MSCs from elderly are senescent and functional defective
To understand the properties of MSCs from elderly (OMSCs), we first compared senescent phenotypes of OMSCs (>65 years old) with MSCs from umbilical cord (UMSCs). As expected, more senescence-associated β-gal positive cells was observed in OMSCs than in UMSCs (Fig 1a-b). In line with these, the senescence-related markers p53, p21 and p16 were markedly upregulated and longevity marker Sirt1 was significantly decreased in OMSCs (Fig 1c). Consistent with these results, the proliferation of OMSCs was significantly reduced as compared to that of UMSCs (Fig 1d). Altogether, these data showed that OMSCs have significant senescent phenotypes, leading to a functional defection.
Exosomes derived from UMSCs ameliorate senescent phenotype of OMSCs
Exosomes from UMSCs (ExoUMSCs) were purified from the conditioned medium of UMSCs. The morphology of ExoUMSCs were typical cup-shaped with double layer membrane structure as visualized by transmission electron microscopy (TEM) (Fig 2a). Western blot analysis confirmed that the isolated particles expressed exosome-specific markers: CD63, CD9 and Alixs (Fig 2b). The particle size of ExoUMSCs was around 60-150nm as shown by Dynamic light scattering (DLS) (Fig S2). After OMSCs were incubated with Dil-labeled ExoUMSCs for 48 h, fluorescence was detected inside the cells (Fig 2c), indicating that ExoUMSCs were efficiently internalized into the target cells. Number of β-gal positive cells were significantly reduced in OMSCs after treated with ExoUMSCs (Fig 2d). Expression of aging-related factors: p53, p21, and p16 was also markedly reduced, and the level of Sirt1 was increased in the ExoUMSCs-treated OMSCs (Fig 2e). The growth rate of OMSCs was significantly increased (Fig 2f), and more OMSCs entered into S phase of cell cycle after treatment with ExoUMSCs (Fig 2g). The numbers of EdU positive OMSCs were significantly higher in the ExoUMSCs-treated OMSCs than the untreated cells (Fig 2h). These results indicate that the exosomes secreted by UMSCs can rejuvenate the senescent MSCs.
ExoUMSCs renewed biological activities of OMSCs
To examine if ExoUMSCs could improve biological functions of OMSCs, OMSCs were pretreated with ExoUMSCs for 48 h (OMSCsExo). The migration capacity was remarkably increased in OMSCsExo measured by either Transwell assay (Fig 3a) or scratch wound assay (Fig S3a-d). The lower levels of Bcl-2/bax confirmed the decreased apoptosis of ExoUMSCs-treated OMSCs (Figure 3b). Furthermore, reduced number of apoptotic cells were observed in the ExoUMSCs treated OMSCs (Figure 3c) as compared with the untreated cells, and the rate of apoptotic cells was almost equivalent to that of UMSCs under the stressed condition. The differentiation potential of OMSCs into osteocytes, chondrocytes and adipocytes were also increased after treated with ExoUMSCs (Figure S4).
Since the therapeutic effects of MSCs are primarily contributed by their paracrine functions, the effect of ExoUMSCs on MSCs paracrine functions was also studied. The conditioned media were collected from OMSCs, OMSCsExo and UMSCs to treat human umbilical vein endothelial cells (HUVECs) and primary mouse cardiomyocytes for assessing the pro-angiogenesis and anti-apoptosis effect of MSCs. Significantly increased tube formation was observed in HUVECs treated with the medium of OMSCsExo as compared with those treated with DMEM or medium of OMSCs (Fig 3d). Similarly, the apoptosis rate of cardiomyocytes (CMs) under hypoxia and serum deprivation condition was attenuated after CMs were treated with medium of OMSCsExo as compared with DMEM or OMSCs groups, and reached a similar level as that of treated with medium of UMSCs (Fig 3e). These results suggested that exosomes secreted by UMSCs can ameliorate the biological function of OMSCs and enhance their paracrine effects.
OMSCs pretreated with ExoUMSCs have better cardioprotective effect in vivo
To investigate the effect of ExoUMSCs on therapeutic functions of OMSCs for cardiac regeneration after myocardium infarction (MI), OMSCs or OMSCsExo treated with or without ExoUMSCs were injected into myocardium immediately after MI. We measured cell survival after MSCs transplantation in heart tissue from different groups at day 7 after myocardial infarction, and found that the survival of UMSC and OMSCExo were significantly higher than those of the OMSC and DMEM groups(P<0.05). There was an increased trend in survival rates of UMSC compared with OMSCExo (Fig S5). We measured and analyzed the immune cells including CD3+ T cells, CD19+ B cells, Ly6G+ neutrophils, F4/80+ macrophages, and inflammatory factors such as IL-1b, Il-6, IL-12, TNFa and MCP-1 in heart tissue from different groups at day 7 after myocardial infarction. No differences in the amount of immune cells and inflammatory factors are detected between OMSC and UMSC groups (Fig S6).
Significant improvement in cardiac functions, both ejection fraction (EF%) and fractional shortening (FS%), was observed in OMSCsExo group as compared with OMSCs group (EF: 56±5% vs 45±12%, p<0.05; FS: 29±3% vs 22±7%, p<0.05) (Fig 4a), while OMSCs treatment also resulted in certain improvement in cardiac function as compared with the control mice treated with DMEM (EF: 45±12% vs 34±7%, p<0.05; FS: 22±7% vs 16±4%, p<0.05). The cardioprotective effect of OMSCsExo reached close to that of UMSCs (Fig 4b). The density of CD31+ capillary and vwF+ cells were significantly increased in mice injected with OMSCsExo (Fig 4c-d and Fig S7) as compared with those injected with OMSCs or DMEM 28 days after infarction. Similar trends were observed in arteriole density as measured by α-SMA staining (Fig 4e-f). The scar size at 28 days after MI was significantly smaller in OMSCsExo group than that of OMSCs and DMEM groups (Fig 4g-h). These results indicate an improvement in therapeutic functions of OMSCs after treated with ExoUMSCs.
miRNA-136 was a key effecter of ExoUMSCs to rejuvenate OMSCs
To study the mechanism of ExoUMSCs–mediated rejuvenation of OMSCs, the profiles of age-related miRNAs were screened from database (Miranda and TargetScan) and literatures [19, 20] as most of the effects of exosome were carried out through exosomal miRNAs. Among these miRNAs, miR-136 was the most abundant in UMSCs as compared with OMSCs, and was increased more than 20-fold in OMSCs after cultured with ExoUMSCs, whereas no significant difference was observed for other miRNAs such as miR-106a, miR-155 and miR-29C among these groups (Fig 5a, Fig S8). Moreover, the level of miRNA-136 in ExoUMSCs was also higher than that in ExoOMSCs (Fig 5b). Next, we examined the role of miR-136 in regulating OMSCs senescence and biological function. The mRNA and protein levels of senescent-associated markers p53, p21 and p16 (Fig 5c-d), as well as the β-gal+ cells were markedly reduced in OMSCs after transfected with miR-136 mimics (Fig 5e). As a hallmark of cellular senescence and DNA damage, γH2AX positive foci were enriched in OMSCs as compared with UMSCs, and were remarkably decreased in miR-136 transfected OMSCs (Fig 5f).
Furthermore, using an EdU incorporation assay to measure DNA synthesis, we found that the proliferation rate of OMSCs was also significantly increased after OMSCs were transfected with miR-136 mimic (Fig 5g and Fig S9a). More cells at S phase of cell cycle (Fig 5h and Fig S9b) and less apoptotic cells (Fig 5i and Fig S10a-c) were observed in the miR-136 overexpressed OMSCs in comparison with the original OMSCs. ExoUMSCs and miR-136 mimics also significantly increased the cell survival rate (Fig 5j) under hypoxia and serum deprivation condition (Fig S11) mimicking the microenvironment of MI in vivo, and resulted in a similar viability as that of UMSCs.
Apaf1 is a functional direct target of miR-136 involving in OMSCs survival
To understand how miR-136 regulates functions of OMSCs, the downstream target of miR-136 in regulating cell senescence and survival was searched from literatures and confirmed via bioinformatic software (Miranda and miRtarbase). Apoptotic peptidase activating factor 1 (Apaf1) gene was found to be one of the predicted target genes for miR-136, and Apaf1 has been reported to play crucial role in regulating cell survival and apoptosis [21]. The expression of Apaf1 mRNA in OMSCs is higher than that in UMSCs (Fig 6a). When OMSCs were treated with agonist (MDK83190) of Apaf1, level of aging-related factors: p53, p21, and p16 were upregulated and level of Sirt1 was downregulated. Whearas, inhibitor (ZYZ-488) of Apaf1 reversed this phenomenon (Fig 6b). There data confirm that Apaf1 negatively affects cell aging.
We then examined the expression of Apaf1 in OMSCs after treated with ExoUMSCs or miR-136 mimic. Declined level of Apaf1 mRNA was observed in ExoUMSCs-treated or miR-136-transfected OMSCs, as compared with untreated OMSCs or miR-NC transfected OMSCs, respectively (Fig 6c). In agreement with this, Apaf1 mRNA was increased in UMSCs after treated with miR-136 inhibitor. These results confirmed that the expression of Apaf1 was negatively regulated by miR-136.
Two putative binding sites of miR-136 were identified at 3’UTR of Apaf1 (positions 138-145 and 1091-1098) through bioinformatic analysis (Fig 6d). To ascertain miR-136 directly binds to the 3’UTR of Apaf1 and causes translational inhibition, such two predicted 3’UTRs of Apaf1 and their mutants were cloned into a luciferase reporter construct pmirGLO (Fig 6d). The luciferase assay revealed that miR-136 mimic significantly reduced the activity of luciferase when the reporter gene was fused with the Apaf1 3’UTR, as compared with the miR-NC (Fig 6e). Mutation of both putative miR-136 sites in the 3’UTR of Apaf1 abrogated the effect of miR-136 on luciferase activity, while single mutation could not achieve such abrogation. Expressions of APAF1 protein and its downstream target Caspase 9 were both suppressed in OMSCs after transfection with miR-136 mimic (Fig 6f). These results indicated that Apaf1 is the downstream effecter of miR-136 through directly binding at the two sites of 3’UTR region.
More miR-136 in cord blood than in adult circulation.
To investigate the expression of miR-136 during aging, blood from 29 healthy male adults (age>40 years) and umbilical cord blood from 9 healthy maternity as young group were collected, and the level of miR-136 in the plasma was measured by qRT-PCR. The expression level of miR-136 was significantly higher in cord blood than that in blood from adults (Fig S12).