Iron Promotes Cardiac Doxorubicin Retention and Toxicity Through Downregulation of the Mitochondrial Exporter ABCB8

Abstract

In several cancers, the efflux and resistance against doxorubicin (DOX), an effective anticancer drug, are associated with cellular iron deficiency and overexpression of the mitochondrial exporter ABCB8. Conversely, decreased ABCB8 expression and disrupted iron homeostasis in the heart have been implicated in DOX-associated cardiotoxicity. While studies have demonstrated that altered iron status can modulate the susceptibility to DOX cardiotoxicity, the exact molecular mechanisms have not been clearly understood. Here, we hypothesized that iron stores influence cardiac ABCB8 expression and consequently cardiac retention and toxicity of DOX. First, we found that ABCB8 deficiency in cardiomyocytes decreased DOX efflux, increased DOX-induced toxicity, and decreased cell viability. Conversely, intracellular DOX retention and toxicity were ameliorated by ABCB8 overexpression. To determine if altered cardiac iron status modifies ABCB8 expression, we treated cardiomyocytes with high iron or iron chelators. Western blot and qPCR analyses revealed that ABCB8 levels were decreased in iron overload and increased in iron deficiency. Subsequently, DOX retention and toxicity were increased in cardiomyocytes with iron overload, whereas iron deficiency ameliorated these effects. Next, we validated our results using a mouse model of hereditary hemochromatosis (HH), a genetic iron overload disorder. HH mice exhibited decreased ABCB8 expression and increased DOX retention and toxicity. These changes were abolished by the treatment of HH mice with a low-iron diet. Finally, cardiac-specific overexpression of ABCB8 in HH mice prevented cardiac DOX accumulation and abrogated DOX-induced cardiotoxicity without altering iron overload in the heart. Together, our results demonstrate that ABCB8 mediates DOX efflux and that iron regulates DOX retention and toxicity by altering cardiac ABCB8 expression. Our study identifies a novel role of iron in DOX-induced cardiotoxicity and suggests potential therapeutic intervention for DOX and anthracycline-based cancer pharmacology.

Keywords: cardiotoxicity; doxorubicin efflux; hemochromatosis; iron chelator; iron overload.

FIGURE 1

ABCB8 effluxes DOX from cardiomyocytes. (A,B) H9C2 cells were treated with ABCB8 siRNA or plasmid DNA and DOX (10 µM) as described in the Methods section. ABCB8 protein expression was determined using western blot and normalized to α-tubulin levels. Discontinuities between non-adjacent lanes of the same membrane were indicated by a solid line. (C–E) ABCB8 KD and ABCB8 OE cells were incubated with DOX (30 µM) for 3 h and DOX fluorescence was measured at the indicated times. Results are representative of four independent experiments (C). DOX fluorescence was quantified over time (D). The area under the curve (AUC) was calculated (E) from the DOX retention plot (D). Data were expressed as mean ± SEM. Statistical significance was assessed using the Student’s t-test or ANOVA with Tukey’s post-hoc comparisons. *p < .05 vs. control. Scale bar = 10 µm.

FIGURE 2

ABCB8 expression negatively influences DOX cardiotoxicity. H9C2 cells were treated with ABCB8 siRNA or plasmid DNA and DOX (10 µM) as described in the Methods section. ROS levels were analyzed by MitoSOX fluorescence (A). Cell viability was determined using the MTT assay (B). Caspase-3 and cleaved caspase levels were determined by western blot analysis (C–F). Discontinuities between non-adjacent lanes of the same membrane were indicated by a solid line. Data were expressed as mean ± SEM. Statistical significance was assessed using the Student’s t-test or ANOVA with Tukey’s post-hoc comparisons. *p < .05 vs. control. Scale bar = 10 µm.

FIGURE 3

High iron promotes cardiac DOX retention. (A,B) H9C2 cells were treated with ferric ammonium citrate (FAC) or deferoxamine (DFO) and DOX (10 µM) as indicated in the Methods section. ABCB8 protein expression was determined by western blot analysis and normalized to α-tubulin levels. Discontinuities between non-adjacent lanes of the same membrane were indicated by a solid line. (C–E) After 48 h of FAC or DFO treatment, cells were incubated with DOX (30 µM) for 3 h DOX fluorescence was measured at the indicated times. Results are representative of four independent experiments (C). DOX fluorescence was quantified over time (D). AUC was determined from the DOX retention plot (E). Data were expressed as mean ± SEM. Statistical significance was assessed using the Student’s t-test or ANOVA with Tukey’s post-hoc comparisons. *p < .05 vs. control. Scale bar = 10 µm.

FIGURE 4

Iron exacerbates DOX cardiotoxicity. H9C2 cells were treated with ferric ammonium citrate (FAC) or deferoxamine (DFO) and DOX (10 µM) as indicated in the Methods section. ROS levels were analyzed by MitoSOX fluorescence (A). Cell viability was measured using the MTT assay (B). Caspase-3 and cleaved caspase levels were determined by western blot analysis (C–F). Data were expressed as mean ± SEM. Statistical significance was assessed using the Student’s t-test or ANOVA with Tukey’s post-hoc comparisons. *p < .05 vs. control. Scale bar = 10 µm.

FIGURE 5

Cardiac DOX retention and toxicity are exacerbated in mice with iron overload hereditary hemochromatosis. Hfe-deficient (Hfe^−/−) and control wild-type (Hfe^+/+) mice were treated with DOX (20 mg/kg) as described in the Methods section (A). Cardiac non-heme iron levels were measured by colorimetric assay using bathophenanthroline disulfonate (B). ABCB8 levels in the heart were measured by western blot analysis and normalized to levels of α-tubulin (C,D). DOX levels were measured in cardiac mitochondria (E) and cardiac tissue (F) by fluorescence and normalized to protein content. Oxidative stress was determined as TBARS levels and measured by colorimetric assay (G). Serum CK-MB levels were measured by ELISA (H). Results are representative of n = 6–7/group for all panels except (C) where n = 5/group. Data were expressed as mean ± SEM. Statistical significance was assessed using two-way ANOVA for 4-group comparisons with Tukey’s post-hoc comparisons and Student’s t-test for two-group comparisons. *p < .05 vs. Hfe^+/+ of the same treatment.

FIGURE 6

Correcting iron overload prevents exacerbated DOX cardiotoxicity in HH. Hfe^+/+ and Hfe^−/− mice were maintained on iron-deficient (ID) diet and administered DOX (20 mg/kg) as described in the Methods section. Cardiac non-heme iron levels were measured by colorimetric assay using bathophenanthroline disulfonate (A). ABCB8 levels in the heart were measured by western blot analysis and normalized to levels of α-tubulin (B,C). Discontinuities between non-adjacent lanes of the same membrane were indicated by a solid line. Mitochondrial (D) and cardiac (E) DOX levels were measured in cardiac tissue by fluorescence and normalized to protein content. Oxidative stress was determined as TBARS levels and measured by colorimetric assay (F). Serum CK-MB levels were measured by ELISA (G). Results are representative of n = 7,8/group except (B) where n = 4,5/group. Data were expressed as mean ± SEM. Statistical significance was assessed using ANOVA with Tukey’s post-hoc comparisons. *p < .05 vs. Hfe^+/+ of the same treatment. ^# p < .05 vs. control diet of the same genotype.

FIGURE 7

Intracardiac ABCB8 mRNA injection prevents mitochondrial DOX retention and cardiotoxicity despite cardiac iron overload. Hfe^+/+ and Hfe^−/− mice received ABCB8 mRNA (15 µg) and DOX (20 mg/kg) as described in the Methods section. ABCB8 expression was determined by western blot analysis and normalized to α-tubulin levels (A,B). Cardiac non-heme iron levels were measured by colorimetric assay using bathophenanthroline disulfonate (C). Mitochondrial (D) and cardiac (E) DOX levels were measured in cardiac tissue by fluorescence and normalized to protein content. Oxidative stress was determined as TBARS levels and measured by colorimetric assay (F). Serum CK-MB levels were measured by ELISA (G). Results are representative of n = 6–7/group except (A) where n = 4-5/group. Data were expressed as mean ± SEM. Statistical significance was analyzed by ANOVA followed by Tukey’s post-hoc comparisons. *p < .05 vs. Hfe^+/+ of the same treatment. ^# p < .05 vs. mRNA control of the same genotype.