Stable isotope applications to monitor starch digestion and fermentation for the development of functional foods.
Fermentation of carbohydrates in the large bowel results in the formation of short chain fatty acids which have generally been accepted to be beneficial to the host whereas fermentation of proteins gives rise to the production of potentially toxic metabolites such as ammonia, phenols, thiol compounds and amines. Most dietary interventions targeting the colon aim at increasing the saccharolytic activity of the colonic microbiota and decreasing the proteolytic activity. We have developed a non-invasive strategy based on stable isotope technology and focussing on the metabolite ammonia to measure quantitatively the influence of non-digestible carbohydrates on the metabolic processes occurring in the colon. After oral administration, the biomarker lactose [15N,15N] ureide is hydrolysed upon arrival in the colon and yields 15NH3 which can be assimilated by the microbiota followed by faecal excretion or can be absorbed through the colon mucosa followed by urinary excretion of 15N. We have shown that the presence of resistant starch (type 3) in the colon resulted in a significant reduction of urinary 15N-excretion and an increase in faecal 15N, indicating that resistant starch efficiently stimulates bacterial activity and removes in this way the toxic metabolite ammonia from the colonic lumen. Secondly, we have shown that administration of resistant starch combined with wheat bran resulted in a less extensive effect in the proximal colon. However, it was not possible to confirm a shift of the site of fermentation to more distal parts of the colon. The technology is also applicable to other fermentable carbohydrates and allows comparative and quantitative evaluation of so-called prebiotic substrates.
Availability of 13C-labelled starch sources for the investigation of starch digestion and fermentation
Starch is the most important source of energy. In spite of that, the information about the metabolic quality of starchy foods is scarce. Much of the discussions about the metabolic quality of starchy food is focused on the discussion about the glycaemic response and the from it derived glycaemic index. However, the glycaemic response is the net result of the influx rate of exogenous glucose, the glucose production of the liver and the glucose disposal in the tissue. To be able to investigate in detail physiological and metabolic effects related to various starchy foods it would be advantageous to be able to monitor the underlying glucose kinetics and to relate this to other metabolic effects. For example the influx of glucose over a long period as compared to a high and short influx not only results in lower glycemia but also in differences in insulinemia, suppression of endogenous glucose synthesis and other metabolic parameters which are likely to be relevant for the prevention of disease. Labelling of starchy food with stable isotopes provides the possibility to study postprandial glucose kinetics in more detail. In the EUROSTARCH project the dual isotope technique was applied which comprises a continuous D-[6,6-2H2]glucose infusion and the ingestion of a starchy meal labelled with the stable isotope 13C. With this technique the rate of appearance of exogenous glucose can be estimated, reflecting the net rate of intestinal glucose absorption and the endogenous glucose synthesis can be calculated. To be able to apply this technique starchy products which are labelled with the stable isotope 13C are thus needed. Carbohydrates, that are not digested in the small intestine, pass into the colon and are (partly) fermented to form a variety of metabolites. This process was once thought to be waste processing, but nowadays the formation of metabolites during fermentation process and subsequent metabolism in the colonocytes are processes which are linked to the occurrence of diarrhoea, constipation, energy salvage, colon cancer, and immune status. More recently evidence is emerging that colonic fermentation not only affects gut metabolism but can also influence metabolic processes in other tissues and organs. Short- and long-term consumption of starch resistant to small intestinal digestion for example has been shown to beneficially increase insulin sensitivity in healthy subjects. The underlying mechanism of this phenomenon is not known, but short-chain fatty acids (SCFA�s), like acetate and propionate, which are products of starch fermentation, could be involved. These SCFA�s are metabolized to some extent by the colonic epithelial cells but also enter the portal circulation. Effects of SCFA s on liver metabolism have been reported and recently SCFA s have been identified to be ligands for the orphan G protein-coupled receptor GPR41 which is primarily expressed in adipose tissue. Adipose tissue is known to secrete various signaling peptides - the adipokines - influencing among others insulin sensitivity, feeding behavior and inflammation and could constitute a link between colonic fermentation and peripheral metabolic effects. However, the scientific evidence of the interrelationship between fermentation products and systemic effects is limited, partly caused by the inaccessibility of the colon. By labelling sources of resistant starch with 13C it is possible to monitor the production of 13C-labelled SCFA�s and relate this to systemic effects or health outcomes. Since there are only limited starchy food sources which are intrinsically labelled - corn and teff - culturing of grains or tubers under 13CO2 atmosphere is necessary. Furthermore the 13C-enrichment of intrinsically labelled starch sources has been shown to be too low to be able to follow fermentation products in vivo. During the project 13C-enriched wheat and barley were produced. Durum wheat has been used to prepare wholemeal bread and the glucose kinetics of after ingestion of bread could be monitored for the first time in healthy volunteers. The magnitude of 13C-enrichment necessary for monitoring in vivo 13C-SCFA s has been established. Highly enriched 13C-labelled barley has then been used to follow 13C-acetate production in vivo. The availability of 13C labelled starch sources opens new avenues for research. The availability of 13C-wheat allows us to study the digestion of 13C-wheat bread. 13C-barley gives us the possibility to analyse in detail the second meal effect. 13C-lactose allows us to make 13C-lactulose an important substrate for colonic fermentation.
Colonic fermentation of dietary fibres, non digestible oligosaccharides or resistant starches by the bacterial microflora results in a production of different organic acids and mainly short chain fatty acids (acetate, propionate and butyrate). These short chain fatty acids are mostly absorbed and partly metabolized in the colonic epithelium and the liver. As a consequence they appear at very low concentration in the stools but can be quantified in the peripheral blood before their captation by different organs. Several methods (long and tedious) are available to quantify SCFAs in the plasma but a precise quantification of acetate (the main SCFA) remains very difficult. Stable isotope techniques are increasingly used to understand metabolism of SCFAs in human. GC/MS is a powerful analytical tool for the stable isotope studies but requires high isotopic enrichments of 13C-labelled SCFA (> 1MPE) in the plasma sample. To measure low isotopic enrichments of 13C-SCFA (or natural abundance < 1 MPE), we have developed a method in GC-C-IRMS using a Solid Phase Micro-Extraction (SPME). After removal of plasma proteins, the SPME fiber was plunged in the liquid sample during 40 min at 40°C. Then, acetate was directly desorbed into GC-C-IRMS. The accuracy of isotopic enrichment measurement was determined using plasma spiked with 13C-acetate and 13C-butyrate solution from 0 to 1 Mol Percent Excess (MPE). Good linearity and repeatability (RSD < 5%) were obtained for acetate and butyrate. Plasma acetate, propionate and butyrate concentrations were also determined relative to 3-Methylvalerate (Internal Standard). Good linearity and repeatability were observed from 0 to 400 µM for acetate, from 0 to 20 µM for propionate and from 0 to 10 µM for butyrate. This method was applied to determine plasma acetate production obtained from the lactulose fermentation in one healthy volunteer over 3 hours.. Acetate concentration was increased twofold, 2 hours after oral lactulose intake. These results are in agreement with those measured by GC/MS in healthy and overweight adults following a lactulose intake by using higher amount of labelled tracers. Natural abundance of 13C-acetate was of 29.93 +/- 0.78 0 (mean n=8) at basal state and 25.87 +/- 0.80 0 (mean n=8) during the production of lactulose. These results illustrate the accuracy of the SPME/GC-C-IRMS method. In conclusion, the combination of SPME in liquid sampling mode with the GC-C-IRMS allows to measure very low enrichments (in the range of natural abundance carbon isotope) of VFAs in human plasma within physiological concentrations. This new method is now used to study the colonic fermentation of 13C-starch in an integrated protocol into human volunteers in collaboration with other Eurostarch partners (HNRC-Nantes, U. Glasgow, RUG, KULRD).
The main focus of this work concerns a procedure to measure volatile fatty acid production in the body. We are now in a much better postion in being able to measure and relate this to potential health benefit of different diets. We use stable isotope tracers to investigate VFA production by isotope dilution. Key steps in validating this protocol have been undertaken. Firstly, we have developed and proven analytical methods. We have developed a robust GCMS method to measure the concentration and 2H-enrichment of VFA in urine and plasma. This new method has been proven during our collaborative studies. We have ensured that the sample preparation procedure yields sufficient sample, in the correct form, for both GCMS and GC-combustion-Isotope Ratio MS (GC-C-IRMS) analysis. GC-C-IRMS is a new method for compound-specific analysis at low 13C-enrichment. This is a novel development made entirely during the Eurostarch project. Most recently we have demonstrated that our procedure can be extended to propionate analysis. We have placed this work in the public domain. A paper has been published in Rapid Communications in Mass Spectrometry, where EU funding was fully acknowledged: Morrison et al (2004) Rapid Communications in Mass Spectrometry 18: 2593-2600.
Knowledge about the digestive characteristics and postprandial glucose kinetics of whole-meal wheat bread
Starchy foods differ considerable in their physiological/metabolic response and consequently also in their potential health benefits. Several studies have shown that diets that contain large amounts of slowly digestible starch - which elicits relatively low postprandial blood glucose and insulin responses - may protect against chronic disease like diabetes mellitus type 2 and cardiovascular disease. To be able to distinguish starchy foods according to their hyperglycemic potential the glycaemic index (GI) is used. Also, in vitro techniques are available that aim to characterize carbohydrate digestion in the gut. With these techniques carbohydrates in food are divided in sugars and starch fractions and are grouped into rapidly available glucose (RAG), slowly available glucose (SAG) and resistant starch. It is generally assumed that the postprandial blood glucose response to starchy foods is mainly determined by the rate of digestion of starch - thus products with slow digestible starch features are considered to be beneficial. However postprandial blood glucose is the net result of various postabsorptive processes and thus - besides absorption of starch derived glucose - also endogenous glucose production and glucose disposal in tissue determine blood glucose response. The contribution of these processes to postprandial blood glucose has so far received little attention in nutritional studies. In many societies bread is the main source of starch. But bread made from refined flour and even whole-meal bread - if made from finely ground whole grains - has been shown to have rapidly digestible starch features. The content of RAG as measured in vitro is high (90 % of total starch) and so is the GI (71 2 mean of 13 studies; glucose = 100) - which implies that it is a less desirable food product. Yet no information is available about the digestive starch characteristics of this important diet component in vivo. Thus, the aim of this study was to characterize starch digestion of wholemeal wheat bread and the underlying glucose kinetics which determine postprandial glucose response in healthy volunteers. In a crossover study 4 healthy men ingested either 13C-enriched wholemeal wheat bread (WB) (133 g) or glucose (55 g) in water. Plasma glucose and insulin concentrations and 13CO2 excretion in breath were monitored during 6 h postprandially. Using a primed continuous infusion of D-[6,6-2H2] glucose, the rate of appearance of glucose was estimated (reflecting glucose influx) and the endogenous glucose production calculated. All test meals were naturally labelled with 13C. Wheat was grown in an atmosphere artificially enriched in 13CO2. Wheat grains were finely milled and whole-meal bread produced. We found that the glucose influx rate after WB was comparable with that after glucose in the early postprandial phase (0-2 h) and higher in the late postprandial phase (2-4 h). Despite the same initial glucose influx rate the 0-2 h AUC of insulin after WB was 41% lower than after glucose (P = 0.037). Paradoxically endogenous glucose production after WB was significantly more suppressed than after glucose (0-2 h AUC: p = 0.015, 2-4 h AUC: p = 0.028). Our data show that starch in WB is partly rapidly and partly slowly digestible which could not be derived from in vitro determination. Postprandial insulin response and endogenous glucose production after WB are not solely determined by the digestive characteristics of starch. Other components of WB seem to be involved and need to be identified. Suppression of endogenous glucose production seems to be largely insulin independent after WB consumption. This is an interesting phenomenon which might be especially relevant in the insulin resistant state. A manuscript is currently prepared to be published in a peer-reviewed international journal. By use of wheat labelled with the stable isotope 13C, which has been produced during the EUROSTARCH project, it was for the first time possible to investigate in vivo the digestive characteristics and postprandial glucose kinetics of wholemeal wheat bread. The study has shown that prediction of in vivo digestibility with in vitro tests needs further evaluation. The use of stable isotopes also made it possible to observe that not only the digestive characteristics of starch are involved in determining postprandial metabolic responses but also other components of the bread, which need to be identified. If this can be confirmed in other starchy products this will have implications for the classification of food products which is currently focused on starch characteristics and will emphasis the need to monitor more metabolic factors than the postprandial glucose to determine the health effect of a given product.
Knowledge about the impact of starchy foods differing in glycemic index and dietary fiber content on the postprandial glucose tolerance of a subsequen
The effect of GI vs. content of indigestible carbohydrates (dietary fibre and resistant starch (RS)) of evening meals or breakfasts on glucose tolerance at subsequent meals ("second-meal effect") was evaluated in healthy subjects. The GI of some of the test products was predicted using an in vitro method (hydrolyse index, HI)in order to allow for an experimental design capable of discriminating between the above characteristics of starchy foods. We concluded that the mechanisms for the second meal effect probably differ depending on the time period in between the meals. In the case of benefits on glucose tolerance in the perspective from breakfast to lunch, the key feature involved is probably low-GI feature per se of the starchy food. However, in the perspective from evening to breakfast, or breakfast to the evening meal, other features are also likely to be involved, such as the type and amount of indigestible carbohydrates. A high dose of barley fibre added to a pasta evening meal slightly improved glucose tolerance at breakfast compared with white bread, whereas ingestion of boiled intact barley kernels in the evening significantly lowered blood glucose responses at breakfast compared with a white wheat bread, or spaghetti. Also, a breakfast consisting of boiled barley kernels improved the glucose tolerance over the course of a whole day (breakfast lunch, and dinner).The improved glucose tolerance over the longer time period (10 h) with the boiled barley kernel meal appeared to emanate from colonic fermentation of indigestible carbohydrates, or to the combination of the low GI features and colonic fermentation. Consequently, breath hydrogen excretion, Short chain fatty acid levels were increased and free fatty acid levels decreased at the time when commencing "the second" meal. The design of low-GI foods capable of improving glycaemic excursions at consecutive meals might be particularly advantageous in reducing risk factors for the insulin resistance syndrome, and the data indicate additive effects of certain low GI foods rich in RS and barley fibre. The studies have pursued previous findings in the Lund research group, and the presently performed work in Lund is the first to demonstrate that the optimization of carbohydrate characteristics of starchy food may improve glucose metabolism over the course of a day. Three human studies have been performed. The results of the first study are published in a scientific journal ("Effects of GI and content of indigestible carbohydrates of cereal based evening meals on glucose tolerance at a subsequent standardised breakfast", Nilsson A, Granfeldt G, et al. (2006), Eur J Clin Nutr.), one is submitted ("Effects of the GI character and content of indigestible carbohydrates at the evening meal for glucose tolerance and metabolic variables the subsequent morning",Nilsson A, Östman E, Björck I), and the third is being prepared for publication.
The aim of the work was to find useful methods to study potential differences in cognitive function during the postprandial period after starchy test meals differing in post meal glycaemia. In addition, the aim was to study cognitive performance in relation to glucose tolerance status. Test meals were designed to give rise to different postprandial blood glucose increments. Cognitive tests of working memory (WM) were performed repeatedly during the 3 h postprandial phase. A test of selective attention (SA) was also included. The cognitive tests were performed in healthy volunteers. In summary, the cognitive tests used (WM and SA) allowed discrimination of cognitive performance as related to differences in postprandial glycaemia. A breakfast that had the capacity to maintain a higher net increment in blood glucose in the later postprandial period, i.e. a low-GI breakfast, showed an advantage on cognitive performance compared with a high-GI breakfast. After adjusting for glucose tolerance, the subjects performed significantly better in the later postprandial period during low-GI condition in both the WM-test and the test of SA compared with the high-GI condition. But, even though a low-GI breakfast seemed to be preferable for the cognitive functions over all, a high GI-breakfast showed to enhance the learning capacity. Thus, it was found that the learning capacity was enhanced when the first learning occasion was performed when the blood glucose increment was most elevated, i.e. in the early postprandial period after the high-GI breakfast. Also, the results showed that the individual glucose tolerance has an important influence on cognitive performance. The subjects with higher glucose tolerance performed better in the cognitive tests, and we therefore suggest that the cognitive performance can be affected by glucose regulation even if the glucose tolerance is within the normal range. The results in the project may result in new knowledge about what foods or meals that is preferable with respect to cognitive functions in the postprandial period, and also provide useful tools to design foods with desirable effects on glycaemic excursion and cognitive function. The results are being prepared for scientific publication.
Knowledge about the impact of substituting high GI starchy foods by low GI starchy foods on metabolic parameters in obese subjects
The glycaemic index has been developed in order to classify food according to the glycaemic response after carbohydrates ingestion. A 5-weeks nutritional, interventional trial was conducted using 2 types of dietary regimens. Overweight healthy subjects of both sexes were given dietary advice to replace their usual starchy foods ad libitum by either Low Glycaemic Index (LGI) or High Glycaemic Index (HGI) starchy products. Foods were considered as having a low GI whenever GI<50% and a high GI whenever GI>70%. Moreover, between Day 1 and Day 36, the subjects from the LGI group have consumed daily a breakfast containing plain biscuits (LGI biscuits) and the subjects from the HGI group have consumed daily a breakfast containing extruded cereals (HGI). The biscuits and the extruded cereals resulted from different technological process and exhibited differences in concentrations of Rapid Available Glucose (RAG) and Slow available Glucose (SAG). Plain biscuits contained 39% of SAG and 61% of RAG, the extruded cereals contained less than 2% of SAG. These breakfasts represented around 20% of daily energy intake. They were isocaloric (around 431 kcal) and contained the same amount of proteins (12%), lipids (26%) and carbohydrates (62%). The only variable parameter was the glycaemic index (45 ± 6 % and 70 ± 12 % for biscuits and cereals, respectively). The 38 subjects were men and women (LGI diet: n= 19; HGI diet: n= 19) [BMI: 27.3 +/- 1.5 kg/m2] aged 20-60 y. The study was a parallel, randomized, intervention trial with 2 matched groups. On the first day (d1) and on the last day (d36) of the nutritional intervention, glycaemic, insulinic, lipid blood profile, body composition and metabolism were measured after ingestion of the same cereals products than usual but whom carbohydrates were labeled with stable isotope 13C. Results: The diet groups did not differ in total energy or macronutrient intakes. Dietary goals were reached in both groups with a significant difference in GI between groups (p<0.0001). Mean body weight decrease was significant in the LGI diet group (-1.1 ± 0.3 kg, p=0.004) and this decrease was significantly greater than in the HGI group (-0.3 +/- 0.2 kg, p=0.04 between groups). No significant differences in body fat mass were observed. However, neither diet altered fasting insulin, insulin resistance indexes (QUICKI, HOMA-IR), lipid profile and substrate oxidations. The LGI diet decreased total cholesterol by 9.6% (p<0.001), HDL cholesterol by 2.5% (p=0.04), LDL cholesterol by 8.6% (p=0.01) and both LDL to HDL ratio (10.1%, p=0.003) and total to HDL cholesterol ratio (8.5%, p=0.001). No significant changes in lipid parameters were observed in the HGI group. At d1, before 5-weeks diet, after ingestion of the plain biscuits or of the extruded cereals, the area under curve of glycemia was lower for LGI than HGI. By using 13C labeled starch, the exogenous glucose appearance at d1 has been measured and was 68.7% of 13C glucose ingested for LGI vs 96.3% for HGI in 270 min (40.2 +/- 1.9 vs 56.4 +/- 2.5 g respectively, p<0.02). Total glucose appearance at d1 trended to be lower after LGI than after HGI breakfasts (63.9 +/- 2.0 vs 69.0 +/- 1.8 g, NS). On the same day, the endogenous glucose production was significantly less inhibited during LGI breakfast despite a similar level of insulinemia (23.7 +/- 2.1 vs 12.6 +/-2.0 g respectively, p=0.001). Comparing d36 and d1, differences between acute and chronically ingestion have been found: the 5-weeks dietary intervention did not alter neither exogenous glucose appearance, total glucose appearance nor endogenous glucose production. Conclusion: The cholesterol profile has been improved in the LGI group after the 5-weeks intervention, and in this group, loss of 1 kg of body weight has been registered. But no effect on the glucose turnover has been observed. GI is useful to classify starches but could not resume all the metabolic properties of the food. The metabolic results could be modulated by the other components of the meal. More studies are needed to conclude on a positive effect of the GI of the diet on the metabolic parameters.