Background: Previous studies have shown that the gut microbiota plays an important role in IBD. However there is no consensus on which bacteria are responsible for the disease. 16S gene profiling studies generate large amounts of information, but are confounded by genetic and environmental factors. Twin studies are instrumental in controlling these variables.
In this study we investigated the microbiota of twin pairs discordant for Crohn's disease (CD) and ulcerative colitis (UC) using 16S rRNA gene sequencing, with the aim of identifying taxa associated with disease.
Methods: Participants were recruited via the UK IBD Twin Registry. Stool samples were collected and frozen using standard methods. Participants who had received antibiotics within 3 months were excluded. Harvey Bradshaw Index and Simple Clinical Colitis Activity Index were recorded. Full medical history was available from the UK IBD Twin Registry.
Samples underwent 16S rRNA sequencing using the Illumina MiSeq platform and analysed using our data analysis pipeline. PERMANOVA was used to evaluate associations with clinical metadata, which included matching of twin pairs for analysis, and STAMP was used to identify taxonomic differences between groups.
Results: 20 twin pairs discordant for CD (5MZ:15DZ mean age 52 years) and 17 discordant for UC (6MZ:11DZ mean age 59.7 years) were recruited. 7 subjects with CD had active disease as did 4 with UC.
Gut microbiota from active CD patients had lower bacterial diversity compared to remission CD patients and healthy twins (Shannon diversity index, p<0.001 healthy vs active CD, active vs remission CD, 1-way ANOVA post-hoc = Tukey). Active UC patients also had lower bacterial diversity compared to remission UC patients and heathy twins (Shannon diversity index, p<0.01 healthy vs active UC, p<0.05 active vs remission). NMDS plots show more definitive clustering to phenotype in CD.
Active CD patients had a higher proportion of Clostridium hylemonae and Lactobacillus delbrueckii compared to healthy twins, and a lower proportion of Bacteroides uniformis, Bacteroides vulgatus, and Faecalibacterium prausnitzii (p<0.05). Active UC patients had a lower proportion of Alistipes spp. compared to healthy co-twins and UC patients in remission (p<0.05).
Conclusions: This study confirms previous findings showing decreased diversity in IBD patients and changes in some bacterial taxa. Our study is the first to show decreases in Alistipes spp. in active UC.