Significant sex and age-related differences in cardiac function and disease risk are well-established. Females exhibit slower cardiomyocyte decline and distinct vascular properties compared to males, while aged hearts show hypertrophy, stiffening, and inflammation. However, the underlying molecular and cellular mechanisms remain unclear. Single-cell methods offer a powerful approach to characterize these variations at high throughput and cell-type resolution. Previous studies using single-cell techniques in humans and model organisms have profiled cardiac cell diversity and regulatory programs. This study aimed to build on these findings by performing a comprehensive single-cell analysis of human heart samples from multiple donors, spanning a wide age and sex range, to identify cell-type-specific molecular programs associated with age and sex.

Prior research extensively documented sex and age-related changes in the human heart at the tissue, cellular, and molecular levels. Studies have demonstrated that female hearts show more modest declines in cardiomyocyte numbers compared to male hearts over time and exhibit different vascular elasticity properties. Aged hearts, in contrast, display characteristics such as ventricular hypertrophy, tissue stiffening, and inflammation. While these macroscopic observations are well-documented, the precise molecular and cellular hallmarks of these differences, and even more so, their causal mechanisms, remain poorly understood. Single-cell approaches in human hearts have profiled the diversity of cardiac cell types and subtypes, mapping cell-type-specific regulatory programs. Similar analyses in model organisms have generated atlases of healthy tissue and investigated alterations in aging and disease. However, large-scale single-cell analyses of human hearts directly comparing age and sex effects are lacking.

The researchers collected heart samples from nine donors (three healthy adults, others with varying ages and sexes). They prepared 15 samples for single-cell analysis, each representing a single anatomical site within a donor. Frozen tissue was powdered and partitioned for nuclei isolation and fixation for both ATAC-seq and RNA-seq. Single-nucleus RNA-seq libraries were generated using a modified 3-level sci-RNAseq workflow, incorporating additional RNase inhibitors, mechanical dissociation, and FACS sorting to reduce noise. Similarly, single-nucleus ATAC-seq libraries were prepared using a 3-level sci ATAC-seq protocol. Data processing involved demultiplexing, read assignment to cells, filtering, and generation of cell-by-gene matrices using custom pipelines. Cell types were annotated using marker genes and a co-embedding approach combining RNA- and ATAC-seq data. Differential expression analyses utilized generalized linear mixed models to account for donor-level variation and identify genes differentially expressed by age and sex. Pathway enrichment analysis was performed to identify biological processes affected by age and sex. Additionally, TF motif enrichment analysis was conducted to identify age- and sex-associated changes in chromatin accessibility. Finally, predictive models of cell-type-specific RNA expression levels were trained using ATAC-seq profiles, incorporating both promoter and distal regulatory sequences to assess the predictive power of a simple TF motif regulatory grammar.

The study generated matched single-cell ATAC-seq (11,738) and RNA-seq (89,804) datasets from 15 samples across 9 donors. Analysis revealed widespread variation in transcriptional and regulatory programs by age and sex. Sex was associated with metabolic alterations, such as changes in oxidative phosphorylation and cholesterol metabolism across multiple cell types. Age was associated with immune changes, including increases in interferon response and inflammation in various cell types. A meta-analysis combining the newly generated data with five external datasets (73 donors total) confirmed many findings, such as metabolic changes by sex and immune changes by age. However, it also showed some discrepancies, highlighting the complexity and variability in human cardiac biology. The meta-analysis also identified significant differences in cardiac cell type composition as a function of sex (altered cardiomyocyte proportions) and age (decreased neurons). Analysis comparing TF motif enrichments in adult and fetal cells revealed mostly similar enrichments, but also highlighted several potentially developmentally specific regulatory factors. Finally, predictive models of gene expression demonstrated that including distal regulatory information from ATAC-seq significantly improved the accuracy of the models, emphasizing the importance of these regions in gene regulation.

This study provides a valuable resource of single-cell multi-omics data, revealing the complexity of age- and sex-associated variations in the human heart. The findings highlight the importance of considering both age and sex as significant covariates when studying cardiac biology. The integration of ATAC-seq and RNA-seq data allowed for a more comprehensive understanding of transcriptional and epigenetic regulation. While the meta-analysis strengthened some findings, inconsistencies also emerged, underscoring the challenges of analyzing data from multiple sources and the need for larger, well-standardized datasets. The identification of potential developmental-stage specific regulatory factors opens new avenues for research into cardiac development and disease. The predictive modeling demonstrated the power of combining distal and proximal information for accurately predicting gene expression, highlighting the importance of non-coding regulatory elements.

This research provides a comprehensive single-cell resolution analysis of age- and sex-related alterations in the human heart. The identification of numerous differentially expressed genes, pathways, and transcription factors highlights the complexity of these effects and underscores the necessity of considering both covariates in future studies. The study’s methodologies and findings lay a foundation for future investigations into personalized medicine and targeted therapies for cardiovascular diseases. Future research could focus on validating developmental stage-specific regulatory factors and developing more sophisticated predictive models to fully leverage the information contained within distal regulatory sequences. Larger and more diverse cohorts are needed to overcome limitations and improve the generalizability of findings.

The study's limitations include the relatively small number of donors in the initial analysis, although this was partially mitigated by the meta-analysis. Confounding factors associated with individual donor health, environmental exposures and medication use were not completely controlled for. The observational nature of the study prevents causal inferences, and larger, more meticulously controlled studies will be necessary to draw firm conclusions. The discrepancies observed between the single-study and meta-analyses highlight the inherent variability in human cardiac biology, the impact of data processing and technical differences, and the challenges of integrating data from multiple sources. Further research is needed to elucidate the specific mechanistic roles of identified factors and fully resolve the complex interplay between age, sex, and cardiac function.