Long-Chain Acylcarnitines and Monounsaturated Fatty Acids Discriminate Heart Failure Patients According to Pulmonary Hypertension Status

Abstract

Defects in fatty acid (FA) utilization have been well described in group 1 pulmonary hypertension (PH) and in heart failure (HF), yet poorly studied in group 2 PH. This study was to assess whether the metabolomic profile of patients with pulmonary hypertension (PH) due HF, classified as group 2 PH, differs from those without PH. We conducted a proof-of-principle cross-sectional analysis of 60 patients with chronic HF with reduced ejection fraction and 72 healthy controls in which the circulating level of 71 energy-related metabolites was measured using various methods. Echocardiography was used to classify HF patients as noPH-HF (n = 27; mean pulmonary artery pressure [mPAP] 21 mmHg) and PH-HF (n = 33; mPAP 35 mmHg). The profile of circulating metabolites among groups was compared using principal component analysis (PCA), analysis of covariance (ANCOVA), and Pearson's correlation tests. Patients with noPH-HF and PH-HF were aged 64 ± 11 and 68 ± 10 years, respectively, with baseline left ventricular ejection fractions of 27 ± 7% and 26 ± 7%. Principal component analysis segregated groups, more markedly for PH-HF, with long-chain acylcarnitines, acetylcarnitine, and monounsaturated FA carrying the highest loading scores. After adjustment for age, sex, kidney function, insulin resistance, and N-terminal pro-brain natriuretic peptide (NT-proBNP), 5/15 and 8/15 lipid-related metabolite levels were significantly different from controls in noPH-HF and PH-HF subjects, respectively. All metabolites for which circulating levels interacted between group and NT-proBNP significantly correlated with NT-proBNP in HF-PH, but none with HF-noPH. FA-related metabolites were differently affected in HF with or without PH, and may convey adverse outcomes given their distinct correlation with NT-proBNP in the setting of PH.

Keywords: acylcarnitines; fatty acids; heart failure; type 2 pulmonary hypertension.

Figure 1

Principal component analysis identified distinct metabolic patterns in PH-HF (n = 33) and noPH-HF (n = 27) and segregated HF patients from healthy subjects (n = 72). A total of 55 variables were included in the analysis and comprised usual biochemical parameters as well as various metabolites measured by a combination of MS-based metabolomics approach targeting fatty acids, acylcarnitines, organic acids and amino acids. For the segregation between healthy subjects (controls) and HF patients, principal component 1 (PC1) and principal component 2 (PC2) accounted for 15.7% and 10.1% of the total variation, respectively. Controls are identified in red, noPH-HF in green and PH-HF in blue. Biplot analysis merging PCA plot and loadings plot identified the most potent metabolites in segregating controls (red) and noPH-HF (green) with PC1 and PC2 accounting for 11.9% and 9.3% respectively, and controls (red) and PH-HF (blue) with PC1 and PC2 accounting for 16.8% and 11.9% respectively. The corresponding loading scores in PC1 and PC2 for both analyses are reported in Table S1.

Figure 2

Acylcarnitines and fatty acids circulating levels are differentially affected in PH-HF (n = 33) and noPH-HF (N = 27) compared to controls (N = 72). Box plots represent top discriminant FA-related metabolites identified using the PCA analysis in the comparison noPH-HF (green) vs. controls (red) and PH-HF (blue) vs. controls. Shown are semi-quantitative analyses of (A) acylcarnitines reported as MS signal ratio normalized to standard, and the quantitative analysis of (B) saturated FA, (C) MUFA, and (D) PUFA. In the boxplots, rectangles represent the SD, the segment inside the rectangle the median and the whiskers above and below the maximum and minimum. * p < 0.05, ** p < 0.01, *** p < 0.001 compared to controls; ^$ p < 0.05, ^$$ p < 0.01 PH-HF vs. noPH-HF. Other more commonly measured metabolites are shown in Figure S1.

Figure 3

Acylcarnitines and fatty acids significantly correlate with NT-proBNP in PH-HF but not in noPH-HF patients. Pearson correlation analysis between NT-proBNP and (A) acylcarnitines, (B) FA (MUFA and PUFA) in noPH-HF (black) or PH-HF (blue) patients. The dotted lines indicate the 95% confidence intervals.