Reactive oxygen species leading to growth inhibition via lipid peroxidation and decreased cell proliferation

You Song, Li Xie, Knut Erik Tollefsen

Norwegian Institute for Water Research (NIVA), Sognsveien 72, 0855, Oslo, Norway

This adverse outcome pathway (AOP 326) describes a linear route by which increased reactive oxygen species (ROS) can lead to decreased organismal growth through lipid peroxidation-mediated impairment of mitochondrial bioenergetics. In this AOP, increased ROS is treated operationally as the molecular initiating event because it represents the earliest common measurable redox perturbation shared by many chemical and non-chemical stressors within the broader ROS-growth AOP network. Increased ROS leads to oxidative stress, which promotes lipid peroxidation. Oxidative damage to membrane lipids, particularly polyunsaturated fatty acids in cellular and mitochondrial membranes, can alter membrane integrity, proton conductance, mitochondrial membrane potential, and respiratory efficiency. These effects reduce coupling of oxidative phosphorylation (OXPHOS), decrease the cellular ATP pool, impair energy-dependent cellular proliferation, and ultimately reduce organismal growth.

    AOP 326 reuses and connects established AOP-Wiki components from two important AOP contexts. The upstream oxidative stress component is associated with AOP 478, in which deposition of energy leads to oxidative stress through increased free radical generation, with subsequent oxidative molecular damage (AOP-Wiki, 2026a). This provides a curated AOP-Wiki context for the use of oxidative stress as a conserved hub KE downstream of free radical or ROS-producing stressors. The downstream bioenergetic and growth segment is directly associated with AOP 263, which causally links decreased coupling of OXPHOS to growth inhibition through ATP depletion and decreased cell proliferation and has been published in the OECD Series on Adverse Outcome Pathways (AOP-Wiki, 2026b; OECD, 2022; Song and Villeneuve, 2021). Thus, AOP 326 links an oxidative-stress/lipid-damage module to an OECD-endorsed bioenergetics-to-growth module. The AOP is relevant to environmental and human health contexts because ROS production, lipid peroxidation, mitochondrial ATP generation, cell proliferation, and organismal growth are broadly conserved in aerobic eukaryotes.    Empirical support is derived from studies in algae, daphnids, mollusks, fish, mammalian systems, and human cells exposed to redox-active chemicals, metals, pesticides, hypoxia-reoxygenation, radiation, and endogenous oxidative stressors. This AOP can support mechanistic interpretation of oxidative stress-mediated growth impairment, assay selection, chemical prioritization, integrated approaches to testing and assessment (IATA), and development of quantitative AOP approaches for mitochondrial and oxidative stress-related toxicity.

Acknowledgement

This project was funded by the Research Council of Norway (RCN), grant no. RCN-315929 “EXPECT: In silico and experimental screening platform for characterizing environmental impact of industry development in the Arctic” (https://www.niva.no/en/projects/expect), the European Partnership for the Assessment of Risks from Chemicals (PARC) through European Union’s Horizon Europe research and innovation programme (Grant Agreement No 101057014, and supported by the NIVA Computational Toxicology Program, NCTP (https://www.niva.no/en/featured-pages/nctp, grant. No. RCN-342628).

AI disclosure

Artificial intelligence (AI) tools were used to support literature prioritization, review and AOP-Wiki page preparation in this work. AOP-helpFinder was used for automated literature mining, and ChatGPT (OpenAI) was used as an auxiliary tool for title and abstract screening, extraction of study metadata, and identification of potential weight-of-evidence indicators. AI-assisted outputs were used only to organize and prioritize information and were verified against the original sources by the authors before inclusion. Additional AI assistance was used for formatting, copy-editing, citation cross-checking, and harmonization of the AOP-Wiki pages. All scientific interpretations, weight-of-evidence judgments, final wording, and conclusions were determined and approved by the authors, who take full responsibility for the content and integrity of the work.

ROS are continuously formed during aerobic metabolism and can also be generated in response to environmental stressors. At controlled levels, ROS participate in redox signaling, whereas excessive ROS can disturb redox homeostasis and initiate oxidative stress (Schieber and Chandel, 2014; Sies et al., 2017). Lipids are important targets of oxidative attack because membrane phospholipids, especially those containing polyunsaturated fatty acyl chains, can undergo radical-driven peroxidation. Lipid peroxidation generates lipid hydroperoxides and secondary reactive products, including malondialdehyde and 4-hydroxy-2-nonenal, which can alter membrane structure, propagate oxidative damage, and affect protein and organelle function (Ayala et al., 2014).

    AOP 326 was developed to represent the lipid peroxidation-driven linear route within the broader ROS-growth AOP network. This route was selected because lipid peroxidation is a well-established consequence of oxidative stress and because mitochondrial membranes are central determinants of OXPHOS coupling. Peroxidative modification of mitochondrial membrane lipids can alter membrane fluidity, proton leak, respiratory control, and mitochondrial membrane potential, providing a mechanistically coherent bridge from oxidative stress to impaired ATP production (Murphy, 2009; Nicholls and Ferguson, 2013; Ouillon et al., 2021). The downstream sequence from decreased coupling of OXPHOS to decreased ATP pool, decreased cell proliferation, and decreased growth is already represented in AOP 263 and provides the growth-relevant terminal module for AOP 326 (AOP-Wiki, 2026b; OECD, 2022; Song and Villeneuve, 2021).

AOP 326 was developed using the principles described in OECD AOP guidance, including modular description of KEs and KERs, reuse of existing AOP-Wiki content where appropriate, evidence evaluation using biological plausibility, empirical support, essentiality, and quantitative understanding, and clear description of the biological domain of applicability (OECD, 2018, 2021). The purpose was to assemble a focused linear pathway from reusable AOP-Wiki elements rather than to create an isolated de novo pathway. This is particularly important because AOP 326 is one branch of the broader ROS-growth AOP network and because its KEs and KERs overlap with oxidative stress, mitochondrial dysfunction, energy metabolism, cell proliferation, and growth-related AOPs.

    Reuse of existing AOP-Wiki content was considered at the beginning of development. AOP 478 was reviewed because it provides an AOP-Wiki precedent for oxidative stress as a central KE downstream of free radical generation, and because it describes oxidative stress as a pathway by which energy deposition can damage biological molecules (AOP-Wiki, 2026a). Although AOP 478 is not a lipid peroxidation-to-growth AOP, its use of KE 1392 (Increase, Oxidative stress) and its oxidative damage context support the upstream portion of AOP 326. AOP 263 was reviewed because it provides the directly relevant downstream module consisting of KE 1446 (Decrease, Coupling of OXPHOS), KE 1771 (Decrease, ATP pool), KE 1821 (Decrease, Cell proliferation), and AO 1521 (Decrease, Growth), together with KERs 2203, 2204, and 2205 (AOP-Wiki, 2026b; OECD, 2022; Song and Villeneuve, 2021). In AOP 326, the novel linking logic is therefore the connection of ROS-driven oxidative stress and lipid peroxidation to the existing OXPHOS-ATP-cell proliferation-growth module.

    The evidence base was assembled through a structured AI-human hybrid workflow. Search terms were first developed for the events in the pathway, including KE names, synonyms, endpoint names, assay terms, taxa, and species. AOP-helpFinder was then used to search PubMed for co-occurrence of KE-related terms, following approaches previously described for literature mining in support of AOP development (Carvaillo et al., 2019; Jornod et al., 2022). The exported AOP-helpFinder results included PMIDs, titles, abstracts, and matched KE terms. Overlap analysis was applied to remove redundant hits and filter literature that was unrelated to the taxa or biological systems considered relevant to this AOP.

    A large language model (LLM) was then used as an auxiliary screening and structuring tool. During abstract pre-screening, the LLM was used to extract study metadata such as stressor, species, biological system, dose or concentration, and exposure time; to identify whether a study provided evidence for biological plausibility, empirical support, essentiality, or quantitative understanding; and to flag indicators of dose-response, temporal, or incidence concordance. Studies were provisionally classified as high, medium, or low priority. High-priority studies were retrieved for full-text review, medium-priority studies were reserved for supporting evidence, and low-priority studies were documented as low relevance. For studies passing the abstract screen, the LLM was used to assist full-text review by organizing information relevant to the KER evidence table. All LLM outputs were checked manually against the full text by expert reviewers before any evidence was accepted.

    The final phase consisted of manual expert curation. Experts verified the LLM-assisted extractions, populated KER evidence tables with methods, endpoints, results, weight-of-evidence categories, and references, and made final calls for biological plausibility, empirical support, essentiality, and evidence gaps. Targeted manual searches were also used to fill gaps identified during evidence curation, especially for the lipid peroxidation to OXPHOS coupling relationship and the downstream AOP 263 KERs. Searches focused on combinations of terms such as reactive oxygen species, oxidative stress, lipid peroxidation, malondialdehyde, 4-hydroxynonenal, mitochondrial membrane potential, proton leak, oxidative phosphorylation, ATP depletion, cell proliferation, growth inhibition, paraquat, copper, cadmium, hypoxia-reoxygenation, Daphnia, algae, bivalves, fish, and AOP. Studies were prioritized when they measured two or more KEs in the same biological system, reported exposure dose or concentration and time, or provided evidence relevant to dose-response, temporal, or incidence concordance.

The overall weight of evidence supporting AOP 326 is considered moderate to high. Biological plausibility is high for all six KERs in the pathway, reflecting well-established mechanistic connections between ROS, oxidative stress, lipid peroxidation, mitochondrial membrane integrity, OXPHOS coupling, ATP production, cell proliferation, and growth. The downstream bioenergetic module from decreased coupling of OXPHOS through decreased ATP pool and decreased cell proliferation to decreased growth (KERs 2203, 2204, 2205) is directly reused from OECD-endorsed AOP 263, which was published in the OECD Series on Adverse Outcome Pathways and has independently established high to moderate WoE for these relationships (OECD, 2022; Song and Villeneuve, 2021). The upstream oxidative stress-lipid peroxidation module is supported by high empirical concordance across algae, crustaceans, bivalves, fish, and mammalian systems. The novel linking relationship from lipid peroxidation to decreased OXPHOS coupling (KER 1599) is rated moderate in empirical support, as studies measuring both endpoints concurrently in the same experimental system are less common, although mechanistic evidence for cardiolipin oxidation and inner mitochondrial membrane integrity is strong. Essentiality is rated moderate to high, with the strongest support for the OXPHOS-to-ATP segment shared with AOP 263. Quantitative understanding is highest for the AOP 263-derived downstream module and low to moderate for the upstream lipid peroxidation-to-OXPHOS segment. The main uncertainties are the directionality and quantitative strength of the lipid peroxidation-to-OXPHOS relationship, given that mitochondrial dysfunction can also promote lipid peroxidation, and the potential for compensatory glycolytic ATP production to buffer depletion. AOP 326 is suitable for mechanistic interpretation, IATA development, and chemical prioritisation for oxidative stress-mediated growth impairment affecting mitochondrial energy metabolism (OECD, 2018; OECD, 2022; Becker et al., 2015).

AOP 326 can support mechanistic interpretation of growth impairment caused by oxidative stressors that induce lipid peroxidation and mitochondrial bioenergetic dysfunction. The AOP is particularly relevant for hazard identification and chemical prioritization when evidence indicates increased ROS or oxidative stress together with lipid peroxidation, mitochondrial membrane potential changes, OXPHOS impairment, ATP depletion, reduced proliferation, or growth inhibition. The AOP may also support IATA development by linking upstream NAM endpoints, such as ROS assays, lipid peroxidation markers, mitochondrial membrane potential, oxygen consumption rate, ATP content, and proliferation assays, to an apical growth endpoint.

AOP 326 can also support chemical grouping and read-across for stressors that share evidence of oxidative lipid damage and mitochondrial bioenergetic impairment. Because oxidative stress and lipid peroxidation are not chemical-specific, this AOP should not be used as a stand-alone basis for regulatory decisions. Instead, it should be used as part of a weight-of-evidence framework that considers stressor mode of action, exposure context, taxonomic relevance, assay specificity, and concordance across multiple KEs. The AOP also highlights important method-development needs, particularly standardized assays for lipid peroxidation, mitochondrial coupling, and ATP depletion that can be applied across taxa and integrated into quantitative AOP approaches.

AOP-Wiki. 2026a. AOP 478: Deposition of energy leading to occurrence of cataracts. Collaborative Adverse Outcome Pathway Wiki. Available from: https://aopwiki.org/aops/478.

AOP-Wiki. 2026b. AOP 263: Uncoupling of oxidative phosphorylation leading to growth inhibition via decreased cell proliferation. Collaborative Adverse Outcome Pathway Wiki. Available from: https://aopwiki.org/aops/263.

Ayala A, Munoz MF, Arguelles S. 2014. Lipid peroxidation: production, metabolism, and signaling mechanisms of malondialdehyde and 4-hydroxy-2-nonenal. Oxidative Medicine and Cellular Longevity 2014:360438.

Barata C, Varo I, Navarro JC, Arun S, Porte C. 2005. Antioxidant enzyme activities and lipid peroxidation in the freshwater cladoceran Daphnia magna exposed to redox cycling compounds. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology 140:175-186.

Bonora M, Patergnani S, Rimessi A, De Marchi E, Suski JM, Bononi A, Giorgi C, Marchi S, Missiroli S, Poletti F, Wieckowski MR, Pinton P. 2012. ATP synthesis and storage. Purinergic Signaling 8:343-357.

Carvaillo JC, Barouki R, Coumoul X, Audouze K. 2019. Linking bisphenol S to adverse outcome pathways using a combined text mining and systems biology approach. Environmental Health Perspectives 127:047005.

Cherkasov AS, Biswas PK, Ridings DM, Ringwood AH, Sokolova IM. 2006. Effects of acclimation temperature and cadmium exposure on cellular energy budgets in the marine mollusk Crassostrea virginica: linking cellular and mitochondrial responses. Journal of Experimental Biology 209:1274-1284.

Conlon I, Raff M. 1999. Size control in animal development. Cell 96:235-244.

Esperanza M, Cid A, Herrero C, Rioboo C. 2015. Acute effects of a prooxidant herbicide on the microalga Chlamydomonas reinhardtii: screening cytotoxicity and genotoxicity endpoints. Aquatic Toxicology 165:210-221.

Jamers A, De Coen W. 2010. Effect assessment of the herbicide paraquat on a green alga using differential gene expression and biochemical biomarkers. Environmental Toxicology and Chemistry 29:893-901.

Jornod F, Jaylet T, Blaha L, Sarigiannis D, Tamisier L, Audouze K. 2022. AOP-helpFinder webserver: a tool for comprehensive analysis of the literature to support adverse outcome pathways development. Bioinformatics 38:1173-1175.

Knauert S, Knauer K. 2008. The role of reactive oxygen species in copper toxicity to two freshwater green microalgae. Journal of Phycology 44:311-321.

Moore TD, Martin-Creuzburg D, Yampolsky LY. 2023. Diet effects on longevity, heat tolerance, lipid peroxidation and mitochondrial membrane potential in Daphnia. Oecologia 202:151-163.

Murphy MP. 2009. How mitochondria produce reactive oxygen species. Biochemical Journal 417:1-13.

Nestler H, Groh KJ, Schonenberger R, Eggen RIL, Suter MJF. 2012. Multiple-endpoint assay provides a detailed mechanistic view of responses to herbicide exposure in Chlamydomonas reinhardtii. Aquatic Toxicology 110-111:214-224.

Nicholls DG, Ferguson SJ. 2013. Bioenergetics 4. London: Academic Press.

Nicotera P, Leist M, Ferrando-May E. 1998. Intracellular ATP, a switch in the decision between apoptosis and necrosis. Toxicology Letters 102-103:139-142. https://doi.org/10.1016/S0378-4274(98)00298-7

OECD. 2018. Users' handbook supplement to the guidance document for developing and assessing adverse outcome pathways. OECD Series on Adverse Outcome Pathways No. 1. Paris: OECD Publishing.

OECD. 2021. Guidance document for the scientific review of adverse outcome pathways. OECD Series on Testing and Assessment No. 344. Paris: OECD Publishing.

OECD. 2022. Uncoupling of oxidative phosphorylation leading to growth inhibition via decreased cell proliferation. OECD Series on Adverse Outcome Pathways No. 28. Paris: OECD Publishing.

Ouillon N, Sokolov EP, Otto S, Rehder G, Sokolova IM. 2021. Effects of variable oxygen regimes on mitochondrial bioenergetics and reactive oxygen species production in a marine bivalve, Mya arenaria. Journal of Experimental Biology 224:jeb237156.

Qian H, Chen W, Sun L, Jin Y, Liu W, Fu Z. 2009. Inhibitory effects of paraquat on photosynthesis and the response to oxidative stress in Chlorella vulgaris. Ecotoxicology 18:537-543.

Schieber M, Chandel NS. 2014. ROS function in redox signaling and oxidative stress. Current Biology 24:R453-R462.

Sies H, Berndt C, Jones DP. 2017. Oxidative stress. Annual Review of Biochemistry 86:715-748.

Sokolova IM, Sokolov EP, Ponnappa KM. 2005. Cadmium exposure affects mitochondrial bioenergetics and gene expression of key mitochondrial proteins in the eastern oyster Crassostrea virginica Gmelin (Bivalvia: Ostreidae). Aquatic Toxicology 73:242-255.

Song Y, Villeneuve DL. 2021. Uncoupling of oxidative phosphorylation leading to growth inhibition via decreased cell proliferation. Environmental Toxicology and Chemistry 40:2951-2963.

Xie L, Gomes T, Solhaug KA, Song Y, Tollefsen KE. 2018. Linking mode of action of the model respiratory and photosynthesis uncoupler 3,5-dichlorophenol to adverse outcomes in Lemna minor. Aquatic Toxicology 197:98-108. https://doi.org/10.1016/j.aquatox.2018.02.005

Xie L, Solhaug KA, Song Y, Brede DA, Lind OC, Salbu B, Tollefsen KE. 2019. Modes of action and adverse effects of gamma radiation in an aquatic macrophyte Lemna minor. Science of the Total Environment 680:23-34. https://doi.org/10.1016/j.scitotenv.2019.05.016

Xie L, Song Y, Petersen K, Solhaug KA, Lind OC, Brede DA, Salbu B, Tollefsen KE. 2022. Ultraviolet B modulates gamma radiation-induced stress responses in Lemna minor at multiple levels of biological organisation. Science of the Total Environment 846:157457. https://doi.org/10.1016/j.scitotenv.2022.157457