Pyridoxal reprograms peroxisomal hydrogen peroxide metabolism to induce lethal oxidative stress in ovarian cancer

Introduction

Ovarian cancer, the most lethal gynecologic malignancy(1) is primarily treated with cytoreductive surgery and platinum-paclitaxel chemotherapy(2). Despite surgical advances,, chemotherapy response remains crucial for survival(3). However, about 70% of advanced-stage patients relapse within three years, with diminishing response durations(3,4). As the first-line treatment reagents, platinum and paclitaxel act through distinct mechanisms (DNA damage and microtubule disruption), both preferentially induce apoptosis in hyperproliferative tumor cells (5,6). This principle also explains their toxicity to rapidly dividing normal tissues, such as bone marrow(7), frequently necessitating dose reduction or treatment discontinuation(8). Consequently, developing novel therapeutic agents with mechanisms of action divergent from current regimens- potentially minimizing toxicity and overcoming resistance - is a paramount goal in ovarian cancer research(9).

Vitamin B6, an essential water-soluble micronutrient abundant in various foods, encompasses a group of related compounds critical for cellular metabolism. This family includes pyridoxine (PN), pyridoxal (PL), and pyridoxamine (PM), along with their phosphate esters(10). Absorbed primarily as PL, dietary vitamin B6 is phosphorylated in the intestine to pyridoxal 5’-phosphate (PLP), its transport form(11). Upon cellular uptake, PLP serves as the central biologically active coenzyme, interconverting with PL, PN, and PM through the actions of enzymes like pyridoxine-5’-phosphate oxidase (PNPO), pyridoxal kinase (PDXK), and tissue-nonspecific alkaline phosphatase (TNSALP) (12).

Clinical standards for assessing vitamin B6 status rely predominantly on PLP levels (12,13), treating the other forms (PL, PN, PM) as its precursors. To date, research on the anti-cancer effects of vitamin B6 has largely centered on epidemiological studies linking PLP intake to cancer incidence, yielding conflicting results. Although PLP may protect against certain cancers (e.g., lung, colon) (14,15), it can also promote tumor progression (16,17)-a property that explains why inhibiting its synthesis can curb tumorigenesis.. Crucially, the cytotoxic effects of non-PLP vitamin B6 vitamers, particularly PL, have been attributed almost exclusively to their intracellular conversion into PLP(18). Consequently, PL’s direct effects in cancer development and treatment are still poorly understood.

Peroxisomes play indispensable roles in cellular metabolic, redox homeostasis, and tumor progression(19), responsible for scavenging reactive oxygen species (20,21), fatty acid oxidation, and ether phospholipid synthesis (22). Given their critical functions, peroxisome number and function are tightly regulated through dynamic processes including de novo biogenesis, growth and fission, and turnover via pexophagy or proteosomal degradation(23,24). As peroxisomes lack their own DNA, they rely entirely on the import of nuclear-encoded proteins, a process coordinated by receptors like PEX5(25,26). Impaired peroxisomal function is pathogenic(27), linked to devastating disorders like Zellweger spectrum disorders, neurodegenerative diseases, however, whether peroxisomal alterations affect ovarian cancer survival remains unexplored(28–30).

In this study, we demonstrate that PL exhibits potent cytotoxicity, independent of its conversion to PLP, using diverse in vitro and in vivo models including patient-derived organoids (PDOs), primary cells, and established ovarian cancer cell lines and shows the strongest anticancer activity among vitamin B6 vitamers, particularly upon blockade of PLP conversion. Mechanistically, PL induces peroxisome proliferation via PEX11A-dependent fission. This concurrently amplifies peroxisomal fatty acid oxidation (FAO)-mediated H₂O₂ production and impairs catalase import and function. Consequently, peroxisomal activity shifts from reactive oxygen species (ROS) scavenging to robust H2O2 generation, inducing overwhelming oxidative stress, mitochondrial damage, and ultimately, ovarian cancer cell death.

Results

PL Exhibits Cytotoxic Activity Against Ovarian Cancer Independent of PLP Conversion

We established PDOs from ovarian cancers that were either relatively sensitive (high grade serous ovarian cancer, HGSOC) or resistant (clear cell ovarian cancer, CCOC) to first-line chemotherapy, and PL demonstrated cytotoxicity against both (Figure 1A). Crucially, among the fundamental vitamin B6 vitamers, PL demonstrated the strongest cytotoxic effect on ovarian cancer cells. In stark contrast, the major dietary precursor pyridoxine (PN) and pyridoxamine (PM) showed minimal efficacy (Figure 1B and Figure S1A). Critically, to rule out nonspecific effects from the salt formulation, we compared pyridoxal with its hydrochloride analog. Both compounds exhibited nearly identical IC50 values across the four cell lines, confirming that the observed cytotoxicity is intrinsic to the PL pharmacophore rather than an artifact of the hydrochloride counterion (Figure S1B). Given its superior aqueous solubility and chemical stability, pyridoxal hydrochloride was selected for all subsequent mechanistic and in vivo studies to ensure reliable and reproducible dosing (PL hereafter refers to pyridoxal hydrochloride).

PL’s superior cytotoxic potency was further validated across 17 primary ovarian cancer cell cultures consistently yielding the lowest IC50 values compared to PLP and PN (Figure 1C). Moreover, utilizing the highly metastatic SKOV3 cell line to establish orthotopic xenograft models, we confirmed PL’s significant inhibitory effect on ovarian cancer growth in vivo (Figure 1D, S1C).

In addition, we further evaluated the therapeutic potential of PL. In tumor-bearing mice, PL at efficacious anti-tumor doses did not induce significant abnormalities in hematological parameters, liver/kidney function, histopathology, or body weight throughout the treatment period (Figure S2A-D). Concurrently, PL demonstrated a high tolerable concentration against a panel of normal cell lines, including human renal proximal tubular epithelial cells (HK-2), human embryonic kidney cells (HEK-293T), and mouse normal hepatocytes (AML-12) (Figure. S2E). Furthermore, real-time cell analysis (RTCA) confirmed that 0.5 mM PL did not inhibit the growth of primary human dermal fibroblasts, underscoring its selective cytotoxicity towards cancer cells (Figure S2F). Importantly, PL retained its cytotoxic potency (as evidenced by unchanged IC50 values) in patient-derived organoids (PDOs) that were resistant to chemotherapeutics (PTX, cisplatin) or the targeted agent olaparib, indicating that its mechanism of action is independent of these common resistance pathways (Figure S3A, B). Furthermore, in multiple ovarian cancer cell lines (SKOV3, OVCAR3, 3AO, ES-2), a sub-toxic concentration of PL significantly sensitized the cells to PTX, markedly reducing the IC50 of PTX in all tested models (Figure S3C). Together, these data indicate that PL possesses a favorable safety profile, circumvents common drug resistance mechanisms, and can potentiate the efficacy of standard chemotherapy, highlighting its potential as a combination therapy agent.

Contrary to the prevailing paradigm that PL acts via conversion to PLP, our data demonstrate a conversion-independent mechanism. PL was superior to PLP in tumor-killing efficacy across histological models. Furthermore, the negligible cytotoxicity of PN and PM, clinical supplements that raise PLP levels, suggests that PL’s activity is independent of PLP conversion. Using a non-phosphorylatable PL tracer biotin-PL (whose biotin moiety was designed to occupy the canonical phosphate binding site at the pyridoxine ring 5’-position), we found that PL distributed widely within the cytoplasm without detectable conversion to PLP (Figure 1E). Notably, biotinylation preserved PL’s cytotoxic activity (Figure 1F), indicating that PL exerts antitumor effects directly through its native structure. Consistently, neither PL nor biotin-PL administration increased cellular PLP content. Furthermore, the accumulation of PL confirmed that the biotin moiety effectively blocked PL phosphorylation (Figure 1G). Collectively, these results establish PL as a potent cytotoxic agent against ovarian cancer in vitro and in vivo, acting independently of its conversion to PLP.

PL Disrupts Cellular Metabolic Processes

To elucidate the mechanisms underlying PL’s cytotoxicity, we performed multi-omics analyses on PL-treated SKOV3 cells. Pathway enrichment of RNA-Seq data revealed significant enrichment for pathways related to cellular metabolism and metabolic diseases, including amino acid biosynthesis, steroid synthesis, carbon metabolism, glucose metabolism, and oxidative phosphorylation. Enrichment also extended to pathways implicated in neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, and prion disease (Figure S4A). Untargeted metabolomics confirmed PL-induced perturbations in cellular metabolism (Figure 2A, B). While metabolite pathway enrichment converged with RNA-Seq data on alterations in the TCA cycle/core carbon metabolism (Figure 2C), the observed metabolic changes showed no significant correlation with transcriptomic changes, suggesting complex regulation.

Proteomic and metabolomic analyses were integrated to delineate PL’s impact on cellular metabolism. Regularized canonical correlation analysis (rCCA) was employed to quantify associations between proteins and metabolites, revealing a concerted rewiring of metabolism upon PL treatment (Figure 2D). Sparse partial least squares discriminant analysis (sPLS-DA) with plotVar visualization illustrated the multivariate relationships, where acute angles between protein (dots) and metabolite (squares) variables indicated positive correlations, and distance from the origin reflected association strength (Figure 2E). Furthermore, a circos plot highlighted robust protein-metabolite correlations (|r| ≥ 0.99) (Figure 2F), and a subsequent Pathway enrichment analysis of the protein-metabolite pairs with a correlation coefficient ≥ 0.99 revealed a significant convergence on several key pathways, including Oxidative phosphorylation, Non-alcoholic fatty liver disease, Thermogenesis (Figure S4B). Together, these analyses linked PL-induced metabolite alterations to specific changes in protein expression, particularly implicating lipid metabolism and the TCA cycle. Corroborating the metabolomics findings, proteomics indicated PL-induced mitochondrial damage in ovarian cancer cells (Figure 2G). Disruption of the TCA cycle was substantiated by the accumulation of key intermediates (α-ketoglutarate, citrate, cis-aconitate) (Figure 2H, I). These integrated multi-omics analyses demonstrate that PL significantly disrupts cellular metabolism, leading to mitochondrial dysfunction.

PL Induces Mitochondrial Damage via ROS Overload

Mitochondrial dysfunction is a recognized driver in neurodegeneration, muscular dystrophies, and cancer(31,32). Consistent with our pathway enrichment results pointing to mitochondrial dysfunction, PL treatment triggered pronounced mitochondrial swelling and a marked loss in mitochondrial membrane potential (ΔΨm), as detected by optical diffraction tomography (ODT) (Figure S4C) and Mito-Tracker Red staining and (Figure S4D, E). Pathways enriched in our DEG analysis also highlighted a significant role for reactive oxygen species (ROS)-natural byproducts of the mitochondrial electron transport chain-consistent with their contribution to this dysfunction (Figure S4A). PL treatment induced an oxidized cellular state in SKOV3 cells, evidenced by increased NAD+ levels (Figure 3A) and a time-dependent surge in intracellular ROS that was confirmed by multiple probes (Figure 3B, C and Figure S5A-C) and recapitulated in vivo and in PDOs (Figure 3D-F). The functional consequence of this ROS elevation was DNA damage, as evidenced by Comet assays and Cleaved-PARP1 immunofluorescence (Figure S5D-F). Crucially, supplementation with the antioxidant glutathione (GSH) significantly attenuated PL’s cytotoxicity (Figure 3G), establishing redox imbalance as a key mediator of PL-induced cell death. This protective effect of GSH was accompanied by the restoration of redox homeostasis and mitochondrial morphology (Figure 3H-K), confirming that ROS are essential for PL’s cytotoxicity in OC.

Given link between ROS and ferroptosis (33), and our pathway enrichment results, we investigated ferroptosis by examining key regulatory factors (GPX4, ACSL4, and NRF2) via MRM and WB (Figure S6A, B), as well as functional hallmarks including intracellular ferrous iron content (Figure S6C, D) and lipid peroxidation (Figure S6E, F) using flow cytometry, but found that PL elicits cytotoxicity in ovarian cancer cells without inducing ferroptosis. Thus, while PL-induced ROS drive lethal oxidative stress and mitochondrial damage critical for cytotoxicity in HGSOC cells, the underlying mechanism is distinct from ferroptosis.

Extramitochondrial H2O2 is the Primary ROS Origin in PL Response

To delineate the specific ROS species elevated by PL treatment, we employed fluorophore probes with distinct specificity, revealing a selective upregulation of hydrogen peroxide (H2O2) (Figure 4A-D). Consistent with this, an Amplex Red-based assay detected a rapid surge in extracellular H2O2 release immediately following PL exposure (Figure 4E). Although H2O2 is less toxic due to efficient scavenging by peroxisomal catalase (CAT), it often originates from the dismutation of the highly reactive superoxide radicals (O2•⁻), which can be generated by mitochondrial respiration or peroxisomal fatty acid β-oxidation(19).

While mitochondrial ETC produces O2•⁻ during respiration, particularly under conditions of electron leak or damage(34), our data exclude this as the primary driver. PL-induced H2O2 originates from a non-mitochondrial source, and observed mitochondrial dysfunction is a downstream effect. This is evidenced by the inability of the mitochondrial ROS scavenger Mito-Tempo to protect cells (Figure 5A) or reduce global ROS (Figure 5B-E), alongside a PL-dependent reduction in mitochondrial O2•⁻ (Mito-O2•⁻) (Figure 5F, G). Assays on isolated mitochondria confirmed that PL suppresses, rather than activates, mitochondrial H2O2 production (Figure 4F). Taken together, these results demonstrate that mitochondrial ROS production is not the initiating event but rather a downstream consequence of extramitochondrial H2O2-driven oxidative stress.

PL-Induced H2O2 Overload by enhancing Peroxisomal Generation and Impairing Scavenging

H2O2 is a major byproduct of peroxisomal fatty acid oxidation (FAO) and is normally rapidly detoxified by catalase (CAT). PL treatment markedly upregulated key enzymes involved in peroxisomal FAO in ovarian cancer cells (Figure 6A). Although mitochondrial FAO enzymes were slightly elevated, the significant accumulation of medium- and long-chain acylcarnitines-intermediates indicative of peroxisomal FAO-suggested a specific enhancement of peroxisomal activity (Figure 6B). The functional necessity of this pathway for PL’s cytotoxicity was confirmed by the ACOX1 inhibitor (10, 12-Tricosadiynoic acid, TDA), which significantly rescued cell viability in RTCA and CCK-8 assays (Figure 6C, D). To directly attribute the ensuing H2O2 overload to peroxisomes, we employed the peroxisome-targeted H2O2 biosensor HyPer7-Peroxi. Flow cytometric analysis revealed that PL treatment specifically increased H2O2 levels within peroxisomes (Figure 6E, F).

Conversely, PL disrupted H2O2 clearance by impairing the PEX5-CAT interaction critical for catalase import and function, despite increased CAT protein expression (Figure 6G-I, Figure S7). This was corroborated by a progressive decline in total catalase enzymatic activity (Figure 6J). Collectively, PL triggers H2O2 overload by concurrently enhancing its peroxisomal generation and impairing its enzymatic degradation.

PL Drives Peroxisome Proliferation by Shifting Dynamics Towards PEX11A-Mediated Fission

Peroxisome abundance is tightly regulated, as uncontrolled proliferation compromises redox homeostasis. PL treatment disrupted peroxisomal homeostasis, leading to a significant increase in peroxisome abundance. This was visually confirmed by fluorescent imaging using a peroxisome-tracker (Figure 7A). Consistent with this morphological expansion, the expression levels of key peroxisomal structural proteins, PEX5 and PMP70, were markedly elevated, as quantified by both semi-quantitative immunofluorescence (Figure S8A, C) and Western blot analysis (Figure S8D). Further corroborating these findings in vivo, immunofluorescence analysis of orthotopic xenograft tumors revealed a concurrent upregulation of both PEX5 and CAT following PL treatment (Figure 7B). Besides, after Peroxisome-Tracker labeling, live-cell ODT/FL (optical diffraction tomography/fluorescence) dual-mode imaging confirmed this rapid proliferation, showing intensifying peroxisomal fluorescence that peaked just prior to cell death (SMovie 1-2), suggesting a potential link between uncontrolled proliferation and cytotoxic outcome.

Having ruled out accelerated degradation via pexophagy, as neither the co-localization of peroxisomes with lysosomes nor PEX5 ubiquitylation levels increased (Figure S8E, F), we focused on peroxisomal biogenesis. Continuous high-resolution live-cell imaging (HiS-SIM) revealed PL treatment drastically reduced membrane fusion events. Concurrently, fission events became the dominant dynamic process, further supported by the specific upregulation of PEX11A, a key regulator of membrane fission (Figure 7C, D, Figure S8B, C, SMovie 3-6). The functional importance of fission was demonstrated by PEX11A knockdown, which specifically attenuated PL-induced fission events (Figure S9A, B, SMovie 7-14). Collectively, these data indicate PL-induced proliferation is primarily driven by PEX11A-mediated fission, not by suppressed turnover. Silencing PEX11A abolished PL-induced peroxisome proliferation and significantly attenuated its anti-tumor efficacy (Figure 8A-C), and conversely, PEX11A overexpression potentiated the cytotoxicity induced by PL (Figure S9C). Furthermore, the generation of nascent peroxisomes with reduced CAT cargo, a hallmark of PL’s action, was markedly blunted upon PEX11A knockdown (Figure 8D, E). Consequently, the critical downstream events-the elevation of total intracellular ROS and the specific surge in H2O2 -were effectively suppressed (Figure 8F, S9D). These data establish that PEX11A-mediated fission is indispensable for PL-induced peroxisome proliferation and the ensuing oxidative cell death. Moreover, clinical database analysis revealed that high PEX11A correlated with poorer overall survival in ovarian cancer patients (Figure S9E). This finding provides direct clinical relevance to our mechanistic findings and positions the perturbation of peroxisomal H2O2 metabolism as a promising therapeutic strategy.

Discussion

Our study unveils a novel, non-canonical anticancer mechanism of pyridoxal (PL) that is independent of its conversion to PLP. We demonstrate that PL triggers peroxisome proliferation via PEX11A-dependent fission. This remodeling shifts the peroxisomal role from a protective organelle to a potent source of oxidative stress by concurrently enhancing H2O2 production via fatty acid oxidation and impairing its clearance by disrupting catalase function. The resulting H2O2 overload induces lethal mitochondrial damage, ultimately driving ovarian cancer cell death (Graphical Abstract, Figure S10).

Traditionally, vitamin B6 has been associated with potential anticancer effects attributed to the antioxidant and anti-inflammatory properties of its vitamers, particularly PLP, PN and PM (35,36). Notably, PLP has been reported to inhibit mitochondrial ROS production(37). While the bioactivities of vitamin B6 vitamers (PN, PM) are linked to antioxidant capacity and PLP is reported to suppress mitochondrial ROS, the cytotoxicity of pyridoxal (PL) has been narrowly attributed to its metabolic conversion to PLP (18), we directly challenge this view by establishing PL’s intrinsic, PLP-independent cytotoxic mechanism. Our comprehensive evidence-including PL’s superior potency over PLP/PN/PM across diverse models (PDOs, primary cells), the efficacy of a non-phosphorylatable analog, and the absence of PLP elevation-definitively uncovers a novel anticancer pathway, extending the preliminary observation of stable PLP levels(38) into a mechanistic paradigm.

Although multi-omics pathway enrichment initially suggested a potential link to ferroptosis detailed investigation ruled out ferroptosis as the primary death mechanism. PL treatment did not modulate key ferroptosis regulators, antioxidant proteins, intracellular ferrous iron levels and lipid peroxidation. Instead, we identified a profound, selective accumulation of H2O2 as the dominant ROS species driving PL’s cytotoxicity. While less reactive than other radicals such as hydroxyl radicals or superoxide (O2•⁻), its accumulation represents a recognized vulnerability in cancer cells, which often exhibit elevated basal H2O2 levels compared to normal cells(39). This has spurred therapeutic strategies exploiting H2O2, such as catalytic nanoparticles for oxygen generation or H2O2-responsive drug carriers(39,40). Our finding that PL rapidly and potently boosts tumor H2O2 production offers a novel approach to leverage this vulnerability for ovarian cancer therapy.

While mitochondria and peroxisomes are primary cellular sources of H2O2, our data decisively exclude mitochondrial ROS as the origin of PL-induced H2O2 overload. This conclusion is supported by the ineffectiveness of the mitochondrial antioxidant Mito-Tempo, the observed reduction in mitochondrial O2•⁻, and the direct measurement of suppressed H2O2 generation in PL-treated isolated mitochondria. Consequently, the mitochondrial dysfunction we observed-including the inhibition of respiratory activity (indicated by reduced O2•⁻) and the attendant disruption of the TCA cycle-is more accurately interpreted as a downstream consequence of oxidative stress. Besides, a live-cell biosensor, peroxisome-targeted H2O2 probe (HyPer7-peroxi), directly visualized a specific increase in H2O2 within peroxisomes after PL treatment. Collectively, these findings reposition mitochondria from initiators to key sensors and executioners of the peroxisome-derived H2O2 signal, integrating metabolic disruption into the cell death pathway. This mitochondrial dysfunction has a direct and catastrophic impact on core metabolism. The observed loss of ΔΨm and, crucially, the suppression of electron transport chain (ETC)-derived ROS production, indicate a severe deceleration of oxidative phosphorylation. Since ETC activity is tightly coupled to and drives the TCA cycle, its inhibition leads to a functional stagnation of metabolic flux. This explains the significant accumulation of TCA cycle intermediates observed in our metabolomic data. Thus, PL-induced peroxisomal H2O2 overload triggers a cascade that culminates in the blockade of central carbon metabolism, effectively starving cancer cells of both energy and biosynthetic precursors.

This redirected our focus to peroxisomes, the other major cellular source of H2O2. According to the established metabolic partnership, peroxisomes initiate the β-oxidation of very-long-chain fatty acids, exporting the shortened products as acylcarnitines for mitochondrial completion(41). In our study, PL-induced peroxisome proliferation, coupled with the significant accumulation of medium- and short-chain acylcarnitines, strongly suggests an enhancement of this peroxisomal phase of FAO. Although mitochondrial fatty acid import machinery (CPT1/2) was upregulated, the concurrent inhibition of mitochondrial respiration likely created a bottleneck, preventing the efficient oxidation of these peroxisomally derived substrates. Further, co-treatment with ACOX1i significantly rescued PL-induced cell death, providing direct evidence that enhanced peroxisomal FAO activity is required for PL’s cytotoxicity. Thus, the acylcarnitine profile reflects a state of disrupted metabolic coordination, where peroxisomal production is accelerated against a backdrop of mitochondrial dysfunction.

Peroxisome abundance is dynamically regulated by biogenesis (de novo, fission) and degradation (pexophagy) to maintain redox homeostasis(42,43). We demonstrate that PL drives peroxisome proliferation predominantly through PEX11A-mediated fission, without altering pexophagy or PEX5-dependent degradation. Critically, nascent peroxisomes generated via fission lack a protein synthesis system and rely on the PEX import machinery to receive functional proteins like CAT. PL treatment disrupted this critical process by impairing the PEX5-CAT interaction, despite increasing total CAT and PEX5 protein levels. This resulted in a population of proliferated but CAT-deficient peroxisomes. The consequent functional shift-enhanced H2O2-generating capacity coupled with crippled scavenging-transforms them into net H2O2 generators. Interestingly, similar peroxisomal alterations (proliferation with reduced CAT import) are linked to redox dysfunction in senescence (21,44,45). However, PL did not induce senescence in cancer cells, indicating that its primary anticancer effect stems from direct cytotoxicity driven by this engineered peroxisomal H2O2 overload. This mechanism distinguishes PL from other known peroxisome proliferators, such as fibrates. While those agents primarily induce peroxisome biogenesis through long-term transcriptional activation via PPARs, PL acts directly and rapidly on the organelle’s dynamics and import machinery to create a lethal redox imbalance. This fundamental difference in the mode of action-direct organelle remodeling versus indirect transcriptional adaptation-may underlie the potent and selective cytotoxicity of PL compared to the generally milder effects of transcriptional proliferators.

Collectively, our study unveils a novel, non-canonical anticancer mechanism of pyridoxal (PL), demonstrating that its potent cytotoxicity in ovarian cancer is independent of PLP conversion. By hijacking peroxisomal dynamics to induce H2O2 overload and subsequent mitochondrial damage, PL thereby exploits a unique metabolic vulnerability distinct from conventional chemotherapeutics. Importantly, this unique mechanism underlies its capacity to circumvent common drug resistance pathways, as evidenced by its potent efficacy in patient-derived organoids (PDOs) from platinum-, PARPi-, and paclitaxel-resistant tumors. Furthermore, a sub-toxic concentration of PL significantly sensitized cells to paclitaxel (PTX), highlighting its potential to synergize with standard chemotherapy. The translational promise of this strategy is bolstered by a favorable therapeutic window: PL exhibited significantly lower toxicity toward normal cells and, at efficacious doses in vivo, showed no significant hematological, hepatic, or renal toxicity. This work not only redefines PL’s biological activity but also establishes peroxisomal H2O2 metabolism as a promising therapeutic axis. Future research should explore combining PL with other H2O2-exploiting strategies(39,40) and investigate its potential in other cancers with peroxisomal dependencies.

Conclusions

Our study redefines pyridoxal (PL) as a potent, PLP-independent cytotoxic agent against ovarian cancer. PL triggers peroxisomal proliferation via PEX11A-dependent fission and impairs catalase function, switching peroxisomes from H2O2 scavenging to massive H2O2 production. This H2O2 surge causes lethal mitochondrial damage and ovarian cancer cell death. These findings reveal a novel peroxisomal mechanism for cancer therapy, positioning PL (or its unphosphorylated derivatives) and peroxisomal H2O2 metabolism as promising therapeutic targets.