Integrated Myocardial and Plasma Lipidome Across the Human Obesity–Heart Failure Spectrum

Technical Summary: Integrated Myocardial and Plasma Lipidome Across the Human Obesity–Heart Failure Spectrum

Problem Statement
Heart failure with preserved ejection fraction (HFpEF) has become the dominant heart failure phenotype, increasingly associated with marked obesity, metabolic syndrome, and systemic inflammation. While a prevailing "lipotoxicity" hypothesis suggests that obesity drives myocardial lipid overload, triggering HFpEF through ceramide-mediated injury and mitochondrial dysfunction, this model relies heavily on animal data. There is a critical lack of rigorous human myocardial lipidomic data across the obesity spectrum, particularly in the presence or absence of heart failure. Furthermore, it remains unknown whether severe obesity alone induces analogous lipid abnormalities in humans or if common preclinical animal models accurately recapitulate the human disease lipidome. Prior human studies were limited by small control groups, narrow lipid class coverage, and a lack of comparison to heart failure with reduced ejection fraction (HFrEF).

Methodology
This study performed a comprehensive, untargeted lipidomic analysis of ventricular samples from 82 phenotyped human individuals, categorized into three groups: non-failing (NF) controls (n=40), HFpEF (n=29), and HFrEF (n=13). The NF cohort spanned a broad Body Mass Index (BMI) range (13.7–51.4 kg/m²), allowing for the interrogation of obesity effects independent of heart failure.

• Lipidomics: Using high-throughput LC-MS/MS, the researchers quantified over 1,000 lipid species across 21 major classes, including phospholipids, sphingolipids, neutral lipids, free fatty acids (FFAs), and acylcarnitines.
• Integration: Myocardial lipidomics were integrated with myocardial transcriptomics, ultrastructural analysis (transmission electron microscopy), and plasma lipidomics from a subset of the same cohorts.
• Animal Model Comparison: The human lipidomic signatures were compared against myocardial tissues from four distinct animal HFpEF models (HFD+L-NAME mice, SAUNA mice, ZSF-1 obese rats, and Göttingen minipigs) and HFrEF models (trans-aortic constriction mice), using the same lipidomic assay.
• Statistical Analysis: Differential abundance was assessed using Mann–Whitney U tests with Benjamini–Hochberg false discovery rate (FDR) correction. Pathway activity was inferred using LipidSig 2.0, mapping lipid species onto biosynthetic networks.

Key Results

1. Distinctive HFpEF Lipid Accumulation: HFpEF myocardium exhibited a massive, asymmetric accumulation of lipids. Approximately 60% of altered lipid species were increased, compared to only 1% decreased (60:1 ratio). This included broad increases in phospholipids, sphingolipids (ceramides, sphingomyelins, dihydroceramides), neutral lipids (triacylglycerols, diacylglycerols), and FFAs. Conversely, medium- and long-chain acylcarnitines were significantly reduced.
2. HFrEF Lipid Remodeling: In contrast, HFrEF showed far fewer lipid changes with a more balanced ratio of up- and down-regulated species (1.3:1). Notably, HFrEF was characterized by the depletion of cardiolipins (CL) and phosphatidylethanolamines (PE), classes that remained unchanged in HFpEF.
3. Obesity Alone Does Not Drive Lipotoxicity: In non-failing hearts, even those from severely obese individuals (BMI ≥35 kg/m²), there were minimal lipidomic alterations. No significant global correlation was found between lipid species abundance and BMI after FDR correction. Ultrastructural analysis confirmed that obese NF hearts lacked the lipid droplet accumulation, mitochondrial swelling, and sarcomere disassembly seen in obese HFpEF hearts.
4. Plasma vs. Myocardial Discordance: Plasma lipidomics revealed few inter-group disparities and showed minimal global correlation with myocardial lipid changes. While some specific lipids (e.g., saturated lysophosphatidylcholines) increased in both compartments, the massive myocardial accumulation in HFpEF was not mirrored in the plasma.
5. Failure of Animal Models to Phenocopy Human Disease: Principal component analysis (PCA) and support vector machine (SVM) classification demonstrated that human HFpEF and HFrEF occupy a distinct multidimensional lipidomic space from all tested animal models.
   • Mouse models (HFD+L-NAME, SAUNA) clustered with obesity-only controls and lacked the broad sphingolipid and phospholipid dysregulation seen in humans.
   • The ZSF-1 rat showed a narrow signature with TAG depletion.
   • The minipig model failed to capture ceramide and acylcarnitine signatures.
   • No animal model recapitulated the specific combination of broad lipid accumulation and acylcarnitine depletion found in human HFpEF.

Key Contributions

• First Comprehensive Human HFpEF Lipidome: This study provides the first broad-scale characterization of the human myocardial lipidome in HFpEF, contrasting it with HFrEF and non-failing hearts across a wide BMI spectrum.
• Refutation of Simple Lipotoxicity: The data challenge the paradigm that obesity alone drives lipotoxicity in HFpEF. The finding that severely obese non-failing hearts do not accumulate lipids suggests that HFpEF involves a specific failure of lipid metabolic adaptation rather than simple substrate overload.
• Metabolic Uncoupling: The study identifies a unique metabolic bottleneck in HFpEF: the coexistence of elevated fatty acid substrates (FFAs, TAGs, DAGs) and reduced acylcarnitines. This suggests increased lipid uptake but impaired mitochondrial transport and oxidation, a pattern distinct from the global suppression of fatty acid oxidation seen in HFrEF.
• Translational Gap Identification: By demonstrating that common preclinical models fail to capture the human HFpEF lipidomic signature, the paper highlights a significant translational gap, cautioning against the direct extrapolation of lipid-targeted therapies from these animal models to human HFpEF.

Significance
The paper concludes that human HFpEF is defined by a specific, broad lipid dyshomeostasis that is distinct from HFrEF and not induced by obesity alone. This lipid profile is not reflected in plasma biomarkers, limiting the utility of plasma lipidomics for diagnosing myocardial metabolic status in HFpEF. The findings reframe HFpEF as a disease of failed lipid metabolic adaptation rather than simple lipid excess. The inability of current animal models to phenocopy this human signature underscores the need for new preclinical models or human-centric approaches to understand the metabolic underpinnings of this prevalent and lethal syndrome.