Aged ASCs showed enhanced cellular senescence and impaired capacities for modulation of fibroblast and HUVEC functions
To evaluate functional alterations in ASCs during the aging process, we conducted a series of fundamental experiments. Compared with Y-ASCs, O-ASCs exhibited elevated levels of SA-β-gal activity (Fig. S1a). Additionally, O-ASCs showed more DNA damage, indicated by the increased enrichment of H2AX foci (Fig. S2a, S2b). Transwell migration assays showed that migration ability was impaired in O-ASCs, compared with Y-ASCs (Fig. S1b). The proportion of EdU-positive cells was greater among Y-ASCs than among O-ASCs (Fig. S1c), which indicated reduced proliferative capacity. Similarly, cell growth curves confirmed an inferior proliferative ability in O-ASCs, compared with Y-ASCs (Fig. S2c). Neovascularization and fibroblast migration are important processes that occur during wound healing. Thus, we examined the effects of CM from Y-ASCs and O-ASCs on fibroblast migration capacity and angiogenesis. Fewer closed tubular structures and less fibroblast migration were observed in the O-ASCs-CM condition than in the Y-ASCs-CM condition (Fig. S1d, S1e). These findings indicated that O-ASCs exhibit reduced cellular proliferation and migration, increased cellular senescence, as well as diminished effects on HUVEC and fibroblast functions.
SAN knockdown reversed cellular senescence and promoted ASC proliferation, migration, and overall function
To explore potential functional lncRNAs related to ASC senescence, ASCs from young and old volunteer donors were subjected to whole transcriptome resequencing. Differentially expressed lncRNAs with high fold change and low p-value are of considerable interest (Additional file 2). We selected the five most altered lncRNAs for further analysis. Among these lncRNAs, we found that the expression levels of NONHSAT035482.2 and NONHSAT150949.1 were significantly elevated and reduced, respectively, in qPCR analysis of O-ASCs (Fig. 1a). NONHSAT035482.2 was substantially more abundant than NONHSAT150949.1 in ASCs (Fig. 1b); therefore, we focused on NONHSAT035482.2 in subsequent experiments; we named this lncRNA “SAN.” The altered expression of SAN was confirmed on the basis of the increased sample volume (Fig. 1c). According to the NONCODE database (http://www.noncode.org/), SAN is located at chromosome 14; it has seven exons, a total length of 2,241 nucleotides, and a poly(A) tail. We performed 5'-RACE and 3'-RACE to identify the 5'-end and 3'-end cDNA sequences in SAN (Fig. S3).
To elucidate the biological role of SAN in ASC senescence, lentivirus vectors were constructed carrying SAN small hairpin RNA and the whole sequence of SAN. First, knockdown efficiencies of small hairpin RNA plasmids were evaluated in 293T cells and ASCs (Fig. S4a, S4b) and the most efficient plasmid was used in subsequent experiments. As shown in Fig. 1d, the proportion of ASCs stained with SA-β-gal was lower in the sh-SAN group than in the sh-NC group, while overexpression of SAN enhanced the level of SA-β-gal. Assessment of p21 protein expression also confirmed differences in cellular senescence among groups (Fig. 1i, 1j). The proliferation and migration capacities were significantly greater in sh-SAN ASCs than in sh-NC ASCs; in contrast, SAN overexpression led to reduced proliferation and migration (Fig. 1e, 1f). To investigate the role of ASCs in regulating HUVEC and fibroblast functions, which is a major component of ASC-based therapy for wound healing, we assessed the effects of CM from those four groups on angiogenesis and migration capacity. Tube formation analysis showed that HUVECs treated with CM from sh-SAN ASCs formed more tubular structures than did HUVECs treated with CM from sh-NC ASCs. Treatment with CM from oe-SAN ASCs led to reduced endothelial network formation capacity, compared with CM from oe-NC ASCs (Fig. 1g). Furthermore, wound scratch assays showed that fibroblast migration capacity was enhanced by treatment with CM from sh-SAN ASCs, compared with CM from sh-NC ASCs; treatment with CM from oe- SAN ASCs led to diminished fibroblast migration capacity, compared with CM from oe-NC ASCs (Fig. 1h). qPCR demonstrated that mRNA expression levels of vascular endothelial growth factor A and fibroblast growth factor were significantly greater in sh-SAN ASCs than in sh-NC ASCs (Fig. S4d). Collectively, these data show that SAN knockdown in ASCs could ameliorate senescence-associated phenotypes and modulate cellular functions in ASCs.
miR-143-3p mediates cellular senescence of ASCs as a competing endogenous RNA of SAN
To further investigate the mechanism by which SAN contributes to ASC senescence, we measured the relative expression of SAN in the nucleus and cytoplasm. The results showed that SAN transcripts were primarily located in the cytoplasm (Fig. 2a). Cytoplasmic lncRNA has been shown to function as competing endogenous RNA [7]. Therefore, we examined SAN target miRNAs by means of bioinformatics prediction and sequencing data analysis. SAN was predicted to bind with seed sequences of several miRNAs, based on differential expression findings (Fig. 2b). Among these miRNAs, miR-143-3p exhibited the highest expression level. qPCR revealed that miR-143-3p was upregulated in Y-ASCs, compared with O-ASCs (Fig. 2c). Furthermore, Pearson correlation analysis showed a negative correlation between expression levels of SAN and miR-143-3p in ASCs (Fig. 2d). Accordingly, we proposed that SAN serves as a sponge for the regulation of miR-143-3p. To assess the direct binding of SAN and miR-143-3p, a dual-luciferase reporter gene assay was conducted, which showed that luminescence activity was significantly inhibited in the SAN-wild-type + miR-143-3p mimic group, compared with the SAN-wild-type + mimic NC group (Fig. 2f, 2g). Additionally, the lncRNA level was reduced after miR-143-3p overexpression, compared with the mimic NC group (Fig. 2e). Besides, a SAN mutant lacking the miR-143-3p binding site could not affect the proliferation, migration and senescence of ASCs (Fig. S5a–d). Overall, these findings demonstrated that SAN could target miR-143-3p and affect its function.
Our findings indicated that miR-143-3p was significantly downregulated in O-ASCs. Therefore, we investigated whether miR-143-3p mediates cellular senescence in ASCs; we examined ASC proliferation, migration, senescence, and cellular function after miR-143-3p overexpression or inhibition. miR143-3p mimic, mimic-NC, miR143-3p inhibitor, or inhibitor NC was separately transfected into ASCs; transfection efficiency was evaluated by qPCR (Fig. S6). Compared with the mimic-NC group, ASCs treated with miR-143-3p exhibited enhanced proliferation and migration capacities, whereas ASC proliferation and migration were reduced in the miR-143-3p inhibitor group, compared with the inhibitor-NC group (Fig. 3b, 3c). In addition, overexpression of miR-143-3p reduced the number of SA-β-gal-positive cells, while inhibition of miR-143-3p led to larger numbers of SA-β-gal-positive cells, compared with the inhibitor-NC group (Fig. 3a). Western blotting analysis showed that treatment with miR-143-3p mimic increased the cell-cycle associated protein cyclin A1 (CCNA1) and the migration-associated protein fibronectin 1 (FN1), while reducing the expression of the senescence-associated protein p21. In contrast, treatment of ASCs with miR-143-3p inhibitor led to reduced expression levels of FN1 and CCNA1, as well as enhanced p21 expression (Fig. 3f, 3g). Furthermore, wound healing and tube formation assays showed that CM from the miR-143-3p-overexpressing group was more effective in modulating fibroblast activity and angiogenesis than was CM from the mimic-NC group; these effects were reduced upon treatment with CM from the miR-143-3p inhibition group, compared with the inhibitor-NC group (Fig. 3d, 3e). Taken together, the findings indicate that miR-143-3p acts as a cellular senescence suppressor in ASCs.
miR-143-3p overexpression attenuated the cellular senescence effect of SAN in ASCs
To detect whether the senescence-promoting effects of SAN were mediated by miR-143-3p, we treated oe-SAN ASCs with miR-143-3p mimics or mimic-NC. As shown in Fig. 4, treatment with miR-143-3p mimics rescued the inhibitory effects of oe-SAN on ASC proliferation and migration and rejuvenated the enhanced senescence phenotype in the oe-SAN group (Fig. 4a–4c). Moreover, CM from oe-SAN ASCs that had been transfected with miR-143-3p exhibited restored fibroblast-modulating and angiogenic effects, compared with CM from oe-SAN ASCs that had been treated with mimic-NC (Fig. 4d, 4e). Western blotting analysis also indicated the restoration of p21 expression (Fig. 4f, 4g). Collectively, the above results showed that overexpression of miR-143-3p reversed the effects of SAN overexpression in ASCs.
ADD3, a target gene of miR-143-3p, served as a senescence-associated gene in ASCs
To elucidate the underlying mechanism of miR-143-3p, bioinformatics databases including TargetScan, miRanda, and RNAhybrid were used to identify potential targets of miR-143-3p. Differentially expressed mRNAs were screened through database prediction and depicted in heatmaps (Fig. 5a). qPCR analysis showed that ADD3 was inhibited after the overexpression of miR-143-3p, whereas it was upregulated in O-ASCs, compared with Y-ASCs (Fig. 5b, 5c). Hence, ADD3 was selected for further validation. Correlation analyses revealed that the expression level of ADD3 was significantly negatively correlated with the expression level of miR-143-3p; it was significantly positively correlated with the expression level of SAN (Fig. 5d, 5e). To test whether miR-143-3p regulates ADD3 protein expression in ASCs, we transfected miR-143-3p mimic or miR-143-3p inhibitor into ASCs; we used mimic-NC or inhibitor-NC as control treatments. Western blotting analysis showed significantly reduced ADD3 protein expression in miR-143-3p mimic-treated ASCs, compared with the mimic-NC group. Similarly, the inhibition of miR-143-3p in ASCs led to enhanced expression of ADD3 (Fig. 5f). Furthermore, SAN overexpression increased the protein expression level of ADD3, while SAN depletion led to reduced ADD3 expression (Fig. S7a, S7b). Luciferase reporter assays were conducted to identify the putative seed-matched sequence in the ADD3 3’-UTR region. Co-transfection of the ADD3-wild-type (WT) plasmid and miR-143-3p mimic into HEK-293T cells led to repression of luciferase activity compared with that of the mimic-NC group, whereas no significant change was observed in the ADD3-mutant-type (MUT) group (Fig. 5j, 5k). These results suggested that ADD3 serves as a target gene for miR-143-3p.
As a downstream gene of miR-143-3p, we presumed that ADD3 might be associated with ASC senescence. Notably, western blotting results indicated that protein expression levels of ADD3 and p21 were elevated in O-ASCs (Fig. 5h, 5i). Similar results were observed in replicative senescent ASCs (Fig. 5g). Following respective transduction of ASCs with either sh-ADD3 or oe-ADD3 lentiviral vector and the corresponding empty vector, obvious reduction or enhancement of ADD3 protein expression level was observed (Fig. S8). Cell proliferation and migration were slightly improved in ADD3 knockdown ASCs (Fig. 6b, 6c); SA-β-gal staining assays also indicated reduced cellular senescence in the sh-ADD3 group (Fig. 6a). Conversely, significant reduction of cellular function and enhancement of senescence were detected in ADD3-overexpressing ASCs, compared with NC-overexpressing ASCs. Consistent with the cell phenotype findings, functional proteins (e.g., FN1, CCNA1, and p21) were influenced by changes in ADD3 protein expression (Fig. 6f, 6g). In addition, CM from ADD3-overexpressing ASCs or ADD3-deficient ASCs was compared with CM from control ASCs in terms of the abilities to induce angiogenesis and fibroblast migration. Importantly, CM from the ADD3-overexpressing group showed inferior abilities to induce angiogenesis and fibroblast migration, compared with the control group, whereas CM from ADD3-deficient ASCs showed superior abilities to induce these processes (Fig. 6d, 6e).
ADD3 downregulation contributed to the anti-senescence effects of miR-143-3p in ASCs
To examine whether ADD3 could reverse the rejuvenation effects of miR-143-3p in ASCs, stem cells containing oe-ADD3 vector or empty vector (oe-NC) were transfected with miR-143-3p mimic and corresponding mimic-NC. They were then subjected to analyses of cellular senescence, cell proliferation, and migration, as well as CM angiogenesis-promoting and fibroblast migration-promoting effects. As shown in Fig. 7b, 7c, compared with the oe-NC + mimic-NC group, the migration and proliferation of ASCs were enhanced in the oe-NC + miR-143-3p mimic group, but reduced in the oe-ADD3 + mimic-NC group. Moreover, compared with the oe-ADD3 + mimic NC group, miR-143-3p mimic treatment rescued the inhibitory effects on ASC proliferation and migration in the oe-ADD3 + miR-143-3p mimic group. Additionally, cellular senescence was significantly enhanced in the oe-NC + miR-143-3p mimic group and reduced in the oe-ADD3 + mimic-NC group (Fig. 7a). Functional proteins (e.g., FN1, CCNA1, and p21) in those groups showed similar tendencies (Fig. 7f, 7g). Furthermore, miR-143-3p mimic treatment in ASCs reversed the inhibitory effects of ADD3 overexpression on modulatory functions in HUVECs and fibroblasts (Fig. 7d, 7e). The findings support the hypothesis that miR-143-3p ameliorates senescence in ASCs by downregulating ADD3.
Transplantation of SAN-depleted aged ASCs accelerates cutaneous wound healing in rats
To evaluate the rejuvenation and therapeutic effects of SAN-deficient ASCs on wound healing in vivo, full-thickness cutaneous wounds were created on the dorsal skin of rats. The edge of the wound area was treated with PBS, control vector-transduced Y-ASCs, control vector-transduced O-ASCs, or SAN-depleted old ASCs (sh-SAN-O-ASCs); wound healing status was then assessed. As shown in Fig. 8a–8c, in contrast with the O-ASC group, the original wound area was significantly smaller in the sh-SAN-O-ASC group at days 7, 10, and 14 post-wounding. Furthermore, cutaneous wounds treated with sh-SAN-O-ASCs were completely healed at approximately 14 days post-wounding, similar to the results in the Y-ASC group.
Wound histological analysis was conducted to assess repair efficacy in each group at day 14 post-wounding. Hematoxylin and eosin staining showed that the sh-SAN-O-ASC group achieved greater re-epithelialization than did the O-ASC group (Fig. 8d, 8e); the mean proportions in the Y-ASC and PBS groups were 97% and 51%, respectively. Analysis of Masson’s trichrome-stained tissue showed that collagen deposition was greater in the sh-SAN-O-ASC group than in the O-ASC group; it was similar to the deposition in the Y-ASC group (Fig. 8f, 8g). Additionally, analysis of frozen skin sections showed that the survival rate was 2.5-fold greater in the sh-SAN-O-ASC group than in the O-ASC group (Fig. S9).
Immunohistochemical and immunofluorescence analyses of wound tissue sections were performed to investigate the abilities of ASCs in each group to promote angiogenesis and cellular proliferation. Compared with the O-ASC group, the sh-SAN-O-ASC group exhibited significantly enhanced cellular proliferation ability, as determined by the proportion of PCNA-positive cells (Fig. 8h, 8i). Moreover, α-smooth muscle actin staining revealed approximately twofold more mature blood vessels in the sh-SAN-O-ASC group than in the O-ASC group (Fig. 8j, 8k). Collectively, these results indicated that SAN knockdown in aged ASCs was able to restore cellular therapeutic capabilities such as neovascularization and cellular proliferation in a cutaneous wound model.