 50% humidity. Standard pellet chow and water were available ad libitum. All experimental procedures were performed upon approval from the Medical Faculty Ethics Committee of Renmin Hospital of Wuhan University (ethics approval number: WDRM‐20241005B).

2.2. Adeno‐Associated Virus 9 Transduction

Recombinant adeno‐associated virus serotype 9 (AAV9) carrying murine Slc40a1 and c‐Maf, as well as short hairpin RNAs (shRNAs) against Slc40a1 and c‐Maf under the control of the neuron‐specific calcium/calmodulin‐dependent protein kinase II (CaMKII) promoter, were engineered by GeneRulor Technology (Zhuhai, China). The shRNA sequences targeting Slc40a1 and c‐Maf were 5′‐GTGGATCCATCCTTAGTATTT‐3′ and 5′‐CCTGTCTTAGAAAGAAAGAAA‐3′, respectively. Anesthetized mice were positioned in a stereotaxic apparatus, and AAVs were microinjected into the bilateral hippocampal cornu ammonis (CA) 1 region at a speed of 0.1 μL/min, with 1 μL per hemisphere. The insertion coordinates were as follows: anteroposterior (AP) −2.0 mm, mediolateral (L) ± 2.3 mm, dorsoventral (DV) −2.0 mm [16, 17]. The animals were allowed to recuperate from viral injection for 14 days before modeling.

2.3. SAE Model

Cecal ligation and perforation (CLP), a standardized method for simulating SAE in rodents, was conducted as previously described [18]. Animals were anesthetized via intraperitoneal injection of pentobarbital sodium (50 mg/kg). After shaving and disinfection, a longitudinal incision was made in the lower abdomen to exteriorize the cecum. A 3–0 silk was employed to ligate half of the distal cecum below the ileocecal valve. The cecum was perforated twice using a sterile 21‐gauge needle, and a small amount of fecal material was gently extruded from each puncture to ensure patency. The peritoneum, muscle and skin were then closed with 6–0 absorbable sutures. Mice were immediately resuscitated through subcutaneous injection of 1 mL of 0.9% saline. The sham‐treated mice underwent laparotomy without ligation or perforation. Thermal support was provided with a heating pad to prevent perioperative hypothermia. Compound lidocaine cream was used to provide pain relief after surgery.

2.4. RNA Sequencing

Three pairs of mouse hippocampi from the sham and CLP groups were collected for RNA sequencing (RNA‐seq) 24 h post‐laparotomy. Following RNA extraction and library preparation, Illumina PE150 and fastp were utilized for RNA‐seq and quality assessment of sequencing data. Paired‐end clean reads were aligned to the reference genome using HISAT2. Differentially expressed genes (DEGs) were analyzed using DESeq2 and identified based on a cutoff standard of |log₂FC| > 1 and adjusted p value < 0.05.

2.5. Novel Object Recognition (NOR) Test

Animals were acclimated to the experimental environment 8 days after CLP. During the familiarization stage, mice were permitted for 5 min to explore two identical cuboids within an open field apparatus (50 × 50 × 40 cm³). After a 24‐h interval, one of the cuboids was substituted with a novel cylindrical object. The discrimination index was calculated using the formula DI = Tn/(Tn + Tf), where Tn and Tf equate the cumulative time spent exploring the novel and familiar objects, respectively [19].

2.6. Morris Water Maze (MWM) Test

The Morris water maze (MWM) test was conducted 11–16 days post‐CLP to evaluate the hippocampus‐dependent spatial memory, as published protocols outlined [20]. The cylindrical tank (50‐cm height and 120‐cm diameter) was equipped with a 6‐cm‐diameter platform in the fourth quadrant. The MWM test comprised two distinct phases: the acquisition phase performed four times daily over five consecutive days, and the retention phase conducted without the hidden platform. Animals were randomly introduced to four different initial positions and given 60 s to locate the platform. If unsuccessful, they were guided to the platform manually. The cumulative time taken to locate the concealed platform was recorded as the escape latency. Duration in the fourth (target) quadrant, and platform crossing times were automatically documented using a video tracking system (XinRuan, Shanghai, China).

2.7. Nissl Staining

Three mice were randomly selected from each group and euthanized under pentobarbital sodium anesthesia to extract hippocampal CA1 specimens on ice. Paraffin‐embedded sections underwent dehydration through a succession of alcohol gradients, followed by a 10‐min incubation with Nissl staining solution. After cover‐slipping with neutral resin, an investigator blinded to the experimental interventions was responsible for morphological observation under a light microscope (Olympus, Tokyo, Japan). The number of bilateral hippocampal surviving neurons of each mouse was quantified using ImageJ software.

2.8. Western Blot

Proteins were isolated from the hippocampal CA1 area using radioimmunoprecipitation assay (RIPA) buffer (ASPEN, Wuhan, China). The lysates were separated via sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE, ASPEN, Wuhan, China), followed by transfer onto polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, USA). The membranes were sequentially incubated with primary antibodies and horseradish peroxidase (HRP)‐conjugated secondary antibodies (1:10,000, ASPEN, Wuhan, China). The primary antibodies utilized in our study were as follows: c‐MAF (1:3000, #ab77071, Abcam), FPN (1:1000, #26601–1‐AP, Proteintech), GPX4 (1:1000, #67763–1‐Ig, Proteintech), Hepcidin (1:500, #orb1254763, Biorbyt), TFR1 (1:1000, #ab214039, Abcam), FTH1 (1:1000, #3998, CST), and β‐actin (1:10,000, #TDY051, TDYbio). Immunodetection was conducted using an enhanced chemiluminescence (ECL) assay kit (ASPEN, Wuhan, China).

2.9. Real‐Time Quantitative Polymerase Chain Reaction (RT‐qPCR)

Total RNA was isolated from hippocampal tissues utilizing TRIpure reagent (ELK, Wuhan, China). Complementary DNA (cDNA) was synthesized using the EntiLink 1st Strand cDNA Synthesis Super Mix (ELK, Wuhan, China). PCR amplification was implemented using the EnTurbo SYBR Green PCR SuperMix kit. The primer sequences for Slc40a1 were 5′‐AGATTAGCAGACATGAATGCTACC‐3′ (sense) and 5′‐GCTTCAGTTCTGACTCCTCTACCT‐3′ (antisense), while those for β‐actin were 5′‐CTGAGAGGGAAATCGTGCGT‐3′ (sense) and 5′‐CCACAGGATTCCATACCCAAGA‐3′ (antisense). The relative Slc40a1 mRNA expression was normalized to that of β‐actin using the 2^−ΔΔCt method.

2.10. Immunofluorescence Staining

Three mice from each group were randomly selected for euthanasia to extract hippocampal CA1 specimens. Paraffin‐embedded sections were incubated with primary antibodies against c‐MAF (1: 200, #55013‐1‐AP, Proteintech), FPN (1: 200, #DF13561, Affinity), NEUN (1: 200, #66836‐1‐Ig, Proteintech), GFAP (1: 300, #16825‐1‐AP, Proteintech), Iba‐1 (1: 5000, #Ab283319, Abcam), Claudin‐5 (1: 400, #29767‐1‐AP, Proteintech), and ZO‐1 (1: 400, #21773‐1‐AP, Proteintech) at 4°C overnight, followed by incubation with Cy3‐labeled goat anti‐rabbit IgG (1: 100, #AS1109, ASPEN) or CoraLite488‐labeled goat anti‐mouse IgG (1: 100, #SA00013‐1, Proteintech) at 37°C in the dark. Finally, the slices were titrated with 4′,6‐diamidino‐2‐phenylindole (DAPI) at room temperature in the dark, and observed under a fluorescence microscope (OLYMPUS, Tokyo, Japan).

2.11. Measurement of Reactive Oxygen Species (ROS) Levels

Hippocampal tissues were promptly dissected and snap‐frozen in liquid nitrogen. After homogenization and centrifugation, the supernatant was collected and incubated with the BBoxiProbe O08 probe (Best Bio, Shanghai, China) at 37°C for 30 min. The hippocampal ROS content was determined by the ratio of ROS fluorescence intensity to protein concentration [21]. Quantitative analysis of ROS was assessed by its ratio to the content of the sham or CLP group, which was standardized to 100%.

2.12. Isolation and Culture of Hippocampal Neurons

Five mice from each group were euthanized under pentobarbital sodium anesthesia, and the hippocampi were rapidly removed under sterile conditions. The hippocampal tissues were digested with 0.125% trypsin in a 5% CO₂, 37°C incubator. The tissues were gently triturated using a sterile pipette to obtain a single‐cell suspension while minimizing mechanical damage. The suspension was filtered through a 70‐μm cell strainer, and the filtrate was collected and centrifuged to pellet the cells. The resulting pellet was resuspended in DMEM/HG complete medium. Neurons were seeded at the appropriate density onto poly‐L‐lysine–coated culture plates. Following a 4–6 h incubation period, the medium was replaced with Neurobasal‐A supplemented with 2% B‐27, 0.5 mM L‐glutamine, and 1% penicillin–streptomycin.

2.13. Lipid Peroxidation Detection by Flow Cytometry

Neurons were seeded to 70% confluency in 6‐well plates and then incubated for 30 min with 2 μM BODIPY 581/591 C11 (Invitrogen, Carlsbad, USA). Neurons were harvested for analysis with a CytoFLEX flow cytometer (Beckman Coulter, Brea, USA). Oxidation of the polyunsaturated moiety of BODIPY‐C11 shifts its fluorescence emission from ~590 nm (reduced) to ~510 nm (oxidized), and the extent of lipid peroxidation was quantified by calculating the fluorescence intensity.

2.14. Determination of 4‐HNE, MDA, GSH, SOD, and Iron

Blood samples were collected in sterilized EP tubes and allowed to coagulate at room temperature, followed by centrifugation at 2000 g for 10 min to obtain the serum. Hippocampal homogenates were assessed for protein concentration using Bicinchoninic acid (BCA) assay (ASPEN, Wuhan, China). The serum level of 4‐hydroxynonenal (4‐HNE) and the hippocampal levels of malondialdehyde (MDA), reduced glutathione (GSH), superoxide dismutase (SOD), and iron were quantified using commercial kits for 4‐HNE (#H268‐1), MDA (#A003‐1), GSH (#A006‐2), SOD (#A001‐3), and iron (#A039‐2) (Jiancheng, Nanjing, China) following the manufacturer's instructions.

2.15. FerroOrange Fluorescent Probe for Fe²⁺ Detection

Intracellular Fe²⁺ levels were measured using the FerroOrange probe (Dojindo, Japan) according to the manufacturer's instructions. Briefly, neurons were stained with 1 μM FerroOrange working solution for 30 min at 37°C in the dark. After rinsing with PBS, neurons were examined on a CytoFLEX flow cytometer (Beckman Coulter, Brea, USA) using a 543‐nm excitation laser.

2.16. Dual‐Luciferase Report Gene Assay

The full‐length (FL) Slc40a1 promoter sequence and its truncations (−2000/−1301, −1300/−601, and −600/+99) were amplified via RT‐qPCR and subsequently inserted into the pGL3‐luciferase reporter vector. HT22 neurons were co‐transfected for 36 h with the specified luciferase plasmids, along with either a pcDNA‐empty vector or pcDNA‐c‐Maf. The primers for Slc40a1‐promoter constructs are detailed in Table 1. Relative luciferase activity was analyzed using a luciferase reporter gene assay kit (Beyotime, Shanghai, China).

TABLE 1.

The primer sequences for Slc40a1‐promoter full length and its truncations.

Slc40a1 promoter		Primer sequences (5′–3′)
FL	Forward	GAGCTCTTACGCGTGCTAGCGTAATAAATATATATTTAAT
FL	Reverse	CTTAGATCGCAGATCTCGAGCAGGTTTGTCAACACTTCAA
−2000/−1301	Forward	GAGCTCTTACGCGTGCTAGCCCCCATTGATGACAATGGAG
−2000/−1301	Reverse	CTTAGATCGCAGATCTCGAGCAGGTTTGTCAACACTTCAA
−1300/−601	Forward	GAGCTCTTACGCGTGCTAGCGTACATTTTATTCTGGCTGG
−1300/−601	Reverse	CTTAGATCGCAGATCTCGAGAGCTTCTGTTCCCCATTGTC
−600/+99	Forward	GAGCTCTTACGCGTGCTAGCGTAATAAATATATATTTAAT
−600/+99	Reverse	CTTAGATCGCAGATCTCGAGTCTGATTGGAAGTTGTTAAC

2.17. Chromatin Immunoprecipitation Coupled With Quantitative PCR (ChIP‐qPCR) Assay

HT22 cells were fixed in formaldehyde to crosslink DNA and c‐MAF, followed by sonication to shear chromatin into fragments. The samples were incubated with either an anti‐c‐MAF (1:100, #sc‐518,062, Santa Cruz) or IgG. The purified DNA sequences were amplified by RT‐qPCR using specific primers targeting the murine Slc40a1 promoter region: 5′‐TTTCCTATGATGGTCATTCTGATG‐3′ (sense) and 5′‐TTCTTCTAGAGGTCCTGAGTTCAAT‐3′ (antisense).

2.18. Statistical Analyses

The Shapiro–Wilk test was utilized to evaluate the normality of the data distribution. All results were denoted as mean ± standard error of the mean (SEM) and analyzed using an unpaired Student's t‐test, one‐way analysis of variance (ANOVA) followed by Tukey's post hoc test, two‐way ANOVA followed by Bonferroni correction for multiple testing. GraphPad Prism version 9.0 was employed to implement statistical analyses, with statistical significance established at p < 0.05.

3. Results

3.1. Cecal Ligation and Puncture Diminishes Slc40a1 Expression

To gain insight into the pathogenesis of SAE, we analyzed our transcriptomic signatures in the hippocampus of sham and CLP‐treated mice. A total of 331 DEGs with thresholds of |log₂FC| > 1 and adjust p value < 0.05 were identified in the mouse hippocampi post‐CLP, of which 213 were upregulated and 118 were downregulated (Figure 1A). Given the crucial pathogenic mechanism of ferroptosis in SAE [22], we further screened two ferroptosis‐related genes by integrating a known ferroptosis gene set [23] and DEGs. NAD(P)H dehydrogenase, quinone 1 (Nqo1) was upregulated, while Slc40a1 was downregulated (Figure 1B,C). NQO1 has been shown to exhibit a context‐dependent dual regulatory function in ferroptosis. Its overexpression (OE) attenuated ferroptosis in circ‐DAPK1‐OE human umbilical vein endothelial cells (HUVECs) under high‐glucose conditions [24], whereas its knockdown reduced system xCT expression to promote ferroptosis in glioma [25]. We therefore focused on Slc40a1 function in the development of SAE. RT‐qPCR and immunoblot analyses showed downregulation of Slc40a1 in the hippocampus of CLP‐induced septic mice (Figure 1D,E). Double‐labeled immunofluorescence in wild‐type mice demonstrated that FPN colocalized strongly with the neuronal marker NEUN and exhibited an intermediate degree of overlap with endothelial markers (Claudin‐5 and ZO‐1), whereas almost no overlap was detected with astrocytic (GFAP) or microglial (Iba‐1) markers (Figure 1F). Reduced FPN immunoreactivity was found in the hippocampal neurons after CLP (Figure 1G,H).

FIGURE 1.

Slc40a1 is downregulated in the hippocampus of mice after CLP. (A) Heatmap plots of the 331 differentially expressed genes (DEGs). (B) Venn diagram of intersected DEGs with the ferroptosis genes. (C) Volcano plots of the 331 DEGs. (D) RT‐qPCR assay for Slc40a1 mRNA in the hippocampus after CLP (n = 6). (E) Western blot analysis of FPN expression in the hippocampus of mice after CLP (n = 6). (F) The colocalization of FPN with neurons, astrocytes, microglia, and endothelial cells was assessed by double immunofluorescence in wild‐type mice (400×, bar: 20 μm) (n = 3). (G) Double immunofluorescence staining of FPN and NEUN in the hippocampus of the sham and CLP groups (200×, bar: 50 μm). (H) Mean fluorescence intensity of FPN was quantified (n = 5). Data are denoted as mean ± SEM. **p < 0.01, unpaired t‐test.

3.2. Selective Neuronal Knockdown of Slc40a1 Aggravates Cognitive Impairment, Oxidative Stress, and Ferroptosis in Septic Mice

To elucidate the role of Slc40a1 in the pathogenesis of SAE, recombinant AAV9‐CaMKII‐shSlc40a1 was employed to target hippocampal neurons in mice subjected to CLP. Survival analysis revealed that the 7‐day survival rate of the CLP group was 52.4% (11 of 21 mice survived). Slc40a1 knockdown (KD) further exacerbated acute mortality post‐CLP (29% mice survived) compared to the CLP + shNC group (50%) (Figure 2A). The NOR and MWM tests were conducted 7 days after CLP to assess the effects of Slc40a1 on cognitive ability. The discrimination index was significantly reduced following CLP, and septic mice receiving AAV9‐shSlc40a1 showed reduced exploration time of the novel object compared to the CLP + shNC group (Figure 2B). During the location navigation phase of the MWM test, CLP‐treated mice exhibited a prolonged escape latency on the 3rd, 4th and 5th days relative to the sham‐treated mice. The latency was further prolonged in the CLP + shSlc40a1 group compared to the CLP + shNC group (Figure 2C). Spatial memory impairments following CLP were evident during the probe trial, as demonstrated by decreased target quadrant stay and platform crossings. Neuron‐specific Slc40a1 knockdown further reduced duration in the target quadrant and platform crossing times (Figure 2D,E). The vacuole‐like neurons in the hippocampal CA1 region of the sham group were densely packed with clearly defined nuclei. However, the number of Nissl‐positive neurons was significantly reduced after CLP. The CLP + shSlc40a1 group demonstrated more conspicuous lesions characterized by neuronal loss and sparse arrangements compared to the CLP + shNC group (Figure 2F,G). The markedly decreased Slc40a1/FPN expression confirmed successful transfection of the AAV9 vector harboring shSlc40a1. Immunoblot analysis revealed diminished levels of GPX4 in the hippocampus of septic mice. Further restraint of FPN and GPX4 expression was noted in Slc40a1‐deficient mice subjected to CLP (Figure 2H; Figures S1 and S2). Excessive ROS production was seen in the hippocampus of CLP‐induced septic mice, which was further augmented by Slc40a1 depletion (Figure 2I). C11‐BODIPY581/591 staining revealed increased lipid ROS in hippocampal neurons extracted from CLP‐treated mice, which was further amplified by Slc40a1 knockdown (Figure 2J). The concentration of the active aldehyde 4‐HNE was prominently elevated in the serum of septic mice. Neuron‐specific Slc40a1 ablation caused more pronounced 4‐HNE accumulation compared to the CLP + shNC group (Figure 2K). The lipid peroxide end‐product MDA levels were substantially increased in the hippocampus of mice treated with CLP, which was further intensified by AAV9‐shSlc40a1. CLP‐induced septic mice exhibited diminished hippocampal GSH concentration and antioxidant activity of SOD, whereas Slc40a1‐deficient mice displayed weaker antioxidant capacity after CLP compared to the CLP + shNC group (Figure 2L–N). Additionally, iron overload was found in the hippocampus of mice exposed to CLP, which was further aggravated by Slc40a1 silencing (Figure 2O,P).

FIGURE 2.

Slc40a1 silencing deteriorates CLP‐induced cognitive impairment, oxidative stress, and ferroptosis. (A) The survival rates of mice after CLP. **p < 0.01, Sham vs. CLP; ^# p < 0.05, CLP + shNC vs. CLP + shSlc40a1. (B) Discrimination index in the NOR test. (C) Escape latency during the acquisition session of MWM test. Two‐way ANOVA followed by Bonferroni's multiple comparisons test. **p < 0.01, Sham vs. CLP; ^## p < 0.01, CLP + shNC vs. CLP + shSlc40a1. Time spent in the target quadrant (D) and platform crossing times (E) during the space exploration phase of the MWM test (n = 9–12 per group). (F, G) Representative Nissl staining images in hippocampal CA1 samples following the indicated treatments (400×, bar: 50 μm) and the number of Nissl‐positive neurons (n = 3). (H) Representative gel bands of FPN and GPX4 in the hippocampus following the indicated treatments (n = 6). (I) Hippocampal ROS levels following the indicated treatments. (J) Lipid peroxidation in hippocampal neurons was assessed using BODIPY‐C11 staining followed by flow cytometric analysis. (K) 4‐HNE concentration in the serum of mice following the indicated treatments. (L−O) Hippocampal MDA, GSH, SOD, and iron levels following the indicated treatments. (P) Ferrous iron in hippocampal neurons was detected by FerroOrange staining followed by flow cytometry analysis (n = 5). Data are denoted as mean ± SEM. **p < 0.01, one‐way ANOVA followed by Tukey's post hoc test.

3.3. Neuron‐Specific Slc40a1 Overexpression Ameliorates Cognitive Impairment, Oxidative Stress, and Ferroptosis in Septic Mice

Recombinant AAV9‐CaMKII‐Slc40a1 was utilized to further illuminate the role of Slc40a1 in sepsis‐induced cognitive dysfunction. Slc40a1 overexpression significantly improved the survival rates of mice subjected to CLP (42.9% for CLP + Vec group and 80% for CLP + Slc40a1 group) (Figure 3A). CLP‐treated mice receiving preemptive injection of AAV9‐Slc40a1 significantly preferred the novel object in the NOR test (Figure 3B). The escape latency was shorter in the CLP + Slc40a1 group on the 3rd, 4th, and 5th days compared to that in the CLP + Vec group (Figure 3C). Slc40a1 overexpression ameliorated CLP‐induced cognitive dysfunction, as evidenced by increased target quadrant residence time and platform crossing times relative to the CLP + Vec group (Figure 3D,E). The number of Nissl‐positive neurons was markedly increased in the CA1 region of the CLP + Slc40a1 group compared to the CLP + Vec group (Figure 3F,G). The markedly elevated Slc40a1/FPN expression suggested successful transfection of the AAV9 vector harboring Slc40a1. Western blot analysis showed upregulation of GPX4 in the hippocampus of Slc40a1‐OE septic mice (Figure 3H; Figures S3 and S4). Slc40a1 overexpression attenuated hippocampal ROS and lipid peroxidation, as well as decreased serum 4‐HNE under septic conditions (Figure 3I–K). Diminished MDA and iron levels were observed in the hippocampus of septic mice receiving AAV9‐Slc40a1 therapy compared to those in the CLP + Vec group. Septic mice receiving AAV9‐CaMKII‐Slc40a1 demonstrated augmented GSH and SOD levels (Figure 3L–P).

FIGURE 3.

Slc40a1 overexpression alleviates CLP‐induced cognitive impairment, oxidative stress, and ferroptosis. (A) The survival rates of mice after CLP. **p < 0.01, Sham vs. CLP; ^# p < 0.05, CLP + Vec vs. CLP + Slc40a1. (B) Discrimination index in the NOR test. (C) Escape latency during the acquisition session of MWM test. Two‐way ANOVA followed by Bonferroni's multiple comparisons test. **p < 0.01, Sham vs. CLP; ^# p < 0.05, ^## p < 0.01, CLP + Vec vs. CLP + Slc40a1. Time spent in the target quadrant (D) and platform crossing times (E) during the space exploration phase of the MWM test (n = 9–12 per group). (F, G) Representative Nissl staining images in hippocampal CA1 samples following the indicated treatments (400×, bar: 50 μm) and the number of Nissl‐positive neurons (n = 3). (H) Representative gel bands of FPN and GPX4 in the hippocampus following the indicated treatments (n = 6). (I) Hippocampal ROS levels following the indicated treatments. (J) Lipid peroxidation in hippocampal neurons was assessed using BODIPY‐C11 staining followed by flow cytometric analysis. (K) 4‐HNE concentration in the serum of mice following the indicated treatments. (L−O) Hippocampal MDA, GSH, SOD, and iron levels following the indicated treatments. (P) Ferrous iron in hippocampal neurons was detected by FerroOrange staining followed by flow cytometry analysis (n = 5). Data are denoted as mean ± SEM. *p < 0.05, **p < 0.01, one‐way ANOVA followed by Tukey's post hoc test.

3.4. Slc40a1 Is Transcriptionally Activated by c‐MAF

To further investigate the molecular mechanisms responsible for Slc40a1 gene downregulation in SAE, we integrated predictions of Slc40a1 promoter‐binding transcription factors with the DEGs. Seven candidate transcription factors (MAFF, MAFK, IRF7, SOX17, GATA2, RORC, and c‐MAF) were selected by intersection analysis (Figure 4A). Online single‐cell RNA‐sequencing data [26] and our experimental evidence indicate that Slc40a1 is predominantly distributed in neurons. c‐MAF emerged as the most likely candidate responsible for regulating Slc40a1 due to insufficient neuronal expression of the other transcription factors (Figure S5). Slc40a1 exhibited a positive correlation with c‐MAF based on the GEPIA database (R = 0.63) (Figure 4B). RT‐qPCR and immunoblot analysis showed downregulation of c‐MAF in the hippocampus of CLP‐induced septic mice (Figure S6 and Figure 4C). Reduced c‐MAF immunoreactivity was found in the hippocampal neurons after CLP (Figure 4D). Dual‐luciferase reporter assay revealed that pcDNA‐c‐Maf significantly enhanced Slc40a1 promoter activity, suggesting a potential interaction between c‐MAF and the Slc40a1 promoter (Figure 4E). The region spanning from 600 bp upstream to 99 bp downstream of the Slc40a1 gene transcription start site (TSS), as well as 1301−2000 bp upstream of the TSS, abrogated c‐MAF‐dependent promoter activation compared to the full‐length Slc40a1 promoter. Conversely, c‐MAF overexpression did not affect the luciferase activity when introducing the luciferase reporter plasmid encoding only the sequence from −601 bp to −1300 bp of the Slc40a1 promoter (Figure 4F). ChIP assay confirmed that the anti‐c‐MAF antibody, but not IgG, specifically enriched DNA fragments encompassing the Slc40a1 promoter region (Figure 4G).

FIGURE 4.

c‐MAF is the transcription factor of Slc40a1. (A) Prediction of Slc40a1 promoter‐binding transcription factors by AnimalTFDB4.0. The 2000 bp region upstream to 99 bp downstream of Slc40a1 gene transcription start site (TSS) is regarded as the promoter region. Venn diagram of 7 intersected transcription factors with DEGs between sham and CLP‐treated mice. (B) Correlation analysis of c‐MAF with Slc40a1 in the hippocampus based on Gene Expression Profile Interactive Analysis (GEPIA; http://gepia2.cancer‐pku.cn). (C) Western blot analysis of c‐MAF expression in the hippocampus of mice after CLP (n = 6). (D) Double immunofluorescence staining of c‐MAF and NEUN in the hippocampus of the sham and CLP groups (200×, bar: 50 μm). Mean fluorescence intensity of c‐MAF was quantified (n = 5). (E) Relative luciferase activity in HT22 cells transfected with luciferase reporter plasmids containing murine Slc40a1 promoter, followed by transfection with either a pcDNA‐Vector or pcDNA‐c‐Maf. (F) Relative luciferase activity in HT22 cells transfected with luciferase reporter plasmids containing full‐length murine Slc40a1 promoter or its truncations, followed by transfection with pcDNA‐c‐Maf. (G) ChIP assay for the binding of c‐MAF to Slc40a1 promoter. Data are denoted as mean ± SEM (n = 6). **p < 0.01, unpaired t‐test or one‐way ANOVA followed by Tukey's post hoc test.

3.5. Selective Neuronal Knockdown of c‐Maf Exacerbates Sepsis‐Induced Cognitive Impairment, Oxidative Stress, and Ferroptosis

AAV9 bearing shRNA against murine c‐Maf under the control of CaMKII promoter was utilized to address the role of c‐Maf in a well‐established SAE model. Survival analysis revealed that c‐Maf knockdown increased acute mortality after CLP (8 of 32 mice survived) compared to the CLP + shNC group (9 of 18 mice survived) (Figure 5A). The discrimination index was further reduced in c‐Maf‐deficient septic mice compared to the negative control CLP mice (Figure 5B). During the training sessions of the MWM test, AAV9‐shc‐Maf further prolonged the latency to find the hidden platform relative to the negative control under CLP conditions (Figure 5C). AAV9‐shc‐Maf‐treated septic mice suffered from worse memory consolidation, as reflected by reduced target quadrant stay and platform crossings compared to the negative control (Figure 5D,E). More pronounced neuronal loss and irregular arrangements were observed in the CLP + shc‐Maf group compared to the CLP + shNC group (Figure 5F,G). The significantly decreased c‐MAF expression indicated successful transfection of AAV9 particles harboring shc‐Maf. Immunoblot analysis revealed further diminution of FPN and GPX4 in c‐Maf‐KD mice compared to the negative control under CLP conditions (Figure 5H; Figures S7 and S8A−C). Moreover, septic mice exhibited elevated hippocampal Hepcidin and TFR1 levels, accompanied by reduced FTH1 expression. AAV9‐shc‐Maf exacerbated the sepsis‐induced upregulation of Hepcidin and TFR1 and further suppressed FTH1 expression (Figure 5I and Figure S8D−F). RT‐qPCR assay showed further restraint of Slc40a1 mRNA in the hippocampus of c‐Maf‐deficient mice subjected to CLP compared to the CLP + shNC group (Figure 5J). Total ROS and lipid ROS further accumulated in the hippocampus of septic mice receiving AAV9‐shc‐Maf relative to the CLP + shNC group (Figure 5K,L). Neuronal knockdown of c‐Maf further augmented the serum concentration of 4‐HNE after CLP (Figure 5M). We found that c‐Maf silencing further exacerbated CLP‐induced lipid peroxidation and iron overload. Septic mice receiving injection of AAV9‐shc‐Maf demonstrated further restraint of GSH and SOD levels (Figure 5N–R).

FIGURE 5.

Neuronal c‐Maf silencing exacerbates CLP‐induced cognitive impairment, oxidative stress, and ferroptosis. (A) The survival rates of mice after CLP. **p < 0.01, Sham vs. CLP; ^# p < 0.05, CLP + shNC vs. CLP + shc‐Maf. (B) Discrimination index in the NOR test. (C) Escape latency during the acquisition session of MWM test. Two‐way ANOVA followed by Bonferroni's multiple comparisons test. *p < 0.05, **p < 0.01, Sham vs. CLP; ^## p < 0.01, CLP + shNC vs. CLP + shc‐Maf. Time spent in the target quadrant (D) and platform crossing times (E) during the space exploration phase of the MWM test (n = 8–12 per group). (F, G) Representative Nissl staining images in hippocampal CA1 samples following the indicated treatments (400×, bar: 50 μm) and the number of Nissl‐positive neurons (n = 3). (H, I) Representative gel bands of c‐MAF, FPN, GPX4, Hepcidin, TFR1, and FTH1 in the hippocampus following the indicated treatments (n = 6). (J) RT‐qPCR assay for Slc40a1 mRNA in the hippocampus following the indicated treatments. (K) Hippocampal ROS levels following the indicated treatments. (L) Lipid peroxidation in hippocampal neurons was assessed using BODIPY‐C11 staining followed by flow cytometric analysis. (M) 4‐HNE concentration in the serum of mice following the indicated treatments. (N−Q) Hippocampal MDA, GSH, SOD, and iron levels following the indicated treatments. (R) Ferrous iron in hippocampal neurons was detected by FerroOrange staining followed by flow cytometry analysis (n = 5). Data are denoted as mean ± SEM. *p < 0.05, **p < 0.01, one‐way ANOVA followed by Tukey's post hoc test.

3.6. Neuron‐Specific c‐Maf Overexpression Mitigates Cognitive Impairment, Oxidative Stress, and Ferroptosis in Septic Mice

To further clarify the role of c‐Maf in the development of SAE, AAV9‐CaMKII‐c‐Maf was delivered in septic mice to yield extensive expression in the hippocampus. Survival analysis revealed that c‐Maf overexpression significantly reduced acute mortality after CLP (12 of 14 mice survived) compared to the CLP + Vec group (9 of 19 mice survived) (Figure 6A). CLP mice treated with AAV9‐c‐Maf displayed higher preference for novel objects compared to the CLP + Vec group (Figure 6B). The escape latency was shorter in the CLP + c‐Maf group compared to that in the CLP + Vec group (Figure 6C). AAV9‐CaMKII‐c‐Maf mitigated CLP‐induced cognitive disturbances, as evidenced by prolonged target quadrant residence time and increased platform crossing frequency (Figure 6D,E). Nissl⁺ neurons were significantly increased in the hippocampal CA1 region of the CLP + c‐Maf group compared to the CLP + Vec group (Figure 6F,G). The markedly elevated c‐MAF expression suggested successful transfection of the AAV9 vectors harboring c‐Maf. AAV9‐c‐Maf enhanced hippocampal FPN and GPX4 protein abundance compared to the empty vector under CLP conditions (Figure 6H; Figures S9 and S10A–C). Moreover, c‐Maf overexpression reduced Hepcidin and TFR1 induction under septic conditions and simultaneously elevated FTH1 expression (Figure 6I and Figure S10D–F). RT‐qPCR assay showed increased Slc40a1 mRNA in the hippocampus of c‐Maf‐OE mice subjected to CLP compared to the CLP + Vec group (Figure 6J). Neuron‐specific c‐Maf overexpression repressed CLP‐elicited ROS generation, lipid peroxidation, and 4‐HNE deposition (Figure 6K–M). Decreased MDA and iron levels were noted in the hippocampus of c‐Maf‐OE mice after CLP. AAV9‐c‐Maf markedly elevated hippocampal GSH and SOD compared to the empty vector under CLP conditions (Figure 6N–R).

FIGURE 6.

c‐Maf overexpression alleviates CLP‐induced cognitive impairment, oxidative stress, and ferroptosis. (A) The survival rates of mice after CLP. **p < 0.01, Sham vs. CLP; ^# p < 0.05, CLP + Vec vs. CLP + c‐Maf. (B) Discrimination index in the NOR test. (C) Escape latency during the acquisition session of MWM test. Two‐way ANOVA followed by Bonferroni's multiple comparisons test. **p < 0.01, Sham vs. CLP; ^# p < 0.05, ^## p < 0.01, CLP + Vec vs. CLP + c‐Maf. Time spent in the target quadrant (D) and platform crossing times (E) during the space exploration phase of the MWM test (n = 8–12 per group). (F, G) Representative Nissl staining images in hippocampal CA1 samples following the indicated treatments (400×, bar: 50 μm) and the number of Nissl‐positive neurons (n = 3). (H, I) Representative gel bands of c‐MAF, FPN, GPX4, Hepcidin, TFR1, and FTH1 in the hippocampus following the indicated treatments (n = 6). (J) RT‐qPCR assay for Slc40a1 mRNA in the hippocampus following the indicated treatments. (K) Hippocampal ROS levels following the indicated treatments. (L) Lipid peroxidation in hippocampal neurons was assessed using BODIPY‐C11 staining followed by flow cytometric analysis. (M) 4‐HNE concentration in the serum of mice following the indicated treatments. (N–Q) Hippocampal MDA, GSH, SOD, and iron levels following the indicated treatments. (R) Ferrous iron in hippocampal neurons was detected by FerroOrange staining followed by flow cytometry analysis (n = 5). Data are denoted as mean ± SEM. **p < 0.01, one‐way ANOVA followed by Tukey's post hoc test.

3.7. Slc40a1 Knockdown Abolishes the Pro‐Cognitive, Antioxidant, and Anti‐Ferroptotic Properties of c‐Maf in SAE

To ascertain whether Slc40a1 silencing blunts c‐Maf‐mediated phenotypes, AAV9‐CaMKII‐shSlc40a1 and AAV9‐CaMKII‐c‐Maf were utilized to map the hippocampal neurons in septic mice. Survival analysis showed that the 7‐day survival rate of the CLP, CLP + shSlc40a1, CLP + c‐Maf, and CLP + c‐Maf + shSlc40a1 group was 55.6%, 25%, 92.3%, and 52.4%, respectively (Figure 7A). The discrimination index was substantially decreased in septic mice receiving concomitant regimens of AAV9‐c‐Maf and AAV9‐shSlc40a1 relative to AAV9‐c‐Maf alone (Figure 7B). Co‐administration of c‐Maf overexpression and Slc40a1 silencing significantly prolonged the escape latency on the 3rd, 4th, and 5th days during the training phase compared to AAV9‐CaMKII‐c‐Maf monotherapy (Figure 7C). CLP mice treated with the combination therapy experienced worse memory performance, as shown by shorter duration in the target quadrant and fewer platform crossing frequencies than AAV9‐c‐Maf alone (Figure 7D,E). Irregular and sparse arrangements reappeared after concurrent administration of AAV9‐c‐Maf and AAV9‐shSlc40a1. Furthermore, the number of Nissl‐positive neurons significantly decreased in the CLP + c‐Maf + shSlc40a1 group compared to the CLP + c‐Maf group (Figure 7F,G). Western blot analysis showed diminished levels of GPX4 in the hippocampus of septic mice receiving AAV9‐c‐Maf and AAV9‐shSlc40a1 compared to AAV9‐c‐Maf alone (Figure 7H and Figure S11). The augmented levels of total hippocampal ROS, neuronal lipid ROS and serum 4‐HNE were observed in septic mice treated with c‐Maf overexpression and Slc40a1 knockdown compared to those in the CLP + c‐Maf group (Figure 7I–K). Additionally, Slc40a1 depletion antagonized the anti‐ferroptotic effects of c‐Maf, as evidenced by elevated concentrations of MDA and iron, along with diminished levels of GSH and SOD in the CLP + c‐Maf + shSlc40a1 group compared to the CLP + c‐Maf group (Figure 7L–P).

FIGURE 7.

Slc40a1 knockdown neutralizes the favorable effect of c‐Maf on cognitive function in SAE. (A) The survival rates of mice after CLP. *p < 0.05, CLP vs. CLP + shSlc40a1; ^# p < 0.05, CLP vs. CLP + c‐Maf; ^& p < 0.05, CLP + c‐Maf vs. CLP + c‐Maf + shSlc40a1. (B) Discrimination index in the NOR test. (C) Escape latency during the acquisition session of MWM test. Two‐way ANOVA followed by Bonferroni's multiple comparisons test. **p < 0.01, CLP vs. CLP + shSlc40a1; ^# p < 0.05, ^## p < 0.01, CLP vs. CLP + c‐Maf; ^&& p < 0.01, CLP + c‐Maf vs. CLP + c‐Maf + shSlc40a1. Time spent in the target quadrant (D) and platform crossing times (E) during the space exploration phase of the MWM test (n = 8–12 per group). (F, G) Representative Nissl staining images in hippocampal CA1 samples following the indicated treatments (400×, bar: 50 μm) and the number of Nissl‐positive neurons (n = 3). (H) Representative gel bands of c‐MAF, FPN and GPX4 in the hippocampus following the indicated treatments (n = 6). (I) Hippocampal ROS levels following the indicated treatments. (J) Lipid peroxidation in hippocampal neurons was assessed using BODIPY‐C11 staining followed by flow cytometric analysis. (K) 4‐HNE concentration in the serum of mice following the indicated treatments. (L–O) Hippocampal MDA, GSH, SOD, and iron levels following the indicated treatments. (P) Ferrous iron in hippocampal neurons was detected by FerroOrange staining followed by flow cytometry analysis (n = 5). Data are denoted as mean ± SEM. *p < 0.05, **p < 0.01, one‐way ANOVA followed by Tukey's post hoc test.

4. Discussion

Our study linked excessive ferroptosis activation with diminished levels of c‐MAF and Slc40a1 in SAE. Neuron‐specific overexpression of c‐Maf and Slc40a1 alleviated CLP‐elicited memory deterioration and rendered potent antioxidant and anti‐ferroptotic capacity. Neuron‐specific knockdown of Slc40a1 reversed the neuroprotection of c‐Maf in SAE progression. Mechanistically, c‐MAF recognized the sequence from −1300 bp to −601 bp upstream of the TSS within the murine Slc40a1 gene and facilitated its transcription.

Iron homeostasis is orchestrated through three core components: transferrin (iron transport), ferritin (iron storage), and ferroportin (Slc40a1‐mediated iron efflux) [27]. Elevated intracellular labile iron pools drive Fenton reaction‐mediated hydroxyl radical production, which culminates in neuronal lesions such as Alzheimer's disease (AD), ischemic stroke, and SAE [28, 29, 30]. Previous work has shown that CLP induces learning and memory deficits in septic rodents, accompanied by elevated ROS, Fe²⁺, and MDA, reduced GSH and SOD, as well as decreased GPX4 expression [31]. These alterations are indicative of a pronounced oxidative and ferroptotic disturbance, which aligns with the pathological changes observed in our study. Consistent with lipopolysaccharide (LPS)‐induced septic models [32, 33, 34], Slc40a1/FPN was downregulated in the mouse hippocampal neurons after CLP. Slc40a1 knockout or knockdown mice manifested worse cognitive performance in the MWM test [28], our data also emphasized that its neuron‐specific ablation impaired hippocampus‐dependent memory in the NOR and MWM tests. In line with earlier evidence that AAV‐mediated Slc40a1 overexpression partially alleviated memory impairment in the AD mouse model [28], we observed more excellent learning and memory performance of septic mice receiving neuron‐specific Slc40a1 overexpression, as evidenced by higher preference for novel objects, reduced escape latency, prolonged duration in the target quadrant, and increased platform crossing times. Slc40a1 inhibition augmented the levels of ROS, MDA, and iron in mouse preosteoblastic cells, whereas Slc40a1 overexpression had the opposite effects [35], coinciding with our results.

The Slc40a1 gene is governed by a bidirectional transcriptional network. Phosphorylated ETS1 had a potent affinity for the Slc40a1 promoter and enhanced its transcription in the brains of senile mice with dementia [36]. CREB‐1 and HLF activated human SLC40A1 transcription by interacting with the promoter regions of SLC40A1 near c.‐662C>T and c.‐8C>G, respectively [37]. NF‐κB coordinated HDAC1 and HDAC3 to repress Slc40a1 transcription in macrophages upon inflammation by binding to the antioxidant response element (ARE) located in the Slc40a1 promoter [38]. COUP‐TF1 bound to the promoter regions of SLC40A1 near c.‐98G>C and functioned as a repressor [37]. We found that c‐MAF recognized the sequence from −1300 bp to −601 bp upstream of the TSS within the murine Slc40a1 gene.

The genetic locus containing the MAF gene was implicated as a clinical AD risk locus [39], suggesting a potential involvement of c‐Maf in cognitive function. Our NOR results showed that septic mice receiving preemptive AAV9‐CaMKII‐c‐Maf preferred the novel object and experienced improved cognitive performance in the MWM tasks. Cataract‐causing variant c.177dupC in c‐MAF triggered excessive ROS generation and mitochondrial‐dependent apoptosis by transcriptional inhibition of crystallin genes [40]. We also demonstrated that neuron‐specific c‐Maf overexpression attenuated ROS accumulation, lipid peroxidation, iron overload, and cognitive decline, which was abrogated by Slc40a1 depletion.

This study has several limitations. First, AAV‐mediated gene manipulation was performed prior to CLP because viral vectors require sufficient time for gene expression. This makes post‐CLP administration technically infeasible within the acute sepsis window. Consequently, the AAV experiments primarily model preventive rather than therapeutic intervention. Although NOR and MWM tests were performed during a recovery period commonly used in CLP models, we were unable to include additional parameters—such as body weight, rectal temperature, and sepsis severity scores (MSS)—to further control for potential behavioral confounders. Therefore, we cannot fully exclude the possibility that general sickness behavior or subtle locomotor alterations may contribute to the observed cognitive deficits. Future studies incorporating open‐field or rotarod assessments, as well as systematic monitoring of sepsis severity, will help more precisely distinguish cognitive impairment from motor deficits or nonspecific illness‐related effects. Despite PCR validation of neuronal Slc40a1 and c‐Maf expression, potential off‐target expression in non‐neuronal cell types cannot be fully excluded. Therefore, additional immunofluorescence analyses will be needed to more accurately characterize AAV‐mediated expression across hippocampal cell types. While promoter truncation assays and ChIP‐qPCR support direct transcriptional regulation of Slc40a1 by c‐MAF, site‐directed mutagenesis of putative MARE motifs in the Slc40a1 promoter could not be conducted due to the lack of a mouse c‐MAF‐matched motif in the JASPAR database. Future studies employing expanded motif resources will be required to precisely map the relevant regulatory sequences.

In conclusion, suppressing ferroptosis represents a promising approach to preventing SAE. Our findings that c‐MAF‐mediated upregulation of Slc40a1 reduces hippocampal sensitivity to ferroptosis highlight c‐MAF as a prospective prevention target in response to SAE.

Author Contributions

Wei Wang and Wenwei Gao designed and directed the study. Wenqin Song, Qianni Shen, and Hui Zhang performed experiments, wrote, and prepared the manuscript. Xueshan Bu performed behavior experiments. All authors provided critical feedback and helped to shape the manuscript.

Funding

This research was supported by the National Natural Science Foundation of China (No. 82401400); the Natural Science Foundation of Hubei Province (No. 2023AFB714); and the Hubei Province Health and Family Planning Scientific Research Project (No. WJ2023F024).

Ethics Statement

The study was reviewed and approved by the Medical Faculty Ethics Committee of Renmin Hospital of Wuhan University (ethics approval number: WDRM‐20241005B).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: RT‐qPCR assay for Slc40a1 mRNA in the hippocampal neurons of mice following the indicated treatments (n = 5). **p < 0.01, one‐way ANOVA test.

Figure S2: Densitometric quantification of WB in Figure 2H (n = 6). **p < 0.01, one‐way ANOVA test.

Figure S3: RT‐qPCR assay for Slc40a1 mRNA in the hippocampal neurons of mice following the indicated treatments (n = 5). **p < 0.01, one‐way ANOVA test.

Figure S4: Densitometric quantification of WB in Figure 3H (n = 6). **p < 0.01, one‐way ANOVA test.

Figure S5: Human brain single cell RNA‐seq analyses of seven candidate transcription factors (MAFF, MAFK, IRF7, SOX17, GATA2, RORC, and c‐MAF) based on the Alzdata database (http://www.alzdata.org/).

Figure S6: RT‐qPCR assay for c‐Maf mRNA in the hippocampus of mice after CLP (n = 6). The primer sequences for c‐Maf were 5′‐AGCAGTTGGTGACCATGTCG‐3′ (sense) and 5′‐TCCTGCTTGAGGTGGTCTACCT‐3′. **p < 0.01, unpaired t‐test.

Figure S7: RT‐qPCR assay for c‐Maf mRNA in the hippocampal neurons of mice following the indicated treatments (n = 5). **p < 0.01, one‐way ANOVA test.

Figure S8: Densitometric quantification of WB in Figure 5H,I (n = 6). **p < 0.01, one‐way ANOVA test.

Figure S9: RT‐qPCR assay for c‐Maf mRNA in the hippocampal neurons of mice following the indicated treatments (n = 5). **p < 0.01, one‐way ANOVA test.

Figure S10: Densitometric quantification of WB in Figure 6H,I (n = 6). **p < 0.01, one‐way ANOVA test.

Figure S11: Densitometric quantification of WB in Figure 7H (n = 6). **p < 0.01, one‐way ANOVA test.

Data Availability Statement

The data reported in this paper have been deposited in the OMIX, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (https://ngdc.cncb.ac.cn/omix: accession no. OMIX013287). The data that support the findings of this study are available from the corresponding author upon reasonable request.