Proteomics-based computational modelling reveals distinct cardiac bioenergetic phenotypes in patients with ischaemic and dilated cardiomyopathy

Background

Ischaemic cardiomyopathy (ICM) and dilated cardiomyopathy (DCM) represent common causes of heart failure with distinct pathophysiological features. While the failing heart is characterised by energetic impairment, the extent to which myocardial metabolism is deranged due to the underlying aetiology remains incompletely defined.

Purpose

To compare myocardial metabolism in a multicentre cohort of patients with ICM, DCM and non-failing controls using proteomics-based computational modelling.

Methods

A total of 118 left ventricular biopsies (DCM, n=41; ICM, n=34; controls, n=43) were analysed using high-resolution mass spectrometry for quantitative proteomics. Patient-specific bioenergetic phenotypes were reconstructed by integrating the individual proteome profile into a comprehensive kinetic model capturing 296 biochemical reactions governing myocardial energy metabolism. The metabolic impact of changes in myocardial workload, circulating substrate availability and enzyme regulation were simulated, and differences across groups were examined by general linear models.

Results

Maximal ATP production capacity was reduced in ICM (mean diff=190.7; 95% CI=70.0, 311.5; P<0.001) and DCM (mean diff=169.4; 95% CI=54.6, 284.3; P=0.002) compared to controls (Figure 1A). This was associated with lower peak O2 consumption (P<0.01 for both diseases), whereas the ATP yield per mole of O2 consumed did not differ across groups.

Compared to ICM (mean diff=1.24; 95% CI=0.06, 2.42; P=0.04) and controls (mean diff=2.12; 95% CI=1.01, 3.23; P<0.001), DCM exhibited a lower fatty acid to glucose utilisation ratio (Figure 1B). The shift in substrate preference was caused by an increased glucose utilisation in DCM compared to ICM (mean diff=18.3; 95% CI=2.24, 34.4; P=0.02) and controls (mean diff=27.5; 95% CI=12.4, 42.7; P<0.001), whereas fatty acid use was decreased in both ICM (mean diff=52.2; 95% CI=11.3, 93.1; P=0.008) and DCM (mean diff=41.5; 95% CI=2.61, 80.4; P=0.03) compared to controls. Maximal glucose uptake capacity was higher in DCM hearts compared to ICM (mean diff=6.41; 95% CI=0.32, 12.5; P=0.04) and controls (mean diff=6.22; 95% CI=0.44, 12.0; P=0.03), providing a potential mechanistic link.

Conclusion

Patients with ICM and DCM shared a similar degree of cardiac energetic failure but demonstrated distinct patterns in the uptake and utilisation of energy-delivering substrates. Identification of patient-specific metabolic phenotypes through computational modelling has the potential to personalise diagnoses and guide metabolism-based therapies in heart failure.Figure 1