Mechanical stress enhances tumor cell ferroptosis by remodeling succinate metabolism

Mechanical stress enhances tumor cell ferroptosis by remodeling succinate metabolism | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Mechanical stress enhances tumor cell ferroptosis by remodeling succinate metabolism Jun Ma, Zhe Li, Nuo-zhou Liu, Rong Wang, Tingxiang He, Ying Qi, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9481973/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The mechanical properties of the tumor microenvironment serve as crucial physical cues that shape cell fate decisions. However, whether microenvironmental mechanical forces modulate ferroptosis to drive radioresistance remains unclear. Here, using PDMS-based hydrogels with tunable stiffness to establish tumor cell culture systems, we found that tumor cells cultured on stiff substrates were more sensitive to both ferroptosis and radiotherapy. Mechanistically, tumor cells grown on stiff matrices showed increased microtubule acetylation and ER sheet-to-tubule remodeling, which increased the formation of mitochondria-associated membranes (MAMs) and led to mitochondrial succinate accumulation. Succinate in turn promoted CPT1A-mediated succinylation of ACSL3, facilitating its degradation and thereby enhancing tumor cell sensitivity to ferroptosis. Collectively, these findings identify MAMs as intracellular mechanosensitive structures that regulate mitochondrial metabolism and ferroptosis in tumor cells, providing new insights into the mechanical control of ferroptosis and its implications for tumor radioresistance. Biological sciences/Chemical biology/Lipids/Phospholipids Biological sciences/Biochemistry/Metabolomics Biological sciences/Cell biology/Cell death Health sciences/Diseases/Cancer/Cancer therapy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Highlights Tumor microenvironmental stiffness remodels mitochondria-associated membranes (MAMs) to reprogram mitochondrial metabolism MAMs expansion drives succinate accumulation through impaired SDH activity, positioning succinate as a mechanosensitive metabolite Succinate fuels CPT1A-mediated succinylation of ACSL3 at K422, promoting its degradation and ferroptosis sensitivity Succinate enrichment signature and lower ACSL3 expression predicts better clinical responses to radiotherapy Introduction The mechanical properties of the tumor microenvironment have emerged as active determinants of cancer progression and therapeutic resistance 1 . Increased extracellular matrix (ECM) stiffness, a hallmark of the desmoplastic tumor stroma, promotes tumor initiation, invasion, immune evasion and chemoresistance through mechanotransduction pathways that reprogram gene expression, metabolism and cell fate decisions 2 . Despite growing appreciation that ECM mechanics influence virtually every aspect of tumor biology, whether and how microenvironmental mechanical forces (MMF) regulate specific modes of regulated cell death to shape radiotherapy response remains poorly understood. Ferroptosis, an iron-dependent form of cell death driven by the peroxidation of polyunsaturated fatty acid-containing phospholipids (PUFA-PLs), has been increasingly recognized as a critical effector of radiotherapy-induced tumor suppression 3-5 . The sensitivity of tumor cells to ferroptosis is governed by the balance between PUFA-PLs abundance in cellular membranes and the cellular antioxidant capacity centered on the GSH-GPX4 axis 6 . Importantly, phospholipid remodeling enzymes, including ACSL4, which channels polyunsaturated fatty acids into membrane phospholipids, and ACSL3, which counteracts this process by incorporating monounsaturated fatty acids, critically determine the ferroptosis-prone or resistant lipid landscape of tumor cells 7, 8 . However, whether extracellular mechanical cues can reshape this phospholipid landscape to alter ferroptosis susceptibility has not been explored. The mitochondria-associated membranes (MAMs) are regions where the ER membrane closely approaches the outer mitochondrial membrane, and these areas are crucial for calcium ion transport, phospholipid transfer, and cell death 9 . The key coupling proteins IP3R and σ1R within MAMs enhance calcium ion exchange between the ER and mitochondria, leading to increased mitochondrial lipid peroxidation and subsequently promoting ferroptosis 10 . Notably, MAMs formation depends on ER tubular morphology maintained by acetylated microtubules 11, 12 , a cytoskeletal feature known to respond to substrate rigidity, raising the possibility that MMF may regulate ferroptosis by remodeling MAMs and reprogramming mitochondrial metabolism. Here, we reveal that the mechanical properties of the tumor microenvironment converge on MAMs to govern ferroptosis sensitivity and radiotherapy response. Using tunable PDMS-based substrates, we show that stiff matrices drive microtubule acetylation, ER sheet-to-tubule remodeling, and MAM expansion, which in turn impairs SDH activity and causes mitochondrial succinate accumulation. Succinate then serves as a metabolic encoder of mechanical information: it fuels CPT1A-mediated succinylation of ACSL3 at K422, triggering ACSL3 degradation and shifting membrane phospholipid composition toward a ferroptosis-prone state. These findings establish MAMs as mechanosensitive organelle interfaces that couple extracellular biomechanics to mitochondrial metabolic reprogramming, and identify the succinate-ACSL3 succinylation axis as a druggable node linking tumor mechanics to radiotherapy efficacy. Results High matrix stiffness enhances radiosensitivity by promoting ferroptosis Radiotherapy remains a cornerstone of tumor treatment, yet its efficacy is frequently undermined by the persistence of radioresistant tumor cell populations within heterogeneous microenvironments. Nasopharyngeal carcinoma (NPC), a malignancy treated primarily with radiotherapy owing to its anatomical location and intrinsic radiosensitivity, represents a paradigmatic model for studying radioresistance: despite high initial response rates, local recurrence driven by radioresistant populations remains a major clinical challenge 13 . We hypothesized that these hallmarks-radioresistance are linked to the MMF and ferroptosis susceptibility of tumor cells. Collagen is the most abundant structural protein in the ECM and is a major determinant of tissue tensile strength 14 , we performed immunofluorescence staining and for Collagen I 15 andnanoindenter measurements in pretreatment biopsy specimens from patients with recurrent NPC (rNPC) and non-recurrent NPC (nrNPC). We observed that the dense fibrotic stromal collagen area of NPC exhibited higher stiffness than the low collagen expressing area (Fig.1a). Meanwhile, significant lower expression of Collagen I was observed in rNPC (Fig.1b), suggesting a potential association between matrix stiffness and radioresistance. To further determine whether MMF regulates the efficacy of tumor radiotherapy, we established an in vitro experimental model using custom-fabricated polydimethylsiloxane (PDMS) gels with defined stiffness levels of approximately 1~1000 kPa (Fig.1c). Using the NPC cell lines HK1 and C666-1, we found that cells on the high (500 kPa) stiffness PDMS gel showed highest radiotherapy sensitivity, and the intermediate and lowest radiotherapy sensitivity occured on modest stiffness (50 kPa) and low stiffness (5 kPa) PDMS matrix (Fig.1d-e). We treated cells with inhibitors targeting distinct cell death pathways, including necrostatin-1 (necroptosis), Z-VAD-FMK (apoptosis), and ferroptosis inhibitors (ferrostatin-1 [Fer-1] and deferoxamine [DFO]), and found that only ferroptosis inhibitors significantly reversed the mechanical force-mediated regulation of radiosensitivity (Fig.1f-g, Extended Data Fig.1a). Interestingly, HKl or C666-1 cells cultured on stiffer matrix showed markedly increased di-oxygenated and tri-oxygenated PUFA-PE species as well as enhanced ferroptosis (Fig.1h-n, Extended Data Fig.1b-e). Consistently, we evaluated the sensitivity of HT-1080, 786-O, A375 and MDA-MB-231 cells toward multiple ferroptosis inducers based on different mechanisms, and also found that cells cultured on stiffer matrix exhibited enhanced ferroptosis sensitivity (Extended Data Fig.1f). Next, we sought to validate these findings in vivo . LOX, a copper-dependent amine oxidase, catalyzes the intra- and intermolecular covalent crosslinking of collagen, thereby contributing to tissue stiffening; conversely, reduced LOX activity has been shown to alleviate tissue stiffness and prevent fibrosis 15, 16 . We therefore established HK1 cells with stable LOX overexpression and generated subcutaneous xenograft tumors in BALB/c nu/nu mice. Nanoindenter measurements demonstrated that LOX-overexpressing HK1 tumors displayed greater tissue stiffness than control tumors (Fig. 1O). Notably, after irradiation, LOX-overexpressing tumors were significantly smaller than wild-type tumors. However, treatment with the ferroptosis inhibitor Liproxstatin-1 abrogated this effect (Fig. 1P). Collectively, these results indicate that MMF can enhance tumor cell radiosensitivity and the effect is associated with increased susceptibility to ferroptosis. Mechanical stiffness reshapes the lipid landscape of tumor cells Next, we investigated why increased matrix stiffness promotes ferroptosis in tumor cells. The phospholipid composition and its redox state critically dictate ferroptosis susceptibility 17 . We performed RNA-seq to profile HK1 cells cultured on matrices of varying stiffness and found no significant changes in ferroptosis-related gene expression (Extended Data Fig. 2a). Consistently, the abundance of proteins involved in lipid redox homeostasis, as well as intracellular GSH and Fe 2+ levels were comparable across stiffness conditions (Extended Data Fig. 2b-d). However, among phospholipid-remodeling factors, ACSL3 was the only protein that displayed stiffness-dependent regulation, with its expression decreasing as matrix stiffness was reduced (Fig. 2a). ACSL3 promotes the activation of monounsaturated fatty acids (MUFAs), such as oleic acid (OA), into their corresponding acyl-CoA species, thereby facilitating their incorporation into membrane phospholipids. This MUFA-driven remodeling reduces the relative abundance of polyunsaturated fatty acid-containing phospholipids (PUFA-PLs), a key determinant of lipid peroxidation propensity and ferroptosis susceptibility (Extended Data Fig. 2e). We next profiled the phospholipid landscape of tumor cells cultured on matrices of different stiffness. Using LC-MS based lipidomic analysis, we found that tumor cells grown on soft matrices displayed a marked reduction in PUFA-PLs compared with those cultured on stiff matrices (Fig. 2b). Notably, this decrease was most prominent in arachidonic acid (AA; C20:4) and adrenic acid (AdA; C22:4) containing phosphatidylethanolamine (PE) and phosphatidylcholine (PC) species, which are strongly linked to ferroptosis execution (Fig. 2c). In addition, we performed airflow-assisted desorption electrospray ionization mass spectrometry imaging (AFADESI-MSI) on pretreatment biopsy specimens from patients with rNPC and nrNPC. Compared with nrNPC, rNPC tumor tissues exhibited a significant reduction in PUFA-containing phospholipids (PUFA-PLs) (Fig. 2d). To validate the contribution of ACSL3 to MMF-regulated ferroptosis sensitivity, we knockouted ACSL3 in HK1, C666-1, and HT1080 cells and cultured these cells on matrices of different stiffness (Extended Data Fig. 2g). Upon treatment with ferroptosis inducers, ACSL3 knockout markedly abrogated the ferroptosis-resistant phenotype observed in cells grown on soft substrates, thereby restoring ferroptosis sensitivity under low-stiffness conditions (Fig. 2e-h, Extended Data Fig. 2h-i). In addition, ACSL3 knockout also markedly reversed the radioresistant phenotype of HK1 cells cultured on soft matrices (Fig. 2i). Collectively, these results indicate that MMF may modulate tumor cell sensitivity to ferroptosis and radiotherapy by altering ACSL3 expression. MMF regulates ferroptosis through MAMs Next, we investigated how MMF regulates ACSL3 protein abundance. We previously showed that ACSL3 mRNA levels were unchanged across stiffness conditions, indicating that MMF did not regulate ACSL3 at the transcriptional level (Extended Data Fig. 2a). Interestingly, we found that ACSL3 showed a decreased half-life in HK1 or C666-1 cells cultured on stiffer matrix (Fig. 3a, Extended Data Fig. 3a). More importantly, the degradation of ACSL3 in HK1 cells cultured on stiff matrix could be blocked by proteasome inhibitors (MG132; carfilzomib) but not lysosome inhibitors (Baf A1; chloroquine), indicating that MMF regulates ACSL3 protein levels depending on proteasome (Fig. 3b). Ubiquitination is a key step for the proteasome-dependent degradation of protein. Consistently, we found that the ubiquitination level of ACSL3 was increased in the cells cultured on stiffer matrix (Fig. 3c, Extended Data Fig. 3b). ACSL3 is primarily localized to the endoplasmic reticulum (ER), lipid droplets (LD) and MAMs 18 . Recent studies have highlighted an essential role for MAMs in ferroptosis execution 10, 19 . We therefore investigated whether MAMs contribute to the MMF-dependent regulation of ACSL3 expression. We observed under electron microscopy that HK1 cells cultured on soft matrix exhibited fewer MAM structures compared to those grown on stiff matrix (Fig. 3d). Additionally, we generated HK1 cells stably expressing mitochondria-targeted GFP (Mito-GFP) and the ER marker Sec61β-mCherry, and cultured them on matrices of defined stiffness. Immunofluorescence analysis showed reduced apparent co-localization between Mito-GFP and Sec61β-mCherry in cells grown on soft substrates, indicating decreased ER-mitochondria proximity under low-stiffness conditions (Fig. 3e). Notably, we stably transfected the HK1 cell line with plasmids encoding GFP1-10-ERT (targeting the ER membrane) and GFP11-TOMM70 (targeting the outer mitochondrial membrane) to construct MAM-GFP reporter cells 20, 21 (Extended Data Fig. 3c). Flow cytometry and immunofluorescence analysis of GFP expression in HK1 MAM-GFP reporter cells revealed that tumor cells cultured on soft matrix formed fewer MAM structures (Fig. 3f-g). These data indicate that the formation of MAMs is reduced in tumor cells cultured on soft matrix. We then employed a rapamycin-induced FKBP-FRB interaction system by co-transfected with CYB5A-GFP-FKBP and TOM20-mCherry-FRB in HK1 cells and C666-1 cells to form chemically inducible MAMs (MAMs-inducing cells) 22, 23 . This also allowed protein GFP and mCherry to localize to MAMs (Extended Data Fig. 3d). Interestingly, MAM-inducing cells exhibited lower ACSL3 protein abundance than parental cells (Fig. 3h). Under treatment with AA-d8, MAM-inducing cells contained more AA-d8-containing PE or PC species compared to parental cells (Fig. 3i). Upon exposure to ferroptosis inducers, MAM-inducing cells showed higher levels of di-oxygenated and tri-oxygenated PUFA-PE species and displayed markedly enhanced ferroptosis compared with their counterparts (Fig. 3j-l). Furthermore, upon radiotherapy treatment, MAM-inducing cells exhibited increased cell death compared with control cells (Fig. 3m). Previous studies have shown that the tubular ER participates in the contacts between the ER and various other organelles (lysosomes, mitochondria, etc.) 11 . Rab10 and the tubular ER-shaping protein reticulon 4 (RTN4) are key proteins that maintain the tubular ER, while Climp63 induces sheet-like ER 23 . In MAM-GFP reporter cells, knockdown of Rab10/RTN4 or overexpression of Climp63 markedly reduced MAM formation and enhanced ACSL3 protein level (Fig. 3n-p, Extended Data Fig. 3e-h). Concurrently, Rab10 or RTN4 knockdown reduced the accumulation of di-oxygenated and tri-oxygenated PUFA-PE species and markedly attenuated ferroptosis relative to control cells (Extended Data Fig. 3i-l). The Tubular ER dynamics depend on microtubules 12 , and MAMs have been reported to occur on acetylated microtubules 24 . Notably, microtubule acetylation and its acetyltransferase α-tubulin acetyltransferase 1 (αTAT1) can respond to substrate rigidity, providing a potential link between MMF and MAMs 25 . We observed a significant reduction in microtubule acetylation levels in HK1 cells and C666-1cells cultured on soft matrices (Extended Data Fig. 4a). Treatment with the microtubule deacetylase inhibitor Tubacin to induce microtubule acetylation led to a marked increase in the abundance of MAMs and reduced ACSL3 protein level (Extended Data Fig. 4b-c). Conversely, knockdown of α-tubulin acetyltransferase 1 (αTAT1), which diminishes microtubule acetylation, markedly reduced MAM formation and increased ACSL3 protein level (Extended Data Fig. 4d-f). Tubacin treatment enhanced ferroptosis and promoted the accumulation of ferroptosis-associated oxidized phospholipids, whereas αTAT1 knockdown exerted the opposite effects (Extended Data Fig. 4g-l). Collectively, these results suggest that stronger MMF may modulate ferroptosis susceptibility by driving ER remodeling and promoting MAMs formation. Mitochondrial succinate accumulation facilitates ferroptosis Next, we investigated how MMF and MAM remodeling regulate ACSL3 protein abundance. MAMs represent specialized hubs for Ca²⁺ transfer and lipid exchange between the ER and mitochondria, and alterations in MAM dynamics can influence mitochondrial fission–fusion balance as well as mitochondrial metabolism. We reasoned that MMF-driven MAM remodeling might reprogram the TCA cycle and the respiratory chain. Accordingly, we assessed mitochondrial function in tumor cells cultured on matrices of defined stiffness, as well as in MAM-inducing cells and their corresponding control cells. Seahorse analysis revealed that cells cultured on stiff matrix showed no statistically significant change in their basal respiration, whereas maximal respiratory capacity and spare respiratory capacity (SRC) were impaired (Fig. 4a-b). Consistently, MAM-inducing cells exhibited a modest reduction in basal oxygen consumption, with a pronounced decrease in maximal respiratory capacity and SRC (Extended Data Fig. 5a-b). Importantly, mitochondria-related metabolite profiling revealed marked succinate accumulation in tumor cells cultured on stiff matrices (Fig. 4c). Succinate is subsequently oxidized to fumarate by SDH, a key enzyme that links the TCA cycle to oxidative phosphorylation (Extended Data Fig. 5c). Further analysis also revealed that succinate-to-fumarate ratio was elevated both in cells grown on stiff substrates and in MAM-inducing cells, indicating a stiffness-dependent regulation of mitochondrial TCA cycle intermediates (Fig. 4d-e). Intriguingly, AFADESI-MSI on tissue specimens from 80 patients with nasopharyngeal carcinoma showed that succinate was significantly elevated in tumor cells from patients with recurrent disease (Fig. 4f). α-Ketoglutarate dehydrogenase (OGDH) catalyzes the conversion of α-KG into succinyl-CoA, which is subsequently converted into succinate by succinyl-CoA synthetase (SCS) (Extended Data Fig. 5c). Upon supplementation with AA-d8, tumor cells treated with dimethyl α-ketoglutarate (DMK) or dimethyl succinate (DES) incorporated higher levels of AA-d8 into PE and PC species than parental cells (Fig. 4g). Consistently, treatment with DMK or DES enhanced ferroptosis and radiotherapy-induced cell death, while also promoting the accumulation of ferroptosis-associated oxidized phospholipids in HK1 and C666-1 cells cultured on soft matrices (Fig. 4h-k, Extended Data Fig. 5d-h). Moreover, treatment with DMK or DES abrogated the elevated protein abundance of ACSL3, as well as its prolonged half-life and reduced ubiquitination observed in tumor cells cultured on soft matrices (Fig. 4l-n, Extended Data Fig. 5i-q). These results suggest that elevated intracellular succinate promotes ACSL3 degradation via the ubiquitin-proteasome pathway. Next, we investigated the mechanisms underlying succinate accumulation in tumor cells cultured on stiff matrices. Proteomic profiling of HK1 cells cultured on matrices of different stiffness revealed no significant changes in the abundance of succinate-associated metabolic enzymes, including OGDH, SCS, and SDH (Extended Data Fig. 6a). Consistently, western blotting showed comparable levels of the SCS subunits SUCLA2, SUCLG1, and SUCLG2, as well as OGDH and the SDH subunits SDHA and SDHB, across stiffness conditions (Extended Data Fig. 6b). Succinate in tumor cells primarily originates from the mitochondrial TCA cycle and can also be acquired from the extracellular environment. Within mitochondria, succinate is exported to the cytosol via the inner membrane transporter SLC25A10 and the outer membrane channel VDAC. In addition, tumor cells can import extracellular succinate through the plasma membrane transporter SLC13A3, and release it into the extracellular space via MCT1 (Extended Data Fig. 6c). RT-PCR analysis showed no significant differences in the expression levels of SLC25A10, SLC13A3, or MCT1 in tumor cells cultured on matrices of different stiffness (Extended Data Fig. 6d). Because the succinate-to-fumarate ratio was elevated in cells cultured on stiff substrates (Fig. 4d), we hypothesized that SDH activity might be impaired under the conditions of stiff substrates. Mitochondrial isolation followed by enzymatic activity assays showed that SDH activity was lower in tumor cells grown on stiff matrices than in those cultured on soft substrates (Extended Data Fig. 6e-f). Similarly, SDH activity was also reduced in MAM-inducing cells (Extended Data Fig. 6g-h). These findings suggest that diminished SDH activity may contribute to succinate enrichment in tumor cells cultured on stiff matrices. Treatment of HK1 or C666-1 cells with the SDH inhibitors dimethyl malonate (DMM) and 3-nitropropionic acid (3-NPA)-the latter irreversibly inactivating SDHA by covalently binding to an arginine residue within its catalytic core effectively abrogated the soft matrix-induced stabilization of ACSL3, restoring ACSL3 ubiquitination and reducing ACSL3 protein abundance (Extended Data Fig. 6i-k), while concomitantly sensitizing tumor cells to ferroptosis (Extended Data Fig. 6l-m). In contrast, pharmacological inhibition of OGDH with CPI-613 markedly reduced ACSL3 ubiquitination and attenuated ferroptosis in the same tumor cell lines (Extended Data Fig. 6j-m). These results suggest a critical role of succinate metabolism in regulating matrix stiffness-induced ferroptotic sensitivity in tumor cells. CPT1A mediates ACSL3 succinylation to promote its degradation Accumulation of succinate or α-KG can increase intracellular succinyl-CoA levels, thereby enhancing protein lysine succinylation 26 . Next, we investigated whether lysine succinylation participates in regulating ACSL3 protein abundance. Using immunoprecipitation (IP) followed by immunoblotting with a pan-anti-lysine succinylation antibody, we detected ACSL3 succinylation across multiple tumor cell types (Fig. 5a, Extended Data Fig. 7a). Notably, ACSL3 exhibited higher levels of succinylation in tumor cells cultured on stiff matrices (Fig. 5b, Extended Data Fig. 7b). DES or DMK supplementation further increased ACSL3 succinylation while concomitantly reducing ACSL3 protein abundance (Fig. 5c, Extended Data Fig. 7c). Conversely, lowering succinyl-CoA levels by glycine treatment restored ACSL3 protein expression (Fig. 5d, Extended Data Fig. 7d). This suggests that succinate-driven succinylation of ACSL3 promotes ACSL3 degradation. Liquid chromatography-mass spectrometry (LC-MS) analysis identified two succinylation sites on ACSL3 in HK1 cells, at lysine 422 (K422) and lysine 660 (K660) (Fig. 5e, Extended Data Fig. 7e). Both sites were highly conserved among different species (Extended Data Fig. 7f-g). We re-expressed inducible ACSL3 mutants with K422/K660 replaced by arginine (K422R/K660R) in ACSL3 -/- HK1 cells (Extended Data Fig. 7h). However, compared to ACSL3-WT, only ACSL3-K422R showed reduced succinylation (Fig. 5f). In the presence of succinyl-CoA, substituting K422 with glutamate (K422E) mimicked succinylation could downregulate ACSL3 levels, whereas K422R had no such effect (Fig. 5g, Extended Data Fig. 7i), indicating that DES and DMK primarily target K422. Consistently, the K422R mutation abrogated the increase in ACSL3 protein abundance observed in tumor cells cultured on soft matrices (Fig. 5h). Moreover, the K422E mutation increased ACSL3 ubiquitination and shortened its half-life, whereas K422R exerted the opposite effects (Extended Data Fig. 7j-k). Consistent with its enhanced stability, re-expression of ACSL3-K422R markedly reshaped the cellular phospholipid landscape relative to ACSL3-WT. Specifically, overall phospholipid (PL) unsaturation was significantly reduced: MUFA-PL species increased by 6.0%, whereas PUFA-PL and saturated fatty acid (SFA)-PL species decreased by 3.4% and 2.7%, respectively (Fig. 5i). Moreover, upon supplementation with AA-d8, cells expressing ACSL3-K422R incorporated substantially less AA-d8 into phosphatidylethanolamine (PE) and phosphatidylcholine (PC) (Fig. 5j). Functionally, ACSL3-K422R re-expression conferred robust protection against RSL3/FIN56-induced ferroptosis, accompanied by reduced accumulation of di-oxygenated and tri-oxygenated PUFA-PE species (Extended Data Fig. 7l-p). Next, we conducted molecular dynamics simulation to explore the effect of K422 succinylation in ACSL3 stability. Compared with the WT model, the succinylated ACSL3 model underwent a marked conformational rearrangement during 100~200 ns before reaching a stable state, whereas the WT protein remained relatively stable throughout the simulation (Extended Data Fig. 8a). Root-mean-square fluctuation (RMSF) analysis showed that these differences were mainly concentrated in Domain 1 and Domain 2, both of which exhibited markedly greater fluctuations in the succinylated model than in the WT model (Extended Data Fig. 8b). More importantly, lysine residues in the succinylated model displayed increased solvent-accessible surface area (SASA), particularly in the C-terminal region encompassing Domain 1 and Domain 2, where nearly all lysine residues displayed increased SASA in the succinylated state (Extended Data Fig. 8c). Consistently, quantification of the number of water molecules within 5, 7, 10, and 15 Å of the ACSL3 centroid further confirmed greater solvent exposure in the succinylated model (Extended Data Fig. 8d-g). Given that increased solvent exposure of lysine residues is generally associated with a higher likelihood of ubiquitination, these results suggest that ACSL3 succinylation may promote its ubiquitination by increasing the accessibility of lysine residues. In terms of protein conformation, Domain 1 and Domain 2, which were adjacent in the WT model, became separated in the succinylated state, generating a prominent cleft that likely facilitated solvent penetration and increased lysine accessibility, thereby potentially enhancing the likelihood of ubiquitination (Fig. 5k, Extended Data Fig. 8h). In addition, the R424-T34 and R424-T37 hydrogen bonds present in the WT model, which connect Domain 1 to the N-terminal region, were disrupted in the succinylated model, where R424 instead formed a stable salt bridge with succinylated K422 (SL422) (Fig. 5l). Collectively, these changes suggest that K422 succinylation destabilizes ACSL3 conformation, increases lysine solvent exposure, and thereby promotes ubiquitination. Succinylation is a dynamic and reversible modification regulated by the balance between succinyltransferases (KAT2A, CPT1A, SUCLA2, OXCT1, etc.) 27 , 28 , 29, 30 and desuccinylases (SIRT5, SIRT7) 31, 32 . To identify ACSL3-interacting proteins, we performed quantitative mass spectrometry on anti-ACSL3 immunoprecipitates from HK1 cell lysates. Among the top 100 most enriched proteins, we identified CPT1A, a succinyltransferase. Subcellular fractionation revealed that both ACSL3 and CPT1A were present in the MAM fraction (Extended Data Fig. 9a). Co-immunoprecipitation assays further confirmed a physical interaction between ACSL3 and CPT1A (Fig. 5m, Extended Data Fig. 9b), and immunofluorescence analysis showed their cytoplasmic co-localization (Fig. 5n). Notably, although CPT1A protein abundance remained unchanged across stiffness conditions (Extended Data Fig. 9c), the interaction between CPT1A and ACSL3 was markedly reduced in tumor cells cultured on soft matrices (Fig. 5o, Extended Data Fig. 9d). CPT1A knockdown increased ACSL3 protein abundance in tumor cells cultured on stiff matrices (Extended Data Fig. 9e-f), while reducing ACSL3 ubiquitination and succinylation and prolonging its half-life (Extended Data Fig. 9g-i). Moreover, DMK or DES supplementation failed to reduce ACSL3 protein abundance in CPT1A-knockdown cells (Fig. 5p). In contrast, re-expression of inducible CPT1A-WT in CPT1A-knockout cells restored succinate-dependent ACSL3 succinylation and reduced ACSL3 protein levels (Extended Data Fig. 9j-k). Next, we re-expressed doxycycline-inducible CPT1A constructs encoding CPT1A-WT, CPT1A-H473A (inactive), or CPT1A-G710E (succinyltransferase-active) in CPT1A-knockout cells 27 . Re-expression of CPT1A-WT or CPT1A-G710E, but not CPT1A-H473A, markedly reduced ACSL3 protein abundance (Extended Data Fig. 9l). Consistently, CPT1A-WT overexpression decreased the level of ACSL3-WT, whereas it had no detectable effect on the AzzzCSL3-K422R mutant (Extended Data Fig. 9m). Moreover, CPT1A-WT and CPT1A-G710E markedly increased ACSL3 succinylation, whereas CPT1A-H473A had no effect (Extended Data Fig. 9n), indicating CPT1A functions as a succinyltransferase for ACSL3. Succinate-ACSL3(K422) axis promotes ferroptosis in vivo Next, to validated the effect of ACSL4 K422 succinylation in in vivo models, we constructed tumor-bearing mice by subcutaneously injecting ACSL3-WT HK1 cells, ACSL3-K422R HK1 cells and ACSL3-K422E HK1 cells. Following irradiation, tumors overexpressing ACSL3-K422R were significantly larger than those overexpressing ACSL3-WT (Fig. 6a), whereas ACSL3-K422E overexpression resulted in smaller tumors relative to ACSL3-WT (Fig. 6b). Importantly, tumor cells isolated from irradiated ACSL3-K422R xenografts contained lower levels of AA-containing PE species as well as di-oxygenated arachidonoyl- and adrenoyl-PE species compared with cells derived from ACSL3-WT and ACSL3-K422E tumors (Fig. 6c-d). Further corroborating our findings, DES or DMK treatment suppressed tumor growth in the ACSL3-WT overexpression group but had little effect in ACSL3-K422R tumors (Fig. 6e-f). Consistently, lipidomic profiling revealed that tumor cells isolated from irradiated ACSL3-WT tumors treated with DES or DMK exhibited increased levels of AA-containing PE species, as well as di-oxygenated arachidonoyl- and adrenoyl-PE species (Fig. 6g-h). Finally, we validated our findings in clinical specimens. In a retrospective radiotherapy cohort of patients with NPC (n=209; SYSUCC NPC cohort; Supplementary Table 1), immunohistochemical staining with an anti-ACSL3 antibody showed that lower ACSL3 expression was associated with a reduced risk of local recurrence after radiotherapy (Fig. 6i-j). Consistently, in public datasets, low ACSL3 expression was associated with improved overall survival in patients with head and neck squamous cell carcinoma and lung adenocarcinoma (Fig. 6k-l). We further defined a nine-gene succinate enrichment signature and found that high signature scores predicted favorable survival in adrenocortical carcinoma and kidney renal clear cell carcinoma (Supplementary Table 2, Fig. 6m-n). Collectively, these findings support a pivotal role for the succinate-ACSL3(K422) axis in regulating ferroptosis sensitivity and indicate that low ACSL3 expression together with high succinate enrichment is associated with improved clinical outcomes. Discussion In this study, we uncovered a mechanistic link between microenvironmental mechanical cues and ferroptosis sensitivity in tumor cells. We found that stiff matrices enhanced ferroptosis and radiotherapy sensitivity by enhancing microtubule acetylation, driving ER sheet-to-tubule remodeling, and promoting MAM formation. This structural reprogramming rewired mitochondrial metabolism, leading to succinate accumulation and reduced SDH activity. Elevated succinate in turn facilitated CPT1A-dependent succinylation of ACSL3 at K422, promoting its ubiquitination and degradation, thereby reshaping phospholipid composition toward a ferroptosis-prone state. We further confirmed the functional importance of this pathway in cellular and in vivo models. More importantly, low ACSL3 expression and high succinate enrichment were associated with favorable clinical outcomes. Collectively, our data identify MAMs as mechanosensitive intracellular hubs that couple tumor biomechanics to mitochondrial metabolism, phospholipid remodeling, ferroptosis, and radiotherapy response. The mechanical properties of the tumor microenvironment are increasingly recognized as active regulators of cancer cell behavior and therapeutic response. ECM composition and stiffness dictate regional susceptibility to tumor initiation 33 , promote tumor progression through mechanosignaling-driven exosome secretion 34 , and confer chemoresistance in pancreatic cancer organoids via CD44-hyaluronan signaling 35 . Beyond tumor cells, biomechanical stress drives CD8 + T cell exhaustion through the Piezo1/CREB/Osr2 axis 36 . Notably, ECM remodeling alters mitochondrial homeostasis in an evolutionarily conserved manner via TGF-β-induced mitochondrial fission and UPR^mt 37 , establishing a direct link between extracellular mechanics and mitochondrial function. However, whether ECM stiffness regulates specific mitochondrial metabolic outputs to control ferroptosis has remained unknown. Our study addresses this gap by revealing that substrate stiffness remodels MAMs through microtubule acetylation and ER morphological transition, driving succinate accumulation and ACSL3 succinylation, thereby positioning MAMs as intracellular mechanosensitive hubs that translate extracellular physical cues into mitochondrial metabolic reprogramming and ferroptosis regulation. The convergence of mechanical signaling on ACSL3 reflects a structural inevitability rather than a stochastic event. ACSL3 is the only known ferroptosis-protective phospholipid remodeling enzyme that resides at MAMs, placing it at the precise subcellular interface where stiffness-driven organelle remodeling, CPT1A-mediated succinylation and succinyl-CoA availability intersect. The evolutionary conservation of K422 at a structurally critical domain hinge further ensures that its succinylation produces maximal conformational destabilization. Thus, ACSL3 represents a uniquely vulnerable node where mechanical, metabolic and structural determinants converge to control ferroptosis sensitivity. This mechanosensitive role of MAMs extends current understanding of ER-mitochondria contacts in ferroptosis. Recent studies have shown that lipid peroxidation during ferroptosis initiates at the ER membrane and progressively spreads to other compartments 19, 38 , and that MAMs serve as the primary hotspots of proferroptotic phospholipid peroxide formation by stabilizing MAMs enhances ferroptosis, whereas untethering them confers protection 39 . The MAMs-resident chaperone σ1R facilitates ER-to-mitochondria Ca²⁺ transfer during ferroptosis, and disruption of MAMs blocks mitochondrial ROS production and ferroptosis execution 10 . Beyond tumor cells, ER–mitochondria contacts regulate immune cell fitness: MFN2-SERCA2 interaction sustains mitochondrial Ca²⁺ homeostasis in tumor-infiltrating CD8 + T cells 40 , and linoleic acid enhances anti-tumor immunity by promoting MAMs formation and mitochondrial energetics 41 . While these studies collectively establish MAMs as signaling hubs integrating Ca²⁺ dynamics, lipid transfer and redox homeostasis, they have focused on MAMs as either executors of chemically induced ferroptosis or regulators of immune cell metabolism. Our study demonstrates that substrate stiffness drives MAMs expansion independent of ferroptosis-inducing stimuli, and that the downstream consequence extends beyond Ca 2+ overload or lipid peroxidation to include SDH dysfunction, succinate accumulation and CPT1A-mediated ACSL3 succinylation, a metabolic axis not previously linked to MAMs. This finding also places our work within the rapidly expanding field of succinate and succinylation signaling. Succinate accumulation drives gut inflammation by switching FOXP3 from succinylation to ubiquitination-mediated degradation 42 , while succinyl-CoA-mediated succinylation of PD-L1 by CPT1A promotes its degradation and enhances anti-tumor immunity 43 . Sustained succinate exposure preserves CD8 + T cell stemness through BNIP3-mediated mitophagy 44 . SIRT5-mediated desuccinylation of TBK1 suppresses inflammatory signaling in aged skeletal muscle 45 , and metabolism-driven succinylation governs resource allocation for antibiotic resistance in bacteria 46 . At the succinate-fumarate node, fumarate regulates mitophagy through succination of Parkin cysteine residues 47 . These studies reveal protein succinylation as a regulatory layer connecting metabolic status to diverse cellular outcomes, yet whether succinylation responds to extracellular physical cues has remained entirely unexplored. Our study demonstrates that substrate stiffness, through EMCS remodeling and SDH dysfunction, drives succinate accumulation that functions as a metabolic encoder of mechanical information. The subsequent CPT1A-mediated succinylation and degradation of ACSL3 directly links mechanotransduction to ferroptosis sensitivity, establishing succinate as a mechanosensitive metabolite and positioning succinylation as a previously unrecognized post-translational mechanism in the mechanical control of cell fate. Methods Clinical specimens We collected 209 paraffin-embedded specimens from patients with locoregionally advanced nasopharyngeal carcinoma (NPC) at Sun Yat-sen University Cancer Center, Guangzhou, China (SYSUCC cohort). All patients were treatment-naïve at the time of biopsy. Tumor staging was determined according to the 8th American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) tumor-node-metastasis (TNM) classification system. All patients subsequently received radical radiotherapy. This study was approved by the Institutional Review Board of SYSUCC (G2025-179-01; G2025-108-01), and the requirement for written informed consent was waived by the ethics committee. Cell culture Human NPC cell lines HK1 and C666-1 were kindly provided by M.-S. Zeng at Sun Yat-sen University Cancer Center (SYSUCC). HT1080 (fibrosarcoma), 786-O (clear cell renal cell carcinoma), A375 (melanoma), MDA-MB-231 (breast cancer), and HEK293T cells were obtained from the China Center for Type Culture Collection. Cells were cultured in RPMI 1640 or DMEM (Invitrogen) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin. All cell lines were routinely screened for mycoplasma contamination. PDMS culture of tumor cells PDMS substrates were prepared by mixing dimethylsiloxane monomer (SYLGARD 184, Dow Corning) with the corresponding curing agent as previously described 36, 48 . To generate matrices with tunable stiffness, the base polymer and cross-linker were mixed at approximate ratios of 50:1 (softer), 40:1 (soft), 30:1 (modest), and 20:1 (stiff), reflecting the viscous nature of the base polymer. After thorough mixing and degassing under vacuum, the PDMS elastomers were cast into 12-well plates, 60-mm culture dishes, or onto glass slides for stiffness calibration, and then cured at 60 °C for 4 h. PDMS forms a planar surface with hydrophobic properties that permit protein adsorption for cell culture applications. Substrate stiffness was measured by nanoindenter (Optics11 life). In this study, the Young’s modulus of the PDMS substrates was approximately 5-10 kPa for the 50:1 formulation, 20-25 kPa for 40:1, 100 kPa for 30:1, and 500 kPa for 20:1. Reagents RSL3 (HY-100218A, MCE), FIN56 (S8254, Selleck), FINO2 (25096, Cayman), Liproxstatin-1(HY-12726, MCE), Doxycycline (HY-N0565, MCE), MG132(S2619, Selleck), carfilzomib (S2853, Selleck), Bafilomycin A1 (S1413, Selleck), Chloroquine (S6999, Selleck), Ferrostatin-1 (Fer1,HY-100579, MCE), Arachidonic Acid-d8 (CAC-390010-5, Cayman), Z-VAD-FMK (S7023, Selleck) and Necrostatin-1 (S8037, Selleck), rapamycin (HY-10219, MCE), Tubacin (GC16386, GLPBIO), Dimethyl 2-ketoglutarate (28394, Cayman), dimethyl succinate (HY-Y0808, MCE), dimethyl malonate (HY-Y1787, MCE), 3-nitropropionic acid (S3652, Selleck), CPI-613(S2776, Selleck), Glycine (HY-Y0966, MCE), Succinyl-CoA (S1129, Sigma-Aldrich). CRISPR-Cas9-mediated genome editing Single-guide RNAs (sgRNAs) targeting the indicated genes were designed using Benchling, and the corresponding target sequences are listed in the Supplementary Table1. Annealed sgRNA oligonucleotides were ligated into the PX458 vector (Addgene) following BbsI digestion. Cells were seeded at approximately 60% confluence and transfected with 1 μg of sgRNA-containing plasmid. The medium was replaced 8 h after transfection. Two or three days later, cells were dissociated with trypsin, and GFP-positive cells were sorted by flow cytometry. Sorted cells were then subjected to single-cell cloning in 96-well plates. Gene knockout was confirmed by western blotting. Lentivirus-mediated gene transfer For lentiviral production, HEK293T cells were co-transfected with the pSin-EF2-Puro-based construct, psPAX2, and pMD2.G. The medium was replaced with UltraCULTURE medium (Lonza) at 8 h after transfection. Viral supernatants were harvested at 48 h, passed through a filter, and applied to the indicated tumor cell lines for overnight infection at an MOI of 100. Western blotting was used to verify ectopic protein expression in the transduced cells. Immunoblotting analysis Protein lysates were prepared using RIPA buffer (Beyotime Biotechnology) containing EDTA-free Protease Inhibitor Cocktail (Beyotime Biotechnology). Following separation by SDS-PAGE (GenScript), proteins were transferred to nitrocellulose membranes. Membranes were then blocked and probed with the indicated primary antibodies (Supplementary Table 2) overnight at 4 °C, followed by incubation with HRP-linked secondary antibodies. Immunoreactive bands were visualized by enhanced chemiluminescence (Thermo Fisher Scientific). Immunofluorescence analysis Cells grown for immunofluorescence analysis were fixed with 4% paraformaldehyde for 15 min at room temperature and washed three times with PBS. After permeabilization in 0.1% Triton X-100/PBS for 15 min and three additional washes with PBST, samples were blocked and incubated with the indicated primary antibodies (Supplementary Table 2). Appropriate Alexa Fluor-conjugated secondary antibodies (Invitrogen) were then applied for 1 h at room temperature. Nuclei were stained with DAPI, and images were captured on a Zeiss LSM 880 confocal laser scanning microscope 49 . IHC analysis Paraffin-embedded samples were sectioned at 3 μm. Antigen retrieval was performed in 0.01 M citrate buffer (pH 6.0) using a pressure cooker for 15-20 min. Sections were then incubated with the indicated primary antibodies (Supplementary Table 2) overnight at 4 °C, followed by DAB-based detection (Dako) on the next day according to the manufacturer’s instructions. Images were captured using an AxioVision Rel.4.6 computerized image analysis system (Zeiss). All sections were scored by two experienced pathologists according to the immunoreactive score (IRS) system 50 . The staining intensity score was defined as follows: 0, negative staining; 1, weak staining; 2, moderate staining; and 3, strong staining. The positive rate score was defined as follows: 1, 70%. The total score of indicated proteins was calculated as staining intensity score multiply by positive rate score. Intracellular ferrous iron (Fe 2+ ) measurements The relative iron concentration in cell lysates was determined with Iron Assay kit (Abcam, #ab83366) and the experiments were performed according to the manufacturer’s instructions. GSH/GSSG ratio assay GSH and GSSG assays were performed, and the GSH/GSSG ratio was calculated using a GSH/GSSG Detection Assay (Abcam, #ab138881) according to the manufacturer’s instructions. Ferroptosis assay For cell viability assay, 2000 cells were plated in replicates in 96-well plates one day before adding the indicated drug. Cell viability was assessed 48h after drug treatment by Cell Counting Kit-8 (TargetMol) and normalized to an untreated control. Dose–response curves and half-maximal inhibitory concentration (IC50) values were generated using GraphPad Prism. To quantify ferroptotic cell death, the proportion of 7-AAD-positive tumor cells was determined after exposure to the indicated drugs or irradiation. Cells were stained with 7-AAD (BioLegend) and subjected to flow cytometric analysis. Data were acquired on a CytoFLEX LX with CytExpert 2.4 and analyzed using FlowJo v10. Identification of oxidized phospholipids by LC-MS Lipids were extracted following the Folch procedure as previously reported 51 , and global oxidized phospholipidomics was carried out as described earlier 52, 53 . Phospholipids (PLs) were analyzed by liquid chromatography-mass spectrometry (LC–MS) on a Dionex UltiMate 3000 LC system coupled to a Q-Exactive mass spectrometer (Thermo Fisher Scientific). Chromatographic separation was achieved on a normal-phase Luna Silica (2) column (3 μm, 150 × 2.0 mm; Phenomenex) at a flow rate of 0.2 ml min-1. The mobile phases consisted of 10 mM ammonium formate in isopropanol/hexane/water (285:215:5, v/v/v; solvent A) and isopropanol/hexane/water (285:215:40, v/v/v; solvent B), with all solvents of LC-MS grade. Gradient elution was programmed as follows: 0 min, 10% B; 23 min, 32%; 32 min, 65%; 35 min, 100%; 70 min, 100%. The column was maintained at 35 °C, and 5 μl of each sample was injected. Data acquisition was performed in negative ion mode at a resolution of 70,000 for full MS scans and 17,500 for data-dependent MS/MS scans. Full-scan spectra were collected over an m/z range of 400–1,800 with a maximum injection time of 200 ms using one microscan. For MS/MS acquisition, the maximum injection time was set to 500 ms, the collision energy to 24 eV, and the isolation window to 1.0 Da. Raw LC–MS data were processed using MZmine v.2.5.3 54 with an in-house analysis workflow and database. Peaks with a signal-to-noise ratio > 3 were extracted and searched against an oxidized PL database. Lipid species were assigned by matching m/z values within 5 ppm, followed by additional filtering based on retention time and confirmation through MS/MS fragmentation patterns, with fragment annotation referenced to lipid maps (https://www.lipidmaps.org). Deuterated PLs (Avanti Polar Lipids) were included as internal standards. Quantification was performed from full-scan LC–MS spectra by ratiometric comparison with the corresponding preselected internal standard, using a standard curve for each PL class. Analysis of AA-d8 containing phospholipids by LC-MS Total lipids were analyzed by LC-MS following separation on a reverse-phase Acquity HSS T3 column (1.8 μm, 100 × 2.1 mm; Waters) at a flow rate of 0.3 ml min-1. The mobile phases were 10 mM ammonium formate in water/acetonitrile (50:50, v/v; solvent A) and isopropanol/acetonitrile (90:10, v/v; solvent B). Elution was achieved with the following gradient: 0 min, 30% B; 5 min, 43%; 5.1 min, 50%; 14 min, 70%; 14.1 min, 70%; 23 min, 99%; 26 min, 99%. The column was held at 40 °C throughout the run. MS and data-dependent MS/MS analyses were performed on a Q-Exactive mass spectrometer (Thermo Fisher Scientific) in both positive and negative ion modes using profile acquisition. Full MS scans were acquired at a resolution of 70,000, whereas MS/MS scans were acquired at 17,500. The scan range for full MS was m/z 114-1,700, with one microscan, a maximum injection time of 100 ms, and an AGC target of 1 × 10 5 . For MS/MS, the maximum injection time was 50 ms, the isolation window was 1.0 Da, and the normalized collision energy was set to 20%, 30%, and 40%. Raw data processing and phospholipid identification were carried out using MS-DIAL 52 . Molecular Dynamics Simulations The three-dimensional structure of ACSL3 predicted by AlphaFoldwas used to construct the wild-type model. Protonation states of charged residues were assigned using the H++ server 55 together with manual inspection of local hydrogen-bonding networks. Histidine residues were modeled in either the ε-protonated or δ-protonated state according to their local environment, whereas Asp/Glu and Lys/Arg residues were maintained in their default charged states. The succinylated ACSL3 model was generated from the wild-type structure by attaching a succinic acid group to the side-chain amino group of K422 using Molecular Operating Environment (MOE2020). The protonation states in the succinylated model were kept identical to those in the wild-type model. Before simulation, each model was hydrogenated using the Leap module in Amber22, neutralized by addition of Na + and Cl - ions with AmberTools, and solvated in a rectangular TIP3P water box with a 10 Å buffer between the protein and the box boundary. Molecular dynamics simulations were performed in Amber22 with the GPU-accelerated pmemd.cuda module. Each system was subjected to stepwise energy minimization, including solvent relaxation with positional restraints on the protein, restrained minimization of the protein backbone, and final unrestrained minimization of the full system. The systems were then gradually heated from 0 to 300 K under the NVT ensemble for 100 ps, followed by 200 ps equilibration under the NPT ensemble to stabilize density. Production simulations were subsequently carried out for 500 ns at 300 K under periodic boundary conditions. The FF19SB force field was applied for ACSL3 56, 57 , and the succinylated K422 residue was defined as a customized residue (SL422). A 12 Å cutoff was used for van der Waals and electrostatic interactions, temperature was controlled with Langevin dynamics, and hydrogen-containing bonds were constrained using the SHAKE algorithm. Construction of tumor xenotransplantation model Six-week-old female specific pathogen-free (SPF) BALB/c nude mice were purchased from Charles River Laboratories. Mice were housed at five animals per cage under a 12-h light/12-h dark cycle at 20-26 °C with 40-70% humidity and had ad libitum access to standard chow and water. For xenograft studies, mice were subcutaneously inoculated with 1×10 6 HK1 cells or stably infected HK1 cells. When tumors reached approximately 5 mm in diameter, mice were subjected to local irradiation. Tumor volume was calculated using the formula length × width 2 × 0.5. At the indicated time points, tumor tissues were collected, paraffin embedded, and processed for immunohistochemical analysis. All animal procedures were approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University and were performed in accordance with institutional guidelines for animal welfare. Every effort was made to minimize animal suffering. The maximal tumor diameter permitted by the ethics committee was 20 mm, and this limit was not exceeded in any animal. All animal experiments were conducted in full compliance with the guidelines and under the approval of the Experimental Animal Ethics Committee of Sun Yat-sen University (SYSU-IACUC-2022-002494). Statistics and reproducibility Data are shown as results from at least three independent experiments. Statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software) or IBM SPSS Statistics version 25. Differences between two groups were assessed using a two-tailed unpaired Student’s t-test, whereas comparisons among multiple groups were performed using one-way or two-way ANOVA followed by Tukey’s multiple-comparisons test. Survival outcomes were analyzed using Kaplan–Meier methods and compared with the log-rank test. Clinical characteristics were compared using the chi-square (χ²) test. Phospholipid species were quantified from full-scan LC-MS spectra by ratiometric normalization to predefined internal standards and calculated using class-specific standard curves. A two-sided P value < 0.05 was considered statistically significant. Declarations Acknowledgments We thank Beijing Viktor Technology Co., Ltd. for the support with AFADESI-MSI platform, Hangzhou Shinning Technology Co., Ltd. for the support with biomechanical testing, and C. Tong (Zhejiang University) for technical assistance. This study was supported by grants from the National Natural Science Foundation of China (82504101), Changping Laboratory Project (2025C-12-04), Guangdong Basic and Applied Basic Research Foundation (2026B1515020031), Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM611), National Natural Science Foundation of China (82430085), Science and Technology Program of Guangzhou (2025B03J0149), Cancer Innovative Research Program of Sun Yat-sen University Cancer Center (CIRP-SYSUCC-0005), Overseas Expertise Introduction Project for Discipline Innovation (111 Project), Natural Science Foundation of China (82522094, 82321004), Guangdong Basic and Applied Basic Research Foundation (2023B1515020020), the Chih Kuang Scholarship for Outstanding Young PhysicianScientists of Sun Yat-sen University Cancer Center (PT22120901), Young Talents Program of Sun Yat-sen University Cancer Center (YTP-SYSUCC-0072). W.Y. Sun gratefully acknowledges the support of the K. C. Wong Education Foundation in Jinan University. A uthor contributions X.-Y.L., Z.L. and W.-Y.S. conceived the experiments. Z.L., N.-Z.L., and Y.Q. carried out and analyzed the data for most of the in vitro experiments. Z.L., and N.-Z.L. designed and performed the animal experiments. W.-Y.S., R.W., and J.-C.F. designed and performed LC-MS/MS analysis. Z.L., N.-Z.L., T.-X.H. and Y.-H.L. collected clinical samples and performed the IHC experiments. H.-M.W., C.-Q.Z., Y.-L.C., Y.-L.W., and Z.-J.D. helped with the data analyses. R.-R.H. provided technical support. Z.L., X.-Y.L. and W.-Y.S. wrote the manuscript. J.M., X.-Y.L., and W.-Y.S. supervised the study. All authors reviewed and discussed the final version of the manuscript. Competing interests The authors declare no competing interests. References Nia, H.T., Munn, L.L. & Jain, R.K. Physical traits of cancer. Science (New York, N.Y.) 370 (2020). Cox, T.R. The matrix in cancer. Nature reviews. Cancer 21 , 217-238 (2021). Lei, G. et al. The role of ferroptosis in ionizing radiation-induced cell death and tumor suppression. Cell Res. 30 , 146-162 (2020). Stockwell, B.R. Ferroptosis turns 10: Emerging mechanisms, physiological functions, and therapeutic applications. Cell 185 , 2401-2421 (2022). Lang, X. et al. Radiotherapy and Immunotherapy Promote Tumoral Lipid Oxidation and Ferroptosis via Synergistic Repression of SLC7A11. Cancer discovery 9 , 1673-1685 (2019). Chen, X., Kang, R., Kroemer, G. & Tang, D. Broadening horizons: the role of ferroptosis in cancer. Nature reviews. Clinical oncology 18 , 280-296 (2021). Grevengoed, T.J., Klett, E.L. & Coleman, R.A. Acyl-CoA metabolism and partitioning. Annual review of nutrition 34 , 1-30 (2014). Magtanong, L. et al. Exogenous Monounsaturated Fatty Acids Promote a Ferroptosis-Resistant Cell State. Cell chemical biology 26 , 420-432.e429 (2019). Phillips, M.J. & Voeltz, G.K. Structure and function of ER membrane contact sites with other organelles. Nat. Rev. Mol. Cell Biol. 17 , 69-82 (2016). Zhang, Z. et al. CGI1746 targets σ(1)R to modulate ferroptosis through mitochondria-associated membranes. Nature chemical biology 20 , 699-709 (2024). Wu, H., Carvalho, P. & Voeltz, G.K. Here, there, and everywhere: The importance of ER membrane contact sites. Science (New York, N.Y.) 361 (2018). Zheng, P. et al. ER proteins decipher the tubulin code to regulate organelle distribution. Nature 601 , 132-138 (2022). You, R. et al. Single-cell and spatial transcriptomics reveal mechanisms of radioresistance and immune escape in recurrent nasopharyngeal carcinoma. Nature genetics 57 , 1950-1965 (2025). Kolácná, L. et al. Biochemical and biophysical aspects of collagen nanostructure in the extracellular matrix. Physiological research 56 Suppl 1 , S51-s60 (2007). Levental, K.R. et al. Matrix crosslinking forces tumor progression by enhancing integrin signaling. Cell 139 , 891-906 (2009). Georges, P.C. et al. Increased stiffness of the rat liver precedes matrix deposition: implications for fibrosis. American journal of physiology. Gastrointestinal and liver physiology 293 , G1147-1154 (2007). Nakamura, T. & Conrad, M. Exploiting ferroptosis vulnerabilities in cancer. Nature cell biology 26 , 1407-1419 (2024). Kimura, H. et al. Syntaxin 17 promotes lipid droplet formation by regulating the distribution of acyl-CoA synthetase 3. Journal of lipid research 59 , 805-819 (2018). von Krusenstiern, A.N. et al. Identification of essential sites of lipid peroxidation in ferroptosis. Nature chemical biology 19 , 719-730 (2023). Yang, Z., Zhao, X., Xu, J., Shang, W. & Tong, C. A novel fluorescent reporter detects plastic remodeling of mitochondria-ER contact sites. Journal of cell science 131 (2018). Cieri, D. et al. SPLICS: a split green fluorescent protein-based contact site sensor for narrow and wide heterotypic organelle juxtaposition. Cell death and differentiation 25 , 1131-1145 (2018). Komatsu, T. et al. Organelle-specific, rapid induction of molecular activities and membrane tethering. Nature methods 7 , 206-208 (2010). Jang, W. et al. Endosomal lipid signaling reshapes the endoplasmic reticulum to control mitochondrial function. Science 378 , eabq5209 (2022). Friedman, J.R., Webster, B.M., Mastronarde, D.N., Verhey, K.J. & Voeltz, G.K. ER sliding dynamics and ER-mitochondrial contacts occur on acetylated microtubules. The Journal of cell biology 190 , 363-375 (2010). Seetharaman, S. et al. Microtubules tune mechanosensitive cell responses. Nature materials 21 , 366-377 (2022). Li, H., Zhang, H., Zhu, J., Tang, Y. & An, J. Protein succinylation in tumors-from biology to clinical therapy. International journal of biological macromolecules 337 , 149489 (2026). Kurmi, K. et al. Carnitine Palmitoyltransferase 1A Has a Lysine Succinyltransferase Activity. Cell Rep. 22 , 1365-1373 (2018). Wang, Y. et al. KAT2A coupled with the α-KGDH complex acts as a histone H3 succinyltransferase. Nature 552 , 273-277 (2017). Ma, W. et al. OXCT1 functions as a succinyltransferase, contributing to hepatocellular carcinoma via succinylating LACTB. Molecular cell 84 , 538-551.e537 (2024). Tong, Y. et al. SUCLA2-coupled regulation of GLS succinylation and activity counteracts oxidative stress in tumor cells. Molecular cell 81 , 2303-2316.e2308 (2021). Shen, R. et al. Lysine succinylation, the metabolic bridge between cancer and immunity. Genes & diseases 10 , 2470-2478 (2023). Du, J. et al. Sirt5 is a NAD-dependent protein lysine demalonylase and desuccinylase. Science (New York, N.Y.) 334 , 806-809 (2011). Bansaccal, N. et al. The extracellular matrix dictates regional competence for tumour initiation. Nature 623 , 828-835 (2023). Wu, B. et al. Stiff matrix induces exosome secretion to promote tumour growth. Nature cell biology 25 , 415-424 (2023). LeSavage, B.L. et al. Engineered matrices reveal stiffness-mediated chemoresistance in patient-derived pancreatic cancer organoids. Nature materials 23 , 1138-1149 (2024). Zhang, J. et al. Osr2 functions as a biomechanical checkpoint to aggravate CD8(+) T cell exhaustion in tumor. Cell 187 , 3409-3426.e3424 (2024). Zhang, H. et al. The extracellular matrix integrates mitochondrial homeostasis. Cell 187 , 4289-4304.e4226 (2024). Xu, L. et al. Live-cell imaging with fluorogenic radical-trapping antioxidant probes reveals the onset and progression of ferroptosis. Nature chemistry 17 , 1963-1973 (2025). Sassano, M.L. et al. Endoplasmic reticulum-mitochondria contacts are prime hotspots of phospholipid peroxidation driving ferroptosis. Nature cell biology 27 , 902-917 (2025). Yang, J.F. et al. Mitochondria-ER contact mediated by MFN2-SERCA2 interaction supports CD8(+) T cell metabolic fitness and function in tumors. Science immunology 8 , eabq2424 (2023). Nava Lauson, C.B. et al. Linoleic acid potentiates CD8(+) T cell metabolic fitness and antitumor immunity. Cell metabolism 35 , 633-650.e639 (2023). Wang, H. et al. Succinate drives gut inflammation by promoting FOXP3 degradation through a molecular switch. Nature immunology 26 , 866-880 (2025). Liang, L. et al. Alterations in PD-L1 succinylation shape anti-tumor immune responses in melanoma. Nature genetics 57 , 680-693 (2025). Ma, K. et al. Succinate preserves CD8(+) T cell fitness to augment antitumor immunity. Immunity 58 , 2505-2523.e2508 (2025). Zhao, Q. et al. SIRT5 safeguards against primate skeletal muscle ageing via desuccinylation of TBK1. Nature metabolism 7 , 556-573 (2025). Wu, J.H. et al. Metabolism-dependent succinylation governs resource allocation for antibiotic resistance. Science advances 11 , eadu2856 (2025). Ham, S.J. et al. Mitochondrial fumarate inhibits Parkin-mediated mitophagy. Molecular cell 85 , 2287-2302.e2289 (2025). Zeng, Y. et al. Substrate stiffness regulates B-cell activation, proliferation, class switch, and T-cell-independent antibody responses in vivo. European journal of immunology 45 , 1621-1634 (2015). Xu, Z. et al. TOLLIP targets GSDME-NT-carrying endocytic vesicles for autophagy to regulate pyroptosis and chemotherapy efficacy. Nature cell biology (2026). Wu, J. et al. Methionine metabolite spermidine inhibits tumor pyroptosis by enhancing MYO6-mediated endocytosis. Nature communications 16 , 2184 (2025). Folch, J., Lees, M. & Sloane Stanley, G.H. A simple method for the isolation and purification of total lipides from animal tissues. The Journal of biological chemistry 226 , 497-509 (1957). Li, Z. et al. Tumor-repopulating cells evade ferroptosis via PCK2-dependent phospholipid remodeling. Nat. Chem. Biol. 20 , 1341-1352 (2024). Sun, W.Y. et al. Phospholipase iPLA(2)β averts ferroptosis by eliminating a redox lipid death signal. Nature chemical biology 17 , 465-476 (2021). Pluskal, T., Castillo, S., Villar-Briones, A. & Oresic, M. MZmine 2: modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. BMC bioinformatics 11 , 395 (2010). Anandakrishnan, R., Aguilar, B. & Onufriev, A.V. H++ 3.0: automating pK prediction and the preparation of biomolecular structures for atomistic molecular modeling and simulations. Nucleic acids research 40 , W537-541 (2012). Tian, C. et al. ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution. Journal of chemical theory and computation 16 , 528-552 (2020). Svozil, D. et al. Geometrical and electronic structure variability of the sugar-phosphate backbone in nucleic acids. The journal of physical chemistry. B 112 , 8188-8197 (2008). Additional Declarations There is NO Competing Interest. Supplementary Files Extendedfigure.docx SupplementaryTable1.docx SupplementaryTable2.docx SupplementaryTable3.docx Graphicalabstract.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9481973","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":632717140,"identity":"541d290e-9a78-4537-be5c-f80037ba6a19","order_by":0,"name":"Jun Ma","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/ElEQVRIiWNgGAWjYHCCBAaGChsGBmbmBpiIARFazqQBtTASr4WBgbHtMIgkUotu+4Fn0gVs56P52xkbGH+21SU2sDdvk2CouYNTi9mZhDTpGTy3c2ccZmxg5m07nNjAc6xMguHYM9xaDgC18Ejczm0AaWFsO5DYIJFjJsHYcBi3lvMPgFoMzuXOPwxzmPwbAlpugGxJOJC7AaiFgbeNGWgLDyEtD5KteQ4k524EajnMc+6wcRtPWrFFwjF8DstJvM37zy533vnDBx/+KKuT7Wc/vPHGhxrcWhgYeBLgzAOMbAwMbCBWAg7FEMB+AInzB6/SUTAKRsEoGKEAAFE9V8wVr4E8AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-1137-9349","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":true,"prefix":"","firstName":"Jun","middleName":"","lastName":"Ma","suffix":""},{"id":632717141,"identity":"f99f9179-99c8-4c6f-91cd-3e9055cdf114","order_by":1,"name":"Zhe Li","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Zhe","middleName":"","lastName":"Li","suffix":""},{"id":632717142,"identity":"45065320-70f6-49ee-8453-c43037afe740","order_by":2,"name":"Nuo-zhou Liu","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Nuo-zhou","middleName":"","lastName":"Liu","suffix":""},{"id":632717143,"identity":"043d2189-fac0-403b-832f-c76ed5d7dabc","order_by":3,"name":"Rong Wang","email":"","orcid":"","institution":"Jinan University","correspondingAuthor":false,"prefix":"","firstName":"Rong","middleName":"","lastName":"Wang","suffix":""},{"id":632717144,"identity":"0e19cef3-88c4-4843-86b7-64f7370f6419","order_by":4,"name":"Tingxiang He","email":"","orcid":"https://orcid.org/0009-0000-9187-2493","institution":"State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Tingxiang","middleName":"","lastName":"He","suffix":""},{"id":632717145,"identity":"b085d40a-5517-435f-ab19-ab263b5d8e34","order_by":5,"name":"Ying Qi","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Ying","middleName":"","lastName":"Qi","suffix":""},{"id":632717146,"identity":"5e35ef49-f051-4941-b4e7-cd74b6c0f6c1","order_by":6,"name":"Chu-qing Zhang","email":"","orcid":"","institution":"State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Chu-qing","middleName":"","lastName":"Zhang","suffix":""},{"id":632717147,"identity":"2d3a86e7-d43d-444e-b070-54d2a85fed74","order_by":7,"name":"Jingcheng Feng","email":"","orcid":"","institution":"State Key Laboratory of Bioactive Molecules and Druggability Assessment","correspondingAuthor":false,"prefix":"","firstName":"Jingcheng","middleName":"","lastName":"Feng","suffix":""},{"id":632717148,"identity":"8a113b76-03a5-4d86-957f-4c3ab1c8ca2c","order_by":8,"name":"Zhenji Deng","email":"","orcid":"","institution":"State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Zhenji","middleName":"","lastName":"Deng","suffix":""},{"id":632717149,"identity":"b64f9ff4-ec24-4ce2-8466-81fd05d237b4","order_by":9,"name":"Hanmiao Wei","email":"","orcid":"","institution":"State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Hanmiao","middleName":"","lastName":"Wei","suffix":""},{"id":632717150,"identity":"658132a5-f2c0-4e2c-86f3-d864d80357a2","order_by":10,"name":"Yun-long Wang","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Yun-long","middleName":"","lastName":"Wang","suffix":""},{"id":632717151,"identity":"23d1a5ff-dbda-4433-b24e-61a36772f7fa","order_by":11,"name":"Ying-lin Cai","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Ying-lin","middleName":"","lastName":"Cai","suffix":""},{"id":632717152,"identity":"4d9468c2-c517-4ebf-a3de-91c03a360cab","order_by":12,"name":"Yu-hao Li","email":"","orcid":"","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Yu-hao","middleName":"","lastName":"Li","suffix":""},{"id":632717153,"identity":"c0d62f40-7223-4b42-b0d4-d250d91634de","order_by":13,"name":"Rong-Rong He","email":"","orcid":"https://orcid.org/0000-0001-5505-5672","institution":"Jinan University","correspondingAuthor":false,"prefix":"","firstName":"Rong-Rong","middleName":"","lastName":"He","suffix":""},{"id":632717154,"identity":"97fa0663-47b2-4fe4-9994-2dbe2759ec03","order_by":14,"name":"Wan-Yang Sun","email":"","orcid":"https://orcid.org/0000-0002-4439-967X","institution":"Jinan University","correspondingAuthor":false,"prefix":"","firstName":"Wan-Yang","middleName":"","lastName":"Sun","suffix":""},{"id":632717155,"identity":"d25eb067-eb0e-44f8-b469-0258fe7f48ac","order_by":15,"name":"Xiaoyu liang","email":"","orcid":"https://orcid.org/0000-0002-8901-9239","institution":"Sun Yat-sen University Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Xiaoyu","middleName":"","lastName":"liang","suffix":""}],"badges":[],"createdAt":"2026-04-21 09:26:52","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9481973/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9481973/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108492907,"identity":"da70742c-e53c-451e-afba-e93bfa7060b8","added_by":"auto","created_at":"2026-05-05 09:58:57","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":951251,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHigh matrix stiffness enhances radiosensitivity by promoting ferroptosis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Left, immunostainings of COL1A1 (green) and CK (red) on NPC paraffin sections. Scale bars, 50 μm. Middle, stiffness maps of yellow outlined area from cryosections of NPC. Right, quantification of average tissue stiffness in COL1A1 high expression NPC (n = 100) and COL1A1 low expression NPC (n = 100).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e, Left, immunostainings of COL1A1(green) and CK (red) on rNPC and nrNPC paraffin sections. Scale bars, 100 μm. Right, Quantification of COL1A1 fluorescence intensity (n = 100 rNPC /100 nrNPC).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, Schematic illustration of custom-fabricated polydimethylsiloxane (PDMS) gels with defined stiffness, designated as stiff, modest and soft.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed-e\u003c/strong\u003e, HK1 \u003cstrong\u003e(d)\u003c/strong\u003e or C666-1\u003cstrong\u003e (e)\u003c/strong\u003e cells were cultured on stiff, modest or soft matrices and pretreated with vehicle or Fer-1 before irradiation. The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef-g\u003c/strong\u003e, HK1 cells were cultured on stiff or soft matrices and pretreated with DMSO, Fer-1, DFO, Z-VAD-FMK (Z-VAD) or necrostatin-1 (Nec-1) before treatment with RSL3\u003cstrong\u003e (f) \u003c/strong\u003eor FIN56 \u003cstrong\u003e(g)\u003c/strong\u003e. The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh-j\u003c/strong\u003e, Dose-dependent cell death induced by the ferroptosis inducers RSL3\u003cstrong\u003e (h)\u003c/strong\u003e, FIN56 \u003cstrong\u003e(i) \u003c/strong\u003eor FINO2\u003cstrong\u003e (g) \u003c/strong\u003ein HK1 cells cultured on stiff, modest or soft matrices. n = 3 replicates from one representative of three independent experiments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek-l,\u003c/strong\u003e HK1 \u003cstrong\u003e(k)\u003c/strong\u003e or C666-1\u003cstrong\u003e (l)\u003c/strong\u003e cells were cultured on stiff, modest, soft or softer matrices, with or without Fer-1, before treatment with RSL3. The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003em-n\u003c/strong\u003e, Levels of the ferroptosis-associated phospholipid species PE (38:4)-OOH\u003cstrong\u003e (m) \u003c/strong\u003eand PE (40:4)-OOH\u003cstrong\u003e (n)\u003c/strong\u003e in HK1 cells cultured on stiff or soft matrices and treated with RSL3, with or without Fer-1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eo\u003c/strong\u003e, quantification of average tissue stiffness in LOX-overexpressing HK1 tumors or control tumors (n =10).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ep\u003c/strong\u003e, HK1 cells stably expressing LOX or vector were implanted subcutaneously into female BALB/c nude mice. Mice were exposed to radiotherapy (8 Gy, arrow) or not. Liproxstatin-1 (Lip-1) was administered to the indicated groups (10 mg/kg, intraperitoneally, once daily; n = 5).\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SD and are representative of at least three independent experiments, unpaired two-tailed \u003cem\u003et\u003c/em\u003e-test (a-b, o), two-way ANOVA (d-g, p) or one-way ANOVA (k-n). ns, not significant.\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/80ca876d3e8645ec2e585f94.jpeg"},{"id":108493561,"identity":"a57253a5-10cf-4f21-b9f5-33cbb58394ec","added_by":"auto","created_at":"2026-05-05 10:00:57","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1040993,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMechanical stiffness reshapes the lipid landscape of tumor cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Immunoblot analysis of PLs-remodelling factors, including ACSL4, ACSL3, LPCAT3, LPCAT1, MBOAT1, iPLA2β, FASN and SCD1, in HK1 and C666-1 cells cultured on matrices of different stiffness.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e, LC-MS-based lipidomic analysis of SFA-containing phospholipids (SFA-PLs), MUFA-containing phospholipids (MUFA-PLs) and PUFA-containing phospholipids (PUFA-PLs) in HK1 cells cultured on matrices of different stiffness. n = 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, Heatmap showing LC-MS analysis of the relative abundance of PE or PC molecular species in HK1 cells cultured on matrices of different stiffness. Row z-scores were calculated from the averaged abundance of each lipid species.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e, MSI-based spatially resolved metabolomic analysis of PUFA-PLs in treatment-naive rNPC and nrNPC tissues.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee-f\u003c/strong\u003e, \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 \u003cstrong\u003e(e) \u003c/strong\u003eor C666-1\u003cstrong\u003e (f) \u003c/strong\u003ecells and parental cells were cultured on stiff or soft matrices and treated with RSL3 or FINO2. n = 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg-h\u003c/strong\u003e, Levels of the ferroptosis-associated phospholipid species PE (38:4)-OOH\u003cstrong\u003e (g) \u003c/strong\u003eand PE (40:4)-OOH \u003cstrong\u003e(h) \u003c/strong\u003ein parental and ACSL3-knockout HK1 cells cultured on stiff or soft matrices and treated with RSL3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e, \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells and parental cells were cultured on stiff or soft matrices and exposed to the indicated doses of irradiation. The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SD and are representative of at least three independent experiments, unpaired two-tailed \u003cem\u003et\u003c/em\u003e-test (e-f), one-way ANOVA (b, g-h) or two-way ANOVA (i). ns, not significant.\u003c/p\u003e","description":"","filename":"image2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/ac4f7a81ecdf397c91278193.jpeg"},{"id":108492590,"identity":"4d6778ad-3a4b-4fae-bf9d-023e9e1c0717","added_by":"auto","created_at":"2026-05-05 09:58:07","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":823953,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMMF regulates ferroptosis through MAMs.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Immunoblots analysis of ACSL3 in HK1 cells cultured on matrices of different stiffness and treated with cycloheximide (CHX) for the indicated times.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e, Immunoblots analysis of ACSL3 in HK1 cells cultured on soft or stiff matrices and treated with MG132, bafilomycin A1 (Baf A1) or chloroquine (CQ) for 6 h.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, HK1 cells cultured on soft or stiff matrices were subjected to immunoprecipitation with an anti-ACSL3 antibody, followed by immunoblotting with anti-ubiquitin (Ub) antibody.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e, Representative electron microscopy images showing MAM structures in HK1 cells cultured on soft or stiff matrices.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e, Immunofluorescence analysis of endoplasmic reticulum (ER; red) and mitochondria (green) co-localization in HK1 cells grown on soft or stiff substrates. Scale bar, 10 μm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e, Flow-cytometric quantification of FITC mean fluorescence intensity (MFI) of the GFP-MAM reporter in HK1 cells cultured on soft or stiff matrices.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e, Immunofluorescence images of GFP-MAM signal in HK1 cells cultured on soft or stiff matrices. Scale bar, 10 μm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e, Immunoblot analysis of ACSL3 in HK1 cells cultured on soft or stiff matrices and treated with vehicle or rapamycin.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e, HK1 cells cultured on soft matrices were treated with vehicle or rapamycin, followed by AA-d8 (10 µM for 36 h) treatment, the relative abundance of PC and PE species containing AA (20:4)-d8 was determined. n = 5. *\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05; **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01; ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e, HK1 cells cultured on soft matrices were pretreated with vehicle or rapamycin before exposure to RSL3 or RSL3 plus Fer-1. The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek-l\u003c/strong\u003e, Levels of ferroptosis-associated phospholipid species, including PE (38:4)-OOH \u003cstrong\u003e(k)\u003c/strong\u003e and PE (40:4)-OOH \u003cstrong\u003e(l)\u003c/strong\u003e, in HK1 cells cultured on soft matrices treated with RSL3 or RSL3 plus Fer-1, with or without rapamycin.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003em\u003c/strong\u003e, HK1 cells cultured on soft matrices were treated with irradiation (IR, 8 Gy), with or without rapamycin treatment.The percentage of dead cells was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003en\u003c/strong\u003e, Flow-cytometric quantification of FITC mean fluorescence intensity (MFI) of the GFP-MAM reporter in HK1 cells cultured on stiff matrices following silencing of \u003cem\u003eRab10\u003c/em\u003e or \u003cem\u003eRTN4\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eo\u003c/strong\u003e, Immunofluorescence images of GFP-MAM signal in HK1 cells cultured on stiff matrices following \u003cem\u003eRab10\u003c/em\u003e or \u003cem\u003eRTN4\u003c/em\u003e silencing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ep\u003c/strong\u003e, Immunoblot analysis of ACSL3 in HK1 cells cultured on stiff matrices following\u003cem\u003eRab10\u003c/em\u003e or \u003cem\u003eRTN4 \u003c/em\u003esilencing.\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SD and are representative of at least three independent experiments, unpaired two-tailed \u003cem\u003et\u003c/em\u003e-test (f, i, m), one-way ANOVA (k-l, n) or two-way ANOVA (j).\u003c/p\u003e","description":"","filename":"image3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/39159144e72f88a3286e0855.jpeg"},{"id":108416409,"identity":"c8711f80-daef-417f-b77c-9cc6012fb604","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1056769,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMitochondrial succinate accumulation facilitates ferroptosis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Oxygen consumption rate (OCR) of HK1 cells cultured on stiff or soft matrices were analyzed. Oligo (oligomycin), FCCP and Rot/AA (rotenone and antimycin A) were used to determine the basal respiration, ATP-coupled respiration, maximal respiratory capacity and non-mitochondrial oxygen consumption. n = 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e, Spare respiratory capacity (SRC) of HK1 cells cultured on stiff or soft matrices. n = 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, Heatmap of mitochondria-related metabolites in HK1 cells cultured on matrices of different stiffness. Row z-scores were calculated from the averaged abundance of each lipid species.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed-e\u003c/strong\u003e, Succinate-to-fumarate ratio in HK1 cells cultured on matrices of different stiffness \u003cstrong\u003e(d)\u003c/strong\u003e or HK1 cells cultured on soft matricest treated with vehicle or rapamycin \u003cstrong\u003e(e)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e, AFADESI-MSI imaging of succinate and fumarate in nrNPC and rNPC tissues.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e, Heatmap showing relative abundance AA-d8 containing PE and PC species in HK1 cells cultured on stiff or soft matrices, with or without DES or DMK supplementation. For each PLs, values were normalized to the first group (stiff) and visualized as log\u003csub\u003e2\u003c/sub\u003e(normalized to control).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e, Percentage of dead cells, assessed as 7-AAD-positive cells by flow cytometry, in HK1 cells cultured on matrices of different stiffness and treated with RSL3 alone, RSL3 plus DMK, or RSL3 plus DMK and Fer-1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e, Percentage of dead cells, assessed as 7-AAD-positive cells by flow cytometry, in HK1 cells cultured on matrices of different stiffness, exposed to the indicated doses of irradiation, and treated with or without DMK.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej-k\u003c/strong\u003e, Levels of the ferroptosis-associated phospholipid species PE (38:4)-OOH \u003cstrong\u003e(j)\u003c/strong\u003e and PE (40:4)-OOH \u003cstrong\u003e(k) \u003c/strong\u003ein HK1 cells cultured on matrices of different stiffness, treated with RSL3 and DMK or DES.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003el\u003c/strong\u003e, Immunoblots analysis of ACSL3 in HK1 cells cultured on matrices of different stiffness and treated with DMK or vehicle.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003em\u003c/strong\u003e, Immunoblots analysis of ACSL3 in HK1 cells cultured on stiff or soft matrices, with or without DES supplementation, and treated with CHX for the indicated times.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003en\u003c/strong\u003e, Cell lysates from HA-ACSL3-WT-expressing HK1 cells cultured on stiff or soft matrices, with or without DES or DMK supplementation, were subjected to immunoprecipitation with an anti-HA antibody, followed by immunoblotting with Ub antibody.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SD and are representative of at least three independent experiments, unpaired two-tailed \u003cem\u003et\u003c/em\u003e-test (b, e), one-way ANOVA (d, h, j-k) or two-way ANOVA (i). ns, not significant.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e","description":"","filename":"image4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/d56b0f8b3360cf14d966937d.jpeg"},{"id":108416412,"identity":"b6ac164c-c270-48b6-9966-28768c974724","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":821641,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCPT1A mediates ACSL3 succinylation to promote its degradation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing lysine succinylation of ACSL3 in \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells reconstituted with ACSL3-WT.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing lysine succinylation of ACSL3 in HK1 cells cultured on soft or stiff matrices.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing lysine succinylation of ACSL3 in re-expressed ACSL3-WT-HA HK1 cells treated with DES or DMK.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e, Immunoblot analysis of ACSL3 and CPT1A in HK1 cells cultured on soft or stiff matrices and treated with MG132 or glycine.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e, Mass spectrometric identification of the succinylation site in ACSL3 in HK1 cells.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing lysine succinylation of re-expressed ACSL3-WT, ACSL3-K422R or ACSL3-K660R in \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells, in the presence or absence of succinyl-CoA.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e, Immunoblots analysis of ACSL3-HA expression in ACSL3-WT, ACSL3-K422R or ACSL3-K660R HK1 cells treated with vehicle, DES or DMK.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e, Immunoblots analysis of ACSL3 expression in \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells reconstituted with ACSL3-WT or ACSL3-K422R cultured on matrices of different stiffness.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e, Bar chart showing the relative abundance of SFA-PLs, MUFA-PLs and PUFA-PLs in \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells reconstituted with ACSL3-WT or ACSL3-K422R.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e, \u003cem\u003eACSL3\u003c/em\u003e-knockout HK1 cells reconstituted with ACSL3-WT or ACSL3-K422R were treated with AA-d8 (10 µM) for 36 h, and the relative abundance of PE and PC species containing AA (20:4)-d8 was determined. n = 5. *\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05; **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01; ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek\u003c/strong\u003e, The structure of ACSL3 was computationally modeled based on the AlphaFold package. N-terminal (green, residues 1-53), Domain 1 (residues 384-451) and Domain 2 (residues 594-720) were used to demonstrate structure differences between ACSL3 WT model and ACSL3 K422\u003cstrong\u003e \u003c/strong\u003esuccinylation model.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003el\u003c/strong\u003e, Representative structures of ACSL3 WT model (left) and ACSL3 K422\u003cstrong\u003e \u003c/strong\u003esuccinylation model (right) and their corresponding key interactions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003em\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing the interaction between CPT1A and ACSL3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003en\u003c/strong\u003e, Representative immunofluorescence images showing the colocalization of CPT1A (green) and ACSL3 (red). Scale bars, 10 μm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eo\u003c/strong\u003e, Immunoprecipitation followed by immunoblotting showing the interaction between CPT1A and ACSL3 in HK1 cells cultured on soft or stiff matrices.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ep\u003c/strong\u003e, Immunoblot analysis of ACSL3 in \u003cem\u003eCPT1A\u003c/em\u003e-knockout cells treated with vehicle, DES or DMK.\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SD and are representative of at least three independent experiments, unpaired two-tailed \u003cem\u003et\u003c/em\u003e-test (j).\u003c/p\u003e","description":"","filename":"image5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/9f9d905ea0ca68eb27f3435d.jpeg"},{"id":108416413,"identity":"a2191305-4a2a-4f19-8fe7-9accd64fed88","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1176410,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSuccinate-ACSL3(K422) axis promotes ferroptosis in vivo.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea-b\u003c/strong\u003e, HK1 cells stably expressing the ACSL3-WT, ACSL3 K422R mutant\u003cstrong\u003e (a)\u003c/strong\u003e or the ACSL3 K422E mutant \u003cstrong\u003e(b)\u003c/strong\u003e were implanted subcutaneously into female BALB/c nude mice to establish xenograft growth models. Mice received radiotherapy (8 Gy; arrows) or not. Tumor volumes are shown for each group (n = 6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e, Levels of esterified AA (C20:4)-PE molecular species in irradiated HK1 xenografts expressing ACSL3 WT, ACSL3 K422R, or ACSL3 K422E (n = 6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e, Levels of PE (38:4)-OOH and PE (40:4)-OOH in irradiated HK1 xenografts expressing ACSL3 WT, K422R, or K422E (n = 6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee-f\u003c/strong\u003e, HK1 cells stably expressing ACSL3-WT or ACSL3 K422R were implanted subcutaneously into female BALB/c nude mice. Mice received radiotherapy (8 Gy; arrows) or not. DES \u003cstrong\u003e(e)\u003c/strong\u003e or DMK\u003cstrong\u003e (f)\u003c/strong\u003e was administered to the indicated groups (500 mg/kg, intraperitoneally; n = 6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e, Heatmap of esterified AA (C20:4)-PE molecular species in HK1 xenografts expressing ACSL3 WT or ACSL3 K422R (KR) treated intraperitoneally with vehicle (veh), DES, or DMK (n = 6). Row z-scores were calculated using the mean abundance of each phospholipid species across tumors in each group.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e, Levels of oxygenated phospholipids PE (38:4)-OOH and PE (40:4)-OOH in ACSL3 WT or K422R xenografts receiving radiotherapy and treated intraperitoneally with vehicle (veh), DES, or DMK (n = 6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e, Representative images of immunohistochemical staining for ACSL3 expression which are graded by the staining intensity in 209 NPC tumor tissues. Scale bars, 50 μm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e, Correlations of locoregional recurrence status with the level of ACSL3 detected by IHC. The \u003cem\u003eP\u003c/em\u003e value was determined using the two-tailed χ\u003csup\u003e2\u003c/sup\u003e test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek-l\u003c/strong\u003e, Kaplan Meier analysis of overall survival according to the ACSL3 expression levels in head and neck squamous cell carcinoma\u003cstrong\u003e (k)\u003c/strong\u003e or lung adenocarcinoma \u003cstrong\u003e(l)\u003c/strong\u003e. The \u003cem\u003eP\u003c/em\u003e value was determined using the log-rank test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003em-n\u003c/strong\u003e, Kaplan Meier analysis of overall survival according to the succinate enrichment signature in adrenocortical carcinoma \u003cstrong\u003e(m) \u003c/strong\u003eor kidney renal clear cell carcinoma \u003cstrong\u003e(n)\u003c/strong\u003e. The \u003cem\u003eP\u003c/em\u003e value was determined using the log-rank test.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData are shown as mean ± SEM (a-b, e-f) or mean ± SD (c-d, h) and are representative of at least three independent experiments, one-way ANOVA (c-d, h) or two-way ANOVA (a-b, e-f).\u003c/p\u003e","description":"","filename":"image6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/84d62580ac9d3c1401c001d6.jpeg"},{"id":108808938,"identity":"3ebbb487-b2fc-4715-80c7-c4ce9d614c32","added_by":"auto","created_at":"2026-05-08 15:47:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6386650,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/983766da-9bf0-4bac-bb04-b8f3f0fd64cb.pdf"},{"id":108416403,"identity":"d0a64aa5-e0b1-4145-b4a7-f76cebeef2d6","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4218415,"visible":true,"origin":"","legend":"","description":"","filename":"Extendedfigure.docx","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/2047dfcc48aebf190069e3a7.docx"},{"id":108416406,"identity":"839010c3-013a-4106-bd78-8788cde3a5e4","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":17373,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/3cbf6c730685376a1f4f8db4.docx"},{"id":108803670,"identity":"663e74b3-9253-4298-abdc-e0f33ea6d418","added_by":"auto","created_at":"2026-05-08 15:03:00","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":16919,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.docx","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/d0de3390d99f8b35ce502183.docx"},{"id":108493351,"identity":"dfd53f08-9050-43c9-96f8-d556187e8ee7","added_by":"auto","created_at":"2026-05-05 10:00:01","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":30209,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.docx","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/8ad4abd592cdfe491781053a.docx"},{"id":108416411,"identity":"151aa80e-8224-4b88-a911-6ced509727d0","added_by":"auto","created_at":"2026-05-04 11:37:13","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":570212,"visible":true,"origin":"","legend":"","description":"","filename":"Graphicalabstract.docx","url":"https://assets-eu.researchsquare.com/files/rs-9481973/v1/82b2334001d031605feff49a.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Mechanical stress enhances tumor cell ferroptosis by remodeling succinate metabolism","fulltext":[{"header":"Highlights","content":"\u003cul start=\"50\"\u003e\n \u003cli\u003eTumor microenvironmental stiffness remodels mitochondria-associated membranes (MAMs) to reprogram mitochondrial metabolism\u003c/li\u003e\n \u003cli\u003eMAMs expansion drives succinate accumulation through impaired SDH activity, positioning succinate as a mechanosensitive metabolite\u003c/li\u003e\n \u003cli\u003eSuccinate fuels CPT1A-mediated succinylation of ACSL3 at K422, promoting its degradation and ferroptosis sensitivity\u003c/li\u003e\n \u003cli\u003eSuccinate enrichment signature and lower ACSL3 expression predicts better clinical responses to radiotherapy\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"Introduction","content":"\u003cp\u003eThe mechanical properties of the tumor microenvironment have emerged as active determinants of cancer progression and therapeutic resistance \u003csup\u003e1\u003c/sup\u003e. Increased extracellular matrix (ECM) stiffness, a hallmark of the desmoplastic tumor stroma, promotes tumor initiation, invasion, immune evasion and chemoresistance through mechanotransduction pathways that reprogram gene expression, metabolism and cell fate decisions \u003cstrong\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e. Despite growing appreciation that ECM mechanics influence virtually every aspect of tumor biology, whether and how microenvironmental mechanical forces (MMF) regulate specific modes of regulated cell death to shape radiotherapy response remains poorly understood.\u003c/p\u003e\n\u003cp\u003eFerroptosis, an iron-dependent form of cell death driven by the peroxidation of polyunsaturated fatty acid-containing phospholipids (PUFA-PLs), has been increasingly recognized as a critical effector of radiotherapy-induced tumor suppression\u003csup\u003e3-5\u003c/sup\u003e. The sensitivity of tumor cells to ferroptosis is governed by the balance between PUFA-PLs abundance in cellular membranes and the cellular antioxidant capacity centered on the GSH-GPX4 axis\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e6\u003c/sup\u003e\u003c/strong\u003e. Importantly, phospholipid remodeling enzymes, including ACSL4, which channels polyunsaturated fatty acids into membrane phospholipids, and ACSL3, which counteracts this process by incorporating monounsaturated fatty acids, critically determine the ferroptosis-prone or resistant lipid landscape of tumor cells \u003cstrong\u003e\u003csup\u003e7, 8\u003c/sup\u003e\u003c/strong\u003e. However, whether extracellular mechanical cues can reshape this phospholipid landscape to alter ferroptosis susceptibility has not been explored.\u003c/p\u003e\n\u003cp\u003eThe mitochondria-associated membranes (MAMs) are regions where the ER membrane closely approaches the outer mitochondrial membrane, and these areas are crucial for calcium ion transport, phospholipid transfer, and cell death\u003csup\u003e9\u003c/sup\u003e. The key coupling proteins IP3R and \u0026sigma;1R within MAMs enhance calcium ion exchange between the ER and mitochondria, leading to increased mitochondrial lipid peroxidation and subsequently promoting ferroptosis \u003csup\u003e10\u003c/sup\u003e. Notably, MAMs\u0026nbsp;formation depends on ER tubular morphology maintained by acetylated microtubules \u003csup\u003e11, 12\u003c/sup\u003e, a cytoskeletal feature known to respond to substrate rigidity, raising the possibility that MMF may regulate ferroptosis by remodeling MAMs and reprogramming mitochondrial metabolism.\u003c/p\u003e\n\u003cp\u003eHere, we reveal that the mechanical properties of the tumor microenvironment converge on MAMs to govern ferroptosis sensitivity and radiotherapy response. Using tunable PDMS-based substrates, we show that stiff matrices drive microtubule acetylation, ER sheet-to-tubule remodeling, and MAM expansion, which in turn impairs SDH activity and causes mitochondrial succinate accumulation. Succinate then serves as a metabolic encoder of mechanical information: it fuels CPT1A-mediated succinylation of ACSL3 at K422, triggering ACSL3 degradation and shifting membrane phospholipid composition toward a ferroptosis-prone state. These findings establish MAMs as mechanosensitive organelle interfaces that couple extracellular biomechanics to mitochondrial metabolic reprogramming, and identify the succinate-ACSL3 succinylation axis as a druggable node linking tumor mechanics to radiotherapy efficacy.\u003cstrong\u003e\u003cbr clear=\"all\"\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eHigh matrix stiffness enhances radiosensitivity by promoting ferroptosis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRadiotherapy remains a cornerstone of tumor treatment, yet its efficacy is frequently undermined by the persistence of radioresistant tumor cell populations within heterogeneous microenvironments. Nasopharyngeal carcinoma (NPC), a malignancy treated primarily with radiotherapy owing to its anatomical location and intrinsic radiosensitivity, represents a paradigmatic model for studying radioresistance: despite high initial response rates, local recurrence driven by radioresistant populations remains a major clinical challenge\u003cstrong\u003e\u003csup\u003e13\u003c/sup\u003e\u003c/strong\u003e. We hypothesized that these hallmarks-radioresistance are linked to the MMF and ferroptosis susceptibility of tumor cells. Collagen is the most abundant structural protein in the ECM and is a major determinant of tissue tensile strength \u003cstrong\u003e\u003csup\u003e14\u003c/sup\u003e\u003c/strong\u003e, we performed immunofluorescence staining and for Collagen I\u003cstrong\u003e\u003csup\u003e15\u003c/sup\u003e\u003c/strong\u003eandnanoindenter measurements in pretreatment biopsy specimens from patients with recurrent NPC (rNPC) and non-recurrent NPC (nrNPC). We observed that the dense fibrotic stromal collagen area of NPC exhibited higher stiffness than the low collagen expressing area (Fig.1a). Meanwhile, significant lower expression of Collagen I was observed in rNPC (Fig.1b), suggesting a potential association between matrix stiffness and radioresistance.\u003c/p\u003e\n\u003cp\u003eTo further determine whether MMF regulates the efficacy of tumor radiotherapy, we established an\u003cem\u003e in vitro \u003c/em\u003eexperimental model using custom-fabricated polydimethylsiloxane (PDMS) gels with defined stiffness levels of approximately 1~1000 kPa (Fig.1c). Using the NPC cell lines HK1 and C666-1, we found that cells on the high (500 kPa) stiffness PDMS gel showed highest radiotherapy sensitivity, and the intermediate and lowest radiotherapy sensitivity occured on modest stiffness (50 kPa) and low stiffness (5 kPa) PDMS matrix (Fig.1d-e).\u003c/p\u003e\n\u003cp\u003eWe treated cells with inhibitors targeting distinct cell death pathways, including necrostatin-1 (necroptosis), Z-VAD-FMK (apoptosis), and ferroptosis inhibitors (ferrostatin-1 [Fer-1] and deferoxamine [DFO]), and found that only ferroptosis inhibitors significantly reversed the mechanical force-mediated regulation of radiosensitivity (Fig.1f-g, Extended Data Fig.1a). Interestingly, HKl or C666-1 cells cultured on stiffer matrix showed markedly increased di-oxygenated and tri-oxygenated PUFA-PE species as well as enhanced ferroptosis (Fig.1h-n, Extended Data Fig.1b-e). Consistently, we evaluated the sensitivity of HT-1080, 786-O, A375 and MDA-MB-231 cells toward multiple ferroptosis inducers based on different mechanisms, and also found that cells cultured on stiffer matrix exhibited enhanced ferroptosis sensitivity (Extended Data Fig.1f). \u003c/p\u003e\n\u003cp\u003eNext, we sought to validate these findings \u003cem\u003ein vivo\u003c/em\u003e. LOX, a copper-dependent amine oxidase, catalyzes the intra- and intermolecular covalent crosslinking of collagen, thereby contributing to tissue stiffening; conversely, reduced LOX activity has been shown to alleviate tissue stiffness and prevent fibrosis \u003cstrong\u003e\u003csup\u003e15, 16\u003c/sup\u003e\u003c/strong\u003e. We therefore established HK1 cells with stable LOX overexpression and generated subcutaneous xenograft tumors in BALB/c nu/nu mice. Nanoindenter measurements demonstrated that LOX-overexpressing HK1 tumors displayed greater tissue stiffness than control tumors (Fig. 1O). Notably, after irradiation, LOX-overexpressing tumors were significantly smaller than wild-type tumors. However, treatment with the ferroptosis inhibitor Liproxstatin-1 abrogated this effect (Fig. 1P). Collectively, these results indicate that MMF can enhance tumor cell radiosensitivity and the effect is associated with increased susceptibility to ferroptosis.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eMechanical stiffness reshapes the lipid landscape of tumor cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, we investigated why increased matrix stiffness promotes ferroptosis in tumor cells. The phospholipid composition and its redox state critically dictate ferroptosis susceptibility\u003csup\u003e17\u003c/sup\u003e. We performed RNA-seq to profile HK1 cells cultured on matrices of varying stiffness and found no significant changes in ferroptosis-related gene expression (Extended Data Fig. 2a). Consistently, the abundance of proteins involved in lipid redox homeostasis, as well as intracellular GSH and Fe\u003csup\u003e2+\u003c/sup\u003e levels were comparable across stiffness conditions (Extended Data Fig. 2b-d). However, among phospholipid-remodeling factors, ACSL3 was the only protein that displayed stiffness-dependent regulation, with its expression decreasing as matrix stiffness was reduced (Fig. 2a). \u003c/p\u003e\n\u003cp\u003eACSL3 promotes the activation of monounsaturated fatty acids (MUFAs), such as oleic acid (OA), into their corresponding acyl-CoA species, thereby facilitating their incorporation into membrane phospholipids. This MUFA-driven remodeling reduces the relative abundance of polyunsaturated fatty acid-containing phospholipids (PUFA-PLs), a key determinant of lipid peroxidation propensity and ferroptosis susceptibility (Extended Data Fig. 2e). We next profiled the phospholipid landscape of tumor cells cultured on matrices of different stiffness. Using LC-MS based lipidomic analysis, we found that tumor cells grown on soft matrices displayed a marked reduction in PUFA-PLs compared with those cultured on stiff matrices (Fig. 2b). Notably, this decrease was most prominent in arachidonic acid (AA; C20:4) and adrenic acid (AdA; C22:4) containing phosphatidylethanolamine (PE) and phosphatidylcholine (PC) species, which are strongly linked to ferroptosis execution (Fig. 2c). In addition, we performed airflow-assisted desorption electrospray ionization mass spectrometry imaging (AFADESI-MSI) on pretreatment biopsy specimens from patients with rNPC and nrNPC. Compared with nrNPC, rNPC tumor tissues exhibited a significant reduction in PUFA-containing phospholipids (PUFA-PLs) (Fig. 2d).\u003c/p\u003e\n\u003cp\u003eTo validate the contribution of ACSL3 to MMF-regulated ferroptosis sensitivity, we knockouted \u003cem\u003eACSL3\u003c/em\u003e in HK1, C666-1, and HT1080 cells and cultured these cells on matrices of different stiffness (Extended Data Fig. 2g). Upon treatment with ferroptosis inducers, ACSL3 knockout markedly abrogated the ferroptosis-resistant phenotype observed in cells grown on soft substrates, thereby restoring ferroptosis sensitivity under low-stiffness conditions (Fig. 2e-h, Extended Data Fig. 2h-i). In addition, ACSL3 knockout also markedly reversed the radioresistant phenotype of HK1 cells cultured on soft matrices (Fig. 2i). Collectively, these results indicate that MMF may modulate tumor cell sensitivity to ferroptosis and radiotherapy by altering ACSL3 expression.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eMMF regulates ferroptosis through MAMs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, we investigated how MMF regulates ACSL3 protein abundance. We previously showed that ACSL3 mRNA levels were unchanged across stiffness conditions, indicating that MMF did not regulate ACSL3 at the transcriptional level (Extended Data Fig. 2a). Interestingly, we found that ACSL3 showed a decreased half-life in HK1 or C666-1 cells cultured on stiffer matrix (Fig. 3a, Extended Data Fig. 3a). More importantly, the degradation of ACSL3 in HK1 cells cultured on stiff matrix could be blocked by proteasome inhibitors (MG132; carfilzomib) but not lysosome inhibitors (Baf A1; chloroquine), indicating that MMF regulates ACSL3 protein levels depending on proteasome (Fig. 3b). Ubiquitination is a key step for the proteasome-dependent degradation of protein. Consistently, we found that the ubiquitination level of ACSL3 was increased in the cells cultured on stiffer matrix (Fig. 3c, Extended Data Fig. 3b).\u003c/p\u003e\n\u003cp\u003eACSL3 is primarily localized to the endoplasmic reticulum (ER), lipid droplets (LD) and MAMs \u003cstrong\u003e\u003csup\u003e18\u003c/sup\u003e\u003c/strong\u003e. Recent studies have highlighted an essential role for MAMs in ferroptosis execution \u003cstrong\u003e\u003csup\u003e10, 19\u003c/sup\u003e\u003c/strong\u003e. We therefore investigated whether MAMs contribute to the MMF-dependent regulation of ACSL3 expression. We observed under electron microscopy that HK1 cells cultured on soft matrix exhibited fewer MAM structures compared to those grown on stiff matrix (Fig. 3d). Additionally, we generated HK1 cells stably expressing mitochondria-targeted GFP (Mito-GFP) and the ER marker Sec61\u0026beta;-mCherry, and cultured them on matrices of defined stiffness. Immunofluorescence analysis showed reduced apparent co-localization between Mito-GFP and Sec61\u0026beta;-mCherry in cells grown on soft substrates, indicating decreased ER-mitochondria proximity under low-stiffness conditions (Fig. 3e). Notably, we stably transfected the HK1 cell line with plasmids encoding GFP1-10-ERT (targeting the ER membrane) and GFP11-TOMM70 (targeting the outer mitochondrial membrane) to construct MAM-GFP reporter cells \u003csup\u003e20, 21\u003c/sup\u003e (Extended Data Fig. 3c). Flow cytometry and immunofluorescence analysis of GFP expression in HK1 MAM-GFP reporter cells revealed that tumor cells cultured on soft matrix formed fewer MAM structures (Fig. 3f-g). These data indicate that the formation of MAMs is reduced in tumor cells cultured on soft matrix. \u003c/p\u003e\n\u003cp\u003eWe then employed a rapamycin-induced FKBP-FRB interaction system by co-transfected with CYB5A-GFP-FKBP and TOM20-mCherry-FRB in HK1 cells and C666-1 cells to form chemically inducible MAMs (MAMs-inducing cells) \u003csup\u003e22, 23\u003c/sup\u003e. This also allowed protein GFP and mCherry to localize to MAMs (Extended Data Fig. 3d). Interestingly, MAM-inducing cells exhibited lower ACSL3 protein abundance than parental cells (Fig. 3h). Under treatment with AA-d8, MAM-inducing cells contained more AA-d8-containing PE or PC species compared to parental cells (Fig. 3i). Upon exposure to ferroptosis inducers, MAM-inducing cells showed higher levels of di-oxygenated and tri-oxygenated PUFA-PE species and displayed markedly enhanced ferroptosis compared with their counterparts (Fig. 3j-l). Furthermore, upon radiotherapy treatment, MAM-inducing cells exhibited increased cell death compared with control cells (Fig. 3m).\u003c/p\u003e\n\u003cp\u003ePrevious studies have shown that the tubular ER participates in the contacts between the ER and various other organelles (lysosomes, mitochondria, etc.) \u003csup\u003e11\u003c/sup\u003e. Rab10 and the tubular ER-shaping protein reticulon 4 (RTN4) are key proteins that maintain the tubular ER, while Climp63 induces sheet-like ER \u003csup\u003e23\u003c/sup\u003e. In MAM-GFP reporter cells, knockdown of Rab10/RTN4 or overexpression of Climp63 markedly reduced MAM formation and enhanced ACSL3 protein level (Fig. 3n-p, Extended Data Fig. 3e-h). Concurrently, Rab10 or RTN4 knockdown reduced the accumulation of di-oxygenated and tri-oxygenated PUFA-PE species and markedly attenuated ferroptosis relative to control cells (Extended Data Fig. 3i-l). \u003c/p\u003e\n\u003cp\u003eThe Tubular ER dynamics depend on microtubules \u003csup\u003e12\u003c/sup\u003e, and MAMs have been reported to occur on acetylated microtubules \u003csup\u003e24\u003c/sup\u003e. Notably, microtubule acetylation and its acetyltransferase \u0026alpha;-tubulin acetyltransferase 1 (\u0026alpha;TAT1) can respond to substrate rigidity, providing a potential link between MMF and MAMs \u003csup\u003e25\u003c/sup\u003e. We observed a significant reduction in microtubule acetylation levels in HK1 cells and C666-1cells cultured on soft matrices (Extended Data Fig. 4a). Treatment with the microtubule deacetylase inhibitor Tubacin to induce microtubule acetylation led to a marked increase in the abundance of MAMs and reduced ACSL3 protein level (Extended Data Fig. 4b-c). Conversely, knockdown of \u0026alpha;-tubulin acetyltransferase 1 (\u0026alpha;TAT1), which diminishes microtubule acetylation, markedly reduced MAM formation and increased ACSL3 protein level (Extended Data Fig. 4d-f). Tubacin treatment enhanced ferroptosis and promoted the accumulation of ferroptosis-associated oxidized phospholipids, whereas \u0026alpha;TAT1 knockdown exerted the opposite effects (Extended Data Fig. 4g-l). Collectively, these results suggest that stronger MMF may modulate ferroptosis susceptibility by driving ER remodeling and promoting MAMs formation.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eMitochondrial succinate accumulation facilitates ferroptosis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, we investigated how MMF and MAM remodeling regulate ACSL3 protein abundance. MAMs represent specialized hubs for Ca\u0026sup2;⁺ transfer and lipid exchange between the ER and mitochondria, and alterations in MAM dynamics can influence mitochondrial fission\u0026ndash;fusion balance as well as mitochondrial metabolism. We reasoned that MMF-driven MAM remodeling might reprogram the TCA cycle and the respiratory chain. Accordingly, we assessed mitochondrial function in tumor cells cultured on matrices of defined stiffness, as well as in MAM-inducing cells and their corresponding control cells. Seahorse analysis revealed that cells cultured on stiff matrix showed no statistically significant change in their basal respiration, whereas maximal respiratory capacity and spare respiratory capacity (SRC) were impaired (Fig. 4a-b). Consistently, MAM-inducing cells exhibited a modest reduction in basal oxygen consumption, with a pronounced decrease in maximal respiratory capacity and SRC (Extended Data Fig. 5a-b). Importantly, mitochondria-related metabolite profiling revealed marked succinate accumulation in tumor cells cultured on stiff matrices (Fig. 4c). Succinate is subsequently oxidized to fumarate by SDH, a key enzyme that links the TCA cycle to oxidative phosphorylation (Extended Data Fig. 5c). Further analysis also revealed that succinate-to-fumarate ratio was elevated both in cells grown on stiff substrates and in MAM-inducing cells, indicating a stiffness-dependent regulation of mitochondrial TCA cycle intermediates (Fig. 4d-e). Intriguingly, AFADESI-MSI on tissue specimens from 80 patients with nasopharyngeal carcinoma showed that succinate was significantly elevated in tumor cells from patients with recurrent disease (Fig. 4f). \u0026alpha;-Ketoglutarate dehydrogenase (OGDH) catalyzes the conversion of \u0026alpha;-KG into succinyl-CoA, which is subsequently converted into succinate by succinyl-CoA synthetase (SCS) (Extended Data Fig. 5c). Upon supplementation with AA-d8, tumor cells treated with dimethyl \u0026alpha;-ketoglutarate (DMK) or dimethyl succinate (DES) incorporated higher levels of AA-d8 into PE and PC species than parental cells (Fig. 4g). Consistently, treatment with DMK or DES enhanced ferroptosis and radiotherapy-induced cell death, while also promoting the accumulation of ferroptosis-associated oxidized phospholipids in HK1 and C666-1 cells cultured on soft matrices (Fig. 4h-k, Extended Data Fig. 5d-h). Moreover, treatment with DMK or DES abrogated the elevated protein abundance of ACSL3, as well as its prolonged half-life and reduced ubiquitination observed in tumor cells cultured on soft matrices (Fig. 4l-n, Extended Data Fig. 5i-q). These results suggest that elevated intracellular succinate promotes ACSL3 degradation via the ubiquitin-proteasome pathway.\u003c/p\u003e\n\u003cp\u003eNext, we investigated the mechanisms underlying succinate accumulation in tumor cells cultured on stiff matrices. Proteomic profiling of HK1 cells cultured on matrices of different stiffness revealed no significant changes in the abundance of succinate-associated metabolic enzymes, including OGDH, SCS, and SDH (Extended Data Fig. 6a). Consistently, western blotting showed comparable levels of the SCS subunits SUCLA2, SUCLG1, and SUCLG2, as well as OGDH and the SDH subunits SDHA and SDHB, across stiffness conditions (Extended Data Fig. 6b). Succinate in tumor cells primarily originates from the mitochondrial TCA cycle and can also be acquired from the extracellular environment. Within mitochondria, succinate is exported to the cytosol via the inner membrane transporter SLC25A10 and the outer membrane channel VDAC. In addition, tumor cells can import extracellular succinate through the plasma membrane transporter SLC13A3, and release it into the extracellular space via MCT1 (Extended Data Fig. 6c). RT-PCR analysis showed no significant differences in the expression levels of SLC25A10, SLC13A3, or MCT1 in tumor cells cultured on matrices of different stiffness (Extended Data Fig. 6d). Because the succinate-to-fumarate ratio was elevated in cells cultured on stiff substrates (Fig. 4d), we hypothesized that SDH activity might be impaired under the conditions of stiff substrates. Mitochondrial isolation followed by enzymatic activity assays showed that SDH activity was lower in tumor cells grown on stiff matrices than in those cultured on soft substrates (Extended Data Fig. 6e-f). Similarly, SDH activity was also reduced in MAM-inducing cells (Extended Data Fig. 6g-h). These findings suggest that diminished SDH activity may contribute to succinate enrichment in tumor cells cultured on stiff matrices.\u003c/p\u003e\n\u003cp\u003eTreatment of HK1 or C666-1 cells with the SDH inhibitors dimethyl malonate (DMM) and 3-nitropropionic acid (3-NPA)-the latter irreversibly inactivating SDHA by covalently binding to an arginine residue within its catalytic core effectively abrogated the soft matrix-induced stabilization of ACSL3, restoring ACSL3 ubiquitination and reducing ACSL3 protein abundance (Extended Data Fig. 6i-k), while concomitantly sensitizing tumor cells to ferroptosis (Extended Data Fig. 6l-m). In contrast, pharmacological inhibition of OGDH with CPI-613 markedly reduced ACSL3 ubiquitination and attenuated ferroptosis in the same tumor cell lines (Extended Data Fig. 6j-m). These results suggest a critical role of succinate metabolism in regulating matrix stiffness-induced ferroptotic sensitivity in tumor cells. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCPT1A mediates ACSL3 succinylation to promote its degradation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccumulation of succinate or \u0026alpha;-KG can increase intracellular succinyl-CoA levels, thereby enhancing protein lysine succinylation\u003cstrong\u003e\u003csup\u003e26\u003c/sup\u003e\u003c/strong\u003e. Next, we investigated whether lysine succinylation participates in regulating ACSL3 protein abundance. Using immunoprecipitation (IP) followed by immunoblotting with a pan-anti-lysine succinylation antibody, we detected ACSL3 succinylation across multiple tumor cell types (Fig. 5a, Extended Data Fig. 7a). Notably, ACSL3 exhibited higher levels of succinylation in tumor cells cultured on stiff matrices (Fig. 5b, Extended Data Fig. 7b). DES or DMK supplementation further increased ACSL3 succinylation while concomitantly reducing ACSL3 protein abundance (Fig. 5c, Extended Data Fig. 7c). Conversely, lowering succinyl-CoA levels by glycine treatment restored ACSL3 protein expression (Fig. 5d, Extended Data Fig. 7d). This suggests that succinate-driven succinylation of ACSL3 promotes ACSL3 degradation. Liquid chromatography-mass spectrometry (LC-MS) analysis identified two succinylation sites on ACSL3 in HK1 cells, at lysine 422 (K422) and lysine 660 (K660) (Fig. 5e, Extended Data Fig. 7e). Both sites were highly conserved among different species (Extended Data Fig. 7f-g). We re-expressed inducible ACSL3 mutants with K422/K660 replaced by arginine (K422R/K660R) in \u003cem\u003eACSL3\u003c/em\u003e\u003csup\u003e-/-\u003c/sup\u003e HK1 cells (Extended Data Fig. 7h). However, compared to ACSL3-WT, only ACSL3-K422R showed reduced succinylation (Fig. 5f). In the presence of succinyl-CoA, substituting K422 with glutamate (K422E) mimicked succinylation could downregulate ACSL3 levels, whereas K422R had no such effect (Fig. 5g, Extended Data Fig. 7i), indicating that DES and DMK primarily target K422. Consistently, the K422R mutation abrogated the increase in ACSL3 protein abundance observed in tumor cells cultured on soft matrices (Fig. 5h). Moreover, the K422E mutation increased ACSL3 ubiquitination and shortened its half-life, whereas K422R exerted the opposite effects (Extended Data Fig. 7j-k). \u003c/p\u003e\n\u003cp\u003eConsistent with its enhanced stability, re-expression of ACSL3-K422R markedly reshaped the cellular phospholipid landscape relative to ACSL3-WT. Specifically, overall phospholipid (PL) unsaturation was significantly reduced: MUFA-PL species increased by 6.0%, whereas PUFA-PL and saturated fatty acid (SFA)-PL species decreased by 3.4% and 2.7%, respectively (Fig. 5i). Moreover, upon supplementation with AA-d8, cells expressing ACSL3-K422R incorporated substantially less AA-d8 into phosphatidylethanolamine (PE) and phosphatidylcholine (PC) (Fig. 5j). Functionally, ACSL3-K422R re-expression conferred robust protection against RSL3/FIN56-induced ferroptosis, accompanied by reduced accumulation of di-oxygenated and tri-oxygenated PUFA-PE species (Extended Data Fig. 7l-p). \u003c/p\u003e\n\u003cp\u003eNext, we conducted molecular dynamics simulation to explore the effect of K422 succinylation in ACSL3 stability. Compared with the WT model, the succinylated ACSL3 model underwent a marked conformational rearrangement during 100~200 ns before reaching a stable state, whereas the WT protein remained relatively stable throughout the simulation (Extended Data Fig. 8a). Root-mean-square fluctuation (RMSF) analysis showed that these differences were mainly concentrated in Domain 1 and Domain 2, both of which exhibited markedly greater fluctuations in the succinylated model than in the WT model (Extended Data Fig. 8b). More importantly, lysine residues in the succinylated model displayed increased solvent-accessible surface area (SASA), particularly in the C-terminal region encompassing Domain 1 and Domain 2, where nearly all lysine residues displayed increased SASA in the succinylated state (Extended Data Fig. 8c). Consistently, quantification of the number of water molecules within 5, 7, 10, and 15 \u0026Aring; of the ACSL3 centroid further confirmed greater solvent exposure in the succinylated model (Extended Data Fig. 8d-g). Given that increased solvent exposure of lysine residues is generally associated with a higher likelihood of ubiquitination, these results suggest that ACSL3 succinylation may promote its ubiquitination by increasing the accessibility of lysine residues. In terms of protein conformation, Domain 1 and Domain 2, which were adjacent in the WT model, became separated in the succinylated state, generating a prominent cleft that likely facilitated solvent penetration and increased lysine accessibility, thereby potentially enhancing the likelihood of ubiquitination (Fig. 5k, Extended Data Fig. 8h). In addition, the R424-T34 and R424-T37 hydrogen bonds present in the WT model, which connect Domain 1 to the N-terminal region, were disrupted in the succinylated model, where R424 instead formed a stable salt bridge with succinylated K422 (SL422) (Fig. 5l). Collectively, these changes suggest that K422 succinylation destabilizes ACSL3 conformation, increases lysine solvent exposure, and thereby promotes ubiquitination.\u003c/p\u003e\n\u003cp\u003eSuccinylation is a dynamic and reversible modification regulated by the balance between succinyltransferases (KAT2A, CPT1A, SUCLA2, OXCT1, etc.) \u003csup\u003e27\u003c/sup\u003e,\u003csup\u003e28\u003c/sup\u003e,\u003csup\u003e29, 30\u003c/sup\u003e and desuccinylases (SIRT5, SIRT7) \u003csup\u003e31, 32\u003c/sup\u003e. To identify ACSL3-interacting proteins, we performed quantitative mass spectrometry on anti-ACSL3 immunoprecipitates from HK1 cell lysates. Among the top 100 most enriched proteins, we identified CPT1A, a succinyltransferase. Subcellular fractionation revealed that both ACSL3 and CPT1A were present in the MAM fraction (Extended Data Fig. 9a). Co-immunoprecipitation assays further confirmed a physical interaction between ACSL3 and CPT1A (Fig. 5m, Extended Data Fig. 9b), and immunofluorescence analysis showed their cytoplasmic co-localization (Fig. 5n). Notably, although CPT1A protein abundance remained unchanged across stiffness conditions (Extended Data Fig. 9c), the interaction between CPT1A and ACSL3 was markedly reduced in tumor cells cultured on soft matrices (Fig. 5o, Extended Data Fig. 9d). CPT1A knockdown increased ACSL3 protein abundance in tumor cells cultured on stiff matrices (Extended Data Fig. 9e-f), while reducing ACSL3 ubiquitination and succinylation and prolonging its half-life (Extended Data Fig. 9g-i). Moreover, DMK or DES supplementation failed to reduce ACSL3 protein abundance in CPT1A-knockdown cells (Fig. 5p). In contrast, re-expression of inducible CPT1A-WT in CPT1A-knockout cells restored succinate-dependent ACSL3 succinylation and reduced ACSL3 protein levels (Extended Data Fig. 9j-k). Next, we re-expressed doxycycline-inducible CPT1A constructs encoding CPT1A-WT, CPT1A-H473A (inactive), or CPT1A-G710E (succinyltransferase-active) in CPT1A-knockout cells \u003csup\u003e27\u003c/sup\u003e. Re-expression of CPT1A-WT or CPT1A-G710E, but not CPT1A-H473A, markedly reduced ACSL3 protein abundance (Extended Data Fig. 9l). Consistently, CPT1A-WT overexpression decreased the level of ACSL3-WT, whereas it had no detectable effect on the AzzzCSL3-K422R mutant (Extended Data Fig. 9m). Moreover, CPT1A-WT and CPT1A-G710E markedly increased ACSL3 succinylation, whereas CPT1A-H473A had no effect (Extended Data Fig. 9n), indicating CPT1A functions as a succinyltransferase for ACSL3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSuccinate-ACSL3(K422) axis promotes ferroptosis\u003cem\u003e in vivo \u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, to validated the effect of ACSL4 K422 succinylation in \u003cem\u003ein vivo \u003c/em\u003emodels, we constructed tumor-bearing mice by subcutaneously injecting ACSL3-WT HK1 cells, ACSL3-K422R HK1 cells and ACSL3-K422E HK1 cells. Following irradiation, tumors overexpressing ACSL3-K422R were significantly larger than those overexpressing ACSL3-WT (Fig. 6a), whereas ACSL3-K422E overexpression resulted in smaller tumors relative to ACSL3-WT (Fig. 6b). Importantly, tumor cells isolated from irradiated ACSL3-K422R xenografts contained lower levels of AA-containing PE species as well as di-oxygenated arachidonoyl- and adrenoyl-PE species compared with cells derived from ACSL3-WT and ACSL3-K422E tumors (Fig. 6c-d). Further corroborating our findings, DES or DMK treatment suppressed tumor growth in the ACSL3-WT overexpression group but had little effect in ACSL3-K422R tumors (Fig. 6e-f). Consistently, lipidomic profiling revealed that tumor cells isolated from irradiated ACSL3-WT tumors treated with DES or DMK exhibited increased levels of AA-containing PE species, as well as di-oxygenated arachidonoyl- and adrenoyl-PE species (Fig. 6g-h). Finally, we validated our findings in clinical specimens. In a retrospective radiotherapy cohort of patients with NPC (n=209; SYSUCC NPC cohort; Supplementary Table 1), immunohistochemical staining with an anti-ACSL3 antibody showed that lower ACSL3 expression was associated with a reduced risk of local recurrence after radiotherapy (Fig. 6i-j). Consistently, in public datasets, low ACSL3 expression was associated with improved overall survival in patients with head and neck squamous cell carcinoma and lung adenocarcinoma (Fig. 6k-l). We further defined a nine-gene succinate enrichment signature and found that high signature scores predicted favorable survival in adrenocortical carcinoma and kidney renal clear cell carcinoma (Supplementary Table 2, Fig. 6m-n). Collectively, these findings support a pivotal role for the succinate-ACSL3(K422) axis in regulating ferroptosis sensitivity and indicate that low ACSL3 expression together with high succinate enrichment is associated with improved clinical outcomes.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we uncovered a mechanistic link between microenvironmental mechanical cues and ferroptosis sensitivity in tumor cells. We found that stiff matrices enhanced ferroptosis and radiotherapy sensitivity by enhancing microtubule acetylation, driving ER sheet-to-tubule remodeling, and promoting MAM formation. This structural reprogramming rewired mitochondrial metabolism, leading to succinate accumulation and reduced SDH activity. Elevated succinate in turn facilitated CPT1A-dependent succinylation of ACSL3 at K422, promoting its ubiquitination and degradation, thereby reshaping phospholipid composition toward a ferroptosis-prone state. We further confirmed the functional importance of this pathway in cellular and in vivo models. More importantly, low ACSL3 expression and high succinate enrichment were associated with favorable clinical outcomes. Collectively, our data identify MAMs as mechanosensitive intracellular hubs that couple tumor biomechanics to mitochondrial metabolism, phospholipid remodeling, ferroptosis, and radiotherapy response.\u003c/p\u003e\n\u003cp\u003eThe mechanical properties of the tumor microenvironment are increasingly recognized as active regulators of cancer cell behavior and therapeutic response. ECM composition and stiffness dictate regional susceptibility to tumor initiation\u003csup\u003e33\u003c/sup\u003e, promote tumor progression through mechanosignaling-driven exosome secretion\u003csup\u003e34\u003c/sup\u003e, and confer chemoresistance in pancreatic cancer organoids via CD44-hyaluronan signaling \u003csup\u003e35\u003c/sup\u003e. Beyond tumor cells, biomechanical stress drives CD8\u003csup\u003e+\u003c/sup\u003eT cell exhaustion through the Piezo1/CREB/Osr2 axis\u003csup\u003e36\u003c/sup\u003e. Notably, ECM remodeling alters mitochondrial homeostasis in an evolutionarily conserved manner via TGF-\u0026beta;-induced mitochondrial fission and UPR^mt\u003csup\u003e37\u003c/sup\u003e, establishing a direct link between extracellular mechanics and mitochondrial function. However, whether ECM stiffness regulates specific mitochondrial metabolic outputs to control ferroptosis has remained unknown. Our study addresses this gap by revealing that substrate stiffness remodels MAMs through microtubule acetylation and ER morphological transition, driving succinate accumulation and ACSL3 succinylation, thereby positioning MAMs as intracellular mechanosensitive hubs that translate extracellular physical cues into mitochondrial metabolic reprogramming and ferroptosis regulation.\u0026nbsp;The convergence of mechanical signaling on ACSL3 reflects a structural inevitability rather than a stochastic event. ACSL3 is the only known ferroptosis-protective phospholipid remodeling enzyme that resides at MAMs, placing it at the precise subcellular interface where stiffness-driven organelle remodeling, CPT1A-mediated succinylation and succinyl-CoA availability intersect. The evolutionary conservation of K422 at a structurally critical domain hinge further ensures that its succinylation produces maximal conformational destabilization. Thus, ACSL3 represents a uniquely vulnerable node where mechanical, metabolic and structural determinants converge to control ferroptosis sensitivity.\u003c/p\u003e\n\u003cp\u003eThis mechanosensitive role of MAMs extends current understanding of ER-mitochondria contacts in ferroptosis. Recent studies have shown that lipid peroxidation during ferroptosis initiates at the ER membrane and progressively spreads to other compartments \u0026nbsp;\u003csup\u003e19, 38\u003c/sup\u003e, and that MAMs serve as the primary hotspots of proferroptotic phospholipid peroxide formation by stabilizing MAMs enhances ferroptosis, whereas untethering them confers protection \u003csup\u003e39\u003c/sup\u003e. The MAMs-resident chaperone \u0026sigma;1R facilitates ER-to-mitochondria Ca\u0026sup2;⁺ transfer during ferroptosis, and disruption of MAMs blocks mitochondrial ROS production and ferroptosis execution \u003csup\u003e10\u003c/sup\u003e. Beyond tumor cells, ER\u0026ndash;mitochondria contacts regulate immune cell fitness: MFN2-SERCA2 interaction sustains mitochondrial Ca\u0026sup2;⁺ homeostasis in tumor-infiltrating CD8\u003csup\u003e+\u003c/sup\u003eT cells \u003csup\u003e40\u003c/sup\u003e, and linoleic acid enhances anti-tumor immunity by promoting MAMs formation and mitochondrial energetics \u003csup\u003e41\u003c/sup\u003e. While these studies collectively establish MAMs as signaling hubs integrating Ca\u0026sup2;⁺ dynamics, lipid transfer and redox homeostasis, they have focused on MAMs as either executors of chemically induced ferroptosis or regulators of immune cell metabolism. Our study demonstrates that substrate stiffness drives MAMs expansion independent of ferroptosis-inducing stimuli, and that the downstream consequence extends beyond Ca\u003csup\u003e2+\u003c/sup\u003eoverload or lipid peroxidation to include SDH dysfunction, succinate accumulation and CPT1A-mediated ACSL3 succinylation, a metabolic axis not previously linked to MAMs.\u003c/p\u003e\n\u003cp\u003eThis finding also places our work within the rapidly expanding field of succinate and succinylation signaling. Succinate accumulation drives gut inflammation by switching FOXP3 from succinylation to ubiquitination-mediated degradation \u003csup\u003e42\u003c/sup\u003e, while succinyl-CoA-mediated succinylation of PD-L1 by CPT1A promotes its degradation and enhances anti-tumor immunity \u003csup\u003e43\u003c/sup\u003e. Sustained succinate exposure preserves CD8\u003csup\u003e+\u003c/sup\u003eT cell stemness through BNIP3-mediated mitophagy \u003csup\u003e44\u003c/sup\u003e. SIRT5-mediated desuccinylation of TBK1 suppresses inflammatory signaling in aged skeletal muscle \u003csup\u003e45\u003c/sup\u003e, and metabolism-driven succinylation governs resource allocation for antibiotic resistance in bacteria \u003csup\u003e46\u003c/sup\u003e. At the succinate-fumarate node, fumarate regulates mitophagy through succination of Parkin cysteine residues \u003csup\u003e47\u003c/sup\u003e. These studies reveal protein succinylation as a regulatory layer connecting metabolic status to diverse cellular outcomes, yet whether succinylation responds to extracellular physical cues has remained entirely unexplored. Our study demonstrates that substrate stiffness, through EMCS remodeling and SDH dysfunction, drives succinate accumulation that functions as a metabolic encoder of mechanical information. The subsequent CPT1A-mediated succinylation and degradation of ACSL3 directly links mechanotransduction to ferroptosis sensitivity, establishing succinate as a mechanosensitive metabolite and positioning succinylation as a previously unrecognized post-translational mechanism in the mechanical control of cell fate.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eClinical specimens\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe collected 209 paraffin-embedded specimens from patients with locoregionally advanced nasopharyngeal carcinoma (NPC) at Sun Yat-sen University Cancer Center, Guangzhou, China (SYSUCC cohort). All patients were treatment-na\u0026iuml;ve at the time of biopsy. Tumor staging was determined according to the 8th American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) tumor-node-metastasis (TNM) classification system. All patients subsequently received radical radiotherapy. This study was approved by the Institutional Review Board of SYSUCC (G2025-179-01; G2025-108-01), and the requirement for written informed consent was waived by the ethics committee.\u003c/p\u003e\n\n\u003cp id=\"_Toc128343217\"\u003e\u003cstrong\u003eCell culture\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHuman NPC cell lines HK1 and C666-1 were kindly provided by M.-S. Zeng at Sun Yat-sen University Cancer Center (SYSUCC). HT1080 (fibrosarcoma), 786-O (clear cell renal cell carcinoma), A375 (melanoma), MDA-MB-231 (breast cancer), and HEK293T cells were obtained from the China Center for Type Culture Collection. Cells were cultured in RPMI 1640 or DMEM (Invitrogen) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin. All cell lines were routinely screened for mycoplasma contamination.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003ePDMS culture of tumor cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp id=\"_Toc128343218\"\u003ePDMS substrates were prepared by mixing dimethylsiloxane monomer (SYLGARD 184, Dow Corning) with the corresponding curing agent as previously described\u003cstrong\u003e\u003csup\u003e36, 48\u003c/sup\u003e\u003c/strong\u003e. To generate matrices with tunable stiffness, the base polymer and cross-linker were mixed at approximate ratios of 50:1 (softer), 40:1 (soft), 30:1 (modest), and 20:1 (stiff), reflecting the viscous nature of the base polymer. After thorough mixing and degassing under vacuum, the PDMS elastomers were cast into 12-well plates, 60-mm culture dishes, or onto glass slides for stiffness calibration, and then cured at 60 \u0026deg;C for 4 h. PDMS forms a planar surface with hydrophobic properties that permit protein adsorption for cell culture applications. Substrate stiffness was measured by nanoindenter (Optics11 life). In this study, the Young\u0026rsquo;s modulus of the PDMS substrates was approximately 5-10 kPa for the 50:1 formulation, 20-25 kPa for 40:1, 100 kPa for 30:1, and 500 kPa for 20:1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReagents\u003c/strong\u003e\u003c/p\u003e\n\u003cp id=\"_Toc128343220\"\u003eRSL3 (HY-100218A, MCE), FIN56 (S8254, Selleck), FINO2 (25096, Cayman), Liproxstatin-1(HY-12726, MCE), Doxycycline (HY-N0565, MCE), MG132(S2619, Selleck), carfilzomib (S2853, Selleck), Bafilomycin A1 (S1413, Selleck), Chloroquine (S6999, Selleck), Ferrostatin-1 (Fer1,HY-100579, MCE), Arachidonic Acid-d8 (CAC-390010-5, Cayman), Z-VAD-FMK (S7023, Selleck) and Necrostatin-1 (S8037, Selleck), rapamycin (HY-10219, MCE), Tubacin (GC16386, GLPBIO), Dimethyl 2-ketoglutarate (28394, Cayman), dimethyl succinate (HY-Y0808, MCE), dimethyl malonate (HY-Y1787, MCE), 3-nitropropionic acid (S3652, Selleck), CPI-613(S2776, Selleck), Glycine (HY-Y0966, MCE), Succinyl-CoA (S1129, Sigma-Aldrich).\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eCRISPR-Cas9-mediated genome editing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSingle-guide RNAs (sgRNAs) targeting the indicated genes were designed using Benchling, and the corresponding target sequences are listed in the Supplementary Table1. Annealed sgRNA oligonucleotides were ligated into the PX458 vector (Addgene) following BbsI digestion. Cells were seeded at approximately 60% confluence and transfected with 1 \u0026mu;g of sgRNA-containing plasmid. The medium was replaced 8 h after transfection. Two or three days later, cells were dissociated with trypsin, and GFP-positive cells were sorted by flow cytometry. Sorted cells were then subjected to single-cell cloning in 96-well plates. Gene knockout was confirmed by western blotting.\u003c/p\u003e\n\n\u003cp id=\"_Toc128343223\"\u003e\u003cstrong\u003eLentivirus-mediated gene transfer\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor lentiviral production, HEK293T cells were co-transfected with the pSin-EF2-Puro-based construct, psPAX2, and pMD2.G. The medium was replaced with UltraCULTURE medium (Lonza) at 8\u0026thinsp;h after transfection. Viral supernatants were harvested at 48\u0026thinsp;h, passed through a filter, and applied to the indicated tumor cell lines for overnight infection at an MOI of 100. Western blotting was used to verify ectopic protein expression in the transduced cells.\u003c/p\u003e\n\n\u003cp\u003e\u003cspan id=\"_Toc128343226\"\u003e\u003cstrong\u003eImmunoblotting analysis\u003c/strong\u003e\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003eProtein lysates were prepared using RIPA buffer (Beyotime Biotechnology) containing EDTA-free Protease Inhibitor Cocktail (Beyotime Biotechnology). Following separation by SDS-PAGE (GenScript), proteins were transferred to nitrocellulose membranes. Membranes were then blocked and probed with the indicated primary antibodies (Supplementary Table 2) overnight at 4\u0026thinsp;\u0026deg;C, followed by incubation with HRP-linked secondary antibodies. Immunoreactive bands were visualized by enhanced chemiluminescence (Thermo Fisher Scientific).\u003c/p\u003e\n\n\u003cp id=\"_Toc111828714\"\u003e\u003cstrong\u003eImmunofluorescence analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp id=\"_Toc128343228\"\u003eCells grown for immunofluorescence analysis were fixed with 4% paraformaldehyde for 15\u0026thinsp;min at room temperature and washed three times with PBS. After permeabilization in 0.1% Triton X-100/PBS for 15\u0026thinsp;min and three additional washes with PBST, samples were blocked and incubated with the indicated primary antibodies (Supplementary Table 2). Appropriate Alexa Fluor-conjugated secondary antibodies (Invitrogen) were then applied for 1\u0026thinsp;h at room temperature. Nuclei were stained with DAPI, and images were captured on a Zeiss LSM 880 confocal laser scanning microscope \u003csup\u003e49\u003c/sup\u003e.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eIHC analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParaffin-embedded samples were sectioned at 3\u0026thinsp;\u0026mu;m. Antigen retrieval was performed in 0.01\u0026thinsp;M citrate buffer (pH 6.0) using a pressure cooker for 15-20\u0026thinsp;min. Sections were then incubated with the indicated primary antibodies (Supplementary Table 2) overnight at 4\u0026thinsp;\u0026deg;C, followed by DAB-based detection (Dako) on the next day according to the manufacturer\u0026rsquo;s instructions. Images were captured using an AxioVision Rel.4.6 computerized image analysis system (Zeiss). All sections were scored by two experienced pathologists according to the immunoreactive score (IRS) system \u003csup\u003e50\u003c/sup\u003e. The staining intensity score was defined as follows: 0, negative staining; 1, weak staining; 2, moderate staining; and 3, strong staining. The positive rate score was defined as follows: 1, \u0026lt; 10%; 2, 10%-35%; 3, 35%-70%; 4, \u0026gt; 70%. The total score of indicated proteins was calculated as staining intensity score multiply by positive rate score. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eIntracellular ferrous iron (Fe\u003csup\u003e2+\u003c/sup\u003e) measurements \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe relative iron concentration in cell lysates was determined with Iron Assay kit (Abcam, #ab83366) and the experiments were performed according to the manufacturer\u0026rsquo;s instructions.\u003c/p\u003e\n\u003cp id=\"_Toc128343229\"\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGSH/GSSG ratio assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGSH and GSSG assays were performed, and the GSH/GSSG ratio was calculated using a GSH/GSSG Detection Assay (Abcam, #ab138881) according to the manufacturer\u0026rsquo;s instructions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFerroptosis assay \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan id=\"_Toc111828689\"\u003eFor cell viability assay, 2000 cells were plated in replicates in 96-well plates one day before adding the indicated drug. Cell viability was assessed 48h after drug treatment by Cell Counting Kit-8 (TargetMol) and normalized to an untreated control. Dose\u0026ndash;response curves and half-maximal inhibitory concentration (IC50) values were generated using GraphPad Prism. To quantify ferroptotic cell death, the proportion of 7-AAD-positive tumor cells was determined after exposure to the indicated drugs or irradiation. Cells were stained with 7-AAD (BioLegend) and subjected to flow cytometric analysis. Data were acquired on a CytoFLEX LX with CytExpert 2.4 and analyzed using FlowJo v10.\u003c/span\u003e\u003c/p\u003e\n\n\u003cp id=\"_Toc128343233\"\u003e\u003cstrong\u003eIdentification of oxidized phospholipids by LC-MS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLipids were extracted following the Folch procedure as previously reported \u003csup\u003e51\u003c/sup\u003e, and global oxidized phospholipidomics was carried out as described earlier\u003csup\u003e52, 53\u003c/sup\u003e. Phospholipids (PLs) were analyzed by liquid chromatography-mass spectrometry (LC\u0026ndash;MS) on a Dionex UltiMate 3000 LC system coupled to a Q-Exactive mass spectrometer (Thermo Fisher Scientific). Chromatographic separation was achieved on a normal-phase Luna Silica (2) column (3\u0026thinsp;\u0026mu;m, 150\u0026thinsp;\u0026times;\u0026thinsp;2.0\u0026thinsp;mm; Phenomenex) at a flow rate of 0.2\u0026thinsp;ml\u0026thinsp;min-1. The mobile phases consisted of 10\u0026thinsp;mM ammonium formate in isopropanol/hexane/water (285:215:5, v/v/v; solvent A) and isopropanol/hexane/water (285:215:40, v/v/v; solvent B), with all solvents of LC-MS grade. Gradient elution was programmed as follows: 0\u0026thinsp;min, 10% B; 23\u0026thinsp;min, 32%; 32\u0026thinsp;min, 65%; 35\u0026thinsp;min, 100%; 70\u0026thinsp;min, 100%. The column was maintained at 35\u0026thinsp;\u0026deg;C, and 5\u0026thinsp;\u0026mu;l of each sample was injected. Data acquisition was performed in negative ion mode at a resolution of 70,000 for full MS scans and 17,500 for data-dependent MS/MS scans. Full-scan spectra were collected over an m/z range of 400\u0026ndash;1,800 with a maximum injection time of 200\u0026thinsp;ms using one microscan. For MS/MS acquisition, the maximum injection time was set to 500\u0026thinsp;ms, the collision energy to 24\u0026thinsp;eV, and the isolation window to 1.0\u0026thinsp;Da.\u003c/p\u003e\n\u003cp\u003eRaw LC\u0026ndash;MS data were processed using MZmine v.2.5.3 \u003csup\u003e54\u003c/sup\u003e with an in-house analysis workflow and database. Peaks with a signal-to-noise ratio \u0026gt; 3 were extracted and searched against an oxidized PL database. Lipid species were assigned by matching m/z values within 5 ppm, followed by additional filtering based on retention time and confirmation through MS/MS fragmentation patterns, with fragment annotation referenced to lipid maps (https://www.lipidmaps.org). Deuterated PLs (Avanti Polar Lipids) were included as internal standards. Quantification was performed from full-scan LC\u0026ndash;MS spectra by ratiometric comparison with the corresponding preselected internal standard, using a standard curve for each PL class.\u003c/p\u003e\n\n\u003cp id=\"_Toc128343234\"\u003e\u003cstrong\u003eAnalysis of AA-d8 containing phospholipids by LC-MS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTotal lipids were analyzed by LC-MS following separation on a reverse-phase Acquity HSS T3 column (1.8\u0026thinsp;\u0026mu;m, 100\u0026thinsp;\u0026times;\u0026thinsp;2.1\u0026thinsp;mm; Waters) at a flow rate of 0.3\u0026thinsp;ml\u0026thinsp;min-1. The mobile phases were 10\u0026thinsp;mM ammonium formate in water/acetonitrile (50:50, v/v; solvent A) and isopropanol/acetonitrile (90:10, v/v; solvent B). Elution was achieved with the following gradient: 0\u0026thinsp;min, 30% B; 5\u0026thinsp;min, 43%; 5.1\u0026thinsp;min, 50%; 14\u0026thinsp;min, 70%; 14.1\u0026thinsp;min, 70%; 23\u0026thinsp;min, 99%; 26\u0026thinsp;min, 99%. The column was held at 40\u0026thinsp;\u0026deg;C throughout the run. MS and data-dependent MS/MS analyses were performed on a Q-Exactive mass spectrometer (Thermo Fisher Scientific) in both positive and negative ion modes using profile acquisition. Full MS scans were acquired at a resolution of 70,000, whereas MS/MS scans were acquired at 17,500. The scan range for full MS was m/z 114-1,700, with one microscan, a maximum injection time of 100\u0026thinsp;ms, and an AGC target of 1\u0026thinsp;\u0026times;\u0026thinsp;10\u003csup\u003e5\u003c/sup\u003e. For MS/MS, the maximum injection time was 50\u0026thinsp;ms, the isolation window was 1.0\u0026thinsp;Da, and the normalized collision energy was set to 20%, 30%, and 40%. Raw data processing and phospholipid identification were carried out using MS-DIAL \u003csup\u003e52\u003c/sup\u003e. \u003c/p\u003e\n\n\n\u003cp id=\"_Toc128343238\"\u003e\u003cstrong\u003eMolecular Dynamics Simulations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe three-dimensional structure of ACSL3 predicted by AlphaFoldwas used to construct the wild-type model. Protonation states of charged residues were assigned using the H++ server\u003cstrong\u003e\u003csup\u003e55\u003c/sup\u003e\u003c/strong\u003e together with manual inspection of local hydrogen-bonding networks. Histidine residues were modeled in either the \u0026epsilon;-protonated or \u0026delta;-protonated state according to their local environment, whereas Asp/Glu and Lys/Arg residues were maintained in their default charged states. The succinylated ACSL3 model was generated from the wild-type structure by attaching a succinic acid group to the side-chain amino group of K422 using Molecular Operating Environment (MOE2020). The protonation states in the succinylated model were kept identical to those in the wild-type model.\u003c/p\u003e\n\n\u003cp\u003eBefore simulation, each model was hydrogenated using the Leap module in Amber22, neutralized by addition of Na\u003csup\u003e+\u003c/sup\u003e and Cl\u003csup\u003e- \u003c/sup\u003eions with AmberTools, and solvated in a rectangular TIP3P water box with a 10 \u0026Aring; buffer between the protein and the box boundary. Molecular dynamics simulations were performed in Amber22 with the GPU-accelerated pmemd.cuda module. Each system was subjected to stepwise energy minimization, including solvent relaxation with positional restraints on the protein, restrained minimization of the protein backbone, and final unrestrained minimization of the full system. The systems were then gradually heated from 0 to 300 K under the NVT ensemble for 100 ps, followed by 200 ps equilibration under the NPT ensemble to stabilize density. Production simulations were subsequently carried out for 500 ns at 300 K under periodic boundary conditions. The FF19SB force field was applied for ACSL3 \u003cstrong\u003e\u003csup\u003e56, 57\u003c/sup\u003e\u003c/strong\u003e, and the succinylated K422 residue was defined as a customized residue (SL422). A 12 \u0026Aring; cutoff was used for van der Waals and electrostatic interactions, temperature was controlled with Langevin dynamics, and hydrogen-containing bonds were constrained using the SHAKE algorithm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConstruction of tumor xenotransplantation model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSix-week-old female specific pathogen-free (SPF) BALB/c nude mice were purchased from Charles River Laboratories. Mice were housed at five animals per cage under a 12-h light/12-h dark cycle at 20-26\u0026thinsp;\u0026deg;C with 40-70% humidity and had ad libitum access to standard chow and water. For xenograft studies, mice were subcutaneously inoculated with 1\u0026times;10\u003csup\u003e6\u003c/sup\u003e HK1 cells or stably infected HK1 cells. When tumors reached approximately 5\u0026thinsp;mm in diameter, mice were subjected to local irradiation. Tumor volume was calculated using the formula length \u0026times; width\u003csup\u003e2\u003c/sup\u003e \u0026times; 0.5. At the indicated time points, tumor tissues were collected, paraffin embedded, and processed for immunohistochemical analysis. All animal procedures were approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University and were performed in accordance with institutional guidelines for animal welfare. Every effort was made to minimize animal suffering. The maximal tumor diameter permitted by the ethics committee was 20\u0026thinsp;mm, and this limit was not exceeded in any animal. All animal experiments were conducted in full compliance with the guidelines and under the approval of the Experimental Animal Ethics Committee of Sun Yat-sen University (SYSU-IACUC-2022-002494).\u003c/p\u003e\n\n\u003cp id=\"_Toc128343242\"\u003e\u003cstrong\u003eStatistics and reproducibility\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData are shown as results from at least three independent experiments. Statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software) or IBM SPSS Statistics version 25. Differences between two groups were assessed using a two-tailed unpaired Student\u0026rsquo;s t-test, whereas comparisons among multiple groups were performed using one-way or two-way ANOVA followed by Tukey\u0026rsquo;s multiple-comparisons test. Survival outcomes were analyzed using Kaplan\u0026ndash;Meier methods and compared with the log-rank test. Clinical characteristics were compared using the chi-square (\u0026chi;\u0026sup2;) test. Phospholipid species were quantified from full-scan LC-MS spectra by ratiometric normalization to predefined internal standards and calculated using class-specific standard curves. A two-sided \u003cem\u003eP\u003c/em\u003e value \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Beijing Viktor Technology Co., Ltd. for the support with AFADESI-MSI platform, Hangzhou Shinning Technology Co., Ltd. for the support with biomechanical testing, and C. Tong (Zhejiang University) for technical assistance. This study was supported by grants from the National Natural Science Foundation of China (82504101), Changping Laboratory Project (2025C-12-04), Guangdong Basic and Applied Basic Research Foundation (2026B1515020031), Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM611), National Natural Science Foundation of China (82430085), Science and Technology Program of Guangzhou (2025B03J0149), Cancer Innovative Research Program of Sun Yat-sen University Cancer Center (CIRP-SYSUCC-0005), Overseas Expertise Introduction Project for Discipline Innovation (111 Project), Natural Science Foundation of China (82522094, 82321004), Guangdong Basic and Applied Basic Research Foundation (2023B1515020020), the Chih Kuang Scholarship for Outstanding Young PhysicianScientists of Sun Yat-sen University Cancer Center (PT22120901), Young Talents Program of Sun Yat-sen University Cancer Center (YTP-SYSUCC-0072). W.Y. Sun gratefully acknowledges the support of the K. C. Wong Education Foundation in Jinan University.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003cstrong\u003euthor contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eX.-Y.L., Z.L. and W.-Y.S. conceived the experiments. Z.L., N.-Z.L., and Y.Q. carried out and analyzed the data for most of the \u003cem\u003ein vitro\u0026nbsp;\u003c/em\u003eexperiments. Z.L., and N.-Z.L. designed and performed the animal experiments. W.-Y.S., R.W., and J.-C.F. designed and performed LC-MS/MS analysis. Z.L., N.-Z.L., T.-X.H. and Y.-H.L. collected clinical samples and performed the IHC experiments. H.-M.W., C.-Q.Z., Y.-L.C., Y.-L.W., and Z.-J.D. helped with the data analyses. R.-R.H. provided\u0026nbsp;technical support. Z.L., X.-Y.L. and W.-Y.S. wrote the manuscript. J.M., X.-Y.L., and W.-Y.S. supervised the study. All authors reviewed and discussed the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eNia, H.T., Munn, L.L. \u0026amp; Jain, R.K. Physical traits of cancer. \u003cem\u003eScience (New York, N.Y.)\u003c/em\u003e \u003cstrong\u003e370\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eCox, T.R. The matrix in cancer. \u003cem\u003eNature reviews. Cancer\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 217-238 (2021).\u003c/li\u003e\n\u003cli\u003eLei, G.\u003cem\u003e et al.\u003c/em\u003e The role of ferroptosis in ionizing radiation-induced cell death and tumor suppression. \u003cem\u003eCell Res.\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 146-162 (2020).\u003c/li\u003e\n\u003cli\u003eStockwell, B.R. Ferroptosis turns 10: Emerging mechanisms, physiological functions, and therapeutic applications. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e, 2401-2421 (2022).\u003c/li\u003e\n\u003cli\u003eLang, X.\u003cem\u003e et al.\u003c/em\u003e Radiotherapy and Immunotherapy Promote Tumoral Lipid Oxidation and Ferroptosis via Synergistic Repression of SLC7A11. \u003cem\u003eCancer discovery\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 1673-1685 (2019).\u003c/li\u003e\n\u003cli\u003eChen, X., Kang, R., Kroemer, G. \u0026amp; Tang, D. Broadening horizons: the role of ferroptosis in cancer. \u003cem\u003eNature reviews. Clinical oncology\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 280-296 (2021).\u003c/li\u003e\n\u003cli\u003eGrevengoed, T.J., Klett, E.L. \u0026amp; Coleman, R.A. Acyl-CoA metabolism and partitioning. \u003cem\u003eAnnual review of nutrition\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, 1-30 (2014).\u003c/li\u003e\n\u003cli\u003eMagtanong, L.\u003cem\u003e et al.\u003c/em\u003e Exogenous Monounsaturated Fatty Acids Promote a Ferroptosis-Resistant Cell State. \u003cem\u003eCell chemical biology\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 420-432.e429 (2019).\u003c/li\u003e\n\u003cli\u003ePhillips, M.J. \u0026amp; Voeltz, G.K. Structure and function of ER membrane contact sites with other organelles. \u003cem\u003eNat. Rev. Mol. Cell Biol.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 69-82 (2016).\u003c/li\u003e\n\u003cli\u003eZhang, Z.\u003cem\u003e et al.\u003c/em\u003e CGI1746 targets \u0026sigma;(1)R to modulate ferroptosis through mitochondria-associated membranes. \u003cem\u003eNature chemical biology\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 699-709 (2024).\u003c/li\u003e\n\u003cli\u003eWu, H., Carvalho, P. \u0026amp; Voeltz, G.K. Here, there, and everywhere: The importance of ER membrane contact sites. \u003cem\u003eScience (New York, N.Y.)\u003c/em\u003e \u003cstrong\u003e361\u003c/strong\u003e (2018).\u003c/li\u003e\n\u003cli\u003eZheng, P.\u003cem\u003e et al.\u003c/em\u003e ER proteins decipher the tubulin code to regulate organelle distribution. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e601\u003c/strong\u003e, 132-138 (2022).\u003c/li\u003e\n\u003cli\u003eYou, R.\u003cem\u003e et al.\u003c/em\u003e Single-cell and spatial transcriptomics reveal mechanisms of radioresistance and immune escape in recurrent nasopharyngeal carcinoma. \u003cem\u003eNature genetics\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 1950-1965 (2025).\u003c/li\u003e\n\u003cli\u003eKol\u0026aacute;cn\u0026aacute;, L.\u003cem\u003e et al.\u003c/em\u003e Biochemical and biophysical aspects of collagen nanostructure in the extracellular matrix. \u003cem\u003ePhysiological research\u003c/em\u003e \u003cstrong\u003e56 Suppl 1\u003c/strong\u003e, S51-s60 (2007).\u003c/li\u003e\n\u003cli\u003eLevental, K.R.\u003cem\u003e et al.\u003c/em\u003e Matrix crosslinking forces tumor progression by enhancing integrin signaling. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e139\u003c/strong\u003e, 891-906 (2009).\u003c/li\u003e\n\u003cli\u003eGeorges, P.C.\u003cem\u003e et al.\u003c/em\u003e Increased stiffness of the rat liver precedes matrix deposition: implications for fibrosis. \u003cem\u003eAmerican journal of physiology. Gastrointestinal and liver physiology\u003c/em\u003e \u003cstrong\u003e293\u003c/strong\u003e, G1147-1154 (2007).\u003c/li\u003e\n\u003cli\u003eNakamura, T. \u0026amp; Conrad, M. Exploiting ferroptosis vulnerabilities in cancer. \u003cem\u003eNature cell biology\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 1407-1419 (2024).\u003c/li\u003e\n\u003cli\u003eKimura, H.\u003cem\u003e et al.\u003c/em\u003e Syntaxin 17 promotes lipid droplet formation by regulating the distribution of acyl-CoA synthetase 3. \u003cem\u003eJournal of lipid research\u003c/em\u003e \u003cstrong\u003e59\u003c/strong\u003e, 805-819 (2018).\u003c/li\u003e\n\u003cli\u003evon Krusenstiern, A.N.\u003cem\u003e et al.\u003c/em\u003e Identification of essential sites of lipid peroxidation in ferroptosis. \u003cem\u003eNature chemical biology\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 719-730 (2023).\u003c/li\u003e\n\u003cli\u003eYang, Z., Zhao, X., Xu, J., Shang, W. \u0026amp; Tong, C. A novel fluorescent reporter detects plastic remodeling of mitochondria-ER contact sites. \u003cem\u003eJournal of cell science\u003c/em\u003e \u003cstrong\u003e131\u003c/strong\u003e (2018).\u003c/li\u003e\n\u003cli\u003eCieri, D.\u003cem\u003e et al.\u003c/em\u003e SPLICS: a split green fluorescent protein-based contact site sensor for narrow and wide heterotypic organelle juxtaposition. \u003cem\u003eCell death and differentiation\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 1131-1145 (2018).\u003c/li\u003e\n\u003cli\u003eKomatsu, T.\u003cem\u003e et al.\u003c/em\u003e Organelle-specific, rapid induction of molecular activities and membrane tethering. \u003cem\u003eNature methods\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 206-208 (2010).\u003c/li\u003e\n\u003cli\u003eJang, W.\u003cem\u003e et al.\u003c/em\u003e Endosomal lipid signaling reshapes the endoplasmic reticulum to control mitochondrial function. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e378\u003c/strong\u003e, eabq5209 (2022).\u003c/li\u003e\n\u003cli\u003eFriedman, J.R., Webster, B.M., Mastronarde, D.N., Verhey, K.J. \u0026amp; Voeltz, G.K. ER sliding dynamics and ER-mitochondrial contacts occur on acetylated microtubules. \u003cem\u003eThe Journal of cell biology\u003c/em\u003e \u003cstrong\u003e190\u003c/strong\u003e, 363-375 (2010).\u003c/li\u003e\n\u003cli\u003eSeetharaman, S.\u003cem\u003e et al.\u003c/em\u003e Microtubules tune mechanosensitive cell responses. \u003cem\u003eNature materials\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 366-377 (2022).\u003c/li\u003e\n\u003cli\u003eLi, H., Zhang, H., Zhu, J., Tang, Y. \u0026amp; An, J. Protein succinylation in tumors-from biology to clinical therapy. \u003cem\u003eInternational journal of biological macromolecules\u003c/em\u003e \u003cstrong\u003e337\u003c/strong\u003e, 149489 (2026).\u003c/li\u003e\n\u003cli\u003eKurmi, K.\u003cem\u003e et al.\u003c/em\u003e Carnitine Palmitoyltransferase 1A Has a Lysine Succinyltransferase Activity. \u003cem\u003eCell Rep.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 1365-1373 (2018).\u003c/li\u003e\n\u003cli\u003eWang, Y.\u003cem\u003e et al.\u003c/em\u003e KAT2A coupled with the \u0026alpha;-KGDH complex acts as a histone H3 succinyltransferase. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e552\u003c/strong\u003e, 273-277 (2017).\u003c/li\u003e\n\u003cli\u003eMa, W.\u003cem\u003e et al.\u003c/em\u003e OXCT1 functions as a succinyltransferase, contributing to hepatocellular carcinoma via succinylating LACTB. \u003cem\u003eMolecular cell\u003c/em\u003e \u003cstrong\u003e84\u003c/strong\u003e, 538-551.e537 (2024).\u003c/li\u003e\n\u003cli\u003eTong, Y.\u003cem\u003e et al.\u003c/em\u003e SUCLA2-coupled regulation of GLS succinylation and activity counteracts oxidative stress in tumor cells. \u003cem\u003eMolecular cell\u003c/em\u003e \u003cstrong\u003e81\u003c/strong\u003e, 2303-2316.e2308 (2021).\u003c/li\u003e\n\u003cli\u003eShen, R.\u003cem\u003e et al.\u003c/em\u003e Lysine succinylation, the metabolic bridge between cancer and immunity. \u003cem\u003eGenes \u0026amp; diseases\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 2470-2478 (2023).\u003c/li\u003e\n\u003cli\u003eDu, J.\u003cem\u003e et al.\u003c/em\u003e Sirt5 is a NAD-dependent protein lysine demalonylase and desuccinylase. \u003cem\u003eScience (New York, N.Y.)\u003c/em\u003e \u003cstrong\u003e334\u003c/strong\u003e, 806-809 (2011).\u003c/li\u003e\n\u003cli\u003eBansaccal, N.\u003cem\u003e et al.\u003c/em\u003e The extracellular matrix dictates regional competence for tumour initiation. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e623\u003c/strong\u003e, 828-835 (2023).\u003c/li\u003e\n\u003cli\u003eWu, B.\u003cem\u003e et al.\u003c/em\u003e Stiff matrix induces exosome secretion to promote tumour growth. \u003cem\u003eNature cell biology\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 415-424 (2023).\u003c/li\u003e\n\u003cli\u003eLeSavage, B.L.\u003cem\u003e et al.\u003c/em\u003e Engineered matrices reveal stiffness-mediated chemoresistance in patient-derived pancreatic cancer organoids. \u003cem\u003eNature materials\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1138-1149 (2024).\u003c/li\u003e\n\u003cli\u003eZhang, J.\u003cem\u003e et al.\u003c/em\u003e Osr2 functions as a biomechanical checkpoint to aggravate CD8(+) T cell exhaustion in tumor. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e187\u003c/strong\u003e, 3409-3426.e3424 (2024).\u003c/li\u003e\n\u003cli\u003eZhang, H.\u003cem\u003e et al.\u003c/em\u003e The extracellular matrix integrates mitochondrial homeostasis. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e187\u003c/strong\u003e, 4289-4304.e4226 (2024).\u003c/li\u003e\n\u003cli\u003eXu, L.\u003cem\u003e et al.\u003c/em\u003e Live-cell imaging with fluorogenic radical-trapping antioxidant probes reveals the onset and progression of ferroptosis. \u003cem\u003eNature chemistry\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 1963-1973 (2025).\u003c/li\u003e\n\u003cli\u003eSassano, M.L.\u003cem\u003e et al.\u003c/em\u003e Endoplasmic reticulum-mitochondria contacts are prime hotspots of phospholipid peroxidation driving ferroptosis. \u003cem\u003eNature cell biology\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 902-917 (2025).\u003c/li\u003e\n\u003cli\u003eYang, J.F.\u003cem\u003e et al.\u003c/em\u003e Mitochondria-ER contact mediated by MFN2-SERCA2 interaction supports CD8(+) T cell metabolic fitness and function in tumors. \u003cem\u003eScience immunology\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, eabq2424 (2023).\u003c/li\u003e\n\u003cli\u003eNava Lauson, C.B.\u003cem\u003e et al.\u003c/em\u003e Linoleic acid potentiates CD8(+) T cell metabolic fitness and antitumor immunity. \u003cem\u003eCell metabolism\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 633-650.e639 (2023).\u003c/li\u003e\n\u003cli\u003eWang, H.\u003cem\u003e et al.\u003c/em\u003e Succinate drives gut inflammation by promoting FOXP3 degradation through a molecular switch. \u003cem\u003eNature immunology\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 866-880 (2025).\u003c/li\u003e\n\u003cli\u003eLiang, L.\u003cem\u003e et al.\u003c/em\u003e Alterations in PD-L1 succinylation shape anti-tumor immune responses in melanoma. \u003cem\u003eNature genetics\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 680-693 (2025).\u003c/li\u003e\n\u003cli\u003eMa, K.\u003cem\u003e et al.\u003c/em\u003e Succinate preserves CD8(+) T cell fitness to augment antitumor immunity. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e58\u003c/strong\u003e, 2505-2523.e2508 (2025).\u003c/li\u003e\n\u003cli\u003eZhao, Q.\u003cem\u003e et al.\u003c/em\u003e SIRT5 safeguards against primate skeletal muscle ageing via desuccinylation of TBK1. \u003cem\u003eNature metabolism\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 556-573 (2025).\u003c/li\u003e\n\u003cli\u003eWu, J.H.\u003cem\u003e et al.\u003c/em\u003e Metabolism-dependent succinylation governs resource allocation for antibiotic resistance. \u003cem\u003eScience advances\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, eadu2856 (2025).\u003c/li\u003e\n\u003cli\u003eHam, S.J.\u003cem\u003e et al.\u003c/em\u003e Mitochondrial fumarate inhibits Parkin-mediated mitophagy. \u003cem\u003eMolecular cell\u003c/em\u003e \u003cstrong\u003e85\u003c/strong\u003e, 2287-2302.e2289 (2025).\u003c/li\u003e\n\u003cli\u003eZeng, Y.\u003cem\u003e et al.\u003c/em\u003e Substrate stiffness regulates B-cell activation, proliferation, class switch, and T-cell-independent antibody responses in vivo. \u003cem\u003eEuropean journal of immunology\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1621-1634 (2015).\u003c/li\u003e\n\u003cli\u003eXu, Z.\u003cem\u003e et al.\u003c/em\u003e TOLLIP targets GSDME-NT-carrying endocytic vesicles for autophagy to regulate pyroptosis and chemotherapy efficacy. \u003cem\u003eNature cell biology\u003c/em\u003e (2026).\u003c/li\u003e\n\u003cli\u003eWu, J.\u003cem\u003e et al.\u003c/em\u003e Methionine metabolite spermidine inhibits tumor pyroptosis by enhancing MYO6-mediated endocytosis. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 2184 (2025).\u003c/li\u003e\n\u003cli\u003eFolch, J., Lees, M. \u0026amp; Sloane Stanley, G.H. A simple method for the isolation and purification of total lipides from animal tissues. \u003cem\u003eThe Journal of biological chemistry\u003c/em\u003e \u003cstrong\u003e226\u003c/strong\u003e, 497-509 (1957).\u003c/li\u003e\n\u003cli\u003eLi, Z.\u003cem\u003e et al.\u003c/em\u003e Tumor-repopulating cells evade ferroptosis via PCK2-dependent phospholipid remodeling. \u003cem\u003eNat. Chem. Biol.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 1341-1352 (2024).\u003c/li\u003e\n\u003cli\u003eSun, W.Y.\u003cem\u003e et al.\u003c/em\u003e Phospholipase iPLA(2)\u0026beta; averts ferroptosis by eliminating a redox lipid death signal. \u003cem\u003eNature chemical biology\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 465-476 (2021).\u003c/li\u003e\n\u003cli\u003ePluskal, T., Castillo, S., Villar-Briones, A. \u0026amp; Oresic, M. MZmine 2: modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. \u003cem\u003eBMC bioinformatics\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 395 (2010).\u003c/li\u003e\n\u003cli\u003eAnandakrishnan, R., Aguilar, B. \u0026amp; Onufriev, A.V. H++ 3.0: automating pK prediction and the preparation of biomolecular structures for atomistic molecular modeling and simulations. \u003cem\u003eNucleic acids research\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, W537-541 (2012).\u003c/li\u003e\n\u003cli\u003eTian, C.\u003cem\u003e et al.\u003c/em\u003e ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution. \u003cem\u003eJournal of chemical theory and computation\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 528-552 (2020).\u003c/li\u003e\n\u003cli\u003eSvozil, D.\u003cem\u003e et al.\u003c/em\u003e Geometrical and electronic structure variability of the sugar-phosphate backbone in nucleic acids. \u003cem\u003eThe journal of physical chemistry. B\u003c/em\u003e \u003cstrong\u003e112\u003c/strong\u003e, 8188-8197 (2008).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9481973/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9481973/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The mechanical properties of the tumor microenvironment serve as crucial physical cues that shape cell fate decisions. However, whether microenvironmental mechanical forces modulate ferroptosis to drive radioresistance remains unclear. Here, using PDMS-based hydrogels with tunable stiffness to establish tumor cell culture systems, we found that tumor cells cultured on stiff substrates were more sensitive to both ferroptosis and radiotherapy. Mechanistically, tumor cells grown on stiff matrices showed increased microtubule acetylation and ER sheet-to-tubule remodeling, which increased the formation of mitochondria-associated membranes (MAMs) and led to mitochondrial succinate accumulation. Succinate in turn promoted CPT1A-mediated succinylation of ACSL3, facilitating its degradation and thereby enhancing tumor cell sensitivity to ferroptosis. Collectively, these findings identify MAMs as intracellular mechanosensitive structures that regulate mitochondrial metabolism and ferroptosis in tumor cells, providing new insights into the mechanical control of ferroptosis and its implications for tumor radioresistance.","manuscriptTitle":"Mechanical stress enhances tumor cell ferroptosis by remodeling succinate metabolism","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 11:37:08","doi":"10.21203/rs.3.rs-9481973/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-chemical-biology","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nchembio","sideBox":"Learn more about [Nature Chemical Biology](http://www.nature.com/nchembio/)","snPcode":"","submissionUrl":"","title":"Nature Chemical Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7f77f04b-853b-4d7f-9552-a181adb724f6","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-11T08:13:28+00:00","index":1,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-10T06:31:48+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-05-01T04:16:14+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-05-01T01:39:40+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"2","date":"2026-04-30T23:09:41+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":67349519,"name":"Biological sciences/Chemical biology/Lipids/Phospholipids"},{"id":67349520,"name":"Biological sciences/Biochemistry/Metabolomics"},{"id":67349521,"name":"Biological sciences/Cell biology/Cell death"},{"id":67349522,"name":"Health sciences/Diseases/Cancer/Cancer therapy"}],"tags":[],"updatedAt":"2026-05-04T11:37:08+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 11:37:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9481973","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9481973","identity":"rs-9481973","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}