Assessing Multiple Features of Mitochondrial Function

Alterations of energy metabolism are key features of insulin resistance, diabetes, and diabetes-related complications. However, the role of mitochondria—the cell’s power plants—in abnormal metabolism is less certain. Mitochondria are responsible for oxidative phosphorylation, which couples substrate oxidation to synthesize ATP. Essentially, the reduction of dinucleotides enables electrons to transfer via the electron transport system across the inner mitochondrial membrane. Energy released from electron transfer is used for proton transport into the matrix, thereby establishing an electrochemical proton gradient (ΔΨ) (1). Mitochondrial function involves various features, each of which can be examined by different methods. The use of different approaches to assess mitochondrial function, coupled with imprecise terminology regarding “mitochondrial dysfunction” can lead to confusing data interpretation.

Several studies have reported lower mitochondrial function expressed per muscle mass in type 2 diabetes as assessed by enzyme activities (2,3) or by high-resolution respirometry (4–7). Although reduced mitochondrial content could serve as explanation for this observation (3,5,8), some studies have also noted lower intrinsic function (4,6,7,9)—that is, reduced respiratory capacity per mitochondrion. Mitochondrial production of reactive oxygen species (ROS) also leads to the formation of lipid peroxides. In turn, these can induce oxidative stress thereby causing cellular and mitochondrial damage (10,11). In this context, it is of interest to simultaneously monitor ΔΨ, ATP, and ROS production in a single tissue sample.

In this issue of Diabetes, Yu et al. (12) describe a method for evaluating the physiology of isolated mitochondria. They used 2-deoxyglucose (2DOG) clamps, which have been used before to assess mitochondrial-bound hexokinase activity (13). The 2DOG clamp is based on the presence of an excess of hexokinase and 2DOG to clamp ΔΨ at defined levels and to quantify ATP production from the conversion of 2DOG to 2DOG-phosphate (2DOGP). Yu et al. report that in vitro ¹H- and ¹H/¹³C-magnetic resonance spectroscopy (MRS) provides 30–40-fold greater sensitivity to detect ATP production than classical in vitro ³¹P-MRS. The method’s strength resides in its ability to measure both ATP and ROS production at fixed ΔΨ. Moreover, the conversion of 2DOG to 2DOGP is an irreversible process occurring in the presence of exogenously added ADP, which effectively controls ΔΨ. The addition of ADP increases mitochondrial energy demand, not supply, a consideration that is important for detailed evaluation of mitochondrial oxidative capacity. Indeed, this method measures oxidative capacity (not resting mitochondrial function [14]) as ATP production transits from state 4 (i.e., no ATP demand) to state 3 respiration (i.e., high ATP demand by stimulating oxidative phosphorylation coupling to energy production). Although the use of mitochondria deprived from their natural environment is a clear limitation, it also circumvents the influence of substrate supply due to changes in local perfusion and substrate uptake. Studies in humans have tried to quantify intact mitochondrial function in vivo by ³¹P-MRS, by either applying the magnetization transfer experiment to assess ATP synthase flux with or without insulin stimulation (15,16) or using the postexercise creatine-phosphate (PCr) recovery rate to assess maximal oxidative capacity (4,17). Obviously, these methods cannot evaluate intrinsic mitochondrial features (14).

Using this novel approach, Yu et al. (12) measured ΔΨ, ATP, and ROS production in muscle of streptozotocin-diabetic rats without (STZ-DM) and with insulin treatment (STZ-INS). The STZ-DM had decreased ATP production and greater ROS production per generated ATP molecule at a lower ΔΨ threshold. Whether these findings hold true to for other diabetes models or human type 1 diabetes, needs to be tested. Of note, the STZ-DM rat exhibits β-cell failure without immunological pathogenesis. We recently reported that muscle mitochondrial function also decreases during development of autoimmune diabetes in nonobese diabetic mice (18). Rising blood glucose and lipolysis indicate that glucolipotoxity could possibly underlie abnormal mitochondrial function and insulin resistance also in type 1 diabetes (Fig. 1). Interestingly, Yu et al. showed that the abnormal mitochondrial function improved with prolonged insulin treatment in STZ-INS (12). In near-normoglycemic patients with type 1 diabetes, muscle ATP synthesis is normal during fasting but impaired during insulin stimulation (15) or deprivation (19). One might speculate that this is due to their long-standing disease with previous extended periods of hyperglycemia and lipolysis. As shown in Fig. 1, elevated lipolysis could lead to oxidative and endoplasmic reticulum stress and accumulation of lipotoxic intermediates, such as diacylglycerol, ceramides, or acylcarnitines, thereby contributing to insulin resistance in type 1 diabetes.

The effect of insulin on mitochondria and ROS is also of interest for obesity and type 2 diabetes because oxidative stress can activate c-Jun N-terminal kinase (JNK) via nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) and protein kinase C (PKC). Both of these have been linked to insulin resistance (20,21) (Fig. 1). Indeed, insulin can stimulate muscle mitochondrial function in lean healthy subjects but not in type 2 diabetic patients (16,22). However, ROS production was not measured in these studies. Reduction of H₂O₂ production by lipid lowering likely improves insulin sensitivity by affecting mitochondrial oxidative capacity in type 2 diabetes (23), which renders the relationship between oxidative stress and insulin action uncertain.

Finally, Yu et al. (12) describe three processes for the generation and use of ΔΨ: 1) generation via proton pumping via the electron transport system complexes, 2) consumption through proton leaks, and 3) ATP synthesis. However, the ΔΨ driving force also significantly depends on the cellular redox environment (24). Moreover, the ratio of reduced glutathione to the oxidized disulfide form of glutathione has recently been suggested to link mitochondrial hydrogen peroxide emission to the control of redox-sensitive phosphatases that may target insulin signaling (24,25) (Fig. 1). Mitochondria are plastic organelles, rapidly adapting to environmental changes (14,26) with altered cellular redox states. Whether this affects insulin signaling or ROS production in the STZ-INS requires further investigation.

Taken together, the new methodology for measuring mitochondrial ATP production from titration of ADP levels at fixed ΔΨ values is a promising tool to explore the interaction between oxidative pathways and insulin action in the fields of insulin resistance, type 1, and type 2 diabetes.