The relationship between mucosal N3 PUFA content, inflammatory response and disease course, in patients with active ulcerative colitis

  • Daniel Simon Pearl

Student thesis: Doctoral Thesis


Ulcerative colitis (UC) is a chronic inflammatory condition of the colon of unconfirmed aetiology. Frequently, inflammation develops in the rectum and extends to proximal areas of the colon and is often characterized by a sharp demarction from endoscopically normal mucosa. Inflamed mucosa is charaterized by progrssive neutrophil infiltration and tissue destruction, whereas biopsies of non-inflamed mucosa demonstrate quiescent infiltrates or mild architectural distortion only, on microscopic examination. Pathogenesis may be related to alteration of cytokines, eicosanoids or their polyunsturated fatty acid (PUFA) precursors. Current treatment of UC is limited to general immune suppressants and systemic biological agents aimed at modifying tumour necrosis factor (TNF)α, but outcome is unpredictable and a subset of patients will be treatment resistant and require emergency or elective surgery.

Published studies suggest that arachidonic acid (AA) and related eicosanoids are elevated in inflamed colonic mucosa and influence local mucosal cytokine production. Fish oil therapy containing eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) utilised experimentally to modulate composition of eicosanoid PUFA precursors and eicosanoid formation has not translated into clinical therapy. There is a fundamental and philosophical disconnect between the experimental evidence supporting the potential for dietry modulation of tissue PUFA composition and linked therapeutic benefit using fish oils, and the clinical evidence showing at best modest benefit.

It was hypothesised that local up regulation of endogenous fatty acid metabolic pathways may be involved in production of the elevated AA levels in inflamed mucosa. Through competition at cellular eicosanoid biosynthetic enzymes, it is suggested that production of Aa-derived eicosanoids exceeds that of EPA-derived eicosanoids, the latter of which are known to possess relative anti-inflammatory effects. In effect, this may limit the therapeutic potential of fish oil, with resultant pathophysiological processes and clinical sequelae.

The study aims to measure mucosal composition of PUFA, eicosanoid production and cytokine levels in inflamed colonic mucosa in UC, to evaluate the role of dietry derived PUFA in UC pathophysiology. A novel method has been developed to abolish between individual variation and measurement bias of in vitro studies, both of which limit the interpretation of published research. This method will allow direct comparison of metabolic and pathphysiological adjacent processes in between diseased and disease-free colonic mucosa.
The primary objective is to determine if tissue levels of PUFA (linoleic acid [LA], α-linoleic acid [αLA], AA and EPA, docosapentaenoic acid [DPA], and DHA) in colorectal mucosa differ significantly above and below the line of demarcation in patients with active UC.
The secondary objective is to map the profile of lipid derived inflammatory mediators derived from PUFA, including AA and EPA, and to measure cytokine production production representative of innate and adaptive immune pathways (IL-8, Il-17A, IFNy, TNFα, IL-4, TGFβ) with adjacent inflamed and non-inflamed colorectal mucosa.
An observational case-control study model was used. Patients with active UC were recruited and mucosal biopsies taken from adjacent inflamed and non-inflamed areas of colonic mucosa, during endoscopic inspection. Non-UC clinical control patients were recruited and biopsies of 'normal' mucosa taken, during endoscopic inspection. Dietry questionnaires to determine fatty acid intake were completed following the consultation.
Mucosal PUFA analysis was by gas chromatography mass spectrometry (GC-MS) following derivatisation and liquid-liquid extraction of mucosal lipids. Cytokine analysis was by enzyme linked immunosorbent assay (ELISA), on thawed homogenates of mucosa. Lipid mediator measurement was by h liquid chromatography tandem linear ion orbitrap mass spectrometry (LC-MS/MS) following homogenisation and liquid-solid extraction. Microscopic examination was by haematoxylin and eosin staining. All investators were blinded to the patients clinical presentation during recruitment.
Biopsy samples were collected from 69 UC patients (54 providing both inflamed and non-inflamedmucosa) and 69 controls. Inflamed mucosa had higher AA (p<0.001) and lower EPA (p<0.010) contents and a higher AA to EPA ratio (p<0.001). Inflamed mucosa also had higher docosapentaenoic acid (DPA) and docosahexaenoic acid (DHA) and lower linoleic acid (LA) and Biopsy samples were collected from 69 UC patients (54 providing both inflamed and noninflamedmucosa) and 69 controls. Inflamed mucosa had higher AA (p<0.001) and lower EPA (p<0.010) contents and a higher AA to EPA ratio (p<0.001). Inflamed mucosa also had higher docosapentaenoic acid (DPA) and docosahexaenoic acid (DHA) and lower linoleic acid (LA) and α-linolenic acid (α-LNA) contents (all p<0.001), compared to non-inflamed and controls. There were significant correlations between severity of inflammation and contents of AA, DPA and DHA (positive correletions) and of LA, α-LNA and EPA (negative correlations). Levels of the lipid mediators PGE2, PGD2, TXB2, 5-HETE, 11-HETE, 12-HETE and 15-HETE were significantly higher in inflamed mucosa with significant correlations to severity of inflammation, determined using validated endoscopic and histological scoring systems. In addition, significantly higher concentration of IL-8 and lower concentration of TGFβ was measured in inflamed mucosa compared with non-inflamed mucosa (p<0.001; p<0.02) and control mucosa (p<0.001; p<0.001); IL-8 was positively correlated (rs=0.481, p<0.01) and TGFβ inversely correlated (rs=0.462; p<0.01) with grade of inflammation. TNFα concentration was not significantly different. Comparisons of inflamed with non-inflamed mucosa also demonstrate significantly lowerconcentration of IFNy(p<0.001), IL-4 (p<0.005) and IL-17A (p<0.002).
These results demonstrate consistent difference in tissue levels of PUFA and altered eicosanoid and cytokine production in inflamed mucosa characterised by neutrophilic infiltration. The pattern suggests up regulation of the in vivo PUFA biosynthetic pathway with elevated AA and reduced EPA production, and with elevation of AA-derived lipid mediators. Cytokine production is consistent with IL-8 mediated neutrophil infiltration and failures of tissue repair through redcced TGFβ, with apparent down regulation of Th1, Th2 and Th17 adaptive pathways.
The current therapeutic paradigm is insufficiently specific to address the complex inflammatory milieu in UC. Patients with UC today run a similar clinical course to that experienced 40 years ago. The pharmaceutical industry is heavily involved in the development of biological therapies and supporting clinical trials for expensive treatments, which bear little relationship to mucosal pathophysiology. This is further supported by the current results which demonstrate no difference in TNFβ levels in mucosa between inflamed and non-inflamed colonic mucosa.

Mucosal PUFA biosynthetic pathways and eicosanoid synthetic pathways are attractive targets for biomedical research into disease modifying therapies. There is sound experimental evidencethat in vivo fatty acid biosynthetic pathways may be under hormonal control in health, and changes in substrate availability modify mucosal PUFA composition and production of eicosanoid PUFA derived inflammatory mediators. However, biomedical research has failed to translate potential therapeutic targets into treatments.
The current work has clearly demonstrated changes in AA composition in inflamed mucosa, and with corresponding changes in other PUFA which supports up regulation of the fatty acid biosynthetic pathway in inflamed tissue. The profile of lipid derived eicosanoid inflammatory mediators clearly demonstrates that AA is the predominant substrate, and that pro-and anti-inflammatory AA derived lipid mediators co-exist in the same tissue matrix. This would support the hypothesis. Critically, this data clearly demonstrates that TNFα cytokine levels are unchanged in colorectal mucosa, and questions whether the clinical effect seen in UC is related to a nonspecific mode of action on the systemic inflammatory response.
Further research is required to determine if Δ5 or Δ6 desaturase enzymes are up regulated in inflamed mucosa, a finding which may suggest involvement of the fatty acid desaturase (FADS)-1 or FADS-2 gene polymorphisms, which encode these enzymes.
Date of AwardJan 2014
Original languageEnglish
SupervisorJames Brown (Supervisor), Jan Shute (Supervisor) & Graham Mills (Supervisor)

Cite this