Metabolomics and biochemical evaluation of skeletal muscle from Ndufs4 knockout mice
Abstract
The dysfunction of mitochondrial complex I (CI) impedes the most efficient mechanism feeding electrons into the respiratory chain (RC) — a system that, together with ATP synthase (CV), is responsible for the majority of cellular energy production via oxidative phosphorylation (OXPHOS). Although CI deficiency is the most common defect in mitochondrial energy metabolism, it is among the most complex, multifactorial, poorly understood disorders that currently lack an effective treatment. One of the most promising tools used to gain insight into this form of mitochondrial disease (MD), is the whole-body Ndufs4 knockout (Ndufs4/) mouse model. Although the neurological phenotype of these animals has been widely studied, the effect of CI deficiency on Ndufs4/skeletal muscle metabolism remains elusive. The lack of research on this tissue is largely owed to the view that these animals do not display strong muscle involvement. However, neuromuscular involvement is a hallmark feature of MD and considering skeletal muscle’s high energetic demand, metabolic activity, and integration with the nervous system, this tissue needs to be investigated. The aim of this study was, therefore, to combine hypothesis-generating metabolic profiling and biochemical strategies to gain insight into the energy metabolism of both glycolytic (white quadriceps) and oxidative (soleus) skeletal muscles from Ndufs4/mice. Profiling methods that utilise various high-content analytical platforms were employed in order to generate a broad metabolic phenotype of CI deficient muscles. This multi-platform metabolomics approach comprised of targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS), untargeted gas chromatography time-of-flight mass spectrometry (GC-TOF-MS) and proton nuclear magnetic resonance (1H-NMR) spectroscopy. However, extensive multi-platform metabolomics of mouse tissues presents a challenge due to the limited-quantity samples obtained — especially in the case of NMR spectroscopy, which is inherently insensitive. Therefore, one of the chief objectives of the study was to develop a 1H-NMR method that enables the analysis of small-quantity biological samples. In the first part of the study, we present a novel miniaturised 1H-NMR method utilising 2 mm NMR tubes, which enables metabolic profiling on a tenth of the sample quantity required by a well-established standard operating protocol (SOP). We demonstrate the miniaturised method’s acceptability regarding precision (CV < 15%), relative accuracy (80–120 %), linearity (R2 > 0.95) and statistical equivalence (p < 0.05) to the SOP when analysing spiked synthetic urine as well as mouse muscle extracts. In addition, we exhibit the novel method’s advantages for large-scale metabolomics when; i) adequate sample quantities are available by analysing increasingly concentrated versions (up to 10×) of samples to expand metabolome coverage; or ii) sample quantities are limited by performing a pilot metabolomics study on minute Ndufs4/(n = 3) and wild-type (WT; n = 3) solei, which identified five metabolites (previously linked to MD) strongly discriminating the two genotypes. In the second part of the study, we investigate the effects of CI deficiency in Ndufs4/skeletal muscles. Through multi-platform metabolic profiling of Ndufs4/(n = 19), and WT (n = 20) white quadriceps and soleus muscles, we provide the first empirical evidence of adaptive responses to CI dysfunction involving non-classical pathways that fuel the ubiquinone (Q)-cycle. By restoring the electron flux to CIII via the Q-cycle, these adaptive mechanisms could maintain adequate oxidative ATP production, despite CI deficiency — providing a possible explanation for the lack of muscle involvement in the Ndufs4/phenotype. We report a respective 48 and 34 discriminatory metabolites between Ndufs4/, and WT white quadriceps and soleus muscles, among which the most prominent alterations indicate the involvement of the glycerol-3-phosphate shuttle, the electron transfer flavoprotein system, respiratory chain CII, and the proline cycle in fuelling the Q-cycle. Enzyme (CI-CIV and CS) activity assays confirmed severely reduced (80 %) CI activity in both Ndufs4/(n = 12) muscle types, compared to WTs (n = 10), along with moderate reductions in CS (12 %) and CIII (18 %) activities in Ndufs4/solei. When comparing muscle fibre types, glycolytic fibres seemed to be more vulnerable to CI deficiency, as greater disturbances in metabolic profiles were evident, along with much lower
residual CI activity (4.58 ± 1.67 nmol/min/mg) compared to oxidative fibres (14.74 ± 6.67 nmol/min/mg). Taken together, this study contributes to the natural science field in two ways. Firstly, through our novel miniaturised 1H-NMR method, we provide a cost-effective alternative solution for the current restrictions of NMR spectroscopy in the analysis of limited-quantity biological samples. Traditionally, cryoprobe technology is required for such studies; however, we show that a typical NMR spectrometer with a standard probe head can be used to analyse small sample quantities with adequate analytical efficiency. Secondly, through discover-phase multi-platform metabolic profiling, we provide novel mechanistic insight into CI deficiency — thereby highlighting the value of metabolomics in MD research. We report skeletal muscle-specific changes in several metabolic pathways that result from whole-body mitochondrial dysfunction, which upon further investigation could provide novel targets for therapeutic intervention in CI deficiency and potentially lead to the development of new treatment strategies.