Aconitine induces cardiomyocyte damage by mitigating BNIP3‐dependent mitophagy and the TNFα‐NLRP3 signalling axis

Abstract Objectives Aconitine, the natural product extracted from Aconitum species, is widely used for the treatment of various diseases, including rheumatism, arthritis, bruises, fractures and pains. However, many studies have reported cardiotoxicity and neurotoxicity caused by aconitine, but the detailed mechanism underlying aconitine's effect on these processes remains unclear. Materials and methods The effects of aconitine on the inflammation, apoptosis and viability of H9c2 rat cardiomyocytes were evaluated by flow cytometry, Western blot, RNA sequencing and bioinformatics analysis. Results Aconitine suppressed cardiomyocyte proliferation and induced inflammation and apoptosis in a dose‐ and time‐dependent manner. These inflammatory damages could be reversed by a TNFα inhibitor and BNIP3‐mediated mitophagy. Consistent with the in vitro results, overexpression of BNIP3 in heart tissue partially suppressed the cardiotoxicity of aconitine by inhibiting apoptosis and the NLRP3 inflammasome. Conclusions Our findings lay a foundation for the application of a TNFα inhibitor and BNIP3 to aconitine‐induced cardiac toxicity prevention and therapy, thereby demonstrating potential for further investigation.

As an indispensable alkaloid, aconitine plays an important role in the bioactivities of analgesic, diuretic, anti-tumour, anti-asthma and anti-inflammatory agents. 7,8 However, the improper use of aconitine poses a high risk of severe poisoning. Cardiotoxicity and neurotoxicity are the main toxic effects of aconitine, and these are due to its effects on the voltage-sensitive sodium channels of the cell membranes of excitable tissues, including the myocardium, nerves and muscles. [9][10][11][12] The typical symptoms of aconitine poisoning include palpitation, vomiting, nausea, arrhythmia, shock, dizziness, hypotension and coma.

| Cell lines and culture
The rat cardiomyocyte H9c2 cell line was derived from rat embryonic heart tissue, respectively, and generally maintains their cardiomyocyte characteristics. The preparation of rat primary cardiocytes was performed according to a previously reported protocol. 19 H9c2 and rat primary cardiocytes cells were cultured in DMEM supplemented with 10% FBS, streptomycin (100 mg/mL) and penicillin (100 U/mL) under a 5% CO 2 atmosphere at 37°C. H9c2 cells were incubated for different F I G U R E 1 A, Chemical structure of aconitine; B, The viabilities of H9c2 cells and rat primary cardiocytes incubated with aconitine, as determined via the MTT method (*P < .05); C, The ROS levels of H9c2 cells and rat primary cardiocytes incubated with aconitine, as determined via the DCF-DA method (*P < .05; **P < .01); D, The viability of H9c2 cells incubated with aconitine or aconitine plus N-acetyl cysteine, as determined via the MTT method. The results are presented as the mean ± SD (n = 3) (**P < .01) times with aconitine (1.0 μmol/L), Enbrel (2.0 μmol/L) or the NLRP3 inhibitor MCC950 (2.0 μmol/L), which were dissolved in 0.1% DMSO alone or together in different combinations. The cell proliferation, apoptosis and reactive oxygen species (ROS) assays were performed according to the methods used in our previous reports, and the experimental procedures are described in the supplementary materials.

| RNA extraction, next-generation sequencing and bioinformatics analysis
Total RNA was extracted from aconitine-or normal saline-treated H9c2 cells by using QIAGEN RNA extraction kits, and all experimental procedures were performed according to the manufacturer's protocol without additional modification. To guarantee a high-quality RNA-sequencing analysis, the quantity of extracted total RNA in each sample was above 2.0 μg, and a RIN value (RNA Integrity
All animal experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals and with approval from the IACUC (Institutional Animal Care and Use Committee) and Ethics Committee of West China Hospital, Sichuan University. All animals were housed in sterilized filter-top cages in an environment of controlled humidity at 22°C with a 12 hours to 12 h day-to-night cycle.
Thirty rats were randomly divided into three groups. There were 10 rats in each group. The experimental group was orally administered with 1.0 mg/kg aconitine per day either alone or in combination with an intracardiac injection of 1 × 10 8 TU BNIP3-adenovirus. Rats in the control group received the same volume of normal saline. After 7 days of treatment, all rats were euthanized, and the heart tissues were sectioned and prepared for immunofluorescent and immunohistochemical analysis.

| Immunohistochemical, immunofluorescence and TUNEL assays
The excised hearts were washed with saline solution, fixed in 10% formalin, and embedded in paraffin. Samples were sectioned (4-6 μm thick) and stained with H&E and Masson's trichrome for microscope observation. Sections were incubated overnight at 4°C with the corresponding primary antibodies followed by incubation F I G U R E 3 A, KEGG pathway enrichment of DEGs in H9c2 cells after aconitine incubation; B, Log 2 Q-value vs enrichment factors plot of top-ranked KEGG pathways; C, GSEA results of DEGs and representative genes in each pathway. NES: net enrichment scores; FDR: false discovery rate with the secondary antibody for 1 hour in the dark. Apoptotic cells were detected with a TUNEL staining kit. All slides were observed and imaged with a fluorescence microscope.

| Aconitine inhibits myocardial cell proliferation and induces apoptosis
Aconitine's cytotoxicity ( Figure 1A) on H9c2 rat cardiomyocyte and rat primary cardiocyte cell lines was assessed by using the MTT assay. The cell viabilities declined significantly after 24 hours of aconitine incubation in a dosage-dependent manner, and H9c2 cell was more sensitive to aconitine incubation than that of rat primary cardiocytes ( Figure 1B). ROS accumulation was also detected by DCF-DA probes. As expected, the incubation of 1.0 μmol/L aconitine for more than 24 hours resulted in significant ROS accumulation with extended exposure time than the corresponding normal saline groups ( Figure 1C). To further identify the relationship between ROS accumulation and aconitine's cytotoxicity, N-acetyl cysteine (NAc) was added to aconitine-treated H9c2 cells; aconitine's cytotoxicity was almost completely diminished by NAc ( Figure 1D). [20][21][22] The apoptotic cell death of H9c2 cells induced by aconitine incubation was determined by Hoechst 33258 staining and Annexin

| Both apoptosis and NLRP3 pathways were involved in aconitine's cytotoxicity
To further study the role of apoptosis in aconitine-induced cardiomyocyte damage, a pan-caspase inhibitor, Z-VAD, was added to the aconitine-treated H9c2 cells and rat primary cardiomyocytes. The cell viability assay results demonstrated that aconitine's cytotoxicity was only partially reversed by Z-VAD ( Figure 4A and Figure   S2A). There might be other cell death subroutines involved in aconitine's cytotoxicity ( Figure 4B and Figure S2B). The detailed relationship and regulatory mechanisms between RIPKs and NLRP3 inflammatory pathways remain unclear. 36 Some recent studies suggested that caspase-8 activation could result in the formation of the F I G U R E 5 A, Western blot analysis of NLRP3 pathway activation after aconitine incubation for 24 h with or without the NLRP3 inhibitor MCC950; B, The subcellular location of NLRP3 and ASC used to detect NLRP3 inflammasome formation after aconitine incubation for 24 h with or without MCC950, scale bar: 6 μm NLRP3-inflammasome and/or secretion of pro-inflammatory interleukins via phosphorylated RIPKs or other unknown mechanisms. 37 According to the Western blot analysis results, the RIPK pathway was stimulated by aconitine after 48 hours or more of incubation, as indicated by the increased expression of RIP1, RIP3 and MLKL ( Figure 4C). The activation of the NLRP3 inflammasome was observed within 24 hours of aconitine incubation and was indicated by the increased expression of NLRP3 and ASC and the delayed activation of the pro-inflammation indicator caspase-1 and IL-1β cleavage ( Figure 4D). As expected, the addition of pan-caspase inhibitor Z-VAD could only suppress the cleavage of Caspase-1 and did not affect the expressions of NLRP3 and ASC. Therefore, the NLRP3 inflammasome pathway (not caspase activation) was the main mechanism underlying the toxicity induced by aconitine.
To further study the role of NLRP3 inflammasomes in aconitineinduced cardiomyocyte damage, the NLRP3 inhibitor MCC950 was added to the aconitine-treated H9c2 cells. [38][39][40] The cleavage of pro-Caspase-1 and pro-IL-1β and expression of NLRP3 after aconitine incubation were severely suppressed by MCC950 rather than the transcript activation of Caspase-1 and IL-1β ( Figure 5A). WB analysis of ASC, which is an NLRP3 partner protein required to assemble the NLRP3 inflammasome, failed several times after repeatedly changing the primary antibodies from different companies. The addition of aconitine potently stimulated the expression and assembly of NLRP3 inflammasome and Caspase-1 activation. Moreover, the activation of NLRP3 inflammasome pathway could be reversed by NLRP3 inhibition of MCC950 in a dose-dependent manner. Therefore, we performed an immunofluorescent assay to simultaneously visualize the protein expression changes of Caspase-1 and ASC. As shown in

| A TNFα inhibitor ameliorates aconitine-induced NLRP3 inflammasomes by activating mitophagy
To determine the effect and regulatory mechanism of TNFα in aconitine-induced inflammation and mitophagy of H9c2 cardiomyocytes, cells were treated with aconitine at 2 μmol/L or with aconitine combined with the TNFα inhibitor Enbrel or BNIP3 shRNA for 24 hours.
F I G U R E 6 A, Western blot analysis of NLRP3 pathway activation after aconitine incubation for 24 h with or without the TNFα inhibitor Enbrel or BNIP3 shRNA. B, Immunofluorescence analysis of BNIP3 and TOM20 to detect mitophagy after aconitine incubation for 24 h with or without the TNFα inhibitor Enbrel, scale bar: 10 μm We considered that BNIP3 was a LC3-interacted protein that played an important role in mitophagy, and the expression of BNIP3 was significantly declined after aconitine treatment. We speculated that BNIP3-mediated mitophagy mighty involved in aconitine-induced TNFα activation and inflammatory response in H9c2 cells. In Western blot analysis, the expression or cleavage of Caspase-1 or IL-1β was not affected by BNIP3 RNA interference alone ( Figure 6A). Moreover, BNIP3 knockdown did not interfere with the aconitine-induced activation of NLRP3, caspase-1 or IL-1β. The addition of TNF-α inhibitor Enbrel could partially reverse the decreased expression of BNIP3 by aconitine or siBNIP3. Conversely, the addition of Enbrel significantly suppressed the Caspase-1 expression and IL-1β and Caspase-1 cleavage. Furthermore, the inhibitory effects of Enbrel on aconitine-induced Caspase-1 and IL-1β activation could be reversed by BNIP3 knockdown. Aconitine-induced inflammation might be mediated by TNFα and BNIP3-dependent mitophagy. [41][42][43][44][45] To further identify the regulatory mechanism of TNFα and BNIP3 in aconitine-induced inflammatory damage, an immunofluorescence assay was performed to visualize the expressions and subcellular locations of BNIP3 and TOM20, a mitochondrial membrane protein ( Figure 6B). Aconitine did not change the expression of TOM20 but decreased the expression of BNIP3. The combination of Enbrel and aconitine strongly activated the expression of BNIP3 and promoted the colocalization of BNIP3 and TOM20. The aconitine-induced cardiomyocyte damage might induce TNFα activation and then suppress BNIP3-dependent mitophagy.

| BNIP3-dependent mitophagy alleviates the mitochondrial damage induced by aconitine
The aconitine-induced mitochondrial damages on H9c2 cells were detected by transmission electronic microscope (TEM) analysis and shown in Figure 7A. Compared with the control group, there were markedly swollen and damaged mitochondria observed in the cytoplasm. To better recognize the influences of BNIP3 and TNFα on F I G U R E 7 A, Morphological changes of mitochondria under TEM analysis, scale bar: 1 μm. B, Immunofluorescence of LC3 and TOM20 to detect mitophagy after aconitine incubation for 24 h with or without the TNFα inhibitor Enbrel under transfection of null vector or BNIP3overexpressed plasmid, scale bar: 10 μm. C, Western blot analysis of the autophagy-related proteins LC3 and p62 after aconitine incubation for 24 h with or without pretransfected BNIP3-overexpressed plasmid F I G U R E 8 A, Masson's trichrome staining of myocardial tissue sections of control, aconitine treated and aconitine plus BNIP3-OE rats, scale bar: 1 mm. B, immunofluorescent staining of TUNEL and TNFα in myocardial tissue sections of control, aconitine treated and aconitine plus BNIP3-OE rats, scale bar: 1 mm. C, The immunohistochemistry analysis of NLRP3, LC3, CTnI and CaMKII in myocardial tissue sections of control, aconitine treated and aconitine plus BNIP3-OE-treated rats, scale bar: 50 μm. *P < .05 vs control. **P < .01 vs control. ##P < .01 to aconitine-treated group protective mitophagy in H9c2 cells with aconitine incubation, [46][47][48][49] TOM20 and LC3 immunofluorescence assays were performed in H9c2 cells transfected with null vector or BNIP3 plasmid, with or without a combination of aconitine or aconitine plus Enbrel treatment, respectively. [50][51][52][53] As expected, BNIP3 overexpression and the TNFα inhibitor significantly stimulated LC3 expression, accumulation and colocalization to mitochondria in cells treated with aconitine ( Figure 7B).
According to the WB results shown in Figure 7C, overexpression of BNIP3 activated LC3 cleavage and ULK1 phosphorylation, which suggested that autophagy was activated. Moreover, the mitochondria and cytoplasm of H9c2 cells were separated, and LC3 and p62 protein expressions were determined by WB analysis. LC3 cleavage and p62 degradation, either inhibited by aconitine or activated by BNIP3 overexpression, occurred in the mitochondrion but not in the cytoplasm. Aconitine-induced cardiomyocyte mitochondrial damage depended on the activation TNFα and inhibition of BNIP3-mediated mitophagy.

| Intracardiac injection of BNIP3-overexpression adenovirus decreased aconitine-induced cardiotoxicity
IHC staining of BNIP3 expression in the heart tissues of rats in the null vector group and the BNIP3-OE adenovirus group displayed the significant overexpression of BNIP3 in the BNIP3-OE adenovirus group ( Figure S3). Masson's trichrome staining was performed to evaluate the degree of cardiac inflammatory fibrosis.
Aconitine-treated rats developed higher cardiac fibrosis, which was significantly prevented by the addition of BNIP3-overexpression adenovirus ( Figure 8A). BNIP3-OE reversed aconitine-induced activation of TNFα, NLRP3, LC3 and cardiac apoptosis as assessed by TUNEL staining (Figure 8B,C). The IHC staining of myocardial damage marker CTnI (Cardiac Troponin I) and CaMKII (Ca2+/calmodulin-dependent protein kinase II) indicated that the overexpression of BNIP3 obviously suppressed aconitine-induced myocardial damage.

| D ISCUSS I ON
Aconitum species are widely used as TCMs for the treatment of various diseases. Thus, the cardiotoxicity of aconitine and its derivatives and/or metabolites have potential risks, but details of the underlying cardiotoxicity mechanism have not been elucidated. Few reports in- In conclusion, we have investigated the molecular mechanisms and prevention/therapeutic strategy of the cardiac toxic natural product aconitine. The BNIP3-mediated mitophagy potently alleviates myocardial injuries of aconitine both in vitro and in vivo. The aconitine cytotoxicity in cardiomyocytes is dependent on the activation of the TNFα and NLRP3 inflammasome pathways, which may provide novel insights into the prevention of aconitine-related toxicity.

ACK N OWLED G EM ENT
We are grateful for financial support from the National Natural

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.