Differential insulin sensitivity of NMR-based metabolomic measures in a two-step hyperinsulinemic euglycemic clamp study

Background: Insulin is the key regulator of glucose metabolism, but it is difficult to dissect direct insulin from glucose-induced effects. We aimed to investigate the effects of hyperinsulemia on metabolomic measures under euglycemic conditions in nondiabetic participants.

Methods: We assessed concentrations of 151 metabolomic measures throughout a two-step hyperinsulinemic euglycemic clamp procedure. We included 24 participants (50% women, mean age = 62 [s.d. = 4.2] years) and metabolomic measures were assessed under baseline, low-dose (10 mU/m²/min) and high-dose (40 mU/m²/min) insulin conditions. The effects of low- and high-dose insulin infusion on metabolomic measures were analyzed using linear mixed-effect models for repeated measures.

Results: After low-dose insulin infusion, 90 metabolomic measures changed in concentration (p < 1.34e^-4), among which glycerol (beta [Confidence Interval] = – 1.41 [- 1.54, – 1.27] s.d., p = 1.28e^-95) and three-hydroxybutyrate (- 1.22 [- 1.36, – 1.07] s.d., p = 1.44e^-61) showed largest effect sizes. After high-dose insulin infusion, 121 metabolomic measures changed in concentration, among which branched-chain amino acids showed the largest additional decrease compared with low-dose insulin infusion (e.g., Leucine, – 1.78 [- 1.88, – 1.69] s.d., P = 2.7e^-295). More specifically, after low- and high-dose insulin infusion, the distribution of the lipoproteins shifted towards more LDL-sized particles with decreased mean diameters.

Conclusion: Metabolomic measures are differentially insulin sensitive and may thus be differentially affected by the development of insulin resistance. Moreover, our data suggests insulin directly affects metabolomic measures previously associated with increased cardiovascular disease risk.