All Sprague-Dawley (SD) rats in this study were purchased from Beijing Vital River Laboratory Animal Technology Company (Beijing, China). Twelve 8-week-old male SD rats with a mean weight of 273.7 ± 3.2 g were randomized into two groups: six rats were subjected to saline (Double Crane Pharmaceutical Co. Ltd, Wuhan, China) peritoneal injection (vehicle group), whereas six rats were intraperitoneally injected with CYP (Jiangsu Hengrui Medicine Co., Ltd. Lianyungang, China) at a dose of 100 mg/kg (CYP treatment group). Echocardiography (echo) and electrocardiography (ECG) were performed at different time points (0, 1, and 3 days).

Continuous recordings of heart rate were obtained with a surface ECG. Rats were anesthetized with 3% isoflurane and were subsequently fixed on a wooden board. ECG recording was performed using the limb lead. Three electrodes on an ECG monitor were inserted into the subcutaneous tissues of the rats' left and right shoulders and the right hind leg. The signal was amplified and recorded on a personal computer using an ECG Processor (EP-2B, Softron Beijing Incorporated, China) and stored on a data acquisition program (SP2006, Softron Beijing Incorporated, China).

ECG measurement was performed as described previously (18). ECG was performed using a Vevo 2,100 system (FUJIFILM VisualSonics, Canada), and the cardiac dimensions and functional parameters were measured. The tissue Doppler imaging (TDI) echo combined with ECG was used to measure the electromechanical coupling time at the lateral wall of the left ventricle as described previously (19).

NRCMs were isolated from newborn SD rats aged 1–2 days as described previously (20). These isolated NRCMs were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 100 U/ml penicillin/streptomycin, and maintained at 37°C in 5% CO₂.

H9 human embryonic stem cells were purchased from the Beijing Cellapy Biological Technology Company (Cellapy, China). H9 cells were cultured and differentiated into cardiomyocytes following previously described procedures (21). In brief, H9 cells were cultured on 35-mm dishes (Corning, USA) with PSCeasy hESC culture medium (Cellapy, Beijing, China). Cells were cultured to reach ~90% confluency and differentiated into ESC-CMs using a chemical method as described previously (22). Immunofluorescent staining with primary antibodies against TNNT2 (Santa Cruz, USA) and α-actinin (Abcam, UK) validated the purity of human cardiomyocytes.

Cells were cultured on glass slides, washed with PBS three times, fixed in 4% paraformaldehyde for 5 min, and then permeabilized with PBS containing 0.5% Triton X-100 (Sigma, USA) for 10 min. After 1 h of blocking with 5% BSA (Amresco, USA), the slides were incubated with primary antibodies followed by incubation with secondary antibodies. After the slides were washed, they were studied with a confocal fluorescence imaging microscope (DMI 4000B, Leica, Germany). The primary and secondary antibodies and their appropriate dilutions are listed in Supplementary Table 1.

MEA recording in cardiomyocytes was performed as described previously (23). In brief, 2 × 10⁴ cells were plated on CytoView MEA plates (Axion Biosystems, USA) pre-coated with 5% matrigel, followed by treatment with CYP at different concentrations (0 and 500 μmoL/L). The experimental data were acquired using a Maestro EDGE (Axion Biosystems, USA) according to the MEA operation manual.

Total RNA was extracted from NRCMs using the TRIzol reagent (Invitrogen) and subjected to reverse transcription (RT) and real-time PCR. The primers used are listed in Supplementary Table 2. RT was performed using a 2,720 Thermal Cycler (Applied Biosystems, USA). Real-time PCR was performed using a QuantStudio 3 apparatus (Applied Biosystems, USA).

Proteins were extracted from cells in RIPA lysis buffer (Solarbio, China) containing 1 mmol/L PMSF (Solarbio, China) and protease inhibitor cocktail (Bimake, China) for western blot analysis. In total, 50 μg of protein were subjected to SDS-PAGE, transferred to a PVDF membrane (Millipore, USA), and incubated with the primary antibodies. These primary antibodies and their appropriate dilutions are listed in Supplementary Table 1. The membrane was then incubated with HRP-conjugated goat anti-mouse IgG (1: 2000, ZSGB-BIO) or HRP-conjugated goat anti-rabbit IgG (1: 2000, ZSGB-BIO). GAPDH was used as a control. Protein levels were determined using the Immobilon® Western Chemiluminescent HRP substrate (Millipore, UK).

In the CYP treatment assays, 250, 500, and 750 μmol/L of CYP (Selleck, USA) were added to the cell complete culture medium for 2 or 4 days for NRCMs and for 2 or 5 days for hESCs-CMs. To impede the expression of METTL3 in cardiomyocytes, adenoviruses harboring the specific small interference RNA (siRNA) sequences of METTL3 were used individually to infect cardiomyocytes at an optimized MOI for 24 h, followed by treatment with 500 μmol/L CYP for an additional 24 h as the MEA assay. The siRNA and negative control (NC) sequences used are listed in Supplementary Table 2.

An RNA m6A dot blot assay was performed as previously described (24). In brief, 1.5 μg of total RNA was spotted onto a positively charged nylon-based membrane (GE Healthcare), blocked with 5% milk at room temperature for 2 h, and incubated with anti-m6A antibodies (1: 2000, Abcam) at 4 °C overnight and secondary antibodies (1: 3000, Abcam) at room temperature for 2 h. The same RNAs were spotted on the positively charged nylon-based membrane and stained with 0.02% methylene blue in 0.3M sodium acetate (pH 5.2), which ensured loading consistency among different samples.

High throughput m6A sequencing was performed with the support of Kangchen Biotech (Shanghai, China). Briefly, total RNA was extracted from NRCMs treated with 500 μmol/L CYP or DMSO (solvent control) for 48 h, followed by random fragmentation to 100–150 nucleotides using RNA fragmentation reagents. Fragmented RNA was subjected to m6A antibody immunoprecipitation following the Magna MeRIP m6A kit protocol (17-10499, Merk Millipore, USA) as described previously (25). An RNA library from immunoprecipitated RNA and input RNA was created on an Illumina HiSeq platform. Differential m6A peaks (fold change ≥1.5 and P ≤ 0.05) between CYP and solvent controls were used for gene ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis.

Ca²⁺ imaging in cardiomyocytes was performed as described previously (21). In brief, hESCs-CMs inoculate with the green fluorescent calcium-modulated protein (GCaMP) calcium sensor (H9-GCaMP-CMs) were seeded onto confocal dishes. Confocal microscope (Leica, TCS5 SP5, Germany) was used for intracellular calcium imaging. Spontaneous Ca²⁺ transients were recorded at 37°C and 5% CO₂ according to the standard line-scan methods (26, 27). A total of 8,192 line scans were acquired for a duration of 8.192 s. The imaging results were analyzed using the Image J and Igor pro software.

All statistical analyses were conducted using the SPSS 20.0 software (IBM Corp., USA) and Graphpad Prism software (version 8.0, GraphPad Software Inc., USA). The data are expressed as the mean ± standard error (SE). A Student's t-test detected the differences between groups. P values of ≤ 0.05 were considered as statistically significant.

To clarify the cellular significance of CYP in cardiomyocytes, we first performed the CCK-8 assay to examine the effects of CYP on the viability of cardiomyocytes. The result confirmed that CYP had no significant effect on NRCMs viability (Supplementary Figure 1A). However, we observed that the levels of atrial natriuretic factor (ANP) and brain natriuretic peptide (BNP) had increased after NRCMs were treated with 500 μmoL/L CYP for 48 h (Supplementary Figures 1B,C). These results suggested that CYP induced slight cardiotoxicity, but did not affect cell viability. CYP was closely associated with cardiac arrhythmias related to QT prolongation and the acute and chronic toxicity of chemotherapy (3). A prolonged QT interval is an important monitoring indicator for myocyte toxicity caused by anticancer agents according to the guidelines issued by the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (28). Therefore, we seeded cardiomyocytes on multielectrode array (MEA) probes to evaluate the effect of CYP on myocardial electrophysiological properties. The time between depolarization and repolarization is the FPD (Figure 1A), which corresponds to the QT interval in an ECG. Compared with solvent control (CON), the FPD of NRCMs treated with CYP increased at 12, 24, and 48 h (Figure 1B). Meanwhile, the impedance of the CON cells showed no significant changes (Figure 1C). We observed that the impedance of NRCMs treated with CYP decreased with time (Figure 1D). These observations indicated that CYP negatively regulated the rhythm and contractility of cardiomyocytes.

Furthermore, we used human embryonic stem cell-derived cardiomyocytes (hESCs-CMs) (Supplementary Figure 1D) to evaluate the effect of CYP on cellular viability. Consistent with this finding in NRCMs, CYP had no significant effect on the viability of hESCs-CMs (Supplementary Figure 1E) but increased the RNA levels of ANP and BNP (Supplementary Figures 1F,G). Intriguingly, exposing hESCs-CMs to CYP (500 μmol/L) for 9 days resulted in a significant increase in FPD (Figure 1E) and reduction of contractile amplitudes (Figure 1F). These results supported QT interval prolongation and cardiac contractile dysfunction in rats.

To further investigate the effect of CYP on cardiac functions, we used intraperitoneally injected CYP to treat the rats with CYP at a dose of 100 mg/kg, which was converted from the clinical dose for treatment of cancer. In vivo ECG recording data (Figure 2A) showed QT interval prolongation in rats after CYP-treatment for 1 day compared with that in rats administered saline (82.17 ± 1.70 vs. 65.17 ± 3.02 ms, P < 0.001; Figure 2B). Corrected QT interval (QTc) prolongation also showed the same variation as QT prolongation (199.83 ± 4.03 vs. 167.67 ± 6.83 ms, P < 0.01; Figure 2C). Further, the prolonged QT and QTc would restore to preadministration levels after CYP-treatment for 3 days (Supplementary Figures 2A,B). These results were consistent with the clinical side effects of CYP.

Electromechanical coupling disturbances were closely related to the long QT syndrome (29, 30). Hence, we further explored the effect of CYP on cardiac electrical and mechanical alterations. Four electromechanical coupling time courses (Qsb, Qst, Rsb, and Rst) were measured with TDI echo combined with ECG (Figure 2D). The measurement results showed that four electromechanical coupling time courses in CYP-treated rats were longer than those in saline controls (Figures 2E–H), particularly in terms of Qst and Rst courses (P < 0.05, Figures 2E,G). Moreover, ultrasound echocardiography (Figure 2I) showed that the fractional shortening percentage (FS%; Figure 2J) and left ventricular ejection fraction (LVEF; Figure 2K) were significantly lower in rats after CYP-treatment for 1 day. These results suggested that CYP induces cardiac electrical and mechanical alterations and decreases the excitation-contraction (E-C) coupling efficiency, leading to cardiac contractile dysfunction. Consistent with the results of QT and QTc, the prolonged electromechanical coupling time courses would restore after CYP treatment for 3 days (Supplementary Figures 2C–F) and the decreased FS and LVEF induced by CYP also showed a regression in CYP-treated rats after 3 days (Supplementary Figures 2G,H).

Previous studies demonstrated that junctophilin-2 (JPH2) that anchor the sarcoplasmic reticulum to T-tubules is the key regulator of Ca²⁺ influx between L-type Ca²⁺ channels (LCCs) and ryanodine receptors (RyRs) and E-C coupling in cardiomyocytes (31, 32), is reportedly associated with atrial fibrillation (33) and arrhythmias (34). Based on the phenomena observed in the above cell and animal experiments, we further investigated the effect of CYP on JPH2 expression in NRCMs and hESCs-CMs at different treatment time points. Notably, a dose-dependent reduction in JPH2 RNA and protein levels occurred in NRCMs treated with CYP for 2 or 4 days (Figures 3A,B). Similarly, different concentrations of CYP treatments decreased JPH2 both in RNA and protein levels at day 2 or 5 in hESCs-CMs (Figures 3C,D). Similarly, JPH2 downregulation occurred in heart tissues of rats treated with CYP (Supplementary Figure 3). These results suggested that CYP induced cardiac electrical and mechanical alterations and cardiac contractile dysfunction by decreasing the expression of JPH2.

To explore the underlying mechanisms involved in the suppression effects of CYP on JPH2 expression in cardiomyocytes, we further investigated the effect of CYP on miR-24 and miR-331 expressions, which were shown to inhibit the expression of JPH2 in our previous studies (31, 35). The real-time PCR analysis revealed that the expression of miR-24 and miR-331 did not significantly change in NRCMs after CYP treatment for 2 days (Supplementary Figures 4A,B). Therefore, it suggested that CYP decreased JPH2 expression through other transcriptional regulatory mechanisms.

N6-methyladenosine (m6A) is the most prevalent modification that widely exists in mRNAs, which is associated with post-transcriptional gene expression regulation (12), and mRNA stabiltity (36). We next investigated whether CYP plays an important role in m6A RNA methylation in NRCMs, considering that CYP can induce nucleic acid methylation. The m6A dot blot testing showed that total m6A levels significantly increased in NRCMs treated with CYP for 2 days (Figures 4A,B). Next, methylated RNA immune precipitation sequencing (MeRIP-seq) was performed to compare the global profiling of m6A target genes between solvent controls and CYP-treated NRCMs. As shown in Figure 4C, the sequence motif “GGAC” was highly enriched in m6A immunoprecipitated RNAs, consistent with the findings of previous studies (37, 38). We found 585 significantly increased m6A peaks distributed in 259 genes, whereas 277 genes had 548 statistically decreased m6A peaks in CYP-treated NRCMs relative to controls. Notably, we observed that reduced m6A peaks were mainly localized in the 5′ untranslated region (5′ UTR), whereas increased m6A peaks were distributed in the coding sequence (CDS) and 3′ untranslated region (3′ UTR; Figure 4D). The pie charts showed that these statistically differentially distributed m6A peaks were mainly noted in the CDS and 3′ UTR of genes in CYP-treated NRCMs regarding CON cells (Figure 4E). To explore the physiological and pathological significance of m6A modification after CYP treatment, we analyzed the KEGG pathway on the significantly altered m6A peaks. Our results showed that upregulated m6A peaks in the CYP-treated NRCMs were significantly related to the cAMP signaling pathway, adrenergic signaling in cardiomyocytes, calcium signaling pathway, GnRH signaling pathway and other dysregulation pathways in cancer (Figure 4F).

Furthermore, RNA sequencing was also performed on NRCMs treated with solvent control (CON) or CYP. Compared with CON, 369 genes were significantly downregulated, and 74 genes were upregulated in the CYP-treated group (Supplementary Figure 5A). The GO enrichment and KEGG analysis of the total DEGs showed that these DEGs were enriched in the NF-κB, TNF, and calcium signaling pathways (Supplementary Figures 5B,C). Remarkably, with the combined MeRIP-seq and RNA-seq results, we found upregulated m6A methylation sites in the 5′UTR and CDS of JPH2 mRNA, accompanied with the downregulation of JHP2 expression on the RNA level. These results suggested that CYP induces calcium signaling changes through JPH2 downregulation caused by increasing m6A modification.

Calcium is a fundamental regulator of E-C coupling and electrophysiological signaling in cardiac myocytes (39). The above MeRIP-seq and RNA-seq results showed that the calcium signaling pathway played an important role in CYP-induced cardiotoxicity. We next verified and analyzed the Ca²⁺ handling properties of hESCs-CMs with CYP treatment by using H9-GCaMP derived cardiomyocytes (H9-GCaMP-CMs) (21). Compared with CON, hESCs-CMs treated with different concentrations of CYP (250, 500, and 750 μmoL/L) demonstrated significant Ca²⁺ transient irregularities, which were virtually absent in CON cells. As shown in Figure 5A, H9-GCaMP-CMs treated with 250 μmoL/L CYP showed no significant changes in the rhythm of Ca²⁺ transient release and reabsorption regarding CON on day 2. As the treatment time prolonged, the cardiomyocytes exhibited longer Ca²⁺ transient durations on day 4 and slower beating rate, lower Ca²⁺ release amplitude, and longer transient durations on days 6 and 8 (Figures 5B–D). We noted a similar pattern of changes in H9-GCaMP-CMs treated with 500 or 750 μmoL/L CYP at different time points. On days 2 and 4, compared with CON, CYP-treated H9-GCaMP-CMs exhibited lower Ca²⁺ release amplitude (Figure 5B) and longer transient durations (Figure 5D). Interestingly, in addition to lower Ca²⁺ release amplitude, slower Ca²⁺ transient durations were observed to occur in H9-GCaMP-CMs treated with CYP at dose of 500 or 750 μmoL/L(Figures 5A,D). On day 8, compared with CON, H9-GCaMP-CMs treated with low, medium and high concentrations of CYP exhibited lower Ca²⁺ release amplitude, and longer time to peak and transient durations (Figures 5A–D). These observations indicated that CYP could induce abnormal electrophysiological and contractile alterations in cardiomyocytes, consistent with the findings of the RNA sequencing and clinical data.

To understand whether m6A RNA methylation plays an important role in the suppression effects of CYP on JPH2 expression in NRCMs, we further investigated the effect of CYP on m6A writers (METTL3, METTL14, and WTAP) and erasers (FTO, and ALKBH5) in CYP-treated cardiomyocytes. Intriguingly, exposing NRCMs to different doses of CYP (250, 500, and 750 umoL/L) for 2 or 4 days resulted in a significant increase in the expression of METTL3 (Figures 6A,B), whereas the expression of METTL14, WTAP, FTO, and ALKBH5 was not significantly altered (Figures 6A,C–F). Similar results were observed in rats treated with CYP. The RNA level of METTL3 increased in hearts (Supplementary Figure 6). These results suggested that CYP induces the m6A methylation of JPH2 mRNA through increasing METTL3 expression, leading to downregulation of JPH2 expression.

To investigate the biological effect of METTL3 on the reduction of JPH2 expression, we designed small interfering RNAs to silence METTL3 in NRCMs. Intriguingly, silencing METTL3 (Figure 6G) resulted in an increase in JPH2 expression (Figure 6H). Additionally, CYP induced JPH2 downregulation in NRCMs transfected with NC sequences. However, there was no reduction effect of CYP on JPH2 expression in si-METTL3 cardiomyocytes (Figure 6H). These data indicated that CYP decreased JPH2 expression by upregulating METTL3.

To determine whether the disruption of METTL3 affects cardiac electrical and mechanical alterations in cardiomyocytes, we performed MEA in si-METTL3 and NC cardiomyocytes treated with CYP. Compared with NRCMs treated with NC, the FPD increased in NRCMs treated with CYP for 1 day (Figure 7A), whereas knock out of METTL3 significantly eliminated the increased FPD induced by CYP (Figure 7A). Similarly, the MEA results showed that the action potential duration (APD) (Figure 7B) was prolonged in NRCMs treated with NC after CYP treatment for 1 day (Figures 7C,D). However, the prolonged APD did not occur in si-METTL3 cardiomyocytes treated with CYP compared with the NRCMs treated with solvent control (Figures 7C,D). The above results demonstrated that the disruption of METTL3 eliminated the electrical alterations of cardiomyocytes induced by CYP.