Fatty acid biomarkers of intake and metabolism and their association with type 2 diabetes

English. Background: The role of fatty acid (FA) intake and metabolism in type 2 diabetes (T2D) incidence is controversial. Some FAs are not synthesised endogenously and, therefore, these circulating FAs reflect dietary intake, for example, the trans fatty acids (TFAs), saturated odd chain fatty acids (OCFAs), and linoleic acid, an n-6 polyunsaturated fatty acids (PUFA). It remains unclear if intake of TFA influence T2D risk and whether industrial TFAs (iTFAs) and ruminant TFAs (rTFAs) exert the same effect. Unlike even chain saturated FAs, the OCFAs have been inversely associated with T2D risk, but this association is poorly understood. Furthermore, the associations of n-6 PUFAs intake with T2D risk are still debated, while delta-5 desaturase (D5D), a key enzyme in the metabolism of PUFAs, has been consistently related to T2D risk. To better understand these relationships, the FA composition in circulating lipid fractions can be used as biomarkers of dietary intake and metabolism. The exploration of TFAs subtypes in plasma phospholipids and OCFAs and n-6 PUFAs within a wide range of lipid classes may give insights into the pathophysiology of T2D. Aim: This thesis aimed mainly to analyse the association of TFAs, OCFAs and n-6 PUFAs with self-reported dietary intake and prospective T2D risk, using seven types of TFAs in plasma phospholipids and deep lipidomics profiling data from fifteen lipid classes. Methods: A prospective case-cohort study was designed within the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study, including all the participants who developed T2D (median follow-up 6.5 years) and a random subsample of the full cohort (subcohort: n=1248; T2D cases: n=820). The main analyses included two lipid profiles. The first was an assessment of seven TFA in plasma phospholipids, with a modified method for analysis of FA with very low abundances. The second lipid profile was derived from a high-throughout lipid profiling technology, which identified 940 distinct molecular species and allowed to quantify OCFAs and PUFAs composition across 15 lipid classes. Delta-5 desaturase (D5D) activity was estimated as 20:4/20:3-ratio. Using multivariable Cox regression models, we examined the associations of TFA subtypes with incident T2D and class-specific associations of OCFA and n-6 PUFAs with T2D risk. Results: 16:1n-7t, 18:1n-7t, and c9t11-CLA were positively correlated with the intake of fat-rich dairy foods. iTFA 18:1 isomers were positively correlated with margarine. After adjustment for confounders and other TFAs, higher plasma phospholipid concentrations of two rTFAs were associated with a lower incidence of T2D: 18:1n-7t and t10c12-CLA. In contrast, the rTFA c9t11-CLA was associated with a higher incidence of T2D. rTFA 16:1n-7t and iTFAs (18:1n-6t, 18:1n-9t, 18:2n-6,9t) were not statistically significantly associated with T2D risk. We observed heterogeneous integration of OCFA in different lipid classes, and the contribution of 15:0 versus 17:0 to the total OCFA abundance differed across lipid classes. Consumption of fat-rich dairy and fiber-rich foods were positively and red meat inversely correlated to OCFA abundance in plasma phospholipid classes. In women only, higher abundances of 15:0 in phosphatidylcholines (PC) and diacylglycerols (DG), and 17:0 in PC, lysophosphatidylcholines (LPC), and cholesterol esters (CE) were inversely associated with T2D risk. In men and women, a higher abundance of 15:0 in monoacylglycerols (MG) was also inversely associated with T2D. Conversely, a higher 15:0 concentration in LPC and triacylglycerols (TG) was associated with higher T2D risk in men. Women with a higher concentration of 17:0 as free fatty acids (FFA) also had higher T2D incidence. The integration of n-6 PUFAs in lipid classes was also heterogeneous. 18:2 was highly abundant in phospholipids (particularly PC), CE, and TG; 20:3 represented a small fraction of FA in most lipid classes, and 20:4 accounted for a large proportion of circulating phosphatidylinositol (PI) and phosphatidylethanolamines (PE). Higher concentrations of 18:2 were inversely associated with T2D risk, especially within DG, TG, and LPC. However, 18:2 as part of MG was positively associated with T2D risk. Higher concentrations of 20:3 in phospholipids (PC, PE, PI), FFA, CE, and MG were linked to higher T2D incidence. 20:4 was unrelated to risk in most lipid classes, except positive associations were observed for 20:4 enriched in FFA and PE. The estimated D5D activities in PC, PE, PI, LPC, and CE were inversely associated with T2D and explained variance of estimated D5D activity by genomic variation in the FADS locus was only substantial in those lipid classes. Conclusion: The TFAs' conformation is essential in their relationship to diabetes risk, as indicated by plasma rTFA subtypes concentrations having opposite directions of associations with diabetes risk. Plasma OCFA concentration is linked to T2D risk in a lipid class and sex-specific manner. Plasma n-6 PUFA concentrations are associated differently with T2D incidence depending on the specific FA and the lipid class. Overall, these results highlight the complexity of circulating FAs and their heterogeneous association with T2D risk depending on the specific FA structure, lipid class, and sex. My results extend the evidence of the relationship between diet, lipid metabolism, and subsequent T2D risk. In addition, my work generated several potential new biomarkers of dietary intake and prospective T2D risk.

German. Die Rolle der Fettsäureaufnahme und des Fettsäurestoffwechsels bei der Prävention von Typ-2-Diabetes (T2D) ist nach wie vor umstritten. Die Fettsäure (FS)-Zusammensetzung in den Blutfettfraktionen kann als Biomarker für die Nahrungsaufnahme und den Stoffwechsel verwendet werden, um die Beziehung zwischen den FS und dem T2D-Risiko besser zu verstehen. Das Hauptziel dieser Arbeit war es, den Zusammenhang zwischen zirkulierenden trans-FS (TFS), ungeradkettigen gesättigten FS (UGFS) und n-6 poly ungesättigte FS (PUFS), die in verschiedenen Lipidklassen angereichert sind, mit dem T2D-Risiko zu untersuchen. Mit einer eingebetten Fall-Kohorten-Studie, die im Rahmen der prospektiven EPIC-Potsdam-Studie konzipiert wurde, untersuchte diese Arbeit zwei Lipidprofile im Hinblick auf das T2D-Risiko: (1) Sieben TFS-Subtypen in Plasma-Phospholipiden und (2) die Zusammensetzung von UGFS und PUFA in 15 Plasma-Lipidklassen. Die Aktivität der Delta-5-Desaturase (D5D) wurde als 20:4/20:3-Verhältnis geschätzt. Assoziationen mit dem Auftreten von T2D wurden mit multivariablen Cox-Regressionsmodellen untersucht. Von den üblicherweise aus Milchprodukten stammenden TFS waren 18:1n-7t und t10c12-CLA mit einer geringeren T2D-Inzidenz, c9t11-CLA mit einer höheren Inzidenz und 16:1n-7t nicht mit dem T2D-Risiko assoziiert. TFS aus industriellen Quellen (18:1n-6t, 18:1n-9t, 18:2n-6t) zeigten keinen statistisch signifikanten Zusammenhang mit dem T2D-Risiko. Die UGFS-Konzentration im Plasma war mit dem T2D-Risiko auf lipidklassen- und geschlechtsspezifische Weise assoziiert, wobei bei Frauen stärkere inverse Zusammenhänge für 15:0 in Monoacylglycerinen (MG), Phosphatidylcholinen (PC) und Diacylglycerinen (DG) sowie für 17:0 in PC, Lysophosphatidylcholinen (LPC) und Cholesterinestern (CE) beobachtet wurden. Höhere Konzentrationen von 18:2 waren in DG, TG und LPC invers mit dem T2D-Risiko assoziiert, während MG(18:2) positiv mit dem T2D-Risiko assoziiert war. Höhere Konzentrationen von 20:3 in Phospholipiden (PC, PE, Phosphatidylinositole (PI)), Fettsäuren (FFS), CE und MG waren mit einer höheren T2D-Inzidenz verbunden. 20:4 stand in den meisten Lipidklassen in keinem statistisch singifikanten Zusammenhang mit dem Risiko, außer bei in FFS und PE angereichertem 20:4, das positiv assoziiert war. Die geschätzten D5D-Aktivitäten in PC, PE, PI, LPC und CE waren invers mit dem T2D-Risiko assoziiert. Zusammenfassend ist die Konformation der TFS für ihren Zusammenhang mit dem Diabetesrisiko entscheidend. Die Assoziationen der UGFS-Plasma-Konzentrationen mit dem T2D-Risiko zeigten lipidklassen- und geschlechtsspezifische Unterschiede. Die Plasma-Konzentrationen der n-6-PUFA waren je nach spezifischer FA und Lipidklasse unterschiedlich mit der T2D-Inzidenz assoziiert. Insgesamt unterstreichen diese Ergebnisse die Komplexität der zirkulierenden FAs und ihren heterogenen Zusammenhang mit dem T2D-Risiko in Abhängigkeit der spezifischen FA-Struktur, der Lipidklasse und des Geschlechtes.