To investigate the effect of HDL on the biological activity of MSCs, we performed flow cytometry and MTS to detect cell differentiation and cell viability respectively. We observed that the MSCs (Passage 4), used in our experiments, expressed typical MSC-related cell surface antigens, which were positive for CD29, CD90 and negative for CD45, CD34; and that the incubation with HDL (100 μg/mL) for 24 h did not exhibit a notably impact on the characteristic phenotypes of MSCs (Figure S1). Meanwhile, we discovered that MSCs viability was not significantly affected by treatment with any of the tested concentrations of HDL (Figure 1A). Therefore, we chose a HDL concentration of 100 μg/mL for subsequent experiments based on methods described in previously published studies [9].

To determine the concentration of H₂O₂, which was employed to induce the cell apoptosis, MSCs was cultured with different concentration of H₂O₂ for 24 h and a MTS assay was performed afterwards. It was found that H₂O₂ impaired the viability of MSCs in a concentration-dependent manner over the tested concentration range (Figure 1B). The effect was statistically significant at concentrations over 200 μM. Based on these findings, 400 μM H₂O₂ was selected for use in the subsequent experiments.

In order to investigate the effects of HDL on H₂O₂-stimulated MSC apoptosis, MSCs were pretreated with or without HDL (100 μg/mL) for 24 h and then cultured with 400 μM of H₂O₂ for an additional 24 h. TUNEL assay showed that incubation with 100 μg/mL HDL did not affect the apoptosis of MSCs (HDL group vs. Control group: (6.83 ± 1.35)% vs. (7.93 ± 1.76)%, p > 0.05). However, H₂O₂ significantly increased MSCs apoptosis (H₂O₂ group vs. Control group: (47.37 ± 7.53)% vs. (7.93 ± 1.76)%, p < 0.01), which was remarkably attenuated after preconditioning with HDL (HDL + H₂O₂ group vs. H₂O₂ group: (28.77 ± 6.91)% vs. (47.37 ± 7.53)%, p < 0.05) (Figure 1C). Caspase-3 activity was significantly increased in H₂O₂-stimulated MSCs compared to controls ((267.4 ± 25.3)% vs. (100 ± 12.2)%, p < 0.01). The increased caspase-3 activity was ameliorated by pre-incubation with HDL (HDL + H₂O₂ group vs. H₂O₂ group: (130.9 ± 19.7)% vs. (267.4 ± 25.3)%, p < 0.05) (Figure 1D). Similarly, compared to H₂O₂ group, HDL pretreatment restored cell viability following exposure to H₂O₂ (HDL + H₂O₂ group vs. H₂O₂ group: (85.3 ± 7.2)% vs. (58.6 ± 6.8)%, p < 0.05) (Figure 1E).

Meanwhile, MSCs aging or senescence, which directly impaired the regenerative capability, was widely reported to be induced by oxidative stress [16,17]. We, therefore, wondered whether HDL treatment could protect MSCs against H₂O₂-stimulated replicative exhaustion. Senescence-associated β-Galactosidase (SA-β-Gal) staining showed that there were no significant differences in the percentages of SA-β-Gal positive cells between the HDL group and the Control group ((102.3 ± 13.7)% vs. (100 ± 12.4)%, p > 0.05). Similarly, there was nearly identical in those between the H₂O₂ group and the HDL+H₂O₂ group ((356.2 ± 25.8)% vs. (367.9 ± 22.37)%, p > 0.05), although there was a significant increment compared the H₂O₂ group with the Control one (p < 0.05) (Figure 2A). Furthermore, the MSCs senescence was confirmed by Western blot assay, suggesting that there were no remarkably differences in the expression of p16^INK4a between the HDL group and the Control ((1.06 ± 0.13) vs. (1.00 ± 0.10), p > 0.05), and between the H₂O₂ group and the HDL + H₂O₂ group ((3.62 ± 0.19) vs. (3.40 ± 0.26), p > 0.05). Likewise, it was found that the expression of p16^INK4a in the H₂O₂ group statistically increased than that in the Control one (p < 0.05) (Figure 2B). Therefore, it indicated that HDL exerted no statistical effects on the senescence of MSCs both under a condition of oxidative stress and under that of control.

As previously reported [18], PI3K/Akt pathway played a critical role in the cell apoptosis. To examine our hypothesis that PI3K/Akt pathway was involved in the protective effect of HDL on MSCs, Western blot analysis was used to assess the changes in Akt phosphorylation in MSCs exposed to HDL with different concentration. The results indicated that incubation with HDL significantly up-regulated Akt phosphorylation. This response began 15 min after exposure, peaked at approximately 30 min (30-min group vs. Control group: (283.8 ± 25.1)% vs. (100 ± 24.71)%, p < 0.05) and was still present at 24 h (24-h group vs. Control group: (197.5 ± 19.37)% vs. (100 ± 24.71)%, p < 0.05) (Figure 3A), suggesting that HDL could enhance the activation of PI3K/Akt pathway in MSCs. To provide further evidence, a PI3K inhibitor, LY294002, was effectively employed to decrease Akt activation (LY294002 group vs. Control group: (12.37 ± 5.19)% vs. (100 ± 24.28)%, p < 0.05) (Figure 3B). Considering that high concentrations of LY294002 could cause a reduction in cell viability or a generation of ROS, a LY294002 group was added to examine its toxicity to MSCs. It was observed that LY294002, at a concentration of 25 μM, did not exhibit a statistical effect on the MSCs viability (LY294002 group vs. Control group: (97.6 ± 6.8)% vs. (100 ± 5.1)%, p > 0.05) (Figure 3C) and apoptosis (LY294002 group vs. Control group: (106.4 ± 13.5)% vs. (100 ± 11.3)%, p > 0.05) (Figure 3D), which was consistent with what was testified previously [19]. Intriguingly, the pre-treatment of LY294002 at this concentration was found to abolish the protective effect of HDL on H₂O₂-stimulated cells. This resulted in decreased cell viability as measured by MTS assay (HDL + H₂O₂ + LY294002 group vs. HDL + H₂O₂ + DMSO group: (51.7 ± 7.1)% vs. (83.6 ± 7.8)%, p < 0.05) (Figure 3C) and increased apoptosis estimated by changes in caspase-3 activity (HDL + H₂O₂ + LY294002 group vs. HDL + H₂O₂ + DMSO group: (286.3 ± 16.7)% vs. (149.5 ± 21.8)%, p < 0.05) (Figure 3D), consistently indicating that HDL protected MSCs against injury induced by H₂O₂ through PI3K/Akt pathway.

The influence of the PI3K/Akt pathway in ROS production based on LY294002 inhibition and assessed using fluorescent probe-DHE demonstrated that the percentage of ROS negative cells with low fluorescence intensity were similarly observed among the Control, the HDL and the LY294002 groups (all p > 0.05) (Figure 4). Cells with strong red fluorescent, indicating ROS accumulation, were more prevalent in the H₂O₂ group than in the Control one ((435 ± 48)% vs. (100 ± 15)%, p < 0.05). HDL pretreatment restored intracellular ROS generation induced by H₂O₂ (HDL + H₂O₂ group vs. H₂O₂ group: (156 ± 25)% vs. (435 ± 48)%, p < 0.05), and the protective effect disappeared when the PI3K/Akt pathway was inhibited by LY294002 (HDL + H₂O₂ + DMSO group vs. HDL + H₂O₂ + LY294002 group: (172 ± 37)% vs. (482 ± 57)%, p < 0.01).

The above results showed that the inhibition of PI3K/Akt pathway attenuated the protective effect of HDL and the decreased ROS, indicating that Akt activation was partly upstream of ROS generation. Excessive production of ROS has been shown to damage stem cells through caspase-mediated pathways [20]. Therefore, these findings suggested that HDL activation of the PI3K/Akt pathway might inhibit ROS generation, and thereby reduce apoptosis of implanted MSCs.

The poor viability of the donor cell, typically <1% at 4 days after transplantation, has been reported in the previous study regarding MSCs-base therapy [5]. Therefore, we testified the effect of HDL preconditioning on the survival of MSCs on Day 4 after transplantation in vivo (Figure 5A,B). After a GFP-labeling and the pre-incubation with or without HDL (100 μg/mL), MSCs (1 × 10⁶ in 100 μL) were respectively transplanted into the infarcted heart of rats. It was discovered that the number of GFP⁺ MSCs per HP in the HDL-MSCs group was 2.58-fold greater than that in the MSCs group (p < 0.05). Furthermore, when we transplanted the male MSCs to the female hearts, it was revealed that the sry DNA level in the HDL-MSCs group was approximately 2.83-fold higher than that in the MSCs group (p < 0.05), indicating that HDL preconditioning could enhance the survival of MSCs after transplantation.

The effect of MSCs preconditioned with HDL on cardiac function was observed by transthoracic echocardiography four weeks after experimentally induced MI in rats (Figure 5C,D). It indicated that left ventricular ejection fraction (LVEF) was 50.8% lower in the PBS group than in the Sham group and LVFS was 53.0% lower in the PBS group than in the Sham group (both p < 0.05); and that LVIDd was 40.9% higher and LVIDs was 90.0% higher in the PBS than in the Sham group (both p < 0.05). Furthermore, compared with PBS group, transplantation of MSCs improved LVEF and LVFS by 13.0% and 21.6%, respectively (both p < 0.05). Injection of HDL-MSCs enhanced the LVEF and LVFS by an additional 13.0% and 20.1%, respectively (both p < 0.05). Transplantation of MSCs decreased the LVIDd and LVIDs by 19.5% and 23.3%, respectively compared to the PBS group (both p < 0.05), and transplantation of HDL-MSCs further decreased LVIDd and LVIDs by approximately 12.3% and 30.0% (both p < 0.05), respectively.

We noticed that Kiya Y et al.[21] had previously reported that continual infusion of HDL following MI prevented LV remodeling and improved myocardial function in rats. Though they attributed the better prognosis to an improvement of the survival of cardiomyocytes resulting from treatment with HDL, they had not well explored the influence of HDL on the cells that migrated to injured myocardium, such as MSCs. And our results from the present study might partly provided the evidences on a novel perspective to support their finding as follows: a higher level of HDL might improve the anti-apoptotic capability of MSCs, which were mobilized to the damaged myocardial region, thus contributing to the myocardial regeneration, bringing about a better cardiac function after MI.

In the recent literature, the effect of transplantation of MSCs on myocardial repair was reported as confined for the low viability of graft cells [1,22]. For instance, even 6 × 10⁷ implanted MSCs in infarcted porcine hearts yielded only limited enhancement of LVEF [23]. A series of pharmacological preconditioning was thought to be an effective approach for promoting MSCs survival [24] but had the potential to cause severe side effects in patients with cardiovascular diseases.

Therefore, we explored whether improving the in-situ physiological environment of MSCs could enhance cells viability and prevent oxidative stress-induced apoptosis after transplantation or mobilization. Epidemiological studies have consistently shown an inverse correlation between HDL levels and outcomes in ischemic heart diseases [8]. We endeavored to elucidate this correlation with a hypothesis that HDL might promote survival of implanted MSCs, ameliorate their capacity to repair damaged tissue and thereby improve clinical prognosis. In support of this hypothesis, our data indicated that HDL protected MSCs from apoptosis induced by H₂O₂in vitro. The mechanisms involved activation of PI3K/Akt pathway, which was consistent with findings from previously reported studies [15,22]. We also demonstrated that the anti-apoptotic effect of HDL enhanced survival of MSCs and improved myocardial repair in vivo, supporting the findings of previous researches showing that the administration of HDL preserved the post-ischemic cardiac function [25].

Beside apoptosis, cellular senescence had been proved to be the other main outcome following oxidative stress stimuli [16], which considerably impaired MSCs therapeutical properties. On the other hand, there were several literature suggesting that a higher level of HDL was associated with a slower rate of leukocyte telomere length shortening per year [26]; and that offspring of individuals with exceptional longevity had significantly larger HDL, which hinted that HDL was probably involved in the cellular aging [27]. Considering those evidences, we carried out the parallel experiments to investigate whether HDL could play a critical role in the protective effect on MSCs senescence stimulated by oxidative stress. H₂O₂ treatment expectably spurred a considerable replicative exhaustion in MSCs; however, MSCs senescence could not be reversed by HDL preconditioning, which indicated that the effectiveness of HDL on the MSCs therapeutical potential in vivo was independent on the influence in cellular aging, while mainly through enhancement in cellular anti-apoptotic property. Indeed, the elucidation of this mechanism is especially critical for physicians engaged in cell-based therapy, whom must be sure that MSCs are devoid of chromosome alterations and that there are no signs of cellular senescence.

The clinical significance of our results is the possible benefit of HDL in patients with myocardial infarction in terms of improving the survival of transplanted MSCs. Elevation of HDL levels in patients, at high risk of cardiovascular disease may also be of vital clinical importance in preventing cardiovascular events. However, we did not identify the specific components of HDL that were responsible for its anti-oxidant and anti-apoptotic effects in the present study, for HDL is a lipoprotein, containing HDL2 and HDL3, with a density ranged from 1.063 to 1.210 g/mL [28]. This will be the subject of future research, along with an investigation of other potential effects of HDL on MSCs (such as promotion of secretion); that may contribute to the restoration of cardiac function in vivo.

All animal procedures were undertaken in accordance with the Guidelines for the Care and Use of Laboratory Animals, published by the National Academy Press (NIH Publication No. 85–23, Revised 1996). The experiments were approved by the Animal Care and Use Committee of Zhongshan Hospital, Fudan University.

MSCs were isolated from four-week-old Sprague-Dawley (SD) rats weighing 100 to 120 g by flushing femurs and tibias with Dulbecco’s modified Eagle’s medium (DMEM; GIBCO), 10% Fetal Bovine Serum (FBS; GIBCO) and penicillin/streptomycin. The bone marrow (BM) solution was cultured at 37 °C in an atmosphere of 5% CO₂. The culture medium was changed every 3–4 days. Non-adherent hematopoietic cells were removed during this process, and the adherent, spindle-shaped MSCs were cultured and expanded. Passage 4 cells were used for subsequent experiments.

An adenoviral vector carrying a green fluorescent protein (GFP) (GeneChem) for cell transduction was used to establish and identify a stable population of MSCs. Transduction was facilitated by transfecting Passage 4 MSCs with purified adenovirus cells that had been exposed to 7 μg/mL polybrene. The transfected cells were incubated for 24 h at 37 °C with 5% CO₂. The medium was removed, replaced with fresh medium and incubated for a further 24 h. Cells were then sorted by flow cytometry for GFP to ensure that a homogenous population of GFP⁺ MSCs was obtained (data were not shown).

MSCs were incubated for 6 h with DMEM supplemented with 1% FBS prior to experimental interventions. The starved MSCs were then respectively exposed to H₂O₂ (0 to 500 μM) or HDL (0 to 200 μg/mL; Merck) for 24 h. The experiments were undertaken in four groups: a control group; a HDL group (MSCs incubated with HDL 100 μg/mL for 24 h); a H₂O₂ group (incubation with 400 μM H₂O₂ for 24 h); a HDL + H₂O₂ group (pre-incubation with HDL 100 μg/mL for 24 h, followed by 24-h exposure to 400 μM H₂O₂). For inhibitor experiments, three additional groups were set as follows: MSCs were pre-incubated with PI3K inhibitor (LY294002, 25 μM) or vehicle (DMSO) for 1 h, and then successively treated with HDL and H₂O₂ as above; the third group was pre-cultured with LY294002 alone. The concentrations of HDL, H₂O₂ and LY294002 used in these experiments were determined from pilot studies and based on similar methods described in the literature [9,19,29].

MSC viability was assessed using a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2- (4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) assay. Cells were cultured in 24-well plates and treated according to the protocol. After 24 h, Cell Titer 96^® Aqueous One Solution Reagent (Promega) was added to each well according to the manufacturer’s instructions. Cell viability was evaluated after 2-h incubation with this reagent, by testing absorbance at 490 nm using a plate reader.

Cell apoptosis was investigated using an In Situ Cell Death Detection kit (Roche). In brief, MSCs were fixed in 3.7% buffered formaldehyde, pretreated with 3.0% H₂O₂ and exposed to TdT enzyme at 37 °C for 1 h. The cells were then incubated with digoxigenin-conjugated nucleotide substrate at 37 °C for 30 min. Nuclei exhibiting DNA fragmentation were identified after staining with 3,3-diamino benzidine (DAB) for 5 min.

The MSCs were counterstained with hematoxylin and viewed by light microscopy. Five fields at high magnification (×200) were randomly chosen to assess the apoptosis index, which was calculated as a percentage of labeled cells to total nuclei.

Caspase-3 activity was determined in MSC lysates (200 μg) using Caspase-3/CPP32 Fluorometric Assay kit (BioVision). The fluorogenic CPP32/caspase-3 substrate was labeled with the fluorochrome 7-amino-4-methyl coumarin. In this assay, the amount of fluorescence produced upon cleavage was proportional to the amount of caspase-3 activity present in the sample.

MSCs cultured in six-well plates were rinsed with phosphate-buffered saline (PBS), fixed and then incubated with freshly prepared senescence-associated β-Galactosidase (SA-β-Gal) staining solution (Beyotime, China) at 37 °C overnight. Five fields at high magnification (×200) were randomly chosen to count the number of blue cells (SA-β-Gal positive cells). At least 500 cells in each group were examined and the percentages of blue cells presented.

MSCs from each group were rinsed with cold PBS and lysed using RIPA buffer containing 1 mM phenylmethanesulfonylfluoride on ice for 30 min. The lysates were transferred to 1.5 mL Eppendorf tubes and centrifuged at 4 °C and 15,000 rpm for 30 min. The protein component was separated by 12% SDS-PAGE and then transferred to polyvinylidene fluoride membranes (Millipore). The membranes were blocked in 5% skimmed milk for 2 h at room temperature, and were incubated overnight at 4 °C with anti-phospho-Akt (Cell Signaling), anti-Akt (Cell Signaling), anti-p16^INK4a. Following a 30 min wash, the membranes were incubated for 1 h with a secondary antibody conjugated to horseradish peroxidase-conjugated IgG (Jackson). Protein expression was determined using an enhanced chemiluminescence system and quantified by densitometry (Image System; Bio-Rad).

The production of intracellular ROS was measured using an oxidation-sensitive fluorescent dihydroethidium (DHE) probe (Vigorous). The assay was based on the principle that DHE crosses cell membranes and is rapidly oxidized in the presence of ROS resulting in the formation of a highly fluorescent form of oxidative ethidium.

In this assay MSCs were washed three-times with PBS and incubated with 10 μM DHE in phenol-red-free MEM medium (Invitrogen) at 37 °C in the dark. After 15 min incubation, the cells were rinsed and the medium was replaced. The fluorescence level, as an indicator of ROS production, was detected using fluorescence microscopy. The excitation and emission filters were 488 nm and 590 nm, respectively.

The MI model was developed in SD rats (220 ± 30 g), as previously described [22]. MSCs (1 × 10⁶ in 100 μL, derived from male rats) (MSCs group) or MSCs preconditioned with HDL (HDL-MSCs group) were intramyocardially injected into female rats 10 min after ligation-induced infarction. The MSCs were injected into 3 to 5 injection sites into the anterior aspect of the viable myocardium that bordered the area of infarction. For the experiments concerning cardiac remodeling, two additional groups were set as follows: rats received intramyocardial injection of 100 μL of PBS without MSCs (PBS group) or received thoracotomy only but were not ligated (Sham group).

To observe the donor MSCs survival in vivo, rats were sacrificed four days after cell implantation. Hearts were harvested and sliced transversally above the ligation suture. Tissue fixation was minimized to reduce autofluorescence. The sample was embedded at optimum cutting temperature and cut by a cryostat from the base (above the ligation) towards the apex until the ligation suture was reached. Beginning at this point, 10 μm sections were collected throughout the entire lesion. The sections were stained using DAPI (4,6-diamidino-2-phenylindole), and the number of GFP⁺ MSCs was estimated under fluorescent microscopy at a magnification of ×200.

Three photomicrographs were taken from the infarct border zone of each section. Six sections were selected for counting, giving a total of 18 high power fields per rat. The numbers of GFP⁺ MSCs were counted and averaged.

Real-time polymerase chain reaction (PCR) was performed for quantification of the male sry DNA levels on Day 4 after cell transplantation. Rats were sacrificed and hearts were excised, frozen in liquid nitrogen and powdered. DNA purification was performed using Genomic DNA Isolation kit (Qiagen). The concentration of the purified DNA was determined by spectrophotometry. Real-time PCR was performed using Takara SYBR Premix Ex TaqTM in a Bio-Rad iQ5 optical module. Primers for amplification of rat Y chromosome sry and β-actin genes were synthesized (Invitrogen) and were listed as follows: sry, forward, 5′-GAG GCA CAA GTT GGC TCA ACA-3′; reverse, 5′-CTC CTG CAA AAA GGG CCT TT-3′. β-actin, forward, 5′-CCA TTG AAC ACG GCA TTG-3′; reverse, 5′-TAC GAC CAG AGG CAT ACA-3′. The cycling conditions were: 30 s at 95 °C, followed by 40 cycles of 5 s at 95 °C, 30 s at 59.5 °C, 30 s at 72 °C. Melting curves were acquired at the end of the reaction by gradually raising the temperature by 1 °C/min from 59.5 °C to 95 °C over a time period of 35 min.

Cardiac remodeling and left ventricular function were assessed by transthoracic echocardiography 4 weeks after MI, using a Vevo 770 high-resolution imaging system (Visual Sonics) with a 17.5-MHz probe. After the induction of light general anesthesia, hearts were imaged in two dimensional and M-modes. The recordings were obtained in the para-sternal long-axis view at the level of the greatest LV diameter. The LV internal end-diastolic diameter (LVIDd) and LV internal end-systolic diameter (LVIDs) were measured from M-mode recordings according to the leading-edge method. All echocardiographic measurements were averaged from at least five separate cardiac cycles.

Statistical analysis was undertaken using SPSS version 16.0. All data were expressed as means and standard errors (SE). Between-group of differences was analyzed using Student’s t-test. Comparisons between more than two groups were performed using one-way analysis of variance (ANOVA) with Bonferroni’s correction. Values of p < 0.05 were considered statistically significant.