The Gut Feeling Behind Autoimmune Encephalitis

Abstract

The Gut Microbiome Associated with LGI1-Antibody Encephalitis

Gilbert E, Binks S, Damato V, Uy C, Colmenero P, Kelly M, Khalil MI, O'Brien M, Claesson MJ, Cryan JF, Delanty N, Irani SR, Cavalleri GL. Epilepsia. 2025;66:4411–4424.

Objective: Autoimmune encephalitis is a cause of brain inflammation characterized by auto-antibodies, which target cell surface neuronal proteins and lead to neuronal dysfunction. The most common form is associated with auto-antibodies to leucine-rich glioma-inactivated 1 (LGI1) protein, the presentation of which includes frequent focal seizures. The exact cause of these auto-antibodies remains unknown, but established predispositions include overrepresented human leukocyte antigen (HLA) alleles. Yet, these HLA alleles are themselves common in the healthy ancestry-matched population. One potential etiological hypothesis is that an environmental trigger, such as the gut microbiome, interacts with a genetically predisposed individual. Methods: To investigate this, we studied 42 patients with LGI1-antibody encephalitis (LGI1-Ab-E) and 27 familial/environmentally matched controls and performed metagenomic shotgun sequencing, to describe the compositional and functional differences in the gut microbiome. Results: We observed that LGI1-Ab-E gut microbiomes exhibited a significant reduction in the ratio of Firmicutes (or Bacillota) and Bacteroidetes phyla, which is associated with the dosage of HLA susceptibility allele count in patients with LGI1-Ab-E. Furthermore, we identified differences in functional gene profiles in the gut microbiome that led to a reduction of neuroinflammatory protective short-chain fatty acids (SCFAs) in LGI1-Ab-E patients. Significance: Taken together, our results suggest that a compositional shift in the gut microbiome of LGI1-Ab-E associates with a neuroinflammatory state, possibly through the reduction of SCFA production. Our study highlights the potential of the gut microbiome to explain some of the complex condition and unravel etiological questions. Validation studies with greater sample sizes are recommended.

Commentary

Autoimmune encephalitis (AE) associated with leucine rich glioma inactivated 1 (LGI1) antibodies is one of the most common forms of AE that can lead to acute symptomatic seizures. ¹ The trigger for this condition that frequently presents in older people remains unknown, with the only clue for underlying mechanism being a high expression of a specific human leukocyte antigen (HLA) haplotype (HLA-DRB1*07:01 allele), which is also common in the general population. ² Why only a proportion of patients with this allele then develop this autoimmune condition has yet to be elucidated. The gut-brain axis has gained traction as a mechanism of disease which may be common along with a spectrum of multiple neurological conditions, given the broad impact of gut microbes on systemic inflammation through the release of small molecules including cytokines, and the production of neuromodulatory metabolites including short-chain fatty acids (SCFA), among other mechanisms. ³ It is therefore tempting to posit that a specific gut microbiome profile might trigger an individual, who may be already genetically prone, to develop AE. If that were to be the case in LGI1-antibody mediated AE (LGI1-ab-AE), implications are broad—one may extend this to other forms of AE, and new potentially complementary interventions to the existing standard immunotherapy may be tested in the clinic. But the question is how to effectively test this hypothesis to generate data that may offer new targets for therapeutic interventions.

In this study, Gilbert et al performed metagenomic shotgun sequencing of the fecal microbiome of 42 patients with LGI1-ab-AE, at various stages of disease (25 chronic, 14 acute), and compared them to their household healthy controls, the traditional form of control in microbiome studies given the impact of geographic region, genetic background and lifestyle (eg, diet), which tend to be shared within a household, on gut microbiome composition. ⁴ A high proportion (88%) of the LGI1-ab-AE cohort harbored the HLA allele found to be associated with this disease, compared to 41% of the control cohort. They found that patients with LGI-ab-AE had lower alpha diversity than their controls, meaning that their gut microbiome hosted fewer types of gut microbes with a lower range of species. They did not find a difference in the types of gut microbes in patients with LGI-Ab-AE compared to their controls, as measured by beta diversity and basic measures of taxonomic abundance comparisons, but did find differential abundance on linear discriminant analysis in the abundance of certain gut microbes, including a depletion of Bifidobacterium longum, a gut microbe also found to be depleted in children with drug resistant epilepsy. ⁵ They then turned to comparing the ratio of Firmicutes to Bacteroides, 2 phyla accounting for the majority of the gut microbiota profile. Low Firmicutes to Bacteroides ratio have been associated with obesity and inflammatory bowel disease, so the authors posited that this marker may be associated with systemic inflammation in LGI-Ab-AE and indeed found a significant reduction in this ratio compared to controls. They then developed a multivariate model accounting for HLA-DRB1*07:01 allele load in the correlation between Firmicutes and Bacteroides ratio with LGI1-ab-AE, and found that the allele was the only remaining significant variable in the model. Covariate analyses accounting for steroid and antiseizure medication (ASM) use were performed and revealed an impact of steroid use on gut microbiome composition, but not of ASMs. Lastly, functional pathway analysis was undertaken, to determine which functions associated with gut microbes were differentially associated with LGI1-Ab-AE compared to controls, and while no functional pathway survived a multiple comparison analysis, certain pathways associated with SCFA were less expressed in the LGI1-Ab-AE group in uncorrected analyses.

Overall, only mild differences were found between disease and control groups in this study. The authors ascribe the subtle findings to a small sample size and state that gut microbiome composition may not be largely shifted in LGI1-ab-AE. But is this a fair conclusion? In a rare disease, it may be challenging to perform gut microbiome analysis in which a large number of comparisons are made and a large sample size is typically needed to account for the methodological process. But when an already small sample is heterogeneous, with early and chronic cases lumped together in an analysis, it may not be surprising to find that no large differences emerged. A prior study only enrolling immunotherapy naive newly diagnosed LGI1-Ab-AE with a smaller sample size (n = 15) similarly observed low alpha diversity in LGI1-Ab-AE but also found a significant difference in beta diversity between LGI1-Ab-AE and healthy controls, albeit not household matched but matched for age, sex, and body mass index. ⁶ They also found a significant depletion in a group of gut microbes associated with SCFA production.

The devil lies in the details even more so in a small sample size, where variations in clinical variables may introduce noise and lower the chances of finding signal that could lend significant insights into disease pathogenesis. In a recent study, ⁷ the gut microbiome of a small population of patients with new-onset refractory status epilepticus (NORSE) was compared to status epilepticus of known cause and chronic epilepsy at various timepoints throughout the illness—significant differences were found despite small sample sizes (12-17 per group), and importantly, the gut microbiome shifted in composition throughout illness in patients with NORSE and status epilepticus, indicating that gut microbiome is dynamic and is likely affected by a number of variables also likely to change over time in patients with AE, including immunotherapy, ongoing seizures, and comorbid medical issues.

Thankfully, large multicenter efforts are underway in North America and Europe to follow patients with AE and other rare neuroinflammatory seizure disorders over time.^8–10 Harnessing such registries in biorepository efforts including biospecimens for microbiome analysis (fecal, saliva) will be crucial to shed light on the potential role of gut dysbiosis in inciting and promoting neuroinflammation in these rare but highly morbid conditions with unique opportunities for disease-targeted interventions.