Bioactive compounds produced during food processing can have strong pro-inflammatory properties with potential health implications. Modulation of chronic inflammation may be the mechanism linking diet to risk of chronic diseases such as diabetes and CVD. Advanced glycation endproducts (AGEs) are a heterogeneous group of pro-inflammatory bioactive compounds produced via Maillard reactions during cooking and processing. It is now well established that AGEs are mainly formed from several dicarbonyl compounds, including methylglyoxal (MGO), glyoxal (GO) and 3-deoxyglucosone (3-DG). The glycation activity of these biologically reactive dicarbonyl compounds is much higher as compared to that of sugars, with MGO as the most reactive precursor in the formation of AGEs. We have recently found high levels of MGO in many different foods. There is increasing evidence that elevated levels of MGO are involved in weight gain and the development of diabetes and other chronic inflammatory diseases including cardiovascular disease. However, bioavailability and physiological consequences of dietary MGO are largely unknown. Our aim is to explore the consequences of dietary MGO on the intestinal microbiota and on the development of metabolic diseases.

We will determine the effect of dietary MGO on the gastrointestinal tract and microbiota (WP1) and on the onset of diabetes, vascular diseases and cognitive function in mice (WP2). We will develop a detailed database of dietary MGO exposures and assess the association of dietary MGO with overweight, weight gain, obesity and risk of associated metabolic diseases (type 2 diabetes, CVD), as well as cognitive function using existing data from 3 large and deep-phenotyping prospective cohort studies (WP3). In WP3, we will also investigate the role of inflammation, endothelial function and micro- and macrovascular function and microbiota composition as potential underlying mechanisms of dietary MGO action.

This comprehensive project will elucidate the role of food-derived MGO as a possible risk factor for overweight and overweight-related metabolic diseases.

For WP1 on the dietary MGO: in vitro and in vivo studies. The in vitro digestion model set up in this WP was used for further investigations on the reactivity of MGO with amino compounds and proteins during simulated digestion. Furthermore, the 1,2-α-dicarbonyl compounds 3-deoxyglucosone (3-DG) and 3-deoxygalactosone (3-DGal) with higher quantitative relevance in food were included. Analytics were performed by using HPLC-UV after derivatization (1,2-α-dicarbonyl compounds) or LC-MS/MS (glycation compounds), respectively. Reference substances were synthesized to allow quantitation of the most prominent MGO-derived glycation compounds MG-HCr, CEL and MG-H1. The putative formation of some less prominent glycations compounds (carbocxyethylcystein, MOLD, MODIC, MOLA) was monitored using LC-qTOF-MS. A decrease of dicarbonyl compounds was observed for all the dicarbonyl compounds of interest depending on time and stage of digestion. Reactions with digestive enzymes and added amino compounds were most prominent for MGO when compared to 3-DG or 3-DGal. The formation of the respective reaction products was observed simultaneously for the added amino compounds, but also for digestive proteins like enzymes that are added in the system to perform the simulated digestion. The amino acids used in the study were found to react in the following order: cysteine > creatine > arginine > lysine. Our data suggest that “digestive” glycation compounds are formed in systems containing MGO, 3-DG, 3-DGal and amino compounds. The major proportion of dietary dicarbonyl compounds ingested with food seem to react with amino compounds and proteins in the bolus and are thus “scavenged” and not available for reactions with endogenous proteins. The investigations of the “downstream reaction sites” of MGO in WP 1.2 and WP 1.3 are ongoing.

For WP2 on the dietary MGO and physiological consequences in mice, where mice were subjected to 1 mM and 50 mM MGO via drinking water, we found significantly increased plasma levels of MGO (2.5 fold) and MG-H1 (1.5 fold) in these mice compared to the control. Related to insulin sensitivity measures, no differences were found in sober glucose levels between the MGO-exposed groups and the control. Only in the low MGO group, the glucose and insulin tolerance test showed a significantly increased sensitivity for insulin (respectively -17% plasma glucose AUC and -39% plasma glucose) in the low MGO group compared to the control. Interestingly, the low dose MGO group showed improved NO-mediated vasorelaxation in small arteries and the femoral arteries, increased NO sensitivity, and indications for more intact nerve endings around the saphenous artery. Thus, results indicate a favorable effect of a chronic low dose MGO for vascular relaxation in both small and large arteries. Concerning the effect of MGO intake on the brain, both MGO and GO levels in the brain were significantly reduced in the low dose MGO group, while no differences were detected for the high dose MGO group, compared to the control. qPCR analysis of the brain revealed a downregulation of the MCP-1 pro-inflammatory response. Next, we will perform experiments including metabolic activity assays to elucidate the mechanism behind the lower MGO and GO concentration in the low-dose MGO group. Interestingly, with respect to the microbiome, findings on the first mice study showed that mice fed with 50 mM methylglyoxal have less Ligilactobacillus and Lactobacillus in their gut microbiota. Measures on the latest mice experiment, including the low MGO (1 mM) group will follow later this year. Also, a large panel of biomarkers, reflecting endothelial dysfunction and low-grade inflammation were analyzed in the plasma of all groups. In the low MGO group, both IL-10 and TNFalpha were significantly increased, while VCAM was significantly lower. CRP levels were significantly increased in both the low and high MGO groups, compared to the control. Currently, a wide range of tissues including the heart, kidney, brain, fat depots, pancreas, muscle, and liver, which were collected directly after sacrificing the animals are being processed for the detection of MGO accumulation or MGO-derived AGEs including the neoformation.

For WP3 on the impact of dietary MGO and its consequences in humans using existing data and resources from large-scale international prospective cohort studies in the second year of the project saw the completion of the development, quality control assessment and integration of the dietary dicarbonyl consumption data into the dietary database of the EPIC, the Maastricht Study  and CODAM cohorts. Integration of the dietary dicarbonyl consumption data for the Maastricht Study and CODAM cohorts was achieved earlier than that of EPIC. Both tasks required the individual level and highly detailed matching of the comprehensive Schalkwijk food dicarbonyl composition database to the individual dietary intake data for each cohort. In EPIC, this task proved much more challenging than initially foreseen given the multi-centric, international composition of the cohort and the observed large variations in consumption of and heterogeneity in foods and food products rich in dicarbonyl compounds. To address this challenge, the nutritionists, dietary database experts and epidemiologists involved in the task collectively decided that the analysis of additional foods and food items was necessary. Thus, a list of 50 most important items that were majorly consumed in EPIC sub-cohorts but which lacked compositional data on dicarbonyl compounds was agreed upon. All these foods were analysed in Professor Schalkwijk’s laboratory by his team, the data were integrated into a revised Schalkwijk food dicarbonyl composition database and then applied to the EPIC cohort data to obtain individual level information on the types and amounts of dicarbonyl compounds consumed by EPIC participants. The data was then integrated into the centralized EPIC database and datasets for statistical analyses were subsequently generated, in line with the ePIDEMic objectives for WP3. Statistical analyses of these datasets have commenced and will continue into the 3rd year of the project.