Inhibition of the long non‐coding RNA ZFAS1 attenuates ferroptosis by sponging miR‐150‐5p and activates CCND2 against diabetic cardiomyopathy

Abstract Diabetic cardiomyopathy (DbCM) is responsible for increased morbidity and mortality in patients with diabetes and heart failure. However, the pathogenesis of DbCM has not yet been identified. Here, we investigated the important role of lncRNA‐ZFAS1 in the pathological process of DbCM, which is associated with ferroptosis. Microarray data analysis of DbCM in patients or mouse models from GEO revealed the significance of ZFAS1 and the significant downregulation of miR‐150‐5p and CCND2. Briefly, DbCM was established in high glucose (HG)–treated cardiomyocytes and db/db mice to form in vitro and in vivo models. Ad‐ZFAS1, Ad‐sh‐ZFAS1, mimic miR‐150‐5p, Ad‐CCND2 and Ad‐sh‐CCND2 were intracoronarily administered to the mouse model or transfected into HG‐treated cardiomyocytes to determine whether ZFAS1 regulates miR‐150‐5p and CCND2 in ferroptosis. The effect of ZFAS1 on the left ventricular myocardial tissues of db/db mice and HG‐treated cardiomyocytes, ferroptosis and apoptosis was determined by Masson staining, immunohistochemical staining, Western blotting, monobromobimane staining, immunofluorescence staining and JC‐1 staining. The relationships among ZFAS1, miR‐150‐5p and CCND2 were evaluated using dual‐luciferase reporter assays and RNA pull‐down assays. Inhibition of ZFAS1 led to reduced collagen deposition, decreased cardiomyocyte apoptosis and ferroptosis, and attenuated DbCM progression. ZFAS1 sponges miR‐150‐5p to downregulate CCND2 expression. Ad‐sh‐ZFAS1, miR‐150‐5p mimic, and Ad‐CCND2 transfection attenuated ferroptosis and DbCM development both in vitro and in vivo. However, transfection with Ad‐ZFAS1 could reverse the positive effects of miR‐150‐5p mimic and Ad‐CCND2 in vitro and in vivo. lncRNA‐ZFAS1 acted as a ceRNA to sponge miR‐150‐5p and downregulate CCND2 to promote cardiomyocyte ferroptosis and DbCM development. Thus, ZFAS1 inhibition could be a promising therapeutic target for the treatment and prevention of DbCM.


| INTRODUC TI ON
Patients with diabetes are vulnerable to a series of cardiovascular complications, one of the most serious complications associated with heart failure (HF). 1,2 As the incidence of diabetes (expected to reach 693 million by 2045) increases, 3 HF due to diabetes has become a worldwide epidemic. 4,5 In the clinical setting, diabetes is present in one third of all HF patients. Further, diabetes has been identified as an independent predictor of adverse outcomes. 6 According to a report published five decades ago by the Framingham Heart Study, a fourfold to fivefold increase in the risk of HF has been identified among patients with diabetes, 2,7,8 where diabetic cardiomyopathy (DbCM) is recognized as a proximate cause. Despite extensive research on DbCM, 9 the full spectrum of its pathogenesis and its relative contribution to the HF phenotype in diabetes has not been revealed.
lncRNAs have been shown to act as competing endogenous RNAs (ceRNAs) to sponge macromolecules, such as micro-RNAs (miRNAs) and proteins, 18 which are associated with a wide range of physiological, biological and pathological processes, including those in the case of cardiac diseases. 19,20 A previous clinical study revealed that miR-150-5p levels were significantly reduced in patients with HF. Thus, miR-150-5p was identified as an independent predictor of HF. 21 The miR-150-5p gene was found to mitigate apoptosis in sepsis-induced myocardial depression, 22 alleviate the progression of myocardial fibrosis 23 and rescue cardiomyocytes from hypoxia-induced injury under the command of the lncRNA, FOXD3-AS1. 24 However, the role of miR-150-5p in DbCM and its relationship with lncRNA-ZFAS1 have not been studied. Cyclin D2 (CCND2) regulates the proliferation of cardiac myocytes, 25 which is beneficial for cardiac dysfunction, 26 and activates cell cycle progression to enhance myocardial repair. 27 The role of CCND2 in DbCM has also not yet been studied.
Ferroptosis, an iron-dependent regulated necrosis associated with a new form of regulatory cell death, was first described in 2012. 28 Ferroptosis can induce the pathological processes of cancer, stroke, cardiovascular disease and kidney failure. 29,30 Glutathione peroxidase 4 (GPX4) can terminate ferroptosis, which occurs in caspase and necrosomal complexes. 31 A recent report demonstrated that inhibiting ferroptosis could decrease mitochondrial iron to alleviate DOX-induced cardiac injury 32 ; however, its role in DbCM has not been explored.
In this study, HG-treated cardiomyocytes and db/db mice were used to simulate DbCM in vitro and in vivo, and lncRNA-ZFAS1 was assessed to demonstrate that ZFAS1 inhibition alleviates the development of DbCM by reducing ferroptosis by stabilizing miR-150-5p to activate CCND2.

| Ethics and animal experiments
Animal experiments were performed in accordance with the Institutes of the First Affiliated Hospital of Wenzhou Medical University Health Guidelines on the Use of Laboratory Animals to ensure the humanitarian care of the experimental animals.
To simulate the animal model of diabetic cardiomyopathy, male db/+ mice and db/db mice (age, 7 weeks, weight, 24 g) were fed a normal diet for 4 weeks and kept at 24℃ under a 14-h light/8-h dark cycle. The animals were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China). Diabetic mice were intracoronarily administered equal volumes (80 μl) of adenoviruses Ad-ZFAS1, Ad-sh-ZFAS1, Ad-CCND2, Ad-sh-CCND2 or Ad-NC. 33 miR-150-5p mimics and mimic control (NC) were injected into the tail vein of mice (50 μg/kg) every 15 days for 12 weeks. Db/db mice were treated with or without ferrostatin-1 (Fer-1, ferroptosis inhibitor; Sigma-Aldrich, 5 mg/kg) for an additional 12 weeks (Figure 1).
Each group consisted of eight mice.

| Microarray-based gene expression data analysis
To obtain differential genetic analyses of diabetes, the microarray data of patients with and without diabetes experiencing HF based on the GSE26887 data set, the micro-RNAs involved in the pathophysiology of DbCM based on the GSE44179 data set, and expression data using rat ventricles on days 3, 28 and 42 after the STZ injection based on the GSE4745 data set were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/ geo/). DbCM-related genes were screened using Excel (Microsoft), |log2FC| > 2.0 and adj. p. Val (p value after correction) <0.05.

| Quantitative reverse-transcriptionpolymerase chain reaction (qRT-PCR)
The expression levels of ZFAS1, CCND2 and miR-150-5p were determined using qRT-PCR. Total RNA from the left ventricular tissue and cardiomyocytes was extracted using the TRIzol reagent
Forty-eight hours after transfection, cardiomyocytes were retrieved for the biotin-based pull-down assay (Thermo Fisher Scientific, USA), which was performed according to the manufacturer's protocol. The expression levels of ZFAS1 were measured using real-time PCR.
Magnetic beads coated with a ZFAS1 probe or a random probe were added to the cardiomyocyte lysate. After washing and enrichment of the bead/RNA complex, miR-150-5p was eluted from streptavidin beads. The expression level of miR-150-5p was determined through Northern blotting (Thermo Fisher Scientific, USA), according to the manufacturer's protocol.

| Immunofluorescence staining
Protein localization of the ferritin heavy chain was determined by immunohistochemical staining. After fixing with 4% paraformaldehyde for 15 min, permeabilization with 0.5% Triton X-100 for 20 min and blocking with 4% goat serum for 30 min at 37℃, adherent experimental cardiomyocytes were incubated with the primary antibody against ferritin heavy chain (FTH1, ab65080) at 4℃ overnight. (P36941; Invitrogen) for 3 min, images were captured using a Nikon Eclipse Ti-U fluorescence microscope.

| Histology and immunohistochemistry
The relative expression of FTH1 and 4-HNE was determined through immunohistochemical staining of the left ventricular myocardial tissues of experimental mice. After dewaxing in a 60℃ incubator, hydration with xylene and anhydrous ethanol, and antigen repair with citrate antigen retrieval solution (Beyotime, China), 5-mm-thick sections of the left ventricular myocardial tissues were incubated with primary antibodies against ferritin heavy chain (FTH1, ab65080) and 4-hydroxynonenal (4-HNE, ab46545) at 4℃ overnight. Sections were then incubated with secondary antibodies at 37℃ for 30 min and stained with 3,3'-diaminobenzidine (Gene Tech, China) at 37℃ for 5 min. Images were then immediately captured in the dark using a Nikon Eclipse Ti-U fluorescence microscope (Tokyo, Japan).

| Masson staining
Cardiac collagen content was measured with Masson staining.
After dewaxing in a 60℃ incubator, hydrated with xylene and anhydrous ethanol, and stained with haematoxylin and Lichun red acid, 5-mm-thick sections of the left ventricular myocardial tissues were finally stained with 1% phosphomolybdic acid. Images were immediately captured in the dark using a Nikon Eclipse Ti-U fluorescence microscope (Tokyo, Japan). Collagen fibres were blue (aniline blue) or green (bright green), and the muscle fibres and cellulose were red.

| Monobromobimane (MBB) staining
MBB staining was used to determine glutathione (GSH) levels in cardiomyocytes. Briefly, cardiomyocytes were stained with MBB (20 μM; Sigma-Aldrich, USA) in PBS for 15 min at 37℃, and images were immediately captured in the dark using a Nikon Eclipse Ti-U fluorescence microscope (Tokyo, Japan).

| JC-1 staining
JC-1 staining was used to determine the mitochondrial membrane potential. Cardiomyocytes were stained with JC-1 (MCE, NJ, USA) in PBS for 30 min at 37℃. Thereafter, the cells were immediately imaged in the dark using a Nikon Eclipse Ti-U fluorescence microscope (Tokyo, Japan).

| Echocardiography
Cardiac function was evaluated using echocardiography. We used a parasternal long-axis view. Transthoracic echocardiography was performed using a Philips iE33 system (Philips Medical, the Each group consisted of eight mice.

| Statistical analysis
SPSS Statistics version 26.0 (SPSS Inc., Chicago, IL, USA) was used for data analysis. Data are expressed as mean ± standard deviation.
Student's t test and analysis of variance were used to analyse the differences between two or more groups. All experiments were repeated at least three times. Statistical significance was set at p < 0.05.

| Upregulated ZFAS1 expression and increased ferroptosis in mice with DbCM-and HG-treated cardiomyocytes
The differential HF-related gene profiles of patients with and without diabetes were screened using the GSE26887 data set from the GEO database. lncRNA-ZFAS1 was found to be significantly upregulated in diabetic patients with HF ( Figure 2A). Therefore, to determine whether ZFAS1 is involved in DbCM, the expression level of ZFAS1 in the left ventricular myocardial tissues of db/db mice and HG-treated cardiomyocytes was determined by qRT-PCR. ZFAS1 was significantly upregulated in the same manner ( Figure 2B, C). FTH1, a key iron storage protein involved in iron metabolism that acts as a ferritinophagy biomarker, was decreased in the left ventricular myocardial tissues of db/db mice, which was alleviated by treatment with ferrostatin-1(Fer-1), a ferroptosis inhibitor, as evaluated by immunohistochemical staining ( Figure 2D). Furthermore, 4-HNE, the final product of lipid hydroperoxidation, was increased in the left ventricular myocardial tissues of db/db mice; however, it was downregulated in the Fer-1 + db/db group, as demonstrated by immunohistochemical staining ( Figure 2D).
Western blotting showed that GPX4, which could terminate the process of ferroptosis, was decreased in the left ventricular myocardial tissues of db/db mice; however, it was upregulated in the Fer-1 + db/db group ( Figure 2E-F). Consistent with the in vivo results, FTH1 was rarely observed in the cytoplasm of HG-treated cardiomyocytes ( Figure 2G). Furthermore, the expression level of GPX4 was reduced in HG-treated cardiomyocytes; however, it was upregulated in the Fer-1 + HG group, as measured by Western blotting ( Figure 2H-I). Collectively, these findings indicate that the expression level of ZFAS1 is upregulated and ferroptosis is increased in DbCM-and HG-treated cardiomyocytes.

| Inhibition of ZFAS1 repressed ferroptosis in mice with DbCM-and HG-treated cardiomyocytes
To further identify the function of ZFAS1 in the DbCM process, we

| lncRNA-ZFAS1 can bind with miR-150-5p to regulate the expression of CCND2
The differential role of the micro-RNA profiles in the pathophysiology of DbCM was screened using the GSE44179 data set from the GEO database, and miR-150-5p was found to be substantially reduced ( Figure 4A). To determine whether miR-150-5p is involved in DbCM, the expression level of miR-150-5p in the left ventricular myocardial tissues of db/db mice and HG-treated cardiomyocytes was determined by qRT-PCR. Accordingly, miR-150-5p was found to be significantly downregulated in the same manner ( Figure 4B, C).
To highlight the potential molecular mechanism by which ZFAS1 and miR-150-5p regulate DM, we explored the underlying target binding sites of ZFAS1 and miR-150-5p. The predicted binding sites of miR-150-5p and ZFAS1 are displayed in Figure 4D. These sites were also analysed using TargetScan software. Based on the dual-luciferase reporter assay, transfection with miR-150-5p mimics significantly reduced the relative firefly luciferase activity of wt-ZFAS1; however, the mut-ZFAS1 luciferase activity remained unaffected ( Figure 4E). Further, based on a biotin-avidin pull-down assay performed to determine whether miR-150-5p could directly bind to ZFAS1, ZFAS1 was found to be pulled down by biotinylated wild-type miR-150-5p. Additionally, miR-150-5p could not pull down ZFAS1 upon introduction of the miR-150-5p mutations that destroy base pairing between ZFAS1 and miR-150-5p. These findings indicate that the identification of miR-150-5p and ZFAS1 is sequencespecific ( Figure 4F). An inverse pull-down assay was also performed to determine whether ZFAS1 could pull down miR-150-5p, and miR-150-5p could thus be co-precipitated by ZFAS1 using a biotinlabelled specific ZFAS1 probe ( Figure 4G).
Analysis of the GSE44179 data set from the GEO database revealed that CCND2 was significantly decreased in rat ventricles after STZ injection ( Figure 4H). To determine whether CCND2 is in- and CCND2 ( Figure 4M). Based on the dual-luciferase reporter assay, transfection with the miR-150-5p mimics significantly reduced the relative firefly luciferase activity of wt-CCND2; however, the mut-CCND2 luciferase activity remained unaffected ( Figure 4N). These findings suggest that ZFAS1 can bind with miR-150-5p to regulate the expression of CCND2.

| ZFAS1 promotes ferroptosis in mice with DbCM and HG-treated cardiomyocytes by modulating miR-150-5p
Owing to the interaction between ZFAS1 and miR-150-5p, we sought to determine whether ZFAS1 could regulate ferroptosis through miR-150-5p. As shown in Figure 5A  As shown in Table 1

| ZFAS1 promotes ferroptosis in mice with DbCM-and HG-treated cardiomyocytes by modulating CCND2
To further examine the role of CCND2 in the positive function of Bax and Bcl-2 in experimental cardiomyocytes was assessed by Western blot analysis. #p < 0.05 versus CON group + Ad-NC +mimic NC or db/+ + Ad-NC + mimic NC group, *p < 0.05 versus HG + Ad-NC + mimic NC or db/db + Ad-NC + mimic NC; data are expressed as mean ± standard deviation (SD, n = 3) TA B L E 1 ZFAS1 improved the cardiac function in the hearts of db/db mice by modulating miR-150-5p    Figure 2A). As expected, ZFAS1 was upregulated in mice with DbCM-and HG-treated cardiomyocytes ( Figure 2B, C). ZFAS1 is a cardiac-related lncRNA. Knockdown of lncRNA-ZFAS1 has been shown to protect cardiomyocytes from MI. 14 However, reports have emphasized that ZFAS1 is upregulated to promote cardiac disease 41 and induce mitochondria-mediated cardiomyocyte apoptosis. 16 Ferroptosis, a unique cell death characterized by the stimulation of reactive oxygen species, results in mitochondrial dysfunction, which is induced by iron catalytic activity and lipid peroxidation. 42 The accumulated evidence demonstrates that ferroptosis is a critical form of cardiomyocyte death. 43,44 Based on immunohistochemical staining ( Figure 2D) and immunofluorescence staining ( Figure 2G), the expression levels of the iron storage protein FTH1 were upregulated in the Fer-1 + db/db and Fer-1 + HG groups, compared with the db/db and HG groups. Additionally, immunohistochemical staining also revealed the expression of 4-HNE ( Figure 2D). Finally, based on Western blotting, GPX4 was found to be upregulated in the Fer-1 + db/db and Fer-1 + HG groups, com-  Figures 5E, 6E). Based on JC-1 staining, the mitochondrial membrane potential was reduced ( Figures 5G, 6G).
According to Masson staining, all changes related to ferroptosis resulted in the accumulation of collagen deposition, which induced cardiac fibrosis in the myocardial tissue ( Figures 5A, 6A). Further, the expression of apoptosis-related genes including cleaved caspase 3, Bax and Bcl-2 was found to be increased as determined by Western blotting (Figures 5C, D, H, I, 6C, D, H, I). These findings suggest that ZFAS1 inhibition impedes cardiomyocyte ferroptosis by sponging miR-150-5p to activate CCND2 against DbCM ( Figure 7).
Collectively, all these findings strongly suggest that ZFAS1 simulation promotes ferroptosis in DbCM. Further, lncRNA-ZFAS1 was recognized to act as a ceRNA to sponge miR-150-5p and could downregulate CCND2. Most importantly, ZFAS1 inhibition suppressed cardiomyocyte ferroptosis and attenuated DbCM progression. Targeting lncRNA-ZFAS1 will enable further development of novel treatments for DbCM.

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