All‐trans retinoic acid protects against doxorubicin‐induced cardiotoxicity by activating the ERK2 signalling pathway

Abstract

Background and Purpose

Doxorubicin is a powerful antineoplastic agent for treating a wide range of cancers. However, doxorubicin cardiotoxicity of the heart has largely limited its clinical use. All‐trans retinoic acid (ATRA) plays an important role in many cardiac biological processes, but its protective effects on doxorubicin‐induced cardiotoxicity remain unknown. Here, we studied the effect of ATRA on doxorubicin cardiotoxicity and the underlying mechanisms.

Experimental Approaches

Cellular viability assays, Western blotting and mitochondrial respiration analyses were employed to evaluate the cellular response to ATRA in H9c2 cells and primary cardiomyocytes. Quantitative PCR and gene knockdown were performed to investigate the underlying molecular mechanisms of ATRA's effects on doxorubicin cardiotoxicity.

Key Results

ATRA significantly inhibited doxorubicin‐induced apoptosis in H9c2 cells and primary cardiomyocytes. ATRA was more effective against doxorubicin cardiotoxicity than resveratrol and dexrazoxane. ATRA also suppressed reactive oxygen species generation and restored expression levels of mRNA and proteins in the phase II detoxifying enzyme system: nuclear factor‐E2‐related factor 2, manganese superoxide dismutase, haem oxygenase‐1, and mitochondrial function (mitochondrial membrane integrity, mitochondrial DNA copy numbers and mitochondrial respiration capacity, biogenesis and dynamics). Both a ERK1/2 inhibitor (U0126) and ERK2 siRNA, but not ERK1 siRNA, abolished the protective effect of ATRA against doxorubicin‐induced toxicity in H9c2 cells. Remarkably, ATRA did not compromise the anticancer efficacy of doxorubicin in gastric carcinoma cells.

Conclusions and Implications

ATRA protected cardiomyocytes against doxorubicin‐induced toxicity, by activating the ERK2 pathway, without compromising its anticancer efficacy. Therefore, ATRA is a promising candidate as a cardioprotective agent against doxorubicin cardiotoxicity.

Abbreviations

ATRA

all‐trans retinoic acid

Bcl‐2

B‐cell lymphoma 2

Bcl‐xl

B‐cell lymphoma‐extra‐large

BrdU

5‐bromo‐2′‐deoxyuridine

Dox

doxorubicin

Drp1

dynamin‐related protein 1

FCCP

carbonyl cyanide‐4‐(trifluoromethoxy)phenylhydrazone

HO1

haem oxygenase‐1

Mfn1

mitofusin 1

Mfn2

mitofusin 2

MMP

mitochondrial membrane potential

MnSOD

manganese superoxide dismutase

MTT (3‐[4,5‐dimethylthiazol‐2‐yl]‐2,5‐diphenyltetrazolium bromide) Nrf2

nuclear factor‐E2‐related factor 2

OCR

oxygen consumption rate

OPA1

optic atrophy 1

RARα

retinoic acid receptor α

ROS

reactive oxygen species

TFAM

mitochondrial transcription factor A

Tables of Links

TARGETS
Nuclear hormone receptors a	Enzymes b
PPARα	Caspase 3
RARα	ERK1
	ERK2
	HO1
	JNK
	MEK
	SIRT3 (HDAC3)

LIGANDS
ATRA	PD98059
Bcl‐2	Resveratrol
Dexrazoxane	SB203580
Doxorubicin	SP600125
LY294002	U0126
	Wortmannin

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2013/14 (^a,bAlexander et al., 2013a,2013b).

Introduction

Various types of human tumours can be treated with anthracyclines. Doxorubicin is one of the most widely used anthracyclines in chemopreventive applications. However, the cardiotoxicity of doxorubicin limits its clinical use (Minotti et al., 2004). An accumulation of doxorubicin‐induced oxidative stress leads to cardiac dysfunction, including cardiomyocyte death (Sterba et al., 2013). A decline in antioxidant enzyme activity (Gu et al., 2012) and mitochondrial dysfunction [energy metabolism, redox balance (Kuznetsov et al., 2011), decrease of mitochondrial biogenesis (Miyagawa et al., 2010) and dynamics (Kuzmicic et al., 2011)] have been reported in doxorubicin cardiotoxicity. ERK activation provides a possible mechanism for protection against doxorubicin‐induced cardiomyocyte damage (Izumi et al., 2006; Simoncikova et al., 2008). These observations suggest that some agents may protect against doxorubicin cardiotoxicity by inhibiting oxidative stress, restoring cardiac mitochondrial function or stimulating ERK activation.

Retinoic acid derivatives are essential for tissue homeostasis, cellular proliferation (Wiggert et al., 1978) and the prevention of various diseases, such as neurotoxicity (Cheng et al., 2013; Kim et al., 2013), hepatitis and hepatic ischaemia (Rao et al., 2010; Nagy, 2012). All‐trans retinoic acid (ATRA) protects cardiomyocytes from various stimuli by inhibiting oxidative stress, preventing cardiomyocyte apoptosis and attenuating p38 MAPK, JNK and NF‐κB activation (Choudhary et al., 2008; Lou et al., 2013; Nizamutdinova et al., 2013). However, the effect of ATRA on doxorubicin cardiotoxicity remains unknown. This study was aimed at evaluating the protective effects of ATRA against doxorubicin‐induced damage of cardiomyocytes and the potential involvement of the ERK signalling pathway in this protection.

Methods

Cell culture and treatment

H9c2 cells were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO₂ in DMEM supplemented with 10% FBS, penicillin (100 U·mL⁻¹), and streptomycin (100 μg·mL⁻¹). H9c2 cells were pretreated with different concentrations (5, 250, 1000 and 2000 nmol·L⁻¹) of ATRA for 24 h and then were treated with 3 μmol·L⁻¹ doxorubicin for another 24 h. In the experiments with inhibitors, H9c2 cells were pretreated for 1 h with U0126 (50 μmol·L⁻¹), SB203580 (50 μmol·L⁻¹), SP600125 (50 μmol·L⁻¹), PD98059 (50 μmol·L⁻¹), LY294002 (50 μmol·L⁻¹) or wortmannin (1 μmol·L⁻¹) before ATRA was added to the growth media.

Primary cultures of rat neonatal cardiomyocytes were prepared from the ventricles of 1‐ to 3‐day‐old Sprague–Dawley rat pups as previously described (Chlopcikova et al., 2001). Cardiomyocytes were seeded on six‐well plates overnight in DMEM with 10% FBS, penicillin (100 U·mL⁻¹) and streptomycin (100 μg·mL⁻¹). BrdU was used to inhibit the growth of cardiac fibroblasts. Primary cardiomyocytes were treated with different concentrations (0.1, 0.5, 1 or 2 μmol·L⁻¹) of ATRA for 24 h and then were treated with doxorubicin.

Human gastric carcinoma SGC‐7901 and AGS cells were maintained at 37°C in RPMI 1640 medium with 10% FBS, penicillin (100 U·mL⁻¹) and streptomycin (100 μg·mL⁻¹). SGC‐7901 and AGS cells were pretreated with ATRA (2 μmol·L⁻¹) for 24 h and then were treated with doxorubicin for 24 h.

MTT assay

For viability assays, the cells were incubated with MTT (5 μg·mL⁻¹) for 1 h at 37°C, and then, the formazan precipitate was dissolved with dimethyl sulfoxide. The optical densities of the resulting formazan solutions were measured using a microplate reader (Thermo 1500) at 570 nm.

Detection of intracellular ROS

Intracellular reactive oxygen species (ROS) generation was analysed using the oxidation‐sensitive fluorescent dye carboxy‐H₂DCF‐DA, which can be oxidized by ROS to emit green fluorescence. H9c2 cells were incubated with DCFH‐DA (1 μmol·L⁻¹) at 37°C for 30 min and then were washed twice with PBS followed by image acquisition using an inverted florescence microscope (Olympus IX71).

Mitochondrial membrane potential evaluation

The mitochondrial membrane potential (MMP) in H9c2 cells was assessed using JC‐1 staining. H9c2 cells were incubated with JC‐1 (1 μmol·L⁻¹) at 37°C for 30 min and then were washed with PBS followed by image acquisition using an inverted florescence microscope (Olympus IX71).

Real‐time PCR

Total RNA was extracted from the cells using TRIzol reagent (Roche, Basel, Switzerland) following the manufacturer's protocol. RT was performed using the PrimeScript RT‐PCR Kit followed by semi‐quantitative real‐time PCR using specific primers.

The primer sequences were as follows:

• Rat HO1:

• (forward) 5′‐TGCTCGCATGAACACTCTG‐3′,

  (reverse) 5′‐TCCTCTGTCAGCAGTGCCT‐3′;

• Rat Nrf2:

• (forward) 5′‐TTCCTCTGCTGCCATTAGTCAGTC‐3′,

  (reverse) 5′‐GCTCTTCCATTTCCGAGTCACTG‐3′;

• Rat MnSOD:

• (forward) 5′‐TGGACAAACCTGAGCCCTAA‐3′,

  (reverse) 5′‐GACCCAAAGTCACGCTTGATA‐3′;

• Rat α‐tubulin:

• (forward) 5′‐TGCTGCCATTGCCACCATCA‐3′,

  (reverse) 5′‐CTCACCCTCACCCTCCACCG‐3′.

mtDNA content assay

Total DNA was extracted using the QIAamp DNA Mini Kit. Quantitative PCR was performed using 18S rRNA as an endogenous reference gene and a mitochondrial D‐loop as the target gene. Quantitative PCR reactions were performed using the SYBR Premix Ex Taq™, 0.5 μL of each primer (10 pmol·L⁻¹) and 2.5 ng of template (DNA) or no template, with RNase‐free water added to a final volume of 10 μL. The cycling conditions were as follows: initial denaturation at 95°C for 5 min, followed by 40 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 20 s.

The following primers were used:

• Rat mitochondrial D‐loop:

• (forward) 5′‐CACCCAAGAACAGGGTTTGT‐3′,

  (reverse) 5′‐TGGCCATGGGTATGTTGTTAA‐3′;

• Rat 18S rRNA:

• (forward) 5′‐TAGAGGGACAAGTGGCGTTC‐3′,

  (reverse) 5′‐CGCTGAGCCAGTCAGTGT‐3′

The mtDNA content was determined as the ratio of the mitochondrial D‐loop to 18S rRNA. The results were presented as a percentage of the control.

Mitochondrial bioenergetics measurement

H9c2 cells were seeded in XF 24‐well microplates (Seahorse Bioscience, Billerica, MA, USA). Medium was changed to unbuffered DMEM after treatment, and the plate was incubated in a non‐CO₂ incubator (37°C; 60 min) before running on the XF24 Analyzer (Seahorse Biosciences) to measure mitochondrial respiration. The final concentrations of the mitochondrial inhibitors were 10 μmol·L⁻¹ antimycin A, 6 μmol·L⁻¹ FCCP and 10 μmol·L⁻¹ oligomycin. Briefly, four parameters of mitochondrial function were calculated from the bioenergetics profile: basal oxygen consumption rate (OCR), ATP‐inked OCR and proton leak OCR. After detection, the cell numbers were calculated, and OCR was adjusted accordingly.

Western blotting

The cell samples were extracted in lysis buffer containing protease and phosphatase inhibitors. The protein concentrations of the lysates were determined using the BCA™ protein assay kit. Samples containing approximately 20 μg of protein were separated by SDS‐PAGE and transferred to nitrocellulose membranes (Millipore, Bedford, MA, USA). The membranes were blocked with 5% non‐fat milk in TBST for 1 h at room temperature, incubated with the indicated antibodies overnight at 4°C and then incubated with the corresponding secondary antibodies for 1 h at room temperature. Western blots were developed using ECL and quantified by scanning densitometry.

siRNA transfection

siRNAs targeting ERK1 and ERK2 were used to silence rat ERK1 and ERK2 respectively. The siRNA transfections were performed with Lipofectamine 2000 according to the manufacturer's instructions. Briefly, 100 nmol of siRNA was mixed with 5 μL of Lipofectamine 2000. The complex was added to the H9c2 cells for 24 h before any other treatments were processed. A scrambled probe was used as a negative control.

The sequences for the siRNAs were as follows:

• Rat ERK1:

• (forward) 5′‐GGCCUCAAGUACAUACACUTT‐3′,

  (reverse) 5′‐AGUGUAUGUACUUGAGGCCTT‐3′;

• Rat ERK2:

• (forward) 5′‐CCUGAGAGGAUUAAAGUAUTT‐3′,

  (reverse) 5′‐AUACUUAAUCCUCUCAGGTT‐3′;

• Rat: negative control:

• (forward) 5′‐UUCUCCGAACGUGUCACGUTT‐3′,

  (reverse) 5′‐ACGUGACACGUUCGGAGAATT‐3′.

Statistical analysis

Data are expressed as the mean ± SEM of at least five independent experiments. The protective effects and inhibition were assessed using graphpad prism 6 software (GraphPad, San Diego, CA, USA). Relative protein semi‐quantification was performed using quantity one software (Bio‐Rad, Hercules, CA, USA). The mitochondrial respiration capacity was assessed using seahorse xf24 extracellular flux analysis software (Seahorse Bioscience, Billerica, MA, USA). The gene relative expression ratio was calculated by the 2^−△△CT method. Differences between groups were assessed by anova followed by Tukey's multiple comparison tests. A P value less than 0.05 was considered to be statistically significant.

Chemicals and reagents

ATRA, resveratrol, dexrazoxane, doxorubicin, SP600125, SB203580, U0126, 5‐bromo‐2′‐deoxyuridine (BrdU), 3‐[4,5‐dimethylthiazol‐2‐yl]‐2,5‐diphenyltetrazolium bromide (MTT), carbonyl cyanide‐4‐(trifluoromethoxy)phenylhydrazone (FCCP) and antimycin A were obtained from Sigma (St Louis, MO, USA). 2′,7′‐Dichlorodihydrofluorescein diacetate (DCFH‐DA), DAPI and JC‐1 were purchased from Invitrogen (Carlsbad, CA, USA). Oligomycin and primary antibodies against B‐cell lymphoma 2 (Bcl‐2), B‐cell lymphoma‐extra‐large (Bcl‐xl), cleaved caspase‐3, PARP and phospho‐p44/42 MAPK (ERK1/2) (Thr²⁰²/Tyr²⁰⁴) were obtained from Cell Signaling Technology (Beverly, MA, USA); antibodies against ERK2, dynamin‐related protein 1 (Drp1), haem oxygenase‐1 (HO1), manganese superoxide dismutase (MnSOD), mitofusin 1 (Mfn1), mitofusin 2 (Mfn2), nuclear factor‐E2‐related factor 2 (Nrf2), mitochondrial transcription factor A (TFAM), optic atrophy 1 (OPA1), α‐Tubulin and β‐actin were obtained from Santa Cruz Biotechnology (Heidelberg, Germany). Antibodies against PPARGC‐1α, OxPhos complex I subunit NDUFS3, OxPhos complex II 30 kDa subunit, OxPhos complex III subunit core 1, OxPhos complex IV subunit I and OxPhos Complex V subunit α were obtained from Invitrogen. Secondary HRP‐conjugated anti‐rabbit IgG (H + L), anti‐mouse (H + L) IgG, and anti‐goat (H + L) IgG antibodies were purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA, USA). Collagenases I and II were obtained from MP Biomedicals (Southern California, USA). Western lysis buffer was obtained from Beyotime Biotechnology (Shanghai, Jiangsu, China). The Pierce ECL western blotting substrate and BCA™ protein assay kit were purchased from Thermo Scientific (Rockford, USA). The QIAamp DNA Mini Kit was obtained from QIAGEN (Düsseldorf, Germany). TRIzol reagent was from Roche (Basel, Switzerland). The PrimeScript reverse transcription (RT)‐PCR Kit and SYBR premix Ex Taq II were obtained from Takara (Tokyo, Japan). The primers were from ShengGong (Shanghai, China). The siRNA constructs were from Gene Pharma (Shanghai, China).

Results

ATRA protects H9c2 cells against doxorubicin‐induced death

The viability of H9c2 cells was reduced after doxorubicin treatment. Pretreatment with ATRA prevented doxorubicin‐induced loss of cell viability in a concentration‐dependent manner. Moreover, ATRA treatment alone did not affect cell viability (Figure 1A). For subsequent experiments, 1 or 2 μmol·L⁻¹ ARTA was used. Furthermore, doxorubicin induced a reduction in Bcl‐2 and Bcl‐xl expression and activation of caspase‐3 and PARP. However, ATRA pretreatment effectively inhibited these changes in apoptosis‐associated proteins in H9c2 cells (Figure 1B and C).

Figure 1.

ATRA protects H9c2 cells from doxorubicin‐induced cell death. (A) ATRA inhibits doxorubicin‐induced viability loss in H9c2 cells (n = 5). (B) Representative images and (C) semi‐quantitative Western blot analysis of apoptosis‐associated proteins (Bcl‐2, Bcl‐xl, cleaved PARP and cleaved caspase‐3) (n = 5). Quantitative values are computed as the ratios of the density of the targeted protein to that of α‐tubulin. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ATRA protects primary cardiomyocytes against doxorubicin‐induced cell death

Pretreatment with ATRA prevented doxorubicin‐induced viability loss in a concentration‐dependent manner (Figure 2A), and the reduction in Bcl‐2 and Bcl‐xl expression and caspase‐3 and PARP activation in primary cardiomyocytes (Figure 2B and C), findings that were similar to those in H9c2 cells.

Figure 2.

ATRA protects primary cardiomyocytes against doxorubicin‐induced cell death. (A) ATRA inhibits the doxorubicin‐induced viability loss in primary cardiomyocytes (n = 5). (B) Western blot analysis of apoptosis‐associated proteins (n = 5). (C) Quantitative results of apoptosis‐associated proteins (Bcl‐2, Bcl‐xl, cleaved PARP and cleaved caspase‐3). Quantitative values are computed as the ratios of density of the targeted protein to that of α‐tubulin. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ATRA is more effective than resveratrol and dexrazoxane against doxorubicin‐induced cardiotoxicity

Both resveratrol and dexrazoxane showed significant protection against doxorubicin‐induced cardiotoxic effects only at a concentration of 100 μmol·L⁻¹, while ATRA started to show protection at a concentration of 2 μmol·L⁻¹, which is much lower than that used for resveratrol (100 μmol·L⁻¹) and dexrazoxane (100 μmol·L⁻¹) (Figure 3A). ATRA was more effective at scavenging ROS (Figure 3B) and inhibiting cell apoptosis (Figure 3C) than resveratrol and dexrazoxane at the same concentration (2 μmol·L⁻¹). Moreover, 100 μmol·L⁻¹ ATRA and dexrazoxane inhibited apoptosis better than resveratrol did at the same concentration (Figure 3D).

Figure 3.

ATRA is more effective than resveratrol (Res) and dexrazoxane (Dex) against doxorubicin‐induced cardiotoxicity. (A) The differences among ATRA, resveratrol and dexrazoxane treatment against doxorubicin‐induced viability loss in H9c2 cells (n = 5). (B) DCFH‐DA (10 μmol·L⁻¹) staining for the detection and evaluation of ROS. Representative images from each group (n = 6) are presented. Western blot analysis of the effects of (C) 2 and (D) 100 μmol·L⁻¹ ATRA, resveratrol and dexrazoxane on doxorubicin‐induced apoptosis (cleaved PARP and cleaved caspase‐3) (n = 5). Above: representative images; below: semi‐quantitative analysis. Quantitative values are computed as the ratios of the density of the targeted protein to α‐tubulin. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ATRA inhibits doxorubicin‐induced oxidative stress

Doxorubicin induced oxidative stress in H9c2 cells by provoking intracellular ROS generation (green fluorescence) and reducing the expression of phase II detoxifying enzymes (Nrf2, HO1 and MnSOD). Pretreatment with ATRA (1 and 2 μmol·L⁻¹) abolished ROS generation (Figure 4A). Furthermore, ATRA induced an up‐regulation of Nrf2, HO 1 and MnSOD mRNA levels and protected against the doxorubicin‐induced decrease in Nrf2 and MnSOD mRNA levels (Figure 4B). ATRA prevented the decrease in the expression levels of the components associated with the phase II detoxifying enzyme system in doxorubicin‐treated H9c2 cells (Figure 4C and D).

Figure 4.

ATRA reduces oxidative stress in doxorubicin‐treated H9c2 cells. (A) DCFH‐DA staining for the detection and evaluation of ROS. Representative images from each group (n = 6) are presented. (B) Western blot analysis of phase II detoxifying enzymes (n = 5). (C) Semi‐quantitative analysis for (Western blot analysis) of phase II detoxifying enzymes (Nrf2, HO1 and MnSOD). (D) PCR analysis of the mRNA levels of phase II detoxifying enzymes in doxorubicin‐treated cells. Quantitative values are computed as the ratios of the density of the targeted protein to that of α‐tubulin. Except as noted, the concentrations of ATRA (1 or 2 μmol·L⁻¹) and doxorubicin 3 μmol·L⁻¹ respectively. Data are expressed as the mean ± SEM. Data were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ATRA attenuates doxorubicin‐induced mitochondrial dysfunction

The mitochondrial membrane integrity in doxorubicin‐treated H9c2 cells was measured by JC‐1 staining. A change from red/orange to green fluorescence indicates mitochondrial membrane integrity loss, which leads to a decrease in the MMP. As shown in Figure 5A, pretreatment with ATRA attenuated the doxorubicin‐induced MMP decrease in H9c2 cells. The mtDNA content was quantified by real‐time PCR and measured as the ratio of D‐loop to 18S rRNA levels (Figure 5B). As shown in Figure 5C, ATRA pretreatment prevented the decrease in the ratio of mitochondrial D‐loop/18S rRNA and mitochondrial respiration capacity (basal respiration and ATP potential) in doxorubicin‐treated H9c2 cells. Doxorubicin also modified the expression of proteins associated with mitochondrial biogenesis and dynamics in H9c2 cells. Specifically, there were decreases in the expression of PPARGC‐1α, TFAM, Mfn1 and OPA1 (long and short isoforms) but no changes in the expression of Mfn2 and Drp1. There were different expression patterns of the respiratory chain complex proteins in doxorubicin‐treated H9c2 cells: a significant decrease in Complex I and Complex II, a slight decrease in Complex III and Complex V and a significant increase in Complex IV. Moreover, ATRA pretreatment inhibited all of these changes in doxorubicin‐treated H9c2 cells (Figure 5D and I).

Figure 5.

ATRA inhibits doxorubicin‐induced mitochondrial dysfunction in H9c2 cells. (A) JC‐1 staining for the MMP. Representative images from each group (n = 6) are presented. (B) Mitochondrial DNA copy number (n = 5). (C) Mitochondrial respiration capability (n = 5). Western blot analysis of (D) mitochondrial biogenesis‐associated proteins (PPARGC‐1α and TFAM), (F) mitochondrial respiratory chain complexes (Complexes I–V) and (H) mitochondrial dynamics‐associated proteins (Mfn1/2, OPA1 and Drp1). (E, G, I) Semi‐quantitative analysis of Western blots. Quantitative values are computed as the ratios of density of the targeted protein to that of α‐tubulin (n = 5). Except as noted, the concentrations of ATRA and Dox are 2 and 3 μmol·L⁻¹ respectively. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ERK signalling is involved in ATRA protection against doxorubicin‐induced cardiotoxicity

Various stimuli, such as cellular stresses and growth factors, can activate the MAPK family, which plays important roles in signal transduction. MAPK pathway inhibitors (U0126, PD98059 for ERK1/2, SB203580 for p38 and SP600125 for JNK) were used to determine which pathways are involved in ATRA cardioprotection. U0126, but not PD98059, SB203580 or SP600125, completely abolished the protective effect of ATRA against doxorubicin‐induced cardiotoxicity (Figure 6A). The Akt signalling pathway also plays an important role in doxorubicin‐induced cardiotoxicity. However, none of the Akt pathway inhibitors (LY294002 and wortmannin) abolished the ATRA protective effect against doxorubicin‐induced cardiotoxicity (Figure 6B). ATRA treatment evoked a marked up‐regulation of the phosphorylated ERK1/2 level in a time‐dependent manner, and the ERK phosphorylation levels were still significantly elevated at 24 h (Figure 6C‐i). The levels of ERK1/2 activated by ATRA were more significant than those of doxorubicin‐induced ERK1/2 activation in H9c2 cells. These results suggest that ATRA activates ERK1/2 to inhibit doxorubicin‐induced cell death signalling (Figure 6C‐ii). Moreover, U0126 almost completely blocked the ERK phosphorylation induced by ATRA (Figure 6C‐iii). Furthermore, U0126 abolished the cardioprotective effect of ATRA against doxorubicin‐induced apoptosis (caspase‐3 and PARP cleavage), oxidative stress (ROS generation), mitochondrial dysfunction (loss of mitochondrial membrane integrity) and ERK2 phosphorylation in H9c2 cells. U0126 significantly inhibited ATRA protection against doxorubicin cardiotoxicity (Figure 6D‐F). PD98059 significantly inhibited ERK1 and slightly inhibited ERK2 activation respectively. However, PD98059 did not abolish the protective effect of ATRA on cardiotoxicity (Figure 6G‐H). Taken together, these results show that ERK2 activation may play a key role in the cardioprotective effect of ATRA.

Figure 6.

U0126 attenuates the cardioprotective effect of ATRA against doxorubicin‐induced damage in H9c2 cells. Effects of (A) MAPK and (B) Akt signalling pathway inhibitors on the protective effect of ATRA against doxorubicin‐induced cardiotoxicity (n = 5) (C) Western blotting for the effect of ATRA and U0126 on phosphorylated ERK1 and ERK2. (i) Doxorubicin and ATRA induced ERK1/2 activation in a time‐dependent manner. (ii) Effect of ATRA on ERK1/2 activation in doxorubicin‐induced cardiotoxicity. (iii) U0126 inhibited ATRA‐induced ERK1/2 activation. Western blotting for the inhibitory effects of (D) U0126 and (G) PD98059 on the protective effect of ATRA on doxorubicin‐induced cardiotoxicity in H9c2 cells, and quantitative analysis of apoptosis‐associated proteins (cleaved caspase‐3 and cleaved PARP). Quantitative values are computed as the ratios of the density of the targeted protein to that of α‐tubulin (n = 5). Western blotting for the effects of (E) U0126 and (H) PD98059 on phosphorylated ERK1 and ERK2 through inhibition of the ATRA effect on doxorubicin‐induced cardiotoxicity and semi‐quantitative analysis of ERK1 and ERK2. Quantitative values are computed as the ratios of the density of the targeted protein to that of α‐tubulin (n = 5). (F) ROS generation (DCFH‐DA staining) and the MMP (JC‐1 staining) abolished the protective effect of ATRA on doxorubicin‐induced cardiotoxicity in H9c2 cells. Representative images from each group (n = 6) are presented. Except as noted, the concentrations of ATRA and Dox were 2 and 3 μmol·L⁻¹, respectively, the concentrations of inhibitors U0126, SB203580, SP600125, PD98059 and LY294002 were each 50 μmol·L⁻¹, and the concentration of wortmannin was 1 μmol·L⁻¹. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ERK2 plays an important role in ATRA protection against doxorubicin‐induced cardiotoxicity

In addition, ERK1 and ERK2 siRNA were designed to confirm the results with U0126. The siRNAs were not toxic to H9c2 cells and showed efficient knockdown ability (Figure 7A and B). Knockdown of ERK1 or ERK2 partially or totally abolished the cardioprotection of ATRA against doxorubicin‐induced viability loss (Figure 7C), oxidative stress (ROS generation), mitochondrial dysfunction (loss of mitochondrial membrane integrity) (Figure 7D and E), and apoptosis (caspase‐3 and PARP cleavage) (Figure 7F and G) in H9c2 cells. Therefore, these results indicate that ERK, specifically ERK2, mediates the cardioprotective effect of ATRA against doxorubicin‐induced cardiotoxicity.

Figure 7.

The role of ERK1 and ERK2 knockdown in the ATRA‐mediated protective effect on doxorubicin cardiotoxicity. (A) Toxicity of ERK1 and ERK2 siRNA in H9c2 cells; (B) silencing of ERK1 and ERK2 and ERK expression and activation under basal and ATRA‐stimulated conditions. Effects of ERK1 and ERK2 silencing on (C) cell viability, (D) ROS (DCFH‐DA staining), (E) MMP (JC‐1 staining) and (F) the levels of apoptosis‐associated proteins (Western blotting) in ATRA cardioprotection (n = 5). (G) Semi‐quantitative results for apoptosis‐associated proteins (cleaved PARP and cleaved caspase‐3). Quantitative values are computed as the ratios of the density of the targeted protein to that of α‐tubulin. Except as noted, the concentrations of ATRA and doxorubicin were 2 and 3 μmol·L⁻¹ respectively. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control and ^P < 0.05 versus Dox.

ATRA effects on doxorubicin‐induced cell death in gastric cancer cells

ATRA induced a loss of viability (Figure 8A) and the generation of ROS (Figure 8B) in both AGS and SGC‐7901 gastric cancer cells, although there were different responses to ERK, caspase‐3 and PARP activation in ATRA‐treated AGS and SGC‐7901 cells (Figure 8C and D). Moreover, ATRA did not compromise the oncological efficacy of doxorubicin in gastric carcinoma cells (Figure 8).

Figure 8.

Effects of ATRA on doxorubicin‐induced cell death in gastric cancer cells. (A) ATRA‐induced viability loss in gastric cancer cells (n = 5). (B) ATRA‐induced ROS generation in gastric cancer cells. Representative images from each group (n = 6) are presented. Western blot and semi‐quantitative analysis for (C) ERK1/2 activation and (D) apoptosis‐associated proteins. Quantitative values are computed as the ratios of the density of the targeted protein to α‐tubulin (n = 5). Except as noted, the concentrations of ATRA and Dox were 2 and 3 μmol·L⁻¹ respectively. Data are expressed as the mean ± SEM and were analysed by anova. *P < 0.05 versus control.

Discussion

The clinical efficacy of doxorubicin, one of the most effective anticancer drugs known, is severely limited for clinical application due to its concentration‐dependent cardiotoxicity (Minotti et al., 2004). The recent search for cardioprotective agents that can alleviate doxorubicin‐induced cardiotoxicity has only led to the identification of dexrazoxane as a possible candidate (Cvetkovic et al., 2005; Kane et al., 2008; Goey et al., 2010). However, its effectiveness is limited, and its long‐term benefits are unclear (Wiseman et al., 1998; Schlitt et al., 2014). The unique biological profile of ATRA has attracted considerable interest for many years, and it has been shown to have both neuro and hepatic protective effects (Nagy, 2012; Kim et al., 2013). In the present study it was revealed that ATRA can protect against doxorubicin‐induced cardiotoxicity. As shown in Figure 1, ATRA demonstrated strong cardioprotection in H9c2 cells. Although H9c2 cells are the typical in vitro cell model for cardiovascular disease research, particularly for doxorubicin‐induced cytotoxicity (Green et al., 2002; Chua et al., 2006; Li et al., 2006; Turakhia et al., 2007; Spagnuolo et al., 2011; Osman et al., 2013; Dong et al., 2014), there are some different characteristics between H9c2 cells and primary cardiomyocytes. Thus, we performed the same tests in primary cells and found that the effects of ATRA were in accordance with the conclusions obtained from the results in the from H9c2 cells (Figure 2).

It is essential to identify agents that can protect against doxorubicin‐induced cardiomyocyte apoptosis but without compromising its chemopreventive effects. Compared with that of resveratrol and dexrazoxane, the cardioprotective effect of ATRA against doxorubicin was mostly achieved at a much lower effective concentration. Although resveratrol can enhance the anticancer activity of doxorubicin and exert cardioprotective effects (Zhang et al., 2011; Xu et al., 2012), the efficacy of the inhibition of apoptosis and ROS generation was much lower than that of ATRA.

Previous research has demonstrated that the induction of oxidative stress, depletion of antioxidants and reduction of phase II detoxifying enzymes in response to doxorubicin have been implicated in many tissues such as the heart, kidney, liver and brain (Pal et al., 2012; Sterba et al., 2013; El‐Moselhy et al., 2014). Nrf2 is responsible for regulating the antioxidant response element (ARE)‐driven expression of genes encoding most antioxidant and phase II detoxification enzymes, such as glutamate–cysteine ligase, glutathione S‐transferase and HO1 (Bryan et al., 2013). In the present work, ATRA reduced doxorubicin‐induced oxidative stress by preventing ROS generation and activating the antioxidant defence system (HO1 and MnSOD). Consistent with our findings, the elevation of Nrf2, HO1 and SOD levels was shown to alleviate cardiotoxicity (Yu et al., 2013; Zhao et al., 2015). It is interesting that ATRA interferes with the recruitment of Nrf2 to the ARE through RARα and by reducing Nrf2 activity to increase the susceptibility to carcinogens in cancer cells (Wang et al., 2007), suggesting that ATRA may be able to evoke cardioprotection simultaneously with chemosensitization in anticancer therapy. The MEK/ERK signalling pathway can activate the expression of Nrf2 activity and nuclear translocation, and MEK/ERK inhibitors block the up‐regulation of Nrf2 expression (Zipper et al., 2003; Yuan et al., 2006; Copple, 2012; Bryan et al., 2013; Richter et al., 2014). Combining ATRA with doxorubicin activated the ERK pathway and induced Nrf2 expression in the present work, and ERK activation was shown to be important for the induction of the phase II detoxification enzyme involved in ATRA's cardioprotective effect.

Mitochondrial dysfunction is involved in cardiac energy deficits, which contribute to heart failure (Rosca et al., 2013). The present work demonstrated that doxorubicinin impaired the MMP, caused mtDNA lesions, had a deleterious effect on the respiratory chain and the expression of mitochondrial biogenesis, as well as decreasing protein dynamics. Hence, the restoration of mitochondrial biogenesis has been shown to contribute to the attenuation of doxorubicin‐induced cardiotoxicity (Miyagawa et al., 2010). Mitochondrial dynamics has emerged as an important active mechanism that affects normal mitochondrial function in heart failure (Marin‐Garcia et al., 2013). Aerobic exercise training and resveratrol have been shown to increase Mfn1/2 in doxorubicin‐treated mice (Dolinsky et al., 2013). Consistent with our findings, mitochondrial function can be improved by ATRA and its derivatives (Berdanier et al., 2001; Chiu et al., 2008; Choudhary et al., 2008; Lin et al., 2008; Klepinin et al., 2014). SIRT3 deacetylates and activates OPA1 to regulate mitochondrial dynamics to protect cells from doxorubicin‐mediated cell death (Samant et al., 2014). SIRT3 is likely to affect a specific response to ATRA (Krupkova et al., 2014). This possibility is interesting and warrants further investigation regarding ATRA‐regulated OPA1 deacetylation in our work. In previous studies, vitamin A was found to enhance the transcription of ATPase 6 (Berdanier et al., 2001). Furthermore, 9‐cis retinoic acid induced the expression of retinoid X receptor localized to the mitochondria and significantly increased the expression of ND1, ND6 and complex I RNA (Lin et al., 2008). However, effects of ATRA on mitochondrial biogenesis and mitochondrial dynamics through RARα in cardiomyocytes have not yet been reported.

There is accumulating evidence that the MAPK superfamily plays an important role in protecting against cardiomyocyte death (Adderley et al., 1999; Zhu et al., 1999; Khan et al., 2006). ERK is particularly implicated in the biological processes of cardiomyocytes, and ERK inhibition exacerbates cardiomyocyte death and heart function impairments (Purcell et al., 2007). ERK activation in doxorubicin‐induced cardiotoxicity offers cardioprotection against apoptosis (Fryer et al., 2001; Su et al., 2006; Xiang et al., 2009). In contrast, the inhibition of ERK activation aggravates oxidative stress‐induced cardiomyocyte apoptosis (Adderley and Fitzgerald, 1999). PD98059 mediates its inhibitory properties by binding to the ERK‐specific MAPK MEK, thereby preventing the phosphorylation of ERK1/2 (p44/p42 MAPK) by MEK1/2(Alessi et al., 1995). PD98059 binds to the inactive forms of ERK1 and prevents activation by upstream activators such as ATRA in our study. U0126 can inhibit both ERK1 and ERK2, while PD98059 inhibits ERK1 more potently than ERK2 (Crews et al., 1992; Favata et al., 1998). U0126 can inhibit both ERK1 and ERK2, while PD98059 inhibits ERK1 more potently than ERK2. PD98059 did not inhibit ERK2 completely, so PD98059 failed to abolish the protective effects of ATRA. Therefore, in our study, U0126 and siRNA inhibited ERK2 and totally abolished ATRA‐mediated cardioprotection in H9c2 cells. Doxorubicin has been shown to activate ERK1/2 to induce cell death previously (Cagnol et al., 2010). In our studies, ATRA‐activated ERK provided a self‐protection mechanism. A difference in the extent of ATRA‐induced ERK activation between cardiomyocytes and gastric cancer cells may be the core reason why ATRA acts differently to prevent doxorubicin‐induced cell death.

ATRA exhibits not only anticancer effects (Mangiarotti et al., 1998; Liu et al., 2003; Rosenberg et al., 2008; Long et al., 2014) but also acts as a new class of cardioprotective agent (Bernard et al., 2011; Guleria et al., 2011; Kim et al., 2013). The results of this study also raise the possibility that ATRA may have therapeutic potential to protect the heart against the toxicity of cancer chemotherapies, suggesting that these compounds, like vincristine (Chatterjee et al., 2008), deserve further extensive pharmacology and preclinical investigations.

Taken together, the present work showed that ATRA has more potential as an effective agent against doxorubicin‐induced cardiotoxicity than dexrazoxane and resveratrol. Moreover, ERK2 activation may mediate the cardioprotective effect of ATRA against doxorubicin‐induced cardiotoxicity by attenuating oxidative stress and restoring mitochondrial function.

Author contributions

L. Y. and C. L. performed the research. L. Y., C. L. and J. L. designed the research study. X. W., C. C. and W. S. contributed essential reagents or tools. L. Y. analysed the data. L. Y., C. L. and J. L. wrote the paper.

Conflict of interest

The authors declare no conflict of interest.

Yang, L. , Luo, C. , Chen, C. , Wang, X. , Shi, W. , and Liu, J. (2016) All‐trans retinoic acid protects against doxorubicin‐induced cardiotoxicity by activating the ERK2 signalling pathway. British Journal of Pharmacology, 173: 357–371. doi: 10.1111/bph.13377.