Proteomics and Precise Exercise Phenotypes in Heart Failure With Preserved Ejection Fraction: A Pilot Study

Figures

• 

  Figure 1. Overview of the study design. CO indicates cardiac output; FHS, Framingham Heart Study; HFpEF, heart failure with preserved ejection fraction; O2, oxygen; PCWP/CO, slope of the pulmonary capillary wedge pressure to cardiac output ratio; PVR, pulmonary vascular resistance; and VO2, oxygen uptake.

• 

  Figure 2. Spearman correlations of parent physiological/hemodynamic exercise measures, proteomic scores, natriuretic peptides, and resting supine PCWP. Exercise measures are labeled “CPET,” proteomic scores are labeled “proteomic.” BNP and NT‐proBNP were measured by Olink. Correlation P values are provided in Table S3. BNP indicates B‐type natriuretic peptide; CO, cardiac output; CPET, cardiopulmonary exercise test; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; O2, oxygen; PCWP, pulmonary capillary wedge pressure; PCWP/CO, pulmonary capillary wedge pressure to cardiac output ratio; PVR, pulmonary vascular resistance; and VO2, oxygen uptake.

• 

  Figure 3. Shared and distinct associations of proteins and exercise phenotypes. A, Shared and distinct associations of the top 50 proteins (by absolute magnitude of ridge coefficient) for each phenotype. B, Circos plot organizing these top 50 proteins across exercise response phenotypes. The outermost color bar signifies the phenotype; the heatmap expresses the ridge model coefficients (scaled); and the inner color bar represents the secreted status of the protein. The brown line connections and their width signify absolute magnitude Spearman correlation >0.6 between proteins of interest with different phenotype associations. Secretome protein annotation was performed using data from the Human Protein Atlas project. CO indicates cardiac output; PCWP/CO, pulmonary capillary wedge pressure to cardiac output ratio; PVR, pulmonary vascular resistance; and VO2, oxygen uptake. Expanded protein names are available in Table S1.

• 

  Figure 4. Shared and distinct protein importance (ridge model coefficient) across phenotypes. The Pearson correlation coefficient is displayed. The ridge coefficients are multiplied by 100 for plotting. Our intention was to label proteins close to the line of equality (dotted line; “shared” mechanisms) and those that had a significant residual from this line (nominal P<0.10). Those proteins that are encircled had a more extreme residual (P≤0.05), signifying “distinct” importance for 1 phenotype over another. The size of each point corresponds to its magnitude for importance for peak VO2 (scaled), and its color represents magnitude of residual (with red as higher residual). Some points may not be labeled due to overplotting. CO indicates cardiac output; PCWP/CO, pulmonary capillary wedge pressure to cardiac output ratio; PVR, pulmonary vascular resistance; and VO2, oxygen uptake. Expanded protein names are available in Table S1.

• 

  Figure 5. Selected phenotypes of central importance in HFpEF physiology. A (left) plots the protein importance (ridge model coefficient) of PCWP/CO slope (left heart filling with exercise) against peak exercise PVR (pulmonary hypertension). Labeling scheme as in Figure 4, with caution that all points may not be labeled to avoid overplotting. The middle panel displays a subset of proteins with significant residual (P<0.10). Each bar is calculated as a log of the ratio of the magnitude of a protein's importance to PVR divided by its importance to PCWP/CO slope. Therefore, a ratio >1 here implies a higher magnitude ridge coefficient for PVR relative to ridge coefficient for PCWP/CO slope (and the log of that ratio >0, or to the right of 0 on the x axis). Effectively, those proteins to the right of 0 are more highly weighted on peak exercise PVR, while proteins to the left of 0 are more highly weighted on PCWP/CO slope. The color represents the sign of the ridge coefficient for PVR (such that an increase in protein would be related to a higher peak exercise PVR if red). The left panel reflects biological annotations. PCWP/CO, pulmonary capillary wedge pressure to cardiac output ratio; PVR,…

• 

  Figure 6. Prospective associations of proteomic signatures of exercise phenotypes with incident HF subtypes in Framingham Heart Study. The association of each proteomic signature with overall HF and with HFpEF and HFrEF are displayed. Logistic regression models were used to relate proteomic scores to overall HF, and multinomial logistic regression was used to relate the proteomic scores to HFpEF and HFrEF. Models were adjusted for age, sex, smoking, diabetes, prevalent coronary heart disease, serum creatinine, body mass index, systolic blood pressure, use of antihypertensive medications, and B‐type natriuretic peptide level. C(a‐v)O2 difference indicates arteriovenous oxygen difference; CO, cardiac output; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; PCWP/CO, pulmonary capillary wedge pressure to cardiac output ratio; PVR, pulmonary vascular resistance; and VO2, oxygen uptake.