Globular Adiponectin Inhibits the Apoptosis of Mesenchymal Stem Cells Induced by Hypoxia and Serum Deprivation via the AdipoR1-Mediated Pathway

Stem cell therapy is a promising new approach for the treatment of ischemic heart diseases [1]. Among all types of cells under investigation, mesenchymal stem cells (MSCs) are most widely studied due to the abundance of autologous sources and the ability to repair the injured myocardium [2,3,4]. Clinical trials revealed that injection of MSCs favorably affected the functional capacity, quality of life, and ventricular remodeling in patients with ischemic cardiomyopathy [5]. However, data indicated that there appeared to be slight improvement of left ventricular ejection fraction after MSCs transplantation for acute myocardial infarction [6,7]. This was probably attributed to the low survival rate of MSCs after being exposed to the hostile engraftment environment with various proapoptotic or cytoxic factors [8], including ischemia, oxidative stress, inflammation and fibrosis [9]. It was demonstrated that hypoxia and serum deprivation (H/SD), combining the two components of ischemic injury, induced programmed MSCs death through the mitochondrial apoptotic pathway [10]. Therefore, identifying factors that prevent this process and enhance the viability of MSCs under the harsh microenvironment in the ischemic heart could be crucial for MSCs' successful utilization in cellular therapy.

Adiponectin (APN, also known as Acrp30) is an adipocytokine predominantly expressed in adipocytes. It exists as two forms: the full-length peptide (fAPN) and the globular C-terminal domain (gAPN). As a product of proteolytic cleavage, gAPN is biologically active and more potent in physiological actions [11,12]. Apart from its well-characterized role in fat tissue metabolism and insulin resistance [13], accumulating evidences established the cardioprotective properties of gAPN [14,15,16,17]. Much of this beneficial effect was attributed to the anti-apoptotic actions of gAPN on cardiovascular cells [18,19,20]. Besides, previous experiments revealed that gAPN played an important modulatory role on the survival and proliferation of several types of stem cells [21,22,23,24,25,26]. More importantly, gAPN has recently been found to promote the proliferation and migration of MSCs [27]. Taken the anti-apoptotic and stem cell-regulatory properties together, it is reasonable to hypothesize that gAPN acts as an anti-apoptotic factor in MSCs under H/SD conditions, which might offer a solution to the efficacy problem of MSCs transplantation in ischemic heart diseases.

Adiponectin receptor 1 and receptor 2 (AdipoR1 and AdipoR2) have been proved to mainly mediate the action of APN on target tissues [28]. AdipoR1 is abundantly expressed and has high affinity for gAPN, whereas AdipoR2 is predominantly expressed in the liver and has high affinity for fAPN [29]. Binding of APN with AdipoR1 or AdipoR2 leads to the subsequent activation of various signaling pathways, especially the adenosine monophosphate-activated protein kinase (AMPK) pathway [30]. AMPK is a metabolic related protein kinase that regulates energy consumption and production, and activated AMPK limits energy utilization to ensure cell survival [31]. Moreover, AMPK is also crucial in the regulation of MSCs proliferation [32]. In this regard, the present study aimed to investigate the impact of gAPN on the apoptosis of MSCs under H/SD conditions, and further identify the main receptor(s) and the intracellular signaling mechanisms that may mediate such actions, focusing particularly on the AMPK pathways.

This study was performed in strict accordance with the Chinese guidelines for the care and use of laboratory animals. All animals received humane care and the experimental protocol was approved by the Care of Experimental Animals Committee of Fuwai Hospital.

Isolation and culture of adult rat bone marrow MSCs were performed as previously described [10]. In brief, bone marrow was harvested from the tibia and femur of Sprague-Dawley rats (60 - 80g, male) and plated into cell culture flasks with complete medium in an incubator set at 37°C containing 5% CO₂ and 95% air. The complete medium was consisted of Iscove's Modified Dulbecco's Medium (IMDM, Gibco, USA), 10% fetal bovine serum (Gibco, USA) and 1% penicillin-streptomycin (Gibco, USA). The d-glucose (dextrose) concentration of IMDM is 25 mM. After 24 hours, non-adherent cells were removed by washing with phosphate-buffered saline (PBS). The medium was changed every two to three days thereafter. When the cells reached 80% confluence, they were detached using 0.25% trypsin-EDTA (Gibco, USA) and subcultured at the ratio of 1:2. All cells used in the experiment were passage 3.

Immunophenotyping analysis was employed to identify the surface markers of MSCs. The cells were harvested, washed with PBS, and incubated with fluorescein isothiocyanate (FΙTC)-labeled anti-CD45, phycoerythrin (PE)-conjugated anti-CD29 and Peridinin-Chlorophyll-Protein (PerCP) Cyanine5.5-labeled anti-CD90 (all from eBioscience, USA), as well as allophycocyanin (APC)-labeled anti-CD11b (from Miltenyi Biotec, Germany) antibodies at 4°C for 10 minutes. After washing with PBS, labeled cells were resuspended in 200 µL PBS and analysed with FACS Calibur System (Becton-Dickinson).

The MSCs were exposed to H/SD conditions for 6 hours as previously described [10]. Briefly, MSCs were washed with PBS twice, and different concentrations (0.01 µg/mL, 0.1 µg/mL, 1 µg/mL) of gAPN (Biovision, USA) were added into the serum free medium. Then MSCs were put into a sealed GENbox hypoxic chamber fitted with an AnaeroPack (Mitsubishi Gas Chemical Company, Japan) to scavenge the free oxygen at 37°C for 6 hours. The oxygen concentration was maintained below 0.1% after 0.5 -1 hour as indicated by an Anaer indicator (Bio-Me'rieux, Marcy I'Etoile, France). MSCs cultured in complete medium under non-hypoxic condition were regarded as the non-apoptotic control group. For pathway studies, the cells were pre-incubated with AMPK inhibitor Compound C (10 µM, Abcam, UK) for 20 minutes before the addition of 1 µg/mL gAPN, and were then exposed to H/SD conditions.

Cells were grown in complete medium in a 6-well plate. After reaching approximately 50% confluence, MSCs were transfected with small interfering RNA (siRNA, Genepharma, China) designed to knockdown rat AdipoR1 or AdipoR2 or the combined of them with lipofectmin2000. Scrambled siRNA (Genepharma, China) was transfected as control. The sequences of siRNAs ultimately used for providing optimal efficiency were AdipoR1, sense 5'-GGC UCU AUU ACU CCU UCU ATT-3' and antisense 5'-UAG AAG GAG UAA UAG AGC CTT-3'; AdipoR2, sense 5'-GAG CCA GAU AUA AGG CUC ATT-3' and antisense 5'-UGA GCC UUA UAU CUG GCU CTT-3', and scrambled control, sense 5'-UUC UCC GAA CGU GUC ACG UTT-3' and antisense 5'-ACG UGA CAC GUU CGG AGA ATT-3'.

Cell nuclear condensation and fragmentation were assessed using chromatin dye Hoechst 33342 as previously described [10]. Briefly, cells were fixed in 4% paraformaldehyde at room temperature for 30 minutes, and washed with PBS twice, then incubated with 5 mg/L Hoechst 33342 fluorochrome at room temperature (protected from light) for 30 minutes. After rewashing with PBS twice, observations were conducted under a fluorescent microscope (Leica, Germany). Apoptotic cells were identified as fragmented and condensed apoptotic nuclei.

Cell apoptosis was assessed with flow cytometry using Annexin V-FITC Apoptosis Detection Kit (Becton-Dickinson, USA) in accordance with the manufacturer's instructions. Briefly, after the desired treatments, MSCs were trypsinized and washed once with PBS. Cells were then resuspended in 100 µL 1× binding buffer. Then, 5 µL Annexin V and 5 µL propidium iodide (PI) solution were added to the binding buffer and incubated in the dark at room temperature for 20 minutes. After the staining, 400 µL 1× binding buffer were added to the cell suspension and 1 × 10⁴ stained cells were analyzed using FACS Calibur System (Becton-Dickinson).

The activity of caspase 3 was measured using the Caspase 3/CPP32 Colorimetric assay kit (Biovision, USA). In brief, after the desired treatment, cells were collected and resuspended in 50 µL of chilled lysis buffer on ice for 10 minutes, and centrifuged at 10,000 g for 1 minute. The supernatant was transferred to a microtube, and the protein concentration was measured using BCA protein assay. Then, 100 µg protein was diluted to 50 µL cell lysis buffer for each assay. Next, 50 µL of 2× Reaction Buffer containing 10 mM DTT were added, followed by adding 5 µL of 4 mM DEVD-pNA substrate (200 µM final concentration) to each sample. After incubation at 37°C for 2 hours, 100 µL of each sample was transferred to a 96-well plate, and the cleavage of the DEVD-pNA was determined by reading the absorbance at 405 nm in a microplate reader.

The loss of mitochondrial membrane potential was assessed using the fluorescent dye 5, 5', 6, 6'-tetrachloro-1, 1', 3, 3'-tetraethyl benzimidazolylcarbcyanine iodide (JC-1) (Beyotime, China) based on the manufacturer's instructions. In brief, after treatment with indicated agents and exposure to H/SD, MSCs were incubated with 0.5 mL complete medium and 0.5 mL JC-1 staining solution at 37°C for 20 minutes, and then rinsed twice with ice-cold JC-1 staining buffer. The cells were viewed and photographed under a fluorescent microscope (Leica, Germany). To quantify the changes of MMP, the red/green fluorescence intensity was assessed with a fluorescence plate reader. Briefly, MSCs were harvested and resuspended in 0.5 mL complete medium. After incubating with 0.5 mL JC-1 staining solution at 37°C for 20 minutes, MSCs were centrifuged at 600 g for 3 minutes and washed twice with ice-cold JC-1 staining buffer. Then, MSCs were resuspended in 200 µL staining buffer and transferred to a black 96-well plate. Finally, the MMP was determined by the ratio of red over green fluorescent emission in a microplate reader. The excitation and emission wavelengths were 490 nm and 530 nm for detection of monomeric form of JC-1 (green fluorescence), and 525 nm and 590 nm for detection of aggregated form of JC-1 (red fluorescence).

Cells were collected and rinsed with ice-cold PBS twice and then lysed in ice-cold lysis buffer on ice for 30 minutes. Cell lysates were centrifuged at 13,000g at 4°C for 15 minutes and the protein concentrations were determined by the BCA protein assays. Equal amounts of proteins (20 µg/lane) were separated on NuPage 4% -12% Bis-Tris Gels (Novex, Life technologies, USA) by electrophoresis for Western blot analysis. The proteins were then transferred to nitrocellulose membranes using dry electroblotting apparatus (Invitrogen, USA), and the membranes were blocked in 5% skim milk at room temperature for 2 hours. Next, the membranes were incubated with primary antibodies in skim milk over night at 4°C. The primary antibodies used were as follows: β-actin (1:1000), pro-caspase-3 (1:1000), Cl-caspase-3 (1:250), Bcl-2 (1:1000), Bax(1:500), phosphorylated (Thr172)-AMPK (1:500) and AMPK (1:1000). All the above antibodies were from Cell Signal Technology. AdipoR1 (1:500) and AdipoR2 (1:500) antibodies were from Santa Cruz Biotechnology. In the following day the membranes were washed three times and secondary antibodies were added and incubated for 2 hours. After washing, the membranes were processed for analysis using an enhanced chemiluminescence detection system (FluorChem M, USA). The target signals were normalized to the β-actin signal and analyzed semi-quantitatively with Quantity One system.

Data are expressed as mean ± standard deviation (SD) from at least 3 independent experiments. Differences among groups were tested by one-way analysis of variance (ANOVA) and Tukey's multiple comparisons via SPSS Statistics software version 22.0. P value less than 0.05 was considered as statistical significance.

The MSCs obtained from the bone marrow of the SD rats exhibited a spindle-like appearance. The results of FACS analysis revealed that the majority of adherent cells from passage 3 were positive for CD29 (99.00 ± 0.20%) and CD90 (98.47 ± 0.81%), but were negative for CD45 (1.5 ± 0.53%) and CD11b (1.62 ± 0.11%) (Fig. 1).

To delineate the effect of gAPN on the apoptosis of MSCs under H/SD conditions, different concentrations of gAPN were added into the serum-free medium, and apoptosis was determined morphologically. As shown in Fig. 2A, cells exposed to H/SD conditions appeared to have shrunken and condensed nuclei. In contrast, inclusion of gAPN dose-dependently reduced the nuclear change induced by H/SD, with 1 µg/mL closely reaching the control level.

Apoptotic cells were then quantified using flow cytometry after Annexin V and PI staining. The Annexin V-FITC assay detected the exposure of phosphatidylserine (PS) to the external surface of the plasma membrane in early apoptosis, whereas the PI assay primarily measured late apoptotic and necrotic cells [33]. The early apoptotic cells (Annexin V+ ⁄ PI-), late apoptotic cells (Annexin V+ ⁄ PI+) and necrotic cells (Annexin V- ⁄ PI+) cells were thus discriminated. Flow cytometry results indicated that, compared with the control group, H/SD significantly increased the percentage of both early (9.05 ± 1.20% vs. 1.13 ± 0.42% in control group, p < 0.05) and late (13.03 ± 0.91% vs. 3.42 ± 0.65% in control group, p < 0.05) apoptotic MSCs. Conversely, gAPN treatment concentration-dependently led to a decrease in apoptotic cells induced by H/SD. The anti-apoptotic effect reached its peak at 1 µg/mL gAPN, with a reduction in early (1.56 ± 0.27% vs. 9.05 ± 1.20% in H/SD group, p < 0.05) and late (4.54 ± 0.96% vs. 13.03 ± 0.91% in H/SD group, p < 0.05) apoptotic cells (Fig. 2B, C).

A typical feature of apoptosis is the activation of caspases, a family of cysteine proteases that plays an essential role in this process. To specify the involvement of gAPN in the survival of MSCs after H/SD, the changes of caspase-3 activity with reference to the control values were analyzed. Results showed that exposure of MSCs to H/SD brought about an approximately 4 to 5 fold increase in the caspase-3 activity, which was inhibited by gAPN in a dose-dependent manner, with 1 µg/mL almost reaching that of the control group (Fig. 2D). The activation of caspase-3 was further expressed as the cleavage of pro-caspase-3 into its active subunit, cleaved caspase-3 (Cl-caspase-3), using the Western blot analysis (Fig. 3A). Compared to the H/SD group, the cleavage of pro-caspase-3 was markedly decreased with the addition of gAPN (Fig. 3B, C). These data suggested that gAPN dose-dependently protected MSCs from apoptosis induced by H/SD, and the maximum effect was observed at 1 µg/mL gAPN, which was used in subsequent experiments.

To investigate the main type of receptor that mediated the actions of gAPN, siRNAs targeting AdipoR1 or AdipoR2 or both of them were used. The knockdown efficiency was confirmed by evaluating the expressions of these proteins using Western blot. As shown in Fig. 4, AdipoR1 and AdipoR2 expressions were significantly suppressed after using corresponding siRNA but not with scrambled siRNA or the unspecific siRNA. Then flow cytometry and caspase-3 activity kit were employed as quantitative assays to decide the functional significance of manipulating AdipoR1 or AdipoR2 expression. As displayed in Fig. 5A-B, compared with the scrambled group, the knockdown of AdipoR1 resulted in a significant increase in apoptotic cells at early (8.12 ± 1.64% vs. 1.58 ± 0.21% in scrambled group, p < 0.05) and late (8.11 ± 1.98% vs. 4.61 ± 0.85% in scrambled group, p < 0.05) phases, whereas no obvious changes were observed with the knockdown of AdipoR2 at early (2.25 ± 0.31% vs. 1.58 ± 0.21% in scrambled group, p > 0.05) and late (4.85 ± 0.97% vs. 4.61 ± 0.85% in scrambled group, p > 0.05) phases. Suppression of both AdipoR1 and AdipoR2 expression in combination by siRNA almost abolished the protective effect of gAPN against early (8.73 ± 1.47% vs. 1.58 ± 0.21% in scrambled group, p < 0.05; 8.73 ± 1.47% vs. 9.05 ± 1.20% in H/SD group, p > 0.05) and late (11.33 ± 0.91% vs. 4.61 ± 0.85% in scrambled group, p < 0.05; 11.33 ± 0.91% vs. 13.03 ± 0.91% in H/SD group, p > 0.05) apoptosis. The differences between si-AdipoR1 group and combined knockdown of AdipoR1 and AdipoR2 group were not statistically significant. In parallel, the caspase-3 activity had a greater increase with the knockdown of AdipoR1 (2.91 ± 0.64 vs. 1.08 ± 0.08 in scrambled group, p < 0.05) than that of AdipoR2 (1.44 ± 0.32 vs. 1.08 ± 0.08 in scrambled group, p > 0.05). The combined knockdown of AdipoR1 and AdipoR2 almost reversed the suppressive effect of gAPN on caspase-3 activity (3.84 ± 0.18 vs. 1.08 ± 0.08 in scrambled group, p < 0.05; 3.84 ± 0.18 vs. 4.65 ± 0.84 in H/SD group, p > 0.05) (Fig. 5C). These findings indicate that it was AdipoR1, rather than AdipoR2 that mainly mediated the anti-apoptosis effect of gAPN against H/SD conditions.

APN exerts its action in various cell systems by activating AMPK signaling pathways [14,18,30,34,35,36]. To determine whether this signaling pathway was stimulated by gAPN in MSCs, phosphorylation of AMPK was assessed following treatment with different concentrations of gAPN under H/SD conditions for 6 hours. Results showed that gAPN induced a dose-dependent increase of phosphorylation of AMPK (p-AMPK) at Thr-172 with 1 µg/mL gAPN reached its peak, and this effect was significantly inhibited by pretreatment with Compound C (10 µM) (0.51 ± 0.09 vs. 1.25 ± 0.15 in 1 µg/mL gAPN group, p < 0.05) (Fig. 6A). Further, compared with the scrambled group, the knockdown of AdipoR1 partly inhibited the p-AMPK (0.75 ± 0.13 vs. 1.25 ± 0.15 in scrambled group, p < 0.05) (Fig. 6B).

We extended our study by examining whether activation of this pathway was critical for the anti-apoptotic actions of gAPN in MSCs. Results showed that pre-incubation with Compound C partly suppressed the anti-apoptotic effect of gAPN on MSCs, evidenced by increased early apoptotic cells (6.38 ± 0.59% vs. 1.55 ± 0.27% in gAPN group, p < 0.05) and late apoptotic cells (8.83 ± 1.96% vs. 4.54 ± 0.96% in gAPN group, p < 0.05) (Fig. 7A, B), as well as elevated caspase-3 activity (3.22 ± 0.29 vs. 1.07 ± 0.2 in gAPN group, p < 0.05) (Fig. 7C). These results illustrated that gAPN inhibited the apoptosis of MSCs against H/SD conditions at least partly through the AMPK pathway.

The balance of anti-apoptotic protein Bcl-2 and pro-apoptotic protein Bax plays an important role in the regulation of mitochondrial integrity and cell survival [37]. To ascertain whether the mitochondrial-dependent apoptotic pathway was involved in the gAPN's protective effect on MSCs, the effect of gAPN on H/SD-induced Bcl-2 and Bax expression changes were examined. Western blot analysis revealed that H/SD significantly decreased Bcl-2 (0.51 ± 0.20 vs. 1.50 ± 0.25 in control group, p < 0.05) and increased Bax (1.29 ± 0.06 vs. 0.80 ± 0.26 in control group, p < 0.05) expressions. Compared to the H/SD group, 1 µg/mL gAPN-treated group significantly increased expression of Bcl-2 (1.42 ± 0.32 vs. 0.51 ± 0.20 in H/SD group, p < 0.05) and decreased expression of Bax (0.70 ± 0.21 vs. 1.29 ± 0.06 in H/SD group, p < 0.05) (Fig. 8A). In addition, the gAPN-induced up-regulation of Bcl-2 and down-regulation of Bax expressions were markedly attenuated with the knockdown of AdipoR1 and inhibition of p-AMPK by Compound C (Fig. 8B, C).

As the loss of mitochondrial membrane potential (MMP) is regarded as one of the inducers of mitochondrial-related apoptosis, whether gAPN influenced the loss of MMP in MSCs subjected to H/SD conditions was examined using JC-1 fluorescence staining assay. JC-1 exists as an aggregated form (red fluorescence) in the matrix of mitochondria with the normal MMP, and it is converted to the monomeric form (green fluorescence) with the loss of MMP. Consistent with the anti-apoptotic effect, treatment with 1 µg/mL gAPN led to a strong protection from MMP loss induced by H/SD, as determined by the reduction of green fluorescence and increase of red fluorescence in gAPN group compared to the H/SD group (Fig. 9A). To quantify the changes of MMP, the red/green fluorescence intensity was assessed with a fluorescence plate reader. Compared with the control group, H/SD induced a significant reduction of red/green fluorescence (0.68 ± 0.05 fold vs. control group, p < 0.05). The addition of gAPN significantly reversed the above change, with the red/green fluorescence closely reaching the control level (0.97 ± 0.04 vs. 0.68 ± 0.05 in H/SD group, p < 0.05). Moreover, the inhibition of the AdipoR1 (0.72 ± 0.04 vs. 0.93 ± 0.05 in scrambed group, p < 0.05) or AMPK pathway (0.82 ± 0.03 vs. 0.97 ± 0.04 in gAPN group, p < 0.05) partly suppressed the protective effect of gAPN on MMP expressed as the red/green fluorescence (Fig. 9B, C).

The therapeutic potential of MSCs for cardiac repair is hampered partly due to their low survival rate within the ischemic myocardial microenvironment into which MSCs are introduced [8]. Hypoxia and serum deprivation, imitating the ischemic engraftment environment, induced the apoptosis of MSCs through the mitochondrial apoptotic pathway [10]. This study found that gAPN could confer a protective effect against the above-mentioned process after binding with AdipoR1 and activating the downstream AMPK pathway. These results highlight an opportunity to improve the survival and therapeutic efficacy of MSCs by targeting gAPN.

APN is an adipocyte-secreted protein that elicits protective effects in the vasculature and myocardium with anti-diabetic, anti-atherosclerotic and anti-apoptotic properties [17]. It is proposed as a tissue-regenerating hormone and regulates the survival, migration and differentiation of several types of stem cells [21]. gAPN, the globular C-terminal domain of APN, is biologically active and easier to manufacture and administer [20,23,24,38]. Our results indicate that gAPN may enhance the survival of MSCs under H/SD conditions, potentially benefit the MSCs transplantation efficacy in addition to the cardioprotective effects. There is indeed precedence suggesting that APN serves as a prosurvival stem cell factor. It was reported that APN promoted the survival of bone marrow mononuclear cells (BM-MNCs) against apoptosis triggered by staurosporine or ceramide [22]. gAPN also counteracted the apoptotic process triggered by growth factor withdrawal in murine mesoangioblasts [24]. Besides, gAPN increased the endothelial repair and angiogenesis by increasing the number and function of endothelial progenitor cells (EPCs) [39,40]. In addition, Hou M et al. showed that C1q tumor necrosis factor-related protein-3 (CTRP3), a newly identified adipokine similar to APN, protected MSCs against H/SD-induced apoptosis through the PI3K/Akt pathway [41]. However, the up-stream molecular / receptor that mediates the action between CTRP3 and PI3K has not been investigated.

In the present study, hypoxia and serum deprivation were combined to imitate the ischemic environment and serve as the inducer of apoptosis in MSCs. It was found that serum deprivation was the predominant factor in this process [10], while hypoxia alone failed to induce apoptosis [32] and exposure to 1-1.5% O₂ even enhanced the proliferation of MSCs [42,43]. The resistance to hypoxia per se is largely due to the up-regulated glycolytic activities as indicated by the high-level expression of glycolytic enzymes in MSCs [44]. Effects of hypoxia are usually mediated by hypoxia inducible factor-1 (HIF-1). It was verified that over-expression of HIF-1alpha conferred resistance to hypoxia and oxidative stress-induced apoptosis in bone marrow-derived MSCs [45]. Wagegg M et al. found that hypoxia (less than 2% O₂) resulted in a significant time-dependent increase in the expressions of HIF-1alpha mRNA and protein in bone marrow-derived MSCs after 72 hours and 2 weeks of incubation when compared to cells under normoxia. However, there were no detectable protein expression of HIF-1 in both normoxia and hypoxia conditions within 24 hours [46]. As the exposure time of MSCs to hypoxia condition was 6 hours in our study, not long enough to induce the expression of HIF-1, and the main contributer of apoptosis is serum deprivation instead of hypoxia [10], HIF-1 might have little involvement in the anti-apoptotic effect of gAPN in our settings. Under serum deprivation condition, APN also exerts anti-apoptotic effects on endothelial cells [34] and ventricular myocytes [14], which may facilitate cardiac repair to a larger extent. Nevertheless, there were studies pointing out that APN could induce apoptosis in some cancer cell lines such as myelomonocytic leukemia cells [47], breast cancer cells [48] and endometrial cancer cells [49]. These findings suggested that the effects of APN on cell survival are highly dependent on cellular type [50].

It was observed in this study that targeted knockdown of AdipoR1 by specific siRNA significantly inhibited the protective effect of gAPN on H/SD-induced apoptosis of MSCs, while AdipoR2 siRNA only slightly suppressed the anti-apoptotic effect. The combined knockdown of AdipoR1 and AdipoR2 almost eliminated the protective effect of gAPN, although this effect was not statistically significant with the only knockdown of AdipoR1. Further, previous studies have shown that gAPN has higher affinity for AdipoR1 receptor subtype [28,51]. On top of that, the mRNA expression level of AdipoR1 was greater than that of AdipoR2 in MSCs, and AdipoR1 played a key role in mediating APN signaling cascade for the regulation of osteoblast differentiation in MSCs [52]. In this regard, it is reasonable to conclude that AdipoR1, rather than AdipoR2, was the main type of receptor mediating the anti-apoptotic effect of gAPN in our study. Our results correlate well with the previous reports suggesting that AdipoR1 primarily mediated the protective effects of gAPN on the H9C2 cells from hypoxia/reoxygenation-induced apoptosis [20]. It was also demonstrated that the metabolic effects of gAPN in primary neonatal cardiomyocytes were mediated via AdipoR1 [53]. In addition, AdipoR1 also mediated the gAPN's unregulation of EPCs function suppressed by high glucose [39].

Binding of gAPN with its specific receptor generates various signaling pathways that regulate cell survival and apoptosis, especially the AMPK pathway. The activation of AMPK by gAPN is critical in cellular responses to metabolic stress involving energy generation and consumption [30,36,54,55]. Under hypoxia, AMPK is normally activated and functions to maintain the balance of catabolism and anabolism in cells like cardiomyocytes [56], which re-establishes energy balance and enhances cell survival in the absence of oxygen. By contrast, hypoxic condition did not alter the phosphorylation of AMPK in MSCs, which was correlated with the maintenance of ATP concentration [32]. In line with the previous results, our data also revealed that hypoxia and serum deprivation alone did not activate AMPK, while gAPN induced sustained phosphorylation of AMPK in MSCs. Further, the application of Compound C, the inhibitor of AMPK, significantly inhibited the protection offered by gAPN against H/SD-induced apoptosis. Thus, taken together, these findings indicate that AMPK pathway might be involved in the anti-apoptotic actions of gAPN. However, Compound C was reported to have some AMPK-independent actions even though it was widely used to verify the role of AMPK in various (patho)-physiological processes [57,58]. Hence, the knockdown of AMPKalpha1 (the sole catalytic isoform expressed in MSCs) by RNA-silencing might provide more potent evidence of the role AMPK played in the above-mentioned process.

Our study is consistent with the previous reports showing that AMPK mediated the protective effects of APN on the apoptosis of endothelial cells induced by serum starvation [34]. Moreover, data suggested that APN counteracted ventricular myocytes and fibroblasts apoptosis triggered by serum deprivation and hypoxia-reoxygenation through AMPK-dependent mechanism [14]. However, there are reports pointing out that the gAPN significantly reduced cell death of cardiomyocytes induced by simulated ischemia and reperfusion in an AMPK-independent fashion [59]. It therefore remains to be elucidated whether the AMPK-dependent anti-apoptotic action of APN is dependent on cell types or different forms of APN.

It has been established that H/SD induces MSCs apoptosis through the mitochondrial-related intrinsic pathway, leading to the accumulation of Bax, loss in mitochondrial membrane potential (MMP) and activation of caspase-3 [10]. The balance of Bax and Bcl-2 expressions play a key role in maintaining the integrity of mitochondrial and suppressing the mitochondrial apoptotic pathway [60,61]. The decrease of Bcl-2 [62] and increase of Bax [63,64] lead to loss of MMP and release of cytochrome c from the mitochondria to the cytoplasm, which activates the caspase-3 effector and leads to apoptotic cell death [37,65]. Importantly, APN was found to reverse neuroblastoma cell apoptosis via regulation of Bcl-2 and Bax expression [61,66] and attenuate the mitochondrial dysfunction in diabetic heart [54].

Mitochondria in undifferentiated MSCs were maintained at a relatively low activity status [44]. Nevertheless, it was found to be highly sensitive to external environment such as changes in the energy status, oxidative stress and apoptotic stimuli [67]. Under H/SD conditions, MSCs exhibited mitochondrial dysfunction with MMP loss [10]. This was confirmed by other studies showing that the mitochondrial-dependent pathway exerted a vital function in other molecules and medications protecting MSCs from apoptosis induced by H/SD, including CTRP3 [41], lysophosphatidic acid [68], angiopoietin-1 [69], lovastatin [70], Tongxinluo [71] and Exendin-4 [72].

Our data showed that Bcl-2 was up-regulated and Bax was down-regulated in proportion to the dose of gAPN and AMPK phosphorylation. On the ground that the ratio between Bcl-2 and Bax proteins plays a pivotal role in the apoptotic cascade, it is not surprising that treatment with gAPN was also associated with the prevention of the downstream loss of MMP, finally preventing caspase-3 activation and apoptotic signaling pathways. These results indicated that cytoprotection of gAPN was mediated through the inhibition of the mitochondrial apoptotic pathway. Our results are in accordance with the protective effects of APN on mitochondrial depolarization and apoptosis of mesoangioblasts induced by growth factors withdrawal [24].

Our study reveals that gAPN protects MSCs against the mitochondrial-related apoptosis under H/SD conditions, which is mainly mediated through the AdipoR1 linked to the downstream AMPK signaling pathways. These findings suggest that gAPN may be a novel survival factor for MSCs in the ischemic engraftment environment and further studies in vivo are needed to verify that gAPN might enhance the therapeutic efficacy of MSCs transplantation in ischemic heart diseases.

MSCs (Mesenchymal stem cells); BM-MNCs (Bone marrow mononuclear cells); HSCs (Hemopoietic stem cells); gAPN (globular adiponectin); AMPK (AMP-activated protein kinase); H/SD (Hypoxia and serum deprivation); AdipoR1 (Adiponectin receptor 1); AdipoR2 (Adiponectin receptor 2); MMP (Mitochondrial membrane potential).

This work was supported by grants from the National Natural Science Foundation of China (81170129,81200107,81300157,81200108),863 Program of China (2011AA020110, 2013AA020101), 973 Program of China (2012CB518602) and the Ph.D Programs Foundation of Ministry of Education of China (20111106110012).

The authors have no potential conflict of interests to declare.