Epilepsy is a widespread and severe neurological disorder marked by sudden, simultaneous abnormal neuron activity in the brain. The rate of active epilepsy typically ranges between 4 and 12 per 1,000 (Thijs et al., 2019). Epilepsy’s incidence shows a bimodal distribution, with the highest risk seen in infants and the elderly (Thijs et al., 2019). Epilepsy remains a significant cause of disability and death, posing a considerable global societal burden. It can be categorized into focal epilepsy (FE) and generalized epilepsy (GE). The latter is subdivided into motor and non-motor (absence) epilepsy. In GE, the epileptogenic networks are widespread, covering thalamocortical structures bilaterally (Asadi-Pooya, 2019). By contrast, FE networks are limited to a single hemisphere, often involving limbic or neocortical regions (Allers et al., 2015). Both childhood absence epilepsy (CAE) and juvenile absence epilepsy (JAE) are part of an epileptic syndrome, commonly presenting with brief staring, rhythmic blinking, or motor automatisms and changes in the electroencephalogram (Kessler and McGinnis, 2019). While 65% of epilepsy patients find relief with antiepileptic medication (Kwan et al., 2011), about a third face drug resistance (Perucca et al., 2023). The frequent seizures, cognitive challenges, psychological disorders, and drug side effects profoundly impact these patients’ quality of life (Wang et al., 2023). Therefore, further exploring the etiology of epilepsy and new therapeutic options for patients has become necessary.

Recently, mounting evidence indicates a strong link between gut microbiomes (GMs) and human health (Järbrink-Sehgal and Andreasson, 2020). Research reveals that GM affects the central nervous system’s physiology, neurochemistry, behavior, and cognitive growth. The theory suggests that communication between the GM and the brain is facilitated through diverse channels, including the central and enteric nervous systems and endocrine and immune pathways (Grenham et al., 2011). This intricate relationship, the gut-brain axis, is maintained through multiple signaling pathways and connectivity networks. Studies have highlighted the role of GM and the gut-brain axis in neurological disorders such as depression, Alzheimer’s disease, and Parkinson’s disease. The hypothesis has been advised that GM dysbiosis is linked to development of epilepsy (Sorboni et al., 2022). The composition of GM in individuals with epilepsy is different from that of healthy control subjects, both before and after treatment (Gong et al., 2022). Additionally, disturbances at the level of specific GM taxa can influence the activity of neurons associated with epilepsy (Darch and McCafferty, 2022). The ketogenic diet (KD), a non-pharmacological option for those with treatment-resistant epilepsy (Socała et al., 2021), can reduce seizure attacks by 50%, as observed in 22–55% of adults and 60–75% of children (Özcan et al., 2022). Yet, whether alteration in GMs induced by KD is protective or detrimental remains debatable.

In general, neuroinflammation is a normal response that helps maintain homeostasis. However, excessive or persistent neuroinflammation can lead to cellular dysfunction (Xanthos and Sandkühler, 2014). Related studies have shown that neuroinflammatory responses are not only the result of epileptic seizures or brain neuropathology but are involved in their production (Prabowo et al., 2013; Pauletti et al., 2019). Activated microglia and astrocytes are the major cell types contributing to neuroinflammation (Vezzani et al., 2011, 2015). Several bioactive products from the GM can trigger the release of inflammatory factors by activating microglia (Kim et al., 2012). Therefore, we hypothesized that cytokines may be mediators in the pathway from GM to epilepsy. Cytokines are inflammatory regulators and, therefore, important intermediate phenotypes in inflammatory diseases. The pro-inflammatory cytokines IL-1β and IL-6 are important contributors to the inflammatory response in the brain. There is evidence that epilepsy is associated with elevated levels of pro-inflammatory cytokines (Vezzani et al., 2013). Moreover, microglia may, with the help of the above cytokines, play a role in the pathology of epilepsy through the IL-1 receptor/Toll-like receptor signaling, COX-2, and the TGF-β signaling process (Vezzani et al., 2016; Alvim et al., 2021; Soltani Khaboushan et al., 2022).

Robust evidence from randomized controlled trials (RCTs) verifies the connection between GM dysfunction and epilepsy onset. However, the nature of this relationship, whether causal or not, requires further study. Mendelian randomization (MR) is an advanced statistical method designed to explore causative associations between variables. It uses single nucleotide polymorphisms (SNPs) as instrumental variables. This method can help negate reverse causality and confounding factor effects. Still, no MR studies have focused on children with juvenile absence epilepsy, who often exhibit higher drug resistance (Symonds et al., 2021). Thus, our research employed a MR analysis using data from a published Genome-Wide Association Study (GWAS) to explore the potential causal link between GM and various epilepsy subtypes.

In this study, we conducted MR analyses to investigate the potential causal relationship between GMs and different epilepsy subtypes. We found that 30 GMs were associated with four epilepsy subtypes using extensive GWAS statistics, and no reverse causality was identified. Notably, the family Veillonellaceae in CAE displayed a significant causal association after FDR correction. Three types of cytokines have significant effects on epilepsy. These results could provide insights into the therapeutic role of epilepsy.

The role of GM in maintaining the state of epilepsy has gained increasing recognition. Yet, due to limited clinical trials examining the differences in the metagenome/metabolome between epilepsy patients and healthy individuals, the influence of GM on epilepsy remains elusive. In our conclusions, no reverse causality was found for epilepsy on GM. Considering modifications related to several external factors (e.g., diet and age), it is not easy to conduct relevant animal or clinical studies. Although some studies have found increased alpha diversity and abundance of thick-walled phyla in patients with drug-resistant epilepsy and altered abundance of Prevotella, Ruminococcus, and other flora in patients with cerebral palsy with epilepsy (Peng et al., 2018; Huang et al., 2019). The results of the available studies are not uniform, and it is therefore difficult to define relevant criteria (Russo, 2022). Previous research has highlighted the importance of Veillonellaceae in developing the intestinal microecosystem during early life (Lou et al., 2022). While direct evidence connecting Veillonellaceae and epilepsy is scarce, its influence on normal brain functions is documented. Furthermore, Veillonellaceae, part of the phylum Firmicutes, shows increased abundance in patients with drug-resistant epilepsy, which also decreases after antiepileptic treatment (Cheraghmakani et al., 2021). Research has identified a prevalent genus, Marvinbryantia, in epileptic animals, which correlates positively with excitatory neurotransmitters (Cheraghmakani et al., 2021). Our findings show that a positive causal relationship between the genus Marvinbryantia and GE aligns with these experimental results. Oliveira et al. observed a negative correlation between Marvinbryantia and d-glucose and lactate. This negative association may disrupt carbohydrate metabolism, possibly contributing to the mechanisms underlying epilepsy (Fei et al., 2020). Additionally, Marvinbryantia is a pro-inflammatory bacterium, possibly exacerbating inflammation and precipitating epileptic conditions (Oliveira et al., 2022). Our MR analysis highlighted Parabacteroides, belonging to the family Tannerellaceae, as a potential protective agent for JAE. Previously linked to increased ketosis and metabolic enhancement in humans (David et al., 2014), this bacterium might explain the seizure control seen with the KD (Bough and Rho, 2007). Olson et al. posited that the KD’s anti-seizure effects arise from increasing specific bacterial species, particularly Parabacteroides merdae. Together, these bacteria reduce the γ-glutamylation of amino acids and boost the hippocampus’s gamma-aminobutyric acid/glutamate ratio, preventing epilepsy (Olson et al., 2018).

Interestingly, 28 other GMs with nominal causal associations, such as Ruminococcus, echoed past findings. Similar to Veillonellaceae, Ruminococcus, another microbiome from phylum Firmicutes, increases in individuals with autism spectrum disorders (Wang et al., 2013). Ruminococcus causes reduced levels of N-acetyl aspartate, a marker of neuronal health, and is typically diminished in individuals diagnosed with epilepsy (Zhang et al., 2016; Mudd et al., 2017). The presence of this microorganism also correlates positively with glutamate and glutamine, both closely associated with the pathogenesis of epilepsy (Sun et al., 2016). Furthermore, Ruminococcus is associated with reduced levels of 5-hydroxytryptophan, which inhibits T-type calcium channels and subsequently reduces epileptiform discharges (Petersen et al., 2017). Additionally, the abundance of the phylum Verrucomicrobia, Bacteroides, and its related genus (including Bacteroides and Barnesiella) was observed to be higher in the drug-resistant epilepsy group than in the control group, aligning with our findings (Peng et al., 2018). Notably, subjects with cerebral palsy and epilepsy, potential risk factors for FE and GE, exhibited an elevated presence of Eubacterium (Peng et al., 2023). Moreover, Anaerostipes was found to be more prevalent in epilepsy patients (Gong et al., 2021). However, it is noteworthy that while Lachnospiraceae is more abundant in healthy individuals (Ding et al., 2021), a reduced presence was noted in temporal lobe epilepsy patients (Wei et al., 2023). This observation contrasts with our MR findings and necessitates further exploration.

Currently, the exact mechanism by which the GM contributes to epilepsy has not been determined. Although the available data do not confirm that cytokines are mediators between GM and epilepsy, we observed many results that are consistent with previous studies. As a protective factor in epilepsy, the IL-1 receptor antagonist anakinra can be used as a treatment for refractory epilepsy (Yamanaka et al., 2021), and vascular endothelial growth factor is considered an attractive target for epilepsy treatment (Lange et al., 2016). On the other hand, we have found that IL-2 increases the risk of epilepsy. As a factor that plays an important immune role, elevated levels of IL-2 over-activate signaling pathways, leading to a worse prognosis in temporal lobe epilepsy (Mazumder et al., 2019).

The precise mechanisms driving the relationship between GMs and epilepsy are not yet fully understood, but emerging evidence provides some initial insights into possible mechanisms. Various bioactive products from GMs have the potential to influence the brain either directly or indirectly. As mentioned above, GM can mediate neuroinflammation, and pro-inflammatory cytokines may contribute to epilepsy by triggering oxidative stress, increased PIC secretion and BBB disruption (Vezzani et al., 2013). For instance, lipopolysaccharides can directly affect the central nervous system by activating Toll-like receptors on microglia and triggering the release of inflammatory factors (Kim et al., 2012). In addition, enteroendocrine signaling and microbial metabolism are among the mechanisms that may contribute to epilepsy. By increasing the release of excitatory neurotransmitters such as glutamate and decreasing inhibitory neurotransmitters such as GABA, neurotransmission can be disrupted (Su et al., 2015; Rana and Musto, 2018). SCFAs are also thought to be protective against epilepsy (Ding et al., 2021). From an autoimmune perspective, GM can regulate microglia maturation as well as astrocyte activation (de Theije et al., 2011). This implies that GM can regulate epilepsy development by modulating innate immunity, adaptive immunity and inflammatory mechanisms (Ding et al., 2021). Notably, Bacteroidetes have been linked to conditions such as encephalitis and autoimmunity (Xu et al., 2020), which could play a role in epilepsy development. A theory also suggests that GMs might affect the hypothalamic–pituitary–adrenal axis, thereby increasing the likelihood of epilepsy (Sudo et al., 2004).

Combined with existing theories, our study has profound clinical implications and provides new ideas for the clinical management of epilepsy. The risk of epilepsy can be reduced by the use of probiotics. In the field of pharmacological microbiology, GMs and their secreted cytokines can be used as drug targets. He et al. used FMT to improve epilepsy in a 22-year-old girl (He et al., 2017). Additionally, GMs may serve as promising biomarkers for identifying the prognosis of epilepsy patients. For instance, a smaller Bacteroidetes/Firmicutes ratio may increase the risk of seizures and lead to poor prognosis (Citraro et al., 2021). The different abundance composition of GMs in the intestine can also be an indicator for evaluating the efficacy of KD (Xie et al., 2017). Furthermore, this study also has several strengths. First, it comprehensively examines the causal relationship between GM and various epilepsy subtypes. Unlike observational studies, this analysis is less vulnerable to confounding factors and reverse causality. Second, the paper’s statistical robustness is evident with a substantial GWAS data sample size for exposure and outcome and significant estimated effects for each genetic variable (F-value >10). Besides, a rigorous FDR correction was applied in the MR analysis to minimize type I errors. Third, we included cytokines to analyze possible mediating factors from GM to epilepsy and annotated genetic variants through FUMA. However, certain limitations warrant consideration. The GWAS data on GMs encompass various diseases and age groups without specific gender and age stratification. We included only individuals of European descent due to the absence of large-scale GWAS data for other ancestries, which might make these findings less generalizable. Lastly, while the MR analysis parallels the insights from RCT studies, further validation through animal experiments remains essential.

In conclusion, our MR study underscores potential causal connections between GMs and various epilepsy subtypes (GE, FE, JAE, and CAE), suggesting that the dysbiosis of the family Veillonellaceae might play a role in CAE onset. A mediating role of the cytokines has not been shown to exist between GM and epilepsy. Further studies are essential to understand the potential mechanisms by which GM might influence epilepsy treatment.

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

YQ: Conceptualization, Formal analysis, Writing – original draft. BS: Conceptualization, Formal analysis, Writing – original draft. MX: Data curation, Writing – original draft. YT: Data curation, Writing – original draft. ZY: Methodology, Software, Writing – original draft. MW: Methodology, Software, Writing – original draft. CM: Supervision, Validation, Writing – review & editing. ZC: Supervision, Validation, Writing – review & editing. ZW: Funding acquisition, Resources, Supervision, Writing – review & editing.