Mitigation of Doxorubicin Cardiotoxicity With Synergistic miRNA Combinations Identified Using Combinatorial Genetics en masse (CombiGEM)

Abstract

Background: Cardiomyocyte loss occurs in acute and chronic cardiac injury, including cardiotoxicity due to chemotherapeutics like doxorubicin, and contributes to heart failure development. There is a pressing need for cardiac-specific therapeutics that target cardiomyocyte loss, preventing chemotherapy complications without compromising chemotherapeutic efficacy.

Objectives: The authors employed massively parallel combinatorial genetic screening to find microRNA (miRNA) combinations that promote cardiomyocyte survival.

Methods: CombiGEM (combinatorial genetics en masse) screening in a cardiomyocyte cell line was followed by validation in the original cell type and screening in primary cardiomyocytes. The top combination was tested in mouse and developing zebrafish models of doxorubicin cardiotoxicity. RNA sequencing provided insight into possible mechanisms.

Results: Multiple miRNA combinations protected cardiomyocytes from doxorubicin in vitro. The most effective (miR-222+miR-455) appeared to act synergistically, and mitigated doxorubicin cardiotoxicity phenotypes in murine and zebrafish in vivo models. RNA sequencing revealed overlapping and synergistic regulation of relevant genes and biological processes in cardiomyocytes, including mitochondrial homeostasis, oxidative stress, muscle contraction, and others.

Conclusions: We identified miR-222 and miR-455 as a combination with potential therapeutic applications for cardioprotection. This study furthers our knowledge of the cardiac effects of miRNAs and their combinations and demonstrates the potential of CombiGEM for cardioprotective combinatorial therapeutic discovery.

Keywords: cardiac myocytes; cardiomyopathy; chemotherapy toxicity; cytoprotection; drug therapy; heart failure; innovation; preclinical study; systems biology.

Graphical abstract

Central Illustration

Combinatorial Screen Leads to a Synergistic Combination That Mitigates Doxorubicin Cardiotoxicity Multiple features make microRNAs (miRNAs) attractive candidates for cardiac-targeted combinatorial therapy. We used combinatorial genetics en masse (CombiGEM) screening in an HL-1 cardiomyocyte cell line to identify broadly protective miRNA combinations. Multiple combinations from the screen protected primary cardiomyocytes against the cardiotoxic chemotherapeutic doxorubicin, with miR-222 and miR-455 showing potent, synergistic protection. This combination also mitigated doxorubicin cardiotoxicity in zebrafish and mouse in vivo models. RNA sequencing suggests overlapping and synergistic regulation of relevant genes and pathways. gDNA = genomic DNA; NGS = next-generation sequencing; PCR = polymerase chain reaction; RNA-seq = RNA sequencing.

Figure 1

Primary Screen for Protective miRNAs in HL-1 Cells Transduced With a Human-Mouse Conserved miRNA Lentiviral Library Volcano plot showing change in abundance of integrated microRNA (miRNA) sequences in the surviving population following (A) adrenergic, (B) hypoxic, and (C) oxidative stress relative to unstressed control samples. The library included controls for assay optimization and quality control, not displayed. Light and dark gray horizontal lines indicate P = 0.05 and P = 0.10, respectively. Green indicates hits. Blue indicates that it met P value but not fold-change criteria. Log₂FC = log₂ fold change.

Figure 2

CombiGEM Screen for Protective MiRNA Combinations in HL-1 Cells Volcano plot showing change in abundance of integrated (A to C) 2-wise and (E to G) 3-wise barcodes in the surviving population following adrenergic, hypoxic, and oxidative stress relative to unstressed control samples. Each data point represents a different combinatorial construct. Hits are labeled with code as annotated in D and H. Gray lines indicate P = 0.05 and Log₂FC = 1, respectively. Blue indicates that it met P value but not fold-change criteria for the condition. Red indicates that it met P value and fold-change criteria for the condition but did not meet criteria for hits based on performance across all stressors (see Supplemental Methods). Green indicates hits based on performance across all conditions. CombiGEM = combinatorial genetics en masse; other abbreviations as in Figure 1.

Figure 3

Several miRNA Combinations Protect Primary Cardiomyocytes Against Doxorubicin (A, B) Relative viability of neonatal rat ventricular cardiomyocytes transduced with lentiviruses expressing the indicated microRNA (miRNA) combinations and treated with (A) vehicle or (B) 2 μM doxorubicin (Dox). In A, data are normalized to uninfected. In B, viability is relative to matching unstressed samples infected with the same virus. ∗∗P < 0.01 and ∗∗∗P < 0.001 vs uninfected, no-miRNA control (YH270) and negative control short hairpin RNA (YH290) combined, by 1-way analysis of variance and Dunnett’s test. Error bars represent SEM. n = 3 replicates/group.

Figure 4

Protection Against Doxorubicin In Vitro and In Vivo by miR-222 and miR-455 Mimics Combined (A) Relative viability of neonatal rat ventricular cardiomyocytes transfected with murine miR-222-3p or miR-455-3p mimics, their combination, or negative control mimic (5 nM each), treated with the indicated doxorubicin concentrations for 4 days. For unstressed samples, data are normalized to the negative control sample. For doxorubicin, viability is relative to unstressed samples treated with the same microRNA(s) (miRNA[s]). Analyzed by 2-way analysis of variance and Bonferroni test. n = 3 replicates/group. (B) The percentage increase in apoptotic (Annexin V positive) neonatal rat ventricular cardiomyocytes measured by flow cytometry after the indicated miRNA and doxorubicin treatment, relative to unstressed cells with negative control miRNA mimic. Analyzed by unpaired 2-tailed t tests with Benjamini-Hochberg correction to control the false discovery rate. n = 3 replicates/group. (C) Fractional shortening quantified from video microscopy and (D) cardiotoxicity severity based on observable phenotypes (see Methods) in zebrafish embryos microinjected with negative control mimic or miR-222-3p and miR-455-3p, treated with 100 μM doxorubicin or dimethyl sulfoxide (DMSO). Analyzed by 1-way analysis of variance and (C) Bonferroni post hoc test and (D) Mann-Whitney U test: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Error bars represent SEM. Created in BioRender. Neg. Ctl. = negative control.

Figure 5

Protection in a Mammalian Dox Cardiotoxicity In Vivo Model (A) Experimental design. (B) Representative echocardiographic images and (C) quantification of ejection fraction, (D) left ventricular end-diastolic diameter (LVEDd), (E) interventricular septum thickness (IVSth), and (F) left ventricular posterior wall thickness (LVPWth) in mice infected with adeno-associated virus 9 (AAV9) encoding miR-222, miR-455, both, or neither, treated weekly with doxorubicin (Dox; 5 mg/kg) or vehicle for 4 weeks. (G) Representative images and quantification of cardiomyocyte cell surface area (CSA) from hematoxylin and eosin–stained heart sections and (H) fibrosis from Sirius red–stained heart sections. n = 5 for vehicle groups; n = 8 for control vector with Dox; n = 10 for miR-222 with Dox, n = 7 for miR-455 with Dox, n = 8 for miR-222+miR-455 with Dox. (I) Cardiac mitochondrial DNA oxidative damage quantified by 8-hydroxy-2′-deoxyguanosine (8-OHdG) enzyme-linked immunosorbent assay. n = 5/group. (J) Quantitative polymerase chain reaction measurement of cardiac mitochondrial cytochrome b (mtCytb) gene expression. n = 5/group. ∗P < 0.05 vs vehicle; †P < 0.05 vs control vector with the same treatment; ‡P < 0.05 vs miR-222 with the same treatment; §P < 0.05 vs miR-455 with the same treatment by (C to F, H to J) 1-way analysis of variance or (G) Kruskal-Wallis test and Holm post hoc tests. Error bars represent SEM. Scale bars = 20 μm (G) and 50 μm (H). Created using BioRender. CSA = cross-sectional area; i.p. = intraperitoneal; mtDNA = mitochondrial DNA; TTE = transthoracic echocardiography.

Figure 6

Synergy and Overlap in Genes and Pathways Regulated by miR-222 and miR-455 Under Doxorubicin (A to C) Differential gene expression after 24-hour doxorubicin treatment in neonatal rat ventricular cardiomyocytes transfected with miR-222 and miR-455 (A) combined or (B, C) individually, vs negative control mimic. Red indicates upregulated. Blue indicates downregulated. Gray/black indicates not differentially expressed. Selected genes implicated in heart disease or doxorubicin toxicity are labeled. (D) Overlap among genes regulated by miR-222 and miR-455 and their combination under doxorubicin. (E, F) Top 10 Gene Ontology terms by fold enrichment from analyses of genes up- or down-regulated by (E) the combination but neither miRNA alone and (F) the combination and both microRNAs (miRNAs) individually. ATP = adenosine triphosphate; ERAD = endoplasmic reticulum–associated protein degradation.