Curry-Derived Flavor Compounds and their Metabolites in Human Milk and Urine

Curry is a globally widespread aroma-intensive dish that differs significantly from region to region and culture to culture in terms of its ingredients and preparation methods. Due to its intense and multifaceted aroma and taste profile, it is ideally suited for studying various aroma- and taste-active substances concerning their metabolization within the human body. The different metabolization pathways including the uptake of substances, their potential molecular modification, and the excretion by the human body are highly complex. A comprehensive approach to characterizing the metabolism of odorants and tastants is therefore of particular interest. This underlying idea forms the basis for the investigation of odorants and tastants in the same samples of human milk and urine. To find out more about the metabolization and excretion of flavor compounds, a standardized curry dish was designed and served to breast-feeding mothers as part of intervention studies. This curry dish was initially characterized with regard to its odor and taste profile and its ingredients. In the first study with n = 18 participants, each participant provided four human milk and urine samples. Samples were taken before and up to three hours after the intervention. In the case of human milk, the samples were characterized by sensory analysis using aroma profile analyses as well as instrumental analysis using GC-MS and LC-MS; in the case of urine, only instrumental analysis was used. Human milk is, on the one hand, the main foodstuff and influential substance for infants and, on the other hand, a comparatively rare substrate with relatively few scientific studies. The aroma profile analysis of the milk samples revealed, among other things, a statistically significant transition of a bergamot-like odor. This is consistent with the instrumental analysis results, in which a statistically significant transition of linalool, which has a bergamot-like odor, was detected. The detected average concentration of 3.5 μg/L also corresponded to an average odor activity value of 6.8 in the samples after the intervention, which is why it can be considered a transition relevant for smell perception. A statistically significant transition of the trigeminally active pepper tastant piperine in the 14–57 μg/L range was detected additionally. Although these concentrations are probably not perceptible for the infant, it is possible that they still activate the TRPV1 receptor and thus contribute to flavor learning. Parallel observation of the transfer of odorants and tastants into human milk showed that this occurred at a similar velocity (within the first hour after consumption of the curry dish) and similar transfer rates. Another critical finding is the considerable inter-individual differences in all substrates and for all substances. On average, the urine samples examined in the same study showed a significant transfer of the odorants linalool, 1,8-cineole, and eugenol 1.5 hours after consumption of the curry dish in the range of 0.5–2 μg/g creatinine. This already indicates that, on average, the substances pass into urine more slowly than into human milk. It also showed that a higher diversity of the target substances transfer into urine than into human milk. As part of the studies on urine, a sub-experiment was conducted with six additional participants to evaluate the contribution of phase II metabolites to the transition of odorants into urine. The samples were divided into one aliquot that was treated with a β-glucuronidase assay and another aliquot in which this assay was performed without β-glucuronidase. For the aroma compounds linalool and eugenol, treatment of the samples with β-glucuronidase resulted, on average, in a 23- and 77-fold increase in concentration, respectively, suggesting that phase II metabolites account for the majority of excretion of these compounds into human urine. However, these initial studies showed that the sampling period for some substances may not have been sufficient to reflect their transition adequately. Therefore, a second study with n = 20 participants was conducted, in which seven human milk samples and seven urine samples were taken within 0.5 hours before and up to 8 hours after the intervention. It was also ensured that the urine and human milk samples were taken with as little time lag as possible to compare both transitions even more accurately. In order to be able to handle such a high number of samples, a Stir Bar Sorptive Extraction Thermal-Desorption GC-MS (SBSE-TD-GC-MS) method was developed as part of this study, which enabled much faster analysis than the previously used solvent-assisted flavour evaporation (SAFE) method. Although the SAFE method provides a more accurate representation of the overall aroma and is therefore preferable for untargeted analysis, the SBSE method proved to be an excellent alternative for targeted studies such as this one due to its faster performance at similar to lower detection limits. On the one hand, the results of this study confirmed the slower but more diverse transition of odorants into urine compared to human milk; on the other hand, this study was able to illustrate the temporal processes more accurately and show significantly increased odorant concentrations even several hours after consumption of the standardized curry dish. This study also included a sub-experiment with n = 12 participants, in which some aliquots of the milk samples were treated with β-glucuronidase while others remained untreated. On average, the sub-experiment showed an increase of 40, 200, and 8500 percent in linalool, eugenol, and vanillin concentrations for the treated milk samples, respectively. Thus, a significant contribution of phase II metabolites to the aroma transfer was demonstrated for human milk for the first time.