BRD4770 protects against DOX-induced cardiotoxicity by inhibiting apoptosis and ferroptosis

Abstract

Doxorubicin (DOX) is an effective anticancer drug, but its clinical utility is limited mainly by cardiotoxicity causing cardiomyocyte ferroptosis and apoptosis. While DOX-induced cardiotoxicity (DIC) involves epigenetic changes, no systematic epigenetic intervention studies have been reported. Here, we identified BRD4770 as a potential small molecule against DIC through the Epigenetics Compound Library screening. BRD4770 inhibited DOX-induced cardiomyocyte ferroptosis and apoptosis by reducing reactive oxygen species (ROS) production and lipid peroxidation and maintaining glutathione homeostasis. BRD4770 treatment alleviated DIC without affecting the antitumor effects of DOX. Mechanistically, BRD4770 promoted nuclear factor erythroid 2-related factor 2 (Nrf2)/activating transcription factor 4 (ATF4)–solute carrier family 7 member 11 (SLC7A11) signaling and suppressed DOX-induced cardiomyocyte death by reducing methylation of lysine 9 on histone 3. Last, we constructed a ROS-responsive nanoliposome loaded with BRD4770 and conjugated with the cardiac-targeting peptide for primary cardiomyocyte, which provided better protection against DIC. These findings suggest that BRD4770 has the potential to prevent cardiomyocyte death–related cardiomyopathy.

BRD4770 protects the heart from DOX-induced cardiotoxicity by inhibiting cardiomyocyte apoptosis and ferroptosis.

INTRODUCTION

Doxorubicin (DOX) is a traditional and widely used anthracycline chemotherapeutic drug that has been used to treat malignancies for more than five decades (1). However, cardiotoxicity occurs in many patients receiving DOX chemotherapy, and DOX-induced cardiac dysfunction and heart failure are collectively referred to as DOX-induced cardiotoxicity (DIC). The effects of DIC substantially restrict the utilization of DOX (2). The DIC-induced cardiomyocyte damage is mainly manifested by the excessive accumulation of reactive oxygen species (ROS) and DNA damage caused by targeting topoisomerase 2β in cardiomyocytes (3). The iron chelator, dexrazoxane (DXZ), is currently the only drug approved by the US Food and Drug Administration for the prevention of DIC in advanced breast cancer (4). DXZ reduces the production of iron-dependent oxidative free radicals by chelating intracellular iron and is also an inhibitor of topoisomerase 2β, which has a notable cardioprotective effect (5). Unfortunately, the use of DXZ is limited because of associated hazards such as secondary malignancies (6); therefore, alternative treatments to alleviate DIC are urgently needed.

Apoptosis and ferroptosis are the two main types of DIC-induced cardiomyocyte death (7–9). Apoptosis is the most well-characterized type of regulated cell death (RCD) and the most studied in DIC (10). DOX induces caspase-3 activation in cardiomyocytes through both the intrinsic and extrinsic pathways, thereby promoting cardiomyocyte apoptosis. Unlike classical RCD apoptosis, ferroptosis is a recently found RCD caused by the peroxidation of membrane phospholipids and iron-dependent cell death (11). Abnormal intracellular iron metabolism, excessive lipid peroxidation, and glutathione (GSH) deficiency occur during ferroptosis (12). Compared to the hearts from patients with non-DIC heart failure or healthy individuals, cardiac biopsy specimens from patients with DIC heart failure show excessive mitochondrial iron (13). Inhibition of ferroptosis or apoptosis has been demonstrated to be an effective method for DIC treatment in animal models (8, 14–16). Ferroptosis and apoptosis are two different types of RCD that may be affected by common factors, such as ROS levels and nuclear factor κB (NF-κB) inflammation (17, 18). Hence, the identification of previously unknown drugs that can simultaneously inhibit cardiomyocyte ferroptosis and apoptosis will be highly valuable for preventing DIC.

Epigenetic modifications, including DNA methylation, histone modifications, and noncoding RNA expression, are integral to regulating gene expression and cellular states. In DIC, DOX-induced stress disrupts gene expression by altering epigenetic modifications in cardiomyocytes (19). The search and development of small-molecule compounds that selectively target epigenetic-modifying enzymes represent a promising strategy for the treatment of DIC, although this aspect requires further research. As a structurally modified small-molecule compound derived from BIX-01338, an S-adenosylmethionine (SAM)–competing inhibitor of a methyltransferase, BRD4770 reduces the di- and trimethylation levels of lysine 9 on histone 3 (H3K9) and induces senescence in the pancreatic cancer cell line PANC-1 (20). BRD4770 and gossypol synergistically induce autophagy-related cell death in PANC-1 cells (21). In contrast, BRD4770 inhibits smooth muscle cell ferroptosis and protects against aortic dissection in mice by maintaining redox homeostasis and inhibiting the inflammatory response (22). These observations suggest that BRD4770 may exhibit disparate functions in different cells when subjected to survival stress. However, the exact function of BRD4770 in DIC-induced cardiomyocyte damage remains unclear.

In this study, we screened an epigenetic compound library and identified BRD4770 as a previously unidentified small-molecule compound that protects against DIC by inhibiting cardiomyocyte ferroptosis and apoptosis. Mechanistically, BRD4770 reduced DOX-induced oxidative stress and NF-κB inflammation via the nuclear factor erythroid 2–related factor 2 (Nrf2)/activating transcription factor 4 (ATF4)–solute carrier family 7 member 11 (SLC7A11) axis. BRD4770 administration protected the heart from dysfunction in DOX-challenged mice while maintaining the efficacy of DOX chemotherapy. DBMP (DSPE-PEG-NHS-DSPE-TK-PEG-BRD4770-MB-PCM), a ROS-responsive nanoliposome containing BRD4770 and a cardiac-targeting peptide, provided precise protection against DOX-induced cardiac dysfunction in mice. These observations indicate that BRD4770 is a promising antiferroptosis and antiapoptosis candidate drug for DIC and that DBMP provides a potential cardiac-targeted drug delivery modality.

RESULTS

Identification of BRD4770 as a potential compound for inhibiting DOX-induced cardiomyocyte death

To identify small-molecule compounds that inhibit DOX-induced cardiomyocyte injury, we screened an epigenetics compound library under DOX or erastin challenge in H9C2 cardiomyocytes (Fig. 1A). Erastin is a small-molecule inducer of ferroptosis by targeting the cystine-glutamate transport receptor and voltage-dependent anion channel. We first tested the concentrations of DOX and erastin that induced H9C2 cell death, and most cardiomyocytes were killed with 2.5 μM erastin or 1 μM DOX (fig. S1, A and B). Compounds that rescued a minimum of 40% of cell viability were considered potential small-molecule compounds. Eight compounds were identified by screening the top 25 compounds that demonstrated the ability to rescue DOX- or erastin-induced cardiomyocyte death (Fig. 1, B to D). Among the eight small-molecule compounds, pyrazolanthrone (SP600125) (23) has been investigated for its role in the DIC pathogenesis. The Cell Counting Kit-8 (CCK-8) assay results showed that BRD4770, among the seven previously uncharacterized small-molecule compounds, markedly suppressed both DOX- and erastin-induced H9C2 cell death (Fig. 1E). Moreover, BRD4770 protected against erastin- and DOX-induced H9C2 cell death in a dose-dependent manner (fig. S1, C and D), and pretreatment for 2 hours exhibited optimal protective efficacy (fig. S1, E to G). These results indicate that BRD4770 is a potential small-molecule compound with the capacity to effectively inhibit DOX- and erastin-induced cardiomyocyte death.

Fig. 1. BRD4770 rescues DOX- and erastin-induced cardiomyocyte death.

(A) Schematic of cell-based high-throughput screening of epigenetic compounds for protection against DOX/erastin-induced cardiomyocyte death. (B) The scatter plot of the effect of epigenetic compounds on DOX (top)– or erastin (bottom)–induced cardiomyocyte death. (C) Venn analysis of a drug compound. The top 25 compounds that protect against DOX- or erastin-induced cardiomyocyte death were selected separately, overlapping a total of 8 compounds. (D) The eight compounds that protect against both DOX- and erastin-induced cardiomyocyte death. Red indicates compounds that have been studied, and green indicates compounds that have not yet been reported in DIC. BDNF, brain-derived neurotrophic factor; JNK1, c-Jun N-terminal kinase 1; TrkA, tropomyosin receptor kinase A; Pim1, proviral integration site for Moloney murine leukemia virus-1 (Pim1) kinase; EGFR, epidermal growth factor receptor; PKA, cAMP-dependent protein kinase; PKC, protein kinase C; HDAC1, histone deacetylase 1. (E) The effects of J147 (5 μM), BRD4770 (5 μM), SCR7 pyrazine (5 μM), SMI-4a (5 μM), daphnetin (5 μM), BG45 (5 μM), and entinostat (5 μM) on cell viability in H9C2 cells after 24 hours of DOX or erastin treatment (n = 6). Statistical significance was determined using (E) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons. *P < 0.05 and **P < 0.01.

BRD4770 protects cardiomyocytes from erastin-induced ferroptosis

We determined the exact role of BRD4770 in erastin-induced ferroptosis in cardiomyocytes. BRD4770 treatment completely reversed erastin-induced cell death in H9C2 cells, as did the ferroptosis inhibitor ferrostatin-1 (Fer-1) (fig. S2A). Similarly, we isolated and identified neonatal rat ventricular myocytes (NRVMs) and found that BRD4770 markedly reversed the erastin-induced decrease in cell viability in NRVMs (fig. S2B). Next, we assessed cell morphology and viability using calcein-AM and propidium iodide (PI) staining. Both bright field (BF) and calcein-AM/PI fluorescence imaging showed that BRD4770 markedly inhibited erastin-induced cell death (fig. S2C). Given that intracellular ROS generation and lipid peroxidation accumulation are hallmarks of ferroptosis, we quantified ROS generation by the reactive oxygen species probe dichloro-dihydro-fluorescein diacetate (DCFH-DA) fluorescence, performed direct quantification of lipid peroxidation by the lipid peroxidation probe C11-BODIPY 581/591 staining, and assessed lipid peroxidation indirectly by determining malondialdehyde (MDA) levels, a terminal byproduct of lipid oxidation, to complement these direct measurements. BRD4770 treatment also markedly suppressed erastin-induced ROS production (fig. S2, D to G), increased the ratio of reduced GSH to oxidized GSH (GSSG) (fig. S2H), decreased erastin-induced lipid peroxidation (fig. S2, I and J), and reduced lipid peroxidation–derived MDA generation (fig. S2K) in erastin-treated H9C2 cardiomyocytes. Accordingly, treatment with BRD4770 reversed the expression of ferroptosis-related markers in erastin-conditioned cardiomyocytes, including the mRNA levels of the ferroptosis-related marker genes heme oxygenase-1 (Hmox1) and prostaglandin-endoperoxide synthase 2 (Ptgs2) (fig. S2L) and the protein levels of glutathione peroxidase 4 (GPX4) and SLC7A11 (fig. S2, M and N). Thus, BRD4770 rescued erastin-induced ferroptosis by reducing ROS production and lipid peroxidation accumulation and maintaining GSH homeostasis in cardiomyocytes.

BRD4770 attenuates DOX-induced ferroptosis and apoptosis in cardiomyocytes

Next, we examined the effects of BRD4770 on DOX-induced ferroptosis in cardiomyocytes. BRD4770 treatment markedly rescued the DOX-induced decrease in the viability of both H9C2 cells (Fig. 2A) and NRVMs (Fig. 2B). As shown by the morphology and calcein-AM/PI staining, BRD4770 inhibited DOX-induced cell death in H9C2 cells (Fig. 2, C and D). BRD4770 treatment also markedly suppressed DOX-induced ROS production (Fig. 2, E to H), increased the ratio of GSH to GSSG (Fig. 2I), and decreased DOX-induced lipid peroxidation (Fig. 2, J and K) and MDA levels (Fig. 2L), by down-regulating the mRNA levels of Hmox1 and Ptgs2 (Fig. 2M) and up-regulating the protein levels of GPX4 and SLC7A11 (Fig. 2, N and O) in DOX-treated H9C2 cardiomyocytes.

Fig. 2. BRD4770 reduces DOX-induced ferroptosis in cardiomyocytes.

(A and B) The effect of BRD4770 (5 μM) on DOX (1 μM)–induced cell death in H9C2 cells (A) and NRVMs (B) (n = 6). H9C2 cells or NRVMs were pretreated with BRD4770 for 2 hours and then stimulated with DOX for another 24 hours. (C and D) Representative images of bright field (C) and Calcein-AM/PI staining (D) showing that BRD4770 prevented DOX-induced cell death in H9C2 cells. Green: Calcein-AM; red: PI. Scale bars, 200 μm. (E and F) The representative flow cytometry images (E) and quantification (F) of DCFH-DA staining in H9C2 cells challenged with DOX and/or BRD4770 (n = 3). (G and H) The representative fluorescence images (G) and quantification (H) of DCFH-DA staining in H9C2 cells challenged with DOX and/or BRD4770. Green: DCFH-DA. Scale bars, 200 μm (n = 6). (I) The GSH/GSSG levels in H9C2 cells treated with DOX and/or BRD4770 (n = 6). (J and K) The representative images (J) and quantification (K) of C11-BODIPY staining in H9C2 cells challenged with DOX and/or BRD4770. Green: Oxidized BODIPY; red: BODIPY; blue: Hoechst. Scale bars, 100 μm (n = 6). (L) The MDA levels in H9C2 cells treated with DOX and/or BRD4770 (n = 6). (M) The relative mRNA levels of Hmox1 and Ptgs2 in H9C2 cells treated with DOX and/or BRD4770 (n = 4). (N and O) Western blot analysis (N) and quantification (O) of the relative protein levels of GPX4 and SLC7A11 in H9C2 cells treated with DOX and/or BRD4770 (n = 4). Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons. *P < 0.05 and **P < 0.01.

As previously reported, DOX induces both ferroptosis and apoptosis in cardiomyocytes (7, 8). We verified that ferroptosis inhibitors (Fer-1 and deferoxamine) and an apoptosis inhibitor (Z-VAD-FMK) reduced DOX-induced H9C2 cardiomyocyte death (Fig. 3A). Notably, the proportion of DOX-induced cell death rescued by BRD4770 was greater than that rescued by either the ferroptosis inhibitor or the apoptosis inhibitor alone (Fig. 3A), indicating that BRD4770 reduced DOX-induced cardiomyocyte death via multiple mechanisms. Therefore, we examined the cell apoptosis rate using annexin V and PI double staining and terminal deoxynucleotidyl transferase (TdT)–mediated deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL) staining and detected the molecular marker, cleaved caspase-3. The apoptosis rate and cleaved caspase-3 levels were increased in DOX-challenged H9C2 cardiomyocytes (Fig. 3, B to G). BRD4770 treatment markedly reduced the apoptosis rate and decreased the cleaved caspase-3 levels in DOX-treated H9C2 cardiomyocytes (Fig. 3, B to G). Together, these data indicated that BRD4770 inhibited DOX-induced ferroptosis and apoptosis in cardiomyocytes.

Fig. 3. BRD4770 reduces DOX-induced apoptosis in cardiomyocytes.

(A) The effect of BRD4770 (5 μM), Fer-1 (2 μM), DFO (100 μM), and Z-VAD (10 μM) on cell viability of H9C2 cells treated with DOX (1 μM) for 24 hours (n = 4 to 5). (B and C) The representative flow cytometric profiles (B) and percentages (C) of annexin V–fluorescein isothiocyanate (FITC)⁺ cells from H9C2 cells treated with DOX and/or BRD4770 (n = 3). (D and E) The representative images (D) of TUNEL staining and percentages of TUNEL-positive cells (E) in H9C2 cells treated with DOX and/or BRD4770. Green: TUNEL; blue: 4′,6-diamidino-2-phenylindole (DAPI). Scale bars, 50 μm (n = 6). (F and G) Western blot analysis (F) and quantification (G) of the levels of cleaved caspase-3 in H9C2 cells treated with DOX and/or BRD4770 (n = 6). Statistical significance was determined using (A) one-way ANOVA with Tukey’s multiple comparisons and (C, E, and G) two-way ANOVA with Tukey’s multiple comparisons. **P < 0.01.

BRD4770 protects against DOX-induced cardiac dysfunction and ferroptosis in mice

To determine whether BRD4770 has a protective effect against DOX-induced cardiac injury, we induced chronic DIC in 8-week-old C57BL/6J male mice and treated them with BRD4770 (Fig. 4A). Compared with the saline group, the DOX-challenged mice showed deteriorated cardiac function and changes in cardiac morphology, manifested by a gradual decrease in the left ventricular ejection fraction (LVEF; 74.37 ± 1.13% versus 49.24 ± 3.10%, P < 0.01) and fractional shortening (LVFS; 42.18 ± 0.98% versus 24.27 ± 1.94%, P < 0.01) (Fig. 4, B and C), decreased heart weight/tibial length (HW/TL) ratio (7.13 ± 0.47 mg/mm versus 5.31 ± 0.33 mg/mm; P < 0.01; Fig. 4D), substantially reduced heart size (Fig. 4E), increased cardiac fibrosis (a typical pathological feature of DIC) (Fig. 4, E and F), and significantly elevated mRNA levels of cardiac stress genes, including natriuretic peptide (Nppa), brain natriuretic peptide (Nppb), and myosin heavy chain 7 (Myh7) (Fig. 4, G to I). In addition, DOX challenge resulted in increased levels of the lipid peroxidation product MDA (Fig. 4J), a decline in the GSH/GSSG ratio (Fig. 4K), and a significant increase in the mRNA levels of Ptgs2 and Hmox1 (Fig. 4, L and M). BRD4770 treatment markedly alleviated DOX-induced cardiac dysfunction, as evidenced by improved cardiac function (Fig. 4, B and C), increased HW/TL ratio and heart size (Fig. 4, D and E), reduced cardiac fibrosis (Fig. 4, E and F), and decreased mRNA levels of Nppa, Nppb, and Myh7 (Fig. 4, G to I). Treatment with BRD4770 also counteracted DOX-induced increases in ferroptosis-related markers, including decreased MDA levels (Fig. 4J), increased GSH/GSSG ratios (Fig. 4K), and decreased mRNA levels of Ptgs2 and Hmox1 (Fig. 4, L and M). Notably, hematoxylin and eosin (H&E) staining of the liver, spleen, lungs, and kidneys showed that BRD4770 administration caused negligible side effects on the major organs of the mice compared with the vehicle group (fig. S3). Furthermore, BRD4770 treatment attenuated DOX-induced structural abnormalities in the hepatic and renal parenchyma, as evidenced by preserved tissue architecture (fig. S3).

Fig. 4. BRD4770 alleviates DOX-induced cardiac injury in mice.

(A) Schematic diagram of BRD4770 treatment of DOX-challenged mice. (B) Representative M-mode echocardiographic images of indicated groups at 3 weeks after DOX challenge in mice. (C) The values of LVEF and LVFS on the indicated days in DOX-challenged mice treated with BRD4770 (n = 6). (D) The ratio of HW/TL in BRD4770-treated mice 21 days after DOX challenge (n = 6). (E) Representative images of hearts, and H&E staining and Masson’s trichrome staining of heart sections. Scale bars, 1 mm (H&E) or 50 μm (Masson’s trichrome). (F) Quantification of the cardiac fibrosis area from BRD4770-treated mice 21 days after DOX challenge (n = 6). (G to I) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis of the mRNA levels of Nppa (G), Nppb (H), and Myh7 (I) in cardiac tissues (n = 6). (J and K) The MDA levels (J) and GSH/GSSG ratios (K) in cardiac tissues (n = 6). (L and M) qRT-PCR analysis of the mRNA levels of Ptgs2 and Hmox1 in cardiac tissues (n = 6). Statistical significance was determined using two-way ANOVA with Šidák’s multiple comparisons (C) and two-way ANOVA with Tukey’s multiple comparisons [(D) and (F) to (M)]. **P < 0.01.

To further investigate whether BRD4770 retains its protective effect after high-dose DOX-induced cardiac injury, we examined the effects of BRD4770 treatment on high-dose DOX-induced cardiac injury (2 and 5 μM) in H9C2 cardiomyocytes (fig. S4A) and high-dose DOX-induced cardiac injury (25 mg/kg) in C57BL/6J mice (fig. S4B). In H9C2 cardiomyocytes, cell viability decreased obviously with increasing DOX concentration, and BRD4770 treatment effectively attenuated the cell viability decline during DOX challenge at different doses (fig. S4A). Compared with the saline group, high-dose DOX-challenged mice exhibited a reduced survival rate (fig. S4C) and impaired cardiac function, evidenced by decreases in the LVEF (74.45 ± 0.70% versus 41.99 ± 2.89%, P < 0.01) and LVFS (41.88 ± 0.69% versus 19.72 ± 1.42%, P < 0.01) (fig. S4, D and E). BRD4770 treatment markedly improved survival and alleviated cardiac dysfunction induced by high-dose DOX in mice (fig. S4, C to E), suggesting its potential utility against high-dose DOX-induced cardiac injury in vivo. Together, these findings suggest that BRD4770 treatment alleviated DOX-induced cardiac dysfunction and ferroptosis.

BRD4770 protects against DIC without affecting the antitumor efficacy of DOX

Next, we tested whether BRD4770 affected the efficacy of DOX chemotherapy in tumors. We selected three common breast cancer cell lines to examine the effects of BRD4770 on DOX-induced cell death: metastatic mouse 4T1, nonmetastatic human MCF-7, and metastatic human MDA-MB-231 cells. BRD4770 significantly reduced the viability of 4T1, MCF-7, and MDA-MB-231 cells both in the presence and absence of DOX (Fig. 5A). To further determine whether BRD4770 administration affects the antitumor efficacy of DOX in vivo, we pretreated BALB/c mice with BRD4770 7 days after implantation of 4T1 breast cancer cells, and then the mice received DOX chemotherapy 2 days later (Fig. 5B). Consistent with the above observations in the C57BL/6J DIC model, DOX administration elicited a processed reduction in body weight in BALB/c mice (Fig. 5C). Comprehensive echocardiographic assessment revealed impaired cardiac function (Fig. 5D), evidenced by diminished LVEF (71.35 ± 1.88% versus 57.43 ± 6.44%, P < 0.01; Fig. 5E) and LVFS (39.54 ± 1.51% versus 29.27 ± 4.29%, P < 0.01; Fig. 5F), accompanied by cardiac atrophy and fibrosis (Fig. 5, G and H). As expected, the administration of BRD4770 rescued DOX-induced weight loss, cardiac dysfunction, and cardiac fibrosis in tumor-bearing mice (Fig. 5, C to H). BRD4770 administration did not alter the inhibitory effect of DOX on tumor volume, indicating that BRD4770 had no effect on the antitumor efficacy of DOX in vivo (Fig. 5, I and J). Collectively, these results suggest that BRD4770 administration prevented DOX-induced cardiac injury without interfering with the chemotherapeutic effects of DOX.

Fig. 5. BRD4770 alleviates DOX-induced cardiac injury without affecting the antitumor efficacy of DOX in mice.

(A) Effect of BRD4770 (5 μM) on the cell viability of 4T1, MCF-7, and MDA-MB-231 cells after DOX (1 μM) treatment (n = 6). (B) Schematic diagram of BRD4770 treatment of tumor-bearing mice with DOX chemotherapy. (C) Changes in body weight in indicated groups of tumor-bearing mice (n = 6). (D) Representative M-mode echocardiographic images of indicated groups of tumor-bearing mice 15 days after DOX chemotherapy. (E and F) Effect of BRD4770 on LVEF (E) and LVFS (F) in tumor-bearing mice 15 days after DOX chemotherapy (n = 4). (G) Representative images of H&E staining and Masson’s trichrome staining of heart sections. Scale bars, 1 mm (H&E) or 50 μm (Masson’s trichrome). (H) Quantification of the cardiac fibrosis area in (G) (n = 6). (I) Representative tumor images of all the mice at the end of the study. (J) Effect of BRD4770 on tumor volume in DOX chemotherapy–treated tumor-bearing mice (n = 6). n.s., not significant. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons [(A), (E), (F), and (H)] and two-way ANOVA with Šidák’s multiple comparisons [(C) and (J)]. *P < 0.05 and **P < 0.01.

BRD4770 improves DOX-induced oxidative stress and inflammatory responses in cardiomyocytes

To further explore the mechanisms underlying the beneficial effects of BRD4770 in DIC, we performed transcriptome sequencing [RNA sequencing (RNA-seq)] analysis on cardiomyocytes treated with DOX in combination with BRD4770 (Fig. 6A). Compared with the dimethyl sulfoxide (DMSO) group, 11,509 differentially expressed genes (DEGs) were identified after DOX treatment [P ≤ 0.05, fold change (FC) ≥ 1.5], including 7723 up-regulated genes, such as typical ferroptosis and oxidative stress–related genes (Gadd45g, Il6, Cxcl1, Ptgs2, Ptgs1, and Alox15), and 3786 down-regulated genes (fig. S5A). In the gene enrichment analysis of DOX–up-regulated genes, arachidonic acid metabolism associated with the induction of ferroptosis was among the top 20 terms enriched in the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis (fig. S5B); gene set enrichment analysis (GSEA) showed that DOX treatment significantly activated the NF-κB pathway (fig. S5C). KEGG pathway analysis of BRD4770–up-regulated genes identified both cardiomyocyte-specific pathways (calcium signaling pathway and cytoskeleton in muscle cells) and potential antitumor pathways (glioma and p53 signaling pathway; fig. S5D), as well as tumor-suppressive genes within these pathways (e.g., Cdkn1a, Gadd45g, and Rb1; fig. S5E), highlighting cell specificity while suggesting possible antitumor mechanisms. Compared to the DOX group, the DOX + BRD4770 group had 1126 DEGs (P ≤ 0.05, FC ≥1.5), including 498 up-regulated genes and 628 down-regulated genes (Fig. 6B); GSEA of these DEGs showed that the NF-κB signaling pathway was down-regulated (Fig. 6C) and the cell redox homeostasis signaling was up-regulated in the DOX + BRD4770 group (Fig. 6D). Further comprehensive analysis identified 413 DEGs that were up-regulated in the DOX group compared to the DMSO group and down-regulated in the DOX + BRD4770 group compared to the DOX group (Fig. 6E). KEGG analysis showed that the NF-κB signaling pathway and arachidonic acid metabolism related to ROS production and lipid peroxidation were enriched (Fig. 6F). We then cross-referenced cell death–related genes with RNA-seq genes enriched for oxidative stress and inflammatory signaling and found 18 typical cell death–related genes (Fig. 6G). Furthermore, we selected classical cell death–promoting genes, including Gadd45g, Ptgs1, Ptgs2, Cxcl1, Il-6, and Cyp1b1, and quantitative reverse transcription polymerase chain reaction (qRT-PCR) also confirmed that BRD4770 treatment markedly inhibited the DOX-enhanced expression of these genes in H9C2 cells (Fig. 6H). In addition, GSEA showed that BRD4770 markedly alleviated the DOX-induced reduction in oxidative phosphorylation (fig. S5, F and G), which is an important indicator of the amelioration of DOX-induced cardiomyocyte injury. Together, these results indicated that BRD4770 mitigated DOX-induced cardiomyocyte injury by alleviating oxidative stress and inflammatory responses.

Fig. 6. BRD4770 improves DOX-induced cardiomyocyte injury by attenuating oxidative stress and inflammatory responses.

(A) Experimental diagram of the RNA-seq in cardiomyocytes treated with BRD4770 and DOX. (B) Volcano plot showing the DEGs between the DOX + BRD4770 group and the DOX group. P adj, adjusted P. (C and D) GSEA plot showing the enrichment of down-regulated genes in “NF-κB signaling pathway” (C) and up-regulated genes in “cell redox homeostasis” (D) signature. (E) Venn diagram of genes down-regulated by BRD4770 and up-regulated by DOX obtained from the comparison of the DOX group and the DMSO group and the DOX + BRD4770 group and the DOX group. (F) Top 20 KEGG enrichment analysis of genes down-regulated in the DOX + BRD4770 group compared to the DOX group. IL-17, interleukin-17. TGF-β, transforming growth factor–β;NOD-like receptor, nucleotide-binding and oligomerization domain (NOD)-like receptor. (G) Heatmap of NF-κB signaling pathway, arachidonic acid metabolism, and cell redox homeostasis–related mRNA profile changes in DMSO, BRD4770, DOX, and DOX + BRD4770 groups. (H) qRT-PCR analysis of the mRNA levels of Gadd45g, Ptgs1, Ptgs2, Cxcl1, Il-6, and Cyp1b1 in H9C2 cells treated with BRD4770 in the presence of DOX (n = 6). Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons. **P < 0.01.

BRD4770 inhibits cardiomyocyte ferroptosis through the Nrf2/ATF4-SLC7A11 signaling pathway

The cystine transporter SLC7A11 is a key molecule in ferroptosis and apoptosis (24), and inhibition of DOX-induced SLC7A11 down-regulation may ameliorate DIC (25). Our results showed that BRD4770 treatment significantly increased the SLC7A11 protein levels (Fig. 2, M and N, and fig. S2, M and N). BRD4770 also increased SLC7A11 mRNA levels in H9C2 cells, suggesting that BRD4770 promoted SLC7A11 gene expression (Fig. 7A). In ferroptotic stress, SLC7A11 transcription is directly regulated by two key transcription factors: Nrf2 and ATF4 (8, 26). qRT-PCR results showed that BRD4770 treatment increased Nrf2 and ATF4 mRNA levels in H9C2 cells (Fig. 7B). RNA-seq and qRT-PCR results indicated that BRD4770 promoted the expression of Nrf2- and ATF4-dependent genes (Fig. 7, C and D). Considering potential NAD(P)H interference in viability assays due to Nrf2/ATF4 modulation, we used NAD(P)H-independent methods, trypan blue staining, and adenosine 5′-triphosphate (ATP) quantification, which consistently confirmed BRD4770’s protection against erastin- or DOX-induced cardiomyocyte death (fig. S6, A to D). Notably, elevated ATF4 levels are often accompanied by endoplasmic reticulum (ER) stress (27). We found that the BRD4770-induced increase in ATF4 expression was independent of ER stress, as evidenced by the absence of an increase in ER stress–related genes (fig. S7, A and B).

Fig. 7. BRD4770 alleviates erastin-induced ferroptosis by Nrf2/ATF4-SLC7A11 pathway in vitro.

(A and B) qRT-PCR analysis of the mRNA levels of Slc7a11 (A) and Atf4 and Nfe2l2 (B) in H9C2 cells treated with BRD4770 (5 μM) for 24 hours (n = 6). (C) Heatmap of ROS-related ATF4- and Nrf2-dependent genes from RNA-seq data. (D) qRT-PCR analysis of the mRNA levels of Phgdh, Psat1, Psph, Cth, Chac1, Mthfd2, Shmt2, and Asns in H9C2 cells (n = 6). (E) Effect of ML385 (5 μM) or/and ISRIB (5 μM) on the cell viability in BRD4770-treated H9C2 cardiomyocytes (n = 6). (F and G) The representative fluorescence images (F) and quantification (G) of DCFH-DA staining in BRD4770-treated H9C2 cardiomyocytes. Green: DCFH-DA. Scale bars, 200 μm (n = 6). (H and I) The representative fluorescence images (H) and quantification (I) of C11-BODIPY staining in BRD4770-treated H9C2 cardiomyocytes. Green: Ox-BODIPY; red: BODIPY; blue: Hoechst. Scale bars, 200 μm (n = 6). (J and K) Western blot analysis (J) and quantification (K) of the levels of histone H3 lysine 9 mono-methylation(H3K9me1), histone H3 lysine 9 di-methylation(H3K9me2), and histone H3 lysine 9 tri-methylation(H3K9me3) in H9C2 cells (n = 3). (L and M) Western blot analysis (L) and quantification (M) of the protein levels of H3K9me3 in H9C2 cells (n = 4). Statistical significance was determined using Mann-Whitney U tests [(A), (B), (D), (E), (G), (I), and (K)] and two-way ANOVA with Tukey’s multiple comparisons (M). **P < 0.01.

On the basis of the above results, we speculated that BRD4770 regulates the transcription of SLC7A11 in cardiomyocytes through Nrf2 or ATF4. To test this hypothesis, we selected ML385 (an inhibitor of Nrf2) and ISRIB (an inhibitor of ATF4) to treat cardiomyocytes with erastin-induced ferroptosis rescued by BRD4770. We found that the protective effect of BRD4770 against erastin-induced cardiomyocyte death was largely abolished by inhibition of ML385 alone, partially abolished by inhibition of ATF4 alone, and completely abolished by simultaneous inhibition of Nrf2 and ATF4 (Fig. 7E). Consistently, the inhibitory effect of BRD4770 on erastin-induced cell death (fig. S7C), ROS production (Fig. 7, F and G), and lipid peroxidation (Fig. 7, H and I) was partially eliminated by the ML385 inhibitor or ATF4 inhibitor alone, whereas these indicators were completely eliminated by the combined inhibition of Nrf2 and ATF4. Furthermore, BRD4770-promoted SLC7A11 expression and GPX4 levels were partially suppressed by the Nrf2 or ATF4 inhibitor and completely suppressed by the combination of Nrf2 and ATF4 inhibitors in H9C2 cells (fig. S7, D to F). These data suggest that Nrf2/ATF4-SLC7A11 signaling was responsible for the protective effects of BRD4770 on cardiomyocytes under stress.

BRD4770 was developed as an inhibitor of histone methyltransferase euchromatic histone lysine N-methyltransferase 2 (EHMT2; also known as G9A) (20); therefore, we focused on the effect of BRD4770 on histone methylation, to further explore the molecular mechanisms of BRD4770-mediated inhibition of cardiomyocyte death and alleviation of DIC in mice. H3K9 methylation in cardiomyocytes was notably inhibited by BRD4770 treatment (Fig. 7, J and K). Moreover, BRD4770 treatment eliminated DOX- or erastin-induced increases in histone H3 lysine 9 tri-methylation (H3K9me3) levels in H9C2 cells (Fig. 7, L and M, and fig. S7, G and H).

DBMP administration attenuates DOX-induced cardiac dysfunction

Next, we explored the potential applications of BRD4770 in DIC. We constructed a BRD4770-loaded nanoliposome (DBMP) conjugated to the cardiomyocyte-targeting peptide (PCM; Fig. 8A). DBMP was formed using the major components 1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE)–thioketal (TK)–polyethylene glycol (PEG), DSPE–PEG–N-hydroxysuccinimide (NHS), BRD4770, methylene blue (MB), cholesterol (CHOL), and hydrogenated soya phosphatidylcholines (HSPCs), followed by surface conjugation with PCM. The TK bond in DSPE-TK-PEG responds to the ROS environment, including oxidation in a high-concentration ROS environment, thereby disrupting the hydrophilic shell of the nanoliposome for drug release. MB is a fluorescent dye that is used to track the distribution of nanoliposomes. The PCM (WLSEAGPVVTVRALRGTGSW) is a 20–amino acid peptide that binds specifically to tenascin-X (TNX) in cardiomyocytes (28, 29).

Fig. 8. DBMP alleviates DOX-induced cardiac injury in mice.

(A) Schematic of the DMP (DSPE-PEG-NHS-DSPE-TK-PEG-MB-PCM) and DBMP design. HSPC, CHOL, DSPE-TK-PEG, DSPE-PEG-NHS, MB, and CeO₂ synthesize DM (DSPE-PEG-NHS-DSPE-TK-PEG-MB), and DM combines with PCM to form DMP. HSPC, CHOL, DSPE-TK-PEG, DSPE-PEG-NHS, MB, CeO₂, and BRD4770 synthesize DBM (DSPE-PEG-NHS-DSPE-TK-PEG-BRD4770-MB), and DBM combines with PCM to form DBMP. (B) Effect of DMP and DBMP (0.5 parts per million) on the cell viability in DOX-treated H9C2 cells (n = 6). (C) Schematic diagram of DBMP treatment of DOX-challenged mice. (D) The survival curves of indicated group of mice (n = 7 to 18). (E) Body weight changes in DOX-challenged mice (n = 6 to 8). (F) Representative M-mode echocardiographic images of the indicated group at 3 weeks after DOX challenge in mice (n = 6). (G and H) The values of LVEF (G) and LVFS (H) on the indicated days in DOX-challenged mice (n = 6). (I) The ratio of HW/TL in mice (n = 6). (J) Representative images of hearts, and H&E staining and Masson’s trichrome staining of heart sections from mice. Scale bars, 1 mm (H&E) or 50 μm (Masson’s trichrome). (K) Quantification of the cardiac fibrosis area from mice (n = 6). Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons [(I) and (K)] and two-way ANOVA with Šidák’s multiple comparisons [(E), (G), and (H)]. *P < 0.05 and **P < 0.01.

DMP (DSPE-PEG-NHS-DSPE-TK-PEG-MB-PCM) and DBMP showed regular spherical morphologies and core-shell structures with good dispersibility, as revealed by transmission electron microscopy (TEM) (fig. S8A). The protein content in liposomes was measured using a protein assay kit and was much higher in DMP and DBMP than in DM (DSPE-PEG-NHS-DSPE-TK-PEG-MB) (fig. S8B). This confirmed the successful conjugation of PCM to DMP and DBMP. DM, DMP, and DBMP had sizes of 43.9, 59.1, and 59.1 nm, respectively (fig. S8C). The average zeta potential values of DM, DMP, and DBMP were −36.0 ± 1.1, −32.0 ± 1.0, and −34.8 ± 0.7 mV, respectively (fig. S8D). The zeta potentials of DM, DMP, and DBMP were similar and within the range of −40 to −30 mV, due to their similar surface chemistries after fabrication. The absorbance properties of DM and DMP were similar, and the characteristic peak of MB was observed at 664 nm, indicating that modifying PCM did not affect the optical properties of DMP. In addition, the characteristic absorption of DBMP at 330 nm was observed in the ultraviolet-visible spectrum, indicating the successful loading of BRD4770 (fig. S8E). The fluorescence properties of these nanoliposomes were consistent with a clear emission at 620 to 649 nm (fig. S8F). To further test the release percentages of ROS-responsive DBMP, we examined the ability of DBMP to release BRD4770 within the DOX-generated ROS environment and found that both DMP and DBMP exhibited a collapsed core-shell structure and fragmented morphology (fig. S8G), and DBMP released approximately 60.2% of BRD4770 at 1 hour and ≥95.4% at 2 hours (fig. S8H). PCM specifically binds to TNX in cardiomyocytes without interfering with normal cellular functions. We labeled DM and DMP with the fluorescent material MB and observed their targeting of H9C2 cells. Flow cytometry confirmed that the proportion of MB-positive H9C2 cells was notably higher in the DBMP group than in the DM group (fig. S8I), indicating that PCM-modified DBMP had better cardiomyocyte-targeting properties. In addition, we performed in vivo studies in mice to investigate the myocardial targeting of DBMP. MB fluorescence was markedly enriched in the hearts of DBMP-treated mice compared with that in the saline and DM groups 12 hours after drug administration (fig. S8J). In conclusion, ROS-responsive liposomal DBMP with PCM-targeting peptide modification was myocardial targeted and suitable for further functional studies.

We explored the effects of DBMP on DOX-challenged cells. Treatment with DMP or DBMP had no effect on cell viability (Fig. 8B), survival, ROS, or lipid peroxidation levels in H9C2 cells (fig. S9, A to E). During DOX stress, the DBMP-treated group showed increased cell viability and survival and reduced ROS levels and lipid peroxidation compared with the DMSO and DMP groups (Fig. 8B and fig. S7, A to E). These results suggest that DBMP, similar to BRD4770, reduced DOX-induced cardiomyocyte injury in vitro. We then assessed the preventive effect of DBMP on myocardial injury in a mouse model of DIC, with DXZ treatment as the positive control (Fig. 8C). As expected, the survival rate (Fig. 8D), body weight (Fig. 8E), and cardiac function (Fig. 8, F to H) were significantly decreased in DOX-challenged mice compared to those in the saline group. Moreover, DOX-treated mice exhibited lower HW/TL ratios (Fig. 8I), smaller hearts, and excessive cardiac collagen deposition in cardiac tissues (Fig. 8, J and K). Notably, similar to the DOX + DXZ group, the DOX + DBMP group showed a markedly improved survival rate, body weight, cardiac function, HW/TL ratios, cardiac morphology, and cardiac fibrosis compared to the DOX+ vehicle or DOX + DMP groups (Fig. 8, D to K). In addition, H&E staining showed that DBMP had no obvious effect on the major organs of mice (fig. S9F). These results confirm the efficacy of DBMP in a murine model of DIC and suggest that DBMP may be a viable method for the targeted cardiac delivery of BRD4770.

DISCUSSION

DIC is one of the most serious adverse effects of DOX in clinical applications, and ideal drugs to prevent adverse cardiac effects are lacking (2, 3, 10). We found that BRD4770 inhibited DOX-induced cardiomyocyte apoptosis and ferroptosis by screening the Epigenetics Compound Library and performing a series of molecular and cellular experiments. BRD4770 inhibited DOX-induced oxidative stress and lipid peroxidation in cardiomyocytes via the Nrf2/ATF4-SLC7A11 pathway. BRD4770 treatment protected the heart against DIC without affecting the efficacy of DOX chemotherapy in tumor-bearing mice. DBMP targeted the injured heart and sustainably released BRD4770, thereby effectively preventing DIC. Collectively, our results demonstrate that BRD4770 alleviates DIC by inhibiting apoptosis and ferroptosis through the Nrf2/ATF4-SLC7A11 pathway, which may provide fresh perspectives for the development of promising drug candidates for DIC prevention (Fig. 9).

Fig. 9. Schematic diagram of BRD4770 attenuating DOX-induced cardiac injury through the Nrf2/ATF4/SLC7A11 pathway.

BRD4770 was identified by high-throughput cell-based screening and then modified to produce a cardiac-targeting nanoliposome DBMP (top and white background). In cardiomyocytes, DOX-induced stress promoted the increase in H3K9 methylation (SAM as the methyl donor), which inhibits the expression of two key transcription factors Nfe2l2 and Atf4 in SLC7A11, reduces cystine transport, and restricts GSH synthesis, resulting in intracellular ROS production, lipid peroxidation accumulation, and inflammatory activation, which, in turn, leads to ferroptosis and apoptosis in cardiomyocytes. DIC causes the progressive loss of cardiomyocytes, leading to cardiac damage and interstitial fibrosis, and ultimately to cardiac dysfunction in mice (bottom left). BRD4770 treatment promoted the transcription of Nfe2l2 and Atf4 by inhibiting the methylation of H3K9, thereby promoting the expression of SLC7A11, which increases cystine transport and GSH synthesis; providing sufficient reducing agent GSH for GPX4 to eliminate ROS and lipid peroxidation caused by DOX, which ultimately reduced ferroptosis and apoptosis of cardiomyocytes, thereby inhibiting the loss of cardiomyocytes and reducing interstitial fibrosis; and ameliorating DOX-induced cardiac dysfunction in mice (bottom middle). DBMP, derived from modified BRD4770, specifically targets cardiomyocytes by binding to TNX receptors located on the cardiomyocyte membrane. In response to DOX-induced ROS, intracellular DBMP releases its loaded BRD4770, reducing ferroptosis and apoptosis in cardiomyocytes, and thereby improving cardiac function after DIC (bottom right). PUFAs-OH, non-toxic lipid-alcohols; PUFAs-OOH, lipid-hydroperoxides. Created in BioRender (J.S., 2025; https://BioRender.com/wbnj5cu).

At present, the molecular mechanism of DIC remains inconclusive, and the more recognized mechanisms include ROS production, calcium dysregulation, mitochondrial dysfunction, and cardiomyocyte death (2, 30). ROS production and other DOX-induced stressors lead to multiple types of cardiomyocyte death including cardiomyocyte apoptosis and ferroptosis (10). Previous studies have demonstrated that large amounts of ROS are generated during the development of DIC and that the inhibition of cardiomyocyte ferroptosis or apoptosis notably ameliorates DIC (8, 9, 14). The inhibition of cardiomyocyte death and improvement of DIC have been research hotspots in the field of cardiooncology. Cardiomyocyte-specific knockout of a disintegrin and metalloproteinase 17 decreases cardiomyocyte apoptosis and left ventricular fibrosis induced by DIC in mice by inhibiting tumor necrosis factor (TNF)–TNF receptor–associated factor 3–transforming growth factor β–activated kinase–mitogen-activated protein kinase signaling (31). THZ1, a small-molecule selective inhibitor of cyclin-dependent kinase 7, alleviates DOX-induced cardiac dysfunction, protects against cardiomyocyte apoptosis, and enhances the anticancer efficacy of DOX in a mouse model (32). Exercise training can promote the expression of Fc γ receptor IIB in B cells, which further regulates the signaling transduction and activation of B cells and improves the anti-inflammatory effect of B cells, thereby reducing DOX-induced cardiomyocyte apoptosis and attenuating DIC in mice (33). In this study, the small-molecule compound, BRD4770, effectively inhibited ferroptosis and apoptosis in cardiomyocytes, eliminated DOX-induced ROS production, and maintained GSH homeostasis. We demonstrated that BRD4770 notably improved DIC in mice but did not affect the antitumor effect of DOX, as previously reported (34).

Epigenetic modifications can rapidly respond to changes in cellular stress by altering gene expression to allow cells to better adapt to stressful environments; therefore, screening small molecules targeting epigenetic modification enzymes is valuable for identifying potential drugs to treat diseases, such as DIC. Among 773 small-molecule compounds targeting epigenetics, we found that BRD4770 is a promising small-molecule compound with the ability to inhibit DOX-induced ferroptosis and apoptosis in cardiomyocytes. BRD4770 is a methyltransferase inhibitor originally reported in 2012 that inhibits H3K9me2 and H3K9me3 catalyzed by G9a (also known as euchromatic histone methyltransferase 2) in pancreatic cancer PANC-1 cells (20). BRD4770 exhibits low cytotoxicity and does not induce cell apoptosis; however, in p53-mutated PANC-1 cells, BRD4770 and the putative BCL-2 homology 3 (BH3) mimetic gossypol act synergistically to induce cell death (20, 21). BRD4770 regulates Sendai virus infection–induced gene transcription in mouse embryonic fibroblasts through the inhibition of G9a (35) and promotes the production of more potent antioxidant and antimicrobial cryptic metabolites by the endophytic fungus Diaporthe longicolla through epigenetic regulation (36). In addition, BRD4770 promotes pyruvate dehydrogenase kinase 4 expression by inhibiting G9a-catalyzed H3K9me2/3 in hepatic cells (37). We found that BRD4770 exerts antioxidant and anti-inflammatory effects and protects cardiomyocytes from DOX-induced ferroptosis and apoptosis by inhibiting H3K9 methylation. We observed the synergistic effect of DOX and BRD4770 on breast cancer cell proliferation in vitro. We found that treatment with BRD4770 alone also had an antitumor effect in tumor-bearing mice, as previously reported (34). However, the synergistic effect was not observed in vivo, which may be due to pharmacokinetic barriers, the complexity of the tumor microenvironment, and dose limitations that prevent the synergistic interaction of DOX and BRD4770 in vivo.

ATF4 and Nrf2 are two key transcription factors that inhibit ferroptosis by promoting the expression of the cystine/glutamate transporter, SLC7A11. ATF4 is expressed in most mammalian cell types and belongs to the ATF/adenosine 3′,5′-monophosphate (cAMP) response element–binding protein family. ATF4 plays a crucial role in ensuring cell survival under stress conditions by regulating gene expression and is involved in various cellular responses to amino acid deprivation, ER stress, and DOX challenges (38–40). Inhibition of lysine-specific demethylase 1 down-regulates ATF4 expression by the epigenetic modification of H3K9me2, thereby inhibiting the expression of SLC7A11, reducing GSH production, and enhancing lipid peroxidation and ROS accumulation to induce ferroptosis in non–small-cell lung cancer cells (41). Nrf2 plays pivotal roles in maintaining intracellular redox homeostasis, regulating inflammatory responses, and determining cell survival by regulating the expression of multiple antioxidant genes (42, 43). A growing body of evidence has indicated that cardiomyocyte death and cardiac pathological changes in DIC are associated with the inhibition of Nrf2, and various clinical drugs and compounds have been shown to improve DIC by activating the Nrf2 signaling pathway (44). Our previous studies demonstrated that the E-prostanoid 1 receptor facilitates the expression of GPX4 and SLC7A11 through the activation of Nrf2, thereby protecting cardiomyocytes from DOX-induced ferroptosis (8). Moreover, BRD4770 treatment increases SLC7A11 expression by moderately inhibiting H3K9 methylation and increasing the binding of Nrf2 to the SLC7A11 promoter, thereby increasing SLC7A11 expression in lung cancer cells (45). However, the role and intrinsic mechanisms of action of BRD4770 in cardiomyocytes remain unknown. Here, we observed that BRD4770 promotes SLC7A11 expression via ATF4 and Nrf2 in DOX-treated cardiomyocytes.

While our findings demonstrate the cardioprotective potential of BRD4770 in DIC, several limitations warrant consideration. First, although BRD4770 demonstrated therapeutic effects against DIC in both C57BL/6J and BALB/c murine models in this study, there are differences between the strains. These variations are likely due to strain-specific physiological metabolic disparities (46), suggesting that the therapeutic potential of BRD4770 in human patients requires further exploration. Second, the long-term safety and potential off-target epigenetic effects of BRD4770 remain uncharacterized, requiring further investigation of chronic toxicity and epigenetic specificity. Third, while the cardiac targeting of DBMP is effective in mice, validation in larger animals with physiological ROS gradients is critical. Thus, validation of BRD4770/DBMP in large animal models with humanlike pharmacokinetics will accelerate clinical translation.

In summary, we identified a small-molecule compound, BRD4770, that can inhibit DOX-induced ferroptosis and apoptosis through the Nrf2/ATF4-SLC7A11 pathway in cardiomyocytes. In addition, we designed DBMP to facilitate the targeted delivery of BRD4770 to the damaged heart, providing a tool for future translational applications of BRD4770. We believe that BRD4770 is a promising compound that rescues cardiomyocyte death under stress and has substantial translational potential for the treatment of DIC and other cardiac diseases associated with cardiomyocyte death.

MATERIALS AND METHODS

Experimental design

This study systematically investigates epigenetic interventions against DIC using multiple in vitro and in vivo models. We used cardiomyocyte cultures and mouse models, combined with pharmacological profiling and RNA-seq, to delineate the cardioprotective mechanisms of BRD4770. To enhance therapeutic specificity, we developed cardiac-targeted ROS-responsive nanoliposomes (DBMP). Sample sizes reflect the number of independent biological replicates as annotated in the figure legends.

Animals

All mice were maintained in a specific pathogen–free room with a 12-hour/12-hour light/dark cycle at a controlled temperature, all animals’ experimental protocols were approved by the Institutional Animal Care and Use Committee of Tianjin Medical University, and all animal procedures conformed to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. For euthanasia, mice are anesthetized with 3 to 5% isoflurane for more than 5 min, followed by tissue collection after confirmation of the absence of an active paw reflex.

Construction of the mouse DIC model and administration of BRD4770 were based on slight modifications of previous reports (22, 47). To induce DOX cardiotoxicity in mice, 8-week-old C57BL/6J male mice were administered DOX (5 mg/kg) or saline via intraperitoneal injection once a week for three consecutive weeks, for a cumulative DOX dose of 15 mg/kg. In experiments using BRD4770 to prevent DIC, mice were treated with BRD4770 (1 mg/kg per day) or a corresponding volume of saline 2 days before DOX administration for three consecutive weeks via intraperitoneal injection. Cardiac function was assessed using echocardiography at the indicated time points, and cardiac tissue was harvested from the euthanized mice on day 22 of the DIC model for further analysis.

Construction of the mouse DIC model (high DOX dosage) was based on slight modifications of previous reports (48). Briefly, 8-week-old C57BL/6J male mice were administered DOX (5 mg/kg) or saline via intraperitoneal injection once weekly for five consecutive weeks, resulting in a cumulative DOX dose of 25 mg/kg. In experiments evaluating BRD4770 for treatment of DIC, mice were treated with BRD4770 (1 mg/kg per day) or a corresponding volume of saline via intraperitoneal injection starting 2 days before DOX administration and continued for five consecutive weeks. Cardiac function was assessed by echocardiography on day 35 of the DIC model.

To evaluate the effect of BRD4770 on the chemotherapeutic capacity of DOX in vivo, tumor-bearing mice were established. Briefly, 8-week-old female BALB/c mice were injected with 5 × 10⁵ 4T1 cells in the right mammary fat pad, followed by intraperitoneal injection of BRD4770 (1 mg/kg per day for 3 weeks) for 2 days before DOX injection (1.5 mg/kg per day) (49). Cardiac function was assessed after the last DOX injection, while the body weight and seeded tumor parameters were recorded every 3 days from the first DOX injection and calculated, as previously described and using the following formula: tumor volume = ¹/₂ × length × width × height.

To evaluate the preventive effect of DBMP on DIC in mice, 25g of 8-week-old C57BL/6J male mice were injected with 100 μl of DMP, DBMP at the MB concentration of 200 μg/ml via tail vein, or DXZ (20 mg/kg) intraperitoneally (50), or via injection with the corresponding volume of saline; DOX challenge was started 2 days later. On day 22 of the DIC model, the mice were euthanized, and cardiac tissues were harvested for subsequent studies.

Reagents and materials

DOX HCl (#S1208), deferoxamine mesylate (#S5742), Z-VAD-FMK (#S7023), ML385 (#S8790), and DXZ HCl (#S1222) were purchased from Selleck Chemicals (Houston, Texas, USA). SCR7 pyrazine (#HY-107845) and Fer-1 (#HY-100579) were purchased from MedChemExpress (Monmouth Junction, NJ, USA). BRD4770 (#T1923), J147 (#T1993), BG45 (#T2294), SMI-4a (#T3058), daphnetin (#T2851), entinostat (#T6233), ISRIB (#T2027), OTS186935 (#T12344), and UNC0638 (#T3257) were purchased from TargetMol (Wellesley Hills, MA, USA).

For DBMP synthetic materials, DSPE-PEG2000-NHS, DSPE-TK-PEG2000, HSPC, and CeO₂ were provided by Xi’an Ruixi Biological Technology Co. Ltd. (Xi’an, China). PCM (WLSEAGPVVTVRALRGTGSW) was synthesized by GL Biochem Ltd. (Shanghai, China). All other chemicals used to synthesize DBMP were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China).

Cell culture

The rat cardiomyocytes (H9C2), mouse breast cancer (4T1), and human breast cancer (MDA-MB-231) cell lines were purchased from the Zhong Qiao Xin Zhou Biotechnology Co. Ltd. (Shanghai, China). The human breast cancer MCF-7 cell line was provided by Lei Shi Laboratory (Tianjin Medical University). H9C2, 4T1,MCF-7, and MDA-MB-231 cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM; C11995500BT, Gibco, USA) supplemented with 10% fetal bovine serum (MN012103, Mengma, China) and penicillin-streptomycin (50 μg/ml; FG101-01, TransGen Biotech, China). All cells used in this study have been tested free of mycoplasma contamination.

Cell viability assay

Cell viability was determined using the CCK-8 kit (C6005, NCM Biotech, China). H9C2 cells were seeded in 96-well plates (8 × 10³ cells per well) and treated with DOX (1 μM) for 24 hours, to induce DOX-induced injury, or erastin (2.5 μM) for 24 hours, to induce cardiomyocyte ferroptosis. Inhibitor compounds used in this study, including BRD4770, J147, SMI-4a, SCR7, BG45, and Fer-1, were added 2 hours before DOX or erastin challenge. At the end of the treatment period, 100 μl of 10% CCK-8 solution was added to the corresponding wells, with the cell-free well serving as a blank control. After incubation for 2 hours at 37°C, the absorbance of each well was measured at 450 nm using a microplate reader (PerkinElmer, USA). The relative percentage of cell viability in each group was normalized to that of the control wells.

High-throughput screening

Rat H9C2 cardiomyocytes were seeded in 96-well plates (8 × 10³ cells per well) and cultured overnight. After cell adherence, 773 compounds from the Epigenetics Compound Library (L1200, TargetMol, USA) were added at a final concentration of 5 μM to the cell culture medium using the Explorer G3 Automated Drug Screening System (PerkinElmer, USA). After cell culture in an incubator for 2 hours, the cells were treated with DOX/erastin for an additional 24 hours, and the cell activity was detected using the CCK-8 assay.

Isolation and culture of NRVMs

NRVMs were isolated from 1- to 3-day-old neonatal Sprague-Dawley rat hearts according to the methods in our previous report (14). Briefly, neonatal rat ventricular tissues were rapidly extracted, cut into small pieces, and transferred to a 0.1% type II collagenase (Worthington, USA) solution for repeated digestion of the cardiac tissue, and individual cells were harvested. The fibroblasts were separated and removed using the differential adhesion method to obtain pure cardiomyocytes. The pure NRVMs were cultured in serum-free medium 199 and placed in a 5% CO₂ standard incubator at 37°C.

Calcein/PI staining

A Calcein/PI Cell Viability/Cytotoxicity assay kit (C2015M, Beyotime Biotech, China) was used to detect live and dead cells, according to the manufacturer’s instructions. After cardiomyocytes were treated with the indicated drugs, the working solution (500 μl + 0.5 μl Calcein + 0.5 μl PI) was added to the 12-well plates and incubated at 37°C for 30 min. The staining was observed and imaged under a fluorescence microscope. Calcein was fluorescent in green [excitation/emission (Ex/Em) = 494/517 nm], and PI was fluorescent in red (Ex/Em = 535/617 nm).

Determination of ROS levels

The fluorescent probe DCFH-DA from a ROS assay kit (S0033S, Beyotime Biotech, China) was used to detect intracellular ROS levels. H9C2 cells were cultured in six-well plates overnight, pretreated with the indicated agents for 2 hours, and then treated with DOX/erastin for another 12 hours. Cells were harvested, washed three times with phosphate-buffered saline (PBS), resuspended in DCFH-DA dilution (1:1000; in serum-free medium), and incubated at 37°C with shaking for 30 min. After washing with PBS, intracellular ROS levels were detected using flow cytometry (BD Biosciences, USA) and analyzed using the FlowJo software.

For fluorescent imaging of the DCFH-DA staining, a 500 μM DCFH-DA dilution (1:1000; in serum-free medium) was added to the treated cells and incubated for 30 min at 37°C. The DCFH-DA dilution was discarded after treatment, and the cells were washed thrice with serum-free medium. Fluorescence signals [fluorescein isothiocyanate (FITC), Ex/Em = 494/517 nm] were detected using a fluorescence microscope (WYS-41XDY, VIYEE, China) and analyzed using ImageJ software 6.0 (Media Cybernetics, USA).

GSH/GSSG ratio measurement

The GSH/GSSG ratios in H9C2 cells and heart tissues were determined using GSH and GSSG assay kit (S0053, Beyotime Biotech, China). Briefly, H9C2 cells and heart tissues were collected, protein removal reagent M was added according to the instructions, and the supernatant was collected for total GSH determination. The GSH scavenging auxiliary and scavenging reagent working solutions were added to some of the prepared samples to remove GSH, and the GSSG content was detected using a microplate reader at 412 nm. The GSSG content was subtracted from the total GSH content to determine the net GSH content, and the GSH/GSSG ratio was obtained.

MDA level assay

MDA levels in H9C2 cardiomyocytes and heart tissues were determined using a Lipid Peroxidation MDA assay kit (S0131S, Beyotime Biotech, China). After drug treatment, cardiomyocytes or heart tissues were lysed with Western Lysis Buffer (P0013C, Beyotime Biotech, China) to obtain supernatants. The protein concentration of the samples was determined using a portion of the supernatant and the BCA protein assay (23225, Thermo Fisher Scientific, USA); the remaining supernatant was heated, reacted with the thiobarbituric acid (a reagent from S0131S kit) solution, and pipetted into a 96-well plate to detect the optical absorbance at 532 nm. MDA levels were calibrated to the protein content of the cells or tissues.

Lipid peroxidation assay

The level of intracellular lipid peroxidation was determined using a lipid peroxidation kit (L267, DOJINDO, Japan) according to the manufacturer’s instructions. Briefly, the cells were washed twice with Hanks’ balanced salt solution (HBSS) buffer (C0219, Bain-Marie Biotech Co. Ltd., China) and then incubated with the prepared working solution (1:1000; 10% DMEM) for 30 min at 37°C. After 10 min of Hoechst nuclear staining (1:1000; 10% DMEM), the cells were washed twice with HBSS and observed under a fluorescence microscope (WYS-41XDY, VIYEE, China). The fluorescent probe emitted red fluorescence under normal conditions; however, the fluorescence changed from red to green upon the onset of lipid peroxidation. The intensity of the green fluorescence of the cells was quantified using ImageJ software 6.0 (Media Cybernetics, USA).

Apoptosis analysis

Apoptosis in H9C2 cells was assessed using a TUNEL Apoptosis Detection Kit (40307ES20, Shanghai Yeison Biotech Co., China) and the annexin V–FITC/PI Apoptosis Detection Kit (MA0220, Meilunbio, China), according to the manufacturer’s instructions. For TUNEL staining, PBS-washed cells were fixed with 4% paraformaldehyde solution in a 12-well plate and incubated with TdT incubation buffer (50-μl system) containing ddH₂O, 5× equilibration buffer, Alexa Fluor 488–12–dUTP Labeling Mix, and recombinant TdT enzyme for 1 hour at 37°C in the dark. The cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) and observed using confocal fluorescence microscopy (Carl Zeiss). For annexin V/PI detection, drug-treated H9C2 cells with supernatant were harvested, resuspended in 100 μl of binding buffer, and incubated with annexin V/PI solution (1:1000) for 15 min in the dark. The stained cells were collected using a FACSVerse flow cytometer and analyzed using FlowJo software (BD Life Sciences).

Echocardiographic analysis

The cardiac function was assessed, as previously described (14). Mice were anesthetized with 1.5% isoflurane (R510, RWD Life Science Co. Ltd., China) at 7, 14, and 28 days after DOX induction and maintained in the supine position. Transthoracic echocardiography was performed using a small animal ultrasound imaging system with an ultrasound probe (Vevo 2100, VisualSonics, Canada), and M-mode images were obtained in parasternal long-axis views of the left ventricle. The LVEF and LVFS were calculated from the measurements.

Histological analysis

At the time of sample collection, the mice were weighed and euthanized, and heart tissues were collected and fixed in 4% paraformaldehyde for ≥24 hours. The hearts were subjected to gradient dehydration and paraffin embedding. The paraffin blocks of heart tissue were sectioned into 5-μm-thick slices; the histologic changes were analyzed using H&E staining, and the degree of fibrosis was analyzed using Masson’s trichrome staining. Stained heart sections were imaged with a microscope (WYS-41XDY, VIYEE, China), digital image analysis was performed with Image-Pro Plus Software 6.0 (Media Cybernetics, CA, USA), and fibrotic tissue was quantified as the percentage of the collagen-positive area.

Western blot analysis

Protein extraction and Western blot detection were performed, as previously described (14). Briefly, total protein was extracted from drug-treated H9C2 cells and quantified using the Pierce BCA Protein Assay Kit (23225, Thermo Fisher Scientific, USA). Protein was separated using SDS–polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (IPVH00010, Merck Millipore, USA). The membranes were blocked with 5% skimmed milk for 1 hour at room temperature and then incubated with primary antibodies overnight at 4°C. The following primary antibodies were used: GPX4 (1:1000; A11243, ABclonal Technology, China), SLC7A11 (1:1000; A2413, ABclonal Technology, China), cleaved caspase-3 (1:1000; 29034, Signalway Antibody, USA), glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1:10,000; AC002, ABclonal Technology, China), H3K9me1 (1:10,000; A2358, ABclonal Technology, China), H3K9me2 (1:10,000; A2359, ABclonal Technology, China), and H3K9me3 (1:10,000; A2360, ABclonal Technology, China). The membranes were then incubated with a horseradish peroxidase–conjugated secondary antibody (1:2000; C1017, Cell Signaling Technology, USA) for 2 hours at room temperature. The immunoreactive bands were visualized using an enhanced chemiluminescence assay (34580, Thermo Fisher Scientific, USA) and imaged using a Tanon Imaging system (Tanon 5200 Multi, China). The fluorescence intensity was quantified using ImageJ software (NIH, Bethesda, MD, USA).

RNA extraction and qRT-PCR

RT-PCR was performed, as previously described. Briefly, total RNA was extracted from drug-treated H9C2 cells and left ventricular heart tissue using TRIzol reagent (15596018, Invitrogen, USA). The RNA concentration was measured using a NanoDrop 2000 (Thermo Fisher Scientific). A total of 1 μg of RNA was reverse transcribed into cDNA using cDNA Synthesis SuperMix (11141es60, Yeasen Biotech, China) according to the manufacturer’s instructions. The cDNA was diluted, mixed with diethyl pyrocarbonate water and SYBR Green Mix (11184es08, Yeasen Biotech, China), and detected using a LightCycler 480 (Roche). The mRNA expression of specific genes was normalized to Gapdh before relative quantitative analysis. The primers used in this study are listed in table S1.

Trypan blue staining

The ratio of live to dead cells after BRD4770 treatment was detected using the Trypan Blue Staining Cell Viability Assay Kit (C0011, Beyotime Biotech, China) according to the manufacturer’s instructions. Drug-treated H9C2 cells were harvested and resuspended in 100 μl of DMEM, mixed with an equal volume of Trypan Blue staining solution (2×; a reagent from the C0011 kit), and stained for 2 min. After staining, 25 μl of the stained cell suspension was pipetted onto a cell counting plate and observed under a microscope (WYS-41XDY, VIYEE, China). The total number of cells and the number of stained cells (dead) were counted, and then the survival rate of cells at each concentration was calculated.

ATP level assay

ATP is the characteristic signature molecule of metabolically active cells and exhibits a good linear relationship with the number of viable cells. Cell viability was measured by ATP levels using the CellTiter-Lumi Luminous Cell Viability Assay Kit (C0065S, Beyotime Biotech, China) according to the manufacturer’s instructions. Briefly, cells were seeded in 96-well black opaque plates. After drug treatment, CellTiter-Lumi II reagent (a reagent from the C0065S kit) was added to each well in a volume equal to the culture medium, followed by gentle mixing. After incubation for 10 min at room temperature, luminescence signals were measured using a multifunctional microplate reader (EnSight, Revvity, Britain).

Transcriptome library construction and sequencing

Total RNA was isolated from postdrug-induced H9C2 cells using TRIzol reagent according to the manufacturer’s standard protocols, and the RNA-containing TRIzol solution was sent to Novozymes Biomedical Technology Co. Ltd. (Shanghai, China) and sequenced on the NovaSeq 6000 platform. DEGs were analyzed using DEGseq, with the criteria |log2FC| ≥ 0.58 and false discovery rate ≤ 0.05. Data were analyzed using the Novozymes cloud platform (Shanghai Novozymes Biomedical Technology Co. Ltd.; http://magic.novogene.com/).

Synthesis of the DM, DMP, and DBMP nanoliposome complexes

To synthesize the DM, DMP, and DBMP nanoliposome complexes, a thin film was obtained by evaporating 10 ml of a chloroform solution containing DSPE-TK-PEG (15 mg), DSPE-PEG-NHS (15 mg), BRD4770 (0.5 mg), HSPC (20 mg), and CHOL (3 mg). The film was then thoroughly mixed with a solution containing MB (1 mg) and CeO₂ (0.5 mg) in water under ultrasonic conditions, and the solution was hydrated for 60 min at 55°C with stirring. The products were then crushed in a cell crusher in an ice bath for 30 min and further purified by ultrafiltration to obtain DBM (DSPE-PEG-NHS-DSPE-TK-PEG-BRD4770-MB) after 20 back-and-forth extrusions using a 400-nm membrane mini-extruder. DM was synthesized using the same procedure but without the addition of BRD4770 (0.5 mg) during filming. A total of 3 mg of PCM-Cys was dissolved in an aqueous solution of methanol (methanol: water = 2:1), which was added to DBM/DM and stirred for 1 day. After complete evaporation of the methanol, the final product DBMP/DMP was purified using ultrafiltration tubes.

Characterization

The nanoparticle morphologies were characterized using a Tecnai G2 TEM. The nanoliposome size and surface potential were characterized using a Malvern Zetasizer Nano ZS analyzer. Absorption and fluorescence spectra were obtained using a TU-1810 Persee spectrophotometer and an RF-6000 SHIMADZU fluorescence spectrophotometer, respectively. The release of BRD4770 was analyzed at a final concentration of 100 μM H₂O₂ to simulate an endogenous ROS environment using LC-16 SHIMADZU high-performance liquid chromatography.

Statistical analyses

Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA) and Statistical Package for the Social Sciences (IBM Corp, Armonk, NY, USA). All data are expressed as the means ± SD unless otherwise noted, and all data were checked for normality using Shapiro-Wilk to check for conformity to the normal distribution. Comparisons between two groups were assessed using the Mann-Whitney U test or the two-tailed Student’s t test. Multiple-group analyses were compared with one- or two-way analysis of variance (ANOVA) with Bonferroni post hoc tests. In all experiments, P values < 0.05 were considered statistically significant.

Study approval

All animal experiment protocols were approved by the Animal Care and Use Committee of Tianjin Medical University (animal ethics approval number: TMUaMEC-2022036).

Supplementary Materials

This PDF file includes:

Figs. S1 to S9

Table S1