In vitro digestion models for dietary phenolic compounds: Dissertation

The aim of this work was to develop in vitro digestion models for mimicking the physiological conditions of upper intestine and microbial conversions in the colon. The main emphasis was on the microbial metabolism of plant phenolic compounds: pure quercetin derivatives, pure anthocyanins and lignans from rye bran and flaxseed. When cereal samples are introduced to an in vitro colon model a removal of digestible components is needed. An enzymatic in vitro digestion model was developed for maximal starch removal from cereal samples. Pepsin, pancreatin and bile concentrations were optimized using an experimental design. Surprisingly, pepsin and bile also affected the extent of starch hydrolysis in synergy with pancreatin. 5-11 % of the original amount of starch remained in the residues of cereal products. Proteins were also partly hydrolysed. The in vitro enzymatic digestion model was used for the pretreatment of rye bran and flaxseed samples. An anaerobic in vitro colon model, conventionally used for the fermentation of non-digestible carbohydrates, was developed further for studying pure phenolic compounds. Human faecal microbiota from several healthy donors was used in the preparation of an inoculum. A low inoculum concentration was used for decreasing the metabolite concentration from the faecal background in the studies concerning pure flavonoids. A dense faecal suspension was suitable for the conversion of rye bran and flaxseed lignans to enterolactone when the plant matrix was present. Flavonoids were deconjugated and degraded to phenolic acids by faecal microbiota. Specific activities of the deconjugative enzymes from the faecal inocula reflected the deconjugation rates of flavonoids. Quercetin aglycone was converted to hydroxyphenylacetic acids, but not to methylated phenolic acids. The extent of metabolism was 60 %, showing that ring-fission was a dominating route in the microbial metabolism of quercetin. Anthocyanins also underwent similar conversion, but the estimated extent of metabolite formation was low (less than 5 %). Protocatechuic acid was identified, and a phenoxyacid or a phenoxyaldehyde was proposed, as ring-fission products of cyanidin. In addition, it was suggested that anthocyanins undergo conjugation with an unknown moiety of 85 mass units. This conjugate was observed for several anthocyanins. Enterolactone production from plant lignans proceeded steadily and slowly for 48 hours in the in vitro colon model using the dense (16.7 %) faecal suspension. Flaxseed lignan conversion to enterolactone was suppressed by the presence of rye matrix. The enterolactone-producing microbiota may be sensitive to non-physiological, low pH values caused by acidic components from rye bran in the presence of microbiota. The presence of rye bran matrix did not interfere with enterolactone formation in an in vivo rat model. The difference in the response to the rye bran matrix may be due to the absorption of the released and metabolised compounds in rats. Rats may also adapt to the diet during their feeding period. This may have enhanced the enterolactone production, and may have further increased the difference between the bioactivity of the microbiota in the in vitro and in vivo models used in this study. Clinical human and animal trials describe end-point metabolism after adaptation to the test diet. The in vitro colon model assists in elucidation of the role of microbiota in the metabolical network of human digestive system and it helps in identification of the crucial reactions. Applications of this method can be extended from the studies of food components to pharmaceutical research.