Mass Spectrometry Imaging-Based Single-Cell Lipidomics Profiles Metabolic Signatures of Heart Failure

Abstract

Heart failure (HF), leading as one of the main causes of mortality, has become a serious public health issue with high prevalence around the world. Single cardiomyocyte (CM) metabolomics promises to revolutionize the understanding of HF pathogenesis since the metabolic remodeling in the human hearts plays a vital role in the disease progression. Unfortunately, current metabolic analysis is often limited by the dynamic features of metabolites and the critical needs for high-quality isolated CMs. Here, high-quality CMs were directly isolated from transgenic HF mice biopsies and further employed in the cellular metabolic analysis. The lipids landscape in individual CMs was profiled with a delayed extraction mode in time-of-flight secondary ion mass spectrometry. Specific metabolic signatures were identified to distinguish HF CMs from the control subjects, presenting as possible single-cell biomarkers. The spatial distributions of these signatures were imaged in single cells, and those were further found to be strongly associated with lipoprotein metabolism, transmembrane transport, and signal transduction. Taken together, we systematically studied the lipid metabolism of single CMs with a mass spectrometry imaging method, which directly benefited the identification of HF-associated signatures and a deeper understanding of HF-related metabolic pathways.

Fig. 1.

A method for lipidomics analysis in mammalian cardiomyocytes at a single-cell level.

Fig. 2.

Images on the isolated mouse cardiomyocytes (CMs). (A) Immunofluorescence staining with calcein (green, live cells positive) of freshly isolated mouse CMs in heart failure (HF) and control groups. (B) Bright-field microscopy images on the CMs in HF and control groups. (C) Total secondary ions images on the CM samples at different groups.

Fig. 3.

Overall lipid profiles of the single-cell mass spectrometry. (A) Negative and (B) positive ToF-SIMS spectra of CM under delayed extraction mode. For a clear display, the intensity of ion signal was amplified by different multiples. a.u., arbitrary units. (C and D) For the imaging components, including C₁₆H₃₁O₂⁻ (FA 16:0, [M-H]⁻, m/z of 255.20), C₃₃H₆₃PO₈⁻ (PC 28:0, [M-TMA]⁻, m/z of 618.48), C₄₁H₆₉NPO₈⁻ (PE 36:6, [M-H]⁻, m/z of 734.47), C₈H₁₉NPO₄⁺ (PC head group, m/z of 224.05), C₂₇H₄₅⁺ (cholesterol, [M-OH]⁺, m/z of 369.33), and C₃₉H₇₇NPO₈⁺(PE34:1, [M+H]⁺, m/z of 718.45).

Fig. 4.

Differential metabolomic analysis of the CMs. (A) Canonical correlation analysis between the expression of detected metabolites of the cytomembrane and the intracellular in negative ion mass spectra. (B and C) OPLS-DA presented stark differences between the metabolites of the cytomembrane of 2 groups (yellow, normal control group; violet, HF group) (B) for the negative ion mass spectra and (C) for the positive ion mass spectra. (D) SHAP plot displayed variables in a top-to-bottom format to demonstrate the feature importance for prediction of HF in cytomembrane. (E and F) Receiver operating characteristic (ROC) curve to quantify the discriminability of metabolites to distinguish between the 2 groups (HF vs. control). AUC, area under the ROC curve. (G) ToF-SIMS imaging of CM’s surface of control and HF. C₁₈H₃₅O₂⁻ is a characteristic fragment of FA 18:0, C₄₆H₉₂N₂PO₆⁻ is a characteristic fragment of SM 42:1, and C₄₁H₇₃NPO₈⁻ is a characteristic fragment of PE 36:4.

Fig. 5.

Functional implications of differential metabolites in the cytomembrane of CMs. (A) Hierarchical cluster analysis was used to visualize the clustering and relevance of the differential metabolites (aquamarine, normal control group; red, HF group). (B) Functional enrichment analysis of these differential metabolites.