Melphalan treatment induces cell death and apoptosis in hiPSC-CMs.
To investigate the cardiotoxicity of melphalan, we generated enriched hiPSC-CMs (Fig. S1) and treated them with melphalan at 4 doses ranging from 0 to 20 µM; the highest dose was slightly above the Cmax of melphalan (15.4 µM) [23]. hiPSC-CMs exposed to 20 µM melphalan contracted weakly after 24 h compared with other groups. After 48 h of treatment, many cells treated with 20 µM melphalan stopped contracting with many turning into round shape and detaching from the plate surface, indicating cell dysfunction and death. Cells treated with 10 µM melphalan presented similar morphology during 3 to 5 days. As shown in Fig. 1a, fewer cells remained following the treatment with 10 and 20 µM melphalan for 5 days.
In order to quantify the cell death, we first validated and optimized two cell viability assays, CellTiter-Blue and CellTiter-Glo 3D Cell Viability Assays, which were reliable and sensitive for the estimation of cell numbers of hiPSC-CMs (Fig. S2). Next, we examined cell viability in cultures after 3- and 5-days melphalan treatment. Based on CellTiter-Blue Cell Viability Assay, 10 µM melphalan treatment for 3 days caused a 15% loss of cells compared with no melphalan treatment and 20 µM melphalan treatment caused a 29% loss. When the treatment duration extended to 5 days, melphalan treatment exacerbated the cell loss, which increased to 28% for 10 µM and 68% for 20 µM (Fig. 1b). The dose-dependent cell death induced by melphalan was validated by CellTiter-Glo 3D Cell Viability Assay (Fig. 1c).
To evaluate if the reduced cell viability in melphalan treated hiPSC-CMs was associated with apoptosis at the early stage, we treated hiPSC-CMs with various doses of melphalan for 24 h and measured activated caspases 3 and 7. As shown in Fig. 1d and e, relative mean fluorescence intensity (MFI) of Caspase3/7 significantly elevated in cells exposed to melphalan in a dose-dependent manner. To further confirm this phenomenon, we examined the expression of apoptosis-related genes by qRT-PCR in cells exposed to melphalan for 3 days. The level of anti-apoptosis gene BCL2 detected was similar in all the groups, but the level of pro-apoptosis gene BAX detected was 5 times higher in hiPSC-CMs treated with 10 µM melphalan compared with no melphalan treatment and 8 times higher in hiPSC-CMs treated with 20 µM melphalan (Fig. 1f).
Melphalan treatment of hiPSC-CMs results in Ca2+ handling defect and alters expression of genes encoding calcium channels and sarcomeric proteins.
Ca2+ is the critical link between electrical excitation and mechanical contraction. Carefully regulated transient rises and reductions of cytosolic Ca2+ correspond to the electrical signals that pervade the heart and control each cycle of contraction and relaxation of CMs. To investigate the effect of melphalan treatment on CM function, we assessed intracellular Ca2+ transients in hiPSC-CMs treated with various doses of melphalan for 3 days. In all conditions, as the representative traces shown in Fig. 2a, two categories of whole cell Ca2+ release events were observed: normal and abnormal Ca2+ transients. Cells were categorized as normal if the Ca2+ transients had mostly consistent amplitudes and rhythmicity, typical cardiac Ca2+ transient morphology (i.e. rapid upstroke and decay kinetics), and no obvious spontaneous Ca2+ release between transients (Fig. 2a-ⅰ). Cells were categorized as abnormal if they exhibited oscillations of the diastolic Ca2+ signal (Fig. 2a-ⅱ and ⅲ), unrecognizable single transient morphology (Fig. 2a-ⅳ), or notable inconsistent amplitudes or beat periods (Fig. 2a-ⅴ, ⅵ). Using these criteria, we counted the numbers of cells exhibiting normal or abnormal Ca2+ transients and calculated the proportion of each category for each culture condition (Fig. 2b). In hiPSC-CMs without melphalan treatment, the majority of the cells exhibited normal Ca2+ transients; whereas in hiPSC-CMs treated with melphalan, the percentage of cells exhibiting abnormal Ca2+ transients increased in a dose-dependent manner. Specifically, 48% of the cells showed abnormal Ca2+ transients when treated with 1 µM melphalan; 57% of the cells showed abnormal Ca2+ transients when treated with 10 µM melphalan; and 67% of the cells showed abnormal Ca2+ transients when treated with 20 µM melphalan. In addition, the treatment of hiPSC-CMs with melphalan at 10 and 20 µM significantly decreased Ca2+ transient amplitude without affecting Ca2+ transient duration compared with no melphalan treatment (Fig. 2c): the amplitude was reduced by 44% in cells exposed to 10 µM melphalan, and 77% in cells exposed to 20 µM melphalan. The maximum upstroke and decay speeds of Ca2+ transients were also significantly decreased in melphalan-treated hiPSC-CMs (Fig. 2c): the maximum upstroke and decay speeds was reduced by 29%-34% in cells exposed to 1 µM melphalan, 44%-47% in cells exposed to 10 µM melphalan, and 67%-74% in cells exposed to 20 µM melphalan. These observations suggest that exposure of hiPSC-CMs to melphalan results in intracellular Ca2+ handling dysfunction in a dose-dependent manner.
We next quantified the expression of genes encoding the components of calcium channels and sarcomere which are crucial to CM function by qRT-PCR in hiPSC-CMs under the above conditions (Fig. 2d). The expression of calcium channel proteins encoding genes RYR2 and CACNA1C was reduced in cells treated with 20 µM melphalan compared with no melphalan treatment. The expression of TNNI1 and MYH7 was also lower in 10 and 20 µM melphalan-treated cells. The expression of light chain of myosin encoding genes MYL2 decreased by 52% in 20 µM melphalan-treated cells but that of MYL7 increased by 63%.
Melphalan treatment alters protein expression levels of hiPSC-CMs identified by proteomic analysis.
To further evaluate the molecular alteration induced by melphalan and to investigate potential mechanisms of melphalan-induced cardiotoxicity, we treated hiPSC-CMs with or without 20 µM melphalan for 3 days and performed proteomic analysis to compare protein expression changes. 68 proteins were significantly upregulated and 185 downregulated in melphalan treated hiPSC-CMs (Fig. 3a). GO analysis showed that melphalan treatment up-regulated proteins associated with response to wounding, stress, and stimulus (Fig. 3c). The up-regulation of proteins involved in apoptotic process and cell death was consistent with the aforementioned results based on cell viability and apoptosis detection at cellular level. More intriguingly, ROS seemed to play an important role due to several significantly enriched GO terms from the up-regulated proteins, such as ROS metabolic process, response to oxidative stress, and response to oxygen-containing compound. In addition, the down-regulated proteins were also related to cell adhesion, cardiovascular system development, actin filament-based process, and heart contraction (Fig. 3c).
Melphalan treatment causes oxidative stress in hiPSC-CMs.
To validate the finding from the proteomic experiments and the hypothesis that oxidative stress could be an underlying mechanism of cardiotoxicity caused by melphalan, we treated hiPSC-CMs with various doses of melphalan for 3 days and measured intracellular ROS by H2DCFDA probe and mitochondrial ROS by MitoSOX probe. As shown in Fig. 4a, increased ROS signals were detected in the cells treated with melphalan in a dose-dependent manner. The relative level of mitochondrial oxidative stress was 0.7 times higher in cells treated with 10 µM melphalan compared with no melphalan treatment, and 1.3 times higher in cells treated with 20 µM melphalan (Fig. 4b).
We next examined the expression of oxidative stress-related genes by qRT-PCR in hiPSC-CMs exposed to melphalan for 3 days. The expression of superoxide dismutase family of genes (SOD1, SOD2, and SOD3), reductase encoding genes (PRDX5 and NQO2), and glutathione related genes (GSR, GPX1) was significantly elevated in cells treated with 20 µM melphalan compared with no melphalan treatment (Fig. 4c, Fig. S3). Particularly, SOD3 level detected was 5.5 times higher in cells exposed to 10 µM melphalan compared with no melphalan treatment and even higher (9.3 times) in cells treated with 20 µM melphalan. These results indicate that melphalan induces ROS production and increases oxidative stress in hiPSC-CMs in a dose dependent fashion.
NAC mitigates cell loss and mitochondrial ROS production in hiPSC-CMs under melphalan treatment.
To further evaluate if ROS production plays a crucial role in melphalan-induced cardiotoxicity, we treated hiPSC-CMs with 0, 10, and 20 µM melphalan in combination with or without 1 mM of ROS scavenger NAC concomitantly, for 3 days, and measured cell viability and ROS production. The dose selection of NAC was based on previous studies in which 1 mM of NAC effectively attenuated the ethanol- and doxorubicin-induced oxidative stress in hiPSC-CMs [22, 28]. As shown in Fig. 5a, treatment of cells with NAC prevented the cell loss caused by melphalan treatment. Furthermore, NAC supplementation dramatically decreased intracellular ROS by 16% in 10 µM melphalan-treated hiPSC-CMs and 37% in 20 µM melphalan-treated hiPSC-CMs (Fig. 5b). More strikingly, treatment of cells with NAC mitigated mitochondrial oxidative stress caused by melphalan treatment to the level similar to that of no melphalan treatment (Fig. 5c). In addition, we observed that hiPSC-CMs exposed to melphalan with NAC supplementation contracted more powerfully and kept better morphology than those without NAC supplementation.
NAC attenuates the alteration of hiPSC-CM beating indexes caused by melphalan treatment.
Normal contraction and relaxation of CMs are essential to maintain normal organ function. To identify the influence of melphalan treatment and NAC supplementation on CM contractility, we recorded spontaneous beating and quantified beating indexes in hiPSC-CMs treated with 0, 10, and 20 µM melphalan with or without 1 mM NAC supplementation for 3 days. As shown in Fig. 6a, recorded traces presented the velocities of contraction and relaxation of each CM beating during 30 s periods under all conditions. We found that treatment of hiPSC-CMs with melphalan at 10 and 20 µM significantly decreased maximum contraction and maximum relaxation without distinct beating rate alteration compared with no melphalan treatment (Fig. 6b). Specifically, the maximum contraction and relaxation in cells exposed to 10 µM melphalan was reduced by 30%-35%, which further dropped by 30% more in cells exposed to 20 µM melphalan. However, with 1 mM NAC supplementation the maximum contraction and maximum relaxation in melphalan-treated cells retained nearly similar levels to the no melphalan treatment. These findings were consistent with microscopic observations of cell behaviors. In addition, we observed an increase in the incidence of irregular beating based on variation of contraction and relaxation velocity, from less than 6% in cells without melphalan treatment to 17%-28% in cells treated with 10 µM melphalan and 56%-61% in cells treated with 20 µM melphalan (Fig. 6c). NAC supplementation attenuated the degree of irregular beating caused by melphalan treatment: to 10%-13% in the 10 µM melphalan-treated cells and 35%-39% in the 20 µM melphalan-treated cells (Fig. 6c). Taken together, these results indicate that melphalan treatment of hiPSC-CMs impairs CM contractility, which could be ameliorated by NAC supplementation.
NAC ameliorates melphalan-induced alteration of hiPSC-CM transcriptomic profiles characterized by RNA-Seq analysis.
To further evaluate the molecular changes associated with melphalan-induced cardiotoxicity and rescue by NAC supplementation, we performed RNA-Seq to analyze global transcriptome profiles of hiPSC-CMs treated with vehicle (Control group), 20 µM melphalan (Mel group), and 20 µM melphalan with 1 mM NAC (Mel+NAC group), respectively, for 3 days. As detected by RNA-Seq, 12,201 genes were commonly expressed in all three groups, and 309 genes were expressed in control and Mel+NAC groups but not in Mel group (Fig. S4a). As shown in Fig. 7a, treatment of the cells with melphalan resulted in up- and down-regulation of 2,097 genes (Mel vs. Control), whereas NAC supplementation to melphalan-treated cells reduced the number of up- and down-regulated genes to 709 (Mel+NAC vs. Control). Interestingly, more genes were down-regulated than up-regulated by the treatment of melphalan (1,422 vs. 675 in Mel vs. Control), whereas NAC supplementation resulted in more genes being up-regulated than down-regulated (567 vs. 66 in Mel+NAC vs. Mel). As shown in Table S5, among the top 10 up-regulated genes by melphalan treatment, 4 were direct p53 effectors (CDKN1A, EDXR, TNFRSF10C, and GDF15). Among the top 10 down-regulated genes by melphalan treatment, 5 were correlated to cell adhesion (CDH13, CNTN1, SDK1, CTNND2, and PARD3B).
Given that more genes were down-regulated by melphalan treatment and more genes were up-regulated by NAC supplementation, we performed GO analysis of DEGs in these groups and examined the degree of the GO terms in these groups overlapped. As shown in Fig. 7b, Tables S5 and S6, melphalan treatment dramatically down-regulated the expression of genes associated with extracellular matrix (121 genes), muscle contraction (78 genes), and synaptic membrane (80 genes). Interestingly, NAC supplementation up-regulated many of the genes involved in these GO terms (extracellular matrix: 83 genes, muscle contraction: 43 genes, and synaptic membrane: 27 genes).
We also examined the signaling pathways regulated by melphalan treatment and NAC supplementation on the basis of KEGG enrichments (Table S5, S6). Noteworthily, several pathways were both regulated by melphalan treatment and NAC supplementation, including apoptosis pathway, p53 signaling, transforming growth factor (TGF)-β signaling and cytokine-cytokine receptor interaction. As shown in Fig. 7c, the genes of apoptosis (e.g., BAX and TNFRSF10C) and p53 signaling pathway (e.g., FAS and CDKN1A) were mostly up-regulated by melphalan (Mel vs. Control), but they were mostly down-regulated by NAC supplementation (Mel+NAC vs. Mel). Those in the TGF-β signaling pathway (e.g., LEFTY2 and THSD4) and cytokine-cytokine receptor interaction (e.g., BMP6 and BMP10) were mostly down-regulated by melphalan treatment (Mel vs. Control), but they were mostly up-regulated by NAC supplementation (Mel+NAC vs. Mel).
In addition, we compared the regulation of genes involved in oxidative stress, cardiac muscle contraction, and cardiac conduction following melphalan treatment and NAC supplementation. As shown in the heatmap (Fig. 7d), the up- and down-regulation of genes involved in oxidative stress (e.g., DUOX2 and NOX4) following melphalan treatment (Mel vs. control) was attenuated with NAC supplementation (Mel+NAC vs. Control). Similarly, the up- and down-regulation of genes involved in cardiac muscle contraction (e.g., Ca2+ handling proteins CACNA1C, RYR2 and CASQ2 and cardiac contractile proteins TNNC1, ACTC1, TNNC1) and cardiac conduction (e.g., ATP2B2 and ABCC9) following melphalan treatment (Mel vs. Control) was attenuated by NAC supplementation (Mel+NAC vs. Control) (Fig. 7d, Fig. S4c).
Finally, we compared the results of proteomics and RNA-Seq analysis. There were 40 genes recognized as DEGs in both analyses, of which 10 were up-regulated and 30 were down-regulated (Fig. 3b). Intriguingly, 6 of the up-regulated genes were involved in the p53 signaling pathway (e.g., CDKN1A and RRM2B), and 11 of the down-regulated genes were relevant to muscle structure (e.g., TTN and TBX20).