Background

Previous studies have reported the cardioprotective effect of adipose-derived stem cells (ASCs) [1, 2]. Evidence suggests that the regenerative effect of transplanted stem cells is mainly mediated by paracrine factors [3, 4], and stem cell-released factors varied depending on the source of stem cells and the stimuli they were exposed to [5, 6]. Therefore, it is important to identify soluble factors contributing to the beneficial effect (i.e., host cell protection) of stem cells under a specified condition because a clinically effective stem cell-based therapeutic strategy can be developed based on such information. Osteoprotegerin (OPG) is known to be produced by stem cells, fibroblasts, and endothelial cells [7]. OPG is a soluble decoy receptor that binds to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), thereby neutralizing TRAIL-mediated apoptotic signaling [8]. TRAIL initiates signaling cascade by binding to corresponding receptors such as TRAIL-R1, TRAIL-R2, TRAIL-R3, and TRAIL-R4 [9]. TRAIL-R1 and TRAIL-R2 can promote cell death signaling cascades, while TRAIL-R3 and TRAIL-R4 antagonize the TRAIL-R1 and TRAIL-R2-mediated apoptotic signaling by competitively binding to TRAIL [10]. In the present study, we examined the effect of OPG on the survival of cardiomyocytes exposed to oxidative stress.

Methods

Culture of ASCs and H9c2 cells

Four vials of ASCs from four different human donors (StemPro human adipose-derived stem cells) were purchased from Invitrogen (Carlsbad, CA, USA), and rat cardiomyoblast cell line H9c2 cells were purchased from ATCC (Manassas, VA, USA). ASCs and H9c2 cells were maintained according to the manufacturer’s instructions. We used low-glucose Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Waltham, MA, USA) and high-glucose DMEM (Gibco) containing 10% fetal bovine serum (FBS; Gibco) and 1% penicillin–streptomycin (Gibco) for ASCs and H9c2 cells, respectively.

Isolation rat neonatal ventricular cardiomyocytes

All experimental procedures for animal studies were approved by the committee for the care and use of laboratory animals of Catholic Kwandong University, and were performed in accordance with the committee’s guidelines and regulations (CKU01-2015-003-1). Neonatal rat ventricular cardiomyocytes were isolated and purified by enzymatic methods following a protocol published previously [11].

Ischemia–reperfusion injury and transplantation of ASCs

Male 8-week-old Sprague–Dawley rats (250 g) were used. The animals were put under anesthesia with zoletil (30 mg/kg) and xylazine (10 mg/kg). The left anterior descending coronary artery was occluded using a 6-0 silk suture (Johnson & Johnson, New Brunswick, NJ, USA). After 60 min of occlusion, the left anterior descending coronary artery was released for reperfusion, followed by cell transplantation. For transplantation, 1 × 106 cells were suspended in 30 μl PBS and injected to border zone using a 0.3-ml insulin syringe (BD Pharmingen, Franklin Lakes, NJ, USA).

Statistical analyses

Quantitative data were expressed as the means ± SD of at least three independent experiments. For statistical analysis, one-way ANOVA with Bonferroni correction was performed using OriginPro 8 SR4 software (version 8.0951; OriginLab Corporation, Northampton, MA, USA) if there were more than three groups. p < 0.05 was considered statistically significant.

Additional methods are available in Additional file 1.

Results

Hydrogen peroxide induced apoptosis of cardiomyocytes

Hydrogen peroxide (H2O2), a member of endogenous reactive oxygen species (ROS), has been used to simulate oxidative stress [12, 13]. When H9c2 and primary cultured rat cardiomyocytes were treated with increasing concentration of H2O2 for 48 hours, the viability of both H9c2 (Additional file 2: Figure S1A) and primary (Additional file 2: Figure S1B) cardiomyocytes significantly decreased with H2O2 at a concentration of 100 μM or higher. However, primary cardiomyocytes were more resistant to ROS than H9c2 cardiomyocytes. Therefore, we mainly used H9c2 cardiomyocytes, which have been used as an alternative to primary cultured cardiomyocytes [14], to clearly demonstrate the prosurvival effect of OPG.

To confirm that the H2O2-induced cell death was oxidative stress mediated, the cells were pretreated with the well-established antioxidant N-acetyl-l-cystein (NAC) [15]. NAC pretreatment attenuated H2O2-induced activation of caspase 3 and 8 (Additional file 2: Figure S1C) and also significantly attenuated H2O2-induced decrease of H9c2 viability (Additional file 2: Figure S1D). Furthermore, NAC pretreatment decreased the number of apoptotic propidium iodide (PI)/Annexin V double-stained cells [16] (Additional file 2: Figure S1E). Interestingly, H2O2 concentration up to 200 μM did not decrease the viability of ASCs (Additional file 3: Figure S2).

Oxidative stress increased OPG production from ASCs

Both H2O2 and TGF-β1, another soluble factor commonly present in damaged heart [17], significantly increased the OPG production of ASCs without decreasing viability (Fig. 1a). Also, H2O2 significantly increased mRNA expression of OPG and TRAIL receptor 3 (TRAIL-R3) (Additional file 4: Figure S3), which attenuates TRAIL-induced apoptosis [18]. In the ischemia–reperfusion (I/R)-injured heart without ASC transplantation, OPG expression was marginal only at day 1 after I/R injury (Fig. 1b). However, in the ASC transplantation group, prominent OPG expression was observed in the vicinity of transplanted ASCs, which gradually decreased with time (Fig. 1c).

Fig. 1
figure 1

Oxidative stress induced OPG secretion from ACSs. (a) Amount of OPG secreted from ASCs exposed to H2O2 (100 μM) determined using an ELISA kit specific to human OPG. Conditioned media of ASCs exposed to H2O2 up to 48 hours were used for the analysis. TGF-β1 (5 ng/ml) was included as a soluble factor commonly present in damaged heart. Inset: viability of ASCs treated with H2O2 (100 μM) or TGF-β1 (5 ng/ml) for 48 hours. *p < 0.05 compared to untreated control. (b) Expression of OPG in ischemia–reperfusion (I/R)-injured rat heart examined by immunofluorescence staining. I/R-injured rat heart was collected 1, 3, and 7 days after the injury without ASC transplantation, and damaged heart was stained with antibodies specific to CD90 (red), a stem cell marker, and OPG (green). Nucleus was stained with DAPI (blue). (c) Expression of OPG in I/R-injured heart with ASC transplantation (1 × 106 cells/head). Two-dimensional immunofluorescence images were converted to 2.5D topological view images for clear presentation of staining intensity among groups. ASC adipose-derived stem cell, OPG osteoprotegerin, TGF-β1 transforming growth factor beta 1 (Color figure online)

Full size image

OPG protected cardiomyocytes from ROS-induced cell death

To examine the effect of OPG on cardiomyocyte viability under oxidative stress, the cells were exposed to H2O2 with or without OPG (1.5 and 3 ng/ml) treatment. OPG treatment significantly protected from H2O2 (Fig. 2a) and suppressed activation of caspase 8 without affecting the expression of TRAIL-R2 (Fig. 2b). Both ASC conditioned media with or without 48 hours of H2O2 (100 μM) conditioning (designated as H2O2CM and NorCM, respectively) significantly attenuated H2O2-induced decrease of viability in both H9c2 (Fig. 2c) and primary (Fig. 2d) cardiomyocytes. Between NorCM and H2O2CM, H2O2CM was more effective to prevent H2O2-induced cardiomyocyte death.

Fig. 2
figure 2

OPG protected cardiomyocytes from ROS-induced cell death. (a) Effect of OPG on H2O2-induced cell death of H9c2 examined using a cell counting kit. H9c2 cells were treated with 100 μM of H2O2 for 24 hours with or without OPG (1.5 and 3 ng/ml) treatment. *p < 0.05, #p < 0.05 compared to H2O2-treated group. (b) Expression of proapoptotic receptor TRAIL-R2 and activated (cleaved) caspase 8 in H9c2 cardiomyocytes exposed to H2O2 (100 μM) for 24 hours with or without OPG treatment (1.5 and 3 ng/ml) examined by western blot analysis. (c) Effect of ASC conditioned media with or without H2O2 conditioning on the viability of H9c2 cardiomyocytes exposed to H2O2. Normal conditioned media of ASC (NorCM) prepared by culturing ASCs for 48 hours and conditioned media with H2O2 conditioning (H2O2CM) prepared by culturing ASCs in the presence of 100 μM of H2O2 for 48 hours. H9c2 cells were cultured in NorCM or H2O2CM with or without addition of 100 μM of H2O2 for 24 hours. *p < 0.05, #p < 0.05 compared to H2O2-treated group. (d) Effect of ASC conditioned media with or without H2O2 conditioning on the viability of primary cardiomyocytes exposed to H2O2. The same experimental procedure used for H9c2 cardiomyocytes was applied to primary cardiomyocytes. *p < 0.05, #p < 0.05 compared to H2O2-treated group. OPG osteoprotegerin, TRAIL-R2 tumor necrosis factor-related apoptosis-inducing ligand receptor 2

Full size image

Downregulation of OPG offset cell protective effect of ASC conditioned media

To validate whether the cell protective effect of the ASC conditioned media was OPG mediated, siRNA specific to OPG was utilized. Delivery of OPG-specific siRNA to ASCs significantly attenuated OPG mRNA expression (Fig. 3a) and decreased OPG production (Fig. 3b). OPG-specific siRNA had no negative effect on the cell viability of ASCs (Fig. 3c). However, when siRNA specific to TRAIL-R3 was codelivered with OPG-specific siRNA, the viability of ASCs significantly decreased (Additional file 5: Figure S4). Finally, H2O2CM obtained from OPG-specific siRNA-treated ASCs showed decreased protective effect on cardiomyocytes exposed to H2O2 (Fig. 3d).

Fig. 3
figure 3

Downregulation of OPG offset protective effect of ASC conditioned media. (a) OPG-siRNA-mediated downregulation of OPG in H2O2-treated ASCs. ASCs were transfected with 20 nM of OPG-specific siRNA for 24 hours, and then exposed to 100 μM of H2O2 for an additional 48 hours. mRNA expression of OPG normalized by 18 s rRNA expression. *p < 0.05, #p < 0.05 compared to H2O2-treated group. (b) Effect of OPG-specific siRNA on the OPG production from ASCs exposed to H2O2. OPG-siRNA (20 nM)-transfected ASCs were cultured in the presence of 100 μM of H2O2 for 48 hours, and the amount of OPG in ASC culture media was determined using an OPG-specific ELISA kit. *p < 0.05 compared to untreated control, #p < 0.05 compared to H2O2-treated group. (c) Viability of OPG-siRNA-transfected ASCs exposed to 100 μM of H2O2. *p < 0.05 compared to H2O2-treated group. (d) Effect of three different types of H2O2CMs on H2O2-induced cell death of H9c2 cells. *p < 0.05. H2O2CM ASC conditioned media prepared by culturing ASCs with 100 μM of H2O2 for 48 hours, H2O2CM/Reagent conditioned media obtained using transfection reagent-treated ASCs, H2O2CM/siRNA conditioned media obtained using OPG-specific siRNA-transfected ASCs, OPG osteoprotegerin

Full size image

Discussion

It has been reported that TRAIL was released from the postischemic heart soon after the onset of reperfusion [19]. Considering ROS production, a major cause of myocardial cell death in I/R injury [20], is an early event in myocardial reperfusion injury, TRAIL released at the onset of reperfusion may by associated with increased ROS and subsequent myocardial cell death in the I/R-injured heart. Therefore, we hypothesized that ASCs transplanted into damaged heart (i.e., I/R-injured heart where excessive oxidative stress causes cardiac cell death [21,22,23]) secrete OPG, and in turn the ASC-released OPG prevents host cardiomyocyte death by antagonizing TRAIL signaling, which has never been tested.

To test the hypothesis in vitro, we used 100 μM of H2O2 to simulate oxidative stress to which both host cardiomyocytes and transplanted ASCs are exposed in vivo. In our study, H2O2 induced cardiomyocyte death via oxidative stress-mediated activation of apoptosis, confirming the validity of our experimental setting (Additional file 2: Figure S1). When ASCs were exposed to H2O2 or transplanted to I/R-injured heart, the production of OPG was significantly increased (Fig. 1), suggesting that ASCs exposed to a hostile microenvironment, such as I/R-injured heart, can produce OPG for the first few days following transplantation. Further experiments showed that OPG can suppress ROS-induced apoptosis of cardiomyocytes (Fig. 2a, b) and demonstrated that H2O2 conditioning of ASCs enhanced the protective effect of conditioned media possibly by increasing OPG production from ASCs (Fig. 2c, d). Experiments using OPG-specific siRNA indicated that the cell protective effect of ASC conditioned media was indeed OPG mediated (Fig. 3).

One interesting observation was that ASCs were less sensitive to H2O2 compared to cardiomyocytes (Additional file 3: Figure S2). This might have something to do with the increased OPG and TRAIL-R3 expression under oxidative stress (Additional file 4: Figure S3) because cotransfection of siRNAs specific to OPG and TRAIL-R3 significantly decreased the viability of ASCs exposed to H2O2 (Additional file 5: Figure S4). This also suggested that OPG and TRAIL-R3 may compensate each other in terms of suppressing TRAIL-mediated cell death in ASCs. Nevertheless, without any further experimental evidence, this can only be speculated at this point. We are currently working to elucidate more detailed underlying mechanisms of OPG-mediated cardiomyocyte protection and to validate and confirm the findings of this study in vivo for further studies.

Conclusion

In the present study we provided evidence that OPG is an ASC-released soluble factor that can protect cardiomyocytes from ROS-induced apoptosis for the first time. The results of this study warrant further studies to fully elucidate the exact nature of stem cell-released paracrine factors and may help us to design an effective stem cell-based therapeutic strategy in the future.