Coeliac disease (CD) is a prevalent (1%) small-bowel immune-mediated disease
involving both genetic and environmental components. Symptoms are often silent
and atypical, and associated with co-morbidity and long-term complications. The only
treatment is lifelong adherence to GFD, which is demanding and sometimes
unsuccessful. We currently lack the tools to non-invasively diagnose CD and the
medications to effectively treat it. Genetic predisposition is necessary but not alone
sufficient for disease development. Published, and our preliminary studies suggest
alterations in small-bowel microbiota and host transcriptome compared to controls;
however, a CD-specific microbial signature is still lacking.
If we could both use and characterize gastrointestinal host and microbiota data to
predict how GFD affects long-term clinical and histological outcome (mainly villous
atrophy), gluten-tolerance could potentially be increased through microbial
modulation. We hypothesize that host genetic variation contributes to shaping
gastrointestinal microbiota, which, in turn, modulate host gene expression through
epigenetic mechanisms potentially impacting pathogenesis. To this end, we will apply
state-of-the-art Machine Learning on multi-omics data from paired duodenum and
antrum biopsies of 75 patients to (a) classify symptoms and/or histology at diagnosis
(baseline), and (b) predict symptoms and/or histology after one year on GFD.
Understanding the (epi)genetic relationships between host and dysbiotic microbiota
will potentially herald disease prognostics, diagnostics and therapeutic interventions.
This prospective study is the first to integrate gastrointestinal host genotypes,
epigenomes and transcriptomes with microbiota (present, active and functional). We
will also evaluate the potential of stool-derived microbiome biomarkers as alternative
non-invasive diagnostic (for CD) and prognostic tools (for GFD outcome). Even if
causality would not be established, particular combinations of these multi-omics data
may still have diagnostic/prognostic value. This study will finally generate a
comprehensive and new host-microbe view on CD pathogenesis, which in its
extension could facilitate therapeutic targets.
