Epilepsy is a chronic central nervous system disease that affects people of all ages, impacting over 70 million people globally (1). It is characterized by persistent seizures originating from abnormal neuronal discharges in the brain, exhibiting paroxysmal, transient, repetitive and stereotyped features (2). Studies have shown that approximately 51 million people worldwide are actively affected by epilepsy as of February 24, 2023 and around 4.9 million new cases emerge annually based on global population and epilepsy incidence rates (3). According to the WHO’s 2010 Global Burden of Disease study, epilepsy is ranked as the second most burdensome neurological disorder worldwide in terms of disability-adjusted life years (4). Despite studies indicating seizure control in over 2/3 patients using antiepileptic drugs, approximately 1/3 develop drug-resistant epilepsy (5). The pathogenesis of epilepsy is complex, often resulting from the combined effect of multiple factors. Age, family history of epilepsy, febrile convulsions, preterm birth, alcohol consumption, central nervous system infection, traumatic brain injury, stroke and sleep deprivation are well-established risk factors for epilepsy (6, 7). However, these traditional risk factors can only partially explain the risk of epilepsy. Therefore, accurately identifying novel epilepsy-related risk factors has become crucial for exploring new therapeutic targets and achieving early diagnosis and treatment.

Recent studies increasingly unveil metabolic or lipid profiles distinguishing responders from nonresponders to anti-seizure medications (ASMs), yet their intricate role in epilepsy remains largely obscure due to complexity (8–10). Lipidomics is an emerging discipline dedicated to the systematic examination of lipids in biological systems, encompassing various fatty acids (FAs) and lipid subclasses. An integrated metabolomics and lipidomics investigation of children with drug-refractory epilepsy (DRE) revealed decreased plasma fatty acids but elevated triglycerides (8). An investigation into kainic acid (KA)-induced acute epilepsy models in mice delineated alterations in phospholipid, endocannabinoid, endocannabinoid-like molecules, arachidonic acid and eicosanoid levels in the brain and periphery (11). Subsequent animal research demonstrated that subchronic administration of palmitoylethanolamide markedly diminished seizure intensity and fostered neuroprotection, which can also modulate plasma and hippocampal endocannabinoid and eicosanoid levels in systemic KA-injected mice (12). An investigation into metabolomic alterations in adult status epilepticus (SE) patients showed that metabolic dysregulation primarily impacted amino acid metabolism, pyrimidine metabolism and lipid homeostasis (13). The existing association between epilepsy and plasma liposomes primarily arises from observational studies characterized by limited sample sizes and numerous confounding variables. The nature of these studies makes it difficult to rule out the possibility of reverse causality, thereby compromising their reliability in establishing causality. In essence, while the correlation between epilepsy and plasma liposomes is evident, the causal relationship remains ambiguous, necessitating further research to elucidate the intricate pathophysiological mechanisms.

Inflammatory cytokines play an important role in infection and inflammation, which serve as crucial mediators for communication between the brain and the nervous system, contributing significantly to epilepsy (14). Neuroinflammation serves as a dual-edged sword, offering protection to the central nervous system as an immune response yet becoming detrimental when overactivated. TNF- α are the main inflammatory factors in the pathogenesis of epilepsy (15). In neurological disorders, including epilepsy, excess TNF damages the central nervous system by causing extracellular glutamate to accumulate to levels sufficient to inhibit synaptic activity or kill neurons (16). In a systematic review, the majority of studies corroborate the heightened expression of TNF- α in epilepsy patients, linked to a higher susceptibility to seizures in the future, albeit with inconsistency (15). However, the results of the study regarding the association between inflammatory cytokines and epilepsy lack consistency, warranting further investigation.

Nevertheless, further exploration is warranted to elucidate the specific contribution of plasma lipidome and inflammatory cytokines to epilepsy. Similar to the randomized controlled trial (RCT), the Mendelian randomization (MR) study is an analytical method of epidemiological etiological inference, relying on Mendelian independent distribution to investigate causal relationships between exposure and outcome variables (17). Bayesian weighted Mendelian randomization (BWMR) integrates the statistical methodologies of MR and Bayesian inference to enhance the precision of assessing causal relationships between exposure factors and outcome variables (18). In the MR study, single nucleotide polymorphisms (SNPs) are regarded as an instrumental variables (IVs) to estimate the causal relationship between exposure and outcome (19). Previous studies have found the causal relationship between human blood metabolites, gut microbiota, circulating lipids, psychiatric disorders and epilepsy (20–24). In recent years, an increasing number of studies have examined the association between plasma lipidome and epilepsy, yet no research has explored this relationship at the genetic level. To the best of our knowledge, the causal association between a broad range of plasma lipidome and epilepsy has not been established using MR. This paper conducted a comprehensive mediation MR analysis of plasma lipidome and epilepsy datasets using a large-scale genome-wide association study (GWAS), introducing inflammatory factors as mediators to ascertain the causal relationship between plasma lipidome and epilepsy.

The comprehensive mediation MR study was the first to explore the causal relationship between Phosphatidylcholine (18:1_18:1) and epilepsy risk, which revealed a significant causal effect of elevated levels of Phosphatidylcholine (18:1_18:1) on increased susceptibility to epilepsy. Additionally, the results of mediation MR analysis suggest that TNFSF12 may account for 4.58% of the mediating protective effect in the causal pathway from Phosphatidylcholine (18:1_18:1) to epilepsy. This analysis highlights the association between Phosphatidylcholine (18:1_18:1) and epilepsy, emphasizing the mediating role of TNFSF12.

Phosphatidylcholine, an amphoteric molecule, consists of a hydrophilic head and a hydrophobic tail, characterized by a bile base group inserted in the head. Comprising phosphate, choline and fatty acids, it serves as a crucial component of biological membranes and can participate in the transmission of nerve signals (33, 34). Studies indicates that Phosphatidylcholine contributes to brain cell repair, neural connectivity and cognitive enhancement, thereby regulating brain function (35, 36). Cholinergic neurons, through the release of acetylcholine (ACh), modulate neuronal communication, synaptic dynamics, neuroplasticity and hippocampal neurogenesis (37). However, inactivation and imbalance of ACh are associated with central nervous system diseases such as epilepsy. Anticholinergic drugs can block and reduce ACh synaptic activity in the central nervous system, which have been used to treat diseases such as Parkinson’s disease and epilepsy (38, 39). A traumatic brain injury study showed that treatment with anticholinergic drugs in the acute phase after brain injury reduced the frequency and severity of seizures, as well as the number of spontaneous recurrent seizures (40). Animal studies demonstrate that the use of an inhibitory acetylcholine engineered channel activated by exogenous and endogenous agonists can control acute and chronic seizures in male mice (41). Our findings indicate that Phosphatidylcholine (18:1_18:1) could increase the risk of epilepsy, while its reduction could potentially suppress seizures. These findings could provide a promising avenue for future epilepsy research.

TNFSF12, also known as tumor necrosis factor-like weak inducer of apoptosis (TWEAK), is a member of the tumor necrosis factor superfamily implicated in inflammation, apoptosis, autoimmunity and nervous system development (42, 43). Inflammatory cytokines serve as crucial mediators in infection and inflammation, facilitating communication between the brain and the immune system, which are pivotal in epilepsy pathogenesis (14). A study demonstrated that TNFSF12 as a pivotal regulator of astrocyte reactivity, inducing mitochondrial oxidative stress along with protein kinase Cδ and signal transducer and activator of transcription activation (44). This process is accompanied by inflammasome activation and the upregulation and release of proinflammatory cytokines. Astrocytes, pivotal in protecting brain tissue through reactive astrogliosis, exhibit compromised function in epileptic foci, contributing to epileptogenesis (45). Astrocytes in the central nervous system show morphological and functional diversity in brain region-specific patterns. Functional alterations in reactive astrocytes are commonly seen in human temporal lobe epilepsy, while reactive astrocyte-mediated neuroinflammation may contribute to the development of hippocampal epilepsy in animal models (46). Although the precise mechanism of TNFSF12 action in epilepsy remains elusive, our study suggests its potential in reducing epilepsy risk, offering a promising avenue for future clinical investigation.

We conducted the first mediation MR study to explore the causal relationship among plasma liposomes, inflammatory cytokines and epilepsy. Our findings from mediation MR analyses suggest that Phosphatidylcholine (18:1_18:1) may elevate epilepsy risk by reducing TNFSF12 levels, as evidenced by data from large GWAS cohorts utilizing genetic variants. In addition to employing various sensitivity analyses, we accounted for confounding factors and reverse causality, thereby bolstering theoretical underpinnings for epilepsy treatment, prevention and management approaches.

Our study possesses several advantages. Firstly, the sample size is substantial, enhancing the efficacy of the analysis, particularly with the inclusion of large-scale GWAS data. Additionally, we employed various sensitivity analysis techniques to mitigate the influence of confounding factors and reverse causality, thus ensuring the reliability of our causal effect estimates derived from observational studies (47, 48).

The limitations of the present study are as follows: (i) Since the number of IVs fulfilling the strict threshold (p < 5 × 10⁻⁸) was extremely small, a more relaxed threshold (p < 1 × 10⁻⁵) was adopted for screening IVs. (ii) Although the study involved a large sample size, it is crucial to acknowledge that the participants were exclusively of European ancestry, potentially limiting the generalizability of our findings to other regions or ethnicities. (iii) Furthermore, our results remain theoretical and await validation through clinical or animal experiments. Further cellular, animal and clinical investigations are warranted to elucidate these mechanisms.

Our mediation MR research suggests potential causal relationships among the plasma lipidomes, inflammatory cytokines and epilepsy. Specifically, TNFSF12 mediates the regulatory effect of Phosphatidylcholine (18:1_18:1) on epilepsy. These findings offer genetic evidence of the association between TNFSF12, Phosphatidylcholine (18:1_18:1) and epilepsy. Additionally, our research may facilitate the identification of biochemical markers for predicting, screening and early diagnosing of epilepsy, opening new avenues for exploring its biological mechanisms.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent from the patients/participants or patients/participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

XW: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. WX: Conceptualization, Data curation, Investigation, Methodology, Writing – review & editing. ML: Methodology, Validation, Writing – review & editing. LW: Conceptualization, Software, Writing – review & editing. YZ: Methodology, Project administration, Supervision, Writing – review & editing. CZ: Methodology, Project administration, Writing – review & editing. WL: Formal analysis, Project administration, Resources, Supervision, Validation, Writing – review & editing. SC: Methodology, Project administration, Resources, Visualization, Writing – review & editing. HH: Funding acquisition, Resources, Supervision, Visualization, Writing – review & editing.