A new study revealed that systemic lupus erythematosus and inflammatory bowel disease may share common gut microbiome characteristics that distinguish them from other autoimmune disorders.

In the study, published in Annals of the Rheumatic Diseases, researchers used machine learning and metagenomic analysis to identify shared microbial signatures and functional alterations across multiple autoimmune diseases.

They analyzed metagenomic datasets from seven studies focusing on various autoimmune disorders, including inflammatory bowel disease (IBD), Graves' disease, systemic lupus erythematosus, ankylosing spondylitis, and myasthenia gravis. The investigators also included four colorectal cancer (CRC) datasets for comparison. The study examined 78 fecal samples, with 32 samples from 14 pretreatment outpatients with lupus and 46 samples from 22 healthy controls. An additional multiple sclerosis (MS) dataset included samples from 25 cases and 26 controls.

The study employed various bioinformatics tools, including MetaPhlAn3 for taxonomic profiling and HUMANN3 for functional profiling. Genomic DNA was extracted using QIAGEN DNeasy 96 PowerSoil Pro Kit, and libraries were prepared using Nextera XT DNA Library Prep Kit. Sequencing was carried out using an Illumina NovaSeq6000 system. Samples with fewer than 10^7 reads were removed from analyses.

Statistical analysis involved generalized linear regression using Maaslin2, with the Benjamini-Hochberg method used to adjust for multiple testing. Random Forest Classifier underwent a five-times repeated 10-fold cross-validation.

Among the key findings were:

• Random forest models trained on microbial compositions from patients with IBD predicted lupus in other cohorts with area under the receiver operating characteristic curve (AUROC) values between 0.60 and 0.74.
• Models trained on lupus microbiome compositions predicted IBD with AUROCs between 0.70 and 0.90.
• Common microbial species were identified in both IBD and lupus cohorts, including Streptococcus oralis, Gemella haemolysans, and Clostridium innocuum.
• Healthy controls exhibited higher abundances of certain species, including Anaerostipes hadrus and Fusicatenibacter saccharivorans.
• Functional analysis revealed shared alterations in carbohydrate-active enzymes (CAZymes) and short-chain fatty acid (SCFA) production pathways between patients with lupus and IBD.
• Host-microbiome protein-protein interactions (PPIs) identified common pathways involved in both diseases, including glucocorticoid receptor signaling and interleukin (IL)-12 signaling.

Taxonomic analysis revealed several species more abundant in patients with IBD and lupus compared with controls. Peptostreptococcus stomatis, Parvimonas micra, Gemella morbillorum, Fusobacterium nucleatum, Solobacterium moorei, and Hungatella hathewayi were predominantly abundant in patients with CRC. In contrast, Eubacterium sp CAG_38, Gemmiger formicilis, C. leptum, and Asaccharobacter celatus were more abundant in healthy controls compared with patients who had IBD and lupus.

Functional analysis identified key protein families (PFAMs) more abundant in healthy controls, including 'PF00404; Dockerin type I domain', 'PF12891; Glycoside hydrolase family 44', and 'PF08672; Anaphase promoting complex subunit 2'. CAZymes enriched in multiple autoimmune diseases included lipopolysaccharide N-acetylglucosaminyltransferase (GT9) and peptidoglycan hydrolase (GH73).

The study found significant differences in SCFA-producing microbial enzymes between healthy controls and patients with autoimmune diseases. Of 66 enzymes related to SCFA production, 33 were common between lupus and IBD, showing higher levels in patient groups. Conversely, 18 enzymes, with 8 common between lupus and IBD, were more prevalent in healthy controls.

Host-microbiome PPI analysis identified interactions involving the glucocorticoid receptor (NR3C1) as significantly associated with both IBD and lupus. There were at least three times more human targets identified in the IBD and lupus cohorts compared with any of the other diseases. Enriched pathways in both lupus and IBD included glucocorticoid receptor signaling, IL-12 signaling, IL-13 signaling, and PI3K/AKT signaling.

Experimental validation through co-immunoprecipitation assays confirmed binding between NR3C1 and predicted bacterial interactors, including proteins involved in vitamin B6 metabolism and oxidative stress response. Proteins from the UniRef50_P37527 (PdxS) cluster showed clear co-immunoprecipitation with FLAG-NR3C1.

Cross-cohort prediction performance varied, with models trained on CRC datasets demonstrating high predictive accuracy primarily within CRC datasets. Intrastudy cross-validation revealed high performance across all datasets, with average AUROC scores ranging from 0.69 to 0.96.

The findings suggested shared microbial mechanisms between lupus and IBD, despite their clinical differences. The study highlighted potential therapeutic implications, including the importance of dietary fiber in maintaining gut health and the role of SCFA-producing bacteria in modulating immune responses.

One study author received salary, consulting fees, honoraria, or research funds from Eligo Biosciences, Enterome, Novartis, Roche, Genentech, Bristol-Meyers Squibb, Sanofi, and AbbVie, and holds a patent on the use of microbiota manipulations to treat immune-mediated diseases. And another is both a salaried employee and a shareholder of Moderna.