Gut microbiota‑driven indole‑3‑propionic acid protects against sepsis‑associated encephalopathy via inhibiting BNIP3L ubiquitination and restoring BNIP3L‑dependent mitophagy as a potential molecular mechanism

Abstract

Sepsis-associated encephalopathy (SAE) is a severe refractory neurological complication occurring in up to 70% sepsis patients with no targeted therapies. Gut microbiota-derived indole-3-propionic acid (IPA), a tryptophan metabolite, exerts neuroprotective effects, but its therapeutic mechanism against SAE remains unclear. BNIP3L-linked mitophagy and mitochondrial quality control are disrupted in septic neurons, while the potential correlative connection between IPA abundance and BNIP3L abundance remains undefined. This study aimed to explore the gut-brain axis mechanism and molecular target of IPA in treating SAE. We established cecal ligation and puncture (CLP) SAE mouse models with fecal microbiota transplantation (FMT) and exogenous IPA intervention; performed in vitro LPS-stimulated HT22 hippocampal neuronal experiments; enrolled a prospective clinical cohort of 111 ICU sepsis patients. 16S rRNA sequencing, untargeted/targeted metabolomics, fiber photometry Ca 2+ recording, transmission electron microscopy, Western blot, co-immunoprecipitation and linkage-specific ubiquitin ELISA were adopted. Statistical analyses included one-way ANOVA, Spearman correlation, multivariate logistic regression and ROC curve analysis. Sepsis induced gut dysbiosis, reduced tryptophan-metabolizing bacteria and depleted endogenous IPA; FMT restored microbial homeostasis and IPA, improving mouse survival, cognitive function and reducing neuronal apoptosis. Clinically, plasma IPA was significantly lower in SAE patients, positively correlated with GCS scores, and exhibited excellent diagnostic value (AUC = 0.891). In CLP mice, IPA supplementation was accompanied by lower systemic inflammatory cytokine concentrations, partial normalization of aberrant EEG power spectra, attenuated excessive Ca 2+ activity in hippocampal CA1 neurons, reduced neuronal loss, improved mitochondrial ultrastructural morphology, and parallel recovery of tissue ATP levels and antioxidant capacity. In HT22 cells, LPS exposure coincided with reduced BNIP3 levels, collapsed mitochondrial membrane potential, suppressed mitophagy and elevated neuronal apoptosis; IPA treatment correlated with parallel improvements in these cellular indicators. Mechanistically, global hippocampal total K48-linked polyubiquitin abundance exhibited an inverse correlation with circulating IPA concentrations, a trend that indirectly suggests potential shifts in BNIP3 ubiquitination profiles during sepsis, though protein-specific modification could not be confirmed herein. Circulating gut-microbiota-derived IPA concentrations showed inverse covariation with total hippocampal K48-linked polyubiquitin abundance, concurrent with parallel shifts in mitophagy-related molecular markers and less severe neuronal injury across clinical and preclinical SAE cohorts; direct evidence for BNIP3L-specific K48 ubiquitination modification was absent in this work in both clinical and preclinical SAE cohorts. IPA shows reliable diagnostic performance for SAE. Collectively, our multi-layer observational data collectively reveal a correlative gut metabolite–BNIP3L–mitochondrial homeostasis signature, which only provides preliminary associative evidence to guide future translational research targeting septic cerebral injury on sepsis-induced cerebral injury. Chinese Clinical Trials Registry, registration number ChiCTR2200067084, prospectively registered before participant enrollment.