Identification and Validation of Lipid-related Genes in Acute Myocardial Infarction

doi:10.21203/rs.3.rs-2249926/v1

Lipid metabolism plays an essential role in the genesis and progress of acute myocardial infarction (AMI). Herein, we identified and verified latent lipid-related genes involved in AMI by bioinformatic analysis. Lipid-related differentially expressed genes (DEGs) involved in AMI were identified using the GSE66360 dataset from the Gene Expression Omnibus (GEO) database and R software packages. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted to analyze lipid-related DEGs. Lipid-related genes were identified by two machine learning techniques: least absolute shrinkage and selection operator (LASSO) regression and support vector machine recursive feature elimination (SVM-RFE). The receiver operating characteristic (ROC) curves were used to descript diagnostic accuracy. Furthermore, blood samples were collected from AMI patients and healthy individuals, and real-time quantitative polymerase chain reaction (RT-qPCR) was used to determine the RNA levels of lipid-related DEGs. Fifty lipid-related DEGs were identified, 28 upregulated and 22 downregulated. Several enrichment terms related to lipid metabolism were found by GO and KEGG enrichment analyses. After LASSO regression and SVM-RFE screening, four genes (ACSL1, CH25H, GPCPD1, and PLA2G12A) were identified as potential diagnostic biomarkers for AMI. Moreover, the RT-qPCR analysis indicated that the expression levels of these four lipid-related genes in AMI patients (n = 50) and healthy individuals (n = 50) were consistent with bioinformatics analysis results. There were 4 key lipid-related genes were identified associated with AMI using machine learning. The validation of clinical samples suggested that ACSL1, CH25H, GPCPD1, and PLA2G12A are expected to be diagnostic markers for AMI and provide new targets for lipid therapy of AMI.

Health sciences/Cardiology/Cardiovascular biology

Health sciences/Diseases/Cardiovascular diseases

Health sciences/Biomarkers

Health sciences/Cardiology

Biological sciences/Biochemistry/Lipids

Acute myocardial infarction (AMI) has become the main cause of hospitalization and death worldwide, seriously threatening human health. Previous studies have suggested that AMI is a complex syndrome with multifactorial disorders. Its risk factors include early family history, smoking, hypertension, dyslipidemia, and diabetes ^1–4. The rupture of vulnerable and lipid-overloaded coronary atherosclerotic plaques can induce the formation of acute thrombus, leading to acute occlusion of blood vessels and progressing to AMI ⁵. Dyslipidemia, especially elevated levels of low-density lipoprotein (LDL) cholesterol, is believed to play a key role in the pathogenesis of atherosclerosis ^6,7. Atherogenesis begins when residues of LDL cholesterol, chylomicron, and very low-density lipoprotein (VLDL) cholesterol molecules enter the artery intima. Then, lipid radicals are oxidized and endocytosed by macrophages, followed by the formation of foam cells ⁶. Previous studies have demonstrated that every 1% reduction in LDL cholesterol levels is associated with a 1% reduction in AMI risk ^8,9. Currently, statin therapy has become the cornerstone for cholesterol-lowering drug therapy for coronary heart disease and AMI patients. Moreover, the combination of PCSK9 inhibitors with statins is recommended to reduce the risk of major adverse cardiovascular events (MACEs) in AMI patients ¹⁰. However, the combination of lipid-lowering therapy still cannot eliminate the risk of MACEs in AMI patients. This might be partly because some lipid-related genes have not been identified. Therefore, identifying new lipid metabolism-related genes associated with AMI will help develop new lipid-lowering drugs to reduce the risk of MACEs in AMI patients.

Microarray analysis is an innovative and practical method to discern susceptibility genes to deal with coronary heart disease ¹¹ and AMI ¹². Nevertheless, microarray analysis using differentially expressed genes (DEGs) might have limitations in reproducibility and sensitivity ^13,14. Machine learning can enhance the prediction and accuracy of these key genes discerned using traditional microarrays or next-generation sequencing data ¹⁵. The most frequently used machine learning techniques include the least absolute shrinkage and selection operator (LASSO) regression and support vector machine recursive feature elimination (SVM-RFE) algorithm ¹⁶. Meanwhile, the combined application of LASSO regression and SVM-RFE algorithm in identifying new lipid-related genes involved in AMI has not been conducted. Therefore, in the present study, we analyzed the GSE66360 dataset from different perspectives: 1) DEGs among lipid-related AMI genes were identified using the “limma” R package and GO and KEGG pathway enrichment analyses. 2) Machine learning methods were applied to large-scale screening and diagnostic identification of AMI-related molecular markers. 3) Lipid-related gene expression levels screened by machine learning were validated using clinical cases.

Identification of lipid-related DEGs in GSE66360

Due to the limitations of chip detection technology, a total of 673 lipid-related genes in 49 AMI samples and 50 normal samples were used to analyze DEGs. After data normalization and removal of batch differences, 50 lipid-related genes were identified, 28 upregulated and 22 downregulated (Table 1). These 50 lipid-related DEGs can be visualized in the volcano plot and heatmap (Figs. 1A, B). The expression pattern between the two groups was highlighted based on the box plot (Fig. 2). Among them, the top three upregulated genes were ACSSL1 (Acyl-CoA Synthetase Long-Chain Family Member 1), PLBD1 (Phospholipase B Domain Containing 1), and CH25H (Cholesterol 25-Hydroxylase). Meanwhile, the top three downregulated genes were ELOVL4 (ELOVL Fatty Acid Elongase 4), TNFAIP8L2 (TNF Alpha Induced Protein 8 Like 2), and CYP2E1 (Cytochrome P450 Family 2 Subfamily E Member 1).

Table 1

The 50 differentially expressed lipid-related genes in AMI samples compared to healthy samples.
Gene symbol	Log₂FC	Changes	p-value	Adjusted. p-value
ACSL1	2.1443758	Up	9.76E-14	6.57E-11
PLBD1	2.0855158	Up	1.33E-08	1.49E-06
CH25H	1.561894	Up	1.67E-10	5.63E-08
PTGS2	1.4727308	Up	1.64E-04	2.56E-03
CYP4F3	1.4307856	Up	2.77E-07	1.35E-05
CD36	1.3377248	Up	1.25E-08	1.49E-06
G0S2	1.3051173	Up	7.36E-06	2.06E-04
LPCAT2	1.274232	Up	1.13E-07	6.95E-06
CYP1B1	1.081695	Up	2.57E-07	1.35E-05
MBOAT2	0.996586	Up	9.09E-08	6.80E-06
DGAT2	0.9506635	Up	4.65E-07	2.09E-05
GPCPD1	0.9145163	Up	2.38E-05	5.52E-04
ABCA1	0.8308084	Up	3.50E-08	2.94E-06
BMX	0.8267136	Up	9.28E-07	3.90E-05
PTGS1	0.8263128	Up	8.69E-05	1.46E-03
MBOAT7	0.8240098	Up	3.66E-06	1.29E-04
SGMS2	0.8165743	Up	5.71E-04	6.86E-03
ABHD5	0.7901841	Up	1.44E-09	3.23E-07
GK3P	0.7584331	Up	1.02E-04	1.68E-03
ALOX12	0.7568656	Up	4.36E-03	3.45E-02
TNFRSF21	0.7529465	Up	2.53E-03	2.27E-02
TBXAS1	0.7461727	Up	3.82E-05	7.34E-04
CYP27A1	0.6990032	Up	2.48E-06	9.27E-05
STARD4	0.6983926	Up	3.56E-03	2.89E-02
ASAH1	0.6750605	Up	4.30E-05	7.82E-04
CYP4F2	0.6721891	Up	4.74E-06	1.59E-04
GK	0.6464259	Up	2.46E-04	3.52E-03
ALOX5AP	0.6346144	Up	1.04E-03	1.13E-02
SLC44A3	-0.590564	Down	1.18E-03	1.20E-02
CEPT1	-0.599105	Down	1.17E-03	1.20E-02
MCEE	-0.609681	Down	2.24E-03	2.12E-02
PLA2G12A	-0.612271	Down	6.79E-06	1.99E-04
PNPLA4	-0.612941	Down	2.48E-03	2.27E-02
NFYB	-0.623073	Down	2.92E-04	4.09E-03
CROT	-0.628778	Down	2.87E-05	6.43E-04
GGPS1	-0.647568	Down	3.19E-05	6.58E-04
ARV1	-0.654258	Down	3.44E-05	6.81E-04
INSIG2	-0.663348	Down	6.01E-03	4.40E-02
THEM4	-0.748033	Down	2.99E-05	6.49E-04
PHYH	-0.755787	Down	2.36E-03	2.21E-02
PIK3C2B	-0.764005	Down	4.21E-04	5.45E-03
PTPN13	-0.771322	Down	3.25E-04	4.47E-03
AGPAT5	-0.799176	Down	8.56E-05	1.46E-03
ORMDL2	-0.803203	Down	1.43E-03	1.44E-02
GDPD3	-0.810763	Down	4.53E-04	5.75E-03
LCLAT1	-0.828645	Down	7.23E-04	8.25E-03
HSD11B1	-0.973808	Down	1.14E-07	6.95E-06
CYP2E1	-1.067542	Down	5.45E-06	1.75E-04
TNFAIP8L2	-1.222639	Down	8.47E-06	2.19E-04
ELOVL4	-1.35438	Down	2.82E-07	1.35E-05

Correlation analysis of lipid-related DEGs

The 50 lipid-related DEGs in the GSE66360 were significantly correlated by the Pearson correlation analysis (Fig. 3). The positive correlation between the CD36 and ACSL1 was strongest; and the CYP4F2 had the most obvious negative correlation with FTPN13. Moreover, the ACSL1 was positively related to GPCPD1 and CH25H; the CH25H was positively related to GPCPD1 and ACSL1; The PLA2G12A was positively related to GPCPD1 while negatively related to CH25H.

Functional analyses of lipid-related DEGs

The biological functions of lipid-related DEGs were determined by GO and KEGG enrichment analyses using R software. The most enriched GO terms were fatty acid metabolic, glycerolipid metabolic, and phospholipid metabolic processes (biological processes, Fig. 4A); intrinsic component of endoplasmic reticulum membrane, integral component of endoplasmic reticulum membrane, and lipid droplet (cellular components, Fig. 4B); oxidoreductase activity, iron ion binding, and acyltransferase activity (molecular functions, Fig. 4C). The KEGG enrichment analyses showed that lipid-related DEGs were involved in the arachidonic acid metabolism, glycerophospholipid metabolism, chemical carcinogenesis-DNA adducts, and the PPAR signaling pathway (Fig. 5).

LASSO logistic regression and SVM-RFE screen of genes

The LASSO logistic regression identified 15 lipid-related genes based on the optimum λ value (Fig. 6A), whereas the SVM-RFE algorithm identified four genes (Fig. 6B). Four overlap genes including ACSL1, CH25H, GPCPD1 (Glycerophosphocholine Phosphodiesterase 1), and PLA2G12A (Phospholipase A2 Group XIIA) were identified by LASSO and SVM-RFE algorithm as key genes for subsequent analysis (Fig. 6C).

ROC curves between AMI and control groups

The expression of the four lipid-related genes between the AMI and control samples of the GSE66360 dataset were analyzed using R software, and the ROC curves were constructed. The area under the curve (AUC), unified with specificity and sensitivity, verified the intrinsic validity of diagnostic tests ¹⁷. Four lipid-related DEGs had a superior diagnostic value for AMI. Among them, the gene with the most significant diagnostic value was ACSL1 (AUC = 0.878). The other genes are: CH25H (AUC = 0.853), GPCPD1 (AUC = 0.819), and PLA2G12A (AUC = 0.747) (Fig. 7). These four lipid-related genes can be considered underlying diagnostic biomarkers for AMI.

Validation by RT-qPCR

To verify the dependability of the GSE66360 dataset, RT-qPCR was used to validate the expression levels of the above four lipid-related genes in clinical samples. The clinical data of AMI and control groups are summarized in Table 2. The RT-qPCR results suggested in Fig. 8, that the expression level of PLA2G12A (p = 2.8e-08) increased in controls compared to the AMI group, while ACSL1 (p = 2.4e-09), CH25H (p = 0.023), and GPCPD1 (p = 1.3e-08) were higher in AMI group. Hence, the RT-qPCR results performed were consistent with the main bioinformatic analysis.

Table 2

Clinical characterists between control and AMI group.
Characteristic	Control (n = 50)	AMI (n = 50)	p
Male [n (%)] ^a	35(70.0)	34(68.0)	0.829
Age (years) ^a	55.16 ± 7.22	56.52 ± 6.79	0.673
BMI (kg/m²) ^a	20.98 ± 2.74	23.69 ± 3.20	0.305
Smoking [n (%)] ^b	11(22.0)	23(46.0)	0.011
Alcohol [n (%)] ^b	13(26.0)	21(42.0)	0.091
hypertension [n (%)] ^b	15(30.0)	30(60.0)	0.003
Type 2 Diabetes [n (%)] ^b	10(20.0)	19(38.0)	0.047
CK (U/L) ^a	125.84 ± 9.35	1162.1 ± 576.30	6.20E-9
CK-MB (U/L) ^a	15.5 ± 3.06	130.35 ± 19.76	1.14E-11
ApoA1 (g/L) ^a	1.38 ± 0.30	0.96 ± 0.19	0.045
ApoB (g/L) ^a	0.82 ± 0.19	1.01 ± 0.29	0.037
TC (mmol/L) ^a	3.96 ± 0.80	4.12 ± 1.14	0.029
TG (mmol/L) ^a	1.29 ± 0.41	1.50 ± 0.66	0.023
HDL-C (mmol/L) ^a	1.63 ± 0.47	1.10 ± 0.29	0.030
LDL-C (mmol/L) ^a	2.25 ± 0.94	3.57 ± 1.26	0.003
Creatinine (µmol/L) ^a	86.14 ± 13.07	88.98 ± 17.49	0.039
Troponin T (µg/L) ^a	0.04 ± 0.03	3.22 ± 1.19	2.87E-5
BMI, Body mass index; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; Apo, Apolipoprotein; TC, Total cholesterol; TG, Triglyceride; CK, creatine kinase; CK-MB, creatine kinase-myocardial band.
^a Mean ± SD determined by t-test.
^b The rate or constituent ratio between the different groups was analyzed by the χ² test.

Verification of the potential biomarkers for AMI

Furthermore, we analyzed the gene expression levels in the AMI group and healthy individuals using the ROC curve to verify the diagnostic value of the four screened lipid-related genes as shown in Fig. 9. The AUC values of ACSL1, CH25H, GPCPD1 and PLA2G12A were 0.846 [95% confidence interval (CI): 0.764–0.929], 0.632 (95% CI: 0.517–0.747), 0.830 (95% CI: 0.749–0.912), and 0.822 (95% CI: 0.744–0.901), respectively. This result showed that these lipid-related DEGs are diagnostic biomarkers for AMI.

Acute myocardial infarction (AMI) refers to hypoxia caused by coronary atherosclerosis stenosis and myocardial necrosis caused by acute and persistent ischemia ¹⁸. Dyslipidemia is a known risk factor for AMI ¹⁹, and lipid-lowering therapy is the treatment cornerstone. Several convincing studies have shown that the combined effect of lowering triglyceride, LDL cholesterol, and total cholesterol levels yield higher cardiovascular risk than lowering LDL cholesterol levels alone ^20–22. The accumulated molecular genetic data indicate that many genes are related to AMI occurrence, including lipid-related genes ²³. However, the lipid-related genes involved in AMI have not been completely identified. Thus, it is necessary to comprehend the role of lipid-related genes in AMI diagnosis and treatment.

Herein, we retrieved data of AMI patients (GSE66360) and subjected it to differential genes analysis, and identified lipid-related DEGs associated with AMI. Lipid-related DEGs were then subjected to GO and KEGG enrichment analyses. LASSO regression is a machine learning method that recognizes variables by looking for a λ value for a minimal classification error ²⁴. SVM-RFE is another machine learning method that finds optimal variables through subtracting SVM-generated feature vectors ²⁵. We used these two algorithms to screen characteristic variables and created an optimal classification model. Four lipid-related genes (ASCL1, CH25H, GPCPD1, and PLA2G12A) were identified based on these two models, which significantly impact AMI diagnosis. Moreover, the findings of CH25H were controversial compared with previous studies and should be interpreted with caution. Nevertheless, the p values of these four lipid-related genes were < 0.05, verified by RT-qPCR and consistent with our bioinformatic analysis results.

ACSL1 is a key rate-limiting enzyme in lipid metabolism ²⁶, catalyzing the energy production of fatty acids or the production of phospholipids, cholesterol esters, and triglycerides ²⁷. Previous studies have shown that heart-specific overexpression of ACSL1 in mice increases triglyceride accumulation in cardiomyocytes ²⁸. Yuanlong Li et al. demonstrated that inhibiting ACSL1 expression in the heart can reduce lipid metabolism and promote the regeneration of cardiomyocytes ²⁹. A cohort study has shown that the expression level of ACSL1 in peripheral blood leukocytes of AMI patients was higher than that of healthy controls, and this high expression was a risk factor for AMI ³⁰. These results supported the findings of our bioinformatic analysis and suggested that ACSL1 might be a promising AMI biomarker. Moreover, PLA2G12A is a secreted phospholipase A2, but its physiological function is largely unclear. In humans, there is a suggestive association between a PLA2G12A polymorphism and response to anti-vascular endothelial growth factor therapy in patients with exudative age-related macular degeneration ³¹. Nicolaou, A et al. showed that PLA2G12A is highly expressed in aortic endothelial cells in vivo and may inhibit atherosclerosis by reducing the adhesion properties of vascular endothelial cells, which confirmed PLA2G12A as a candidate gene for atherosclerosis protection ³². This was consistent with our findings that PLA2G12A was downregulated in AMI samples and was a protective gene, possibly by reducing vascular adhesion to decrease AMI incidence.

CH25H regulates cholesterol and lipid metabolism by converting cholesterol to 25-HC, and plays an important role in regulating cellular inflammatory states and cholesterol biosynthesis in endothelial cells and monocytes ³³. Elizabeth S et al. showed that 25-HC production promotes the formation of macrophage foam cells and increases susceptibility to atherosclerosis, thereby increasing AMI risk ³⁴. However, the pro-inflammatory role of CH25H in atherosclerosis remains controversial. Other studies have shown that CH25H is involved in macrophages' functional endothelium and anti-inflammatory phenotype and that CH25H ablation increases susceptibility to atherosclerosis ³⁵. Our current study suggested that CH25H was upregulated in AMI samples, consistent with the Elizabeth S et al. results. This contradiction might be partly due to different experimental conditions requiring further study. Additionally, GPCPD1 is a key enzyme in choline and phospholipid metabolism. It can cleave glycerophocholine to form glycerol-3-phosphate and choline ³⁶. GPCPD1 has been reported to promote cell migration, metastasis, adhesion, and diffusion in breast, endometrial, and ovarian cancers. However, its biological role in cardiovascular disease remains unclear. Hence, more studies are needed to further verify our current findings.

The functional analyses were also performed to evaluate the potential biological functions of lipid-related DEGs. The GO enrichment analysis showed that these genes were closely related to fatty acid metabolism. Furthermore, the KEGG enrichment analysis revealed that the lipid-related genes were primarily associated with the PPAP signaling pathway. PPAR is activated by fatty acids and their derivatives, thereby creating a lipid signaling network between the cell surface and the nucleus ³⁷. As lipid sensors and master regulators, PPAR controls the expression of genes that function in lipid metabolism ³⁸. The PPAR signaling pathway, a crossing regulator of lipid signaling and inflammation, ³⁷ was enriched, indicating that it plays a crucial role in lipid metabolism response to AMI. A previous study has found that the downregulation of PPARγ contributes to the activation and aggregation, eventually forming micro-thromboses, finally leading to myocardial dysfunction ³⁹. These results indicated that these lipid-related genes might affect AMI occurrence through the PPAR signaling pathway. Further research is required to confirm the biological functions of lipid-related DEGs.

However, our current study also has some limitations. First, we used the dataset from circulating endothelial cells to perform the bioinformatics analysis, and used the peripheral blood mononuclear cells from myocardial infarction and normal people for verification. Although there was some sample heterogeneity, our research, like other studies using GSE66360 dataset ^40–42, obtained a satisfactory result, which fully supported our conclusion. However, more studies are needed to further confirm our findings. Second, the included clinical samples were relatively small. Therefore, our conclusions must be verified by a larger AMI cohort. Third, lipid-related DEGs were only confirmed in clinical samples, and their potential functions were not demonstrated in AMI cells or animal models. Hence, more in vivo and in vitro studies are needed to clarify the underlying mechanisms of these key genes in AMI.

In summary, four lipid-related genes involved in AMI were confirmed by bioinformatics analysis and machine learning methods. These genes might influence AMI occurrence by regulating lipid metabolism. Our current findings might help understand the mechanisms of lipid metabolism-related genes in AMI and develop future lipid-lowering treatment strategies for AMI.

Lipid-related gene dataset

The workflow diagram of this research was shown in Supplementary Figure S1. A total of 742 lipid-related genes were retrieved from Gene Set Enrichment Analysis (https://www.gsea-msigdb.org/gsea/index.jsp) (Supplementary Table S1). The AMI dataset (GSE66360) was downloaded from the GEO website (https://www.ncbi.nlm.nih.gov/geo/). The GSE66360 dataset, including 49 AMI and 50 control samples, was based on the GPL570 platform (Affymetrix Human Genome U133 Plus 2.0 Array).

Identification of lipid-related DEGs

The “limma” R package was used to identify lipid-related DEGs between AMI patients and normal participants. The threshold values were p < 0.05 and |log fold change (FC)| > 0.585 ^{43 44,45}. Heatmaps, volcano plots, and boxplot charts were plotted using “heatmap” and “ggplot2” R packages.

Functional enrichment analysis of lipid-related genes

The GO and KEGG pathway enrichment analyses were conducted using the “enrich plot” R package. Cell composition, biological processes, and molecular functions were included in the GO analysis.

Screening of lipid-related genes through SVM-RFE and LASSO logistic regression

The “glmnet” R package was used to perform LASSO logistic regression which the response type set as binomial and alpha set as 1 to identify lipid-related genes ⁴⁶. LASSO regression is a regularized penalty regression method, combining ridge regression and subset selection. It applies ordinary least squares, but the sum of absolute values of the regression coefficients is less than the predetermined constant value ⁴⁷. LASSO logistic regression is a generalization of the binomial distribution of the LASSO output variable. Herein, we used LASSO to screen lipid-related genes. Moreover, SVM-RFE is a machine learning method based on support vector machines that identify optimal variables by removing SVM-generated feature vectors ⁴⁸, and the thresholds were set as follows: halve.above = 100 and k = 5. The “E1071” R package was used to establish the SVM module to sift lipid-related genes. Then, the intersections of lipid-related genes sifted by LASSO and SVM-RFE were applied to AMI diagnostic analysis, and the ROC curves were plotted.

Clinical validation samples

From September 2021 to May 2022, 50 AMI patients (AMI group) and healthy participants (control group) were recruited from the Hunan Provincial People’s Hospital. The blood samples were collected from AMI patients within hours of admission with chest pain and before using antiplatelet or anticoagulant to eliminate the influence of possible changes in blood status after pharmacological intervention. All AMI patients underwent percutaneous coronary intervention (PCI) within 12 h of the chest pain onset. The AMI patients were diagnosed based on the 2018 guidelines for diagnosing AMI patients ⁴⁹. A total of 50 healthy individuals were enrolled in the hospital physical examination center in the same period. The exclusion criteria were: (i) active inflammation; (ii) patients receiving thrombolysis and with other underlying heart diseases (e.g., severe valvular abnormalities, cardiomyopathy, or congenital heart disease); and (iii) patients who had hepatic and/or renal dysfunction, tumors, and autoimmune diseases. All participants provided written informed consent before the beginning of the study and all experiments were performed in accordance with relevant named guidelines and regulations. This research was approved by the Ethics Committee of the Hunan Provincial People’s Hospital (approval number: [2021]-41).

Real-time quantitative polymerase chain reaction (RT-qPCR)

Peripheral blood was obtained from blood samples of patients using RNeasy™ Mini Kit (QIAGEN, Frankfurt, Germany) to extract total RNA. Total RNA was reverse transcribed into cDNA using the PrimeScript RT reagent kit (Takara Bio, Japan). RT-qPCR was performed with a LightCycler 480 II Real-time PCR instrument (Roche, Switzerland) using the TransStart Top Green qPCR SuperMix (AQ131-03, Transgen, Beijing, China).

Statistical analysis

All bioinformatics and Pearson’s correlation analyses were performed using R software (version 4.6.0, http://www.R-project.org). SPSS software (version 22.0) was used to analyze clinical data. Clinical characteristic data were analyzed using Student’s t-test and χ² test. R and Grap Pad Prism software were used for ROC curve analysis. A p < 0.05 was considered statistically significant.

Data availability

All data are provided within the manuscript or as Supplemental File.

Acknowledgements

We thank all the participants of this study.

Author contributions

Z.Y.L and F.L. designed the study, performed the statistical analyses and drafted the manuscript. P.-F.Z. and X.Q.H participated in the design and helped to draft the manuscript. F.L. collected the research data. H.W.P. performed the statistical analyses. S.L.P and Y.C. revised the manuscript. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have no competing interests.

Yamada, Y., Matsui, K., Takeuchi, I. & Fujimaki, T. Association of genetic variants with coronary artery disease and ischemic stroke in a longitudinal population-based genetic epidemiological study. Biomed Rep 3, 413–419, doi:10.3892/br.2015.440 (2015).
Abram, S. et al. Frequency, Predictors, and Implications of Abnormal Blood Pressure Responses During Dobutamine Stress Echocardiography. Circ Cardiovasc Imaging 10, doi:10.1161/circimaging.116.005444 (2017).
Chiu, M. H. et al. Coronary artery disease in post-menopausal women: are there appropriate means of assessment? Clin Sci (Lond) 132, 1937–1952, doi:10.1042/cs20180067 (2018).
Madhavan, M. V., Gersh, B. J., Alexander, K. P., Granger, C. B. & Stone, G. W. Coronary Artery Disease in Patients ≥ 80 Years of Age. J Am Coll Cardiol 71, 2015–2040, doi:10.1016/j.jacc.2017.12.068 (2018).
Anderson, J. L. & Morrow, D. A. Acute Myocardial Infarction. N Engl J Med 376, 2053–2064, doi:10.1056/NEJMra1606915 (2017).
Andreadou, I. et al. PCSK9 in Myocardial Infarction and Cardioprotection: Importance of Lipid Metabolism and Inflammation. Front Physiol 11, 602497, doi:10.3389/fphys.2020.602497 (2020).
Michos, E. D., McEvoy, J. W. & Blumenthal, R. S. Lipid Management for the Prevention of Atherosclerotic Cardiovascular Disease. N Engl J Med 381, 1557–1567, doi:10.1056/NEJMra1806939 (2019).
Grundy, S. M. et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III Guidelines. J Am Coll Cardiol 44, 720–732, doi:10.1016/j.jacc.2004.07.001 (2004).
Gotto, A. M., Jr. & Brinton, E. A. Assessing low levels of high-density lipoprotein cholesterol as a risk factor in coronary heart disease: a working group report and update. J Am Coll Cardiol 43, 717–724, doi:10.1016/j.jacc.2003.08.061 (2004).
Gencer, B. et al. Eligibility for PCSK9 Inhibitors According to American College of Cardiology (ACC) and European Society of Cardiology/European Atherosclerosis Society (ESC/EAS) Guidelines After Acute Coronary Syndromes. J Am Heart Assoc 6, doi:10.1161/jaha.117.006537 (2017).
Sinnaeve, P. R. et al. Gene expression patterns in peripheral blood correlate with the extent of coronary artery disease. PLoS One 4, e7037, doi:10.1371/journal.pone.0007037 (2009).
Muse, E. D. et al. A Whole Blood Molecular Signature for Acute Myocardial Infarction. Sci Rep 7, 12268, doi:10.1038/s41598-017-12166-0 (2017).
Ntzani, E. E. & Ioannidis, J. P. Predictive ability of DNA microarrays for cancer outcomes and correlates: an empirical assessment. Lancet 362, 1439–1444, doi:10.1016/s0140-6736(03)14686-7 (2003).
Ein-Dor, L., Kela, I., Getz, G., Givol, D. & Domany, E. Outcome signature genes in breast cancer: is there a unique set? Bioinformatics 21, 171–178, doi:10.1093/bioinformatics/bth469 (2005).
Chen, R., Liu, X., Jin, S., Lin, J. & Liu, J. Machine Learning for Drug-Target Interaction Prediction. Molecules 23, doi:10.3390/molecules23092208 (2018).
Wei, S. et al. Gastric Cancer Tumor Microenvironment Characterization Reveals Stromal-Related Gene Signatures Associated With Macrophage Infiltration. Frontiers in genetics 11, 663–663, doi:10.3389/fgene.2020.00663 (2020).
Kumar, R. & Indrayan, A. Receiver operating characteristic (ROC) curve for medical researchers. Indian Pediatr 48, 277–287, doi:10.1007/s13312-011-0055-4 (2011).
Moreira, D. M. et al. Role of vascular inflammation in coronary artery disease: potential of anti-inflammatory drugs in the prevention of atherothrombosis. Inflammation and anti-inflammatory drugs in coronary artery disease. Am J Cardiovasc Drugs 15, 1–11, doi:10.1007/s40256-014-0094-z (2015).
Boateng, S. & Sanborn, T. Acute myocardial infarction. Dis Mon 59, 83–96, doi:10.1016/j.disamonth.2012.12.004 (2013).
Ferrières, J. et al. Total lipid management and cardiovascular disease in the dyslipidemia international study. Cardiology 125, 154–163, doi:10.1159/000348859 (2013).
Chapman, M. J. et al. Triglyceride-rich lipoproteins and high-density lipoprotein cholesterol in patients at high risk of cardiovascular disease: evidence and guidance for management. Eur Heart J 32, 1345–1361, doi:10.1093/eurheartj/ehr112 (2011).
Can, M. et al. Is direct method of low density lipoprotein cholesterol measurement appropriate for targeting lipid lowering therapy? Int J Cardiol 142, 105–107, doi:10.1016/j.ijcard.2008.11.141 (2010).
Zdravkovic, S., Wienke, A., Pedersen, N. L. & de Faire, U. Genetic susceptibility of myocardial infarction. Twin Res Hum Genet 10, 848–852, doi:10.1375/twin.10.6.848 (2007).
Zhao, Y. P., Wang, J. J., Li, X. Y., Peng, G. J. & Yang, Z. Extended least squares support vector machine with applications to fault diagnosis of aircraft engine. ISA Trans 97, 189–201, doi:10.1016/j.isatra.2019.08.036 (2020).
Wan, J. J., Chen, B. L., Kong, Y. X., Ma, X. G. & Yu, Y. T. An Early Intestinal Cancer Prediction Algorithm Based on Deep Belief Network. Sci Rep 9, 17418, doi:10.1038/s41598-019-54031-2 (2019).
Goldenberg, J. R., Wang, X. & Lewandowski, E. D. Acyl CoA synthetase-1 links facilitated long chain fatty acid uptake to intracellular metabolic trafficking differently in hearts of male versus female mice. J Mol Cell Cardiol 94, 1–9, doi:10.1016/j.yjmcc.2016.03.006 (2016).
Li, L. O. et al. Liver-specific loss of long chain acyl-CoA synthetase-1 decreases triacylglycerol synthesis and beta-oxidation and alters phospholipid fatty acid composition. J Biol Chem 284, 27816–27826, doi:10.1074/jbc.M109.022467 (2009).
Chiu, H. C. et al. A novel mouse model of lipotoxic cardiomyopathy. J Clin Invest 107, 813–822, doi:10.1172/jci10947 (2001).
Li, Y. et al. Targeting ACSL1 promotes cardiomyocyte proliferation and cardiac regeneration. Life Sci 294, 120371, doi:10.1016/j.lfs.2022.120371 (2022).
Yang, L. et al. High expression of long chain acyl-coenzyme A synthetase 1 in peripheral blood may be a molecular marker for assessing the risk of acute myocardial infarction. Exp Ther Med 14, 4065–4072, doi:10.3892/etm.2017.5091 (2017).
Wang, V. M. et al. Suggestive association between PLA2G12A single nucleotide polymorphism rs2285714 and response to anti-vascular endothelial growth factor therapy in patients with exudative age-related macular degeneration. Mol Vis 18, 2578–2585 (2012).
Nicolaou, A. et al. Quantitative trait locus mapping in mice identifies phospholipase Pla2g12a as novel atherosclerosis modifier. Atherosclerosis 265, 197–206, doi:10.1016/j.atherosclerosis.2017.08.030 (2017).
Tuong, Z. K. et al. RORα and 25-Hydroxycholesterol Crosstalk Regulates Lipid Droplet Homeostasis in Macrophages. PLoS One 11, e0147179, doi:10.1371/journal.pone.0147179 (2016).
Gold, E. S. et al. ATF3 protects against atherosclerosis by suppressing 25-hydroxycholesterol-induced lipid body formation. J Exp Med 209, 807–817, doi:10.1084/jem.20111202 (2012).
Li, Z. et al. Krüppel-Like Factor 4 Regulation of Cholesterol-25-Hydroxylase and Liver X Receptor Mitigates Atherosclerosis Susceptibility. Circulation 136, 1315–1330, doi:10.1161/circulationaha.117.027462 (2017).
Glunde, K., Penet, M. F., Jiang, L., Jacobs, M. A. & Bhujwalla, Z. M. Choline metabolism-based molecular diagnosis of cancer: an update. Expert Rev Mol Diagn 15, 735–747, doi:10.1586/14737159.2015.1039515 (2015).
Wahli, W. & Michalik, L. PPARs at the crossroads of lipid signaling and inflammation. Trends Endocrinol Metab 23, 351–363, doi:10.1016/j.tem.2012.05.001 (2012).
Michalik, L. et al. International Union of Pharmacology. LXI. Peroxisome proliferator-activated receptors. Pharmacol Rev 58, 726–741, doi:10.1124/pr.58.4.5 (2006).
Zhou, H. et al. Melatonin suppresses platelet activation and function against cardiac ischemia/reperfusion injury via PPARγ/FUNDC1/mitophagy pathways. J Pineal Res 63, doi:10.1111/jpi.12438 (2017).
Zheng, P. F. et al. Genes associated with inflammation may serve as biomarkers for the diagnosis of coronary artery disease and ischaemic stroke. Lipids Health Dis 19, 37, doi:10.1186/s12944-020-01217-7 (2020).
Cheng, M., An, S. & Li, J. Identifying key genes associated with acute myocardial infarction. Medicine (Baltimore) 96, e7741, doi:10.1097/md.0000000000007741 (2017).
Ji, Z. et al. Accuracy of triggering receptor expressed on myeloid cells 1 in diagnosis and prognosis of acute myocardial infarction: a prospective cohort study. PeerJ 9, e11655, doi:10.7717/peerj.11655 (2021).
Xia, C. et al. Total saponins from Paris forrestii (Takht) H. Li. show the anticancer and RNA expression regulating effects on prostate cancer cells. Biomed Pharmacother 121, 109674, doi:10.1016/j.biopha.2019.109674 (2020).
Huang, J., Liu, L., Qin, L., Huang, H. & Li, X. Weighted Gene Coexpression Network Analysis Uncovers Critical Genes and Pathways for Multiple Brain Regions in Parkinson's Disease. Biomed Res Int 2021, 6616434, doi:10.1155/2021/6616434 (2021).
Xu, M. et al. Elucidation of the Mechanisms and Molecular Targets of Sanhuang Xiexin Decoction for Type 2 Diabetes Mellitus Based on Network Pharmacology. Biomed Res Int 2020, 5848497, doi:10.1155/2020/5848497 (2020).
McEligot, A. J., Poynor, V., Sharma, R. & Panangadan, A. Logistic LASSO Regression for Dietary Intakes and Breast Cancer. Nutrients 12, doi:10.3390/nu12092652 (2020).
Pierre, J. F. et al. Discovery and predictive modeling of urine microbiome, metabolite and cytokine biomarkers in hospitalized patients with community acquired pneumonia. Sci Rep 10, 13418, doi:10.1038/s41598-020-70461-9 (2020).
Sundermann, B. et al. Support Vector Machine Analysis of Functional Magnetic Resonance Imaging of Interoception Does Not Reliably Predict Individual Outcomes of Cognitive Behavioral Therapy in Panic Disorder with Agoraphobia. Front Psychiatry 8, 99, doi:10.3389/fpsyt.2017.00099 (2017).
Thygesen, K. et al. Fourth Universal Definition of Myocardial Infarction (2018). Circulation 138, e618-e651, doi:10.1161/cir.0000000000000617 (2018).

No competing interests reported.

Identification and Validation of Lipid-related Genes in Acute Myocardial Infarction

Status:

Version 1

Abstract

Figures

Introduction

Results