ICG-001 Provides Cardioprotection Against Doxorubicin-Induced Cardiotoxicity and Enhances Cancer Cytotoxicity

Visual Abstract

Highlights

• •This study established a mitochondria-rich patient-derived hiPSC-CM model of DOX-induced cardiotoxicity, which recapitulates patient response to treatment.
• •ICG-001 can protect against DOX-induced cardiotoxicity and enhance tumor cytotoxicity by targeting DRP1 and Wnt signaling in CMs and cancer, respectively.
• •These results support future clinical trials involving ICG-001 as an adjunct treatment that may improve the safety and efficacy of chemotherapy for cancer patients.

Summary

Doxorubicin is effective against cancer but can cause doxorubicin-induced cardiotoxicity (DCT). Drug discovery efforts against DCT are hampered by the need to balance cardioprotection and cancer control. This study demonstrates that ICG-001 suppressed DCT in patient-derived human induced pluripotent stem cell–derived cardiomyocytes in vitro and in mice in vivo, comparable to conventional treatment, dexrazoxane. Unlike dexrazoxane, ICG-001 was cytotoxic to cancer cells. Mechanistically, ICG-001 protected the mitochondria in cardiomyocytes via DPR1 inhibition, but suppressed cancer by repressing Wnt signaling. These dual mechanisms underscore the potential of ICG-001 as an adjunct treatment to doxorubicin to improve its safety and efficacy.

Anthracyclines, such as doxorubicin (DOX), are effective chemotherapeutic agents for many cancers, but their use is limited by dose-dependent myocardial damage that can lead to cardiomyopathies and death.^1,2 Clinically, more than one-half of childhood cancer patients receive anthracyclines as part of their treatment, and these patients often experience cardiac complications.² Years of animal experimentation and numerous clinical trials have yielded only 1 clinically approved cardioprotective treatment: dexrazoxane. However, administration of dexrazoxane has been associated with reduced tumor response rates to DOX, and an increased prevalence of secondary malignancies.^3,4 Given these limitations, dexrazoxane treatments are approved by the U.S. Food and Drug Administration for only a subset of patients undergoing anthracycline-based treatments.

Doxorubicin-induced cardiotoxicity (DCT) results in part from the accumulation in, and damage to mitochondria caused by DOX. The mitochondria occupy ∼30% of the volume of adult cardiomyocytes (CMs), and produce ∼90% ATP via oxidative phosphorylation. Mitochondrial defects can compromise cardiac bioenergetics and function, leading to cell death. Mitochondrial homeostasis is maintained by a balance between fusion, which generates elongated network; and fission, which promotes the breakdown of mitochondria. Various reports implicated excessive mitochondrial fission in DCT and showed that inhibition of dynamin-related protein 1 (DRP1), a key mediator of fission, is cardioprotective.^5,6 ICG-001 is a small molecular inhibitor of the Wnt/β-catenin signaling pathway and of cAMP response element-binding-protein-binding protein (CBP)–dependent gene transcription. It also inhibits DRP1 in colorectal cancer cells.⁷ ICG-001 can protect against myocardial infarction and pressure overload; however, these effects are thought to act on noncardiac cells.^8,9 The effects of ICG-001 on CMs and its ability to protect against DCT is unknown.

Human pluripotent stem cell (hPSC)–derived cardiomyocytes (CMs) are often used to model cardiac diseases, but their application is limited by their developmental immaturity.10, 11, 12 Human PSC-CMs resemble fetal CMs and have lower mitochondrial density and metabolic function compared with adult CMs.10, 11, 12 These immature properties render conventional hPSC-CMs inadequate to evaluate treatments that counter DCT.¹³ In fact, dexrazoxane can be toxic to in vitro differentiated hPSC-CMs, contrary to clinical data.¹³ We recently showed that CD36 is a marker of cardiac maturation.¹⁰ A subset of hPSC-CMs isolated based on high levels of CD36 (CD36^hi CMs) have improved mitochondrial function, and are more sensitive to damage induced by H₂O₂, hypoxia/reoxygenation, and DOX relative to CD36^lo CMs.¹⁰ Unlike conventionally grown CMs, CD36^hi CMs generated from generic hPSCs respond to the protective effects of dexrazoxane.

In this study, we established and validated an in vitro model of DCT using 2 patient-derived CD36^hi hiPSC-CMs (abbreviated as CD36^hi CMs). The 2 hiPSC lines were derived from breast cancer patients who had experienced DCT.¹³ CMs differentiated from these hiPSCs showed increased sensitivity to DOX in vitro and are considered “DOX-sensitive” compared with hiPSC-CMs from patients who did not develop DCT.¹³ We reasoned that our patient-derived CD36^hi CMs would have 2 important advantages: recapitulation of the effect of clinical treatment (dexrazoxane), and patient-specific drug sensitivity. We then used our CD36^hi CM model to evaluate the effects of ICG-001, and its ability to function through inhibition of DRP1.¹⁰ We report that ICG-001 protects against cardiotoxicity induced by DOX and related anthracyclines, comparable to dexrazoxane, with additional benefits of tumor suppression. Cardioprotective effects were validated using an in vivo model of DCT. We conclude that ICG-001 is a promising adjunct therapy to enhance the safety and efficacy of anthracycline treatment.

Methods

Human iPSC culture and cardiac differentiation

Human iPSCs [lines: 40FSDNC15 (DOXTOX2) and 79FSDNC1(DOXTOX4)] were obtained from Stanford University Cardiovascular Institute, and were cultured and differentiated using established protocols.^13,14 Cardiac cultures underwent metabolic selection to promote maturation^15,16 and were cultured for 35 ± 5 days, after which the CD36^hi CMs were isolated by FACS.¹⁰

Phenotypic evaluation

Mitochondrial function

Mitochondrial morphology was evaluated following staining with the MitoTracker deep red dye (Thermo Fisher Scientific), and costaining with the Hoechst 33342 dye (Thermo Fisher Scientific). Mitochondrial morphology was scored as elongated (densely organized, filamentous mitochondrial networks spread throughout the cytoplasm), punctate (fragmented and punctate staining throughout the cell) or perinuclear (mitochondria mainly around the nucleus, with little staining throughout the cell), as described.¹⁶ Representative images are shown in Supplemental Figure S1. Redox potential was measured using the PrestoBlue assay (Thermo Fisher Scientific). ATP production was determined using the CellTiter-Lumi Luminescent Cell Viability Assay Kit (Beyotime). Mitochondrial respiration was evaluated with Mito Stress Test (Agilent Technologies) on Seahorse extracellular flux analyzer (XFe-96) as described.¹⁶

Nonmitochondrial phenotype

DNA damage was identified by staining for γH2AX. Sarcomeres were qualitatively assessed by α-actinin and MLC2V staining.

Apoptosis and cell death

Caspase 3/7 activity was measured using the Caspase-Glo 3/7 Assay System (Promega). Nuclei condensation was determined after staining with the Hoechst 33342 dye. Nuclei were scored as normal (relatively large nuclei with regular shape) or condensed (intensely stained, small nuclei with irregular shape). Representative images are shown in Supplemental Figure S1.

Animal experiments

All mice were kept in specific pathogen-free environment at a controlled temperature (22 ± 2 °C) under 12 hours of dark/light cycle, with free access to water supply and standard rodent diet during the study. The 9- to 13-week-old male C57BL/6 J mice were given weekly intraperitoneal injections of the following: 1) saline and PEG300; 2) DOX, 5 mg/kg; 3) DOX, 5 mg/kg, and ICG-001, 20 mg/kg administered ∼1 hour before DOX; and 4) DOX, 5 mg/kg, and dexrazoxane 50 mg/kg administered ∼1 hour before DOX, over 4 weeks. Echocardiography examinations were done using the Prospect T1 ultrasound equipment (Scintica) under anesthesia with continuous inhalation of 0.5% to 1% isoflurane, at a heart rate of 295 ± 4 beats/min. Histopathological analysis was performed after haemotoxylin and eosin (H&E) staining of paraffin-embedded sections. Fibrosis was examined by Masson’s Trichrome staining.

Ethical approval

Animal experiments complied with all institutional and national requirements for the care and use of laboratory animals and were approved by the Animal Ethics Committee of the Chinese University of Hong Kong (23-026-GRF).

Statistical analysis

GraphPad PRISM 10 (GraphPad Software) was used for statistical analysis. Data are presented as the mean ± SEM. Normality was assessed using a Shapiro-Wilk normality test. For normally distributed data, 1- or 2-way analysis of variance was performed followed by Tukey's or Dunnett's post hoc test for multiple pairwise comparisons, as specified. For non-normally distributed data, a Friedman test with Dunn's post hoc test was performed. A P value <0.05 was considered the threshold of statistical significance.

Details of experimental procedures can be found in supplementary methods and Supplemental Tables S1 and S2.

Results

Establishment and validation of a patient-derived CD36^hi hiPSC-CM model for DCT

In this study, we utilized 2 patient-derived hiPSCs to establish a clinically relevant in vitro model of this disease. Both hiPSC lines were derived from breast cancer patients who developed DOX-induced cardiotoxicity and recapitulate patient sensitivity to DOX, ie, “DOX-sensitive.”¹³

Human iPSC-CMs were differentiated through modulation of the Wnt signaling pathway, followed by metabolic selection (Figure 1A).^14,15 CMs were cultured until day 35 ± 5 of differentiation, at which time 40% to 75% CMs were positive for CD36. CD36^hi CMs were defined as CMs with top ∼25% of CD36 staining intensity (Figure 1B).

Figure 1.

Validation of CD36^hi hiPSC-CM Model

(A) Schematic of differentiation and CD36^hi cardiomyocyte (CM) isolation. Human induced pluripotent stem cell (iPSC)-CMs were stained with mitotracker and BODIPY dyes, and anti-CD36 antibody. (B) Cells were gated as CD36^hi and CD36^lo. Isotype control in red, CD36 in blue. (C) Normalized fluorescence intensities of mitotracker and BODIPY staining, and forward scatter (FSC) and size scatter (SSC) measurements, n = 5, 4, 5, 5. (D-G) CD36^hi CMs were treated with solvent control (CTRL), doxorubicin (DOX) (1 μmol/L, 26 hours), ± dexrazoxane (DRZ) (200 μmol/L, 1-hour pretreatment, followed by cotreatment with DOX). (D) Mitotracker dye staining. Percentage (%) of cells with elongated (asterisk), punctate (arrowhead), and perinuclear (arrow) mitochondria, n = 12. Significance is shown relative to DOX. (E) γH2AX staining, n = 3. (F) Redox potential measured using the PrestoBlue assay, n = 7. (G) % CMs with condensed nuclei, n = 4. Data are presented as mean ± SEM. Statistical significance was calculated using the 1- or 2-way analysis of variance with Tukey’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Scale bar = 20 μm.

We first validated that these patient-derived CD36^hi CMs had more adult-like mitochondrial properties. Flow cytometry showed that the CD36^hi subpopulation had significantly and qualitatively higher mitotracker and BODIPY staining compared with CD36^lo and mixed CMs, respectively, indicating that the former had more abundant mitochondria and greater fatty acid accumulation (Figure 1C). Imaging also revealed denser and more elongated mitochondria among CD36^hi compared with CD36^lo CMs (Supplemental Figure S2A). Forward and side scatter were significantly higher in CD36^hi compared with CD36^lo CMs, indicating larger cell size and organellar complexity in the former (Figure 1C). Our results are consistent with CD36^hi CMs having improved mitochondrial and adult-like traits.

We validated CD36^hi DOX-sensitive CMs as a model of DCT by testing if dexrazoxane could protect against DOX-induced damage. CD36^hi CMs were isolated by FACS and treated with 1 μmol/L DOX, which approximates the plasma concentration of DOX in patients.¹⁷ Dexrazoxane was applied at 200 μmol/L. This dose approximates the mean peak plasma concentration of dexrazoxane (∼36.5 mg/L, 137 μmol/L) after a 15-minute intravenous infusion at 500 mg/m²,¹⁸ and we have previously shown that dexrazoxane protects generic CD36^hi hPSC-CMs against DCT at this concentration.¹⁰ We examined mitochondrial homeostasis using the mitotracker dye. We observed 3 patterns of mitochondrial morphologies among the CD36^hi CM samples, but in different proportions. The morphologies include the following: ‘elongated” mitochondrial networks thought to represent a healthy balance of fission and fusion; punctate, fragmented mitochondria suggestive of increased fission; and perinuclear mitochondria consistent with increased fission and mitochondria dysfunction (Supplemental Figure S1). Control CD36^hi CMs (solvent treated) had mostly healthy, elongated mitochondria networks, while punctate and perinuclear mitochondria predominated in DOX-treated CD36^hi CMs (Figure 1D). Dexrazoxane prevented mitochondrial fragmentation induced by DOX, and significantly increased the percentage of cells with healthy, elongated mitochondria from 30.0 ± 6.6% to 45.1 ± 8.7% (Figure 1D). We also examined redox potential as a surrogate measure of mitochondrial function and found that dexrazoxane significantly preserved redox potential in DOX-treated CD36^hi CMs (Figure 1F). Staining of CD36^hi CMs for γH2AX, a marker of DNA damage, showed that dexrazoxane reduced γH2AX signal by 1.8 ± 0.2-fold compared with DOX alone (Figure 1E). The frequency of apoptosis was assessed based on the appearance of brightly stained, small and “condensed” nuclei following Hoechst nuclear staining. A reduction of nuclear condensation upon dexrazoxane + DOX (7.3 ± 3.2%) was observed compared with DOX alone (15.3 ± 3.6%) (Figure 1G). These experiments were performed using CD36^hi CMs from both patient-derived hiPSC lines. Examples of data from individual cell lines are shown in Supplemental Figure S3, and results were combined for analysis (Figure 1). All of the data are consistent with dexrazoxane having a protective role in DOX-sensitive, CD36^hi CMs. For comparative purposes, we evaluated the effects of dexrazoxane in CD36^hi and CD36^lo CMs. Although CD36^hi and CD36^lo CMs had similar levels of topoisomerase 2β (TOP2B), which was proposed to be molecular target of dexrazoxane; the latter drug protected CD36^hi CMs but not CD36^lo CMs, analogous to what we observed in generic hPSC-CMs¹⁰ (Supplemental Figures S2B and S2C). In summary, patient-derived DOX-sensitive CD36^hi CMs can recapitulate known clinical responses to dexrazoxane, a finding that validates their use as a suitable in vitro human model for the identification of novel cardioprotective agents against DOX.

ICG-001 protects against DCT in CD36^hi CMs

We used this model to test if ICG-001, an established CBP/β-catenin inhibitor, could protect against DCT through inhibition of mitochondrial fission. CD36^hi CMs were treated with DOX and multiple concentrations of ICG-001 before an assessment of mitochondrial fragmentation. Compared with control, DOX-treated CD36^hi CMs were significantly less likely to have elongated mitochondria and more likely to have perinuclear mitochondrial localization (Elongated: control 65.9 ± 5.8% vs DOX 26.7 ± 7.5%; perinuclear: control 12.6 ± 5.8% vs DOX 44.3 ± 7.8%) (Figure 2A). ICG-001 (2 μmol/L) significantly increased the percentage of elongated mitochondria (ICG-001+DOX 48.6 ± 6.9% vs DOX 26.7 ± 7.5%) and reduced the proportion of perinuclear mitochondria (ICG-001+DOX 23.7 ± 6.4% vs DOX 44.3 ± 7.8%). DOX also reduced redox potential, which was significantly alleviated by ICG-001 (Figure 2B). To see if attenuation of mitochondrial damage is sufficient to comprehensively protect CMs, we assayed control, DOX and ICG-001 + DOX treated CD36^hi CMs for DNA damage. Relative to DOX-treated cells, ICG-001 + DOX significantly reduced γH2AX staining by 1.4 ± 0.1-fold (Figure 2C and 2D) although the signal intensity was still significantly greater than untreated controls. Structural integrity of CD36^hi CMs was qualitatively examined by staining for α-actinin and MLC2V, which are thin and thick filament sarcomeric proteins important for contraction. Although brightly stained, striated structures could be clearly seen in control cells, staining intensity was reduced and sarcomeres became disorganized after DOX treatment, a structural phenotype that was ameliorated by ICG-001 (Figures 2E and 2F). ICG-001 also significantly suppressed caspase 3/7 activity (Figure 2G) and decreased the proportion of cells with nuclear condensation (Figure 2H). Cardioprotective effects were observed in CMs from both independent hiPSC lines. Examples of data from individual cell lines are shown in Supplemental Figure S4, and results were combined for analysis (Figure 2). For comparison, we also evaluated the response of CD36^lo CMs to ICG-001. CD36^hi CMs were significantly more sensitive to DOX than CD36^lo CMs. ICG-001 protected CD36^hi CMs but not CD36^lo CMs (Supplemental Figures S2D and S2E), a result that is similar to the divergent effects we observed for dexrazoxane in CD36^hi vs CD36^lo CMs (Supplemental Figure S2C).

Figure 2.

ICG-001 Protects Against DOX-Induced Cardiotoxicity in CD36^hi CMs

CD36^hi CMs were treated with solvent CTRL, DOX (1 μmol/L, 26 hours) ± ICG-001 (ICG, 1-hour pretreatment). (A) Mitotracker dye staining. Percentage (%) of cells with elongated (asterisk), punctate (arrowhead), and perinuclear (arrow) mitochondria. Significance is shown relative to DOX. n = 10. (B) Redox potential measured using the PrestoBlue assay, n = 11. (C and D) Graph and images of γH2AX staining n = 8. (E) α-actinin staining, n = 4 and (F) MLC2V staining, n = 4. Arrow indicates striations and asterisk indicates disorganized structure. (G) Normalized caspase 3/7 activity, n = 12. (H) Percentage (%) of cells with condensed nuclei, n = 6. Data are presented as mean ± SEM. Statistical significance was calculated using the 1- or 2-way analysis of variance with Tukey’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Scale bar = 20 μm. Abbreviations as in Figure 1.

Although most studies on chemotherapy-induced cardiotoxicity have focused on DOX, other members of the anthracycline family are cardiotoxic. Daunorubicin is indicated for the treatment of acute lymphoblastic and acute myeloid leukemia in adults, and it is associated with dose-dependent cardiotoxocity.¹⁹ We thus tested if ICG-001 could protect CD36^hi CMs against daunorubicin. Similar to results with DOX, ICG-001 preserved mitochondrial integrity (Figure 3A), significantly restored redox potential (Figure 3B and Supplemental Figure S4B) and suppressed nuclear condensation (Figure 3C) in daunorubicin-treated cells, showing that the cardioprotective effects of ICG-001 is not limited to DOX.

Figure 3.

ICG-001 Protects Against Daunorubicin-Induced Cardiotoxicity

CD36^hi CMs were treated with solvent control (CTRL), daunorubicin (DNX: 2 μmol/L, 26 hours) ± ICG-001 (ICG: 2 μmol/L) (A) Mitotracker dye staining, n = 5. (B) Redox potential measured using the PrestoBlue assay, normalized to control and presented as fold change (FC), n = 10. (C) Percentage (%) of cells with condensed nuclei, n = 3. Data are presented as mean ± SEM. Statistical significance was calculated using the 1-way analysis of variance with Tukey’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Scale bar = 10 μm.

We benchmarked the efficacy of ICG-001 against dexrazoxane in CD36^hi CMs. Dexrazoxane and ICG-001 both significantly suppressed DOX-induced mitochondrial fragmentation (Figure 4A), increased redox potential (Figure 4C), ameliorated DNA damage (Figure 4D), and inhibited cell apoptosis (Figure 4E), whereas ICG-001 significantly increased ATP production (Figure 4B). We also examined the effects of ICG-001 and dexrazoxane on oxygen consumption following the additions of inhibitors of mitochondrial complexes and uncoupler. ICG-001 qualitatively improved basal and maximal respiration, and ATP production, consistent with improved activities of mitochondrial complexes I-V (Supplemental Figure S5). Some qualitative differences were observed between the effects of ICG-001 and dexrazoxane. ICG-001 could better protect against mitochondrial fragmentation, increase ATP level and oxygen consumption, and suppress nuclear condensation, whereas dexrazoxane was more effective against DNA damage, consistent with mitochondrial and DNA damage being the primary targets of ICG-001 and dexrazoxane respectively. Nonetheless, no statistically significant differences could be demonstrated between the 2 compounds, showing that ICG-001 may be similar to dexrazoxane in its protective effects against DCT.

Figure 4.

ICG-001 Is Comparable to Dexrazoxane in Cardioprotection

CD36^hi CMs were treated with solvent CTRL, DOX (1 μmol/L, 26 hours), ± ICG-001 (ICG: 2 μmol/L), or DRZ (200 μmol/L). (A) Percentage (%) of CD36^hi CMs with elongated, punctate and perinuclear mitochondria were quantified, n = 8. Significance is shown relative to DOX. (B) ATP assay normalized to control, n = 6. (C) Redox potential was measured using the PrestoBlue assay, normalized to control and presented as FC, n = 7. (D) The intensity of γH2AX staining, n = 3. (E) % cells with condensed nuclei, n = 6. Data are presented as mean ± SEM, and n refers to biological replicates. Statistical significance was calculated using the 1- or 2-way analysis of variance with Tukey’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Abbreviations as in Figures 1 and 3.

We performed RNA-sequencing to determine the transcriptomic profiles of control, DOX, and DOX + ICG-001 treated CD36^hi CMs. Gene ontology analysis of differentially expressed genes (DEGs) showed that DOX significantly reduced the expression of genes encoding nuclear, extracellular and mitochondrial proteins, compared with control, in line with DOX as an inhibitor of TOP2B and inducer of mitochondrial damage (Supplemental Table S3). DEGs implicated in transcription were enriched in control cells, while DEGs involved in mRNA splicing, transcription, and protein ubiquitination were higher in DOX-treated CD36^hi CMs (Supplemental Table S4). We next compared DOX and DOX + ICG-001–treated CD36^hi CMs. In terms of cellular components, up-regulated DEGs in DOX + ICG-001–treated CD36^hi CMs were enriched for genes encoding proteins located in the z-disc, the mitochondria, and the sarcoplasmic reticulum, while down-regulated DEGs comprise extracellular region genes and others (Figure 5A, Supplemental Table S5). Analyses of biological process and KEGG pathway revealed up-regulation of genes important for sarcomere organization, contraction, and metabolic pathways (Figure 5B, Supplemental Table S6). Focusing on the mitochondria and metabolism, we observed increased expression of genes important for oxidative phosphorylation (eg, NADH:ubiquinone oxidoreductase subunit AB1 [NDUFAB1], ATP synthase membrane subunit f [ATP5MF], and cytochrome c oxidase subunit 6A2 [COX6A2]), fatty acid metabolism (eg, acyl-CoA dehydrogenase short chain [ACADS]), energy transfer (creatine kinase, mitochondrial 2 [CKMT2]), and antioxidant response (eg, peroxiredoxin 3 [PRDX3]). Structural genes such as cardiac troponin T (TNNT2) and Myomesin-2 (MYOM2) were also up-regulated. This is consistent with our cell biology analysis showing preserved mitochondrial function and sarcomeric structure after ICG-001 treatment (Figure 2). Conversely, genes associated with cytokine production and extracellular matrix were decreased after ICG-001 treatment, in line with previous reports that ICG-001 could suppress inflammation and fibrosis.^8,9

Figure 5.

ICG-001 Protects CD36^hi CMs by Inhibiting DRP1 and Preserving the Mitochondria

CD36^hi CMs were treated with DOX (1 μmol/L, 26 hours) ± ICG-001 (ICG: 2 μmol/L). Gene ontology analysis of differentially expressed genes between DOX vs ICG+DOX. (A) Cellular components and (B) biological processes and pathways. Enrichment is indicated by −log(p), where p refers to P value. Large −log10(p) implies significant enrichment. (C) Ingenuity Pathway Analysis showing upstream regulators (blue), which positively (triangles) or negatively (diamonds) regulate downstream targets (oval shapes) involved in metabolism (green), Ca²⁺ handling (pink), contraction (purple), and electrophysiology (yellow). (D) qPCR analysis of Wnt target genes. Expression was normalized to glyceraldehyde 3-phosphate dehydrogenase and to control, n = 3. (E) Quantification and representative images of DRP1 staining, normalized to control and presented as FC, n = 4. CD36^hi CMs were treated with DOX ± mdivi-1 (MD). (F) Fluorescence images of mitotracker dye staining (green). The percentage (%) of cells with elongated (asterisk), punctate (arrowhead) and perinuclear (arrow) mitochondria were quantified, n=13. Significance is shown relative to DOX. (G) %CD36^hi CMs with condensed nuclei, n = 8. Data are presented as mean ± SEM. Statistical significance was calculated using the 1- or 2-way analysis of variance Tukey’s post hoc test, or the Friedman test with Dunn’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Scale bar = 20 μm. Abbreviations as in Figures 1 and 3.

We performed Ingenuity Pathway Analysis to identify upstream regulators induced by ICG-001 based on prior knowledge of expected effects between transcriptional regulators and their target genes.²⁰ Our analysis predicted the activation of ESRRA/ESRRG, GATA4, MYOCD, HAND2, and TBX5, and the inhibition of SIX1 in DOX + ICG-001 vs DOX treated hiPSC-CMs (Figure 5C). Together, these factors may cooperatively regulate downstream targets important for contraction (eg, ACTA1, TNNT2, TNNI3), electrophysiology (eg, KCNQ1 and KCNH2), Ca²⁺ handling (CASQ2 and SLC8A1) and metabolism (THRA and CKM) (Figure 5C).

Mechanistically, ICG-001 has been shown to inhibit Wnt/β-catenin signaling and DRP1. From the RNA-seq data set, genes related to Wnt signaling were not enriched among DEGs between DOX and DOX + ICG-001 treated CD36^hi CMs. There was no significant difference in the expression of the mediators (CREBBP and EP300) or downstream targets of Wnt signaling (CCND1 [cyclin D1], MYC [MYC Proto-Oncogene, BHLH Transcription Factor], MEIS1 [Meis Homeobox 1] and BIRC5 [survivin]). We also independently evaluated the mRNA levels of Wnt target genes by RT-qPCR and found no significant difference between CD36^hi CMs treated with DOX vs DOX + ICG-001 (Figure 5D). ICG-001 is thought to inhibit the transcription of Wnt target genes by blocking the interaction between β-catenin and CBP. Our results are therefore inconsistent with Wnt signaling being the target of ICG-001–mediated protection in CD36^hi CMs.

We examined DRP1 mRNA and protein levels in DOX vs DOX + ICG-001 treated CD36^hi CMs. RNA-seq analyses revealed no enrichment of mitochondrial fission genes among DEGs, or any significant difference in the mRNA levels of DRP1 (encoded by DNM1L) or in other mediators of mitochondrial fission (FIS1 and MFF) or fusion (MFN1, MFN2, and OPA1). However, DOX increased DRP1 staining, which was significantly ameliorated by ICG-001 (Figure 5E). To further test the role of DRP1 in cardioprotection against DOX, we inhibited DRP1 using an established small molecule inhibitor, mdivi-1.²¹ Mdivi-1 phenocopied ICG-001. It improved mitochondrial morphology and significantly promoted mitochondrial elongation (Figure 5F); and significantly suppressed nuclear condensation (Figure 5G) in DOX-treated CD36^hi CMs. Together, our results are consistent with suppression of DRP1 being a contributor of ICG-001-mediated cardioprotection.

ICG-001 protects against DCT in prolonged culture of hiPSC-CMs

Prolonged culture is known to promote structural and functional maturation of hiPSC-CMs.²² To determine if the cardioprotective effects of ICG-001 could be observed in older CMs that had not been sorted for CD36, we used patient-derived DOX-sensitive hiPSC-CMs cultured for 100 to 250 days. Day 100 to 250 hiPSC-CMs mimic adult CMs in their elongated morphology (Supplemental Figure S6A). DOX-induced mitochondrial fragmentation, reduced redox potential and ATP production, and these were attenuated by ICG-001 (Supplemental Figures S6B to S6D). These data show that the effects of ICG-001 are conserved among mature hiPSC-CMs based on either CD36 expression or time in culture.

ICG-001 protects against DCT in vivo

To determine if the cardioprotective effects observed in vitro are applicable in vivo, we injected mice with DOX once a week × 4 to simulate 4 rounds of chemotherapy treatment (Figure 6A) (5 mg/kg; intraperitoneal, cumulative dose 20 mg/kg).²³ This regimen was chosen because the pharmacokinetic properties of DOX after ∼5 mg/kg in mice is similar to that reported in patients after a standard treatment of DOX (60 mg/m²).23, 24, 25 ICG-001 and dexrazoxane were administered 1 hour before DOX treatment, analogous to our in vitro pretreatment protocol. ICG-001 was applied at 20 mg/kg because this dosage of ICG-001 has previously been shown to induce antitumor effects.^26,27 Dexrazoxane is applied clinically at a 10-20:1 ratio of dexrazoxane to DOX. Here, we used 50 mg/kg of dexrazoxane to achieve a 10:1 ratio relative to DOX (5 mg/kg). This dosage of dexrazoxane has been used in prior studies of DCT to demonstrate cardioprotection.^28,29

Figure 6.

ICG-001 Can Protect Against DOX-Induced Cardiotoxicity In Vivo

(A) Schematic of treatment. C57BL/6 mice were intraperitoneally injected with DOX ± pretreatment (1 hour) of ICG or DRZ, n = 4-7. (B) Change in body weight (BW) relative to Day 0. Significance was calculated based on %BW on day 28, n = 4-7. (C) Heart weight (HW) normalized to tibial length (TL). (D) Examination of cardiac pathology by H and E staining. Degeneration of the myocardial tissue (black arrow), sarcoplasmic vacuolation (yellow arrow), inflammatory cell infiltrations (red arrow). (E) Fibrosis was measured by Masson’s Trichrome staining and quantified using image analysis platform. Echocardiography examinations were performed on day 28. (F) Representative tracings. (G) Ejection fraction (EF). (H) Fractional shortening (FS). Left ventricular internal diameter (LVID) at the end of (I) systole (s) and (J) diastole (d). (K) qPCR analysis of Wnt target genes, normalized to glyceraldehyde 3-phosphate dehydrogenase and to control mice, n = 5. (L and M) Gene set enrichment analysis (GSEA) of RNA-sequencing data. (L) Top 5 biological processes and cellular components altered in DOX vs control and ICG-001+DOX vs DOX animals, ranked by the magnitude of normalized enrichment score (NES). Positive/negative NES represent up/down-regulation. (M) Top/bottom 25 biological processes, ranked by NES, were grouped into broad functional categories and color coded. Data are presented as mean ± SEM, and n refers to biological replicates. Statistical significance was calculated using the 1-way analysis of variance with Tukey’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Abbreviations as in Figures 1 and 3.

Control mice treated with vehicle gained body weight over the treatment regimen. Conversely, the body weight of DOX-treated mice gradually declined, and this was qualitatively ameliorated by ICG-001 and by dexrazoxane treatment (control: 105.7 ± 1.5%, DOX: 92.4 ± 2.0%, ICG-001 + DOX: 96.5 ± 2.9%, DRZ + DOX: 93.0 ± 2.0%) (Figure 6B). A significant reduction in heart weight also was observed in DOX-treated mice, but this reduction was qualitatively lessened in the DOX + ICG-001 cohort (Figure 6C). Cell size was assessed using wheat germ agglutinin staining and was similar among all groups (Supplemental Figure S7A). Histopathological examination revealed the presence of intracellular vacuoles, immune inflammatory cell infiltrations, and myofibril degeneration after DOX, analogous to clinical reports, and these were ameliorated by ICG-001 (Figure 6D). Furthermore, Masson trichome staining showed increased collagen deposition among the hearts of DOX-treated animals relative to controls. This increase was significantly attenuated in the ICG-001 + DOX treatment group (Figure 6E).

Cardiac function was evaluated by echocardiography (Figures 6F to 6J). DOX treatment decreased left ventricular ejection fraction compared with control but this was significantly reversed by ICG-001 treatment (control: 76.4 ± 2.4%, DOX: 55.6 ± 5.4%: ICG-001+DOX: 74.4 ± 3.2%, DRZ+DOX: 67.6 ± 2.6%) (Figures 6F and 6G). ICG-001 treatment also significantly attenuated DOX-induced reduction in fractional shortening (control: 45.8 ± 2.2%, DOX: 29.3 ± 3.6%, ICG-001+DOX: 43.1 ± 3.1%, DRZ+DOX: 36.9 ± 2.3%) (Figures 6F and 6H). Other cardiac parameters also showed improvements: the left ventricular internal diameter at the end of systole (LVIDs) and left ventricular internal diameter at the end of diastole (LVIDd) were both significantly increased after DOX injection, indicative of LV dilatation, but this was significantly and qualitatively attenuated by ICG-001 respectively (Figure 6I and 6J). Comparison between ICG-001 and dexrazoxane revealed comparable (LVIDs) or qualitatively superior (body weight, heart weight, left ventricular ejection fraction, and fractional shortening) protection by ICG-001 (Figures 6B to 6J).

To determine the molecular mechanisms underlying the protective effects of ICG-001, we examined the effects of this compound on Wnt signaling and DRP1 expression. Consistent with in vitro results, ICG-001 (and dexrazoxane) did not noticeably affect the mRNA levels of Wnt signaling targets (Figure 6K). Conversely, ICG-001, but not dexrazoxane, qualitatively reduced DRP1 protein level (Supplemental Figure S7B). We next profiled the transcriptomes of cardiac samples from control, DOX, and DOX + ICG-001–treated animals by RNA sequencing, followed by gene set enrichment analysis³⁰ (Figures 6L and 6M). Relative to control, DOX-treated hearts had greatly reduced levels of mitochondrial and ribosomal transcripts (eg, inner mitochondrial membrane protein complex and mitochondrial protein-containing complex, organellular ribosome), while genes involved in immune response (eg, antigen processing) were up-regulated (Figure 6L). These changes were mostly reversed upon ICG-001 treatment. In fact, the 5 most up-regulated biological processes and cellular components in DOX + ICG-001 hearts were all related to the mitochondria, eg, aerobic respiration and oxidative phosphorylation (Figure 6L). To look more broadly at genes altered by DOX and ICG-001, we manually grouped the top 25 biological processes with the most positive and negative enrichment scores into functional categories (Figure 6M, Supplemental Table S7). Genes important for mitochondrial function and energy production were down-regulated by DOX. This includes genes involved in ATP synthesis, electron transport chain, and mitochondrial translation. Immune and inflammatory genes were up-regulated. These changes were reversed in DOX + ICG-001 vs DOX. Genes important for the mitochondria, respiration, and energy production were up-regulated, whereas those involved in inflammation and extracellular matrix were down-regulated, in line with the improved histopathological features of DOX + ICG-001 vs DOX animals (Figure 6D). These in vivo transcriptomic changes are consistent with our in vitro data showing preserved mitochondrial gene expression (Figure 5) and function (Figure 2).

ICG-001 suppresses tumor growth by inhibiting the Wnt signaling pathway

The use of dexrazoxane is limited partly because of prior reports that dexrazoxane can promote cancer. We thus evaluated the effect of ICG-001 in a panel of cell lines from solid tumors and leukemia of adult and pediatric origins (Figures 7A to 7I, Supplemental Figure S8). Consistent with the known antitumor effects of ICG-001,^7,26,31 treatment with this drug alone reduced the viability of MV4-11 leukemia, SKOV3 ovarian cancer, BE2C neuroblastoma, MDA-231 breast cancer, and Raji and HS-Sultan lymphoma cells (Figures 7A to 7F, Supplemental Figure S8) in a concentration-dependent manner. Reduced viability was also observed when ICG-001 was used as a cotreatment with DOX compared with DOX alone in MV4-11, SKOV3, MDA-231, Raji, HS-Sultan, BE2C, HL60 leukemia, HeLa cervical and A549 lung cancer cells (Figures 7B to 7F and Supplemental Figure S8). These effects were statistically significant in most of the cell lines tested (MV4-11, BE2C, MDA-231, Raji, HeLa, and A549) at 2 μmol/L ICG-001, which was the concentration needed to protect CD36^hi CMs (Figure 2). Furthermore, the viability of MV-411 cells was significantly lower after ICG-001 + DOX treatment compared with dexrazoxane + DOX, showing that the former has superior anticancer efficacy (Figure 7C). A significant increase in trypan blue staining and morphological signs of cell death were detected in ICG-001 + DOX vs DOX treated MV4-11 cells (Figures 7D and 7E), showing that the reduction of viability induced by ICG-001 is associated with increased cytotoxicity of cancer cells.

Figure 7.

ICG-001 Suppresses Human Cancer Cells by Inhibiting the Wnt Signaling Pathway

MV4-11 (A-E, H-J) and SKOV3 (F and G) cells were treated with solvent CTRL, DOX (MV4-11: 0.2 μmol/L and SKOV3: 1 μmol/L, 24 hours) ± ICG-001 (ICG: 2 μmol/L unless otherwise stated) ± DRZ (200 μmol/L) ± mdivi-1 (MD). (A-C, F, G, and J) Normalized viability measured using the PrestoBlue assay, presented as FC, n = 4-7. (D) %Trypan blue⁺ cells, n = 8. (E) Representative phase-contrast of MV4-11 cells after treatment. (H) qPCR analysis of Wnt target genes, normalized to glyceraldehyde 3-phosphate dehydrogenase and control, n = 6. (I) Western blot of DRP1, normalized to β-actin and control, n = 3. (J) The effect of mdivi-1 (MD) measured using the PrestoBlue assay, normalized to control, n = 4. Data are presented as mean ± SEM. Statistical significance was calculated using the 1-way analysis of variance with Tukey’s or Dunnett’s post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Significance is shown relative to CTRL (A and F) or DOX (B and G) or both (C and D, H-J). Abbreviations as in Figures 1 and 3.

To test if the antitumor effects of ICG-001 with DOX are mediated via Wnt signaling, we evaluated the expression of Wnt downstream targets in MV4-11 cells. We found that ICG-001 significantly reduced the mRNA levels of CCND1, MEIS1, and BIRC5 compared with DOX alone (Figure 7H). We next tested if ICG-001 could inhibit DRP1 in cancer cells.⁷ No significant change in DRP1 protein level was detected when ICG-001 was applied alone or with DOX (Figure 7I). Suppressing DRP1 with mdivi-1 had no notable effect on cell viability (Figure 7J), unlike its effect on CMs (Figures 5F and 5G). Overall, these results consistently supported the Wnt signaling pathway, rather than DRP1, as the target of ICG-001 in cancer cells.

Discussion

DOX is effective against many types of cancer but can induce dose-dependent, potentially lethal cardiotoxicity. In the United States, dexrazoxane is only indicated for women with metastatic breast cancer who have received a cumulative DOX dose of 300 mg/m². Restriction on the use of dexrazoxane is partly caused by prior reports that it can compromise the antitumor cytotoxicity of DOX and increase the incidence of secondary malignant neoplasms.^3,4 Recent studies lend support to the safety of dexrazoxane, although others still advise limited use.32, 33, 34 Despite the controversy surrounding the use of dexrazoxane, it is indisputable that cardioprotection against DCT must be balanced by cancer control. In this study, we utilized patient-derived DOX-sensitive, CD36^hi hiPSC-CMs and an in vivo mouse model to identify potentially novel and efficacious treatments for DCT. We report that the Wnt inhibitor, ICG-001, can provide the dual benefits of cardioprotection and enhanced tumor cytotoxicity.

Burridge et al¹³ first showed that hiPSC-CMs recapitulate the predilection of patients to DCT and firmly establishes the importance of patient-derived hiPSC-CMs as a clinically relevant model of this disease. However, the same report also showed that dexrazoxane was detrimental to hiPSC-CMs,¹³ contrary to the known effects of the latter in patients. The discrepancy between results in hiPSC-CMs, and known clinical effects suggest that the use of conventional hiPSC-CMs may lead to false negatives and compromises the discovery of new therapy. This difference was attributed by the authors to lower mitochondrial density in immature hiPSC vs adult CMs. We obtained 2 of the cell lines used in this study and used 2 strategies to improve the maturation of these hiPSC-CMs. First, we optimized our differentiation and maturation protocols to improve purity and maturation.¹⁴ Next and most importantly, we isolated the CD36^hi CM subpopulation from patient-derived hiPSC cardiac cultures. Patient-derived CD36^hi CMs were more mature, had higher mitochondrial mass, and had denser and more elongated mitochondria than control and CD36^lo CM. Because DOX accumulates in and damages the mitochondria, we believe that improved mitochondrial mass and function can critically influence cellular responses to DOX and dexrazoxane, in line with what Burridge et al¹³ proposed. Although the 2 subpopulations do not differ in their expression of TOP2B, patient-derived CD36^hi, but not CD36^lo hiPSC-CMs, correctly responded to the cardioprotective effects of dexrazoxane, analogous to what we reported in generic hPSC-CMs.¹⁰ Similar discrepancies were seen with ICG-001, ie, CD36^hi, not CD36^lo hiPSC-CMs were protected by this drug. These results with dexrazoxane and ICG-001 show that the use of immature hiPSC-CMs may give rise to false negative results and impede the discovery of new therapy. Our data highlight the importance of using mature hiPSC-CM models for drug discovery and validates our patient-derived CD36^hi CMs as a clinically relevant model to study DCT.

Using this and an established mouse model of DCT, we showed that ICG-001 can protect against DCT in vitro and in vivo through direct effects on CMs. ICG-001 had previously been reported to have beneficial effects in other cardiac disorders; however, most of the benefits were attributed to effects on non-CMs. In a model of infarcted myocardium, ICG-001 was shown to suppress fibrosis.⁸ In the setting of pressure overload injury,⁹ ICG-001 was antihypertrophic and alleviated inflammation by reducing macrophage accumulation.⁹ The lack of emphasis on CMs may be caused by the fact that ICG-001 is a Wnt inhibitor, and the Wnt pathway is thought to be inactive in adult CMs, despite its importance during cardiac development and differentiation.^14,35,36 Consistent with this, ICG-001 did not significantly alter the level of established Wnt target genes in CD36^hi CMs or in mouse heart, showing that ICG-001 does not protect CMs via this pathway. Instead, we found that ICG-001 protects CD36^hi CMs by preserving the mitochondria possibly via suppression of DRP1 and maintaining mitochondrial integrity. Impaired mitochondrial dynamics has been implicated in numerous cardiac diseases including ischemia-reperfusion injury, dilated and metabolic cardiomyopathies,³⁷ and in drug-induced cardiotoxicity.¹⁶ DOX can induce excessive mitochondrial fission by up-regulating or activating DRP1,³⁸ while suppression of the latter via siRNA, heterozygous deletion, and pharmaceutical intervention can suppress DCT in H9c2 rat cardiac myoblast cells and in mice.^6,38 We advanced these studies by showing that preserving mitochondrial integrity using ICG-001 could protect against DOX (and daunorubicin) in patient-derived in vitro and mouse in vivo models of DCT.

Consistent with the mitochondria being the target of ICG-001-mediated protection, we observed an up-regulation of mitochondrial and metabolic genes, as well as genes critical for multiple aspects of cardiac function after ICG-001 treatment. Our bioinformatics analysis predicts that the activation of ESRRA/ESRRG, GATA4, MYOCD, HAND2, TBX5, and the inhibition of SIX1 may contribute to the transcriptomic phenotype induced by ICG-001. The transcription regulators identified here are key to heart function.39, 40, 41 For instance, the cardiac specific deletion of TBX5 and GATA4 in mice can lead to cardiac dysfunction and compromised CM survival,^40,41 while postnatal knockdown of ESRRA/ESRRG results in cardiomyopathy and mitochondrial dysfunction.⁴² Conversely, SIX1 is a proinflammatory, interferon-stimulated transcription factor implicated in adverse cardiac remodeling.⁴³ Altered activities of these factors by ICG-001 may therefore help preserve cardiac function upon DOX injury. Indeed, overexpression of GATA4 can prevent DOX-induced CM death in cultured neonatal rat cardiomyocytes.⁴⁴ Although we show that ICG-001 can suppress DRP1, the latter is not known to be a transcriptional regulator. Inhibition of DRP1 by mdivi-1 during hPSC-CM differentiation can increase the mRNA levels of TBX5 and GATA4, but this may reflect enhanced differentiation toward the cardiac lineage, rather than direct effect.⁴⁵ The role of DRP1 in the transcriptional program induced by ICG-001 awaits further investigation. Nonetheless, our analysis uncovered a transcriptomic regulatory network that contributes to ICG-001–mediated cardioprotection.

Therapeutically, our pairwise comparison showed that ICG-001 is comparable to dexrazoxane in its ability to protect hiPSC-CMs in vitro and mouse heart in vivo. Unlike dexrazoxane, however, ICG-001 alone or with DOX reduced viability in a diverse panel of cancer cells likely through inhibition of the Wnt/β-catenin signaling pathway, which has been shown to suppress multiple cancers.^7,26,31,46 Consistently, we find that ICG-001 suppresses Wnt targets in cancer cells. Cyclin D1 and MEIS1³¹ are important for proliferation, while survivin, is a key antiapoptotic protein. Reduced expression of these genes likely contributed to the tumor suppressive effects shown. In contrast to the results in hiPSC-CMs, ICG-001 did not alter DRP1 levels in cancer cells. Currently, PRI-724, an isomer/analogue of ICG-001, is being/was tested in clinical trials and was shown to be well-tolerated among patients with liver fibrosis, primary biliary cholangitis and cancer.47, 48, 49 Future clinical studies designed to evaluate the cardiac safety and efficacy of ICG-001 and PRI-724 as cotreatment with anthracyclines may therefore be warranted.

Study limitations

HiPSC-CMs are known to be immature.10, 11, 12 Using an optimized differentiation protocol coupled with CD36 isolation, we find that ICG-001 affects DRP1 in CMs and is protective against DOX. It is, however, possible that our CD36^hi CM model may not fully recapitulate the response of adult CMs to DOX. To overcome this limitation, we also employed prolonged culture of hiPSC-CMs, which is known to promote a more adult-like phenotype. Key experiments were repeated using patient-derived hiPSC-CMs cultured for 100 to 250 days. Consistent with our data with CD36^hi CMs, the results confirm that ICG-001 can protect older, mature human CMs against DCT. Second, the heart rate at which echocardiography was performed is outside the physiologic range. Thirdly, our in vivo experiments were done using wild-type mice without cancer. Ideally, comparison between DOX and DOX + ICG-001 treatment in tumor-bearing mice should be performed to fully validate our in vitro findings of ICG-001 cytotoxicity and protection on cancer cells vs CMs, respectively. In summary, our results are consistent with a protective effect of ICG-001 on the heart in preclinical models. Although the clinical significance of these findings should be further tested in clinical trials, our study may provide novel strategies for protecting the hearts of cancer patients who undergo anthracycline treatment.

Conclusions

We show that ICG-001 suppresses the growth of a broad range of cancer cell lines through Wnt inhibition, while simultaneously providing cardioprotection through preserving the mitochondria. These properties make ICG-001 a potentially superior alternative to conventional treatment to protect against DCT among cancer patients.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: DOX is a first-line treatment against cancer but can cause cardiotoxicity. There is only one U.S. Food and Drug Administration–approved treatment against DCT, dexrazoxane, but its use is limited because of prior reports that it may compromise the efficacy of DOX and induce secondary malignancy.

TRANSLATIONAL OUTLOOK: We showed that ICG-001 can protect against DCT with efficacy similar to that of dexrazoxane, but it offers the added benefit of tumor suppression. These results support future clinical trials involving ICG-001 as an adjunct treatment to DOX that may lead to improved clinical outcomes for cancer patients.

Funding Support and Author Disclosures

This work was supported by funding from the Children’s Cancer Foundation (6905997) and the Health and Medical Research Fund (06173796), Early Career Scheme (24100421) and The Chinese University of Hong Kong Direct Grant (4054671) (to Dr Poon), and Research Grant Council Theme-based Research Scheme (T13-602/21-N) and collaborative research fund (C5045-20EF) (to Dr Wang). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.