MicroRNA-22 promoted osteogenic differentiation of valvular interstitial cells by inhibiting CAB39 expression during aortic valve calcification

Abstract

Calcific aortic valve disease (CAVD) is a common valve disease characterized by the fibro-calcific remodeling of the aortic valves, which is an actively regulated process involving osteogenic differentiation of valvular interstitial cells (VICs). MicroRNA (miRNA) is an essential regulator in diverse biological processes in cells. The present study aimed to explore the role and mechanism of miR-22 in the osteogenic differentiation of VICs. The expression profile of osteogenesis-related miRNAs was first detected in aortic valve tissue from CAVD patients (n = 33) and healthy controls (n = 12). miR-22 was highly expressed in calcified valve tissues (P < 0.01), and the expression was positively correlated with the expression of OPN (r_s = 0.820, P < 0.01) and Runx2 (r_s = 0.563, P < 0.01) in VICs isolated from mild or moderately calcified valves. The sustained high expression of miR-22 was also validated in an in-vitro VICs osteogenic model. Adenovirus-mediated gain-of-function and loss-of-function experiments were then performed. Overexpression of miR-22 significantly accelerated the calcification process of VICs, manifested by significant increases in calcium deposition, alkaline phosphate activity, and expression of osteoblastic differentiation markers. Conversely, inhibition of miR-22 significantly negated the calcification process. Subsequently, calcium-binding protein 39 (CAB39) was identified as a target of miR-22. Overexpression of miR-22 significantly reduced the expression of CAB39 in VICs, leading to decreased catalytic activity of the CAB39–LKB1–STRAD complex, which, in turn, exacerbated changes in the AMPK–mTOR signaling pathway, and ultimately accelerated the calcification process. In addition, ROS generation and autophagic activity during VIC calcification were also regulated by miR-22/CAB39 pathway. These results indicate that miR-22 is an important accelerator of the osteogenic differentiation of VICs, and a potential therapeutic target in CAVD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04177-6.

Introduction

Calcific aortic valve disease (CAVD) is the most common primary valve disease, and it has an increasing prevalence due to an aging population [1, 2]. CAVD has been found to be a multifaceted and actively regulated process driven by various molecular mechanisms, including endothelial injury and apoptosis, osteogenic reprogramming of valvular interstitial cells (VICs), oxidative stress and inflammation, and lipid metabolism [3]. Echocardiography can be used to confirm and assess the presence of aortic stenosis and the degree of valve calcification, which serve as the key criteria for the diagnosis and evaluation of CAVD [2]. Surgery and catheter intervention are the only treatment options for CAVD. No therapy is effective for halting the progression of CAVD [4]. Therefore, new therapeutic targets and treatment strategies are critically needed for CAVD.

VICs are important for maintaining valvular structure and function owing to their ability to undergo trans-differentiation to valvular osteoblast-like cells (VOCs). Because VICs are the prevalent cells in aortic valves, their osteogenic differentiation may underlie the formation of calcified nodule in CAVD pathogenesis [5]. In calcified aortic valves, the majority of VICs are activated and undergo trans-differentiation into other cell types, including osteoblasts, that express bone markers and matrix proteins, including osteopontin (OPN) and runt-related transcription factor 2 (Runx2), and lead to valve calcification [6]. Therefore, osteogenic differentiation of VICs is implicated as the key pathological features in CAVD [5]. Understanding the molecular mechanisms underlying the phenotypic transition of VICs is urgently needed for developing therapeutic strategies in CAVD.

MicroRNAs (miRNAs) are an important class of small non-coding RNAs that are essential in diverse biological processes. Some miRNAs have been identified as potential therapeutic targets in cardiovascular diseases, including pathological cardiac hypertrophy, atherosclerosis, and heart failure. Studies on miRNA expression profiles and function in calcific aortic valves and VICs have suggested that aberrantly expressed miRNAs might be responsible for the pro‐osteogenic phenotype of VICs in CAVD pathogenesis. In particular, miR-22, which is universally expressed in various tissues, has been confirmed to be associated with multiple human disorders, including diverse tumors and cardiovascular diseases [7–9]. In addition, an miR-22-based therapy for elderly patients with myocardial infarction has been preclinically developed [10]. The present study aimed to investigate the role and mechanism of miR-22 during CAVD pathogenesis. The results could provide insights into the molecular mechanism underlying CAVD pathogenesis.

Materials and methods

Clinical samples

Calcified aortic valve tissues were obtained from patients with CAVD who were undergoing aortic valve replacement at the Department of Cardiovascular Surgery of Changhai Hospital between September 2015 and February 2018. Patients with a history of rheumatic heart disease, endocarditis, or congenital heart disease were excluded. Non-calcified aortic valves with normal echocardiographic analyses were obtained from patients undergoing heart transplantation and served as controls. The protocol was approved by the Medical Ethics Committee of Changhai Hospital and the study conforms to the Declaration of Helsinki.

Assessment of aortic valve calcification

Aortic valve calcification was assessed by transthoracic echocardiography based on a previously reported aortic valve calcification scoring system (Table S1) [11]. The calcified aortic valves were divided into three groups: mild (< 1/3 of the valve leaflet area highly echogenic), moderate (1/3–2/3 of the valve leaflet area highly echogenic), and severe (> 2/3 of the valve leaflet area highly echogenic).

Isolation and culture of aortic VICs

Pathological VICs and normal VICs were isolated from the non-calcific areas of mildly or moderately calcified and non-calcified aortic valves, respectively [12]. Valve leaflets were digested in type I collagenase solution (1.0 mg/mL) for 30 min, and scratched to eliminate valvular endothelial cells (VECs). Then, the valve leaflets were digested in fresh collagenase solution for 6 h with shaking, followed by centrifugation and resuspension. Cells were maintained in M199 growth medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. The purity of VICs was confirmed by immunofluorescence assay, and cultures were passaged for 4–6 generations for experiments.

Histological and immunohistochemical assay

The histological and immunohistochemical assays were performed in paraffin sections of aortic valve tissues, as described previously [13]. For immunohistochemistry staining, the deparaffinized and rehydrated sections were incubated with primary antibody overnight, followed by incubation with the appropriate secondary antibody. The visualization of staining signaling was accomplished using a DAB horseradish peroxidase color development kit (Beyotime, Shanghai, China). For histological examination, the deparaffinized and rehydrated sections were stained with Alizarin Red S (Sigma-Aldrich, St. Louis, MO, USA) or von Kossa solution for ionized calcium deposits. Images of the tissue sections were captured by an optical microscope (Olympus, Tokyo, Japan).

Fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) for miR-22 detection was performed on fresh-frozen sections of aortic valves, as described previously [14]. Briefly, the tissue sections were acetylated in triethanolamine buffer plus 0.25% acetic anhydride. After prehybridization, hybridization was conducted at 56 °C overnight in the hybridization buffer containing digoxigenin-labeled locked nucleic acid probe for miR-22 (Exiqon, Vedbaek, Denmark). Then, the slides were sequentially washed with hybridization buffer, 1 × SSC in hybridization buffer, and MABT buffer, and incubated in blocking solution, followed by incubation with Anti-Digoxigenin-POD (Roche, Basel, Switzerland) overnight at 4 °C. Finally, the signals were developed using Alexa Fluor 633 Dye-conjugated secondary antibody.

In vitro calcification assessment and quantification

The in vitro calcification of VICs was induced by osteogenic medium (M199 complete medium containing 0.1 μmol/L insulin, 50 μg/mL ascorbic acid, and 2 mmol/L NaH₂PO₄) for 7 days [15]. For Alizarin Red S staining, 4% formaldehyde-fixed VICs were stained with Alizarin Red S solution for the observation of calcium deposits. For calcium deposition quantification, VICs were decalcified with 0.6 N HCl for 24 h, and the calcium content in the supernatants was determined with a calcium colorimetric assay kit (Sigma-Aldrich). The total protein content of VICs was used for data normalization. The alkaline phosphate (ALP) activity of VICs was detected with an ALP test kit (Beyotime). The expression of osteogenic markers (OPN and Runx2) was detected by quantitative real-time PCR (qRT-PCR) and western blot assays.

Quantitative real-time PCR (qRT-PCR)

Total RNA was isolated from VICs or aortic valve tissues by MiniBEST Universal RNA Extraction Kit (TaKaRa, Dalian, China) according to the manufacturer’s protocol. Equal amounts (200 ng) of RNA samples were reverse transcribed to complementary DNA (cDNA) with PrimeScript RT reagent Kit (TaKaRa). qRT-PCR was performed on a LightCycler 480 II PCR system (Roche, Basel, Switzerland) using TB Green Premix Ex TaqII (TaKaRa). The relative gene expression level was analyzed using the 2^(−ΔΔCt) method. β-actin and RNU6b were used as loading controls for mRNA and miRNA, respectively. The sequences of primers are presented in Table S2.

Western blot analysis

Equal amounts of protein (about 30 μg) extracted from cells or aortic valve tissues were resolved on SDS-PAGE gel, and then transferred to polyvinylidene difluoride membranes. After blocking, membranes were incubated with diluted primary antibodies at 4 °C overnight, followed by incubation with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody. The immunoreactive bands were visualized by ECL Plus Western Blotting Substrate (Thermo Fisher Scientific, Waltham, MA, USA) on ChemiDoc MP system (Bio-Rad, Hercules, CA, USA). Immunoblots were quantified using ImageJ software. β-Actin was used as the loading control. Information on the used antibodies is presented in Table S3.

Construction and infection of recombinant adenovirus

Recombinant adenoviruses were produced using the AdMax system (Hanbio Biotechnology, Shanghai, China) according to the manufacturer’s instructions. The adenovirus titer in plaque-forming units (pfu) was determined by plaque formation assay following the infection of 293 cells. The optimal multiplicity of infection (MOI) of recombinant adenovirus in VICs was determined as 5 pfu/cell.

Bioinformatics analysis

The potential targets of miR-22 were predicated by the PITA algorithm [16], and are listed in Table S4. The RNAhybrid and miRanda algorithms were used to analyze the complementary matching and minimum free energy of hybridization between miR-22 and 3ʹ-untranslated region (3ʹ-UTR) of target mRNA [17, 18]. GO and KEGG analyses were performed with OECloud tools (https://cloud.oebiotech.cn) based on p value to gene number, gene numbers involved in selected term, and enrichment score.

Co-immunoprecipitation assay

The cells were lysed in RIPA buffer and centrifuged at 4 °C. The supernatants were incubated with anti-LKB1 antibody (1:100, Abcam, Cambridge, UK) overnight, and then mixed with protein A beads (Thermo Fisher Scientific) at 4 °C. After centrifugation, the beads were washed with RIPA buffer. The complexes were eluted with SDS-sample loading buffer and subjected to western blot analysis.

Reactive oxygen species (ROS) detection assay

Intracellular ROS level was determined with non-fluorescent probe 2ʹ,7ʹ-dichlorofluorescein diacetate assay (DCFH-DA. Beyotime). The treated VICs were incubated with 10 μmol/L DCFH-DA for 20 min at 37 °C. After washing, fluorescent signal was observed and photted under a fluorescence inverted microscope (Olympus), or measured by a microplate reader (BioTek, Winooski, VT, USA) at 535 nm.

Statistical analysis

Statistical analyses were performed with SPSS v22.0 software. Shapiro–Wilk test was used to evaluate the normality of the quantitative data. The difference between two groups was analyzed by two-tailed Student’s t test (normal distribution) or Mann–Whitney U test (non-normal distribution). The difference between multiple groups was analyzed by one-way ANOVA followed by post hoc Tukey test (homogeneity of variance) or one-way ANOVA followed by Dunnett’s T3 test (heterogeneity of variance). The qualitative data were compared with Fisher’s exact test. A P value < 0.05 was considered to indicate a statistical significance.

Results

Baseline and clinical characteristics of human participants

The baseline characteristics and echocardiographic data of all participants are shown in Table 1. A total of 33 CAVD patients were enrolled and analyzed, with groups defined by mild calcification (n = 10), moderate calcification (n = 12), and severe calcification (n = 11). No significant differences in clinical parameters between control (n = 12) and CAVD (n = 33) groups (Table 1). Alizarin Red S (Fig. 1A) and Von Kossa (Fig. 1B) staining showed that calcified valves from CAVD patients had significantly more calcification nodules. Meanwhile, elevated expression of osteogenic markers, including OPN and Runx2, was detected in calcified valves by qRT-PCR (Fig. 1C, D) and western blot assays (Fig. 1E and Fig. S1).

Table 1.

Clinical characteristics of subjects enrolled in this study

Parameters	CAVD group (n = 33)	Control group (n = 12)	P
Age (years)	63.5 ± 8.6	68.0 ± 8.1	0.119
Male, n (%)	15 (45.5)	5 (41.7)	1.000
BMI (kg/m2)	23.7 ± 4.1	24.3 ± 2.2	0.693
History
Hypertension, n (%)	14 (42.4)	2 (16.7)	0.164
Diabetes mellitus, n (%)	8 (24.2)	3 (25.0)	1.000
Coronary heart disease, n (%)	12 (36.4)	4 (33.3)	1.000
Hyperlipidaemia, n (%)	7 (21.2)	2 (16.7)	1.000
Smoking, n (%)	11 (33.3)	5 (41.7)	0.728
Laboratory findings
HDL-C (mmol/L)	1.25 ± 0.30	1.28 ± 0.38	0.782
LDL-C (mmol/L)	2.83 ± 0.93	2.50 ± 0.92	0.297
Triacylglycerols (mmol/L)	1.38 ± 0.70	1.31 ± 0.55	0.747
Creatine (μmol/L)	77.76 ± 20.11	75.08 ± 13.23	0.672
Medication
ACEI/ARB, n (%)	11 (33.3)	1 (8.3)	0.136
Beta-blockers, n (%)	9 (27.3)	3 (25.0)	1.000
Diuretics, n (%)	5 (15.2)	1 (8.3)	1.000
Echocardiography
LVEF (%)	56.4 ± 9.2	59.9 ± 4.7	0.211
Aortic valve area (mm2)	0.80 ± 0.38	/	/
Aortic mean gradient (mmHg)	70.2 ± 37.5	/	/

BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, HDL-C high-density lipoprotein-cholesterol, LDL-C low-density lipoprotein-cholesterol, ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin II receptor blocker, LVEF left-ventricular ejection fraction

Fig. 1.

Detection of calcification in aortic valve tissues from control (n = 12) and CAVD patients (n = 33). A Representative images of Alizarin Red S staining (red) for detection of calcified nodules in aortic valve tissues. Scale bar = 50 μm. B Representative images of Von Kossa staining (black) for detection of calcified nodules in aortic valve tissues. Scale bar = 50 μm. C, D,Relative mRNA expression of OPN and Runx2 in aortic valve tissues from control (n = 12), mild (n = 10), moderate (n = 12), and severe (n = 11) groups. β-actin was used as the loading control. **P < 0.01. E Representative images of western blots for OPN and Runx2 protein expression in aortic valve tissues. β-actin was used as the loading control

Upregulated expression of miR-22 was detected in aortic valves from CAVD patients

Based on the previous findings on miRNA function [19, 20], 12 miRNAs, that were reportedly involved in calcification or osteogenic differentiation, were selected as candidate targets. To identify the aberrant expression of miRNAs in calcified aortic valve, qRT-PCR was used to detect the expression of osteogenic miRNAs in a small cohort of CAVD patients. In comparison with the control group (n = 5), miR-21, miR-22, and miR-125a were significantly increased (P < 0.01), while miR-10, miR-214, and miR-204 were significantly decreased in the CAVD group (n = 5, P < 0.01). No significant changes were found in the expression of miR-138, miR-148a, miR-496, and miR-375 (P > 0.05. Figure 2A). The highly expressed miR-22 in calcified aortic valves was then confirmed in a larger cohort of CAVD patients (n = 33, P < 0.01. Figure 2B). However, no significant difference in miR-22 expression was found in valve tissues from CAVD patients with different calcification severity (P > 0.05. Fig. S2A). FISH assay confirmed that miR-22 was observed throughout the calcified valves and was localized in VICs, as indicated by the co-expression of vimentin (Fig. 2C and Fig. S2B).

Fig. 2.

miR-22 was highly expressed in aortic valve tissues from CAVD patients. A Relative expression level of osteogenesis-related miRNAs in aortic valve tissues from control and CAVD patients. n = 5 in each group. RNU6b was used as the loading control. *P < 0.05, and **P < 0.01. B Relative expression level of miR-22 in aortic valve tissues from control (n = 12) and CAVD patients (n = 33). RNU6b was used as the loading control. **P < 0.01. C Representative images of fluorescence in situ hybridization for miR-22 colocalization in aortic valve tissues. miR-22 (red) could be detected in vimentin-positive VICs (green). Nucleus was stained with DAPI (blue). The white asterisk represents the signal of miR-22 in VICs. Scale bar = 50 μm

The elevated miR-22 level was correlated with osteogenic differentiation of VICs

To further identify the role of miR-22 in the osteogenic differentiation of VICs, primary aortic VICs were isolated from mildly or moderately calcified valves of CAVD patients. Immunofluorescence staining confirmed that the isolated cells were positive for vimentin and negative for CD31 (Fig. 3A and Fig.S3). qRT-PCR revealed that the expression of miR-22 was significantly higher in VICs from moderate group (n = 7) than that from mild group (n = 8, P < 0.01. Figure 3B). Based on the expression of miR-22, OPN, and Runx2 in these VICs (n = 15), Spearman test demonstrated that miR-22 expression was positively correlated with OPN (r_s = 0.820, P < 0.01) and Runx2 (r_s = 0.563, P < 0.01. Figure 3C, D). Next, an in vitro osteogenic model was established in VICs isolated from normal valves using osteogenic medium. Alizarin Red S staining showed significant calcium nodules could be detected after 3 days of calcification induction (Fig. 3E). The significantly elevated calcium deposition, ALP activity, and osteogenic markers expression (OPN and Runx2) could be detected after 3 days of stimulation (Fig. 3F–I). qRT-PCR assay revealed that expression of miR-22 was significantly upregulated after 3 days of stimulation, and was maintained at a high level in the following days (Fig. 3J).

Fig. 3.

The elevated miR-22 level was correlated with VICs’ osteogenic differentiation. A Representative images of immunofluorescence staining for primary isolated VICs. The left panel is the representative images of optical light microscope observation, and the right panel is the fluorescence microscope observation under the same field of view. VICs were positive for vimentin (red) and negative for CD31 (green). Nucleus was stained with DAPI (blue). Scale bar = 100 μm. B Relative expression level of miR-22 in VICs isolated from aortic valves with different degrees of calcification, including mild (n = 8) or moderate (n = 7). **P < 0.01. C, D Correlation analysis between miR-22 and osteogenic markers (OPN and Runx2) in the above-mentioned primary VICs (n = 15). E Detection of mineralized nodule formation in VICs after calcification induction by Alizarin Red S staining. The lower panel shows the optical microscope images with magnification. Scale bar = 50 μm. F Detection of the calcium concentration in VICs after Alizarin Red S staining. **P < 0.01. G Detection of ALP activity in VICs after calcification induction. **P < 0.01. H, I, Relative mRNA expression level of OPN and Runx2 in VICs after calcification induction. β-actin was used as the loading control. **P < 0.01. J Relative expression level of miR-22 in VICs after calcification induction. RNU6b was used as the loading control. **P < 0.01

Overexpression of miR-22 promoted the osteogenic differentiation of VICs

To clarify the function of miR-22 in VICs osteogenic differentiation, adenovirus-mediated gain- and loss-of-function approaches were used. qRT-PCR analysis confirmed that the infection of VICs with Ad-miR-22 could induce a significant increase in miR-22 expression over the endogenous level for 7 days (Fig. S4A). As analyzed by Alizarin Red S staining, significantly increased calcium nodules could be observed in miR-22-overexpressing VICs after calcification induction (Fig. 4A). Meanwhile, the overexpression of miR-22 resulted in further increases in calcium deposition (Fig. 4B), ALP activity (Fig. 4C), and mRNA and protein levels of osteoblastic markers (Fig. 4D–F and Fig. S4B) in VICs. In contrast, silencing of miR-22 in VICs negated the formation of calcified nodules (Fig. 4G), calcium deposition (Fig. 4H), ALP activity (Fig. 4I), and expression of osteoblastic markers (Fig. 4J–L and Fig. S4C).

Fig. 4.

miR-22 regulated the osteogenic differentiation process of VICs. A Detection of mineralized nodule formation in VICs with miR-22 overexpression by Alizarin Red S staining after 7 day calcification induction. The lower panel shows the optical microscope images with magnification. Scale bar = 50 μm. B Detection of the calcium concentration in VICs after Alizarin Red S staining. **P < 0.01. C Detection of the ALP activity in VICs with miR-22 overexpression. *P < 0.05. D, E Relative mRNA expression levels of OPN and Runx2 in VICs with miR-22 overexpression. β-actin was used as the loading control. **P < 0.01. F Representative images of western blots for OPN and Runx2 protein expression in VICs with miR-22 overexpression after 2 day calcification induction. β-actin was used as the loading control. G Detection of mineralized nodule formation in VICs with miR-22 inhibition by Alizarin Red S staining after 7 day calcification induction. The lower panel shows the optical microscope images with magnification. Scale bar = 50 μm. H Detection of the calcium concentration in VICs after Alizarin Red S staining. **P < 0.01. I Detection of ALP activity in VICs with miR-22 inhibition. *P < 0.05. J, K Relative mRNA expression levels of OPN and Runx2 in VICs with miR-22 inhibition after 2 day calcification induction. β-actin was used as the loading control. *P < 0.05. L Representative images of western blots for OPN and Runx2 protein expression in VICs with miR-22 inhibition after 2 day calcification induction. β-actin was used as the loading control

Calcium-binding protein 39 was identified as a direct target of miR-22

It has been acknowledged that miRNAs fine-tune protein expression by binding to 3ʹ-UTR of targets’ mRNA. Therefore, bioinformatics analysis was applied to search for the potential targets of miR-22. The prediction results showed that miR-22-binding sites were presented in the 3’-UTR of many genes (Table S4). Based on the enrichment score, these potential target genes were associated with the neuromuscular junction and postsynaptic density in the cellular component analysis, linked with protein domain specific binding and transcription factor binding in the molecular function analysis, and related to osteoblast development and negative regulation of pri-miRNA transcription by RNA polymerase II in the biological process analysis (Fig. 5A). The top 20 pathways of enrichment in KEGG analysis are shown in Fig. 5B, including Oocyte meiosis, signaling pathways regulating pluripotency of Stem Cells, endocrine resistance, and longevity regulating pathway-multiple species, etc. Among them, LKB1 pathway, an important upstream serine/threonine kinase of AMPK, was also determined as an enriched pathway. Then, we focused on these putative miR-22 targets involved in the LKB1 pathway. Combining RNAhybrid and miRanda algorithms analysis, calcium-binding protein 39 (CAB39) was selected as a putative target of miR-22. One miR-22-binding site with tight binding potential was identified in the 3ʹ-UTR of CAB39 by RNA-hybrid program (Fig. 5C). Dual-luciferase reporter assay further confirmed that overexpression of miR-22 significantly decreased the activity of the luciferase reporter containing wild-type sequences of CAB39 3ʹ-UTR, but had no obvious effect on the activity of the luciferase reporter containing mutant sequences (Fig. 5D). Furthermore, compared with the Scramble group, the protein expression of CAB39 was significantly decreased in the Ad-miR-22 group, while significantly increased in the shmiR-22 group (Fig. 5E, F, and Fig. S5).

Fig. 5.

Calcium-binding protein 39 (CAB39) was a direct target of miR-22. A, B Gene Ontology (GO) and KEGG pathway enrichment analysis of the miR-22 potential targets predicted by bioinformatics. Top 20 terms of enrichment were listed, respectively. C A schematic of the full-length information of human CAB39 mRNA. One 13-nt binding site of miR-22 is located in the 3ʹ-UTR of CAB39. UTR untranslated regions, CDS coding sequence, wt wild type, mt mutant type. D Dual-luciferase reporter assay verified the binding of miR-22 and wild- or mutant‐type 3ʹ-UTR of CAB39. **P < 0.01. E, F Representative images of western blots for CAB39 protein expression in VICs with miR-22 overexpression or inhibition after 2 day calcification induction. β-actin was used as the loading control

miR-22 promoted osteogenic differentiation of VICs by inhibiting CAB39 expression

To further substantiate whether miR-22 relies on CAB39 to promote VICs osteogenic differentiation, the expression profile of CAB39 in VICs osteogenic model was first investigated. Western blot assay confirmed that the expression of CAB39 was significantly decreased in VICs treated with osteogenic medium (Fig. 6A and Fig. S6A). Next, adenovirus-mediated CAB39 gain-of-function and loss-of-function studies were performed in VICs. Overexpression of CAB39 could delay the osteogenic differentiation process of VICs, as indicated by the results from Alizarin Red S staining and assays for calcium deposition, ALP activity, and detection of osteogenic marker expression. In contrast, inhibition of CAB39 accelerated the process of osteogenic differentiation (Fig. 6B–E and Fig. S6B–D). Rescue experiments were performed in VICs by co-overexpression of miR-22 and CAB39. The results showed that overexpression of CAB39 could neutralize the pro-osteogenic differentiation effect of miR-22 in VICs, which was demonstrated by the reduced calcium nodule formation and ALP activity (Fig. 6F–I, and Fig. S6E–F), as well as the decreased expression of OPN and Runx2 (Fig. 6I and Fig. S6G). Moreover, overexpression of miR-22 had almost no pro-osteogenic effect on CAB39 silenced VICs (Fig. 6J–M and Fig. S6H), suggesting the importance of CAB39 in miR-22 pro-osteogenic function.

Fig. 6.

miR-22 promoted osteogenic differentiation of VICs by inhibiting CAB39 expression. A Representative images of western blots for Cab39 protein expression in VICs after calcification induction. β-actin was used as the loading control. B Detection of mineralized nodule formation in VICs with CAB39 expression intervention after 7 day calcification induction. The lower panel shows the optical microscope images with a magnification. Scale bar = 50 μm. C The intracellular calcium concentration in VICs with CAB39 expression intervention after 7 day calcification induction. *P < 0.05, and **P < 0.01. D ALP activity in VICs with CAB39 expression intervention after 7 day calcification induction. **P < 0.01. E Representative images of western blots for OPN and Runx2 protein expression in VICs with CAB39 expression intervention after 2 day calcification induction. β-actin was used as the loading control. F Detection of mineralized nodule formation in VICs with miR-22 and CAB39 co-overexpression after 7 day calcification induction. The lower panel shows the optical microscope images with a magnification. Scale bar = 50 μm. G The intracellular calcium concentration in VICs with miR-22 and CAB39 co-overexpression after 7 day calcification induction. *P < 0.05, and **P < 0.01. H ALP activity in VICs with miR-22 and CAB39 co-overexpression after 7 day calcification induction. **P < 0.01. I Representative images of western blots for OPN and Runx2 protein expression in VICs with miR-22 and CAB39 co-overexpression after 2 day calcification induction. β-actin was used as the loading control. J Detection of mineralized nodule formation in CAB39 silenced VICs with miR-22 overexpression after 7 day calcification induction. The lower panel shows the optical microscope images with a magnification. Scale bar = 50 μm. K The intracellular calcium concentration in CAB39 silenced VICs with miR-22 overexpression after 7 day calcification induction. *P < 0.05. L, ALP activity in CAB39 silenced VICs with miR-22 overexpression after 7 day calcification induction. M Representative images of western blots for OPN and Runx2 protein expression in CAB39 silenced VICs with miR-22 overexpression after 2 day calcification induction. β-actin was used as the loading control

miR-22 regulated the AMPK–mTOR pathway by CAB39–LKB1–STRAD catalytic activity

CAB39 has been reported to markedly increase LKB1 catalytic activity by enhancing the binding of STRAD to LKB1 [21]. Therefore, the binding relationship among CAB39, STRAD, and LKB1 was first verified in VICs by co-immunoprecipitation assay. The results showed that the endogenous CAB39 and STRAD in VICs could be co-immunoprecipitated by LKB1 antibody. The infection of shCAB39 significantly inhibited the protein expression of CAB39, leading to a reduction in STRAD binding to LKB1, but it did not affect the expression of STRAD and LKB1 (Fig. 7A). Meanwhile, the expression of LKB1 and STRAD also did not significantly change in VICs with overexpression of miR-22 and CAB39 (Fig. 7B and Fig. S7A). Next, we detected the expression of the CAB39–LKB1–STRAD complex during osteogenic differentiation of VICs. Consistent with the change in CAB39 expression, western blot assay results confirmed that LKB1 and STRAD protein expression was also decreased in VICs after osteogenic induction (Fig. 7C and Fig. S7B). The CAB39–STRAD–LKB1 complex was identified as a kinase of AMP-activated protein kinase (AMPK). Therefore, we further detected the expression of AMPK and its downstream mTOR signaling pathways in VICs with miR-22 and CAB39 overexpression under osteogenic induction condition. The expressions of phosphorylated AMPK (p-AMPK) and phosphorylated TSC2 (p-TSC2) were significantly decreased in VICs after miR-22 overexpression, while the expression of phosphorylated mTOR (p-mTOR) was significantly increased. However, overexpression of CAB39 was able to neutralize the role of miR-22 in the regulation of AMPK and mTOR signaling pathways (Fig. 7D and Fig. S7C, D).

Fig. 7.

miR-22 regulated the AMPK–mTOR pathway by CAB39–LKB1–STRAD catalytic activity. A Immunoprecipitation assay confirmed the binding relationship between CAB39, STRAD, and LKB1 in VICs with miR-22 overexpression. B Representative images of western blots for CAB39–LKB1–STRAD protein expression in VICs with miR-22 and/or CAB39 overexpression. β-actin was used as the loading control. C Representative images of western blots for CAB39–LKB1–STRAD protein expression in the osteogenic differentiation process of VICs. D Representative images of western blots for the AMPK–mTOR pathway signaling in VICs with miR-22 and/or CAB39 overexpression. β-actin was used as the loading control

miR-22/CAB39 regulated ROS production and autophagy in VIC calcification

It has been reported that miR-22 could regulate ROS accumulation and autophagy in bone marrow mesenchymal stromal cells [22]. Therefore, the role of miR-22 on ROS production and autophagy was further investigated in VIC calcification. After 2 day calcification induction, ROS production was significantly increased in VICs. Overexpression of miR-22 significantly accelerated ROS accumulation in VICs, while overexpression of CAB39 could counteract the promotion role of miR-22 (Fig. 8A, B). Then, the dynamical autophagic activity was detected in the calcification process of VICs (Fig. 8C and Fig. S8 A, B). Under calcification induction conditions, overexpression of miR-22 could reduce the expression of beclin 1 and conversion of soluble LC3β-I to lipid-bound LC3β-II (LC3 II/I) in VICs (Fig. 8D and Fig. S8C, D), indicating that overexpression of miR-22 could inhibit the autophagic activity. Moreover, on the 3rd day of calcification induction, it could be found that the inhibition role of miR-22 on autophagic activity could be partially abolished by CAB39, demonstrated by the decrease of beclin 1 expression and LC3 II/I (Fig. 8E and Fig. S8E, F). However, there were no significant changes in the expression of Beclin 1 and LC3 II/I in VICs overexpressing miR-22 in the late stage of calcification (day 7), suggesting that autophagy was basically in the final phases in severely calcified VICs.

Fig. 8.

miR-22/CAB39 regulated ROS production and autophagy in VIC calcification. A, B The intracellular ROS production was detected by 2,7‐dichlorodi‐hydrofluorescein diacetate (DCFH‐DA, 10 μmol/L) in VICs after 2 day calcification induction. The fluorescent intensity was acquired and analyzed by Scanning Systems. Scale bar = 50 μm. **P < 0.01. C Representative images of western blots for Beclin 1 and LC3 protein expression during VIC calcification. β-actin was used as the loading control. D Representative images of western blots for Beclin 1 and LC3 protein expression in VICs with miR-22 overexpression after 1-, 3-, or 5-day calcification induction. β-actin was used as the loading control. E Representative images of western blots for Beclin 1 and LC3 protein expression in VICs with miR-22 and/or CAB39 overexpression after 3- or 7-day calcification induction. β-actin was used as the loading control

Correlation between CAB39 expression and valve calcification in CAVD

To investigate the clinical significance of CAB39 in CAVD, we detected its expression profile in valve tissues from CAVD patients with different levels of calcification severity. Intense immunohistochemistry staining of CAB39 was found in normal and mildly calcified valve tissues, while weak CAB39 staining was found in moderately calcified valve tissues. The immunohistochemical staining for CAB39 was rarely detectable in severely calcified valve tissues (Fig. 9A and Fig. S9A). Then, western blot assay was used to detect the expression of CAB39 and osteogenic markers (OPN and Runx2) in valve tissues from 33 CAVD patients and 12 control participants. Consistent with the immunohistochemistry results, the protein expression of CAB39 was significantly higher in normal valves than in calcified valve tissues (Fig. 9B and Fig. S9B–D). Moreover, the CAB39 protein level was significantly negatively correlated with OPN (r_s =  − 0.623, P < 0.01) and Runx2 (r_s =  − 0.438, P < 0.01) expression in CAVD patients (Fig. 9C, D), suggesting that decreased CAB39 expression was associated with valve calcification severity.

Fig. 9.

Correlation between CAB39 expression and valve calcification in CAVD. A Representative images of immunohistochemistry staining for CAB39 in aortic valve tissues from control (n = 12), mild (n = 10), moderate (n = 12), and severe (n = 11) groups. Scale bar = 50 μm. The lower panel shows the optical microscope images with magnification. B Representative images of western blots for CAB39, OPN, and Runx2 protein expression in aortic valve tissues from control (n = 12) and CAVD patients (n = 33). β-actin was used as the loading control. C, D Correlation analysis between CAB39 and osteogenic markers (OPN and Runx2) in calcified aortic valve tissues. Spearman test demonstrated that CAB39 expression was negatively correlated with OPN (r_s =  − 0.623, P < 0.01) and Runx2 (r_s =  − 0.438, P < 0.01)

Discussion

In the present study, we found that the expression of miR-22 was markedly altered in calcified aortic valve tissues from CAVD patients, and altered expression of miR-22 was responsible for the human aortic VIC pro-osteogenic phenotype. Furthermore, we demonstrated that CAB39 was a novel target gene for miR-22, and overexpression of CAB39 was protective against progression of osteogenic differentiation in aortic VICs through activation of the AMPK–mTOR signaling pathway. Thus, these findings revealed potential new mechanisms of aortic valve calcification, indicating that miR-22 and CAB39 are important mediators of osteogenic transition in VICs.

Although the detailed mechanism underlying the phenotypic transition of VICs remains poorly understood, differential gene expression profiles have been detected in calcified valves and VICs through high-throughput methods [23, 24]. Genetic polymorphisms and expression levels of specific genes, such as ENPP1 and oxidized phospholipids, have been identified to be associated with CAVD [25, 26]. Epigenetic regulatory networks have also been found to control pro-osteogenic activation in human aortic VICs [23]. In the present study, we found a close correlation between miR-22 expression and aortic valve calcification, and identified the role of miR-22 in the osteogenic differentiation of VICs. Other miRNAs, such as miR-125b and miR-204-5p, were also found to be involved in regulating the calcification process of VICs [27, 28]. Therefore, miRNA-mediated regulation networks might be important mechanisms for CAVD.

Previous reports have indicated that dysregulation of miR-22 is implicated in the differentiation of mesenchymal stem cells [29, 30]. Huang et al. found that overexpression of miR-22 promoted osteogenic differentiation of mesenchymal stem cells derived from human adipose tissue, but inhibited adipogenic differentiation, indicating that miR-22 could function as a regulator balancing adipogenic and osteogenic differentiation of mesenchymal stem cells [31]. However, Yin et al. also found that inhibition of miR-22 promoted osteogenic differentiation of bone marrow-derived mesenchymal stem cells and improved bone formation in mice [32]. These results suggested that the role of miR-22 in osteogenic differentiation might be related to the source from which mesenchymal stem cells were derived. In our study, overexpression of miR-22 accelerated the phenotypic transition from aortic VICs to valvular osteoblast-like cells.

In a previous study on ionizing radiation-induced bone marrow mesenchymal stromal cell injury, miR-22 accelerated cellular apoptosis due to mitochondrial ROS accumulation and autophagy attenuation [22]. Accumulation of ROS leads to imbalance of cellular homeostatic, resulting in oxidative stress and mitochondrial dysfunction. In this process, autophagy is also induced and reduce oxidative damages by engulfing and degrading oxidized substance. However, excessive high levels of ROS could also activate autophagic cell death and have destructive effects on cells [33]. Although the detailed role of ROS-mediated autophagy modulation on the pathogenesis of CAVD has not been unclear, recent reports have linked ROS generation and autophagy to the calcification of aortic valves. Evidence of increased ROS generation has been found in calcified valves, and ROS-induced DNA damage response is dysfunctional in early asymptomatic stages of CAVD [34]. The upregulated autophagic flux was found in the calcified tissues from CAVD patients [35], but it was also reported that VICs from normal aortic valves displayed significantly higher autophagic activity than those from calcified valves [36]. In the present study, we found that overexpression of miR-22 could accelerate ROS accumulation and reduce the autophagic activity in early stage of VIC calcification. Therefore, the pro-osteogenic action of miR-22 might also be partly mediated by ROS accumulation and autophagy inhibition.

Recent studies on mineralization mechanisms have identified a calcification-prone population of extracellular vehicles (EVs) derived from VICs as putative mediators of calcification in diseased valves [37, 38]. In a calcifying milieu, VICs highly expressed osteogenic genes (PiT-1, RUNX2, Msx2, and TNAP, etc.) [37], and pro-calcific VIC-derived EVs were enriched in annexins II, V, and VI [39]. Intercellular communication occurs when VECs take up these VIC-derived EVs [39]. Recently, the emerging role of miRNAs in intercellular communication led to potential perspectives on valve calcification mediated by EVs. Compared with cells, miRNAs are more concentrated in EVs, accounting for 50% of total RNA content [40]. Moreover, the findings on the vast differences in miRNA subsets present in EVs and originating cells have led to speculation that the loading mechanism of miRNAs into EVs is selective [41, 42]. The elevated miR-22 level was also detected in osteogenic VICs and their derived EVs (Fig. S10), suggesting that miR-22 may be an important pro-calcific genetic information transferred from calcifying VICs to VECs or other types of cells.

AMPK is a central guardian maintaining cellular energy homeostasis by orchestrating diverse processes, including lipogenesis, glycolysis, tricarboxylic acid cycle, and mitochondrial dynamics [43]. When the intracellular energy decreases, AMPK is phosphorylated and activated, and inhibits the activity of mTORC1 by directly phosphorylating TSC2 and RAPTOR (a key binding subunit of mTORC1), thereby inducing autophagy [44]. It has been reported that pharmacological activation of AMPK can significantly inhibit smooth muscle cell calcification [45]. Although the detailed function of AMPK pathway in VIC calcification was unknown, as the first activated pathway in cells upon energy stress, AMPK pathway might also play a vital role in CAVD progression, including triggering autophagy, attenuating endoplasmic reticulum stress, and inhibiting Runx2 signaling pathways. CAB39–LKB1–STRAD complex is an important upstream serine/threonine kinase of AMPK [46, 47], and its catalytic activity was closely related to miR-22 expression in VICs. Accumulation of miR-22 would reduce the catalytic activity of CAB39–LKB1–STRAD complex on AMPK, leading to the acceleration of VIC calcification. Therefore, intervention of miR-22/CAB39 might become a potential strategy for the prevention and treatment of CAVD.

The in-vitro model could not fully mimic the valve calcification process. Thus, it is essential to confirm the role of miR-22/CAB39 by an in vivo CAVD model in the future. The present study yielded new insights into the mechanism underlying CAVD pathogenesis, and gave the evidence that miR-22 and CAB39 might be developed as therapeutic targets for CAVD.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

GW, ZX, and LZ conceived and designed the study. FY, SL, and YG performed the experiments. FY and XD analyzed the data. FY, and YY collected and analyzed the clinical information. GW and FY wrote the manuscript.

Funding

This work was supported by National Natural Science Foundation of China (81800341, 81800342, and 82000364), National Key Research and Development Program of China (2016YFC1100900), and Shanghai Science and Technology Innovation Action Plan “Science and Technology Support Project in Biomedical Science” (21S11906000).

Availability of data and materials

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Conflict of interest

The authors declare that they have no conflict of interests.

Ethics approval

The study was approved by the Medical Ethics Committee in Changhai Hospital, and certify that the study was performed in accordance to the principles of the Declaration of Helsinki. Written informed consent was obtained from all the participants prior to enrollment.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent for publication

Patients signed informed consent regarding publishing their data and photographs.