Skeletal muscle mitochondrial function: is it quality or quantity that makes the difference in insulin resistance?

Mitochondria are the combustion engines of life, producing ATP and CO₂ at the expense of nutrient substrates and molecular O₂. Given its role in locomotion, skeletal muscle ATP production can increase by more than 100-fold during near maximal contraction. Since ATP is not stored within skeletal muscle in great abundance (∼20–25 mmol(kg dry muscle)⁻¹), during heightened metabolic demand (i.e. contraction) there is an essential requirement for rapid and sustained ATP production, a role fulfilled primarily by mitochondria. As such, skeletal muscle cells are densely populated with these complex organelles.

With an increasing prevalence of inactivity and obesity, type 2 diabetes has become a major public health concern. Given their central role in energy production, it is perhaps not surprising that mitochondrial dysfunction has been implicated in the aetiology of skeletal muscle insulin resistance, a precursor of frank diabetes. Furthermore, interventions which stimulate skeletal muscle mitochondrial biogenesis can improve insulin sensitivity in obese and/or insulin resistant individuals. While it seems clear that a reduction in skeletal muscle mitochondrial content is synonymous with insulin resistance, disagreement exists within the literature as to whether insulin resistance is accompanied by intrinsic alterations in skeletal muscle mitochondrial function. It is plausible that at least some of the disagreement as to whether insulin resistance is associated with intrinsic alterations in skeletal muscle mitochondrial function arises from (i) the different methodologies used to determine mitochondrial function in muscle samples and (ii) the surrogates of mitochondrial content used to normalise these functional measures.

In the first instance, the choice of making functional measures in isolated mitochondria or permeabilised myofibres may be crucial. For example, in the context of ageing, extraction from their host tissue alters the morphology and function of mitochondria when compared to those which remain in situ within myofibres, exaggerating the impact of senescence on mitochondrial function (Picard et al. 2010). Furthermore, measures of oxygen consumption or ATP production rates, whether determined in isolated organelles or permeabilised myofibres, are frequently normalised to mitochondrial content in order to give a qualitative index of mitochondrial function. Consequently, accurately determining mitochondrial content becomes crucial to any measure of mitochondrial function. While transmission electron microscopy (TEM) is generally accepted to be the most robust means of determining both the size and number of mitochondria within a skeletal muscle biopsy sample, such methodologies are not available to all researchers. Thus, a number of markers of mitochondrial content which can be determined biochemically are often used to infer mitochondrial content or normalise measures of mitochondrial function. Common surrogate measures of mitochondrial content include mtDNA copy number, cardiolipin content and maximal activities of enzymes such as citrate synthase (CS) or cytochrome c oxidase; however, which of these best represents the true mitochondrial content is not known. With this in mind, a better understanding of which proxy of mitochondrial content best correlates with morphological determinations of mitochondrial size and number will likely lead to a better understanding of the impact of numerous pathophysiological states on mitochondrial physiology, while allowing the more reliable comparison of data generated by independent laboratories.

In a recent issue of The Journal of Physiology, Larsen and colleagues (2012) used TEM to determine mitochondrial fractional area and high resolution respirometry (HRR) to determine maximal coupled respiration in skeletal muscle biopsies obtained from young healthy human volunteers with the aim of assessing the validity of a number of proxy measures of mitochondrial content and oxidative capacity (OXPHOS). A main finding of their work was that several surrogate measures of mitochondrial content correlated well with morphological measures of mitochondrial fractional area made by TEM. In particular, both content of cardiolipin (r= 0.86, P < 0.001), a phospholipid located in the inner mitochondrial membrane, and activity of CS (r= 0.84, P < 0.001), a key enzyme in the Krebs cycle, were strongly correlated with mitochondrial fractional area (Larsen et al. 2012). Intriguingly, these researchers were also able to show that mitochondrial respiration stimulated by the complex IV electron donor tetramethyl-benzenediamine dihydrochloride and ascorbate also correlated well with mitochondrial fractional area, suggesting that specific HRR protocols have the potential to give a measure of mitochondrial content which correlates well with TEM data. Perhaps of most interest though was the fact that mtDNA copy number did not correlate with morphological measures of mitochondrial content (r= 0.35), a finding which is perhaps surprising if not worrying given the fact that several studies have reported differences in skeletal muscle mitochondrial content and function when mtDNA copy number was used as a proxy of mitochondrial content.

The novel findings presented by Larsen and colleagues are instructive in a number of ways. For example, future research concerning mitochondrial bioenergetics should strive to make appropriate surrogate measures of mitochondrial content. For laboratories without access to TEM, both cardiolipin and CS activity appear to be valid measures of mitochondrial content, whereas mtDNA copy number seems a poor indicator of mitochondrial content.

An additional aim of the work of Larsen and colleagues was to identify biomarkers which correlate well with mitochondrial OXPHOS. OXPHOS was determined in permeabilised myofibres following the addition of glutamate (10 mm), malate (2 mm), ADP (5 mm) and succinate (10 mm). It was found that both succinate dehydrogenase (complex II) and cytochrome c oxidase (complex IV) enzyme activities were strongly correlated with maximal coupled respiration. Arguing that succinate, the direct substrate for complex II, exerts a strong control over the electron transport chain and thus OXPHOS, Larsen and colleagues conclude that complex IV activity is likely to be the more appropriate proxy of skeletal muscle OXPHOS. This particular finding is likely to be most relevant to researchers interested in skeletal muscle bioenergetics that do not have the capability to measure mitochondrial oxygen consumption or ATP production rates, particularly as complex IV activity can be determined relatively easily spectrophotometrically. Accordingly, determining complex IV activity in muscle homogenates provides a valid index of mitochondrial oxidative capacity.

In our view, the work presented by Larsen and colleagues offers a strong explanation as to why there appears to be disagreement in the literature as to whether insulin resistance is accompanied by intrinsic deficits in mitochondrial function within skeletal muscle. By way of example, Kelley et al. (2002) previously concluded that skeletal muscle mitochondria of insulin resistant type 2 diabetics have impaired oxidative capacity when compared to healthy lean individuals. In this study mitochondrial function was assessed by determining rotenone-sensitive NADH:O₂ oxidoreductase (complex I) activity normalised to protein content via creatine kinase (CK) activity (Kelley et al. 2002). However, in light of Larsen and colleagues data, it seems that complex I activity is actually a better proxy of mitochondrial content (r= 0.78, P= 0.002) rather than OXPHOS (r= 0.53, P= 0.048), at least in young healthy individuals (Larsen et al. 2012). Furthermore, Kelley and colleagues study also reported that CS activity (normalised to CK activity) was significantly lower in diabetic patients compared to lean controls, also pointing to a lower mitochondrial content (Kelley et al. 2002), Thus, it is probable that these researchers may, in fact, actually have detected quantitative deficits in skeletal muscle mitochondria in diabetic patients, as opposed to real changes in intrinsic function.

In contrast to the findings of Kelley and co-workers, Boushel and colleagues found no intrinsic deficits in mitochondrial function in patients with type 2 diabetes when compared to healthy volunteers (Boushel et al. 2007). While Boushel et al. showed that mitochondrial respiration in permeabilised myofibres was lower in type 2 diabetics compared to healthy controls, normalising respiratory rates to mitochondrial content eliminated these differences (Boushel et al. 2007). Although a weakness of this study was that mtDNA was used to normalise these functional data, the same researchers have since demonstrated that skeletal muscle mitochondrial respiration is not different between lean, obese and diabetic subjects when CS activity is used as a surrogate of mitochondrial content (Larsen et al. 2009). Taken together, the findings of the studies detailed above clearly highlight the need for researchers to select the most appropriate surrogate measures of OXPHOS and mitochondrial content when assessing mitochondrial function to avoid conflicting experimental results and therefore differing scientific conclusions.

In summary, whether the decline in skeletal muscle mitochondrial function associated with insulin resistance is the result of both a reduction in mitochondrial content and intrinsic deficits in mitochondrial function remains unclear. The different methodological approaches used to determine mitochondrial function and the various surrogates of mitochondrial content used to normalise mitochondrial function to mitochondrial content likely play a role in the lack of consensus in this area. Progress will be partially (or largely) dependent on researchers standardising outcome measures and selecting the most valid surrogates of mitochondrial content in their future experiments. Regarding the latter, Larsen and colleagues have made a valuable contribution in identifying surrogates which offer the best indication of true mitochondrial content and function. While these data were obtained from skeletal muscle taken from young, healthy individuals, until equivalent data is available in various patient groups, this laudable work should direct future research concerning skeletal muscle mitochondrial physiology in a number of settings such as insulin resistance, ageing, sepsis and burn injury.