Heart failure (HF) arises from pathological multicellular remodeling in response to increased biomechanical stress and is characterized by impaired cardiac structure, metabolism, and function. Mechanical unloading can induce reverse remodeling, yet recovery is often incomplete, and the molecular determinants of reversibility versus persistence remain poorly defined. Here, we combined a reversible murine pressure-overload model, O-ring aortic banding, followed by de-banding, with phenotypic profiling and bulk and single-nucleus transcriptomics to resolve the cellular and intercellular programs governing adverse and reverse cardiac remodeling. Pressure overload induced robust HF, while unloading promoted partial functional recovery, accompanied by regression of fibrosis and cardiomyocyte hypertrophy. Bulk transcriptomics revealed a continuous damage-to-recovery transcriptional trajectory, marked by partial normalization of extracellular matrix and metabolic pathways but persistent proinflammatory and mitochondrial dysfunction signatures. Single-nucleus transcriptomics uncovered recovery dynamics that are highly cell-type-specific: cardiomyocytes and endothelial cells largely re-established homeostatic states, whereas fibroblasts retained activated, revealing remodeling-associated programs that dominated residual pathology. Unbiased analysis identified two coordinated multicellular programs, a reversible metabolic stress response and a persistent inflammatory-mitochondrial dysfunction, shared across cardiac cell types. Cell-cell communication analysis pointed at non-myocyte populations as drivers of reverse remodeling . Lymphatic endothelial and vascular endothelial cells emerged as major sources of pro-recovery signals, including Reelin (RELN). Functional validation in neonatal rat and human induced pluripotent stem cell-derived cardiomyocytes demonstrated that RELN suppresses pathological fetal gene activation under basal and pro-hypertrophic conditions. Collectively, these findings define reverse cardiac remodeling as an active yet constrained multicellular process and identify lymphatic-derived RELN as a previously unrecognized regulator of cardiomyocyte recovery in chronic HF.
Pre-print available soon!
This repository contains the code used for the analysis of bulk RNA-seq and single-nucleus RNA-seq (snRNA-seq) data associated with the manuscript.
- Scripts for bulk RNA-seq analysis are available in analysis/bulk
- Scripts for snRNA-seq analysis are available in analysis/singlenuc
A Conda environment YAML file is included to facilitate installation of all dependencies required to run the scripts and notebooks.
The bulk RNA-seq counts and metadata, as well as the processed snRNA-seq AnnData object are publicly available for download at Zenodo.
This repository is provided under the GNU General Public License v3.0.