Cardiovascular disease is often managed through broad clinical categories: heart failure, cardiomyopathy, ischemic injury, arrhythmia risk. These categories are essential for diagnosis and treatment, but they do not fully explain why patients with similar clinical presentations can progress differently or respond unevenly to therapy.
A new single-nucleus atlas of the adult human heart aims to bring that biology into sharper focus. In a study published in Nature Cardiovascular Research, researchers from the Broad Institute of MIT and Harvard and Mass General Brigham developed HeartMap, an integrated atlas spanning more than 2.4 million cardiac nuclei from 209 individuals.
The resource brings together data from nine studies and includes eight anatomical regions and seven healthy or disease states. Its value lies not only in scale, but in its attempt to make cardiac single-cell data more comparable across studies.
Why cardiology needs cellular resolution
Precision medicine depends on identifying the biological mechanisms that matter for a specific patient or disease subtype. In oncology, this has increasingly meant matching molecular alterations to targeted therapies. Cardiology has made important progress with genetics, imaging, biomarkers, and risk stratification, but many therapeutic decisions still operate at a relatively high level of disease classification.
Heart disease is not driven by cardiomyocytes alone. Fibroblasts, endothelial cells, immune cells, vascular smooth muscle cells, pericytes, and other populations all contribute to remodeling, inflammation, fibrosis, and tissue dysfunction. A patient’s clinical phenotype may therefore reflect different combinations of cellular programs.
HeartMap was designed to help researchers interrogate those programs. The authors write that differentiating transcriptional signatures between cardiovascular disease groups using HeartMap “may aid in precision medicine approaches” by informing biomarker and therapeutic target discovery.
Separating robust disease signals from dataset noise
Single-cell and single-nucleus sequencing have already revealed important differences between healthy and diseased cardiac tissue. However, individual studies vary in how samples are collected, processed, sequenced, and analyzed. Those differences can obscure whether a signal is truly disease-related or specific to a cohort or protocol.
To address this, the HeartMap team reprocessed and harmonized published datasets, using computational integration to reduce technical variation while preserving biological signal. The final atlas resolved 14 broad cardiac cell types and 52 clusters.
Across the atlas, diseased and non-failing hearts separated by gene expression patterns. Disease-versus-control comparisons produced more differentially expressed genes than comparisons between disease states, suggesting that different cardiovascular conditions share broad remodeling programs while retaining more specific molecular features.
That distinction is important for translation. Shared injury programs may help explain common features of heart failure progression, while disease-specific or cell-state-specific programs may point to more selective therapeutic opportunities.
Fibroblasts show why cell state matters
One of the most clinically relevant findings involves fibroblasts. These cells are central to fibrosis and cardiac remodeling, but they are not a single uniform target. Some fibroblast activity may be part of necessary repair, while other states may contribute to scar formation, tissue stiffening, inflammation, and progressive dysfunction.
HeartMap identified 29 fibroblast subclusters, including activated fibroblast populations that differed across cardiomyopathies. Two activated populations, enriched for COL22A1 or TNC, were validated in human heart tissue using RNAscope.
This matters because anti-fibrotic drug development has long faced a precision problem: suppressing fibrosis broadly may not be the same as targeting the cell states that drive pathological remodeling. The authors note that identifying these activated fibroblast populations brings the field closer to understanding which subpopulations are “ideal candidates for therapeutic intervention.”
A research reference, not a clinical test
HeartMap is not a diagnostic assay and is not ready to guide care for individual patients. The atlas is built from previously generated datasets, so the researchers could not fully control sample collection, nuclei isolation, sequencing methods, or clinical metadata. The study also notes limited ancestral diversity, with samples largely from White or unclassified ethnic backgrounds. Most disease samples represented chronic or end-stage disease, limiting insight into early disease initiation.
Even so, the atlas points to an important direction for precision cardiology. Future patient datasets could be compared against resources such as HeartMap to determine which cellular programs are active in a given disease context. Over time, that could help connect clinical phenotypes with cell-specific biomarkers, drug targets, and mechanisms of progression.
Rather than treating the diseased heart as a single failing organ, HeartMap frames it as a dynamic cellular system. That shift may be essential if cardiovascular medicine is to move beyond broad disease labels toward therapies guided by the cells and pathways driving disease in individual patients.
The post HeartMap Atlas Offers New Single-Cell View of Human Heart Disease appeared first on Inside Precision Medicine.