Mitochondrial adaptation in obesity is a ClpPicated business

Abstract

Quality control systems that maintain mitochondrial oxidative phosphorylation (OXPHOS) include rescue by mitochondrial fusion, elimination of dysfunctional mitochondria by mitophagy, and degradation of damaged proteins by proteases. ClpP is an ATP‐dependent protease located in the mitochondrial matrix and mutated in Perrault syndrome, causing gonadal atrophy and hearing loss. Given that hearing loss is common in mitochondrial diseases caused by mtDNA mutations, ClpP was proposed to be part of the quality control system to maintain proper mitochondrial OXPHOS function. Two recent studies independently report that deletion of ClpP in mice protects from insulin resistance and obesity by increasing mitochondrial OXPHOS capacity and browning in gonadal white adipose tissue and mitochondrial coupling in brown adipose tissue 1, 2. Furthermore, liver‐ and muscle‐specific deletion of ClpP has no major effects on insulin resistance. These studies reveal that ClpP might be involved in tissue‐specific mitochondrial remodeling in response to metabolic demands, rather than exclusively removing damaged proteins to maintain OXPHOS capacity.

Mitochondrial oxidative phosphorylation (OXPHOS) is maintained by a number of different quality control mechanisms that facilitate mitochondrial turnover and functional regeneration. With such a long list of redundant systems involved in quality control, it is surprising that mitochondrial dysfunction can even occur. However, it is becoming increasingly clear that mitochondrial proteins executing quality control have other roles not related to the prevention of mitochondrial dysfunction 3. The lack of understanding of these additional roles might be the underlying reason behind our inability to explain why a reduction in mitochondrial function in some cases is compensatory, delaying the outcome of metabolic and age‐related diseases, whereas in other cases worsens the outcome of these diseases. Accordingly, one could hypothesize that a reduction in mitochondrial function induced by an alteration and/or malfunction of the quality control mechanisms is more likely to contribute to disease. On the other hand, a reduction in mitochondrial function achieved by a normal cellular response and with the mitochondrial quality control mechanisms functioning adequately is most likely to be compensatory.

One of these mitochondrial quality control components is the caseinolytic peptidase P (ClpP). ClpP is a protease located in the mitochondrial matrix highly conserved from bacteria to C. elegans and mammals. Mutations in human ClpP are associated with Perrault syndrome and cause sensorineural hearing loss (SNHL) in males and females, and ovarian dysfunction in females 4. As SNHL is present in mitochondrial diseases caused by mtDNA mutations reducing OXPHOS capacity 5, it was deemed that mutations in ClpP cause mitochondrial OXPHOS dysfunction. These findings also suggested that ClpP activity removing damaged proteins in the mitochondria might not be efficiently compensated in tissues highly dependent on OXPHOS, leading to tissue‐specific deterioration (i.e., neurons). These conclusions were partially reinforced by the generation of ClpP‐deficient mice, which recapitulated the symptoms of Perrault syndrome. However, the respiratory defects in ClpP knockout (ClpP KO) mice were concluded to be minor, with a strong inflammatory component attributed to the accumulation of mitochondrial components 6.

Later work indeed demonstrated that ClpP function is not exclusively linked with OXPHOS function maintenance by removing dysfunctional proteins. ClpP deletion was shown to ameliorate the deleterious consequences of dysregulating mitochondrial translation induced by deleting mitochondrial aspartyl aminoacyl‐tRNA synthetase. As a consequence, these results demonstrated that ClpP activity is actively involved in regulating mitochondrial translation and adaptation to metabolic demands, discarding that ClpP is exclusively acting as a disposal unit to degrade dysfunctional proteins and/or regulating the mitochondrial unfolded protein response in mammals 7, 8.

In light of these additional roles of ClpP, Becker et al 1 and Bhaskaran et al 2 report that ClpP KO mice are protected from diet‐induced obesity and insulin resistance, diseases that remodel mitochondria in liver, muscle, and adipose tissue. These three tissues are indeed major contributors to hyperglycemia associated with systemic insulin resistance. On the other hand, tissue‐specific ClpP deletion in muscle or in liver did not have any protective effects against obesity, hyperglycemia, and insulin resistance and did not induce growth retardation 1. These latter experiments performed by Becker et al excluded the effects of missing ClpP activity on hepatic and muscle mitochondria as drivers of improved insulin sensitivity, as well as being the drivers of growth retardation in whole body ClpP KO mice. The absence of improved insulin sensitivity in these tissue‐specific ClpP KO mice could potentially exclude the secretion of Fgf21 or other molecules from muscle and liver as drivers improving systemic glucose metabolism in ClpP KO mice, as reported for models of muscle‐ and liver‐specific mitochondrial dysfunction.

Instead, both Becker et al and Bhaskaran et al report an increase in mitochondrial OXPHOS capacity and browning in white adipose tissue (WAT), as the mechanism driving improved insulin sensitivity in ClpP mice (Fig 1) 1, 2. Interestingly, the WAT deposit showing the largest increase in mitochondrial mass and function was gonadal, which is the deposit less susceptible to browning, when compared to inguinal. This selective effect on different deposits suggests that ClpP deletion induces a remodeling of mitochondrial function within WAT, rather than just recruiting beige adipocytes.

Figure 1. The role of the mitochondrial protease ClpP in obesity and insulin resistance.

Mice with the mitochondrial protease ClpP inactivated in all tissues (ClpP ^−/− mice) are protected from diet‐induced obesity and insulin resistance, while being cold intolerant. This protection is associated with increased mitochondrial oxidative phosphorylation capacity (OXPHOS) in white adipose tissue (WAT) and healthy whitening of brown adipose tissue (BAT). ClpP KO BAT harbors reduced expression of thermogenic Ucp1 but improved OXPHOS efficiency. The absence of ClpP increases total levels of the mitochondrial matrix protein VLCAD (very long‐chain acyl‐CoA dehydrogenase) in BAT and WAT, an enzyme that catalyzes fatty acid oxidation.

While both studies do not present the effects of an adipose tissue‐specific ClpP knock out, the results shown suggest that ClpP deletion within adipose tissue, and not in other tissues, is responsible for the mitochondrial changes in WAT and brown adipose tissue (BAT). Our rationale behind this conclusion is that BAT and WAT remodeling in ClpP KO mice resemble the effects of BAT‐specific deletion of the mitochondrial quality control and fusion protein, mitofusin 2 (Mfn2) 9. Becker et al showed that ClpP KO BAT is hypertrophied, accumulates more lipids, and reduces Ucp1 expression in chow diet‐fed mice, as reported in BAT‐specific Mfn2 knockout (BAT‐Mfn2 KO) mice. In addition, both Mfn2‐ and ClpP‐deficient BAT mitochondria show improved mitochondrial ATP synthesis, with ClpP KO mitochondria only showing improved efficiency and Mfn2 KO mitochondria showing improved capacity. Furthermore, this BAT mitochondrial remodeling promoting coupled respiration, expanding BAT, and inducing cold intolerance is associated with resistance to diet‐induced obesity and hyperglycemia both in BAT‐Mfn2 KO and ClpP KO mice. These similarities mean that we can add ClpP inhibition to the list of BAT manipulations that promotes healthy whitening of BAT to prevent hyperglycemia, likely driven by coupled fat oxidation to sustain the ATP demand of hypertrophied BAT.

Another aspect in common between the ClpP KO mice and BAT‐Mfn2‐KO is the paradoxical result of increased mitochondrial OXPHOS activity, while the protein levels of certain complex I subunits are decreased 2, 9. Bhaskaran et al report a marked decrease in Ndufs1 protein levels in gonadal WAT, despite complex I‐driven respiration was unchanged and complex II‐driven respiration was increased, even after high‐fat diet (HFD) feeding. We observed the same pattern in Mfn2 KO mitochondria isolated from BAT of HFD‐fed females, an increased complex II‐driven respiration, no changes in complex I‐driven respiration, and a marked decrease in complex I subunit Ndufb8. The mitochondrial remodeling induced by ClpP and Mfn2 deletion toward coupled respiration (i.e., OXPHOS instead of thermogenesis) can explain the reduction in complex I protein subunits. Coupled respiration fueled by fatty acids was shown to saturate the CoQ pool and induce reverse electron transport chain (RET), which together triggers complex I protein degradation 10.

Therefore, we can conclude that, as a consequence of ClpP deletion, mitochondria are remodeled in gonadal WAT and BAT to promote coupled fat oxidation, which is associated with resistance to obesity and hyperglycemia. Accordingly, ClpP expression is decreased in obesity, suggesting that ClpP reduction in adipose tissue is an adaptive response to prevent hyperglycemia 2. Mechanistically, how would deleting a mitochondrial protease induce this mitochondrial remodeling? In this regard, Becker et al identify that removing ClpP markedly increases the levels of the protein responsible of oxidizing very long‐chain fatty acids inside the mitochondria, VLCAD, both in WAT and in BAT 1. It is tempting to speculate that ClpP deletion promotes coupled fat oxidation by increasing VLCAD, but at the same time promoting the degradation of Cpt2 and Ucp1. Ucp1 uncouples mitochondrial fat oxidation from ATP synthesis to generate heat and is selectively expressed in BAT, beige adipocytes, and browned WAT. Cpt2 is also located in the inner membrane and catalyzes a non‐rate‐limiting step of long‐chain fatty acid oxidation. More specifically, Cpt2 is required to transform the fatty‐acyl‐carnitine to fatty‐acyl‐CoA in the mitochondrial matrix, an essential step for enzymes such as VLCAD to start fat oxidation. This compensatory mechanism degrading Cpt2 and Ucp1 to auto‐limit fat oxidation would only make sense if mitochondria are coupled, meaning that fat oxidation is limited by the ATP demand and synthesis. Indeed, excessive fuel provision leads to mitochondrial membrane hyperpolarization in coupled mitochondria, resulting in increased ROS production and reductive stress. Consequently, it would be interesting to test whether antioxidants or FCCP could restore Ucp1 and Cpt2 levels.

In summary, these studies reveal that a protease involved in quality control has additional roles to regulate mitochondrial protein composition in a tissue‐specific manner and to adjust mitochondria to different metabolic demands. Since systemic abrogation of ClpP causes Perrault syndrome in humans, adipose tissue‐specific inhibition of ClpP might be a potential strategy to prevent insulin resistance in obesity, as it enhances both healthy whitening of BAT and coupled mitochondrial oxidation in gonadal WAT.

EMBO Reports (2018) 19: e46295