The LHD monomer (purity > 95%) (Figure 1A), provided by the Department of Pharmaceutical and Chemical Research, Harbin Medical University, was dissolved in dimethyl sulfoxide (DMSO). Palmitic acid (PA) was purchased from Sigma (St. Louis, MO) and dissolved in 0.1 N sodium hydroxide and 15% bovine serum albumin (BSA), filtered, and then stored at -20°C. The high-fat diet (HFD) feed, consisting of 77.6% maintenance feed, 10% lard, 10% yolk powder, 2% cholesterol, 0.2% bile salt, and methylthiouracil 0.2%, were purchased from HuaFuKang Biotechnology (Beijing, China).

The animal procedures in this study were approved by the Animal Experimental Ethics Committee of Harbin Medical University. To avoid the complicating effects of estrogen on cardiovascular disease (Sabbatini and Kararigas, 2020; Meng et al., 2021), healthy male Wistar rats (180–220 g), purchased from the Experimental Animal Center of the Affiliated Second Hospital of Harbin Medical University (Harbin, China), were used in the study. The rats were maintained on a HFD for 4 weeks to establish hyperlipidemia, which was determined by analyzing the serum levels of TC, TG, LDL-C, and high-density lipoprotein cholesterol (HDL-C) in blood samples from orbital region of rats (Supplementary Figure S1). Following the method of previous studies (Su et al., 2020), the rats were divided into four groups: HFD, HFD with LHD 50 mg kg⁻¹, HFD with LHD 100 mg kg⁻¹, and HFD with atorvastatin 7.2 mg kg⁻¹ (Figure 1B). In control (CTL) group, an equal volume of distilled water was administered. Heart tissue and serum samples were collected after 6 weeks of continuous daily intragastric administration for the corresponding follow-up experiments.

Human cardiomyocyte AC16 cell line was kindly gifted by Prof. Ben-zhi Cai (Department of Pharmacy at The Second Affiliated Hospital) and grown in DMEM containing 10% fetal bovine serum (Biological Industries, Israel), 1% penicillin, and 1% streptomycin at 37°C in a humidified atmosphere containing 5% CO₂ (v v⁻¹). Cells were treated with LHD (25 µM) for 30 min and then added PA (500 µM) for 16 h to establish an in vitro model of inflammatory response induced by HFD.

For transfection, AC16 cells were grown to 60–70% confluence. We used Lipofectamine RNAiMAX Reagent (Thermo Fisher Scientific, CA, United States, 13778075) to transfect small interfering RNAs (siRNAs) and Lipofectamine 3,000 (Thermo Fisher Scientific, CA, United States, 3000015) to transfect plasmids for gene overexpression. The transfection was performed in accordance with the manufacturer’s instructions (Han et al., 2021). Briefly, 2 µl targeting siRNA or negative control (NC) siRNA was mixed with 150 µl Opti-MEM (Thermo Fisher Scientific, CA, United States, 31985070) and 9 µl Lipofectamine RNAiMAX reagent was mixed with 150 µL Opti-MEM together at room temperature. For the overexpression analysis, the pHG-CMV-Kan2-FTO plasmid (Yingrun Biotechnology, Changsha, China, HG-HO080432) and pHG-CMV-Kan2-CD36 plasmid (Yingrun Biotechnology, Changsha, China, HG-HO000072) were diluted and transfected into AC16 cells. After 24 or 48 h of transfection, the cells were harvested for analysis. The siRNA sequences are shown in Supplementary Table S1.

Total RNA was isolated from the heart tissue and cells by using Trizol reagent (Invitrogen, Carlsbad, CA, United States); 0.5–1 µg of RNA was used to prepare cDNA using the Reverse Transcription Kit (Vazyme Biotech, Nanjing, China, R223-01) in accordance with the manufacturer’s protocol. qRT-PCR was performed in a 10 µl volume with SYBR Green PCR Master Mix (Roche, Switzerland, RC-4913914001). The qRT-PCR analysis was performed on a LightCycler^® 480 II (Roche, Switzerland) comprising initial denaturation, annealing, and extension steps. The real-time PCR conditions were: denaturation at 95°C for 10 s, followed by 40 cycles of 95°C for 10 s and 55°C for 30 s. The sequences of the specific primers used for qRT-PCR are shown in Supplementary Table S2.

Abdominal aortic blood was collected and centrifuged at 3,000 rpm for 15 min at 4°C to gather serum samples. To verify the protein expression of TNF-α and IL-6 in rat serum, the rat TNF-α ELISA kit (ABclonal, Wuhan, China, ab208348) and the rat IL-6 ELISA kit (ABclonal, Wuhan, China, ab100712), respectively, were used in accordance with the kit instructions. The supernatant was centrifuged at 1,000 rpm for 5 min prior to the analysis.

The CETSA experiment was based on previously reported methods (Martinez Molina et al., 2013; Jafari et al., 2014). Briefly, after 100 µM LHD treatment for 3 h, the cells were collected in PBS and centrifuged at 300 g for 5 min, and the supernatant was discarded. Aliquots of the cells were transferred to separate Eppendorf tubes, subjected to a temperature gradient, and kept at room temperature for 3 min. After three freeze-thaw cycles, the supernatant was centrifuged and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

The DARTS assay was performed in accordance with a previously reported method (Lomenick et al., 2009; Pai et al., 2015). Briefly, we divided the cells into three groups: control (CTL), LHD, and input. First, we centrifuged the cells at 18,000 g for 10 min at 4°C in lysis solution. An appropriate volume of 10× TNC was added to the supernatant to detect the protein concentration using the bicinchoninic acid assay. Cell lysates were incubated with 83.4, 166.7, and 250.0 µM LHD at room temperature for 1 h before pronase digestion. Finally, the lysates were analyzed by western blotting to determine the binding between LHD and FTO.

Western blotting analysis was performed as described in the previous report (Zhao et al., 2021). The cells were lysed in lysis buffer. After centrifugation, the supernatant was collected and the protein concentration was quantified by BCA assay kit. The samples were boiled at 95°C for 5 min and immediately frozen in a –80°C refrigerator. The primary antibodies were used as follows: anti-CD36 (ABclonal, Wuhan, China, A5792, 1:1,000), anti-GAPDH (ABclonal, Wuhan, China, AC033, 1:1,000), anti-FTO (ABclonal, Wuhan, China, A1438, 1:1,000), anti-P65-S536 (ABclonal, Wuhan, China, AP0475, 1:1,000), anti-IKB-α (Proteintech, Wuhan, China, 10268-1-AP, 1: 1,000), and anti-β-actin (ABclonal, Wuhan, China, AC026, 1: 1,000).

The Wistar rats were anesthetized by intraperitoneal injection with avertin and the hair near the chest area was removed, as previously described (Cai et al., 2020). The cardiac function and heart dimensions were determined by two-dimensional echocardiography. Echocardiography was performed using a Vevo 1,100 VisualSonics device (VisualSonics, Toronto, ON, Canada). The left ventricular ejection fraction (EF%) and left fractional ventricular shortening (FS%) were calculated using M-mode images.

The heart tissue was fixed with 4% paraformaldehyde, embedded in paraffin, and sectioned into 5 μM coronal slides for further analysis. Hematoxylin eosin (H&E) staining (Solarbio, Beijing, China, g1260) was performed according to the method recommended by the manufacturer.

The tissues were fixed with paraformaldehyde and cryosectioned. The slides were first rinsed with PBS to wash off the embedding medium, and then soaked with 60% isopropyl alcohol. Staining with Oil Red O (Solarbio, Beijing, China, G1120) before washing with 60% isopropyl alcohol until the background was colorless. After counterstaining with hematoxylin, the slides were sealed with glycerol-gelatin.

To determine the level of m⁶A in cells, the dot blot assay were performed on total RNA or poly (A)⁺ RNA, as described previously (Nagarajan et al., 2019). In brief, AC16 cells were treated with LHD (25 µM) for 16 h and transfected with FTO siRNA for 24 h before the dot blot assay. The RNA samples were diluted in RNase-free buffer, denatured at 95°C for 5 min, and immediately cooled, and then crosslinked by UV irradiation following stained with methylene blue (Solarbio, Beijing, China, G1300). After incubating with 5% skim milk, the membrane was detected with m⁶A antibody (ABclonal, Wuhan, China, A19841, 1:1,000). Finally, the membrane was analyzed using an Odyssey (LICOR Biosciences, Lincoln, NE, United States).

All data were expressed as mean ± standard deviation of mean unless otherwise noted. Data analysis was performed using GraphPad Prism 7.0 software. The significance of the differences was analyzed using Studentʼs t-test or one-way analysis of variance (*p < 0.05, **p < 0.01, ***p < 0.001 compared with the CTL group, ^# p < 0.05, ^## p < 0.01, ^### p < 0.001 compared with the HFD or PA group). p < 0.05 was considered statistically significant.

To determine the role of LHD (Figure 1A) in hyperlipidemia and the associated cardiac function, we employed an HFD-induced rats model of hyperlipidemia. The in vivo experimental protocol is shown in Figure 1B. First, we analyzed the serum concentrations of TC, TG, LDL-C, and HDL-C in Wistar rats fed on a HFD for 4 weeks. Compared with the CTL group, the circulating concentrations of TC, TG, and LDL-C were higher in the HFD group (Supplementary Figure S1). After 6 weeks of continuous treatment with LHD, the serum levels of TC, TG, and LDL-C were significantly reduced, but there were no significant change in HDL-C levels. The same changes were observed in the atorvastatin treatment group (Figure 1C). Next, we performed Oil Red O staining to analyze lipid aggregation in the heart, and found that the administration with HFD for 10 weeks led to the accumulation of excess lipids in cardiomyocytes; however, treatment with LHD or atorvastatin significantly alleviated this accumulation (Figure 1D). Moreover, cardiac echocardiography showed that, compared with the CTL group, the administration of HFD clearly decreased the EF% and FS%; in contrast, LHD treatment significantly mitigated the hyperlipidemia-induced cardiac systolic and diastolic dysfunction (Figures 1E,F). These results indicated that LHD treatment altered the hyperlipidemia-associated changes in lipids and cardiac function in vivo.

Hyperlipidemia-induced dysfunction is usually accompanied by a spontaneous inflammatory response in cardiomyocytes (Wang Y. et al., 2017). Therefore, we performed an ELISA to detect the expression of pro-inflammatory cytokines. As shown in Figure 1G, hyperlipidemia resulted in higher serum concentrations of TNF-α and IL-6, but LHD and atorvastatin treatment significantly reduced the expression of inflammatory factors. Furthermore, H&E staining revealed that LHD and atorvastatin treatment reduced infiltrative inflammation in the heart of HFD-fed rats (Figure 1H). These results implied that LHD effectively reduced the inflammation associated with hyperlipidemia.

Hyperlipidemia and other lipid metabolic diseases induce inflammation that are predominantly associated with saturated fatty acids. Of the various fatty acids, PA (16:0) plays an essential role in this process (Liu et al., 2017). To detect the anti-inflammatory effect of LHD in vitro, we stimulated AC16 cardiomyocytes with PA to mimic hyperlipidemia-induced inflammation. First, we observed the cell morphology after PA and LHD treatment. We found that, compared with the CTL group, PA treatment led to cardiomyocyte more rounded, but this was alleviated by PA treatment followed by LHD treatment (Figure 2A). Moreover, PA treatment stimulated the mRNA expression of the proinflammatory factors IL-6 and TNF-α, and LHD treatment significantly reduced this abnormal elevation (Figure 2B). However, LHD prevented the PA-induced phosphorylation and degradation of NF-κB P65 and IκB-α (Figures 2C,D).

Cardiomyocyte damage caused by hyperlipidemia disrupts the balance between metabolic enzymes and fatty acids (Son et al., 2018). To determine whether LHD rescued the hyperlipidemia-induced inflammatory phenotype by altering metabolism-related enzyme activity, we investigated the expression of oxidative phosphorylation-related enzymes and fatty acid metabolism enzymes in cardiomyocytes. The results showed that PA stimulation increased the expression of most metabolism-related enzymes; the most significant change occurred in CD36 expression and this increase was notably inhibited by LHD treatment (Figures 2E,F). Consistent with this, Western blot dected that CD36 expression was clearly suppressed in LHD-treated cells (Figures 2G,H). Collectively, these data indicate that CD36 may play an important role in the LHD-mediated regulation of the inflammatory response in cardiomyocytes.

To further investigate the effect of CD36 in PA-induced cardiomyocyte inflammation, we used siRNA to silence the expression of CD36 in AC16 cells. Two CD36 siRNA oligomers were tested and significant reductions were achieved in the mRNA and protein expression; siRNA2 was selected for subsequent studies (Figures 3A,B). As expected, the effect of silencing CD36 was similar to that of LHD treatment: the reduced expression of CD36 restored NF-κB phosphorylation and IκB-α expression in PA-treated cells (Figures 3C,D; Supplementary Figure S2A). Moreover, qRT-PCR results suggested that the expression of pro-inflammatory factors was markedly reduced by CD36 loss of function (Figure 3E). To further verify the important role of CD36 in LHD anti-PA-induced cardiac inflammation, we transfected the exogenous CD36 plasmid into AC16 cells. qRT-PCR analysis showed that CD36 overexpression significantly interfered with the mRNA expression of CD36, IL-6 and TNF-α rescued by LHD (Supplementary Figure S2B–D). These results suggest that CD36 is required for the LHD anti-PA stimulation of cardiomyocytes.

Next, to explore the underlying anti-inflammatory mechanism of LHD in cardiomyocytes, we used SwissTargetPrediction, a small molecule compound and protein binding prediction website, to predict potential binding partners of LHD. We found that FTO, as an adiposity and obesity-related gene and an adipose metabolism-related gene (Wang et al., 2015), was the most promising LHD-binding candidate and may play a regulatory role in PA-induced inflammation process (Figure 4A; Supplementary Table S3). To understand the binding of LHD and FTO, we performed molecular docking studies of LHD and the FTO protein using AutoDockTools 1.5.6 and PyMol 2.4 software. We found that LHD bound directly within the FTO catalytic pocket (Figure 4B). In addition, CETSA and DARTS assays were used to confirm the binding between LHD and FTO in vitro. As expected, the FTO protein resisted pronase activity in a dose-dependent manner in the presence of LHD (Figure 4C), and direct binding with LHD increased the thermal stability of the FTO protein (Figure 4D).

Notably, we found that expression of the FTO protein expression was not altered by the PA-induced inflammatory response or the anti-inflammatory effect of LHD (Supplementary Figure S3A,B). We then investigated whether LHD disrupted the enzymatic action of FTO. We observed the structure of LHD bound to FTO and found that LHD binds specifically to amino acids R96 and E234 of FTO (Figure 4E). The accumulated evidence shows that the N atom on the m⁶A purine ring interacts with the R96 and E234 residues of the FTO protein via H-bonding, which locks the m⁶A base in place (Zhang et al., 2019). To confirm these findings, we assessed whether LHD regulated FTO-mediated m⁶A demethylation in cells. We performed a m⁶A dot blot to analyze mRNA and found that m⁶A methylation was increased in LHD-treated cells (Figure 4F; Supplementary Figure S3C). Overall, these results suggested that LHD effectively bonded with the FTO protein and inhibited m⁶A demethylation by FTO.

To further investigate the effect of FTO on the PA-induced inflammatory response and CD36 expression, we used siRNA to silence intracellular FTO expression. The qRT-PCR and western blotting results showed that the two siRNAs tested, which targeted different regions of FTO, significantly reduced FTO expression, and the more effective siRNA, siRNA1, was selected for the subsequent experiments (Figures 5A,B; Supplementary Figure S4A). As reported in previous studies, FTO silencing significantly enhanced intracellular levels of the m⁶A modification (Supplementary Figure S4B). The loss of function of FTO significantly enhanced the inhibitory effect of LHD in CD36 mRNA expression (Figure 5C), and the mRNA expression of IL-6 and TNF-α was markedly suppressed by FTO siRNA treated cells (Figure 5D; Supplementary Figures S4C–E).

To verify the effect of FTO in LHD against PA-induced cardiac inflammation, we transfected exogenous FTO plasmid and confirmed that the mRNA and protein expression of FTO was significantly increased (Figures 5E,F). From the qRT-PCR analysis, we found that the mRNA expression of CD36, IL-6 and TNF-α was significantly restored in LHD-treated cells. However, this was altered by FTO overexpression (Figures 5G,H), Moreover, the FTO overexpression plasmid significantly enhanced intracellular CD36 protein expression (Supplementary Figure S4F), which indicated that FTO is required to mediate the anti-inflammatory effects of LHD in PA-treated cardiomyocytes.

It was shown that m⁶A binding proteins selectively recognize the dynamic m⁶A modification to regulate the stability of mRNA and the translation status (Wang et al., 2014). In combination with the above experiments showing that LHD, as an FTO inhibitor, regulates intracellular m⁶A levels and mitigates the PA-induced increase in CD36 mRNA expression. Therefore, we hypothesized that the m⁶A modification regulates the expression of CD36 in cardiomyocytes through its impact on mRNA transcription. To verify this hypothesis, we predicted the association between the m⁶A modification sites and CD36 transcript by SRAMP, a sequence-based predictor of m⁶A modification sites. The results showed that there were many m⁶A modification sites on the CD36 transcript (Figure 6A). Next, we determined the stability of CD36 mRNA after cells were transfected with FTO siRNA. We found that the half-life of CD36 mRNA was not significant change in FTO expression (Figure 6B). Thus, we examined the stability of CD36 mRNA following PA-induced inflammation. As expected, FTO silencing significantly decreased the half-life of CD36 mRNA compared with PA treatment (Figure 6C). Together, these results indicate that LHD/FTO-mediated m⁶A demethylation impairs the stability of CD36 mRNA and protects against fatty acid-induced inflammation.