Serum metabolite patterns of adipose tissue distribution and body composition subphenotypes

Background Obesity prevalence is increasing globally, accompanied by increases in obesity-related diseases. While obesity is usually defined by BMI, the development of obesity-related disease might be better characterized by specific adipose tissue (AT) distributions or body composition subphenotypes. Serum metabolite patterns reflecting AT distribution could provide insights into potential underlying pathophysiological pathways and the interplay between AT depots. We therefore aim to identify metabolite signatures associated with specific AT depots and body composition subphenotypes. Methods Targeted metabolites (Biocrates p180 kit) were measured in fasted serum for N = 390 individuals from the population-based KORA-FF4 cohort (42% women, average age 56y). AT was measured by magnetic resonance imaging. Association of n = 29 AT depots (visceral (VAT), subcutaneous (SAT), pancreas, bone marrow, skeletal muscle, heart, kidney) and five body composition subphenotypes with 146 metabolites and 40 derived indicators were investigated by linear regressions with confounder adjustment for traditional cardiovascular disease risk factors and life style parameters. Results Subphenotypes were associated with 59, and single ATs with 275 metabolites or indicators, with VAT and SAT showing most associations. Compared to subphenotype I (low overall ATs), subphenotype II (average ATs) showed positive associations with diacylglycerophospholipids with differently saturated C32 fatty acid and sphingomyelins. Subphenotype III (high muscle and bone marrow fat) was negatively associated with total lysophosphatidylcholines (lyso-PCs) and total monounsaturated lyso-PCs, while showing a positive association with total long-chain acylcarnitines (C14–C18). Subphenotype IV (high SAT, high VAT and high liver fat) exhibited positive associations with short-chain acylcarnitines, alanine and aromatic amino acids. Subphenotype V (high pancreas fat fraction) was related to arginine and the ratio of ornithine and arginine as surrogate for ornithine synthesis. Three metabolites or indicators (lysoPC C 18.2, total polyunsaturated lyso-PC, phospholipase A2 as ratio of lyso-PC/diacyl- and acylalkylglycerophospholipids) were associated with all subphenotypes. These results were supported by the associations of individual ATs with metabolites or indicators. Conclusions ATs, including ectopic fat depots such as pancreas fat, and subphenotypes of body composition show distinct serum metabolite patterns, which can serve as a first step to characterize potential obesity-related pathophysiological pathways.