Abstract Objective The aim of this study was to explore the potential mechanism underlying the effects of a high-fat diet (HFD) on heart failure (HF), utilizing the integration of serum metabolomics. Method A Sprague-Dawley rat model for HF induced by pressure overload via abdominal aortic contraction was established. A metabolomics approach based on liquid chromatography–mass spectrometry (LC-MS) was performed to analyze the serum biomarkers from control + chew diet (CCD) group, model + chew diet (MCD) group, and model + HFD(MHFD) group. Principal component analysis (PCA) and orthogonal projection to latent structures-discriminan* t analysis (OPLS-DA) were employed to identify differences in metabolic profiles among the three groups. Results There were significant differences in OPLS-DA scores between the CCD vs. MCD groups and MCD vs. MHFD groups. In comparison to the CCD group, the MCD group demonstrated significant alterations in multiple metabolites, such as Lyso PE (20:4 (5Z, 8Z, 11Z, 14Z))/0:0, Lyso PE (16:0/0:0), L-glutamine, N2-γ-glutamine, butyrylcarnitine, PHOOA-P, and so on. In addition, the levels of metabolites such as cholic acid, (9 S, 10E, 12 S, 13 S) −9,12,13-trihydroxyoctadec-10-enoylcarnitine, and 8-deoxy19,20-epoxycytidine C in the MHFD group were significantly higher than those in the MCD group. 17 metabolites showed a significant restorative trend from HF to a normal condition, implying their significant relationship with the positive effects of HFD on HF. These metabolites can be considered potential biomarkers for HFD treatment. Conclusions These results can be used to better understand the effects of HFD on HF rats, which will provide new ideas for treatment. Supplementary Information The online version contains supplementary material available at 10.1186/s12872-025-04932-0. Keywords: High fat diet, Rat, Serum metabolites, Heart failure, Non-targeted metabolomics Introduction Heart failure (HF) is an intricate clinical syndrome that results from diminished cardiac output due to structural or functional cardiac abnormalities [[32]1]. It affects an estimated 26 million people across the globe [[33]2] and is progressively acknowledged as a critical public health concern internationally [[34]3]. HF is associated with significant clinical burdens and economic costs globally [[35]4], accounting for approximately 2–3% of total healthcare expenditures worldwide [[36]5, [37]6]. Despite these enormous costs, mortality from HF still remains high. Therefore, there is an urgent need to find effective measures to treat HF. HF is characterized by a metabolic imbalance of fatty acids (FA) and glucose metabolism [[38]7]. Dietary intake patterns, which are linked to disorders in glucose and lipid metabolism [[39]8], may serve as a novel therapeutic target for HF beyond pharmacological interventions. The preponderance of epidemiological evidence from the past few decades has associated a high intake of fats with an increased risk of cardiovascular disease. However, recent studies have challenged this association [[40]9, [41]10]. Robert E. et al. have found that high-fat feeding in mice does not exacerbate HF or induce cardiac dysfunction [[42]10]. Yan Zhen Tan et al. found that short-term feeding of a high-fat diet (HFD) prevents HF induced by pressure overload through the activation of mitophagy [[43]11]. Metabolomics is a comprehensive approach that involves studying metabolites in systems biology [[44]12], and it has been widely used for identifying biomarkers, monitoring therapeutic responses, and investigating pathophysiological pathways [[45]13, [46]14]. The objective is to identify specific metabolites that play roles in health and disease, thus enabling the development of diagnostic and mechanistic biochemical biomarkers capable of monitoring changes in individual metabolic homeostasis. To the best of our knowledge, no studies have been conducted to investigate the effects of a HFD on HF in rats using a non-targeting metabolomics approach. In this study, we analyzed the metabolic profile of serum from rats with heart failure using liquid chromatography–mass spectrometry (LC-MS), and the data were processed with Progenesis QI v3.0 software and the Metabo Analyst 5.0 platform to identify metabolites and metabolic pathways. Concurrently, we investigated the effects of a HFD on HF rats, identifying metabolites that exhibited significant changes, thereby providing preliminary evidence of the HF rats’ response to HFD treatment and elucidating potential underlying mechanisms. Materials and methods Animal models Eight-week-old male Sprague Dawley rats, obtained from Wu’s Laboratory Animal Company (SPF grade, animal number: SCXK (Jing) 2019-0008), were used in the study. The animal experiments were conducted with the approval of the Ethics Committee of Fujian Medical University, and all methods were carried out in accordance with the relevant guidelines and regulations (approval number: 2022KYLLD03093). This study was carried out in compliance with the ARRIVE guidelines. The rat model of HF was established as our prior reported [[47]15]. The rats were anesthetized with 2% pentobarbital (50 mg/kg intraperitoneal) and their abdomen was sterilized. A laparotomy was performed 1 cm below the xiphoid process of the abdominal midline, carefully exposing the abdominal cavity to facilitate the separation of the abdominal aorta. Subsequently, the abdominal aorta was constricted to a diameter of 0.7 mm. The abdominal cavity of the rats was closed using 6 − 0 absorbable suture, and the surgical site was disinfected. In the sham group, rats underwent the same procedures without abdominal aorta ligation. Animal grouping and treatment After ten weeks of the surgery, the surviving rats were confirmed to have HF through echocardiography, serum BNP levels, and hematoxylin and eosin (H&E) staining. Fifteen rats with HF were then randomly divided into two groups (Fig. [48]1). The groups included the model chow diet (MCD) group and the model high-fat diet (MHFD) group. Additionally, three rats from the sham group were randomly selected to serve as the control chow diet (CCD) group. The animal chow, including HFD (45% fat, 20% protein, and 35% carbohydrate, XTHF45 according to the Research Diet D12451). The diets high in fat derived 45% of their energy from lard (88%) and soybean oil (12%). The fatty acid composition of the lard was 27% palmitate, 11% stearate, 44% oleic acid, and 11% linoleic acid. a matched control diet (10% fat, 20% protein, and 70% carbohydrate, XTCON50H), was purchased from Jiangsu Synergy Pharmaceutical & Biological Engineering Co (Nanjing, China). Rat were anesthetized in a plexiglass chamber using 5% isoflurane (USP, #NDC 13985-046-60, VetOne, Boise, ID, USA) for 5 min and then decapitated once fully sedated, as determined by the absence of a paw withdrawal reflex. Fig. 1. [49]Fig. 1 [50]Open in a new tab Schematic of the experimental design Samples collection After being fed different diets for 16 weeks, the rats were injected intraperitoneally with 2% pentobarbital (50 mg/kg). Blood was then collected from the orbital vein and left at room temperature for 1 h before being centrifuged at 3000 rpm for 10 min. The supernatant was subsequently stored at −80°C until further use. Preparation of extraction samples Preparation of extraction samples was described as prior researchs [[51]16, [52]17]. The samples, which were stored at −80 °C, were thawed, and 100 µL of each was transferred into a 1.5 mL Eppendorf tube. Subsequently, 300 µL of protein precipitant, consisting of methanol-acetonitrile in a 2:1 volume ratio and containing L-2-chlorophenylalanine at a concentration of 2 µg/mL, was added. The mixture was vortexed for 1 min. Subsequently, ultrasonic extraction took place in an ice-water bath for 10 min, followed by a resting period at −40°C for 30 min. After centrifugation at 4 °C for 10 min at 13,000 rpm, 200 µL of the supernatant was transferred into an LC-MS injection vial and then evaporated. The residual content was reconstituted in 300 µL of methanol-water (V: V = 1:4), vortexed for 30 s, subjected to ultrasound for 3 min, and then stored at −20°C for 2 h. Subsequently, the samples were centrifuged at 4 °C at 13,000 rpm for 10 min. From each tube, 150 µL of the supernatant was collected using crystal syringes, filtered through 0.22 μm microfilters, and transferred into LC vials. To monitor the stability and reproducibility of the analytical process, internal quality control (QC) samples were used in this study. The specific preparation method was as follows: an equal volume (100 µL) of each serum sample was mixed, vortexed and shaken homogeneously, and then the QC samples were prepared according to the same processing procedures as the experimental samples (including protein precipitation, centrifugation, re-dissolution and filtration). This method is based on the high performance liquid chromatography-mass spectrometry (HPLC-MS) metabolomics QC strategy proposed by Sangster et al. [[53]17]. to ensure the stability of the system and the reliability of the data. LC-MS/MS analysis A detailed description of the analysis process was presented at relevant researches [[54]18, [55]19] To analyze the metabolic profiling in both ESI positive and ESI negative ion modes, an ACQUITY UPLC I-Class Plus (Waters Corporation, Milford, USA) was used, which was fitted with a Q-Exactive mass spectrometer equipped with a heated electrospray ionization (ESI) source from Thermo Fisher Scientific (Waltham, MA, USA). An ACQUITY UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm) was employed in both positive and negative modes. The binary gradient elution system consisted of (A) water (containing 0.1% formic acid, v/v) and (B) acetonitrile (containing 0.1% formic acid, v/v) and separation was achieved using the following gradient: 0.01 min, 95% A; 2 min, 95% A; 4 min, 70% A; 8 min, 50% A; 10 min, 20% A; 14 min, 0% A; 15 min, 0% A; 15.1 min, 5% and 16 min, 95% A. All samples were kept at 4℃ during the analysis, and the injection volume was 2 µL. The mass range was set from m/z 100 to 1200. The resolution for full MS scans was set at 70,000, while for HCD MS/MS scans, it was set at 17,500. Collision energy levels were set at 10, 20, and 40 eV. The mass spectrometer operated with a spray voltage of 3800 V (+) and 3000 V (-). The sheath gas flow rate was set at 35 arbitrary units, and the auxiliary gas flow rate was set at 8 arbitrary units. The capillary temperature was maintained at 320°C, and the Aux gas heater temperature was set at 350 °C. The S-lens RF level was set at 50. Data preprocessing and statistical analysis The original LC-MS data were processed using Progenesis QI v3.0 software (Nonlinear Dynamics, Newcastle, UK) for baseline filtering, peak identification, integration, retention time correction, peak alignment, and normalization. A detailed description of the data preprocessing and statistical analysis process was provided in the relevant research [[56]20]. Main parameters of 5 ppm precursor tolerance, 10 ppm product tolerance, and 5% product ion threshold were employed to identify compounds. Identification was performed based on precise mass-to-charge ratio (m/z), secondary fragments, and isotopic distribution using various databases such as the Human Metabolome Database (HMDB), LipidMaps (V2.3), Metlin, and self-built databases. The extracted data went through further processing steps that involved removing peaks with missing values (ion intensity = 0) in more than 50% of groups, replacing zero values with half of the minimum value, and screening according to the qualitative results of the compound. Additionally, compounds with resulting scores below 36 out of 60 points were considered inaccurate and removed. Finally, a data matrix was generated by combining the positive and negative ion data. The matrix was imported into R to perform principal component analysis (PCA) and observe the overall distribution among the samples, as well as the stability of the analysis process. Orthogonal partial least squares discriminant analysis (OPLS-DA) was then used to distinguish the metabolites that differed between groups. The variable importance in projection (VIP) values obtained from the OPLS-DA model were utilized to rank the overall contribution of each variable to group discrimination. A two-tailed Student’s t-test was subsequently performed to verify whether the identified metabolites differed significantly between groups. Differential metabolites were selected based on VIP values greater than 1.0 and p-values less than 0.05. KEGG pathway analysis enrichment of the differential metabolites was then conducted using the KEGG database ([57]https://www.kegg.jp/). Fisher’s exact test was employed to analyze and calculate the significance level of metabolite enrichment for each pathway in order to determine which metabolic pathways were significantly affected. All values are presented as mean ± SD. Data were compared with one-way ANOVA or two-way ANOVA, with all ANOVA tests followed by an unpaired t-test, as appropriate. Normal distribution of data was analyzed by Kolmogorov–Smirnov normality test. Bonferroni’s correc tion for multiple comparisons was used. Differences were considered significant when *, P < 0.05; **, P < 0.01; ns, P non significance. Results Cardiac phenotypic and functional outcomes of HFD intervention in rats with heart failure To clarify the intervention effect of high-fat diet on the process of heart failure, the present study systematically assessed cardiac morphologic and functional parameters (Fig. [58]2A). Echocardiographic results of rats in the MHFD group compared with the MCD group showed that high-fat feeding in interfered with cardiac remodeling in rats (Fig. [59]2B). Compared with the MCD group, the MHFD rats exhibited significant cardiac remodeling and a remitted recovery of left heart function compared with the MCD group (Fig. [60]2C). And, food intake and growth curves during the experimental period were counted (Figure S1A-B). These results help readers to investigate the appropriate dose of high-fat diet for heart failure. Fig. 2. [61]Fig. 2 [62]Open in a new tab Heart failure relief parameters and functional validation. A Cardiac-related parameters; B Echocardiography Comparison. C LVEF index. **, P < 0.01; ns, P non significance Analytical assessment of metabolomics Serum metabolite profiles were obtained for each group in positive and negative ion mode. The base peak chromatograms for each group in positive and negative ion mode are shown in (Fig. [63]3A-F). A number of endogenous metabolites were identified from the chromatograms, including PHOOA-PE, LysoPE (16:0/0:0), LysoPE (20:4 (5Z, 8Z, 11Z, 14Z)/0:0), 2-[4,7,10-Tris(carboxymethyl)−6-[4-[3-(2,5-dioxopyrrol-1-yl)propanoylam ino]butyl]−1,4,7,10-tetrazacyclododec-1-yl]acetic acid, 1-O-Hexadecyl-sn-glycero-3-phosphocholine, and (2,3-Dimethoxyphenyl)-[1-[2-(4-fluorophenyl)ethyl]piperidin-4-yl]methan ol. The PCA scores plots (Fig. [64]4) revealed a tight clustering of QC samples, suggesting minimal instrumental drift throughout the analysis. The distance between the samples and the sample sets exceeded the distribution of the QC samples, indicating that the observed differences between samples were genuine biological variations rather than instrument drift. To further validate the robustness of the analytical platform, the relative standard deviation (RSD) of peak intensities for all metabolites in QC samples was calculated. Over 85% of metabolites exhibited an RSD < 15% in both positive and negative ion modes, confirming high repeatability of the measurements. Fig. 3. [65]Fig. 3 [66]Open in a new tab Base Peak Chromatogram in positive-ion mode (A, B, C) and negative-ion mode (D, E, F) in serum samples of CCD, MCD, and MHFD groups. A, B, C: in positive-ion mode; D, E, F: negative-ion mode. Abbreviation: CCD, control +chew diet; MCD, model+ chew diet; MHFD, model+ high fat diet Fig. 4. [67]Fig. 4 [68]Open in a new tab Differential metabolic profiles in mice serum among the CCD, MCD and MHFD groups. Abbreviation: QC, quality control; CCD, control+chew diet; MCD, model+ chew diet; MHFD, model+ high fat diet Multivariate data analysis based on MS data The standard criteria for screening significantly different metabolites between the CCD vs. MCD groups and the MCD vs. MHFD groups are a VIP value of > 1 for the first principal component of the OPLS-DA model and a p-value of < 0.05 for the T-test (Fig. [69]5A and B). We evaluated the predictive abilities of the constructed OPLS-DA model using 7-fold cross-validation, with the parameters R2X, R2Y, and Q2. The OPLS-DA models for the MCD vs. CCD groups (Fig. [70]5A: R2X = 0.838, R2Y = 1, Q2 = 0.977) and the MCD vs. MHFD groups (Fig. [71]5B: R2X = 0.682, R2Y = 1, Q2 = 0.981) were well separated. The groups were clearly isolated in each comparison condition, and samples of the same categories were grouped together among different groups. Fig. 5. [72]Fig. 5 [73]Open in a new tab OPLS-DA analysis of serum metabolic profiling variation of CCD vs. MCD groups and MCD vs MHFD groups. A Score plot of OPLS-DA based on serum profiling of the CCD group and MCD group. R2X = 0.838, R2Y = 1, Q2 = 0.977; B Score plot of OPLS-DA based on serum profiling of the MCD vs MHFD groups. R2X = 0.682, R2Y = 1, Q2 = 0.981. Abbreviation: CCD, control +chew diet; MCD, model+ chew diet; MHFD, model+ high fat diet Based on the OPLS-DA results, the screening results of differential metabolites are presented in (Fig. [74]6A and B). In the MCD vs. CCD groups, a total of 283 differential metabolites were identified, with 90 up-regulated (31.8%) and 193 down-regulated (68.2%). Similarly, in the MCD vs. MHFD groups, 106 differential metabolites were screened, including 37 up-regulated (34.9%) and 69 down-regulated (65.1%). The metabolite expression heat maps for different groups are displayed in (Figure S3 and S4). Further analysis of metabolite classes revealed distinct patterns: Lipid-related metabolites (e.g., LysoPE, LysoPC) were predominantly upregulated in the MCD group compared to CCD (red clusters in Figure S2), aligning with the known lipid metabolism dysregulation in heart failure. Amino acids such as L-glutamine and N2-γ-glutamine showed significant downregulation in MCD, suggesting altered nitrogen metabolism. Organic acids and nucleotides (e.g., 8-Deoxy19,20-Epoxycytidine C) exhibited mixed trends, with no dominant class-specific pattern. These findings highlight the systemic metabolic perturbations in HF and the partial restoration by HFD intervention (Figure S3). Fig. 6. [75]Fig. 6 [76]Open in a new tab Metabolic characteristics of serum in rat. A and B Volcano plot and heat map of normalized metabolites in serum samples of CCD vs MCD groups and MCD vs MHFD groups; Columns represent the samples, and rows represent the metabolites. Red dots represent significantly up-regulated differential metabolites in the experimental group, blue dots represent significantly down-regulated differential metabolites, and gray dots represent non-significant differential metabolites Pathways analysis For pathway analysis, more representative biological pathways were distinguished based on P value and Rich Factor. The results were demonstrated in (Fig. [77]7A and B). In this study, the original P value < 0.05 for the significant pathway. In CCD vs. MCD groups and MCD vs. MHFD groups, there were 26 and 1 pathways with p value < 0.05 respectively. Such as D-glutamine and D-glutamate metabolism, glutamatergic synapse, cocaine addiction, GABAergic synapse, long-term depression, synaptic vesicle cycle, and so on were significant pathways in CCD vs. MCD groups. Fig. 7. [78]Fig. 7 [79]Open in a new tab KEGG pathway enrichment bubble diagram. A CCD vs MCD groups; B MCD vs MHFD groups; The vertical axis represents the KEGG metabolic pathway, and the horizontal axis represents the rich factors (the ratio of the number of differential metabolites to the number of annotated metabolites in this pathway). The color of the bubble from green to red indicates that the p-value decreases in turn; the larger the bubble, the more the number of metabolites enriched in the pathway Serum metabolic characterization of HDL Cholesterol-Intervened heart failure rats In order to further explore the potential mechanism of the effect of a HFD on HF, we further explored the differential metabolites between the three groups and performed pathway enrichment analysis. The same procedures were used to analyze the serum samples derived from three group. Total of 20 metabolites were putatively identified. The Venn-diagram was drawn to show the overlap between MCD vs. CCD groups and MHFD vs. MCD groups (Figure S4). According to the results of the FA metabolism feature and Venn analysis, 17 of them were regarded as potential biomarkers of HF treatment, because they showed a significantly restoring trend from HF to normal condition. The detailed results are displayed in Table [80]1. Hierarchical cluster analysis of metabolites showed the levels of all 17 metabolites in different groups. Figure [81]8 illustrates that the CCD group and MHFD group exhibited similar patterns when compared to the MCD group, suggesting that the high-fat diet (HFD) may help restore the levels of endogenous metabolites in heart failure (HF). Among these 17 metabolites, we identified 365 metabolites strongly associated with 16 of them (abs (R) > 0.8, p < 0.05). As depicted in Fig. [82]9A, there are 365 metabolite-metabolite pairs, which include metabolites such as PHOOA-PE, LysoPE (16:0/0:0), LysoPE (20:4 (5Z, 8Z, 11Z, 14Z)/0:0), 2-[4,7,10-Tris(carboxymethyl)−6-[4-[3-(2,5-dioxopyrrol-1-yl)propanoylam ino]butyl]−1,4,7,10-tetrazacyclododec-1-yl]acetic acid, 1-O-Hexadecyl-sn-glycero-3-phosphocholine, and (2,3-Dimethoxyphenyl)-[1-[2-(4-fluorophenyl)ethyl]piperidin-4-yl]methan ol. These, along with their related metabolites, cluster together effectively, suggesting that they may share similar biological functions. The results also showed that the metabolites associated with L-Homoserine were the most abundant (Fig. [83]9B). Functional enrichment analysis revealed that the metabolites associated with these 16 compounds were predominantly enriched in the cystine-methionine metabolic pathway, D-amino acid metabolism, and other pathways. Table 1. Potential biomarkers of serum metabolic profiles based on MS measurements [84]graphic file with name 12872_2025_4932_Tab1_HTML.jpg [85]Open in a new tab Fig. 8. [86]Fig. 8 [87]Open in a new tab Clustering heatmap of 17 potential biomarkers in three groups. Rows: samples; Columns: biomarkers. Colors ranging from purple to red indicate the expression abundance of metabolites, ranging from low to high. Abbreviation: CCD, control +chew diet; MCD, model+ chew diet; MHFD, model+ high fat diet Fig. 9. [88]Fig. 9 [89]Open in a new tab A Network diagram Associations between 16 metabolites and other metabolites; B KEGG pathway enrichment bubble diagram. A The red bubble represents 16 metabolites, whereas the blue bubble corresponds to 365 metabolites associated with the former; B The vertical axis represents the KEGG metabolic pathway, and the horizontal axis represents the rich factors (the ratio of the number of differential metabolites to the number of annotated metabolites in this pathway) Discussion Prior to conducting this study, we designed an experiment to investigate the effect of abdominal aortic constriction operation in establishing a rat model of HF [[90]15]. After ten weeks of abdominal aortic ligation, noninvasive measurements such as ejection fractions (EF), fraction shortening (FS), and BNP were obtained using echocardiography to evaluate cardiac function. Based on the results, it can be concluded that the abdominal aortic constriction operation successfully and effectively established a rat model of HF. The normal heart primarily relies on the oxidation of FA to meet its continuous high-energy demands [[91]21]. However, when HF occurs, the substrate required for the heart to maintain energy metabolism becomes more dependent on glucose, accompanied by the inhibition of FA utilization [[92]22]. Recent studies suggest that a HFD may reduce cardiac dysfunction, prevent the progression of HF, and decrease mortality by increasing the utilization of FA and activating mitophagy [[93]10, [94]23, [95]24]. This study validates the conclusion that among specific markers of the rat model of HF, 17 potential biomarkers showed a significant trend towards recovery from HF to a normal state when subjected to a HFD. This finding suggests that an HFD may have a beneficial effect on HF. In this study, we found that high-fat diet (HFD) intervention significantly restored the serum expression levels of 17 metabolites in heart failure rats by untargeted metabolomics analysis, and this metabolic remodeling was significantly associated with improved cardiac function and reduced cardiac hypertrophy. Notably, the HFD formulation in this study was dominated by polyunsaturated fatty acids (PUFA, 55%), with only 20% saturated fatty acids (SFA), which is mechanistically different from the negative effects of lard-based HFD (SFA 40%) in the study by William et al. [[96]25] PUFA enhances mitochondrial β-oxidation efficiency through activation of PPARα pathway, upregulates CPT1, MCAD and other key enzyme activities, and simultaneously inhibits NF-κB-mediated inflammatory response and generation of lipid peroxidation products, thus circumventing the risk of endoplasmic reticulum stress and excessive opening of the mitochondrial membrane permeability transition pore that may be triggered by high SFA [[97]25]. In addition, it is generally recognized that a high-fat diet is a risk factor for heart failure. Previous studies have also shown that short-term (8 weeks) high-fat diets improve heart failure, but high-fat diet interventions lasting more than 16 weeks adversely affect heart failure rats [[98]26]. In contrast to the study by Tan et al., the percentage of specific unsaturated fatty acids could not be clearly compared [[99]26]. The choice of a 16-week intervention cycle was therefore based on the metabolic adaptation characteristics of rats: metabolic adaptation can be observed in the short term (8 weeks), whereas oxidative stress may be induced in the long term (> 20 weeks); this design balances the adequacy of the pathological process (10 weeks of postoperative modeling + 16 weeks of intervention) with the species-specific risk. However, the present study did not directly validate the mechanisms of PUFA regulation of autophagy or mitochondrial dynamics, which will need to be further addressed in the future in conjunction with gene editing modeling and dynamic metabolic flow analysis. In conclusion, PUFA-based HDL cholesterol diets may improve heart failure phenotypes through metabolic modulation rather than directly reversing structural damage, but their clinical translation needs to be individualized with patients’ fatty acid profiles and comorbidities. In this study, LC-MS technology was employed to investigate the non-targeted metabolomics of serum in HF rats. The metabolic profiles of HF rats were compared with those of normal rats through multivariate statistical analysis, aiming to identify potential biomarkers of HF. These findings open up new possibilities for the early detection and treatment of HF. Compared with CCD group rats, the MCD group rats exhibited 183 differentially expressed metabolites. Among them, several significantly different identifiable markers included LysoPE (20:4 (5Z, 8Z, 11Z, 14Z))/0:0, Lyso PE (16:0/0:0), L-glutamine, N2-γ-glutamine, butyrylcarnitine, and PHOOA-PE. LysoPE (16:0/0:0) and LysoPE (20:4(5Z, 8Z, 11Z, 14Z)/0:0) are two crucial forms of LysoPEs. Specifically, LysoPE (16:0/0:0) is a saturated fatty acid possessing a chain length of 16 carbon atoms, whereas LysoPE (20:4(5Z, 8Z, 11Z, 14Z)/0:0) is an unsaturated FA featuring a chain length of 20 carbon atoms. Our results are consistent with the conclusion reached by Gema Marin-Royo et al. that biomarkers can serve as an important tool for diagnosis and monitoring of disease when plasma biomarker levels reflect directional changes or correlations similar to those observed in affected tissues [[100]27]. LysoPE is a type of phospholipid molecule that plays a crucial role in cell signal transduction, regulation of inflammatory response, lipid metabolism, cell membrane stability, and antioxidant activity [[101]28, [102]29]. It has been reported that LysoPEs could inhibit the fat oxidation of hepatocytes and facilitate lipid deposition [[103]30]. Considering the biological characteristics of myocardial FA oxidation inhibition during HF [[104]22], we can infer that LysoPE is involved in the onset and progression of HF. Pathway analysis revealed a significant enrichment of D-glutamine and D-glutamic acid metabolic pathways in HFD rats compared to normal rats. Glutamine is a conditionally essential amino acid that predominantly exists in skeletal muscle and represents the most abundant free amino acid in the body [[105]31]. Several studies have reported that phenylacetylglutamine (PAGln), formed by the binding of glutamine to phenylacetic acid, can promote cardiovascular disease (CVD)-related phenotypes through host G protein-coupled receptors (α2A, α2B, and β2-adrenergic receptors (ADR)) [[106]32]. These studies have also found an association between PAGln and an increased risk of adverse cardiovascular events in patients with HF [[107]33]. Therefore, we suspect that this pathway may be related to the onset and progression of HF. This study also discusses the differences in metabolites between the MCD group and MHFD group, aiming to identify the precise metabolic mechanism underlying the effect of HFD on HF. The differential metabolites between the two groups were mainly lipids and lipid-like molecules such as hyocholic acid, (9 S,10E,12 S,13 S)−9,12,13-Trihydroxyoctadec-10-enoylcarnitine, 9-methoxy-pentadecanoic acid, and 8-desoxy-19,20-epoxycytochalasin C, among others. Regarding the pathway analysis, it is notable that we did not identify any significantly enriched pathways. The following reasons may explain this outcome: the biological hypothesis underlying pathway enrichment analysis posits that alterations in upstream genes within a given pathway can affect downstream related genes, resulting in changes to the expression of numerous genes within said pathway and achieving statistical enrichment. However, in numerous pathways, genes may not necessarily regulate each other; instead, they may participate in different stages or components of a specific biological process [[108]34]. In addition, the information on pathways included in databases such as KEGG is not yet comprehensive or exhaustive [[109]35]. The majority of pathways solely cover a single regulatory pathway, and there is limited knowledge regarding the transcription factors involved in intermediate stages or the formation of other metabolites. To gain a deeper understanding of the direct regulation of HFD treatment, we conducted metabolomic analysis and identified 17 hub metabolites as potential targets. Consistent with our observations of metabolic profiles, we found that 9 hub metabolites were elevated in the MCD group and exhibited a decline after HFD treatment. Conversely, 8 hub metabolites were initially found to be low in the MCD group but showed an increase following HFD treatment. The physiological significance of several crucial hub metabolites is elaborated upon below. Butyrylcarnitine (C4-acyl carnitine) is categorized as a short-chain acylcarnitine, which is an ester compound formed through the conjugation of free carnitine and acyl-coenzyme A (acyl-CoA) generated by short-chain FA [[110]36]. Under normal circumstances, human cardiomyocytes primarily rely on FA oxidation to provide the substantial amount of energy required [[111]37]. In HF, myocardial fatty acid oxidation is hindered, leading to the accumulation of acyl-CoA and the conversion of a significant amount of carnitine-bound acyl-CoA into acylcarnitine [[112]22]. The accumulation of acylcarnitine has been recognized as an indicator of incomplete fatty acid oxidation in individuals experiencing heart failure with reduced ejection fraction [[113]38, [114]39]. HFD can counteract the fatty acid oxidation that occurs due to HF [[115]23, [116]24], ultimately diminishing the accumulation of acyl-CoA and reversing the build-up of acylcarnitine. PHOOA-PE belongs to the category of oxidized phospholipids. Previous studies have indicated that the continuous elevation of oxidized phospholipids in the bloodstream can induce pro-inflammatory effects and harm blood cells as well as vascular wall cells [[117]40, [118]41]. Consequently, it is often considered an indicator of chronic vascular disease and acute inflammation [[119]42, [120]43]. Intriguingly, our study discovered a significant reduction in the concentration of PHOOA-PE in plasma when subjects were placed on a HFD. As a result, we hypothesize that a HFD could be an effective approach in ameliorating HF by mitigating oxidative stress. Conclusions In summary, this study utilized a non-targeted LC-MS serum metabolomics method to examine the metabolic characteristics of rats with HF. Numerous significant metabolites associated with multiple metabolic pathways were detected. Upon evaluating the impact of a HFD on a rat model of HF, clear grouping of the three cohorts was achieved through PCA and OPLS-DA. Combined with biochemical measurements, the serum metabolic profile demonstrated that the HFD was capable of restoring the profile to its normal state. This restoration implicated the significant involvement of 17 potential biomarkers of HFD treatment, primarily associated with the cardioprotective effects of HFD. Serum metabonomics improved our comprehension of the precise mechanisms by which HFD affects HF. Supplementary Information [121]Supplementary Material 1.^ (237.6KB, docx) Acknowledgements