Doxorubicin‐induced cardiotoxicity is maturation dependent due to the shift from topoisomerase IIα to IIβ in human stem cell derived cardiomyocytes

Abstract Doxorubicin (DOX) is widely used to treat various cancers affecting adults and children; however, its clinical application is limited by its cardiotoxicity. Previous studies have shown that children are more susceptible to the cardiotoxic effects of DOX than adults, which may be related to different maturity levels of cardiomyocyte, but the underlying mechanisms are not fully understood. Moreover, researchers investigating DOX‐induced cardiotoxicity caused by human‐induced pluripotent stem cell‐derived cardiomyocytes (hiPSC‐CMs) have shown that dexrazoxane, the recognized cardioprotective drug for treating DOX‐induced cardiotoxicity, does not alleviate the toxicity of DOX on hiPSC‐CMs cultured for 30 days. We have suggested that this may be ascribed to the immaturity of the 30 days hiPSC‐CMs. In this study, we investigated the mechanisms of DOX induced cardiotoxicity in cardiomyocytes of different maturity. We selected 30‐day‐old and 60‐day‐old hiPSC‐CMs (day 30 and day 60 groups), which we term ‘immature’ and ‘relatively mature’ hiPSC‐CMs, respectively. The day 30 CMs were found to be more susceptible to DOX than the day 60 CMs. DOX leads to more ROS (reactive oxygen species) production in the day 60 CMs than in the relatively immature group due to increased mitochondria number. Moreover, the day 60 CMs mainly expressed topoisomerase IIβ presented less severe DNA damage, whereas the day 30 CMs dominantly expressed topoisomerase IIα exhibited much more severe DNA damage. These results suggest that immature cardiomyocytes are more sensitive to DOX as a result of a higher concentration of topoisomerase IIα, which leads to more DNA damage.

Children are more susceptible to the cardiotoxic effects of DOX than adults, especially children younger than 4 years. [8][9][10][11][12] This may be due to the level of maturity of cardiomyocyte. It has been shown in previous studies that DOX-induced apoptosis gradually decreases during cell maturation. [13][14][15] However, the underlying mechanism remains unclear due to a lack of in vitro cellular maturation models.
Since human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) can mimic many aspects of human cardiomyocytes, such as contractile function, cardiac genes expression and electrophysiological phenotypes, they have become an important technology in human cardiovascular disease modelling. 16 Moreover, hiPSC-CMs provide a drug-testing platform for large-scale screening of compounds. The US Food and Drug Administration initiated a new program, the Comprehensive In Vitro Proarrhythmia Assay (CiPA), to assess the clinical cardiac safety of compounds prior to Phase I clinical trials. 17 One essential requirement of the CiPA is to assess the electrophysiological effects of new drugs on hIPSC-CMs assays. In 2016, Burridge et al established the DOX-induced cardiomyocyte injury model based on hiPSC-CMs. 18 However, they found that dexrazoxane (DEX), a recognized cardioprotective drug used for treating DOX-induced cardiotoxicity, whose properties were confirmed by animal experiments and clinical trials, could not alleviate the toxicity caused by DOX on hiPSC-CMs cultured for 30 days. 18 The immaturity of hiPSC-CMs might be the leading cause of different responses to the same drug in the heart in vivo and in vitro.
To better understand the mechanism of DOX-mediated cardiotoxicity in cardiomyocytes of different maturation levels, we chose 30-day-old (day 30) and 60-day-old (day 60) hiPSC-CMs to represent immature and relatively mature cardiomyocytes, respectively.
We suggest that the level of maturity may affect the modelling of DOX-induced cardiotoxicity.
Cell lines were used between passages 20 and 85. All cultures were routinely tested for mycoplasma using a MycoAlert Plus Kit (Lonza).
The media were changed on day 4 and every other day for CDM3.
Contracting cells were observed from day 8. On day 10, the media were changed to a purification medium made using RPMI 1640 (no glucose) (Corning), 500 µg/mL recombinant human albumin, and 213 µg/mL L-ascorbic acid 2-phosphate. The medium was replaced with RPMI 1640 (Corning) supplemented with 500 µg/mL recombinant human albumin 48 hours before the experiment in order to avoid antioxidant effects.

| Flow cytometry cardiac differentiation efficiency
For assessment of cardiac differentiation efficiency, cells on day 15 of differentiation were dissociated using Cell Dissociation Solution (Cellapy) for 25 minutes at 37°C and transferred to flow cytometry tubes (BD Biosciences). Cells were then fixed with 4% paraformaldehyde (PFA) for 10 minutes, permeabilized with 0.1% saponin for 20 minutes, and stained using 1:100 mouse monoclonal IgG1 TNNT2 (Santa Cruz) for 30 minutes at RT. Isotype control Alexa Fluor 594 mouse IgG (Life Technology) was used to establish gating. Cells were then analysed using a flow cytometre.

| Cardiomyocyte plating and drug treatment
Cells were separately cultured to 25 and 55 days and then immediately used for experiments. The hiPSC-CMs were dissociated using Cell Dissociation Solution (Cellapy) for 25 minutes at 37°C, centrifuged at 1000 rpm for 2 minutes and plated onto coverslips, 12-well cell culture plates or 96-well culture plates coated with matrigel (Corning) 5 days before experimentation. Doxorubicin hydrochloride (Sigma-Aldrich) was resuspended to stock solution 1000 μmol/L in PBS. For treatments on hiPSC-CMs, 2.5 µmol/L DOX was diluted in P11a and cells were treated for 24 hours. For DEX treatment, hiPSC-CMs were co-treated with 100 μmol/L of DEX (Selleck) with DOX. For N-Acetyl-L-Cysteine (NAC) treatment, hiPSC-CMs were co-treated with 1 mmol/L NAC (Sigma-Aldrich) with DOX.

| Quantitative real-time PCR
Total mRNA was isolated using TRIzol, and 1 µg of the mRNA was used to synthesize cDNA using the GoScript Reverse Transcription System (Promega). A concentration of 0.25 µL of the reaction mixture was used to quantify gene expression by qPCR using SYBR ® Premix Ex Taq (TaKaRa). Quantitative real-time (RT) PCR conditions were as follows: initial denaturation at 94°C for 4 minutes followed by 40 cycles at 94°C for 1 minute, annealing for 1 minute at 56 and 72°C for 1 minute, with a final extension at 72°C for 10 minutes. Expression values were normalized to the average expression of GAPD. Primer sequences are shown in Table S1.

| Immunofluorescence staining
The cells were plated on 20 mm coverslips and were fixed with 4% PFA for 20 minutes. Then, after washing with PBS three times for

| Cell viability assay
The hiPSC-CMs were passaged and cultured in 96-well plates at 8×104 cells/well. After DOX treatment for 24 hours, 10 µL of CCK-8 (Cell Counting Kit-8, Dojindo) was added directly to each well in the 96-well plates, which were then incubated at 37°C for 3 hours; absorbance was read at 450 nm.

| TUNEL staining
Cells were stimulated with DOX of different concentrations for 36 hours. Apoptosis was measured using a TUNEL assay kit (Promega). The cells were plated on 20 mm coverslips, fixed with 4% PFA for 20 minutes at RT and washed twice for 5 minutes with PBS. Next, cells were treated with 0.2% Triton X-100 for 30 minutes and washed twice for 5 minutes with PBS. Then the cells were pre-incubated with terminal deoxynucleotidyl transferase buffer for 10 minutes at room temperature. In the absence of light, the reaction buffer was added to the cells, and the coverslips were incubated in a humid atmosphere at 37°C for 1 hour. Coverslips were washed again three times for 5 minutes, mounted with Fluoroshield Mounting Medium with DAPI, imaged using the confocal microscope and analysed using Imagej software.

| Measurement of mitochondrial transmembrane potential (Δψm) loss
Mitochondrial depolarization was monitored with the potentiometric dye JC-1 using the Mitoprobe assay kit (Invitrogen) in accordance with the manufacturer's instructions. JC-1 accumulates in polarized mitochondria with a resting membrane potential and fluoresces red. However, during ΔΨm loss, JC-1 aggregates are released from the mitochondria, which results in a green fluorescence. Thus, in order to assess mitochondrial depolarization, treated cells were stained with 2 μmol/L JC-1 for 20 minutes at 37°C at 5% CO2, washed and resuspended in fresh media; and then red and green fluorescence were monitored using High Content Analysis (MetaXpress).

| Mitochondrial reactive oxygen species assay
In order to determine the levels of mitochondrial superoxide using fluorescence microscopy, the cells were grown in a 20 mm glass slide.
Following treatment, MitoSOX (Invitrogen) reagent was dissolved in dimethyl sulfoxide (5 mmol/L), diluted to 5 µmol/L in serum-free medium, and was then added to the cells; this was followed by incubation for 10 minutes at 37°C; the cells were then washed twice with PBS. Subsequently, the cells were fixed with 4% PFA for 20 minutes at RT, washed again, mounted with Fluoroshield Mounting Medium with DAPI, imaged using the confocal Microscope and analysed using High Content Analysis.

| Intracellular reactive oxygen species assay
The intracellular reactive oxygen species (ROS) level was determined using 2', 7'-dichlorodihydrofluorescein diacetate (DCFH-DA) (Beyotime). The cells were grown in the 20 mm glass slide. Following treatment, DCFH-DA reagent was diluted to 10 µmol/L in serum-free medium and was then added to the cells; this was followed by incubation for 20 minutes at 37°C; the cells were then washed twice with PBS. Subsequently, the cells were fixed with 4% PFA for 20 minutes at RT, washed again, mounted with Fluoroshield Mounting Medium with DAPI, imaged using the confocal microscope and analysed using High Content Analysis.

| Mitochondria staining
In order to determine the distribution and level of mitochondria, the cells were grown in the 20 mm glass slide. Mitotracker (Invitrogen) reagent was diluted to 1 µmol/L in serum-free medium, and was then added to the cells; this was followed by incubation for 30 minutes at 37°C; the cells were then washed twice with PBS. Subsequently, the cells were fixed with 4% PFA for 20 minutes at RT, washed again, mounted with Fluoroshield Mounting Medium with DAPI, imaged using the confocal microscope and analysed using High Content Analysis. (mean ± SEM). Comparisons were conducted using the one-way ANOVA test, followed by either the All Pairwise Multiple Comparison Procedures (Sidak) method or an unpaired, two-tailed Student's t test.

| Generation of hiPSC-CMs
In an effort to ensure reproducibility, we cultured hiPSCs under chemically defined conditions, and hiPSCs showed normal colony morphology ( Figure 1A,B). Then we examined the pluripotency of hiPSCs using immunofluorescence staining for pluripotent markers OCT4, SSEA-4, NANOG and TRA-1-60 ( Figure 1D). Using a small molecule-based directed differentiation method ( Figure 1E), we were able to generate high purity CMs from the hiPSCs cultured as a monolayer. The expression of multipotency related genes POU5F1, SOX2, NANOG, LIN28 and FOXD3 decreased significantly after differentiating into hiPSC-CMs ( Figure 1C). Approximately 8 days after induction, the differentiated cells started to contract spontaneously ( Figure 1F,G). Flow cytometry analysis indicated that the hiPSC-CMs were highly pure cellular populations, with more than 95% of cells expressing the cardiac marker TNNT2 ( Figure 1H).

| hiPSC-CMs matured over time when cultured in vitro
We found that the maturity of hiPSC-CMs increased with culture time, so we chose hiPSC-CMs that were 30 days old and 60 days old; we compared their characteristics in terms of molecular expression, myofilament structure, gap junctions and calcium transients.
We first compared the changes in the gene expression profiles of structural and functional genes using quantitative real-time PCR.
The expression of sarcomeric genes (such as MYL2, LRRC1 and After measuring the spontaneous calcium transient, we found that the amplitude of day 60 CMs was higher than day 30 CMs  Figure S2A,B). These results suggested that the day 60 hiPSC-CMs were more mature in terms of structural and functional phenotypes than the day 30 hiPSC-CMs.

| Maturation decreased the cytotoxicity of DOX in hiPSC-CMs
We have suggested that the degree of maturity influences the cells' response to DOX and affects the modelling of DOX-induced cardiotoxicity. Therefore, we measured the cell viability of

| Maturation led to more ROS production in hiPSC-CMs
Doxorubicin-induced cardiotoxicity is known to have two major  (Figure 4F,G). In order to determine the impact of ROS overproduction on cell viability reduction, we analysed the cell viability following co-treatment with antioxidant NAC. Our results showed that NAC could restore cell viability in both groups, but more evidently so in the day 60 group ( Figure 4H).
Consequently, we may conclude that DOX can induce ROS production in hiPSC-CMs in a dose-dependent manner and that DOX leads to more ROS production in the day 60 hiPSC-CMs.

| DOX led to more severe DNA damage in less mature hiPSC-CMs through TOP2α
However, the above results do not explain the phenomenon of more severe toxicity in the day 30 hiPSC-CMs, so we considered the other pathway of DNA damage. DOX, targeting Top2, plays its cytotoxic role due to the formation of a Top2-DOX-DNA complex.
Top2 plays an important role in maintaining the DNA topology. 28 There are two Top2 isozymes in human cells: Top2α and Top2β. 29 Top2α is only expressed in proliferative cells and tumour cells, and is highly expressed in the G2/M period of cell cycle. DOX is thought to bind to Top2α, which has a highly elevated expression in cancer cells. 30 By contrast, Top2β is present in all cells. It is worth noting that adult mammalian cardiomyocytes express TOP2β but contain no detectable TOP2α. There is substantial evidence that Top2β is predominantly responsible for DOX-induced cardiotoxicity via DNA damage. 31 The role of Top2β was verified by a previous study showing that cardiomyocyte-specific deletion of Top2β protects mice from the development of DOX-induced progressive heart failure. 32 We assessed the level of double-stranded DNA damage by detecting phosphorylated H2A histone family member X (γ-H2A.X).
We observed a dose-dependent increase in DNA damage, which was significantly higher in the day 30 hiPSC-CMs ( Figure 5A,B).
Given that the day 30 hiPSC-CMs were more proliferative than the day 60 cells, it was necessary to explore the gene expression  Previous studies have confirmed that as the heart of a human or a mouse develops, the expression of TOP2A gradually decreases, and the expression of TOP2B generally increases. [35][36][37] We speculate that the sensitivity of age-related DOX-induced cardiotoxicity may be related to the different levels of expression of these TOP2 genes, which might help guide clinical medication and prevent DOX cardiotoxicity.
We aimed to clarify this phenomenon, rather than studying DOX-induced cardiotoxicity itself, and so we used the hiPSC-CMs of a healthy individual instead of those taken from patients suffering from DOX-induced cardiomyopathy. However, it is better to choose hiPSC-CMs from patients who are suffering from DOX-induced cardiomyopathy for pathogenesis studies and drug screening. The technology used for hiPSC-CM analyses has developed rapidly over the past 10 years. HiPSC-CM analyses take into account the overall genetic background of a specific individual, and thus have many advantages over animal models. However, hiPSC-CMs do not match the phenotype of adult cardiomyocytes in terms of maturity. There are a number of ways to promote the maturation of hiPSC-CMs, but these methods only allow F I G U R E 6 Schematic illustration of doxorubicin (DOX)-induced cardiotoxicity in human-induced pluripotent stem cellderived cardiomyocytes (hiPSC-CMs) cultured 30 d and 60 d respectively.
The day 60 hiPSC-CMs were relatively mature, with well-developed and distinct sarcomeres, defined T-tubules, and more abundant connexin-43 and mitochondria, compared to the day 30 hiPSC-CMs. When treated with Doxorubicin (DOX), the day 60 hiPSC-CMs produced more reactive oxygen species (ROS) and better simulated the physiological processes of ROS overproduction, although both groups responded to N-Acetyl-L-Cysteine. In addition, the day 30 hiPSC-CMs mainly expressed TOP2α rather than TOP2β, and failed to mimic DOX-induced cardiotoxicity in the pathophysiological process of DNA damage hiPSC-CMs to approach the phenotype of late foetal cardiomyocytes, which is far from the phenotype of adult cardiomyocytes. 23 This is an emerging technology, and there remains considerable room for improvement.