The loss of microRNA-26b promotes aortic calcification through the regulation of cell-specific target genes

Abstract

Aims: Vascular calcification is the abnormal deposition of calcium phosphates within blood vessels. This condition is significantly associated with the development of cardiovascular disease, yet the underlying mechanisms remain largely unknown. MicroRNAs (miRNAs) may be crucial in initiating vascular calcification by regulating a network of specific cellular targets. In this study, we explored for the first time the potential role of microRNA-26b (miR-26b) in vascular calcification.

Methods and results: Using micro-positron emission tomography and computed tomography (micro-PET/CT) imaging with 18F-sodium fluoride, we measured aortic calcification in miR-26b knockout mice (miR-26bKO). We conducted bulk RNA sequencing (RNA-seq), single-cell RNA sequencing, and network analysis to identify cell-specific targets and the cellular complexity contributing to the observed phenotype. Additionally, we examined aortic tissues from patients with aortic aneurysm or valvular-related aortopathy to determine how the expression levels of miR-26b and its targets correlate with calcification. Our findings revealed that miR-26b is downregulated in the aortic tissues of patients with aortic calcification, whereas miR-26b expression negatively correlates with calcification levels. Similarly, miR-26bKO mice developed spontaneous age-related aortic microcalcifications. Combining single-cell transcriptomics with network analyses, we identified and mapped cell-type specific targets of miR-26b and regulatory pathways. Furthermore, we validated the cell-specific expression of Smad1 in smooth muscle cells (SMCs) and characterized the cell-cell communication between aortic cells, exposing the bone morphogenetic protein (BMP) pathway. The development of microcalcification was attributed to Bmp4 released from fibroblasts (FBLs), leading to Smad1 phosphorylation and calcium accumulation in SMCs of miR-26bKO mice. We found that aortic microcalcification could be pharmacologically reversed by disrupting cellular communication. Lastly, we demonstrated an inverse correlation between miR-26b and SMAD1 levels in calcified aortic tissues.

Conclusion: The deficiency of miR-26b is crucial for initiating and promoting aortic calcification, revealing new therapeutic targets for aortic disease.

Keywords: aortic calcification; aortic disease; cell–cell communications; microRNAs; single-cell transcriptomics.

Graphical Abstract

Figure 1

miR-26b-5p is downregulated in the aortic tissues of patients with aortopathy and aortic valve disease. (A) Histological characterization of human aortas (aortopathy concomitant or not with aortic valve disease, see Supplementary material online, Table S1) was performed by alizarin red staining (top panel), Sirius red/alcian blue double staining in bright field (middle panel) and polarized light (bottom panel). Scale bar = 200 μm. (B) Calcification was quantified as % of alizarin red staining, with a significant increase of calcium content within the ‘medium/high’ vs. the ‘low’ group. Extensive ECM was evidenced using Sirius red/alcian blue double staining. (C) Collagen and (D) GAG accumulation were quantified in different calcification groups. (E) The expression of miR-26b-5p in low, medium/high calcification samples and (F) correlation with the content of calcium (% alizarin red) (r = −0.4329, P = 0.0094; Spearman correlation; n = 35). For (B–E): n = 19 low (5 female/14 male) and n = 16 medium/high (4 female/12 male); *P < 0.05; **P < 0.01 vs. low calcification; Student’s unpaired t-test. All data are mean ± SEM.

Figure 2

miR-26bKO mice develop spontaneous aortic microcalcifications. (A) Representative ex vivo micro-PET/CT scans of the aortic arch and thoracic aorta in WT and miR-26bKO mice at 6 months. The colours correspond to the strength of the ¹⁸F-NaF (hydroxyapatite marker) signal. (B) Quantification of microcalcification in the aortic arch and thoracic aorta (n = 5 per group). (C) Representative images of alizarin red staining from WT and miR-26bKO aortas (left) and relative fold-change of calcified areas (right) (n = 5 per group). Scale bar = 50 μm. (D) Calcium deposition in the aortas (n = 5 per group). (E) Representative images of Elastic Van Giesson (EVG) staining from WT and miR-26bKO male aortas (left) and measured elastic fibre density (right), n = 7–8 per group. Scale bar = 50 μm. (F) Images of Sirius red-stained mouse aortas in brightfield (left) and polarized light (right), along with medial fibrosis, Intima-media thickness and Thick/thin fibres ratio measurement, n = 7–8 per group. Scale bar = 50 μm. Black arrows = interstitial fibrosis deposits. Dotted line = medial thickness. Luminal (A), medial (B), and adventitial (C) areas. Dotted lines delimit the luminal area. (G) Abdominal aorta stiffness was studied by pressure myography, (n = 5–6 per group; male mice). (H) Young’s modulus image mappings within mouse aorta (left) and media stiffness measurements measured by atomic force microscopy (right) (n = 6 per group). Male mice have been used in these experiments. For (B–F and H): ns = non-significant; *P < 0.05 vs. WT; **P < 0.01 vs. WT. Student’s unpaired t-test. For (G): *P < 0.05 vs. WT; two-way ANOVA. All data are mean ± SEM. Low-magnification images of the aortic rings have been displayed in Supplementary material online, Figure  S3A and B.

Figure 3

Clustering and mapping of cell groups in miR-26bKO mouse aorta. Single-cell transcriptomic analysis was performed in cells isolated from mouse aortas from 6-month-old WT and KO male mice (n = 5191). (A) UMAP plot composed of whole aortas from WT and KO mice, showing 12 cell clusters. (B) UMAP plot, composed of whole aortas from WT and KO mice, shows the total cell distribution. (C) Cluster frequency analysis in the aorta of WT and KO mice. (D) UMAP projection of the annotated clusters representing main cell types residing in the aorta. (E) Dot plot displaying the gene enrichment for gene ontologies (GO: Biological Process) (log10 of the false discovery rate) for cell types: ECs, fibroblasts (FBLs) and vascular smooth muscle cells (SMCs), comparing GEX levels between WT and miR-26bKO. The gene ratio (x-axis) indicates the number of overlapping genes divided by the total number of reference genes in the gene ontology (GO) category.

Figure 4

Network analysis of miR-26b target genes in the aorta of miR-26bKO mice. (A) Network analysis of DEGs between WT vs. KO mouse in smooth muscle cells (SMCs), ECs and fibroblasts (FBLs) (scRNA-Seq) and (B) whole aorta (bulk RNA-Seq). miR-26b targets are circled in red (P_adj < 0.05). (C) Quantification of GEX of miR-26b-5p and Smad1 within ECs, FBLs and SMCs from WT and miR-26bKO mouse aortas (n = 5 per group). (D) Representative immunohistochemical images and quantification of p-Smad1 media localization across WT and miR-26bKO mouse aortas samples (n = 5 per group). White arrows: p-Smad1 positive cells. Scale bar = 100 μm. (E) Representative Western blots for p-Smad1 (Ser463/465), total Smad1 and β-actin in WT and mir-26bKO mouse aortas and quantification of signal intensity compared to β-actin (loading control) (n = 5 per group). Male mice have been used in these experiments. For (C–E): **P < 0.01 vs. WT; One-way ANOVA. All data are mean ± SEM.

Figure 5

Analysis of cell–cell communication in the aorta of miR-26bKO mice. (A) Chord Diagram representing the inferred BMP network: the inner thinner bar colours represent the targets that receive a signal from the outer bar. The inner bar size is proportional to the signal strength received by the targets. Edge colours are consistent with the sources as senders, and edge weights are proportional to the interaction strength. A thicker edge line indicates a stronger signal. (B) Signalling roles for BMP pathways in the aorta of miR-26bKO mice in the form of dominant senders, receivers, mediators and influencers for each cell cluster/cell type are inferred from network centrality measures. (C) Dot plot presenting average gene expression of BMP pathway target genes in aortic ECs, fibroblasts (FBLs), and smooth muscle cells (SMCs). Average expression is represented by colour, with increased expression represented by a darker colour vs. decreased expression with a lighter colour. Percentage expression (%) is displayed according to the size of the dot, with a larger dot representing a higher percentage of a cell expressing a gene. (D) Expression analysis of Bmp2, Bmp4, and Bmp6 in isolated aortic SMCs and FBLs from WT and miR-26bKO mice (n = 6 per group [3 male/3 female]). (E) Bmp4 quantification in WT and miR-26bKO SMC or FBLs media (n = 5 per group [3 male/2 female]). (F) Dot plot presenting average gene expression of Smad1 target genes in aortic SMCs from WT and miR-26bKO mice. (G) Expression of Mfap4 and Id3 in WT and miR-26bKO SMCs treated with Bmp4 for 48 h (n = 5 per group [2 male/3 female]). Male and female mice have been used in (D, E, and G). For (D and E): **P < 0.01 vs. WT; One-way ANOVA. For (G): **P < 0.01 vs. WT + Bmp4; Two-way ANOVA. All data are mean ± SEM.

Figure 6

Inhibition of ALK receptors reduces calcium deposition in smooth muscle cells lacking miR-26b and microcalcification in miR-26KO aortas. (A) Analysis of p-Smad1 nuclear translocation in miR-26bKO SMCs treated with WT or miR-26bKO FBL-conditioned media in the presence or absence of LDN-193189 (500 nM) for 30 min (n = 5 per group [3 male/2 female]). Scale bar: 50 μm. (B) Representative Western blots for p-Smad1 (Ser463/465) and total Smad1 in miR-26bKO SMCs treated as above and quantification of signal intensity compared to Smad1 (n = 5 per group [3 male/2 female]). (C) Expression of Mfap4 and Id3 in miR-26bKO SMCs treated with WT or miR-26bKO FBL-conditioned media in the presence or absence of LDN-193189 for 48 h (n = 5 per group [3 male/2 female]). (D) Analysis of calcium deposition in WT and miR-26bKO SMCs culture in high-phosphate media (2.6 mM). Cells were treated with WT or miR-26bKO FBL-conditioned media in the presence or absence of LDN-193189 for 10 days (n = 5 per group [3 male/2 female]). (E) Aortic SMCs and FBLs isolated from WT and miR-26bKO mice were co-cultured in high-phosphate media in a Transwell™ with SMCs on the abluminal side of the insert membrane and FBLs on the luminal side. Cells were cultured in the presence or absence of LDN-193189 for 10 days (n = 5 per group [3 male/2 female]). (F) Representative ex vivo micro-PET/CT scans of the heart and descending aorta in wildtype, miR-26bKO mice at 6 months. The colours correspond to the strength of the ¹⁸F-NaF (hydroxyapatite marker) signal. (G) Quantification of microcalcification in the aortic arch and thoracic aorta (n = 5 per group). (H) Calcium deposition within the mouse aorta (n = 6–8 per group). (I) Immunostaining of p-Smad1 within the aorta of WT, miR-26bKO and miR-26bKO mice treated with LDN-193189. Scale bar = 100 μm. White arrows: p-Smad1 positive cells and quantification of p-Smad1 positive aortic media cells (n = 5 per group). (J) Representative Western blots for p-Smad1 (Ser463/465), total Smad1 and β-actin in aortas of WT, miR-26bKO and miR-26bKO mice treated with LDN-193189 and quantification of signal intensity compared to β-actin (loading control) (n = 5 per group). For (A–D): **P < 0.01 vs. WT FM; ^##P < 0.01 vs. KO FM; For (E): **P < 0.01 vs. WT FBLs; ^##P < 0.01 vs. KO FBLs; Two-way ANOVA. Male and female mice have been used in these experiments. For (G–J): ns, non-significant; *P < 0.05; **P < 0.01 vs. WT; ^#P < 0.05; ^##P < 0.01 vs. KO; one-way ANOVA. All data are mean ± SEM. Male mice have been used in these experiments.

Figure 7

Inverse correlation between miR-26b and SMAD1 in calcific aortic tissue. (A) The expression of SMAD1 in low and medium/high calcification samples (B) and correlation with the level of miR-26b-5p (r = −0.3536, P = 0.0372; Spearman correlation). (C) In situ hybridization for miR-26b-5p, immunostaining of von Kossa and EVG and immunohistochemistry of p-SMAD1 within low and moderately calcified human ascending thoracic aortic aneurysm tissue. Scale bar = 500 μm. (D) Quantification of miR-26b-5p positive signals per cell or p-SMAD1 positive signals per nucleus in patient samples with low or moderate microcalcification and association between miR-26b-5p and SMAD1 expression within the same sample. For (A) n = 19 low (5 female/14 male) and n = 16 medium/high (4 female/12 male); *P < 0.05; **P < 0.01 vs. low calcification; Student’s unpaired t-test. For (D), *P < 0.05 vs. low calcification. Student’s unpaired t-test (n = 5 per group). All data are mean ± SEM.