The lymphoid tissue of the intestinal mucosa contains 70%–80% of all immune cells in the body (69). The GM affects immune cells; for example, germ-free (GF) mice show immune abnormalities with a decreased population of T and B cells and a reduced cytokine production (70). Further, the GM appears as one of the most important factors for the maturation of microglial cells, as well as the astrocyte activation (71), which is age- and sex-dependent (72). The GM regulates innate immunity, adaptive immunity, and inflammatory mechanisms to modulate the development of epilepsy.

The intestinal mucosal barrier and blood–brain barrier (BBB) work together to prevent GM and its secretions from entering the brain. “Leaky gut” syndrome is characterized by increased intestinal permeability, which allows bacteria, toxic metabolites, and small molecules to translocate into the bloodstream (73). Under gut inflammation, bacteria can directly release factors into the systemic circulation, which activates peripheral immune cells, alters BBB integrity and thus transport rates, and can even induce “leaky brain” (74). Stress can increase intestinal mucosal permeability, and lipopolysaccharides and other cytokines in the lumen enter the blood circulation stimulating toll-like receptors, producing inflammatory cytokines that could increase BBB permeability and damage the brain (75, 76).

Astrocytes are the most abundant glial cells in the brain and have a variety of functions, including regulating the integrity of the BBB, the recycling of neurotransmitters, and participating in immune responses (77). Microglial cells are the resident macrophages of the CNS, mediating the innate immune response (78). Microglia with a larger, less ramified, amoeboid morphology can promote inflammation, neuronal activity regulation, phagocytic neuron clearance, and chronic seizures (79). Microglia and astrocytes are involved in the pathogenesis of epilepsy by releasing excess cytokines (67), and interact with each other: microglia can modulate astrocytes’ phenotype and function (80), while mouse microglia can regulate astrocytes’ behavior through, for example, VEGF-B, which promotes the pathogenic response and inflammatory response of astrocytes, and TGF-α, which promotes the opposite (81). Gut microbes metabolize dietary tryptophan into aryl hydrocarbon receptor agonists and interact with its receptor to control microglial activation and TGF-α and VEGF-B expression, thereby modulating astrocyte pathogenic activity (81, 82). Inflammatory cytokines and chemokines released by astrocytes enhance microglial activities, including migration, phagocytosis of apoptotic cells, and synaptic pruning (83). The interaction between astrocytes and microglia leads to increased pro-inflammatory cytokine production and BBB permeability, which results in the infiltration of peripheral blood immune cells and cytokines into the CNS, and subsequent chronic neuroinflammation (84). GF and antibiotic-treated (Abx-treated) animals have also altered microglial morphology and defects in maturation, activation, and differentiation, resulting in an inadequate immune response to a variety of pathogens, which could be repaired after GM recolonization, which suggests that intestinal microbial diversity is critical for microglial and CNS function (85). In addition to glial cells residing in the CNS, peripheral immune cells, such as T cells and monocytes invading the brain tissue, are also involved in epilepsy’s morbidity. Monocytes can differentiate into macrophages and invade the brain, where they differentiate into “microglia-like cells” and contribute to epilepsy (86). As the organ with the largest population of immune cells, the gut is likely to play a role in this process, but the exact mechanism requires further investigation.

The GM can induce epilepsy through the innate immune pathway. BBB permeability increases throughout the life of GF mice related to the decreased expression of occludin and claudin-5 proteins in the endothelium (87). GM dysbiosis decreases claudin production and increases the permeability of the intestinal lining, leading to the escape of microorganisms, metabolites, and toxins from the intestinal lumen (88). GM dysbiosis also reduces SCFAs, which increases BBB permeability and promotes neuroinflammation (89). If these two barriers are broken, the immune cells and factors released by the microbiota enter the brain and induce seizures. Peptidoglycan (PGN) is a component of the bacterial cell wall that mainly exists in the human intestinal tract (90). PGN, as a driver of chronic encephalitis, has also been detected in brain microglia (90). Chronic inflammation, such as sclerosis-related temporal lobe epilepsy, is also a cause of epilepsy (91). Therefore, we conclude that PGN may translocate from the gut to the CNS by promoting gut leakage and brain leakage, leading to chronic inflammation and contributing to the occurrence of epilepsy.

The GM also contributes to epilepsy generation by inducing adaptive immunity. The GM can induce immune cells to produce cytokines that enter the brain through the intestinal mucosa and BBB and activate brain immune cells to participate in the immune response. T helper cell 17 (Th 17) cells are a proinflammatory CD4+ T cell subtype and key components of adaptive immunity (92). IL-17 is a cytokine produced by Th 17 cells, which can be modulated by specific GM phyla, such as Bacteroidetes (92, 93). As recently shown, both in the CSF and in the peripheral blood of patients with epilepsy, IL-17 levels were higher than in controls, and highly correlated with the frequency and severity of seizures (94–97). Therefore, the GM can influence the occurrence of epilepsy by mediating IL-17. GM metabolites, such as SCFAs, can affect the synthesis and secretion of immunoglobulins by regulating B lymphocyte differentiation (98, 99). The absence of commensal microbiota downregulates IgA and IgG1, and upregulates IgE, which leads to increased susceptibility to diseases (100, 101).

Therefore, GM can induce an immune response through the gut–brain axis, which leads to epileptogenesis. However, only a few studies have directly focused on the relationship between the gut, immune responses, and epilepsy, and many issues remain to be explored.

Neurotransmitter imbalance is closely related to epilepsy. Neurotransmitter imbalance exists in epileptic foci, such as GABA with hypoactivity and glutamate with hyperactivity, dopamine and norepinephrine (NE) with hyperactivity, and serotonin with hypoactivity (108). In the gastrointestinal tract, neurotransmitters can be secreted directly by the GM or produced by gastrointestinal cells under the stimulation of GM metabolites. Different GM can produce different neurotransmitters (Enterococcus spp., Streptococcus spp., and Escherichia spp. produce serotonin; Lactobacillus spp. and Bifidobacterium spp. produce GABA, Escherichia spp. and Bacillus spp. produce NE and dopamine). The various neurotransmitters produced by the GM can pass through the intestinal mucosa but rarely through the BBB, with the exception of GABA (109). In hippocampal injury, or epileptic status, GABA produced by GM can lead to an imbalance between the GABA and glutamate systems, causing seizures. Sun et al. showed that the relative abundance of the genera Coprococcus, Ruminococcus, and Turicibacter was positively correlated with glutamate and glutamine levels (110). The GM can affect the glutamine–glutamate–GABA cycle, produce neurotransmitters, and mediate the expression of GABA and NMDA receptors in specific brain regions (hippocampus, amygdala, and locus coeruleus) (111). A. mucinophilia and Parabacteroides colonization could alter the level of amino acids in the serum and gut lumen to modulate the levels of seizure-associated neurotransmitters, such as GABA and glutamate, in the hippocampus, thus providing protective anti-seizure effects in mice (20, 112). Enterochromaffin cells (ECs) produce approximately 90% of 5-hydroxytryptamine (5-HT) (113). In GF mice, certain intestinal microbiota, such as spore-forming clostridia taxa, can promote 5-HT biosynthesis in the gut by upregulating colonic tryptophan hydroxylase 1, a rate-limiting enzyme for 5-HT production (114, 115). Previous studies have shown that patients with temporal lobe epilepsy have a 5-HT deficiency. A drug combination that increases 5-HT, such as selective serotonin reuptake inhibitors, may improve seizure control in patients with epilepsy (116). The 5-HT decrease induced by reserpine appears to increase susceptibility to minimal electroshock seizures in rats (117). However, changes in intestinal 5-HT levels could not directly affect the brain, as 5-HT cannot cross the BBB (118). Chemotherapeutic drugs often cause nausea and vomiting, caused by the release of large amounts of 5-HT in the intestine and the subsequent activation of vagal afferents (119). 5-HT released by ECs may have a potential impact on brain–gut axis signal transduction by regulating intestinal vagal afferent activity (120) and inflammatory responses (121). Alterations in 5-HT signaling are associated with irritable bowel syndrome (122). Therefore, we speculate that a change in 5-HT levels in the intestine may be related to epilepsy, but there is no evidence to support this. N-acetyl aspartic acid (NAA) levels are reduced in patients with epilepsy, and in the epileptic suckling pig model, Austin et al. found that low NAA levels were associated with fecal Ruminococcus, and this process may be mediated by serum cortisol (123). NE has a double effect on epilepsy onset depending on its amount, NE at low doses has pro-epileptic effects, while high doses could inhibit epilepsy (124). Dopamine, serotonin, and acetylcholine are closely correlated with epilepsy and could indirectly affect brain function through the enteric nervous system, the vagus nervous system, and by regulating the expression of peripheral receptors (125).

SCFAs, including acetate, propionate, and butyrate, can be produced by some gut bacteria (mainly Bacteroides and Firmicutes) through the fermentation of insoluble dietary fibers (126). SCFAs play an essential role in microglial maturation, the gut–brain nervous system, BBB permeability, and stress responses through direct or indirect pathways, all closely related to epilepsy (127). As mentioned earlier, SCFAs were reduced in WAG/Rij rats, and butyrate had an anti-absence seizure effect (16). The protective effects and mechanisms of different SCFAs in epilepsy were further studied in a PTZ-induced epileptic mouse model (8, 128, 129). Sodium butyrate may exhibit antiepileptic effects in PTZ-induced epileptic mice by alleviating intestinal inflammation and oxidative stress (8). Butyrate also ameliorates mitochondrial dysfunction and protects brain tissue from oxidative stress and neuronal apoptosis through the Keap/Nrf2/HO-1 pathway, thereby increasing seizure threshold and reducing seizure intensity (129). Propionate treatment may alleviate seizure intensity and prolong the incubation period of seizures by reducing mitochondrial damage, hippocampal apoptosis, and neurological deficits (128). These studies show that SCFAs are reduced in epilepsy models and their protective effect on epilepsy through different mechanisms.

Diet, especially KD, could regulate the occurrence of epilepsy by shaping GM as discussed in detail in Section 3. KD is a high-fat, low-carbohydrate, and adequate protein diet used since 1921 in patients with refractory epilepsy (140). KD also has positive effects on other neurological diseases, such as multiple sclerosis, Parkinson’s disease, Alzheimer’s disease, and migraine (141–143). The classic ratio of fat to protein and carbohydrates in KD is 4:1 (18), which triggers a metabolic pattern shift from glucose metabolism toward the metabolism of fatty acids (53). The classic KD could relieve epilepsy by multiple pathways, including modulation of neurotransmitters, brain energy metabolism, oxidative stress, ion channels, and GM (53, 54, 144).

In addition to the classical form, several modified KD diets have arisen, including the modified Atkins diet, medium-chain triglyceride diet, low-glycemic index treatment, and modified Mediterranean ketogenic diet (MMKD), with various composition ratios of fat, protein, and carbohydrates (145). In medium-chain triglyceride diet, medium-chain triglycerides are used instead of long-chain triglycerides (145). MMKD is characterized by olive oil as a source of monounsaturated fatty acids (146). Olive oil contains antioxidant molecules such as monounsaturated fatty acids and polyphenols, which have beneficial effects on inflammation, cardiovascular disease, and oxidative status of the body (147). MMKD can regulate the GM of patients with mild cognitive impairment, especially fungal flora (146). A 12-month KD based on olive oil could alleviate symptoms in 83.1% of patients with refractory epilepsy, comparable to the effect of a traditional KD treatment (147). However, there is no comparative study between MMKD and traditional KD on the efficacy and side effects on refractory epilepsy. At present, most KDs are individualized modified KDs balanced between ketogenic effect and palatability.

In KD, saturated fat has always been the prominent fat used (148); however, animal and human studies have demonstrated the anti-epileptic effects of polyunsaturated fatty acids (PUFAs), especially omega-3(n-3) PUFAs (149). Dietary n-3 PUFAs are found in flaxseed, nuts, marine fish, and marine mammals. N-6 PUFAs are mainly derived from animal products and vegetable oils and constitute the majority of PUFAs in the modern Western diet (149). Docosahexaenoic acid (DHA, 22:6n-3), the primary n-3 PUFA in the brain, participates in the regulation of neural function through a variety of pathways, such as interaction with ion channels and neurotransmitter release (150, 151). A case–control study indicated a lower serum omega 3/omega 6 ratio in children with epilepsy than in healthy children (152). Both in vitro and in vivo studies demonstrated that a diet rich in n-3 fatty acids is beneficial for epilepsy control, but the results of clinical studies are somewhat contradictory (149). A meta-analysis of seven clinical trial studies in 2021 indicated that omega-3 supplementation significantly reduced seizure frequency and was more effective in adults than in children (153). Therefore, adjustment of dietary n-3/n-6 levels could be associated with seizure control. A recent study indicated that a high [fat]:[carbohydrate + protein] ratio is not indispensable for the treatment of epilepsy (154). A new combined diet with low [fat]:[proteins + carbohydrates] ratio, including medium-chain triglycerides, PUFAs, low glycemic index carbohydrates, and a high branched-chain amino acids/aromatic amino acids ratio also reduces excitatory drive and protects against seizures in rodent models (154). Although only tested on animals, it is a promising diet with fewer side effects.

Dietary intervention is an effective and perspective way to control epilepsy, and further research on the bacterial–gut–brain axis would contribute to develop the more effective dietotherapy.

Probiotics are living microorganisms that, at the appropriate dose, are beneficial to the host health (155). Most common probiotics include Bifidobacterium and Lactobacillus (156). In 2020, the International Scientific Association for Probiotics and Prebiotics defined prebiotics as a “substrate selectively utilized by host microorganisms conferring a health benefit” (157). Synbiotics has been defined as “a mixture comprising live microorganisms and substrate(s) selectively utilized by host microorganisms that confers a health benefit on the host” consisting of two subsets: synergistic synbiotic (the prebiotic is selectively utilized by the co-administered live microorganisms) and complementary synbiotic (each component works independently) (157). In a prospective study, the frequency of seizures was reduced by ≥50% in 28.9% of DRE patients treated with a probiotic mixture for 4 months, and 76.9% of these improved patients maintained a lower seizure frequency 4 months after discontinuation (158). This study indicated that adjuvant probiotics reduced the frequency of seizures and could be used as a complementary treatment to antiepileptic therapy (158). In the PTZ-induced chemical kindling mouse model, the probiotic supplementation group did not show full kindling, and GABA increased in mouse brain tissue, which indicated that probiotic supplementation could substantially reduce seizure severity (159).When treating PTZ-induced seizures in mice with KD, synbiotics or Lactobacillus fermentum MSK 408 could reduce the side effects of KD without disturbing its antiepileptic effects (21, 160). Both KD and MSK 408 increase GABA metabolism by regulating the GM (160). Several SCFAs, such as propionate and butyrate, have antiepileptic effects. After a month of intervention with the classic KD, the total SCFAs were significantly reduced, especially acetate, propionate, and butyrate, which may be due to a reduction in the intake of fermentable carbohydrates or a reduction in fermenting bacteria by KD (161). Synbiotics can lead to GM enrichment associated with SCFAs (21). MSK 408 could influence SCFAs and restore serum lipid profile and tight junction protein mRNA expression in the gut and brain independently by modulating the GM (160).

These studies are preliminary observations of supplementary probiotics in the treatment of DRE, and further theoretical validation and mechanism exploration in larger placebo-controlled trials and more rigorous animal experiments should be conducted. Probiotics have the potential to be a complementary treatment for refractory epilepsy and can be used in combination with KD therapy to reduce side effects.

Antibiotics can have either good or negative effects on the treatment of inflammatory bowel disease, irritable bowel syndrome, hepatic encephalopathy, and other diseases (162). Despite the little evidence on the link between the use of antibiotics and seizures, a retrospective study of six refractory epileptic patients treated with antibiotics showed that certain antibiotics could reduce the frequency of seizures in the short term (163). They hypothesized that antibiotics might induce seizure freedom or decrease seizure frequency by interfering with the intestinal flora and the gut–brain axis (163). However, certain antibiotics can also induce epilepsy; for example, lactam antibiotics, including penicillin, cephalosporins, and carbapenems, are most likely to cause seizures (164). Unsubstituted penicillin, like fourth-generation cephalosporins, imipenem, and ciprofloxacin in combination with renal dysfunction, brain lesions, and epilepsy could lead to an increased risk of symptomatic seizures. Therefore, serum levels and EEG should be closely monitored when using antibiotics in these patients (165). Antibiotic application exerts short- or long-term effects on GM composition in both humans and animals (162). Although some antibiotics can disrupt the balance of intestinal microorganisms and cause diseases, others increase the abundance of beneficial microorganisms and play a positive role in the GM (162). Different antibiotics lead to different patterns of GM alteration; for example, macrolides induce the reduction of Actinobacteria (mainly Bifidobacteria) (166, 167), oral vancomycin reduces Firmicutes and increases Proteobacteria (168), whereas penicillin only has a weak effect on the human microbiota (168). The extent of amoxicillin-induced epilepsy is not parallel to GM changes, which is contradictory to the hypothesis that GM acts as a bridge in antibiotic-induced epilepsy. However, the influence of antibiotics on GM is also related to the initial GM composition (169) and habits of the host (170). In the future, multi-center cooperation is needed to further clarify the specific effects and mechanisms of various antibiotics on epilepsy.

The GM contains a rich variety of drug-metabolizing enzymes that influence their pharmacology, resulting in interpersonal differences in drug efficacy and toxicity (171). For example, clonazepam is an anticonvulsant and anti-anxiety drug reduced and metabolized by the GM, resulting in drug toxicity (172). Non-antibiotic drugs alter the GM to some extent. In a large study involving the effects of 1,197 non-antibiotic drugs on 40 GM, 24% of the drugs with human targets inhibited the growth of ≥1 strain in vitro (173). Antiepileptic drugs, such as carbamazepine, valproic acid, and lamotrigine affect GM composition (174, 175). Valproic acid treatment during pregnancy in mice resulted in altered fecal microbiota (176, 177), with increased Firmicutes and decreased Bacteroidetes (178), which may be associated with ASD-like behavior in offspring (176, 177). Lamotrigine might decrease the growth of E. coli by inhibiting bacterial ribosome biogenesis (179). Further research on the relationship between antiepileptic drugs and gut microorganisms will help to develop new antiepileptic drugs based on the principle of GM regulation. Adjusting GM composition could alter the metabolic process of antiepileptic drugs to improve their efficacy and reduce side effects.

FMT has been proven to be an extremely effective treatment for recurrent or refractory Clostridium difficile infections (176, 177, 180, 181). Moreover, FMT has been extensively studied as a potential treatment for GM-related diseases. Its effectiveness has been demonstrated in a range of diseases, such as ulcerative colitis (182), hepatic encephalopathy (183), irritable bowel syndrome (184), obesity (185), and even neurological disorders (186, 187). Recent studies have revealed a correlation between epilepsy and GM; thus, the value of FMT administration in patients with epilepsy has been further investigated. FMT has been shown to prevent the recurrence of epilepsy after discontinuation of antiepileptic drugs in patients with Crohn’s disease and a long history of epilepsy (188). In a rat model of epilepsy with basolateral amygdala kindling following chronic stress, GM transplantation from chronically stressed rats to sham-stressed rats accelerated the kindling epileptogenesis process, whereas transplantation from sham-stressed group to stressed rats reduced the pro-epileptic effects of stress (42). Olson et al. observed an increased seizure threshold in a GF temporal lobe epilepsy mouse model after transplantation with microbiota from KD-treated mice or long-term administration of microbiota associated with KD treatment (20). However, there are still some challenges to the FMT process. Currently, FMT is performed by placing small amounts of liquefied or filtered feces directly into the colon or through a feeding tube, enema, or capsule (189). Therefore, FMT may result in the transmission of bacteria, viruses, or diseases that cannot be detected by screening (189, 190). Furthermore, FMT may disrupt the baseline microbiota diversity, resulting in the breakdown of colonization resistance to a broad spectrum of harmful microorganisms (190). Therefore, more long-term follow-up studies are needed to determine the efficacy and safety of FMT in patients with epilepsy before large-scale clinical application.