Search in the Abstract Database

Search Abstracts 2016

* = Presenting author

P019 Informing ulcerative colitis pathophysiology and outcomes through metabolic pofiling

S. Fong*1, C. Le Roy2, J. Sanderson1, S. Claus2

1Guy’s and St Thomas’ NHS Foundation Trust, IBD Centre, London, United Kingdom, 2University of Reading, Department of Food and Nutritional Sciences, Reading, United Kingdom


Ulcerative colitis (UC) is a chronic gastrointestinal disorder involving a complex interplay between host genetics, mucosal immunology, gut microbiome and the environment. A complete understanding of aetiology and biomarkers predicting outcome are lacking. We used high-resolution magic angle spinning nuclear magnetic resonance (HR MAS NMR) spectroscopy to generate metabolic profiles from intact colonic biopsies from UC patients and controls to inform us about disease pathophysiology and severity.


Colonic tissue biopsies were acquired from 20 individuals (10 UC patients, and 10 controls who had normal mucosa on colonoscopy) at the time of routine colonoscopy. Metabolic signatures were obtained by HR MAS NMR followed by multivariate pattern-recognition analysis to investigate differences between cohorts, and 2-dimensional correlation spectroscopy (COSY) was performed to aid metabolite identification.


Principal component analysis (PCA) identified one outlier that was excluded from further analysis. Orthogonal projection to latent structures analysis (OPLS) was applied to remaining cases and distinguished UC and control cohorts with significant predictive accuracy. The OPLS model supported good fit (R2Y 0.8362) and predictability (Q2Y 0.5026) parameters. The data analysis showed that colonic biopsies from UC patients contained relatively higher levels of antioxidants (ascorbate) and lysine compared with those obtained from controls. A novel finding was that the model was capable of stratifying the samples based on severity (mild, moderate and severe) which has not been demonstrated in previous metabonomic studies on colonic samples of UC.

Figure 1. OPLS model predicts UC severity.


HR-MAS NMR–based metabolic profiling has identified distinct differences in UC patients and controls, in particular, increased levels of antioxidants and amino acids which may be required in greater quantities during catabolic conditions such as UC. This study shows the utility of HR-MAS NMR in exploring diagnostic possibilities and generating a better understanding of UC pathophysiology.