Systemic antibody responses against gut microbiota flagellins implicate shared and divergent immune reactivity in Crohn’s Disease and chronic fatigue syndrome

Background: Patients with Crohn’s disease (CD) and myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) exhibit elevated antibody responses against gut microbiota flagellins. However, flagellin-specific antibody repertoires and functional roles in the diseases remain incompletely understood. Bacterial flagellins can be categorized into three types depending on their interaction with toll-like receptor 5 (TLR5): (1) “stimulator” and (2) “silent” flagellins, binding TLR5 through a conserved N-terminal motif, with only stimulators activating TLR5 due to a specific C-terminal domain; (3) “evader” flagellins of pathogens, which circumvent TLR5 activation via mutated N-terminal TLR5 binding motifs. Here we studied the characteristics, epitope binding, and sequence (dis)similarity of anti-flagellin antibody responses in CD and ME/CFS.

Methods: Since conventional antibody profiling methods like enzyme-linked immunosorbent assays [ELISAs] do not allow for large-scale measurements of antibody repertoires, we leveraged phage-display immunoprecipitation sequencing (PhIP-Seq) to characterize 344,000 rationally selected peptide antigens in 256 patients with CD, 40 patients with ME/CFS and in two equally sized groups of age- and sex-matched healthy controls from population-based cohorts in the Netherlands and U.K., respectively. Different sequence alignment strategies were employed to compare flagellin peptide structures with observed antibody-bound flagellin peptide reactivity.

Results: Both patients with CD and ME/CFS exhibited elevated antibody responses against distinct regions of flagellin peptides compared to healthy individuals (P<0.001). N-terminal binding to Lachnospiraceae flagellins was comparable in both diseases, while C-terminal binding was more prevalent in CD. N-terminal antibody-bound flagellin sequences were similar across CD and ME/CFS, resembling ‘stimulator’ and ‘silent’ flagellins more than evaders. However, C-terminal antibody-bound flagellins showed higher resemblance to stimulator than to silent flagellins in CD, but not in ME/CFS. This group of antibody-bound flagellins was exclusively identified in a subset (10-20%) of patients with CD and characterized by its strong overrepresentation (exceeding 20-fold), underscoring its potential significance in distinguishing pathophysiologic subtypes of CD.

Conclusion: Antibody binding to the N-terminal domain of stimulator and silent flagellins may impact TLR5 activation in both CD and ME/CFS patients. Furthermore, elevated antibody binding to the C-terminal domain of stimulator flagellins in CD may explain pathophysiological differences between diseases. Our results highlight the diagnostic potential of these antibody responses and their impact on innate/adaptive immunity balance.