Epilepsy is a chronic neurological disease characterized by recurrent and transient brain dysfunction. It affects approximately 65 million patients worldwide and poses a substantial global health challenge (Fisher et al., 2014). Despite extensive research into the pathogenic mechanisms of epilepsy, its etiology remains unclear. Approximately one-third of patients with epilepsy who do not respond to currently available antiseizure medications develop drug-resistant epilepsy (DRE). Moreover, some daily factors, such as fatigue, hunger, alcohol or stimulant beverage consumption, insomnia, or infections can aggravate seizures. These observations underscore the complexity of epilepsy and the need for further exploration to understand its etiology and management.

Studies of the proposed concept of the microbiota-gut-brain axis (MGBA) have demonstrated its vital role in the pathological process of epilepsy (Yue et al., 2022). The gut microbiota (GM) is a complex group of symbiotic microorganisms (bacteria, viruses, and fungi) dwelling in the gastrointestinal tract, and is considered the second genome in humans (Ursell et al., 2012). MGBA facilitates communication between the brain and GM through various pathways. These include the autonomic and enteric nervous systems, immune system and neuroimmunity, neurotransmitters, short-chain fatty acids (SCFAs), spinal mechanisms, hypothalamic-pituitary-adrenal axis, and peptidoglycans, etc. (Cryan et al., 2019). Studies on the relationship between GM and epilepsy initially emerged with a ketogenic diet (Olson et al., 2018; Gong et al., 2021). Several clinical studies have shown substantial differences in the GM profiles among epilepsy patients, healthy individuals, and those with DRE (Arulsamy et al., 2020; Gong et al., 2020). The role of the GM in epileptogenesis and the development of epilepsy has been gradually confirmed in different epileptic animal models (Citraro et al., 2021; Mengoni et al., 2021; Mu et al., 2022b). The aforementioned evidence underscores the important role of the GM in epilepsy.

In clinical practice, it has been observed that the use of certain antibiotics may increase susceptibility to seizures, particularly within the β-lactam, fluoroquinolone families, fourth-generation cephalosporins, and carbapenems (Wanleenuwat et al., 2020). Ciprofloxacin (CPF) is an antibiotic belonging to the fluoroquinolone class that is effective against a wide range of bacteria. Previous studies have shown that quinolone antibiotics, including CPF, affect seizure susceptibility in various animal models and patients with epilepsy (Abdel-Zaher et al., 2012; Arafa et al., 2013; Cheraghmakani et al., 2021; Sivarajan and Ramachandran, 2023). When CPF is applied, fecal concentrations are high and CPF exerts a long-lasting effect on the GM (Zimmermann and Curtis, 2019). Previous studies on the mechanism of quinolone antibiotic-induced seizures have primarily focused on their potential to change neurotransmitter levels, disrupt neurotransmitter-receptor binding, have a chemical structure similar to that of epileptogens, and increase oxidative stress due to drug interactions (Ilgin et al., 2015; Wanleenuwat et al., 2020). However, the vital role of the GM has been neglected.

Therefore, the above-mentioned evidence predominantly examines the variations in GM between individuals with epileptic status and healthy controls, or among those undergoing different drugs or diet treatments. There is a notable lack of investigation into the therapeutic effects of GM modulation and the underlying mechanisms of the MGBA. Additionally, there has been a lack of recognition regarding the significant impact of CPF on the GM. Our study aims to combine the two aspects and fill this gap by focusing on elucidating the pathway of CPF induce seizure susceptibility through the GM.

To date, no studies have examined whether CPF increases seizure susceptibility from the perspective of GM and metabolite changes in animals. Recent studies have revealed that the GM is a major determinant of plasma metabolome, potentially playing a more dominant role than genetics (Loh et al., 2024). Therefore, based on the aforementioned research progress, we hypothesized that the GM plays a pivotal role in the mechanism by which CPF enhances seizure susceptibility, potentially via metabolic pathways in the MGBA. Our investigation is the first to delve into the GM and its relationship with seizure susceptibility induced by CPF using 16S ribosomal RNA amplicon sequencing. Furthermore, we explored the potential therapeutic effects of transplanting fecal samples from healthy controls. And we aimed to identify the core genera through joint analysis with lithium pilocarpine-induced epilepsy. Finally, we aimed to unravel the intrinsic mechanisms through serum untargeted metabolomic analysis to shed light on how the GM influences epilepsy. Our study contributes to a better understanding of how the GM functions in epilepsy via MGBA. These insights may pave the way for developing novel therapeutic approaches from a new perspective.

In this study, we demonstrated for the first time that the therapeutic dose of oral CPF increases susceptibility to epilepsy by directly altering the GM and its metabolites, without increasing CPF concentrations in the brain tissue. Moreover, in a more detailed analysis, we observed the diversity, composition and identical changes in some core genera in CPF-induced seizure susceptibility and lithium pilocarpine-induced epilepsy rat models. These changes may play a crucial role in epilepsy. Finally, we identified one of the metabolites of the tryptophan-indole metabolic process, IPA, as potentially significant in the pathological process of epilepsy. This study provides valuable insights into the intricate interplay between the GM and epilepsy, opening avenues for refined treatment strategies for epilepsy.

Numerous clinical studies have demonstrated that some antibiotics can induce epilepsy or increase susceptibility to epilepsy, including β-lactam, fluoroquinolone families, fourth-generation cephalosporins, and carbapenems (Wanleenuwat et al., 2020). In our study, we focused on CPF primarily because clinical observations have demonstrated that quinolone antibiotics can induce seizures (Slavich et al., 1989; Springuel, 1998; Kisa et al., 2005; Striano et al., 2007; Sutter et al., 2015). This phenomenon has also been corroborated in animal experiments, such as mice and zebrafish (Abdel-Zaher et al., 2012; Arafa et al., 2013; Sivarajan and Ramachandran, 2023). In fact, the research on the mechanism of antibiotic-induced epilepsy has been investigated for decades, which can be traced back to the 1990s. Here is a summary of some previous findings. First, some antibiotics have been found to inhibit GABA synthesis or block GABA receptors. β-lactams (i.e., penicillins, cephalosporins, and carbapenems) and quinolones all have been shown to act on these process (Segev et al., 1988; Hantson et al., 1999). Second, activation of excitatory transmission through activation of N-methyl-D-aspartate (NMDA) receptors, as may occur with carbapenems and quinolones (Akahane et al., 1993; Schmuck et al., 1998; Lode, 1999). Additionally, the structural similarities between certain antibiotics and epileptogenic compounds have been explored (De Sarro et al., 1989), with specific attention given to substituents at the C7 position that enhance binding to GABA receptors (Lode, 1999). Furthermore, interactions between older generation ASMs and certain antibiotics are known to influence the pharmacokinetics (Patsalos and Perucca, 2003). Coadministration of valproic acid and carbapenem will decrease the plasma valproic acid concentration by 58%, and increase the valproic acid clearance (Tobin et al., 2009). Finally, some antibiotics increase oxidative stress, potentially compromising the efficacy of epilepsy treatment (Ilgin et al., 2015; Wanleenuwat et al., 2020). However, many of these studies often induced acute and subacute seizures using large doses of CPF administered intraperitoneally, intravenously, or even intrathecally. Consequently, they may have overlooked the significant impact of CPF on the GM. Our study is the first to reveal that prolonged, clinically relevant oral administration of CPF promotes seizure susceptibility through the GM. Additionally, despite its lower BBB permeability than that of levofloxacin, CPF exhibits a higher proconvulsive potential, suggesting that its increased susceptibility to epilepsy may involve other pathways (Akahane et al., 1993). CPF is known to have a high concentration in the gastrointestinal tract (Zimmermann and Curtis, 2019), and induces gut dysbiosis (Stewardson et al., 2015; Zhu et al., 2020). Finally, considering the widespread use of CPF in the treatment of urinary tract and reproductive system infections (Thai et al., 2023), complete avoidance is challenging in certain clinical scenarios. Therefore, CPF is an appropriate instrument for studying the relationship between epilepsy and MGBA, and our study provides valuable insights for practical applications in a clinical setting.

Analysis of GM revealed a decrease in α-diversity, an estimate of species richness in CPF-14 d and Pilo group. This result is corresponding with previous animal and clinical studies (Gong et al., 2020; Oliveira et al., 2022). Decreased diversity and evenness have often been considered to reflect disease status. At the genus level, we found that some probiotic strains decreased. Lactobacillus can produce tryptophan and indole derivatives that regulate the immune process (Xia et al., 2021; Sun et al., 2022; Yan et al., 2022). In addition, it can modulate several neurotransmitters (Wu and Shah, 2017) and enhance the production of non-volatile acids and SCFAs (Dalile et al., 2019). Administration of probiotics, including Lactobacillus and Bifidobacterium, can reduce the seizure severity in the chronic epilepsy model induced by PTZ (Bagheri et al., 2019; Kilinc et al., 2021). Additionally, Butyricicoccus, Parabacteroides (Liao et al., 2024), and Ruminococcaceae (Peled et al., 2024) can produce SCFAs to promote health. Specifically, we found consistent changes in five genera across the two rat models: Akkermansia, Bacteroides, Marvinbryantia, Oscillibacter, and Ruminococcaceae_NK4A214_group. We observed similar findings in previous studies we observed similar findings on the GM and epilepsy (Russo, 2022; Yue et al., 2022). Among these, our particular focus was on Akkermansia because it ranked at the front position in the LEfSe analysis in the two rat models, which has been discussed in many previous studies. Bacteroides can produce SCFAs and are beneficial to the host; however, they can be opportunistic pathogens and are related to some pathological status (Shin et al., 2024). It could also be related to regulate the secretion of IL-6 and IL-17 in dendritic cells to associated with seizure severity (Xie et al., 2017). The functions of Marvinbryantia and Oscillibacter remain unclear, but they are often reported to be associated with health status or serve as protective factors (Fang et al., 2016; Xu et al., 2020; Crossland et al., 2023; Guo et al., 2023). This may be related to the production of SCFAs, which contribute to the maintenance of GM homeostasis. This suggests that these genera could be core contributors, play a substantial role in epileptogenesis, and serve as potential target genera for epilepsy treatment from the perspective of GM.

In our study, Akkermansia increased significantly and was ranked first in the LEfSe analysis of the two rat models. In fact, Akkermansia stands out as a known member of the GM and has been extensively investigated, particularly in the context of metabolic disorders and inflammatory processes (Cani et al., 2022). In studies of metabolic disorders, including obesity (Everard et al., 2014), diabetes (Plovier et al., 2017), and liver steatosis (Kim et al., 2020), Akkermansia has been found to be positively associated with improved metabolic health and better clinical outcomes. The same result was also observed in some inflammatory bowel diseases (Liu et al., 2021) and in response to cancer checkpoint immunotherapies (Santoni et al., 2018). Therefore, it is regarded as a paradigm for next-generation probiotics (Abuqwider et al., 2021; Cani et al., 2022). However, the “success story” of Akkermansia has not been without its share of controversies. In our study, the relative abundance of Akkermansia spanning the entire phylum significantly increased in CPF-14 d and Pilo groups. From the phylum to genus level, Verrucomicrobia, Verrucomicrobiae, Verrucomicrobiales, Akkermansiaceae, and Akkermansia displayed increased levels across all strata in the context of epilepsy. This observation is consistent with numerous studies on epilepsy (Huang et al., 2019, 2022; Arulsamy et al., 2020; Gong et al., 2021; Yue et al., 2022). Furthermore, Akkermansia increases have been noted in other neurological conditions, such as anxiety and depression induced by chronic stress (Li et al., 2019), multiple sclerosis (Cox et al., 2021), and Parkinson's disease (Zhao et al., 2021).

Akkermansia muciniphila, as the name suggests, is an exclusive mucin-degrading specialist present in the human intestine from early life (Derrien et al., 2008). A. muciniphila produces oligosaccharides, acetate, propionate, 1,2-propylene glycol, and ethanol via the breakdown of mucus and mucin (Derrien et al., 2004). Elevated levels of A. muciniphila are associated with reduced intestinal mucosal thickness, increased gut permeability, and inflammation. These factors may contribute to systemic inflammation and subsequent brain damage (Citraro et al., 2021). Previous findings suggest a potentially positive association with inflammatory status (Everard et al., 2013). The increase in Akkermansia in the lithium pilocarpine-induced epilepsy rat model suggests that inflammation and metabolic changes may play a role within the GM in the traditional model of chronic epilepsy. Overall, we believe that Akkermansia may play a complex role in the development of epilepsy, depending on the different populations and treatment choices. The relative abundance of Akkermansia may need to be maintained at a specific level to strike a balance between excess and deficiency.

Similar to the shared alterations observed in the GM, we conducted a comparative analysis of common changes in metabolites between the CPF and Pilo groups. Tryptophan-indole metabolism has emerged as a common phenomenon. In particular, one of the products, IPA, exhibited a consistent decreasing trend in both rat models. This change was more pronounced in the Pilo group. This discrepancy may be attributed to the fact that the lithium pilocarpine-induced epilepsy model manifests evident spontaneous seizures, whereas CPF primarily lowers the seizure threshold, representing a condition that is not strictly synonymous with an epilepsy model.

Tryptophan (TRP) metabolism pathways leading to serotonin, kynurenine, and indole derivatives are under the direct or indirect control of the GM (Agus et al., 2018). GM encodes enzymes that metabolize TRP to 5-HT, KYN, and indole metabolites, which affect TRP availability (O'Mahony et al., 2015). TRP is the only amino acid that contains an indole. In contrast to other TRP metabolites, indoles result exclusively from microbial metabolism (Kennedy et al., 2017). Indoles and their derivatives can be used as ligands to regulate inflammation and autoimmune responses in vivo, and play important roles in MGBA (Zhou et al., 2023b). In our studies, we have found some genera significantly related to this process (Figure 8A). These genera have been reported to be associated with metabolites involved in tryptophan metabolism, such as IPA, indoleacrylic acid (IAA), L-kynurenine, and L-tryptophan (Wikoff et al., 2009; Zelante et al., 2013; Roager and Licht, 2018; Su et al., 2022).

TRP metabolism has been explored in the context of epilepsy, with some studies focusing on kynurenic acid metabolism. We found that the concentration of kynurenic acid decreased significantly in the serum of the Pilo group, which corresponds to the findings of (Dey et al., 2022), who discovered that in a lithium-pilocarpine rat model, there was a reduction in kynurenic acid levels in the hippocampus and anterior temporal lobe. Furthermore, the exogenous application of kynurenic acid inhibits glutamatergic activity in slice preparations from rats with temporal lobe epilepsy. Promoting kynurenic acid production has been suggested as a strategy to mitigate spasms in animal models of infantile spasms (Mu et al., 2022a). In a pilot study of patients with epilepsy, tryptophan and kynurenine were shown to have the potential as diagnostic markers for epilepsy (Zhou et al., 2023a).

While evidence linking epilepsy to indole and its derivatives is currently limited, a growing body of research is shedding light on their relationship with the central nervous system (CNS), which serves as a neuroprotective factor. In 1980, Young et al. (1980) conducted an experiment on rats, indicating that IPA in the CSF is not derived from the CNS, but from bacterial metabolism in the gut. It has a high permeability to the BBB and does not exhibit a gradient. IPA is a small molecule of indoles that can passing through the lipid bilayer via free diffusion (Piñero-Fernandez et al., 2011). IPA was detected in the gut, serum, and brain of SPF mice compared to germ-free mice, thereby its production was interpreted to be fully mediated by the GM (Sarkar et al., 2016; Loh et al., 2024). The GM affects the diversity of indole compounds in serum, and GM induction is sufficient to introduce IPA into the host (Wikoff et al., 2009). Probiotics increase the level of IPA, reduce the inflammatory process, lower the HPA axis regulation factors, and reduce depressive-like behavior (Abildgaard et al., 2017). All of these showed a close correlation between the GM and IPA. In CNS studies, plasma IPA levels were significantly lower in patients with Huntington's disease (Rosas et al., 2015). Liu et al. (2020) found that IPA and 5-HT treatment significantly attenuated cognitive deficits in diabetic mice. In Alzheimer's disease (AD), IPA can protect against Aβ-induced neuronal death as a scavenger of free radicals (Chyan et al., 1999). Sun et al. (2022) found that the indoles derivates inhibited the activation of the NF-κB signal pathway as well as the formation of the NLRP3 inflammasome, reduced the release of inflammatory cytokines, and then reduced the neuroinflammation in AD transgenic mice. Xie et al. (2022) found that the administration of IPA attenuated the activity of neurotoxic reactive A1 astrocytes in a mouse model of ischemic stroke. And IPA protects neurons from ischemia-induced damage by reducing DNA damage and lipid peroxidation (Hwang et al., 2009). In recent years, it has become evident that indole derivates, including IAA and IPA, are acting as ligands of the aryl hydrocarbon receptor (AHR). AHR is a transcription factor extensively expressed in astrocytes and microglia (Gutiérrez-Vázquez and Quintana, 2018). Its activation leads to alterations in both the innate and adaptive immune responses, subsequently regulating intestinal epithelial function, gut barrier integrity, and GM composition (Dong et al., 2020; Scott et al., 2020; Li et al., 2021). Previous study on multiple sclerosis mouse model showed that AHR acts as a negative regulator of NF-κB activation (Rothhammer et al., 2016). In recent years, more and more researchers have confirmed that neuroimmunity and inflammation also plays a crucial role in epilepsy (Vezzani et al., 2011, 2015). The changes in GM and tryptophan-indole metabolism mentioned above may collectively contribute to the development of epilepsy (Figure 9).

In this study, we presented new evidence for the first time that CPF can directly alter GM, leading to increased epilepsy susceptibility in rats. The reversal of FMT on the susceptibility to epilepsy confirmed that GM and MGBA may be pivotal mechanism underlying the increased susceptibility to epilepsy induced by CPF. Changes in the genera Akkermansia, Bacteroides, Marvinbryantia, Oscillibacter, and Ruminococcaceae_NK4A214_group have been shown to be associated with epileptic activity in two epilepsy related animal models. In addition, GM-produced amino acid metabolites may be critical to the mechanisms of MGBA in relation to seizures and epilepsy. IPA, a product of tryptophan-indoles metabolism, may play a crucial role in this process from the perspective of neuroimmune regulation.

In future research on the mechanisms or clinical translation potentials referring the GM and epilepsy, a starting point could be the core GM genera or the tryptophan-indole metabolic pathway mentioned above. For instance, modulation of the relative abundance of core GM genera through probiotics or targeted drugs, or supplementation of tryptophan metabolites such as kynurenic acid and IPA, could be investigated. Further molecular mechanism studies are necessary to elucidate how GM influence seizure susceptibility or the development of epilepsy, and to identify molecular-level treatment targets.

The original contributions presented in the study are publicly available. This data can be found here: NCBI BioProject, accession PRJNA1056293, PRJNA1056516.

The animal study was approved by the Institutional Ethical Committee for Animal Welfare of the Seventh Affiliated Hospital of Sun Yat-sen University. The study was conducted in accordance with the local legislation and institutional requirements.

SZ: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Writing – original draft, Writing – review & editing. YL: Conceptualization, Data curation, Methodology, Supervision, Writing – review & editing. QZ: Data curation, Investigation, Methodology, Writing – review & editing. MY: Data curation, Methodology, Writing – review & editing. HLi: Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Writing – review & editing. RN: Formal analysis, Methodology, Writing – review & editing. HLa: Formal analysis, Supervision, Writing – review & editing. JW: Methodology, Writing – review & editing. XY: Conceptualization, Funding acquisition, Resources, Supervision, Validation, Visualization, Writing – review & editing. LZ: Conceptualization, Funding acquisition, Investigation, Resources, Supervision, Validation, Writing – review & editing.