Intramuscular lipid droplets in type 2 diabetes and obesity: effects of exercise and training on subcellular distribution, morphology, and mitochondrial contact

In the skeletal muscle, the excessive storage of lipids, deposited as lipid droplets (LDs), hasmultiple detrimental effects on cellular functions and systems and is associated with thedevelopment of type 2 diabetes. However, LD morphology, subcellular distribution, andmitochondrial contact display a high level of heterogeneity, which may impede the currentunderstanding of lipid-induced insulin resistance. By employing transmission electronmicroscopy (TEM), this thesis aimed to provide a comprehensive analysis to investigateintramuscular LDs and mitochondria and associations with insulin sensitivity in patients withtype 2 diabetes and glucose-tolerant obese and lean controls. We further aimed to investigatethe effects of 1) eight weeks of high-intensity interval training (HIIT) targeting major musclegroups and 2) 60 min acute exercise on the morphological characteristics of LDs.

In study I, insulin-stimulated glucose disposal rate (GDR), as assessed by the hyperinsulinemic-euglycemic clamp, was ~40% lower in the patients with type 2 diabetes than in the obese and lean controls at baseline. After HIIT, insulin-stimulated GDR increased by ~30–40% in all three groups, accompanied by a clinically relevant decrease in HbA1c (4 ± 2 mmol/mol) and fasting plasma glucose (1.0 ± 0.4 mmol/l) in patients with type 2 diabetes. HIIT also increased V̇ O2max (~8–15%), decreased total fat mass, and increased lean body in all three groups. There were no correlations between these training adaptions nor group-specific differences in these responses.

In study II, the excess storage of intramuscular lipids in patients with type 2 diabetes was present as extremely large individual LDs situated in single muscle fibers poor of subsarcolemmal mitochondria. HIIT improved this intramuscular deficiency by redistributing LD size and subcellular distribution while increasing mitochondrial volumetric content. Morphological shape analysis further showed that individual LDs were better described as ellipsoids than spheres. Moreover, analysis of contact between LDs and mitochondria revealed an altered interaction between organelles in insulin-resistant conditions. However, there were no robust correlations between morphological LD analysis and insulin sensitivity.

In study III, acute exercise did not mediate changes in LD volumetric content, numerical density, or profile size. However, when evaluating the fraction of LDs touching mitochondria and the magnitude of inter-organelle contact, exercise increased the LD-mitochondrial contact irrespective of obesity or type 2 diabetes. This effect was profound in the subsarcolemmal region of type 1 fibers with an increased contact length from ~275 to ~420 nm. Interestingly, the absolute contact length before exercise (ranging from ~140 to ~430 nm) was positively associated with fat oxidation rates during exercise.

Together, results from study I, II, and III demonstrates that type 2 diabetes displays cellular heterogeneity in intramuscular LD storage, highlighting the relevance of single-cell modalities in clinical research. HIIT changed this muscle LD storage toward non-diabetic characteristics while efficiently improving insulin sensitivity with intact responses in all three groups. Although this is too early to state, our data further suggest that the increased contact between LDs and mitochondria with acute exercise is not disturbed in obesity or type 2 diabetes.