Dkk1 exacerbates doxorubicin-induced cardiotoxicity by inhibiting the Wnt/β-catenin signaling pathway

Doxorubicin (Dox) is one of the most highly effective anti-cancer agents available for a wide range of malignancies, including breast cancer, osteosarcoma and Hodgkin's lymphomas. However, the major adverse effect of Dox is cardiotoxicity, which restricts its clinical application (Chatterjee et al., 2010). Many factors are involved in the pathogenesis of Dox-induced cardiotoxicity, including apoptosis, mitochondrial dysfunction and the accumulation of reactive oxygen species (ROS) (Octavia et al., 2012; Takemura and Fujiwara, 2007; Zhou et al., 2001). At present, the ameliorative effects of prevailing drugs used to treat Dox cardiomyopathy are unsatisfactory (Gianni et al., 2008; Takemura and Fujiwara, 2007), and the exact mechanism through which Dox acts needs to be further investigated.

The Dickkopf (Dkk) family of secretory glycoproteins, comprises four main members (Dkk1, Dkk2, Dkk3 and Dkk4) (Niehrs, 2006). Dkk1, the founding member of Dkk family, is best studied and functions as an antagonist of canonical Wnt/β-catenin signaling pathway. Dkk1 is involved in many physiological and pathological processes, and exerts an important role in the development of many diseases (Huang et al., 2018). It has been reported that Dkk1 promoted ischemia-induced DNA damage by downregulating basal LRP5 and LRP6 (LRP5/6) levels and then altering G-protein-coupled receptor (GPCR) signals (Wo et al., 2016). Overexpression of Dkk1 led to a lack of heart looping and blood regurgitation, and a block in endocardial cushion formation because endogenous Wnt/β-catenin signaling is inhibited (Hurlstone et al., 2003). Dkk1 also served as an atherogenic factor and enhanced the endothelial–mesenchymal transition in primary aortic endothelial cells (Cheng et al., 2013; Kim et al., 2011; Ueland et al., 2009). However, the function of Dkk1 in Dox-induced cardiomyopathy remains largely unexplored.

The canonical Wnt/β-catenin signaling pathway is essential in a multitude of developmental processes and for the maintenance of homeostasis via cell proliferation and migration, apoptosis, and genetic stability and instability (Kahn, 2014; Nusse and Clevers, 2017). β-catenin is the main factor modulating in this pathway, and its stabilization leads to activation of canonical Wnt cascade. The abnormal regulation of Wnt/β-catenin pathway is associated with a variety of cardiovascular diseases. The canonical Wnt/β-catenin signaling pathway is indispensable for heart development (Andersen et al., 2018; Foulquier et al., 2018; Merks et al., 2018; Sadahiro et al., 2018). Deficiency of β-catenin leads to a failure in the differentiation of cardiac components and heart regeneration (Ozhan and Weidinger, 2015; Ruiz-Villalba et al., 2016). Wnt/β-catenin signaling is also important to prevent cardiac senescence caused by Dox cardiotoxicity (Xie et al., 2018).

This present study sought to investigate whether the Dkk1–Wnt/β-catenin signaling axis plays a role in the process of Dox-induced cardiomyopathy. We found that both overexpressing Dkk1, via infection with adenovirus (Ad-Dkk1), and adding of active Dkk1 protein aggravated Dox-induced apoptosis, mitochondrial injury and cardiac dysfunction by inhibiting the canonical Wnt/β-catenin signaling pathway. Conversely, extracellular addition of specific anti-Dkk1 antibody attenuated Dox-induced cardiotoxicity. This might be helpful to better understand the mechanism of Dox-induced cardiomyopathy and provide a potential therapy for the detrimental effects of Dox treatment.

To detect the changes in Dkk1 during Dox-induced cardiotoxicity, rat embryonic ventricular myoblastic H₉C₂ cells were treated with Dox (1 μmol/l) for the times indicated. As shown in Fig. 1A–C, Dox treatment significantly reduced the cell viability and increased the proportion of cells undergoing apoptosis at 12 h, as implied by morphological changes, and MTS assay and flow cytometry analysis. Western blotting results showed that Dox caused an increase in apoptosis biomarkers (Bax relative to Bcl2, cleaved caspase 3 relative to total caspase 3, and cleaved PARP1 relative to total PARP1) (Fig. 1D,E). In addition, Dox induced the decline of mitochondrial membrane potential (by Rh123 and TMRE staining), as well as nuclear condensation (as determined by Hoechst 33342 staining) and mitochondrial matrix swelling/disorganization (by Mitotracker staining) (Fig. 1F–J). The results from quantitative real-time RT-PCR (qRT-PCR) and western blot analysis suggested that the mRNA and protein expression of Dkk1 was remarkably increased after exposure to Dox for 12 h (Fig. 1K,L). Compared with the control group, an increase in the Dkk1 protein concentration was discovered in the culture medium after 12 h Dox treatment through an ELISA assay (Fig. 1M), suggesting that Dox promoted the Dkk1 secretion process.

To further assess the changes in Dkk1 during Dox-induced cardiomyopathy in vivo, SD rats were intraperitoneally injected with three equal doses (each containing 5 mg/kg body weight) every 4 days with a total cumulative dose of 15 mg/kg body weight Dox, and the control group received the same volume of normal saline in parallel. The hearts of Dox-treated rats were obviously smaller than those of control group (Fig. 2A). Hematoxylin and eosin (HE), Masson and Picro Sirius Red (PSR) staining revealed that Dox treatment also increased the disorganization of myoctyes and cardiac fibrosis (Fig. 2B–D; Fig. S1A). The heart weight to the tibia length ratio (HW:TL) was obviously decreased in the Dox group (Fig. 2F). Echocardiography illustrated a tendency of decrease in ejection fraction (EF %), fractional shortening (FS %), stroke volume (SV), cardiac output (CO), interventricular septum (IVS), left ventricular posterior wall thickness (LVPW), left ventricular diameter (LVID) and left ventricular volume (LVV) for Dox-treated rats (Fig. 2E,G–O). Western blot analysis and Tunel staining showed that the proportion of apoptotic cardiomyocytes was significantly increased in the Dox group (Fig. 2P,Q). Moreover, we detected that these were abnormal structural changes in the nucleus (such as condensation and paramorphia) and mitochondria (including irregular arrangement, swelling, vacuolation and disrupted cristae) in Dox-treated rat hearts by using transmission electron microscopy (TEM) (Fig. 2S). All of these results show the model of Dox-induced cardiomyopathy had been successfully established in vivo. In accordance with results from H₉C₂ cells, Dox injection remarkably increased the mRNA and protein levels of Dkk1 (Fig. 2R,T,U). We also determined that Dox induced the secretion of Dkk1 into the serum (Fig. 2V).

To investigate whether Dkk1 plays a role in Dox-induced cardiotoxicity, recombinant adenovirus was used to overexpress Dkk1 (Ad-Dkk1) in H₉C₂ cells. As shown in Fig. 3A–G, overexpressing Dkk1 exacerbated Dox-induced cardiomyocyte apoptosis, mitochondrial membrane depolarization, nuclear condensation and mitochondria swelling. Given that Dox treatment promoted Dkk1 protein secretion extracellulary (Fig. 2V), we added active Dkk1 protein or anti-Dkk1 antibody (hereafter anti-Dkk1) to culture medium to explore the role of extracellular Dkk1 in Dox-induced cardiotoxicity. The active Dkk1 protein aggravated Dox-induced cardiomyocyte apoptosis and mitochondrial disorder (Fig. 3H–L). Conversely, extracellular addition of anti-Dkk1 attenuated Dox-induced apoptosis and mitochondrial dysfunction (Fig. 3M–Q). In addition, treatment with the Dkk1 inhibitor [WAY-262611 (WAY) at 1.25 μmol/l for 48 h] reversed the effects of Ad-Dkk1 in Dox treatment (Fig. 3R). These results suggested that extracellular Dkk1 aggravated Dox-induced cardiotoxicity.

Adenovirus encoding Dkk1 was transduced into the rat left ventricle through intramyocardial injection. After 14 days, we performed another 2-week Dox treatment; this gives a total in vivo overexpression duration of 28 days. Compared with the Ad-GFP group, the mRNA expression, the intracellular and serum protein levels of Dkk1 were obviously increased in the Ad-Dkk1 overexpression rats (Fig. 4A–C), suggesting that Dkk1 protein was successfully overexpressed in the Ad-Dkk1 group. The heart size of rats with Dox treatment was much smaller than in those overexpressing Dkk1 (Fig. 4D). The myocyte disorganization and fibrosis induced by Dox was worse in the hearts of Ad-Dkk1-treated rats, as shown by HE, Masson and PSR staining (Fig. 4E–G). There was no significant difference in heart weight (HW), body weight (BW) and tibia length (TL) between Ad-Dkk1 overexpression plus Dox treatment and Dox-treated rats (Fig. 4H–J). However, the HW:TL ratio was distinctly reduced after Ad-Dkk1 overexpression in the Dox-treated group (Fig. 4K). The echocardiography data revealed that both Ad-Dkk1 and Dox treatment caused cardiac dysfunction, and overexpression of Dkk1 exacerbated Dox-induced heart injury (Fig. 5A–I; Fig. S2). In addition, the level of apoptosis was exacerbated when rats were infected with Dkk1, as revealed through a Tunel assay and western blot analysis (Fig. 5J,K). As shown in TEM images, the delivery of Ad-Dkk1 exacerbated Dox-induced damage of nuclei and mitochondria (Fig. 5L).

As it has been reported that Dkk1 is an inhibitor of the canonical Wnt/β-catenin signaling pathway (Huang et al., 2018), we determined the β-catenin level after Dox treatment and found that the protein expression of β-catenin was downregulated in Dox-treated H₉C₂ cells and heart tissues (Fig. 6A,B). To determine the role of the canonical Wnt/β-catenin signaling pathway in Dox-induced cardiomyopathy, cells were pre-treated with a specific activator of β-catenin (LiCl) or its inhibitor [KYA1797K (KYA)] followed by Dox for 12 h. As shown in Fig. 6C–F, LiCl suppressed the protein expression of apoptosis biomarkers, and reduced the level of Dox-induced apoptosis in cardiomyocytes. By contrast, KYA exacerbated the Dox-triggered cardiotoxicity response (Fig. 6G–J). Furthermore, active Dkk1 protein reduced the protein expression of β-catenin, while anti-Dkk1 alleviated the downregulation of β-catenin caused by Dox (Fig. 7A,B). Dkk1 overexpression aggravated the Dox-induced reduction in β-catenin protein levels, and the Dox-induced increase in apoptosis and mitochondrial damage, which was reversed by treatment with the β-catenin activator LiCl (Fig. 7C–I). These results suggested that Dkk1 aggravates Dox-induced cardiotoxicity via inhibiting canonical Wnt/β-catenin signaling pathway.

Dox is an effective chemotherapy drug for many tumors, but its clinical application is limited by its serious cardiotoxicity (Chatterjee et al., 2010). The pathological characteristics of Dox-induced cardiotoxicity are similar to those of dilated cardiomyopathy, including cardiomyocyte apoptosis, mitochondrial swelling, energy metabolism abnormality and excessive accumulation of ROS (Green and Leeuwenburgh, 2002; Santulli et al., 2015; Štěrba et al., 2013; Umanskaya et al., 2014; Zhou et al., 2001). Existing drugs focusing on reduction of ROS accumulation fail to reverse this side effect (Gianni et al., 2008; Takemura and Fujiwara, 2007), suggesting that the mechanism of Dox-induced cardiomyopathy is very complicated, with contributions from multiple factors and pathways, and deserves further study (Hulmi et al., 2018; Li et al., 2018; Yarana et al., 2018). In this study, we found that both overexpressing Dkk1, through infection of Ad-Dkk1, and the addition of active Dkk1 protein aggravated Dox-induced apoptosis, mitochondrial injury and cardiac dysfunction. Conversely, extracellular addition of specific anti-Dkk1 antibody attenuated Dox-induced apoptosis and mitochondrial damage. Knockdown of Dkk1 and treatment with the Dkk1 inhibitor WAY in H9C2 cells showed the similar effects to that of anti-Dkk1 antibody (Figs S3, S4).

Dkk1 is well known as an inhibitor of canonical Wnt/β-catenin signaling cascade, and acts by inducing endocytosis of the ternary Dkk1–LRP5/6–Kremen1 complex or by competitively binding to the Wnt co-receptor (Foulquier et al., 2018). Dkk1 destabilizes atherosclerotic lesions and promotes plaque formation and vulnerability through inhibiting Wnt signaling during atherosclerosis (Di et al., 2017). Baicalin and geniposide exerts anti-inflammation effects to prevent the development of atherosclerosis by elevating the ratio of Wnt1 to Dkk1 (Wang et al., 2016). Dkk1 induces the endocytosis and degradation of LRP5/6 to promote cardiac ischemic injury (Wo et al., 2016). It has also been found that high circulating levels of Dkk1 are related to type 2 diabetes with cardiovascular disease (Garcia-Martín et al., 2014). Here, our results show that the mRNA expression, the intracellular and serum protein levels of Dkk1 are substantially increased after exposure to Dox treatment, suggesting that Dox promotes the Dkk1 secretion process.

Canonical Wnt/β-catenin signaling is required in embryogenesis and organogenesis and for homeostasis, and participates in many disease processes, including cardiac development and cardiovascular diseases (Kahn, 2014; Nusse and Clevers, 2017). The disruption of Wnt/β-catenin signaling in epicardial cells and cardiac fibroblasts impairs cardiac function after acute ischemic cardiac injury (Duan et al., 2012). Overexpression of β-catenin decreases apoptosis and the size of myocardial infarctions (MIs), and improves ventricular function, suggesting that canonical Wnt signaling is important in MI healing (Dawson et al., 2013). The activation of Wnt signaling inhibits adipogenic transcription factors, and is beneficial to arrhythmogenic right ventricular cardiomyopathy (Dawson et al., 2013). In addition, the long intergenic noncoding RNAs (lincRNA) p21 (also known as TRP53COR1) exerts an anti-senescence effect on Dox-associated cardiotoxicity, while that effect is widely abolished upon silencing β-catenin (Xie et al., 2018). We found that the protein levels of β-catenin were reduced in our Dox-induced cardiotoxicity model. Moreover, LiCl (a specific activator of β-catenin) suppressed Dox-induced cardiomyocyte apoptosis and mitochondrial injury, while KYA1797K (an inhibitor of β-catenin) exacerbated the Dox-triggered cardiotoxicity response, implying that activation of Wnt signaling has a protective effect on Dox-induced cardiomyopathy.

In view of the close relationship between Dkk1 and the canonical Wnt/β-catenin signaling pathway in cardiovascular diseases, we speculated that Wnt/β-catenin signaling was involved in the process of Dkk1 aggravated Dox-induced cardiotoxicity. Here, we found that overexpression of Dkk1 aggravated the Dox-induced reduction in β-catenin protein levels, and the Dox-induced increase in apoptosis and mitochondrial damage, which was reversed by treatment with the β-catenin activator LiCl. Dkk1 levels inversely co-related with β-catenin in cardiomyocytes, but β-catenin induction did not affect the protein expression of Dkk1 (Fig. S5).

In summary, this study uncovered that Dkk1 plays a key role in Dox-induced cardiomyopathy via inhibiting the canonical Wnt/β-catenin signaling pathway. Although there were some limitations, our study suggests that monitoring and lowering extracellular Dkk1 protein may serve as a potential therapeutic approach for cardiomyopathy caused by clinical application of Dox.

Doxorubicin (Dox, purity 99.37%) was bought from TargetMol (Target Molecule Corp., USA) and was dissolved in water to 10 mmol/l and stored at −20°C. KYA1797K and WAY-262611 were obtained from MCE (MedChemexpress) and diluted in DMSO and stored at −80°C. The rat Dkk1 ELISA kit was from Mlbio (Shanghai, China). Primary antibodies against Dkk1 (diluted 1:3000, goat; cat. no AF4010) were purchased from R&D systems. Rabbit antibodies against caspase 3 (diluted 1:1000; cat. no 19677-1-AP) and PARP1 (diluted 1:1000; cat. no 13371-1-AP) were from Proteintech Group (Chicago, IL). Rabbit anti-cleaved caspase 3 (diluted 1:1000; cat. no 9661) and -β-catenin (diluted 1:1000; cat. no 8480) antibodies were obtained from Cell Signaling Technology. Anti-Bax (diluted 1:1000, rabbit; cat. no ab32503), Bcl2 (diluted 1:200, rabbit; cat. no BA0412) and α-tubulin (diluted 1:5000, mouse cat. no T6199) antibodies were products from Abcam and Sigma-Aldrich, respectively.

Rat embryonic ventricular myoblastic H₉C₂ cells were received from the Cell Bank of the Chinese Academy of Sciences (CAS, Shanghai, China). The cells were maintained in Dulbecco's modified Eagle's medium (DMEM; GIBCO, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) and incubated at 37°C under 5% CO₂.

After extracting from cells or tissues, proteins were separated by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane (Millipore). Afterwards, the membranes were blocked for 1 h at ∼25°C with 5% skimmed milk [dissolved in Tris-buffered saline with 0.1% Tween 20 (TBST)] and incubated with the indicated primary antibodies at 4°C overnight, followed by incubation with horseradish peroxidase-conjugated secondary antibodies for 1 h at ∼25°C. Protein bands were detected with enhanced chemiluminescent substrate, and the intensity analysis was performed by LabWorks software (Bio-Rad).

RNA from H₉C₂ cells or rat cardiac tissues was isolated with Trizol reagent (Invitrogen). cDNA was then prepared with a one-step RT kit (Thermo Fisher Scientific) in 20 μl reactions. A SYBR-Green quantitative PCR kit in an iCycler iQ system (Bio-Rad) was used to detect the mRNA levels of each different gene. The amplification conditions were 15 min at 95°C, then followed by 40 cycles of 30 s at 95°C, 1 min at 55°C and 30 s at 72°C. All PCRs were performed in triplicate. Each rat-specific primer was synthesized by Sangon (Shanghai, China). β-actin was used as a housekeeping gene for normalization.

The sequences of rat-specific primers were as follows: Dkk1, 5′-CCCTCTGACCACAGCCATTT-3′ and 5′-AGAGCCTTCTTGCCCTTTGG-3′; and β-actin, 5′-ACAACCTTCTTGCAGCTCCTC-3′ and 5′-CTGACCCATACCCACCATCAC-3′.

All animal experiments were performed according to the Guide for the Care and Use of Laboratory Animals (NIH Publication No.85-23, revised 1996) and approved by the Research Ethics Committee of Sun Yat-Sen University. Male Sprague-Dawley (SD) rats (7–8 weeks of age, weighing 180–220 g, Certification No. 44005800007189, SPF grade) were provided by the Experimental Animal Centre of Guangzhou University of Chinese Medicine. After a 3-day quarantine, 16 rats were randomized into two groups, and received Dox and normal saline (NS), respectively. Over a period of 14 days, Dox was intraperitoneally injected in three equal dosages (each containing 5 mg/kg body weight) every 4 days and the final cumulative dose was 15 mg/kg body weight. Rats in the control group received an equal volume of NS administrated in parallel.

After the 2-week treatment, two-dimensional-guided M-mode echocardiography was performed with a Technos MPX ultrasound system (ESAOTE, SpAESAOTE SpA, Italy) and the basic hemodynamic parameters were measured. Then the animals were killed and their fresh hearts were rapidly removed. For morphometric measurement, histological cross sections (5 μm thick) of the heart tissues were fixed in 4% paraformaldehyde and then treated with hematoxylin-eosin (HE), Masson and Sirus Red stain. The rest of the tissues was quickly frozen in liquid nitrogen and then stored at −80°C for further assay.

SD rats were anesthetized with 7% chloral hydrate (0.35 mg/kg body weight) before endotracheal intubation. Then, left thoracotomy was conducted to expose the heart for direct gene delivery. A total of 200 μl Ad-Dkk1 or Ad-GFP (10⁹ particles) were injected into five or six sites in the left ventricular walls randomly. After the operation, the surgical wound was carefully sutured and gentamicin was given to prevent infection.

Following the manufacturer's protocol, a TdT-mediated dUTP nick-end labeling (TUNEL) apoptosis detection kit (Keygen Biotech, China) was used to detect DNA fragmentation in the heart sections, and the percentage of apoptotic cells were quantified as the ratio of TUNEL-labeled cells to total cells in the heart fractions. Hoechst 33342 staining was performed to observe nuclear condensation of H₉C₂ cells.

To monitor the mitochondrial membrane potential (Δψ_m), H₉C₂ cells were loaded with Rh123 (10 μg/ml, Sigma) or tetra-methylrhodamine ethyl ester (TMRE) (10 nmol/l, Invitrogen) at 37°C for 10 min. In addition, H₉C₂ cells were stained with 1 μmol/l Mitotracker Red (Invitrogen) at 37°C for 10 min for analysis of mitochondrial matrix swelling. Afterwards, the cells were washed three times and replaced by DMEM without Phenol Red. A laser scanning microscope (EVOS FL Auto, Life Technologies, Bothell, WA) was used to image the cells.

To morphologically assess mitochondria in the hearts of rats, tissues were fixed with 2.5% cold glutaraldehyde for 30 min and then 1% osmium tetroxide was added. Uranyl acetate and lead citrate were used to stain ultrathin sections for observation under an electron microscope (JEM-1400, JEOL Ltd., Japan) in different visual fields.

In order to detect the cell viability of H₉C₂ cells after treatments, MTS was supplemented into cell cultures in 96-well plates and was incubated away from light for 2 h at 37°C. Then the absorbance was measured with a microplate reader (TECAN, Switzerland; 490 nm wavelength) and the viability of cardiomyocytes was determined by the assessing the absorbance of control and treatment groups.

Cell apoptosis was assessed with an annexin V/propidium iodide (PI) apoptosis assay kit (BestBio, Shanghai, China). In short, the collected cells were washed twice (300–500 g, 5 min) with ice-cold phosphate-buffered saline (PBS), and then the 1× binding buffer from the kit (4×10⁶ cells/ml) was added to them. These mixtures were collected in a 2 ml culture tube, and then 5 μl annexin V and 10 μl PI were added into each tube in the dark for, respectively, 15 min and 5 min before the mixture was transferred into a 5 ml culture tube, which was used for testing. Finally, the cells were analyzed by flow cytometry (excitation at 488 nm; emission at 530 nm, EPICS XL, Beckman Coulter, USA). The proportion of apoptotic cells was determined as the percentage of positively stained cells.

Data are presented as means±s.e.m. Statistical analysis of two groups was performed with an unpaired Student's t-test, and one-way analysis of variance (ANOVA) with Bonferroni post-tests were used for multiple groups. A value of P<0.05 was considered statistically significant when analyzing the difference of groups.