Spironolactone and Fibrosis in Heart Failure Risk: Machine Learning Analysis of HOMAGE Trial Plasma Proteomics
Susana Ravassa , Nicolas Girerd , Frank Edelman , Begoña López , João Pedro Ferreira , Daniela Zurkan , Gorka San José , Iñigo Latasa , Pierpaolo Pellicori , Franco Cosmi , Johannes Petutschnigg , Stephane Heymans , Hans-Peter Brunner-La Rocca , Burkert Pieske , Christian Delles , Andrew L. Clark , Javier Díez , Faiez Zannad , John G. F. Cleland , Arantxa González
MedComm ›› 2026, Vol. 7 ›› Issue (3) : e70634
In the HOMAGE (Heart Omics in AGEing) trial, spironolactone reduced serum concentrations of procollagen Type I C-terminal propeptide (PICP), a fibrosis biomarker, in patients at risk of heart failure. To elucidate the underlying mechanisms, multidimensional analyses including proteomics were conducted. Olink cardiovascular and inflammation panels (n = 276 proteins) were measured in plasma from 488 HOMAGE participants at baseline, 1 month, and 9 months after randomization. Proteins associated with PICP changes were identified using machine learning algorithms (MLAs). Selected candidates were further analyzed in patients with heart failure and preserved ejection fraction (Aldo-DHF trial). Linear regression and mediation analyses assessed which MLA-selected proteins mediated spironolactone's effects on PICP. MLAs consistently linked PICP reduction to changes in biomarkers of collagen (e.g., decreased COL1A1), fatty acid metabolism (e.g., increased FABP4), immune function (e.g., increased CCL24 and IL6RA, and decreased FLT3L), neurological function (e.g., increased DNER), cell–matrix interactions (e.g., increased galectin-9 [GAL9] and decreased thrombospondin-2 [THBS2]), and reduced NT-proBNP. Mediation analysis suggested that changes in GAL9 and THBS2 were associated with spironolactone-induced PICP reduction, which was confirmed in Aldo-DHF patients. This study raises the hypothesis that spironolactone inhibits collagen synthesis via inflammatory, metabolic, and extracellular matrix pathways, and particularly through modulation of GAL9 and THBS2.
biomarkers / fibrosis / heart failure / machine learning algorithms (MLA) / procollagen Type I C-terminal propeptide (PICP) / spironolactone
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
|
| [21] |
|
| [22] |
|
| [23] |
|
| [24] |
|
| [25] |
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
| [30] |
|
| [31] |
|
| [32] |
|
| [33] |
|
| [34] |
|
| [35] |
|
| [36] |
|
| [37] |
|
| [38] |
|
| [39] |
|
2026 The Author(s). MedComm published by Sichuan International Medical Exchange & Promotion Association (SCIMEA) and John Wiley & Sons Australia, Ltd.
/
| 〈 |
|
〉 |