MiR-483 induces senescence of human adipose-derived mesenchymal stem cells through IGF1 inhibition

Characterization and verification of senescent MSCs

The senescence phenotypes of hADSCs grown in vitro were assessed by isolating hADSCs from two different donors and subjecting to long-term proliferation under the same culture conditions. After passage 9 (P9), hADSCs gradually increased in size, and acquired a flat and irregular shape, which is the typical morphology of senescent cells (Supplementary Figure 1A, 1B). Flow cytometry analysis showed that hADSCs from early (P9) and late (P16) passages were both positive for the MSCs surface markers CD73, CD90, and CD105, but negative for CD34, CD45, CD11b, CD19 and HLA-DR (Supplementary Figure 1C, 1D).

Another key criterion for the identification of MSCs is their potential to differentiate into osteoblasts and adipocytes. Here, we analyzed the effect of replicative senescence on the differentiation potential of hADSCs. Our findings revealed that hADSCs undergoing long-term culture had an increased adipocyte differentiation ability but a reduced osteocyte differentiation ability (Figure 1A, 1B).

Figure 1. Characterization and verification of senescent MSCs from human adipose tissues from two different donors (#1 and #2). (A) Oil Red O staining of hADSCs at P9 and P16 (10×; scale: 50 μm); MRNA expression of PPARγ was evaluated by RT-qPCR. (B) Alkaline phosphatase staining of hADSCs at P9 and P16 (10×; scale: 50 μm); MRNA expression of RUNX2 was evaluated by RT-qPCR. (C) MRNA expression levels of p16, p21, and p53 were evaluated by RT-qPCR; western blot analysis of p53. (D) SA-β-gal staining of hADSCs at P9 and P16 (10×; scale: 50 μm). (E) The proliferation of hADSCs at P9 and P16 was detected by CCK-8 assay. (F) Telomere length was demonstrated by qPCR.

Furthermore, the expression of three cell senescence-related molecular markers (p16, p21, and p53) were significantly increased from P9 to P16 (Figure 1C). Additionally, the number of cells that positively stained with SA-β-gal was significantly increased in P16 (Figure 1D). Cell counting kit-8 (CCK-8) assay results showed that the proliferation of hADSCs was significantly downregulated in P16 (Figure 1E). Changes in telomere length were evaluated using quantitative PCR (qPCR) [16], which revealed that telomere shortening occurred during the aging of hADSCs in vitro (Figure 1F). These data generally indicated that hADSCs exhibit a constant increase in senescence with successive passaging in culture.

MiRNAs expression in the adipogenic differentiation of hADSCs

Previous studies have revealed that osteoblasts and adipocytes share a common mesenchymal ancestor and that the distinct processes of osteogenic and adipogenic differentiation from MSCs are competitive [17]. We found that adipogenic differentiation of hADSCs increased during cellular senescence which was consistent with previous findings [18]. Comparative sequencing of small RNA using adipogenic differentiation samples from day 0 (undifferentiated) and day 14 were performed to identify any putative miRNAs that simultaneously affect adipogenic differentiation and replicative senescence in hADSCs (Figure 2A, 2B).

Figure 2. MiR-483 expression during the adipogenic differentiation of hADSCs. (A, B) Small-RNA sequencing was performed to profile miRNAs with variant expression during adipogenic differentiation of hADSCs. Data are expressed as the number of reads and (C) differentially expressed miRNA genes. (D, E) The relative expression of miR-483-5p and miR-483-3p was validated by RT-qPCR. (F) The location and classification of miR-483.

Apart from miR-1908, which has been reported to be highly expressed during adipogenic differentiation of hADSCs [19], we found that miR-483-3p and miR-483-5p expression were also significantly increased during adipogenesis (Figure 2C). To further verify these results, miR-483-5p and miR-483-3p expression were analyzed using RT-qPCR (Figure 2D, 2E). As pre-miR-483 is located in intron 2 of the IGF2 transcript and encodes two mature miRNAs (miR-483-3p/miR-483-5p) (Figure 2F), and IGF2 modulates the differentiation of stem cells [20, 21], we also measured the IGF2 expression using RT-qPCR. Consistently, IGF2 and miR-483 displayed similar expression pattern in adipogenic differentiated hADSCs that were isolated from two different donors and cultured under the same conditions (Supplementary Figure 2). Therefore, we speculated that miR-483 plays a vital role in the differentiation of hADSCs.

MiR-483 promotes adipogenesis of hADSCs in vitro

To further investigate the role of miR-483 in the differentiation of hADSCs, we transfected hADSCs with 483-3p-I (an inhibitor of miR-483-3p), 483-5p-I (an inhibitor of miR-483-5p), or negative control inhibitor (I-CT). The transfection efficiency was quantified at 48 hours after transfection by RT-qPCR. The success of the inhibition was demonstrated by low miR-483-5p and miR-483-3p expression (Figure 3A, 3B). We induced adipogenic differentiation of hADSCs transfected with 483-5p-I, 483-3p-I, or I-CT, then measured the adipogenic induction efficacy. It was found that inhibition of miR-483-5p and miR-483-3p in hADSCs decreased the number of lipid droplets (Figure 3C). To confirm these results, we investigated the expression of two primary adipogenic transcription factors, lipoprotein lipase (LPL) and peroxisome proliferator-activated receptor-gamma (PPARγ), by RT-qPCR (Figure 3D, 3E) and western blot analysis (Figure 3F). Compared with the control group, the mRNA and protein levels of LPL and PPARγ were significantly decreased in the 483-3p-I group, but not in the 483-5p-I group. Because miR-483-3p had the stronger effect of the two miRNAs, we selected it for further characterization.

Figure 3. Effects of miR-483 on adipogenic differentiation of hADSCs. (A, B) The efficiency of miR-483-5p and miR-483-3p inhibition was assayed by RT-qPCR. (C) The effect of inhibition of miR-483-3p and miR-483-5p on adipogenic differentiation of hADSCs was evaluated on day 14 using Oil Red O staining (10×; scale: 50μm). (D–F) RT-qPCR and western blot analysis of LPL and PPARγ expression levels. I-CT: negative control inhibitor; 483-5p-I: inhibitor of miR-483-5p; 483-3p-I: inhibitor of miR-483-3p.

MiR-483-3p directly targets the 3′UTR of IGF1 mRNA

The biological process underlying the regulation of adipogenesis by miR-483-3p was further illustrated by the bioinformatics analysis of potential miR-483-3p target genes using the TargetScanHuman 7.0 and miRWalk online databases. Through database prediction, we selected IGF1, O-linked N-acetylglucosamine transferase (OGT), and chloride voltage-gated channel (CLCN3) as potential targets of miR-483-3p. We cloned the 3′ UTRs of each of these three genes separately into a dual-luciferase-based reporter plasmid and cloned miR-483-3p into pSUPER. The miR-483-3p plasmid or control plasmid was then co-transfected with one of the three reporter plasmids into HEK-293T cells. While the presence of miR-483-3p significantly decreased the luciferase reporter activity of IGF1, the decreases in the activities of OGT and CLCN3 were not significant (Supplementary Figure 3). We also created a mutation of eight nucleotides within the mRNA binding site of miR-483-3p to interrupt the hypothetical connection between the IGF1 mRNA and miR-483-3p (Figure 4A, 4B). As shown in Figure 4B, the mutant miR-483-3p restored the luciferase activity of IGF1.

Figure 4. Identification of IGF1 as a direct target of miR-483-3p. (A) Putative miR-483-3p target sites in IGF1 and design of mutation sites. (B) Activity of the luciferase reporter gene in HEK-293T cells co-transfected with a luciferase reporter plasmid containing IGF1 3′ UTRs and wild-type pSUPER-miR-483-3p, mutant pSUPER-miR-483-3p, or empty pSUPER vector. Luciferase activity was assayed 32 hours after transfection. (C, D) RT-qPCR analysis of IGF1 mRNA levels in hADSCs transfected with 483-3p-M or 483-3p-I. mut: mutant; M-CT: mimic control; 483-3p-M: mimic of miR-483-3p.

HADSCs were also transfected with 483-3p-I or 483-3p-M (mimic of miR-483-3p), and RT-qPCR was used to quantify the mRNA levels of IGF1. IGF1 expression was downregulated by miR-483-3p overexpression (Figure 4C) and upregulated by miR-483-3p inhibition (Figure 4D). These observations generally suggested that IGF1 is a direct target gene of miR-483-3p.

MiR-483-3p induces senescence of hADSCs through IGF1

Given that miR-483-3p was upregulated during adipogenic differentiation and that it controlled the expression of adipogenic differentiation markers and phenotypes, we sought to determine whether it was involved in hADSC senescence in vitro. The expression levels of miR-483-3p during early and late passages of hADSCs taken from two different donor samples were examined using RT-qPCR. MiR-483-3p expression was significantly increased in senescent cells (Figure 5A). To further investigate the biological effect of miR-483-3p on hADSC senescence phenotypes, hADSCs were transfected with 483-3p-I or 483-3p-M to alter the in vitro expression levels of miR-483-3p, then stained for SA-β-gal. Transfection with 483-3p-M increased the activity of SA-β-gal while 483-3p-I decreased it compared with the corresponding control group (Figure 5B, 5C). Interestingly, we also found that IGF1 expression was significantly reduced in senescent cells (Figure 5D).

Figure 5. Effects of miR-483-3p and IGF1 on senescence of hADSCs. (A) MiR-483-3p expression at early-passage (P7) and late-passage (P14) hADSCs was detected by RT-qPCR. (B) SA-β-gal staining of hADSCs (P14) transfected with 483-3p-M and corresponding control (10×; scale: 100μm). (C) SA-β-gal staining of hADSCs (P14) transfected with 483-3p-I and corresponding control (10×; scale: 100μm). (D) IGF1 expression in hADSCs at P7 and P14 was detected by RT-qPCR. (E, F) Efficiency of IGF1 inhibition was assayed by RT-qPCR and ELISA. (G) SA-β-gal staining of hADSCs (P14) transfected with si-CT or si-IGF1 (4×; scale: 100μm). (H) SA-β-gal staining of hADSCs (P14) co-transfected with 483-3p-I and si-IGF1 (10×; scale: 100μm).

Based on these results, we transfected hADSCs with the small interfering RNA (siRNA) si-IGF1 to inhibit endogenous IGF1 expression. The results of RT-qPCR and ELISA showed that si-IGF1 effectively reduced IGF1 expression (Figure 5E, 5F). Additionally, the SA-β-gal activity of the si-IGF1 transfection group was significantly higher than that of the si-NC group (Figure 5G).

To further verify the relationship between miR-483-3p and IGF1 in the in vitro aging of hADSCs, we co-transfected hADSCs with 483-3p-I and si-IGF1. The knockdown of miR-483-3p reduced SA-β-gal activity, but the additional transfection of si-IGFI restored this activity in hADSCs compared with the control group (Figure 5H).

Isolation and characterization of hADSCs

Samples of subcutaneous adipose tissues were obtained from patients without malignant tumors, autoimmune diseases, congenital diseases, or genetic diseases, at the age of 30 (#1) and 34 (#2) undergoing fracture surgery. The operation protocols were approved by the Ethics Research Committee, Tongji University School of Medicine.

Tissue was excised from blood vessels and sliced into approximately 1 mm³ size, digested in a 0.2% collagenase type I/II mixture (Gibco, New York, USA) at 37°C for 1 hour, filtered with 70-μm nylon mesh and then subjected to centrifugation at 1,500 rpm for 10 minutes. The cells were cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum, 1% non-essential amino acids, 100 U/mL penicillin and 100 μg/mL streptomycin, in a 5% CO2 environment at 37°C. After one week of culture when the cells reached about 80% confluence, they were passaged and transferred one-third of cells to a new culture dish (passage 1, P1). At P2, surface antigen sorting was performed on adherent and spindle-shaped hADSCs according to the following criteria: positive for CD73, CD90, and CD105 and negative for CD34, CD45, CD11b, CD19 and HLA-DR (BD, Human MSC Analysis Kit). Adipogenic and osteogenic differentiation kits (Gibco, New York, USA) were used to induce adipogenesis and osteogenesis, respectively. The adipocyte phenotype was evaluated using Oil Red O staining according to the previously described protocols [23]. Osteoblast phenotype was assessed using BCIP/NBT Alkaline Phosphatase staining assay kit according to the manufacturer’s protocols (Beyotime, Shanghai, China).

Cell proliferation CCK-8 assay

Cell proliferation ability was evaluated using CCK-8 assay (Dojindo, Shanghai, China) in early-passage and late-passage hADSCs according to the manufacturer’s instructions. Briefly, cells were cultured in 96-well plates at a concentration of 1 × 10⁴ cells/well. After 24 hours of culture, 10 μL CCK-8 reagent was added to each well and further incubated for 4 hours. The absorbance was determined at a wavelength of 450nm using SpectraMax M5 (Molecular Devices, CA, USA).

MiRNAs sequencing

The miRNAs samples were extracted using the mirVana™ miRNA Isolation Kit (Ambion, Waltham, USA) following the manufacturer's instructions. Small RNA libraries were prepared using the Small RNA Sample Preparation kit (Illumina, SanDiego, USA) according to the manufacturer’s protocols. The quality of the purified small RNA sequencing libraries was confirmed on an Agilent 2100 Bioanalyzer. The small RNA sequencing data were first cleaned using small RNA sequencing data cleaning pipeline. Next, the clean sequence data were aligned to miRBase to detect and estimate the expression of microRNAs. Subsequently, the remaining read sequences were aligned to the fRNAdb to get the expression of various non-coding RNAs. The miRNAs sequencing data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO) under accession number GSE138879.

Cell transfection

Mimics of miR-483-3p (483-3p-M) and miR-483-5p (483-5p-M), inhibitors of miR-483-3p (483-3p-I), and their corresponding negative controls (mimic-NC (M-CT) and inhibitor-NC (I-CT)), and the siRNA of IGF1 (si-IGF1) and its corresponding negative control (si-NC) were synthesized by Genepharma (Shanghai, China). hADSCs were grown until 50%–60% confluence then transfected with 100 nM miRNA mimic, inhibitor or NC using Lipofectamine 2000 reagent (Invitrogen, CA, USA). The culture medium was changed after 12 hours of transfection and cellular differentiation was induced 48 hours later according to the manufacturer’s recommendations.

Dual-luciferase reporter assay

A partial sequence of IGF1 containing the predicted miR-483-3p binding site was amplified by PCR and cloned into the dual-luciferase plasmid. The miR-483-3p gene and a mutant ‘seed’ sequence were synthesized by Thermo Fisher Scientific and cloned into the pSUPER vector. Reporter assays were performed in HEK-293T cells, which were harvested 32 hours after transfection. Firefly and Renilla luciferase activities were measured using the Dual-Glo Luciferase assay system (Promega, Madison, USA). A signal from the green Renilla luciferase intracellular protein was normalized to the red firefly luciferase signal.

RT-qPCR analysis

Total RNA was extracted using Trizol reagent (Invitrogen, California, USA). The quality and concentration of total RNA were determined using a nanodrop 2000 and separation on a 1% agarose gel. Reverse transcription and detection of miRNAs were accomplished using the all-in-one miRNA qRT-PCR detection kit (Genecopoeia, Maryland, USA) according to the manufacturer’s instructions. MRNA expression was quantified via an initial synthesis of cDNA using a PrimeScript RT reagent kit (Takara, Dalian, China), followed by RT-qPCR using SYBR Green SuperMix (Bio-Rad, California, USA) on the Quantstudio 7 Flex Real-Time PCR system (Thermo Scientific, MA, USA). Each sample was tested in triplicate. The relative expression levels of miRNAs and mRNA were normalized to the expression of U6 and β-actin, respectively. The relative gene expression was calculated using the comparative CT (2^-ΔΔCT) method, and the sequences of the primers used were listed in Table 1.

Western blot analysis

Cells were washed twice with cold phosphate-buffered saline, then lysed with RIPA lysis buffer containing protease inhibitor and shaken for 15 minutes on ice. Protein samples were subjected to centrifugation at 12,000 ×g for 15 minutes, and the supernatants were collected. Protein concentration was determined using bicinchoninic acid protein assay kit (Beyotime). Samples containing equal amounts of protein were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred onto a 0.22-μm polyvinylidene difluoride membrane. After blocking with 5% non-fat milk (Sigma, St. Louis, USA), the membranes were incubated with a primary antibody (PPARγ: ab191407, Abcam, Cambridge, UK; LPL: ab172953, Abcam; α-tubulin: T9026, Sigma) at 4°C overnight, then further incubated with the corresponding secondary antibody (170-6515 and 170-6516, Bio-Rad) according to the manufacturer’s instructions. Protein bands were quantified using the Amersham Imager system.

Assay of cellular senescence (SA-β-gal staining)

Cells were cultured in 24-well plates at a concentration of 2 × 10⁴ cells/well. SA-β-gal activity was determined using the Cell Senescence β-Galactosidase Staining Kit (Beyotime) according to the manufacturer’s instructions. The nuclei were stained with DAPI, and blue-stained senescent cells were counted using Image J.

ELISA

The protein content of IGF1 in cells was determined using the Human IGF1 enzyme-linked immunosorbent assay (ELISA) kit (Elabscience, Wuhan, China) according to the manufacturer’s instructions.

Statistical analysis

Results are express as the means ± standard deviation (n=3). Student's t-test was used to determine significance between two groups. A P-value <0.05 was considered statistically significant.