Identification of candidate long non-coding RNAs and mRNAs associated with heart aging in mice

Animal model and grouping

C57BL/6 wild-type male mice were used in this study. Young mice aged 10 weeks were obtained from Shanghai Jiesijie Laboratory Animal Co., while elderly mice (approximately 18 months old) were sourced from Jiangsu Lingfei Biotechnology Co. The latter were raised to 20 months of age at our facility to establish a natural aging model. Animals were grouped based on age into:

• 3m group:3-month-old male mice (n = 5)

• 20m group:20-month-old male mice (n = 5).

This sample size has been widely used in similar exploratory studies comparing age-related differences in cardiac molecular expression and is considered sufficient to identify significant differential expression of target genes, while adhering to ethical guidelines minimizing animal usage (Keele et al., 2023). Although no formal power calculation was performed prior to the study due to its exploratory nature, the chosen number aligns with standard practices in molecular and transcriptomic profiling studies, allowing adequate statistical power for detecting biologically meaningful differences between age groups.

Mice were housed individually in specific pathogen-free (SPF) conditions at the Laboratory Animal Center of Zhongshan Hospital, Fudan University. The environment was maintained at 23 ± 1 °C with 50% ± 10% humidity and a 12-hour light/dark cycle. Mice had ad libitum access to sterile water and standard rodent chow (Shanghai SLAC Laboratory Animal Co., Ltd.) with environmental enrichment provided. The study was approved by the Animal Ethics Committee of Zhongshan Hospital, Fudan University (Approval No. 2021-20). Single housing was used for all animals.

Mice were included if they appeared healthy, with no signs of illness or weight loss. No animals or data points were excluded. Due to the distinct physical characteristics of age groups, full randomization and blinding were not feasible for animal handling, but all outcome assessors for molecular analyses were blinded to group identity.

Echocardiographic assessments were conducted 24–48 h prior to euthanasia to allow sufficient recovery from isoflurane anesthesia while minimizing physiological variability. All mice were euthanized by cervical dislocation without prior anesthesia to avoid anesthetic-induced transcriptomic alterations. Cardiac tissue was immediately harvested and snap-frozen in liquid nitrogen for microarray and downstream analysis.

Echocardiography

Mice were anesthetized with 1% isoflurane (R510-22, RWD Life Science) and examined using a Vevo 2100 High-Resolution Imaging System (FUJIFILM Visual Sonics, Toronto, Canada) equipped with an 707B 30 MHz probe (Moreno-Fernandez et al., 2021). Parasternal long- and short-axis views were acquired. Parameters including heart rate (HR), interventricular septum in systole and diastole (IVSs, IVSd), left ventricular internal dimension in systole and diastole (LVIDs, LVIDd), left ventricular posterior wall in systole and diastole (LVPWs, LVPWd), rejection fraction (LVEF%) and fractional shortening (FS%) were measured using Vevo LAB software (Version 3.2.6).

Cardiac diastolic function was assessed by the myocardial performance index (MPI or Tei index) (Lakoumentas et al., 2005). Calculated as of Tei index = (isovolumic relaxation time (IVRT) + isovolumetric contraction time (IVCT))/ejection time (ET).

Histological staining

Tissue sampling for histological analyses occurred immediately following euthanasia, at a single time point for each animal.

For hematoxylin-eosin (HE) staining, cardiac tissues were fixed with 10% paraformaldehyde (PFA; Sigma-Aldrich, Cat# P6148) and embedded in paraffin. A 4 µm was stained with hematoxylin for 2 min, washed, and then counterstained with eosin. For Masson staining, tissues were sequentially treated with Masson trichrome kit reagents (Solarbio, Cat# G1340), including phosphotungstic acid, bright green, and acetic acid washes. Slides were dehydrated, cleared, mounted, and visualized under a light microscope.

Apoptosis detection

Apoptosis was performed using a terminal deoxynucleotide transferase dUTP nick end labeling (TUNEL) assay kit (Roche, Cat# 11684795910). Cardiac sections were permeabilized, incubated with TUNEL reaction mixture, and developed with 3,3’-Diaminobenzidine (DAB) substrate. Positive apoptotic nuclei were visualized using light microscopy.

Senescence-associated β-galactosidase staining

Tissues were fixed with 10% PFA and embedded in OCT compound. Frozen sections (10 µm) were cut, air-dried, and treated for β-galactosidase staining. After antigen retrieval and blocking, sections were incubated with anti-β-galactosidase (Abcam, cat#ab9361, 1:500) overnight at 4 °C, followed analyzed by Image J software (v1.53k,National Institutes of Health, Bethesda, MD, USA).

Microarray and transcriptome analysis

Myocardial tissue samples were collected from 3-month-old and 20-month-old mice. Each animal was sampled at a single time point. Total RNA was extracted using TRIzol reagent (Invitrogen, USA), and its quantity and purity was assessed using NanoDrop One spectrophotometry (Thermo Fisher Scientific, USA). RNA integrity was confirmed by denaturing agarose gel electrophoresis.

Gene expression profiling was performed using the Agilent-085983 Arraystar Mouse LncRNA Microarray V4.0 platform (GPL26962), which enables simultaneous detection of both long non-coding RNAs (lncRNAs) and protein-coding mRNAs. Sample labeling, hybridization, washing, and scanning were conducted by Aksomics Bio (Shanghai, China), following the manufacturer’s standard protocol.

Briefly, total RNA was treated to remove ribosomal RNA and amplified into fluorescently labeled complementary RNA (cRNA) using the Arraystar Flash RNA Labeling Kit. The labeled cRNAs were hybridized to the microarray slides and scanned with the Agilent G2505C Scanner. Raw signal intensities were extracted using Agilent Feature Extraction Software (version 10.7.3.1).

Downstream data processing and statistical analyses were performed using R version 3.6.0 and Bioconductor release 3.10. Raw data were background-corrected and quantile-normalized. Probes flagged as “Present” or “Marginal” in at least 50% of samples were retained for further analysis. Differentially expressed genes (DEGs), including both lncRNAs and mRNAs, were identified using the limma package (version 3.42.2), applying linear models and empirical Bayes moderation. Genes with an adjusted p-value (false discovery rate, FDR) <0.05 and —log2 fold change— ≥ 1 were considered statistically significant.

Functional enrichment analyses, including Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, were conducted using the clusterProfiler package (version 3.14.3). All annotated genes from the platform were used as the background gene set.

No technical replicates were included. Five biological replicates per group were used to ensure statistical robustness and account for biological variability.

The microarray data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO) under accession number GSE256034.

Protein–protein interaction network and functional annotation

To investigate potential functional relationships and molecular interactions among the most DEGs mRNA associated with cardiac aging, we selected the top 100 DEGs based on adjusted p-value and log2 fold-change. Protein–protein interaction (PPI) network analysis was performed using the STRING database (version 12.0; https://string-db.org) with the organism set to Mus musculus and a medium confidence score threshold of 0.4.

The network statistics, including average node degree and enrichment p-values, were extracted directly from the STRING output. Unconnected nodes and sparse interactions were retained to reflect the full distribution of DEGs.

In addition, we selected the top 20 DEGs for individual functional annotation. Gene symbols were queried using the GeneCards, UniProt, and PubMed databases to obtain descriptions of biological roles, especially in pathways related to immune response, mitochondrial function, cellular transport, or transcriptional regulation.

Statistics

As an exploratory observational study, formal sample size calculation based on a primary outcome measure was not performed. The chosen sample size was determined based on previously published literature and preliminary experimental experience, ensuring sufficient sensitivity for identifying biologically meaningful differences in molecular profiles and phenotypic characteristics between age groups.

Data are expressed as mean ± standard error or standard deviation. For comparisons of continuous variables between groups, the student’s t-test was used for Gaussian data and the Mann–Whitney test for non-Gaussian data.

For microarray transcriptomic analyses, differentially expressed genes were identified using appropriate statistical tests with multiple hypothesis testing correction. Adjusted p-values were calculated using the Benjamini–Hochberg false discovery rate (FDR) method, and genes with FDR <0.05 and —log2 fold change— ≥ 1 were considered statistically significant.

Statistical analyses were performed using SPSS 26.0 (SPSS Inc., Chicago, Illinois, U.S.A.) and GraphPad Prism 9.20 software (GraphPad Software, Inc., La Jolla, CA, U.S.A.) and transcriptomic analyses were conducted using R (version 3.6.0) with standard bioinformatics packages. P-values <0.05 were considered significant.

Cardiac phenotype in aged mice

We first characterized the cardiac phenotype in a natural aging mouse model. Compared to young (3-month-old), aged mice (20-month-old) exhibited increased heart size (Fig. 1A), as demonstrated by higher heart weight/body weight ratio (HW/BW) (Fig. 1B) and heart weight/tibial length ratio (HW/TL) (Fig. 1C).

Cardiac ultrasound examination was then performed on mice from both age groups. At comparable heart rates (Fig. 1D), aged mice showed no significantly differences in ejection fraction (EF) (Fig. 1E) or FS (Fig. 1F) compared to young mice. However, the Tei index, an indicator of global myocardial performance, was significantly elevated in the aged group (P < 0.01, Fig. 1G), indicating impaired myocardial function. Specific echocardiographic measurements revealed that, the IVSd (Fig. 1H) and LVPWd (Fig. 1K) were increased in aged mice. Further both left ventricle mass (Fig. 1L) and corrected heart mass (Fig. 1J) were significantly greater in the 20-month-old group. Other echocardiographic parameters showed no significant difference between groups (extended Fig. 1).

Histologic alterations in the heart of aged mice

Histological examination was performed on left ventricles sections from both 3-month-old and 20-month-old mice. HE staining revealed that the myocardium of young mice exhibited neatly arranged myocardial fibers, uniformly distributed nuclei, and intact cellular morphology (Fig. 2A). In contrast, myocardium from aged mice showed disrupted alignment of myocardial fibers, enlarged cardiomyocyte volume, and irregular cellular arrangement, indicating significant structural remodeling (Fig. 2B).

Masson’s trichrome staining demonstrated increased collagen deposition within the myocardial tissue of 20-month-old mice compared to young controls, reflecting enhanced myocardial fibrosis during aging (Figs. 2C–2E). Additionally, TUNEL staining identified significantly increased cardiomyocyte apoptosis in the hearts of aged mice relative to young mice (Figs. 2F–2G). Furthermore, aged mice exhibited significantly senescence-associated β-galactosidase activity, indicating enhanced cellular senescence within the aged myocardium (Figs. 2H–2J).

Identification of candidate lncRNAs and mRNAs associated with cardiac aging

To systematically identify transcriptomic changes associated with cardia aging, we performed microarray-based transcriptome profiling of cardiac tissues from 3-month-old (young) and 20-month-old (aged) C57BL/6 mice. A total of 67,236 lncRNA probes and 16,720 mRNA probes were detected. Differential expression analysis revealed 1,293 upregulated and 1,551 downregulated lncRNAs in aged hearts compared to young controls, as visualized by the volcano plot (Fig. 3A). Similarly, 663 mRNAs were upregulated and 1,597 were downregulated in aged hearts compared to young hearts (Fig. 3B). The top 100 differentially expressed lncRNAs and mRNAs are listed in extended Tables S1 and S2, respectively.

To further explore the biological significance of these transcriptional changes, we performed GO and KEGG enrichment analyses on the differentially expressed mRNAs. GO analysis indicated the upregulated mRNAs were significantly enriched in biological processes such as defense response, stress response, leukocyte cell–cell adhesion, inflammatory response, and negative regulation of endopeptidase activity (Fig. 3C). Regarding cellular components, upregulated mRNA was primarily associated with membranes, extracellular matrix, and cell surface structures (Fig. 3D). Enriched molecular functions included passive transmembrane transporter activity, channel activity, endopeptidase regulatory activity, and cyclin-dependent protease activity (Fig. 3E).

In contrast, downregulated mRNAs were predominantly enriched in biological processes related to cellular metabolic processes, cellular organization, and general cellular function (Fig. 3F). Cellular component enrichment highlighted structures such as organelles, cytoplasm, and intercellular junctions (Fig. 3G), whereas molecular function enrichment included catalytic activity and structural molecule activity (Fig. 3H).

KEGG pathway analysis further revealed that upregulated mRNAs were significantly involved in the inflammation- and stress-related signaling pathways, including NFκB signaling, IL-17 signaling, Hippo signaling, cytokine receptor interactions, and C-type lectin receptor signaling pathways (Fig. 3I). Conversely, pathways associated with the downregulated mRNAs in aged hearts encompassed key aging-related pathways such as mitogen-activated protein kinase (MAPK) signaling, ubiquitin-mediated proteolysis, ErbB signaling, insulin signaling, and autophagy (Fig. 3J).

PPI network reveals sparse connectivity among aging associated DEGs

Protein–protein interaction (PPI) analysis of the top 100 DEGs mRNA from aged mouse hearts revealed a sparsely connected network. Using the STRING database with a medium confidence threshold (score >0.4), we identified 100 nodes and only 29 edges, resulting in an average node degree of 0.58 (Fig. 4). This indicates that most genes have few or no direct interactions with each other, and that the DEGs do not form a cohesive or densely interconnected network.

Furthermore, no statistically significant enrichment was detected for GO terms or KEGG pathways, suggesting that these DEGs are not confined to a specific biological process or molecular pathway.

This lack of connectivity reflects the functionally heterogeneous nature of gene expression changes during cardiac aging. Rather than being organized into a central signaling hub, the transcriptomic alterations appear to be distributed across a diverse range of biological processes. Consistent with this, our functional annotation of the top 20 DEGs (Extended Table S3) revealed involvement in immune regulation (Cxcl13, Il3ra, Chil3), mitochondrial function (Timm50), transcriptional control (Vgll2, Zfp65), cell adhesion (Iglon5), protease activity (Klk1b16, Wfdc21), and cell cycle regulation (Cenpu).

Notably, several of these gene are closely related to inflammaging, a hallmark of aging characterized by chronic low-grade inflammation (Ramos et al., 2025). Specifically, Cxcl13 (a chemokine mediating B-cell chemoattraction), Skint7 (T-cell selection), Tspan32 (immune cell surface signaling), Wfdc21 (a WAP-domain protease inhibitor), and Mmd2 (monocyte-to-macrophage differentiation) were identified as inflammaging-associated DEGs. These findings underscore the contribution of immune remodeling and persistent inflammation to the transcriptome landscape of the aging heart.