miR-155-5p inhibition rejuvenates aged mesenchymal stem cells and enhances cardioprotection following infarction.

Abstract Aging impairs the functions of human mesenchymal stem cells (MSCs), thereby severely reducing their beneficial effects on myocardial infarction (MI). MicroRNAs (miRNAs) play crucial roles in regulating the senescence of MSCs; however, the underlying mechanisms remain unclear. Here, we investigated the significance of miR‐155‐5p in regulating MSC senescence and whether inhibition of miR‐155‐5p could rejuvenate aged MSCs (AMSCs) to enhance their therapeutic efficacy for MI. Young MSCs (YMSCs) and AMSCs were isolated from young and aged donors, respectively. The cellular senescence of MSCs was evaluated by senescence‐associated β‐galactosidase (SA‐β‐gal) staining. Compared with YMSCs, AMSCs exhibited increased cellular senescence as evidenced by increased SA‐β‐gal activity and decreased proliferative capacity and paracrine effects. The expression of miR‐155‐5p was much higher in both serum and MSCs from aged donors than young donors. Upregulation of miR‐155‐5p in YMSCs led to increased cellular senescence, whereas downregulation of miR‐155‐5p decreased AMSC senescence. Mechanistically, miR‐155‐5p inhibited mitochondrial fission and increased mitochondrial fusion in MSCs via the AMPK signaling pathway, thereby resulting in cellular senescence by repressing the expression of Cab39. These effects were partially reversed by treatment with AMPK activator or mitofusin2‐specific siRNA (Mfn2‐siRNA). By enhancing angiogenesis and promoting cell survival, transplantation of anti‐miR‐155‐5p‐AMSCs led to improved cardiac function in an aged mouse model of MI compared with transplantation of AMSCs. In summary, our study shows that miR‐155‐5p mediates MSC senescence by regulating the Cab39/AMPK signaling pathway and miR‐155‐5p is a novel target to rejuvenate AMSCs and enhance their cardioprotective effects.


| INTRODUC TI ON
Despite the existing treatments, including percutaneous coronary intervention and coronary artery bypass grafting, myocardial infarction (MI) is still the main cause of morbidity and mortality worldwide, particularly in elderly patients . Over the last few decades, mesenchymal stem cell (MSC)-based therapy has emerged as a novel alternative treatment for MI (Kim et al., 2018;Teerlink et al., 2017). Accumulating evidence has demonstrated that transplantation of MSCs can attenuate cardiac remodeling and improve heart function recovery following MI by inhibiting cardiomyocyte apoptosis, increasing angiogenesis, and rejuvenating cardiac muscle cells Zhang et al., 2017Zhang et al., , 2019. However, the functions of MSCs dramatically decline with aging, as evidenced by increased cellular senescence, impaired proliferative capacity, and decreased paracrine effects, thereby heavily limiting their cardioprotective effects following MI (Liu et al., 2014;Zhang et al., 2018). Therefore, exploring a novel strategy to rejuvenate aged MSCs (AMSCs) to enhance their therapeutic effects for elderly patients with MI is urgently needed.
MicroRNAs (miRNAs), a class of ~21-23 nucleotide long noncoding RNAs, are critical repressors of gene expression by virtue of binding to the 3′-untranslated region (UTR) of target mRNAs . miRNAs have been reported to be involved in mediating multiple biological processes of stem cells, including cell division, differentiation, and survival Zhou et al., 2019).
Recently, increasing evidence has demonstrated that miRNAs play crucial roles in regulating the cellular senescence of MSCs (Meng et al., 2018;Yoo, Kim, Jung, Lee, & Kim, 2014). miR-495 targets Bmi-1 and induces MSC senescence, as evidenced by enhanced β-galactosidase activity and reduced cell proliferation .
It has been reported that miR-155-5p is significantly enhanced with age in bone marrow-derived extracellular vesicles (Davis et al., 2017), which prompted us to study the relationship of miR-155-5p and aging. Furthermore, miR-155-5p was found to be elevated in dermal MSCs of psoriatic patients, indicating that miR-155-5p could impair the functions of MSCs (Hou et al., 2016). However, whether miR-155-5p regulates the cellular senescence of MSCs has not been determined.
It is well known that the imbalance between mitochondrial fusion and fission is closely associated with cellular senescence (Nishimura et al., 2018). TGF-β induces vascular progenitor cell senescence by stimulating mitochondrial fusion . However, whether miR-155-5p mediates MSC senescence by regulating mitochondrial dynamics and the potential underlying mechanisms remain to be addressed.
In the current study, we aimed to investigate the role of miR-155-5p in regulating the senescence of MSCs and explored the related molecular mechanisms. Furthermore, we also examined whether inhibition of miR-155-5p could rejuvenate AMSCs and improve cardioprotection when AMSCs were transplanted into a mouse model of MI.

| AMSCs exhibit increased cellular senescence
We first examined the surface antigens of young MSCs (YMSCs) and AMSCs. Flow cytometry analysis showed that both YMSCs and AMSCs had similar surface markers and were positive for CD73, CD90, and CD105 and negative for CD31 and CD45 ( Figure   S1A). Next, we evaluated the differentiation capacity of YMSCs and AMSCs. The results showed that both YMSCs and AMSCs differentiated into adipocytes and osteocytes, as confirmed by oil red O staining and alizarin red staining, respectively (Figure S1B,C).
In addition, AMSCs exhibited increased adipogenic capacity and decreased osteogenic capacity compared with YMSCs, suggesting that the functions of AMSCs were impaired (Figure S1B,C). A previous study showed that senescent MSCs display increased adipogenic and decreased osteogenic differentiation capacities (Ma et al., 2018); thus, we examined the cellular senescence of YMSCs and AMSCs. Cell growth curves showed that AMSCs exhibited lower proliferative ability and arrested at passage 7, of AMSCs. In summary, our study shows that miR-155-5p mediates MSC senescence by regulating the Cab39/AMPK signaling pathway and miR-155-5p is a novel target to rejuvenate AMSCs and enhance their cardioprotective effects.
Furthermore, compared with YMSCs, AMSCs exhibited increased levels of senescence-associated β-galactosidase (SA-β-gal) activity ( Figure 1b) and increased expression levels of p53 and p21 protein ( Figure 1c). Additionally, Ki67 immunostaining showed a lower proliferative ability of AMSCs than YMSCs (Figure 1d). A wound healing assay demonstrated decreased migration ability of AMSCs compared with YMSCs ( Figure S1D). Promoting neovascularization is one of the major mechanisms underlying MSCbased therapy for MI. We therefore evaluated the angiogenic capacity of conditioned medium (CdM) from YMSCs and AMSCs.
As shown in Figure 1e, tube length was significantly increased in the CdM from MSCs compared with DMEM. Notably, compared with YMSCs-CdM treatment, AMSCs-CdM treatment presented decreased endothelial network formation capacity (Figure 1e).
Taken together, these data show that AMSCs exhibit increased cellular senescence.
Additionally, miR-155-5p levels were much higher in AMSCs than  Figure   S2E) and the angiogenic capacity of CdM from AMSCs ( Figure S2F).
Collectively, these findings suggest that miR-155-5p mediates the cellular senescence of MSCs.

| miR-155-5p induces cellular senescence of MSCs by regulating mitochondrial dynamics
Our previous study showed that mitochondrial fusion contributes to replicative senescence of MSCs ; thus, we explored whether miR-155-5p induces cellular senescence of MSCs by regulating mitochondrial dynamics. We first examined the mitochondrial morphology in YMSCs and AMSCs. As shown in

| miR-155-5p regulates mitochondrial dynamics via the Cab39/AMPK signaling pathway
Our previous studies showed that the AMPK signaling pathway plays a critical role in regulating mitochondrial dynamics; thus, we aimed to determine whether miR-155-5p regulates mitochondrial dynamics via the AMPK signaling pathway Li et al., 2019). We used TargetScan (http://www.targe tscan.org/) to predict the target genes of miR-155-5p and found a potential binding sequence in the 3′UTR of calcium-binding protein 39 (Cab39) To evaluate whether endogenous miR-155-5p regulated Cab39, we transfected YMSCs and AMSCs with wild-type pGL3-Cab39-3′-UTR luciferase reporter, respectively, and then examined the luciferase activity. As shown in Figure S4, compared with YMSCs, the luciferase activity was dramatically reduced in AMSCs, indicating a negative relationship between endogenous miR-155-5p and Cab39 ( Figure S4)

| Transplantation of anti-miR-155-5p-AMSCs improves cardiac function following infarction in aged mice
To showed that although the number of surviving MSCs was the highest in heart tissue from the YMSC group, MSC survival was much higher in the anti-miR-155-5p-AMSC group than in the AMSC group ( Figure 5d). To further confirm the survival of transplanted MSCs in ischemic heart tissue of mice, we detected the human repeat sequences Alu-sx in heart tissue using PCR. As shown in Figure 5f, Alu-sx was detected in all MSC groups, but not in the sham group and MI group (Figure 5f). Notably, the expression of Alu-sx was greatly enhanced in the anti-miR-155-5p-AMSC group than in the AMSC group (Figure 5f). Taken together, these data show that inhibition of miR-155-5p in AMSCs can improve cardioprotection following MI in mice.

| Anti-miR-155-5p-AMSC transplantation inhibits cardiomyocyte apoptosis and enhances angiogenesis in infarcted mouse hearts
To evaluate the anti-apoptotic effects of MSC transplantation, the apoptosis of cardiomyocytes in the ischemic area was deter-  (Figure 6a,b). Moreover, fewer apoptotic cardiomyocytes were observed in the hearts from the anti-miR-155-5p-AMSC group than in the AMSC group (Figure 6a,b). To determine the angiogenic effects of MSC transplantation, the arteriole densities and capillary densities were examined by α-SMA staining and CD31 staining in mouse hearts at 28 days post-transplantation, respectively. As shown in Figure 6c,d, the arteriole density was significantly increased in all MSC-treated groups compared with the MI group, and the arteriole density was the highest in the YMSC group (Figure 6c,d). Notably, more arterioles were found in the anti-miR-155-5p-AMSC group than in the AMSC group (Figure 6c,d). Consistent with these findings, a similar result was shown in the capillary densities from the different MSC-treated groups. The capillary density was the highest in the heart tissue from the YMSC group, and more capillaries were formed in the anti-miR-155-5p-AMSC group than in the AMSC group (Figure 6e,f). Collectively, these findings suggest that anti-miR-155-5p-AMSC transplantation inhibits cardiomyocyte apoptosis and enhances angiogenesis in infarcted mouse hearts.

| D ISCUSS I ON
The current study presented several major findings. First, miR- F I G U R E 5 Transplantation of anti-miR-155-5p-AMSCs improved heart function following infarction in aged mice. (a) Representative echocardiography images taken 28 days after MI in aged mice that received injections of PBS, YMSCs, AMSCs, or anti-miR-155-5p-AMSCs or control mice. (b) The LVEF and LVFS were evaluated at baseline (before MI), 1 and 28 days in control or aged mice with MI that received injections of PBS, YMSCs, AMSCs or anti-miR-155-5p-AMSCs. (c) Representative images of Masson's trichrome staining and quantitative analysis of infarction size in control or aged mice with MI that received injections of PBS, YMSCs, AMSCs, or anti-miR-155-5p-AMSCs. Scale bar = 2.5 mm. (d) Representative images of HNA staining and quantitative analysis of cell survival in aged mice that received injections of YMSCs, AMSCs or anti-miR-155-5p-AMSCs at 28 days post-MI. Scale bar = 50 μm. (e) Representative PCR image of Alu-sx in the ischemic heart tissue in aged mice that received injections of YMSCs, AMSCs or anti-miR-155-5p-AMSCs at 28 days post-MI. Data are expressed as the mean ± SEM. n = 6-7. **p < .05; **p < .01; ***p < .001 with MSCs collected from young mice. Consistent with these findings, we observed that miR-155-5p was significantly increased in human AMSCs and serum from aging donors in the current study, suggesting that miR-155-5p may be a potential factor in regulating MSC senescence. We further found that overexpressing miR-155-5p in YMSCs enhanced the cellular senescent phenotype, including increased SA-β-gal activity and expression of p21 and p53 and decreased Ki67-positive cells. Furthermore, the angiogenic capacity of CdM from miR-155-5p-treated YMSCs was also downregulated. In contrast, inhibition of miR-155-5p in AMSCs reduced SA-β-gal activity and increased cell proliferation and angiogenesis. Transplantation of anti-miR-155-5p-AMSCs had a better capacity to attenuate cardiac remodeling and restore heart function in mice following infarction than transplantation of AMSCs.
The exact mechanism underlying miR-155-5p regulation of MSC senescence, however, is still largely unknown.
Mitochondrial fusion and fission are essential to maintain cell function, and abnormal mitochondrial dynamics accelerate cellular senescence (Rizza et al., 2018). Our previous study showed that late passage MSCs exhibited large tubular mitochondria compared with early passage MSCs, suggesting that mitochondrial fusion contributes to replicative cellular senescence. Furthermore, knockdown of FGF21 in the early passage MSCs accelerated mitochondrial fusion, leading to cellular senescence ).
In the current study, we found that AMSCs also exhibited large tubular mitochondria accompanied by decreased p-Drp1 (Ser616) and increased Mfn2 levels, suggesting that an imbalance of mitochondrial dynamics mediates the physiological senescence of MSCs. Furthermore, treatment with a miR-155-5p mimic-induced mitochondrial fusion and senescence of YMSCs, and these effects were abrogated in part by Mfn2-siRNA, suggesting that miR-155-5p induced MSC senescence by regulating mitochondrial dynamics. Accumulating evidence has shown that AMPK signaling regulates mitochondrial dynamics (Hang et al., 2018;He et al., 2019). Whether miR-155-5p mediates mitochondrial fusion by regulating AMPK signaling remains unclear. We sought to identify miR-155-5p targets and found that Cab39 is a direct target gene of miR-155-5p. It has been reported that Cab39, an upstream coactivator of AMPK, exhibits cell protective functions (Kuwabara et al., 2015). In the current study, we found that the expression of Cab39 and p-AMPK was greatly reduced in AMSCs compared with YMSCs, suggesting that Cab39/AMPK signaling is associated with the cellular senescence of MSCs. Furthermore, miR-155-5p mimic treatment greatly downregulated the expression of Cab39 and p-AMPK. Combined with miR-155-5p mimic-induced mitochondrial fusion, we sought that miR-155-5p-induced mitochondrial fusion may be regulated by Cab39/AMPK signaling. Moreover, we found that AICAR, an AMPK activator, partially rescued miR-155-5p-induced mitochondrial fusion. These results further confirmed that miR-155-5p induces mitochondrial fusion by partially targeting Cab39/AMPK signaling. However, AMPK possess various bypass signaling cascades to regulate its function. In addition to Cab39, we cannot exclude that miR-155-5p regulates AMPK activation via other pathways or molecules.
There are some limitations in the current study. First, only miR-155-5p in AMSCs was investigated. The functions of other miRNAs that are significantly enriched in AMSCs require further investigation. Second, it has not been investigated yet whether miR-155-5p regulates other targets to mediate MSC senescence in addition to Cab39. Third, the long-term impact of anti-miR-155-5p-AMSC on heart function recovery following infarction was not examined in this study. Fourth, it would be important to confirm the alteration of miR-155-5p-dependent biological pathways also in heart tissue of MI. Last but not least, more strong and more nonbiased data such as phenotype observation of Cab39-KO/TG mice and omics-based screening are needed to verify our findings in the future study.
In summary, our study demonstrates that inhibition of miR-155-5p, partially via the Cab39/AMPK signaling pathway, rejuvenates aged MSCs by regulating mitochondrial dynamics and provides a candidate target to enhance the cardioprotection of MSCs for the aged heart following infarction.

| Cell culture
Human YMSCs and AMSCs were isolated from the bone marrow of young and aged volunteer donors, respectively, as previously de- YMSCs and AMSCs at passages 3-4 were used in the current study.

| Characterization of MSCs
The

| Scratch wound assay
YMSCs and AMSCs were cultured in a 6-well plate with complete culture media until they were 100% confluent. Then, scratches of the same width cross the entire well were made using a 200-µl pipette tip. Subsequently, MSCs were carefully washed with PBS to remove cell debris and then incubated with serum-free medium in an incubator with 5% CO 2 at 37°C. After 24 hr of incubation, the migration of MSCs into the wound area was examined. The experiments were repeated at least three times.

| Preparation of CdM derived from MSCs
Both YMSCs and AMSCs were seeded into 6-well plates with growth medium and cultured until 70%-80% confluence. After different treatments, the medium was replaced with 2 ml per well serum-free medium. After 48 hr, the CdM was harvested, centrifuged, filtered, and stored at −80°C until use.

| Immunofluorescence staining
Mesenchymal stem cells were fixed with 4% paraformaldehyde

| Luciferase assay
The 3′-UTR of human Cab39 was inserted into the pGL3 luciferase

| Viral vector construction and infection
The lentiviral plasmid constructs for the inhibition of miR-155-5p in AMSCs were purchased from GenePharma. The plasmid contained an expression cassette consisting of a CMV promoter followed by cDNA encoding eGFP and an anti-miR155-5p sequence ( Figure S7).
The lentivirus was packaged as previously reported . For stable transduction, AMSCs at a confluence of 70%-80% were infected by lentivirus at a multiplicity of infection of 10 with polybrene (8 μg/ml). The infection efficiency was determined by eGFP fluorescent signal viewed under the microscope at 48 hr after infection and the cells were labeled as anti-miR-155-5p-AMSCs.

| MI model and transplantation of MSCs
All animal procedures in the current study were approved by the used to induce an acute MI model by ligation of the left anterior decedent coronary artery (LAD) as previously described . After LAD ligation, mice randomly received one of the following treatments: (a) phosphate-buffered saline (PBS) (MI group, n = 12); (b) 3 × 10 5 YMSCs (YMSC group, n = 11); (c) 3 × 10 5 AMSCs (AMSC group, n = 12) or 3 × 10 5 anti-miR-155-5p-aged MSCs (anti-miR-155-5p-AMSC group, n = 12). All MSCs were suspended in 100 μl PBS and were injected intramuscularly at four sites around the border zone of the infarcted heart. Another group of mice that underwent thoracotomy without LAD ligation served as the control group (Sham group, n = 6). Cardiac function in each mouse was assessed by transthoracic echocardiography (Ultramark 9; Soma Technology) at baseline (before MI), 1, or 28 days following MI. LVEF and LVFS were calculated as previously described .

| Masson's staining
After echocardiography assessment at 28 days post-MI, all the mice were sacrificed, and the heart tissues were harvested, embedded, and sectioned. The infarction size of the mouse heart, as evidenced by fibrosis, was examined by Masson's staining kit according to the manufacturer's protocol (HT15, Sigma). The percent infarct size was calculated as the ratio of fibrosis area to total LV area ×100%.

| TUNEL staining
The apoptosis of cardiomyocytes in the heart tissue from different groups was evaluated by TUNEL staining as previously described (Zhang et al., 2015). The sections were mounted with 4′, 6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Inc.) and imaged using a fluorescence microscope. The apoptotic rate was calculated as the ratio of TUNEL-positive cells to DAPI-positive cells ×100%.

| Statistical analysis
All data are expressed as the mean ± SEM. Statistical analyses were performed using Prism 5.04 software (GraphPad Software for Windows). Comparisons between two groups were analyzed by unpaired Student's t test, and comparisons between more than two groups were analyzed by one-way ANOVA followed by Bonferroni test. A value of p < .05 was considered statistically significant.

ACK N OWLED G M ENTS
This research was supported in part by the National Natural Science Cell Clinical Technology Transformation Platform" (ZJ2018-ZD-004).

CO N FLI C T O F I NTE R E S T
The authors confirm that they have no conflicts of interest.