Loss of TRIM21 alleviates cardiotoxicity by suppressing ferroptosis induced by the chemotherapeutic agent doxorubicin

Abstract

Background: Doxorubicin, an anthracycline chemotherapeutic agent, is widely used in the treatment of many cancers. However, doxorubicin posts a great risk of adverse cardiovascular events, which are thought to be caused by oxidative stress. We recently reported that the ubiquitin E3 ligase TRIM21 interacts and ubiquitylates p62 and negatively regulates the p62-Keap1-Nrf2 antioxidant pathway. Therefore, we sought to determine the role TRIM21 in cardiotoxicity induced by oxidative damage.

Methods: Using TRIM21 knockout mice, we examined the effects of TRIM21 on cardiotoxicity induced by two oxidative damage models: the doxorubicin treatment model and the Left Anterior Descending (LAD) model. We also explored the underlying mechanism by RNA-sequencing of the heart tissues, and by treating the mouse embryonic fibroblasts (MEFs), immortalized rat cardiomyocyte line H9c2, and immortalized human cardiomyocyte line AC16 with doxorubicin.

Findings: TRIM21 knockout mice are protected from heart failure and fatality in both the doxorubicin and LAD models. Hearts of doxorubicin-treated wild-type mice exhibit deformed mitochondria and elevated level of lipid peroxidation reminiscent of ferroptosis, which is alleviated in TRIM21 knockout hearts. Mechanistically, TRIM21-deficient heart tissues and cultured MEFs and H9c2 cells display enhanced p62 sequestration of Keap1 and are protected from doxorubicin-induced ferroptosis. Reconstitution of wild-type but not the E3 ligase-dead and the p62 binding-deficient TRIM21 mutants impedes the protection from doxorubicin-induced cell death.

Interpretation: Our study demonstrates that TRIM21 ablation protects doxorubicin-induced cardiotoxicity and illustrates a new function of TRIM21 in ferroptosis, and suggests TRIM21 as a therapeutic target for reducing chemotherapy-related cardiotoxicity.

Funding: NIH (CA129536; DK108989): data collection, analysis. Shanghai Pujiang Program (19PJ1401900): data collection. National Natural Science Foundation (31971161): data collection. Department of Veteran Affairs (BX004083): data collection. Tianjin Science and Technology Plan Project (17ZXMFSY00020): data collection.

Keywords: Antioxidant; Cardiotoxicity; Doxorubicin; Ferroptosis; TRIM21.

Fig. 1

Loss of TRIM21 has a protective effect on doxorubicin-induced cardiotoxicity in mice.TRIM21^+/+ or TRIM21^−/− male mice (10 to 12 week-old) were treated with doxorubicin (20 mg/kg, i.p.). (a) Kaplan–Meier survival curves were generated. Log rank test was used to compare the difference in survival between TRIM21+/+ (n=16) and TRIM21-/- (n=14), showing significant difference with p-value < 0.0001. (b) Mice were injected with saline or doxorubicin (20 mg/kg, i.p.). 5 days later, mice were photographed. Representative pictures are shown. Note doxorubicin-treated WT mice displayed smaller body weight and hair loss. (c) 5 days after doxorubicin injection, mice were subjected to echocardiography. (d) M-mode images from nine cardiac cycles for each animal were used to calculate ventricular measurements: ejection fraction, fractional shortening, heart rate, and LVIDd (left ventricular internal diameter at end-diastole). Individual data points and the mean ± standard deviation are shown. Student's t-test was used to compare the results. n.s., non-significant; **p<0.01; ***p<0.001; ****p<0.0001. (e) Abdominal photographs of mice after 5-days after doxorubicin injection. Abdominal ascites are indicated by red arrows. (f) Hearts from each group were photographed. Representative photos are shown. (g) Mice were weighted at the time of doxorubicin injection and 5 days after, relative body weight (normalized to body weight at the injection) and the heart/body weight ratio was measured. Each individual data point represents a mouse. Individual data points and the mean ± standard deviation are shown. Student's t-test was used to compare the results. n.s., non-significant; **p<0.01.

Fig. 2

TRIM21 ablation alleviates doxorubicin-induced cardiotoxicity. TRIM21^+/+ or TRIM21^−/− mice (n=5 in each group) were injected with saline or doxorubicin (20 mg/kg via i.p.). 5 days later, hearts were isolated. (a) Representative H&E staining images are shown. (b) Hearts were collected and examined with 4-HNE IHC staining. The intensity was quantified by Image J. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; **p<0.01. (c) The level of MDA in the heart tissue was determined and the mean ± SD of triplicate experiments is shown. Student's t-test was used to compare the results. n.s., non-significant; ***p<0.001. (d) Transmission electron microscopy of heart tissues. Note deformed/damaged mitochondria in doxorubicin-treated TRIM21^+/+ hearts. The extent of mitochondrial damage was quantified using the Flameng score. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; ****p<0.0001.

Fig. 3

TRIM21 negatively regulates p62-Keap1 aggregation and anti-ferroptotic effect in heart tissue.TRIM21^+/+ or TRIM21^−/− mice (n=5 in each group) were i.p. injected with saline or doxorubicin (20 mg/kg). Heart tissues were harvested 5 days later. (a) Heart tissues were lysed in RIPA buffer containing 1% SDS and probed for indicated proteins. (b) IHC staining was performed for indicated antibodies. IHC intensity was determined and quantified by Image J. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; ***p<0.001, ****p<0.0001. (c) Heart tissues were lysed and separated into detergent (1% Triton X-100)-soluble ‘‘S’’ and insoluble ‘‘I’’ fractions and subjected to IB. The ratio of insoluble/soluble fractions was normalized to the corresponding tubulin band and is shown on the bottom. Densitometry of immunoblot bands was determined by the Odyssey Infrared Imaging System. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; ***p<0.001, ****p<0.0001.

Fig. 4

Doxorubicin-induced cardiomyopathy and ferroptosis are suppressed in TRIM21^−/− hearts. (a) Relative TRIM21 expression in human heart tissue samples under indicated conditions. TRIM21 expression levels were estimated by normalized probe intensity obtained from previously published microarray datasets, with Gene Expression Omnibus (GEO) accession number shown in each sub-figure. P values less than 0.2 are annotated. (b) Correlation between TRIM21 and indicated genes in indicated datasets. (c) Relative TRIM21 expression in mouse heart tissue samples upon doxorubicin (Doxo) treatment. TRIM21 expression levels were estimated by normalized probe intensity obtained from previously published microarray datasets, with Gene Expression Omnibus (GEO) accession number shown. P values less than 0.2 are annotated. (d) Ten to twelve week-old TRIM21^+/+ and TRIM21^−/− male mice were treated with PBS or 20 mg/kg Doxo by intraperitoneal injection. Heart tissues were collected after 5 days and used for RNA-seq. Hierarchical clustering and heatmap illustration of differentially expressed genes under indicated conditions. Color key donates Z-scores after normalization. (e) Venn diagram showing overlap of genes differentially expressed in Doxo versus PBS-treated TRIM21^+/+ and TRIM21^−/− mice. Analysis was restricted to genes with a 2-fold difference. (f) Dot plots showing the enriched terms from Gene Ontology analysis using differentially expressed genes identified from Doxo versus PBS treatment in TRIM21 wild-type and knockout mice. Note that pathways involved in redox regulation and metal ion homeostasis are identified. (g) Total mRNA was extracted and analyzed by qRT-PCR (n=5 in each group). Relative mRNA levels of indicated genes are shown as mean ± standard deviation. Note that A549 human lung carcinoma cell line was used as a control to demonstrate that SLC7A11 is expressed at a very low level in mouse heart tissues. Student's t-test was used to compare the results. n.s., non-significant; **p<0.01.

Fig. 5

TRIM21-deficient MEFs are protected from doxorubicin-induced ferroptosis. (a) TRIM21^+/+ or TRIM21^−/− MEFs were treated with doxorubicin (0.1 μM) alone or together with ferrostatin-1 (Fer-1, 2 μM) for the indicated times. Cells were observed under phase-contrast microscope, and cell death determined by trypan blue staining. Data shown are the mean plus SD of three random countings (500 cells each). Student's t-test was used to compare the results at indicated times. ***p<0.001. (b) Cells were treated with doxorubicin (0.1 μM) in the presence of ferrostatin-1 (Fer-1, 2 μM), dexrazoxane (DXZ, 100 μM), zVAD (10 μM), or necrostatin-1 (Nec, 0.5 μM) for 36 h. Cell death was determined by trypan blue staining. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; ****p<0.0001. (c) Cells treated with doxorubicin (1 μM) for indicated times, and probed for caspase-3. (d) TRIM21^−/− MEFs were reconstituted with vector, HA-TRIM21WT, HA-TRIM21C16A or HA-TRIM21W381/383A mutants. (e)TRIM21^−/− MEFs reconstituted with WT TRIM21 or the mutants were treated with doxorubicin (1 μM) for 16 h. Immunofluorescence of indicated proteins was performed. Note that doxorubicin treatment induced Keap1 and p62 aggregation in TRIM21^−/− MEFs, which was abrogated by the reconstitution of TRIM21 WT but not the ligase-dead C16A mutant or the p62 binding-deficient W381/383A mutant. (f) Quantification of (e). Data shown are mean ± standard deviation. Student's t-test was used to compare the results. n.s., non-significant; ****p<0.0001. (g) Cells were lysed in IP lysis buffer containing 1% Triton X-100. The insoluble fraction was dissolved in RIPA buffer containing 1% SDS. Both the Triton X-100 soluble and insoluble fractions were subjected to WB. α-tubulin was probed as a marker of the soluble fraction. The WB bands were quantified by ImageJ, and the ratio of insoluble to soluble p62 and Keap1, respectively, is shown at the bottom.

Fig. 6

TRIM21 negatively regulates p62-Keap1 aggregation and anti-ferroptotic effect in cardiomyocytes. (a) H9c2 immortalized rat cardiomyocytes were infected with lentiviral none-target control (shNTC) or two independent shRNAs for TRIM21. Relative mRNA level of TRIM21 was measured by quantitative RT-PCR. Data shown are mean ± standard deviation. Student's t-test was used to compare the results. ****p<0.0001. (b) Cells were treated with doxorubicin (1 μM) for 48 hours and observed under phase-contrast microscope. (c) Cell death was determined by trypan blue staining. Data shown are the mean plus SD of three countings (500 cells each). Student's t-test was used to compare the results. n.s., non-significant; ****p<0.0001. (d) H9c2 cells infected with lentiviral shNTC or shTRIM21 were treated with or without doxorubicin (Doxo, 1μM) alone or together with ferrostatin-1 (Fer-1,10 μM), DXZ (100 μM), zVAD (10 μM), or necrostatin-1 (Nec, 0.5 μM). Cell death was determined by trypan blue staining after 36 h. Shown is the average plus SD of triplicate experiments. Student's t-test was used to compare the results. n.s., non-significant; ****p<0.0001. (e) H9c2 cells infected with lentiviral shNTC or shTRIM21 were treated with doxorubicin (1 μM) for 4 h. Cells were lysed in IP lysis buffer containing 1% Triton X-100. The insoluble fraction was dissolved in RIPA buffer containing 1% SDS. Both the Triton X-100 soluble and insoluble fractions were subjected to IB. The ratio of insoluble to soluble bands was determined by ImageJ. (f) HA-TRIM21WT, HA-TRIM21C16A or HA-TRIM21W381/383A mutants was overexpressed in H9c2 cells. (g) H9c2 cells reconstituted with vector, HA-TRIM21WT, HA-TRIM21C16A or HA-TRIM21W381/383A mutants were treated with doxorubicin (1 μM) for indicated times. Cells were lysed in IP lysis buffer containing 1% Triton X-100. The insoluble fraction was dissolved in RIPA buffer containing 1% SDS. Both Triton X-100 soluble and insoluble fractions were subjected to IB. The ratio of insoluble to soluble bands was determined by ImageJ.