Abstract Excitotoxicity is a prevalent pathological event in neurodegenerative diseases. The involvement of ferroptosis in the pathogenesis of excitotoxicity remains elusive. Transcriptome analysis has revealed that cytoplasmic reduced nicotinamide adenine dinucleotide phosphate (NADPH) levels are associated with susceptibility to ferroptosis-inducing compounds. Here we show that exogenous NADPH, besides being reductant, interacts with N-myristoyltransferase 2 (NMT2) and upregulates the N-myristoylated ferroptosis suppressor protein 1 (FSP1). NADPH increases membrane-localized FSP1 and strengthens resistance to ferroptosis. Arg-291 of NMT2 is critical for the NADPH-NMT2-FSP1 axis-mediated suppression of ferroptosis. This study suggests that NMT2 plays a pivotal role by bridging NADPH levels and neuronal susceptibility to ferroptosis. We propose a mechanism by which the NADPH regulates N-myristoylation, which has important implications for ferroptosis and disease treatment. Keywords: Neurodegenerative diseases, Ferroptosis, Excitotoxicity, NADPH, FSP1, Myristoylation 1. Introduction Neurodegenerative conditions are typified by the occurrence of lipid peroxidation and disrupted iron homeostasis [[43][1], [44][2], [45][3], [46][4], [47][5], [48][6], [49][7], [50][8], [51][9], [52][10], [53][11]]. Investigations employing iron chelating agents and lipid peroxide trapping agents provide initial confirmation of the involvement of ferroptosis in neurodegeneration [[54][12], [55][13], [56][14]]. Therapeutic strategies primarily focus on reducing iron overload. Although these chelating agents demonstrate effectiveness in vitro, they are constrained in clinical treatment by their insufficient ability to cross the blood-brain barrier [[57]14,[58]15]. The protracted clinical utilization of iron-chelating agents, for instance, deferoxamine (DFO), may engender systemic metal ion depletion, resulting in anemia and associated complications [[59]16,[60]17]. The clinical administration of iron chelators for curing neurological disorders necessitates stringent regulation to avert unfavorable outcomes [[61]18]. Advanced alternative therapeutics will be a hopeful avenue for the treatment of neurological impairments such as ischemic damage and neurodegenerative conditions. Excitotoxicity, which happens in stroke and neurodegenerative disorders, is widely acknowledged to be an oxidative and iron-dependent process [[62]19,[63]20]. High doses of glutamate block cystine input by inhibiting system x[c]^− cystine/glutamate antiporter activity [[64]19,[65]21]. Nevertheless, apart from the function of glutamate per se, the effect of activating glutamate receptors in ferroptosis remains incompletely understood. In order to attain a comprehensive understanding, it is imperative to differentiate between the distinct functions of glutamate as a metabolite and the excitotoxic cascade signal in the process of ferroptosis. Based on transcriptome analysis, NADPH abundance can serve as a predictive factor for sensitivity to ferroptosis inducers [[66]22]. In addition to facilitating the biosynthesis of mevalonate and fatty acids, NADPH is responsible for maintaining the antioxidant system which relies on glutathione (GSH) and thioredoxin. Both processes intricately play a crucial role in regulating ferroptosis [[67][23], [68][24], [69][25], [70][26]]. HD domain containing 3 (HDDC3/MESH1) is a widely expressed cytoplasmic NADPH phosphatase [[71]27]. Administrating Erastin or depriving cysteine upregulates MESH1, decreases NADPH and renders cells susceptible to ferroptosis. Conversely, MESH1 knockout shields cells against lipid peroxidation by preserving NADPH [[72]27,[73]28]. NAD kinase silencing decreases NADPH and aggravates erastin-, FIN56- and RSL3-induced ferroptosis [[74]27]. Overall, these advances elucidate the importance of regulating cytoplasmic NADPH when ferroptosis occurs. Until now, the correlation between NADPH and ferroptosis has been based on indirect interventions, such as theoretical deductions considering its role as a reducing flux [[75]29]. To date, no investigation has been conducted into the effect and mechanism of direct NADPH supplementation in ferroptosis. NADPH is well known for its unreplaced role in electron transfer, metabolism and redox homeostasis. Little is known about its effect in signaling transduction, especially protein modification. Here we show that exogenous NADPH, besides being reductant, interacts with N-myristoyltransferase 2 (NMT2) and upregulates the N-myristoylated ferroptosis suppressor protein 1 (FSP1). It proposes a mechanism by which the NADPH regulates N-myristoylation, which has important implications for ferroptosis and disease treatment, such as cancers and neurodegeneration diseases. Historically, it was widely accepted that NADPH must be generated endogenously, as it was thought to be inaccessible from exogenous sources [[76]25]. The observed transportation of NADPH through the P2X purinoceptor 7 receptor (P2X7R) challenges the existing paradigm, which has yet to be published. This discovery establishes exogenous NADPH supplementation as a dependable treatment for enhancing ferroptosis resistance. This study aims to investigate the effectiveness and mechanisms of NADPH as a therapeutic intervention for preventing ferroptosis in neurodegenerative neurons. 2. Materials and methods 2.1. Bioinformatics analysis The dataset [77]GSE111434 was obtained from the Gene Expression Omnibus (GEO) in MINiML format. Differences in mRNA expression were analyzed with R software's limma package (version: 3.40.2). Corrected P values were calculated in GEO to account for false-positive results. "Adjusted P < 0.05 and absolute value of log[2] (fold change) > 1" were considered significant mRNA expression differences. The Cluster Profiler package in R software was used to analyze Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment. We downloaded the ferroptosis-related gene dataset from the FerrDb website ([78]http://www.zhounan.org/ferrdb/) and utilized an interactive online tool to calculate the intersection between differentially expressed genes (DEGs) and this dataset in order to obtain the DEGs that are induced by KA-mediated ferroptosis. The expression heat map of 10 up-regulated genes and 10 down-regulated genes was presented using the R package pheatmap. 2.2. Cell culture and treatment The cortex of 18-day-old Institute of Cancer Research (ICR) mouse embryos was harvested for preparation of primary cortical neurons. The cortical tissue was subjected to enzymatic digestion utilizing 0.25 % trypsin until a flocculent suspension was achieved. Subsequently, complete medium was introduced to terminate the enzymatic reaction. Addition of DNase was performed, and the cells were subjected to gentle agitation and mechanical disruption. The pellet was resuspended after centrifugation at 1500 rpm for 5 min. Cells was passed through a cell filter with a pore size of 40 μm and diluted to 1 × 10^6 cells/mL. As described previously, neural cultures were cultured for a duration of 6–8 days in vitro for most experiments. 2.3. Cell viability assay We measured the viability of neurons with a Cell Counting Kit (CCK). After treatment, detection reagent was added into wells and incubated at 37 °C. A 450 nm absorbance measurement was performed. There were three independent blank wells in this experiment. 2.4. Animal treatment and drug administration ICR mice (male, 25–30 g, 6–8 weeks old) were provided by Zhaoyan (Suzhou) New Drug Research Center. The mice were housed under standard environmental conditions, including regulated temperatures, humidity, and light-dark cycles, and were provided with unrestricted access to food and water. The use of animals complied with institutional animal healthcare regulations. We secured the necessary authorization for all animal-based experiments through the Institutional Animal Care and Use Committee of Soochow University. We dissolved 0.625 nmol KA (K0250, Sigma Aldrich) (Pubchem CID:10255) in 1 μL of normal saline and infused it into the right striatum via a glass micropipette over 2 min. Then, the needle was allowed to remain in situ for another 5 min. The injection coordinates were consistent with a previous report [[79]10]. NADPH (Pubchem CID: 5884) was administered as previously described [[80]10,[81]28]. 2.5. Transmission electron microscopy (TEM) Mice were anesthetized and perfused with precooled 1 × phosphate buffered saline (PBS) and 4 % paraformaldehyde after 48 h of KA treatment. We sectioned the striatum into 1 mm^3 pieces and fixed them in 2.5 % glutaraldehyde. Subsequently to paraffin sectioning, the obtained samples were subjected to observation through TEM imaging for analysis. Five images per mouse were selected for mitochondrial area quantification, and the number of mitochondria used for statistical analysis in each group exceeded 90. 2.6. BODIPY 581/591C11 analysis Primary neurons or HT22 cells were rinsed with 1 × PBS and incubated with BODIPY 581/591C11 for 10 min at 37 °C following treatment. Upon undergoing a redox reaction with reactive oxygen species in the cell membrane, the unsaturated butadiene moiety of BODIPY 581/591C11 exhibited a redshift in fluorescence emission peak from 590 nm to 510 nm. Flow cytometry was employed to analyze the resultant fluorescence. Following the afore mentioned administration, primary neurons or HT22 cells underwent cotreatment with BODIPY 581/591C11 and Hoechst 33342. Subsequently to the washing procedure, the cells were subjected to examination under a confocal microscope. Quantification was executed according to procedures previously described [[82]32]. 2.7. NADPH assay The cellular specimens were subjected to a double washing procedure and subsequently, 200 μL NADPH extraction solution was introduced. Gentle agitation was utilized to promote cell lysis, followed by centrifugation at 12000g at 4 °C for 10 min. Supernatant was collected and subjected to heating at 60 °C for 30 min to decompose nicotinamide adenine dinucleotide phosphate (NADP^+). The resultant mixture was then centrifuged at 10000 g at 4 °C for 5 min, with the supernatant being retained for NADPH detection. A 50 μL sample was then combined with 100 μL glucose-6-phosphate dehydrogenase (G6PDH) working solution and incubated at 37 °C for 10 min. Subsequently, 10 μL coloring solution was introduced to each well and incubated at 37 °C for 10–20 min. The absorbance was measured at 450 nm. The detection outcomes were expressed as NADPH content per unit weight of protein. 2.8. Western blot analysis Western blotting was performed as previously described. The following primary antibodies were used: anti-β-actin (A5441, Sigma Aldrich), anti-GAPDH (ab37168, Abcam), anti-glutathione peroxidase 4 (GPX4) (ab125066, Abcam), anti-ferroptosis suppressor protein 1 (FSP1) (39498, SAB), anti-Na^+/K^+-ATPase (05–369, Millipore), and anti-NMT2 (ABclonal, A7042). 2.9. Intracellular Fe^2+ determination We utilized a fluorescent probe named FerroOrange (F374, Dojindo) to evaluate intracellular Fe^2+. After incubation with 100 μM kA for 8 h, neurons were stained with FerroOrange (1 μM) for 30 min and counterstained with Hoechst 33342 for another 10 min. Fluorescence imaging of primary neurons was observed under a confocal microscope. 2.10. Iron assay As directed by the manufacturer, striatal iron contents (Fe^2+ and Fe^3+) were detected using an iron assay colorimetric kit (I291, Dojindo). The striatum of mice was collected and subsequently homogenized with 350 μL assay buffer. The mixture that ensued was exposed to ultrasonic homogenization and centrifuged at 16,000 g for 10 min at 4 °C. Subsequently, 100 μL the supernatant was combined with 5 μL assay buffer for the quantification of divalent iron, while another 100 μL supernatant was mixed with 5 μL reducer solution for the measurement of total iron. Following this, all samples were incubated in 37 °C for 15 min. Mixed 100 μL sample with 100 μL probe solution and incubated at 37 °C for 1 h. Measured the absorbance at 593 nm and calculated the content based on the standard curve. 2.11. FSP1 activity assay The FSP1 activity assay was performed as previously described [[83]33]. A 100 μL PBS system containing 500 μM (nicotinamide adenine dinucleotide) NADH, 200 μM Coenzyme Q (CoQ), and purified FSP1 was prepared. The activity of FSP1 responds to the rate of NADH consumption, which is detectable through its absorption at a wavelength of 340 nm. To establish a normalized baseline, it is recommended to include both NADH-free and enzyme-free control groups in the analysis. 2.12. Immunofluorescence Preparation brain sections was as previously described [[84][30], [85][31]]. Primary antibodies were diluted in 1 % bovine serum albumin (BSA) solution, applied overnight at 4 °C, and washed three times with PBST (PBS solution containing 0.2 % Tween-20). Then, the sections were incubated with secondary antibodies for 3 h at room temperature. After washing three times with PBST, the nuclei were stained utilizing 4',6-diamidino-2-phenylindole (DAPI) for 15 min. We used a laser scanning confocal microscope to acquire images. The primary antibodies used were as follows: anti-FSP1 (39498, SAB) and anti-Na^+/K^+-ATPase (05–369, Millipore). 2.13. Plasma membrane fractionation We used a membrane and cytosol protein extraction kit (P0033, Beyotime) to extract the plasma membrane. The protocol was carried out in accordance with the manufacturer's instructions. 2.14. Labelling of N-myristoylated proteins Primary neurons were incubated with Click-IT® Myristic Acid Azide (50 μM, [86]C10268, Thermo Fisher Scientific) in the dark for 5 h to label N-myristoylated proteins. For details of usage, refer to the manufacturer's instructions. 2.15. Click chemistry After labeling N-myristoylated proteins, primary neurons were fixed with 4 % paraformaldehyde and permeabilized with PBS containing 0.05 % Triton X-100, followed by washing in PBS containing 2 % BSA. Then, they were reacted with Alexa Fluor® 594 alkyne (A10275, Thermo Fisher Scientific) by using a Click-iT® Cell Reaction Buffer Kit ([87]C10269, Thermo Fisher Scientific). The cells were blocked with 5 % BSA, washed with 0.2 % PBST, and incubated overnight at 4 °C with anti-FSP1. After secondary antibody incubation and DAPI staining, images were taken by confocal fluorescence microscopy to detect N-myristoylation of FSP1. 2.16. siRNA, plasmid and transfection The plasmids expressing mouse wildtype or mutated NMT2 were procured from Genomeditech (Shanghai, China). Cells were transfected with plasmids or small interfering RNA (siRNA) by using Lipofectamine 3000 according to the manufacturer's instructions. The sequence of NMT2-siRNA is as follows: 5′-GCACUAGAUUUGAUGGAAATTUUUCCAUCAAAUCUAGUGCTT-3′ (1), 5′-CAUAAGAAACUGAGAUCAATTUUGAUCUCAGUUUCUUAUGTT-3′ (2) and 5′-UCUACAACAUUCACACAGATTUCUGUGUGAAUGUUGUAGATT-3′ (3). 2.17. Molecular docking The study employed the CDOCKER module of the Discovery Studio 2016 software to conduct docking of NADPH and NMT2 active sites. The NADPH and NMT2 conformations were imported and adjusted, encompassing hydrogenation, CHARMM force field, and Momany-Rone charges, and the structures were optimized accordingly. The software recognized the active sites and carried out CDOCKER docking studies. Subsequently, the best structure was identified through analysis of the CDOCKER interaction energy. 2.18. Affinity chromatography of GST-tagged proteins Plasmids were transformed into Escherichia coli cells. Protein expression was induced with 0.1 mM IPTG and incubated with shaking at 120 rpm for 48 h. The bacteria were disrupted by ultrasonication, and the collected supernatant was subjected to GST affinity chromatography. Finally, the GST-NMT2 fusion protein was eluted using a GSH solution. 2.19. 2′, 5′-ADP pull down The protein and ADP resin were incubated at 4 °C for 2 h, followed by three washes with binding buffer. 20 μL of 2 × loading buffer was added, heated, and the supernatant was collected for western blot analysis. 2.20. NMT activity assay A 100 μL PBS system containing purified GST-NMT2, 1 mM NADPH, 4 μM myristoyl CoA, and 40 μM CPM was added to the 96-well plate. A 20 μM substrate peptide was added and the fluorescence intensity was measured at 470 nm emission wavelength upon excitation at 362 nm. 2.21. Statistical analysis This study employed the one-way analysis of variance (ANOVA) with Tukey's test and Student's t-test for comparing multiple groups and two groups, respectively. The data were presented as the mean ± standard error of the mean (SEM). A P value of ≤0.05 was deemed statistically significant. 3. Results 3.1. Exogenous NADPH supplementation alleviates KA-induced ferroptosis by inhibiting lipid peroxidation The role of ferroptosis in the neuronal excitotoxicity has not yet been well characterized. The GEO dataset ([88]GSE111434) is reanalyzed to investigate the effects of KA treatment on primary mouse neurons [[89]34]. Pathway enrichment analysis reveals that KA significantly perturbs ferroptosis and some related regulatory pathways ([90]Fig. 1A), including iron-sulfur cluster assembly [[91]35], cysteine and methionine metabolism [[92]36], iron uptake and transport [[93]37], AMP-activated protein kinase (AMPK) signaling pathway [[94]38], intracellular iron homeostasis [[95]37], nicotinate and nicotinamide [[96]22], oxidative phosphorylation [[97]39], hypoxia-inducible factor-1 (HIF-1) signaling pathway [[98]40], biosynthesis of unsaturated fatty acid [[99]41], response to oxidative stress [[100]42], mitophagy [[101]43], glutathione metabolism [[102]44] and cholesterol metabolism [[103]45]. The identified DEGs are compared with the 503 ferroptosis-related genes, as listed in the FerrDb database, among which 237 show significant changes. Further classifying them into 129 ferroptosis drivers and 108 ferroptosis inhibitors using the FerrDb online tool, some classic regulatory genes are shown in the figure ([104]Fig. 1B). This indicates that the neuronal excitotoxicity induced by KA elicites noteworthy perturbations in ferroptosis. Fig. 1. [105]Fig. 1 [106]Open in a new tab KA-induced excitotoxicity triggered ferroptosis. (A) RNA sequencing data from primary cortical neurons treated with KA (100 μM, 8 h) were collected from the dataset [107]GSE111434. Pathway enrichment analysis was performed on the differentially expressed genes (DEGs) to confirm their potential function, with color indicating the significance of enrichment results and the bar chart length indicating the enrichment score. (B) The identified DEGs were compared with the ferroptosis dataset from the FerrDb database, and 237 differentially expressed genes related to ferroptosis were identified. The heat map of differentially expressed genes shows 10 upregulated genes and 10 downregulated genes, with color indicating gene expression levels. (C) TEM images of the perinuclear region of striatal neurons from KA (0.625 nmol, 48 h)-treated mice (n = 3). Images c-d are representative areas indicated by the red wireframes in a-b. Images e-f show representative mitochondria indicated by red arrows in c-d. a-b, scale bar = 5 μm; c-d, scale bar = 2 μm; e-f, scale bar = 500 nm. (D) Mitochondrial area frequency in the perinuclear region. More than 90 mitochondria were calculated from 5 pictures of 3 mice per group. (E) The frequency of mitochondrial area in each interval was counted. **P < 0.01 vs control. (For interpretation of the references to color