Cardiac dysfunction and peri-weaning mortality in malonyl-coenzyme A decarboxylase (MCD) knockout mice as a consequence of restricting substrate plasticity

Inhibition of malonyl-coenzyme A decarboxylase (MCD) shifts metabolism from fatty acid towards glucose oxidation, which has therapeutic potential for obesity and myocardial ischemic injury. However, ~ 40% of patients with MCD deficiency are diagnosed with cardiomyopathy during infancy. Aim To clarify the link between MCD deficiency and cardiac dysfunction in early life and to determine the contributing systemic and cardiac metabolic perturbations. Methods and results MCD knockout mice (−/−) exhibited non-Mendelian genotype ratios (31% fewer MCD−/−) with deaths clustered around weaning. Immediately prior to weaning (18 days) MCD−/− mice had lower body weights, elevated body fat, hepatic steatosis and glycogen depletion compared to wild-type littermates. MCD−/− plasma was hyperketonemic, hyperlipidemic, had 60% lower lactate levels and markers of cellular damage were elevated. MCD−/− hearts exhibited hypertrophy, impaired ejection fraction and were energetically compromised (32% lower total adenine nucleotide pool). However differences between WT and MCD−/− converged with age, suggesting that, in surviving MCD−/− mice, early cardiac dysfunction resolves over time. These observations were corroborated by in silico modelling of cardiomyocyte metabolism, which indicated improvement of the MCD−/− metabolic phenotype and improved cardiac efficiency when switched from a high-fat diet (representative of suckling) to a standard post-weaning diet, independent of any developmental changes. Conclusions MCD−/− mice consistently exhibited cardiac dysfunction and severe metabolic perturbations while on a high-fat, low carbohydrate diet of maternal milk and these gradually resolved post-weaning. This suggests that dysfunction is a common feature of MCD deficiency during early development, but that severity is dependent on composition of dietary substrates.

Absorbance was recorded at 510 nm. A standard curve covering 1-5 mmol/L was used for the determining the sample concentrations.
In a separate series of experiments, tissue lipids were extracted from KO, WT and Het mouse tissue samples (LV, skeletal muscle, liver). Powdered tissue was homogenised in ice-cold chloroform-methanolwater mixture (2:1:0.8) and centrifuged for 15 min (13 500 rpm, 4°C). The upper aqueous layer was decanted, the lower chloroform layer dried under 100% gaseous N 2 and re-suspended in 2-propanol.

Real-time quantitative Reverse Transcriptase-Polymerase Chain Reaction
Total RNA (1ng) was used as input in one-step RT and amplification reactions using the Qiagen Quantitect SYBR Green RT-PCR kit (Qiagen, UK) on the Rotor-Gene system (Corbett Research Ltd, Qiagen, UK).
The oligonucleotide sequences are listed in Table 1. Either the sense or the antisense oligonucleotide per pair, were designed to span intron-exon boundaries on the cDNA sequence. For data analysis, the doublestandard curve method was employed, in which standard curves spanning five log dilutions of heart RNA were constructed for both the reference and the genes of interest. For quantification, the relative quantities of the above genes were normalized against the reference gene 36B4.

CardioNet metabolic network reconstruction
Flux distributions were predicted for different nutritional supplies while making the assumption that in an environment with restricted access to resources cardiac metabolism is following the concept of optimality.
Optimal solutions were calculated for a limited set of 10 nutrients while demanding a complex metabolic target function reflecting important cellular functions of the cardiomyocyte as previously described [2]. MCD knockout was simulated by restricting the corresponding network reaction to carry a zero flux.
Constraints were applied to the exchange of substrates to reflect the diet depending variability of available nutrients, such as saturated and unsaturated, medium and long-chain fatty acids, glucose, ketone bodies, lactate and pyruvate.
The dietary composition for fatty acids, glucose and ketone bodies were added according to experimentally obtained plasma concentrations for high-fat diet ( Figure 5). As the plasma metabolite concentrations were not available under low-fat diet conditions, substrate availability were constrained according to the dietary composition of normal laboratory mouse chow fed to adult MCD animals (Table 2 online supplement).
Flux balance analysis was used to simulate MCD deficiency using a PERL based algorithm in conjunction with the solver CPLEX. Cardiac efficiency was calculated based on (i) the oxygen demand and (ii) endogenous glucose demand and (iii) total substrate uptake as previously described [2]. Flux distributions were analyzed in R statistics and Cytoscape [3].  Attempted Echo, but died within minutes of anaesthesia (no images obtained). Liver had nutmeg appearance. Data is mean ± SEM. CSA is cross-sectional area.* denotes P < 0.05 between MCD +/+ and MCD -/by oneway ANOVA with Bonferroni's correction for multiple comparisons. Data is mean ± SEM. There are no significant differences for any parameter using one-way ANOVA.