Role of dopamine D2 receptors in ischemia/reperfusion induced apoptosis of cultured neonatal rat cardiomyocytes.

BACKGROUND: Myocardial ischemia/reperfusion injury is the major cause of morbidity and mortality for cardiovascular diseases. Dopamine D2 receptors are expressed in cardiac tissues. However, the roles of dopamine D2 receptors in myocardial ischemia/reperfusion injury and cardiomyocyte apoptosis are unclear. Here we investigated the effects of both dopamine D2 receptors agonist (bromocriptine) and antagonist (haloperidol) on apoptosis of cultured neonatal rat ventricular myocytes induced by ischemia/reperfusion injury.
METHODS: Myocardial ischemia/reperfusion injury was simulated by incubating primarily cultured neonatal rat cardiomyocytes in ischemic (hypoxic) buffer solution for 2 h. Thereafter, these cells were incubated for 24 h in normal culture medium.
RESULTS: Treatment of the cardiomyocytes with 10 μM bromocriptine significantly decreased lactate dehydrogenase activity, increased superoxide dismutase activity, and decreased malondialdehyde content in the culture medium. Bromocriptine significantly inhibited the release of cytochrome c, accumulation of [Ca2+]i, and apoptosis induced by ischemia/reperfusion injury. Bromocriptine also down-regulated the expression of caspase-3 and -9, Fas and Fas ligand, and up-regulated Bcl-2 expression. In contrast, haloperidol (10 μM) had no significant effects on the apoptosis of cultured cardiomyocytes under the aforementioned conditions.
CONCLUSIONS: These data suggest that activation of dopamine D2 receptors can inhibit apoptosis of cardiomyocytes encountered during ischemia/reperfusion damage through various pathways.

Background

Dopamine receptors belong to the family of seven transmembrane domain G-protein coupled receptors. The diverse physiological actions of dopamine are mediated by at least five distinct G protein-coupled receptor subtypes. Two D1-like receptor subtypes (D1 and D5) activate adenylyl cyclase. The other receptor subtypes belong to the D2-like subfamily (D2, D3, and D4), inhibit adenylyl cyclase, and activate K+ channels [1]. The activation of D2-like receptors decreases intracellular calcium levels. The underlying mechanisms for this effect may be related to D2-like receptor-induced activation of potassium currents and the subsequent alterations in membrane potential and activation of G proteins. The latter directly inhibits some calcium channels [2].

After myocardium undergoes severe ischemia, restoration of blood flow is a prerequisite for myocardial salvage. However, reperfusion may also induce deleterious changes, such as decreased myocardial contraction and arrhythmias. These changes occur at the time of reperfusion, termed as ''reperfusion injury''. The mechanisms of ischemia/reperfusion injury include the production of reactive oxygen species, abnormal lipid metabolism, and calcium overload [3]. The resulting alteration in cellular metabolism and generation of toxic molecules contribute to tissue damage in ischemia/reperfusion injury, which is characterized by the presence of necrotic and apoptotic areas in the affected organs [4].

Dopamine receptors exist in cardiac tissue [5]. The activation of dopamine D2 receptors decreased heart rate, arterial blood pressure [5]. The expression of dopamine D2 receptors was reduced in myocardial hypertrophy [6]. Dopamine D2 receptors also had the protective effect on cerebral ischemia/reperfusion injury [7]. However, the role of dopamine D2 receptors on myocardial ischemia/reperfusion injury has not been clear.

In this study, we applied myocardial ischemia/reperfusion injury in neonatal rat cardiomyocytes and explored the effects of dopamine D2 receptors agonist and antagonist on cell apoptosis as well as the underlying signal transduction pathways.

Materials and methods

Materials

Bromocriptine (dopamine D2 receptors agonist), haloperidol (dopamine D2 receptors antagonist), and fluo-3/AM were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Antibodies against DR2, caspase-3 and -9, Bcl-2, Fas, Fas ligand (Fas-L), and cytochrome c (Cyt c) were from Santa Cruz (Heidelberg, Germany). The terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling (TUNEL) kit was purchased from Roche (Mannheim, Germany). The Western blot kit was from Promega (Madison, WI, USA). The antibody against β-actin was purchased from Boster (Wuhan, China), and the antibody against annexin-V was from Promega. Dulbecco's modified Eagle's medium (DMEM) was purchased from Gibco (Carlsbad, California, USA). Assay kits for malondialdehyde (MDA), superoxide dismutase (SOD) and lactate dehydrogenase (LDH) were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).

Cell culture and treatment

Primary cultures of neonatal rat cardiomyocytes were performed as previously described [8]. Newborn Wistar rats (1-3 days) were used for this study. The rats were handled in accordance with the Guide for the Care and Use of Laboratory Animals published by the China National Institutes of Health. Briefly, hearts from male Wistar rats (1-3 days old) were minced and dissociated with 0.25% trypsin. Dispersed cells were seeded at 2×105 cells/cm2 in 60-mm culture dishes with Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and then cultured in a 5% CO2 incubator at 37°C. Three days after the cells were seeded, the cultured cardiomyocytes were randomly divided into the following four groups: control group, cardiomyocytes continuously cultured for 26 h in DMEM; ischemia/reperfusion group, cells subjected to 2 h of simulated ischemia followed by 24 h of reperfusion; bromocriptine and haloperidol treated groups. In the treatment groups, 10 μM bromocriptine or haloperidol was added to the culture medium at the beginning of reperfusion. Drugs were dissolved in pre-warmed medium and added directly to the culture. For controls, equivalent volumes of medium were added.

Simulated Ischemia/Reperfusion

The experimental protocol used to simulate ischemia/reperfusion was a modified version of the method proposed by Han et al. [9]. The cells were treated with ischemic buffer solution (mM: 5.37 KCl, 0.44 KH2PO4, 136.89 NaCl, 4.166 NaHCO3, 0.338 Na2HPO4, 5 D-glucose, pH 7.3-7.4 at 37°C) saturated with 95% N2 and 5% CO2. The pH was adjusted to 6.8 with lactate to mimic ischemic conditions. The dishes were put into a hypoxic incubator that was equilibrated with 1%O2/5%CO2/94%N2. After hypoxic treatment, the culture medium was rapidly replaced with fresh DMEM with 10% FBS (10%FBS/DMEM) to initiate reoxygenation.

Cell viability assay

Cell viability was determined by the 3-(4,5- dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT) assay, as described previously [10]. Cells were cultured in 96-well plates. MTT (final concentration, 0.5 mg/ml) was added to each well under sterile conditions, and the plates were incubated for 4 h at 37°C. The supernatant was removed, and dimethyl sulfoxide (150 μl/well) was added. The plates were then agitated on a plate shaker. The absorbance of each well was measured at 490 nm with a Bio-Rad automated EIA analyser (Bio-Rad Laboratories, Hercules, CA, USA). The viability of control cells was considered 100%, and the others were expressed as percentages of control.

Flow cytometric assay

The apoptotic ratio was measured by flow cytometry as described previously [11,12]. Cells were washed three times with ice-cold PBS, and then stained with annexin V-fluorescein isothiocyanate for 15 min at room temperature in 200 μl binding buffer. Next, 300 μl binding buffer was added, and the cells were stained with propidium iodide for 30 min at 4°C. The fluorescence of the cells was analyzed by flow cytometry. The percentage of apoptotic cells was determined using Mod Fit LT software (Verity Software House Inc., Topsham, ME, USA).

TUNEL staining

Apoptosis was detected by the TUNEL method. In accordance with the manufacturer's protocol, cells in 24-well plates were fixed with 4% paraformaldehyde for 30 min at room temperature. After two washes with PBS, cells were permeabilized with 0.1% Triton X-100 for 2 min at room temperature, then incubated with 50 μl TUNEL reaction mixture for 60 min at 37°C. 3,3'-Diaminobenzidine was used to generate an insoluble coloured substrate at the site of DNA fragmentation. Finally, the cells were counterstained with methyl green for morphological evaluation and characterisation of normal and apoptotic cells. All cells were analyzed under a microscope. The percentage of apoptotic cells was calculated as the ratio of the number of TUNEL-positive cells to the total number of cells, counted in three different random fields.

Transmission electron microscopy

Cells were harvested and fixed with 3.0% glutaraldehyde and 1.5% paraformaldehyde, washed in phosphate buffer saline (PBS), and fixed in osmium tetroxide. Then, cells were dehydrated in an ethanol series, embedded in epoxy resin and examined under a transmission electron microscope (JEM-2000EX).

Measurement of [Ca2+]i

Cultured cardiomyocytes in 96-well plates were loaded with 10 μM Fluo-3/AM for 60 min at 37°C. Cells were then rinsed with Ca2+-free PBS three times to remove the extracellular fluo-3/AM, and 200 μl DMEM was added. Excitation was set at 488 nm, and emission was monitored at 530 nm. Fluorescence images indicating the [Ca2+]i were recorded using a confocal laser scanning microscope (Leica, Germany). The loaded cells were randomly divided into the following three groups: ischemia/reperfusion group in which cells were incubated for 8 min with DMEM and 60 min with simulated ischemia solution, then reperfused for 30 min with DMEM; bromocriptine or haloperidol treated groups in which the similar protocol for the ischemia/reperfusion group was used but 10 μM bromocriptine or haloperidol was added to the medium at the beginning of reperfusion.

Detection of cytochrome c release from mitochondrial

To quantify cytochrome c release, western blot analysis of cytochrome c in the cytosolic fraction was performed as described previously [13,14]. Briefly, cells were harvested, washed twice with ice-cold PBS, and incubated in ice-cold Tris-sucrose buffer (0.35 mM sucrose, 10 mM Tris-HCl at pH 7.5, 1 mM EDTA, 0.5 mM dithiothreitol, 0.1 mM phenylmethylsulphonyl fluoride). After a 40 min incubation, cells were centrifuged at 1000×g for 5 min at 4°C and the supernatant was further centrifuged at 40,000×g for 30 min at 4°C. The supernatant was retained as the cytosolic fraction and analyzed by Western blot with a primary rat anti-cytochrome c monoclonal antibody and a secondary goat anti-rat immunoglobulin G (Promage). β-actin expression was used as the control.

Detection of the expression of DR2, caspases-3 and -9, Fas, FasL, and Bcl-2 by Western blot

Total proteins were prepared from neonatal rat myocytes as described previously [15,16]. To each lane of a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, 20 μg of total protein was applied, electrophoresed and transferred to a polyvinylidene fluoride membrane. Membranes were blocked with Tris-buffered saline containing 5% non-fat milk at 37°C for 1 h, then incubated overnight at 4°C with anti-DR2 (1:500), anti- caspase-3 (1:500), anti- caspase -9 (1:500), anti-Fas (1:500), anti-Fas-L (1:500), anti-Bcl-2 (1:500) antibodies (from Santa Cruz of Heidelberg, Germany). The membrane was then washed three times in 1×Tris-buffer saline-Tween20 (TBST) buffer and incubated with anti-IgG antibody conjugated with alkaline phosphatase diluted 1:1000 in TBST for 1 h at room temperature. Antibody-antigen complexes were detected using Western Blue Stabilized Substrate for alkaline phosphatase. β-actin expression was used as the control. The intensities of the protein bands were quantified by a Bio-Rad ChemiDoc™ EQ densitometer and Bio-Rad Quantity One software (Bio-Rad Laboratories).

Measurement of MDA level, LDH and SOD activities

The level of malonaldehyde (MDA) in the culture medium was detected with thiobarbituric acid-reactive substances assay and the experiment was carried out with a commercial kit (Jiancheng, Nanjing, China) according to manufacturer's introduction. MDA values were expressed as nmole per gram protein.

Superoxide dismutase activity (SOD) in the culture medium was determined by monitoring the inhibition of the autoxidation of hydroxylamine using previously described methods [17]. The protein in the culture medium was detected by the Coomassie brilliant blue method.

The activity of Lactate dehydrogenase activity (LDH) in the culture medium, as an indicator of cytotoxicity, was measured spectrophotometrically with a commercially available assay kit (Jiancheng Bioengineering Institute).

Statistical analysis

All experiments were performed at least three times for each determination. Data were expressed as mean ± S.E.M. Comparisons among the groups were performed by KruskaleWallis one-way ANOVA. Values of < 0.05 were considered as significant.

Results

Measurement of LDH, SOD and MDA levels

In the ischemia/reperfusion group, the levels of LDH and MDA were significantly increased but SOD activity was decreased compared with the control group (P < 0.01). In the bromocriptine-treated group, LDH and MDA levels were significantly decreased but SOD activity was increased (P < 0.05 or P < 0.01 versus ischemia/reperfusion group). No changes of LDH, MDA and SOD were observed in the haloperidol-treated group in comparison with that in ischemia/reperfusion group (P > 0.05) (Table 1).

Table 1

The levels of LDH and MDA as well as SOD activity in Culture medium from different groups

Groups	LDH (U/mL)	SOD (U/mL)	MDA (nmol/mL)
Control	29.7 ± 2.2	36.1 ± 1.2	8.8 ± 0.6
I/R	56.4 ± 1.8a	30.4 ± 2.3a	17.4 ± 1.2a
Bromocriptine	44.3 ± 2.3c	36.6 ± 2.1c	14.2 ± 0.8b
Haloperidol	58.1 ± 1.9	31.7 ± 2.1	17.2 ± 1.3

Data are means ± S.E.M. of 10 determinations. aP < 0.01 versus control; bP < 0.05, cP < 0.01 versus ischemia/reperfusion (I/R).

Cell viability

MTT analysis showed that the viability of the bromocriptine-treated cells (81% ± 3.0%) was significantly higher than that of the ischemia/reperfusion group (74% ± 3.3%) (P < 0.01). However, haloperidol treatment had no effect on cell viability (73% ± 3.8%) compared with the ischemia/reperfusion group (P > 0.05) (Figure 1A).

Figure 1

The effects of dopamine D. Cardiomyocytes were subjected to 2 h of simulated ischemia, and then incubated under normal conditions in the presence or absence of 10 μM bromocriptine or haloperidol for 24 h. The data are means ± S.E.M. of five determinations.▼▼P < 0.01 versus control; ★★P < 0.01 versus ischemia/reperfusion. (A) MTT assay for cell viability. (B) DNA fragmentation determined by quantitative TUNEL staining of apoptotic cardiomyocytes. (C) Flow cytometry analysis.

Rate of apoptotic cell

Only 6 ± 0.7% TUNEL positive nuclei were detected in control cells at the end of the experiment, and ischemia/reperfusion significantly increased the percentage of apoptotic cells to 18 ± 1.6% (P < 0.01 versus control group). Compared with the ischemia/reperfusion group, bromocriptine treatment reduced the percentage of TUNEL-positive cells to14 ± 1.0% (P < 0.01). The percentage of TUNEL-positive cells in the haloperidol-treated group was not significantly different from that in the ischemia/reperfusion group (P > 0.05) (Figure 1B).

Flow cytometry analyses detected few apoptosis in the control group. Apoptosis was significantly increased when cells were exposed to ischemia/reperfusion (P < 0.01 versus control group), and pretreatment with bromocriptine decreased the apoptosis rate significantly (P < 0.01 versus ischemia/reperfusion group). In contrast, pre-treatment with haloperidol did not significantly affect the apoptosis induced by ischemia/reperfusion (P > 0.05 versus ischemia/reperfusion group) (Figure 1C).

Morphological changes

The results of transmission electron microscopy showed that structure of nuclear membrane was clear with even distribution of nuclear chromatin, and mitochondria structure was intact in the control group. Morphological change of apoptosis, including nuclear pycnosis, nuclear chromatin margination, aggregation and condensation, and swelling and vacuolization of mitochondria were observed in the ischemia/reperfusion and the haloperidol-treated groups. Compared with ischemia/reperfusion, these morphological changes were less severe in the bromocriptine-treated group (Figure 2).

Figure 2

Ultrastructural changes in cardiomyocytes. (A) Control group; (B) ischemia/reperfusion group; (C) The bromocriptine-treated group; and (D) The haloperidol-treated group (magnification ×8000). In the ischemia/reperfusion group, nuclear chromatin margination, aggregation, and condensation and mitochondrial swelling and vacuolisation were observed. There was no significant change in nuclear chromatin in the control group. Cardiomyocyte injury was lessened in the bromocriptine-treated group compared with the ischemia/reperfusion group but was not significantly changed in the haloperidol-treated group.

The concentration of intracellular free calcium was recorded continuously by laser confocal scanning microscopy. The average cytosolic Ca2+ concentration in a single cell was determined using the fluorescent probe fluo-3/AM, which is a sensitive Ca2+ probe. The Ca2+ peak was increased in the ischemia/reperfusion group (P < 0.01 versus control group). When bromocriptine was added to the medium at the beginning of reperfusion, the Ca2+ peak was significantly decreased compared with the peak in the ischemia/reperfusion group (P < 0.05 versus ischemia/reperfusion group). However, haloperidol did not alter ischemia/reperfusion-induced change in the Ca2+ peak (P > 0.05) (Figure 3).

Figure 3

Changes of intracellular free calcium concentration. (A) Fluorescent images of cardiomyocytes incubated in DMEM, or incubated in simulated ischemia solution and re-incubated in DMEM in the presence or absence of 10 μM bromocriptine or haloperidol, respectively. (B) Changes in fluorescent intensities of intracellular Ca2+ in different treatment groups were recorded continuously with confocal laser scanning microscopy. Intracellular Ca2+ was recorded for a total of 98 min. ★P < 0.05 versus ischemia/reperfusion.

Expression of DR2

The expression of DR2 was increased markedly in the ischaemia/reperfusion group compared to the control group (P < 0.01) (Figure 4).

Figure 4

Expression of DR2. The intensity of each band was quantified by densitometry, and data were normalized to the β-actin signal. The expression levels in the control group were considered the basal levels, and the others are expressed as fold change from the control group. The fold change values represent the means ± S.E.M. of five determinations. ▼▼P < 0.01 versus control.

Expression of Bcl-2

The expression of Bcl-2 was increased significantly in the ischemia/reperfusion group compared with the control group (P < 0.01). However, Bcl-2 expression was increased more in the bromocriptine group than the ischemia/reperfusion (P < 0.01). In the haloperidol-treated group, Bcl-2 expression was similar to that of the ischemia/reperfusion group (Figure 5C).

Figure 5

Expression of caspase-9 (A), release of cytochrome . The intensity of each band was quantified by densitometry, and data were normalized to the β-actin signal. The expression levels in the control group were considered the basal levels, and the others are expressed as fold change from the control group. The fold change values represent the means ± S.E.M. of five determinations. ▼▼P < 0.01 versus control; ★★P < 0.01 versus ischemia/reperfusion.

Expression of Caspases-3 and -9, Fas, Fas-L, and release of Cyt c

Compared with the control group, the expression of caspases-3 and -9, Fas, Fas-L, and the mitochondrial release of cytochrome c were significantly increased in the ischemia/reperfusion group (P < 0.01), and bromocriptine reversed these increases (P < 0.05 or P < 0.01 versus ischemia/reperfusion group). In contrast, the expression of caspases-3 and -9, Fas, Fas-L, and cytochrome c release were not changed obviously in the haloperidol-treated group compared with the ischemia/reperfusion group (P > 0.05) (Figure 5A-B, 6A-C).

Figure 6

Expression of Fas (A), Fas-L (B) and caspase-3 (C) determined by Western blot analysis. The intensity of each band was quantified by densitometry, and data were normalized to the β-actin signal. The expression levels in the control group were considered the basal levels, and the others are expressed as fold change from the control group. The fold change values represent the means ± S.E.M. of five determinations. ▼▼P < 0.01 versus control; ★P < 0.05, ★★P < 0.01 versus ischemia/reperfusion.

Discussion

Cardiomyocyte apoptosis is a major pathogenic mechanism underlying myocardial ischemia/reperfusion injury [18]. However, very little is known about the effect of dopamine D2 receptors activation or suppression on cardiac apoptosis-induced by ischemia/reperfusion injury. In this paper, we examined whether bromocriptine (a selective agonist of dopamine D2 receptors) or haloperidol (dopamine D2 receptors antagonist) play a role in cardiomyocyte apoptosis induced by simulated ischemia/reperfusion injury. The results showed that bromocriptine significantly decreased LDH and MDA levels, increased SOD activity, reduced morphological injury of cardiomyocytes, increased the cell viability, and decreased apoptosis. We also observed that DR2 expression was increased markedly during ischaemia/reperfusion. These results suggested that activation of dopamine D2 receptors significantly inhibit apoptosis induced by ischemia/reperfusion injury.

Intracellular calcium is regarded as a trigger for cardiomyocyte apoptosis in ischemia/reperfusion injury. Massive intracellular calcium accumulation can increase free radical generation, damage mitochondria, release apoptotic promoters, and activate the caspase cascade [19]. It was reported that D2-like receptors might suppress activation of adenylate cyclase, down-regulate cyclic AMP (cAMP) levels, then inhibit the activation of protein kinase A (PKA), inhibit L-type voltage-gated channels, and finally decrease [Ca2+]i [20]. Our study observed that [Ca2+]i was increased significantly after ischemia/reperfusion, while bromocriptine markedly decreased [Ca2+]i compared with the ischemia/reperfusion group. Moreover, bromocriptine increased SOD activities and decreased MDA contents. These results indicate that activation of dopamine D2 receptors may play an anti-apoptotic role by inhibiting the calcium overload and scavenging free radical.

Apoptosis can be initiated through two basic pathways. One is the mitochondrial or intrinsic pathway, which is initiated by stress signals through the release of cytochrome c from the mitochondrial intermembrane space. This triggers cytosolic caspase-3 activation through formation of the cytochrome c/Apaf-1/caspase-9- containing complex apoptosome. The other one is the so-called extrinsic pathway, which is a receptor mediated cell death initiation process. Stimulation of death receptors of the tumor necrosis factor receptor superfamily (Fas and TNF receptor) results in activation of the initiator caspase-8, which can propagate the apoptosis signal by direct cleavage of downstream effector caspases such as caspase-3[21-23]. Caspase-3, a central 'executioner' or 'downstream' caspase, is an important effector molecule in apoptosis [24]. Bcl-2, a potent inhibitor of apoptosis, can prevent disruption of the mitochondria and the subsequent release of cytochrome c, and caspase activation [25]. It is well-known that ischemia/reperfusion can cause increase of oxygen free radical and intracellular calcium overload, which induce apoptosis by mitochondrial permeability transition pore opening and increase of Fas/Fas-L (TNF-α) transcription through activation of MAPK-NF-κB pathway.

Although the effective concentration of bromocriptine to protect reperfusion-induced cell apoptosis is 60 to 100 fold higher than the peak therapeutic plasma concentration, the inhibition of reperfusion-induced increase of cellular calcium by bromocriptine could be antagonized by 10 μM haloperidol. This result indicate that the inhibition of hypoxia/reperfusion induced apoptosis by bromocripline may be related to the stimulation of dopamine D2 receptor.

Our experimental results showed that bromocriptine down-regulated the expression of pro-apoptotic genes (caspase-3 and -9, Fas and Fas-L), decreased release of cytochrome c, and increased the expression of anti-apoptotic gene (Bcl-2). These data suggest that the activation of dopamine D2 receptors inhibit apoptosis by down-regulating both cytochrome c-caspase-3 and Fas/Fas-L signaling pathways.

Interestingly, we found that haloperidol had no effect on the aforementioned indices related to apoptosis, compared with the ischemia/reperfusion group. The precise role of dopamine D2 receptor inhibition in apoptosis induced by ischemia/reperfusion need to be explored further.

Conclusions

Our study indicated that activation of dopamine D2 receptors could inhibit apoptosis by reducing calcium overload, scavenging free radical, down- regulating cytochrome c-caspase-3 mitochondrial and Fas/Fas-L death receptor signaling pathways. Attenuation of apoptosis by dopamine D2 receptors activation may open a new therapeutic window in the treatment of ischemia/reperfusion injury.

Abbreviations

DR: dopamine receptors; DR2: dopamine D2 receptors; Bro: Bromocriptine; Hal: haloperidol; I/R: ischemia/reperfurion; Cyt c: cytochrome c; MDA: malondialdehyde; SOD: superoxide dismutase; LDH: lactate dehydrogenase; MTT: 3-(4,5-dimethyl thiazol-2yl)-2,5-diphenyltetrazolium bromide; cAMP: cyclic AMP; PKA: protein kinase A; TUNEL: TdT-mediated dUTP nick end labeling; MAPK: mitogen-activated protein kinase;

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

HZL, JG and CQX drafted the manuscript, LPH, CMJ, HXL, SZB and GWL participated in the design of the study and did most of the experiments, LNW and HL conceived of the study, YJZ, WHZ and YL participated in its design and coordination, YT, BFY, GDY, RW and LYW revised the paper and gave some suggestions. All authors read and approved the final version of the manuscript.