| BackgroundChronic heart failure (CHF), as a complicated clinical syndrome, takes place at the serious stage of cardiovascular diseases due to various causes. In the21st century, the CHF morbidity rate has increased with the aging population and improved medical treatments, and has become one of the most important public health problems in China.CHF is a progressive disease. People that have previously developed CHF would continue to develop CHF through ventricular remodeling, even after the removal of conditions that caused CHF in the first place. Although the heart failure treatment strategies have been significantly improved in the past10+years, its prognosis remains poor, with5-year mortality rate similar to that of cancer. As a result, study on the pathogenesis of congestive heart failure to explore new therapeutic target has become a main research direction in congestive heart failure. In this field, myocardial metabolic remodeling in recent years has caught significant attentions. Metabolic remodeling by definition is loss of order in myocardial substrate use and energy metabolism that happens in heart failure. Studies have shown that in the development of heart failure, changes in energy metabolism and substrate utilization often occur in cardial muscles. These metabolic changes reflect the increased myocardial energy demand, as well as the compromised capacity of ATP synthesis and utilization.A large number of clinical and experimental evidence has shown that the blocked substrate utilization and the deficiency of energy substances may lead to myocardial remodeling and the progressing of chronic heart failure, and conversely, metabolic intervention can improve cardiac function in heart failure. How to detect myocardial energy metabolism is a key issue that metabolic treatment needs to solve. The current main methods for the monitoring of myocardial metabolism are imaging methods such as PET, MRI etc, but these methods are difficult to implement in clinical settings due to their complexity and high cost. We have, in the early stage of our research, established a method to evaluate the myocardial energy consumption (MEE) by echocardiography. In this study, we set out to detect serum metabolic spectrum in patients with heart failure by metabolomics technique, to analyze the correlation of serum metabolic spectrum with the MEE obtained via echocardiography, and therefore, to study the feasibility of evaluating myocardial metabolic remodeling in heart failure via serum metabolomics. After obtaining these results, we also intend to select a serum metabolic biomarker that is simple to use and objectively represents MME, and therefore, establish a practical, simple, economical and effective serology method for MEE monitoring. We also hope that the studies on the mechanism of the energy metabolism in heart failure, via these metabolic biomarkers, will provide clues in discovering new therapeutic targets in heart failure. This study is comprised of three parts:Chapter Ⅰ1H-NMR metabolomics in chronic heart failureMetabolism is a fundamental life activity and provides the material basis that all living organisms cannot live without. When the body is affected by exogenous factors (such as drugs, poisons, etc) or endogenous factors (metabolic disturbance, pathophysiological changes), metabolic changes occur in cells, tissues and even the whole body, causing changes in the components and concentration of body fluid metabolites. Metabolomics studies the holistic and dynamic change of metabolites, establishes appropriate classification models and extracts biomarker clusters that related to myocardial metabolism remodeling, and therefore, are useful in reaching the correct diagnosis and understanding of the disease development. In this chapter, we analyze the overall features and complete information during heart failure, by applying metabolomics methods and studying and the function of metabolites in the serum with time, in order to study the feasibility of using these methods in chronic heart failure. MATERIALS AND METHODS:This study included61participants with an average age of61.5years.43were males and18were females. The normal control group had15participants (n=15). The heart failure group had46participants(n=46), which included11patients of NYHA class grade Ⅱ,18patients of grade Ⅲ,17patients of grade IV. All participants were required to fast8hours prior to blood sampling, and to have free of tea, coffee and other beverage a week before. The blood sample collected was naturally coagulated for1hr at room temperature. From which, the serum sample was then collected and centrifuged at3000g for10mins at4℃. The upper serum was collected and store at-80℃.The detection instrument used in this study was a Varian (Varian, Inc.) INOVA600MHz superconducting NMR spectrometer equipped with pulsed field gradient with the gradient field of three-resonance probe. Prior to testing, the serum sample was thawed at room temperature, and300μl of it was then transferred to a centrifuge tube.100μl TSP weight aqueous solution (1mg/ml) and200μl heavy water (D2O) were then added into the centrifuge tube and were fully mixed with the serum sample. The mixture was centrifuged at13,000g for10mins.550μl supernatant was taken and transferred into a5mm NMR tube, and stood at room temperature till it became homogeneous and transparent. Then, the treated serum sample was analyzed using the relaxation editing pulse sequence (CPMG) and diffusion editing pulse sequence (LEDs) respectively. NMR spectra were obtained by the Fourier transform of the raw data followed by phase adjust and baseline correction. CPMG data, within the range of δ0.4～4.4, was integrated every0.01ppm segment; LED data, within the range of δ0～6.0, was integrated every0.04ppm segment, with region of δ4.6～5.0 excluded from integration. The integral normalized to the total integral intensity of each spectrum. The data was output, converted into Excel files and saved. SIMCA-P+software (v11.5, Umetrics, Umea, Sweden) was used for multivariate statistical analysis. Prior to analysis, the data was averaged, centered and automatically scaled. We first used principal component analysis (PCA) to develop the discriminant model, followed by data processing through orthogonal signal correction (OSC). The subsequence data was then treated with regression analysis by using partial least squares method. Results:1. Typical serum’H-NMR patterns displayed significant difference in control group and heart failure group, including chemical shifts peak area δ0.84～0.90ppm,δ1.26～1.30ppm,δ1.32～1.33ppm,δ4.09～4.10ppm, δ2.22ppm δ3.40～4.00ppm,δ1.19～1.20ppm and δ1.46～1.48ppm. These peak assignments indicated that the concentration of serum lipoproteins and the relative content of the glycosylated compounds in the heart failure group are higher than that in the control group, while the concentration of organic acid content of alanine was lower compared with the control group;2. PCA analysis showed that the difference in serum metabolomics between the control group and severe heart failure group was the most significant, while difference in the other two groups was difficult to distinguish;3. Using OSC-PLS method greatly improved the data resolution among these three groups.4. Through cross-validation and variable exchange methods, the classification model we designed was verified to have more than85%probability in accuracy and predictability, and therefore, can be used for further analysis.5. Three methods (vaiable importance in projection [VIP], loading weights and correlation coefficients) were used to screen for serum metabolites markers, and their results were basically consistent with each other. Conclusion:1.The differences in serum metabolomics between the normal control group and severe heart failure group were significant; While the differences in serum metabolomics between mild heart failure group and normal group were not obvious;2. Combination of orthogonal signal correction PLS method and1H-NMR technology efficiently identified the differences among the serum metabolic groups and significantly improved the data resolution;3. The high accuracy and predictability of the OPLS classification model is demonstrated by a variety of authentication methods.4. Metabolic markers selected using different screening methods are consistent.Chapter Ⅱ The use of metabolomics methods to evaluate myocardial metabolic remodeling and screen serum markersWe have proved in the first chapter that metabolomics in the serum of patients with severe heart failure has undergone significant changes compared to the normal person. In this chapter, we continued to study serum metabolic in different myocardial energy consumption levels using1H-NMR metabolomics method. Based on this study, we intend to screen for biomarkers that can be used as indicators for myocardial energy consumption. Identification of these biomarkers will provide important diagnostic tools for the objective quantitative assessment of myocardial metabolic remodeling, and thus will provide clues for the mechanistic study of metabolism in heart failure. MATERIALS AND METHODS:the subjects and serum metabolomics methods are the same as those described in the first chapter. The myocardial energy expenditure (MEE) in participants was measured by Echocardiography ultrasound method that has been established in our previous study, and was used as the categorical variable in the OPLS model. Results:1. PCA showed that the serum metablomics in high-MEE group was significantly different from those in low-MEE group;2. PLS model built for MEE achieved the best separation after orthogonal signal correction with more than90%accuracy and predictability, indicating the model is stable and reliable;3.18metabolites, which have a (variable importance priority) VIP value greater than1, were selected as the candidates of biomarkers. Using oneway-ANONA statistical method, metabolites that showed no obvious differences in concentration among the three participant groups were excluded, and8remaining metabolites (3-hydroxybutyrate, acetone, succinate, valine alanine, glutamine, methionine, creatine) were selected as serum metabolic biomarkers;4. Among the8selected metabolites, acetone,3-hydroxybutyrate, and succinate were positively correlated with MEE; while alanine and methionine were negatively correlated with MEE. The concentration of Acetone,3-hydroxybutyrate, succinate, and alanine, methionine in patients was not affected by ACEI and other drugs. The analysis of traditional heart failure risk factors indicated that Acetone,3-hydroxybutyrate and succinate were positively correlated to brain natriuretic peptide, creatinine and heart rate (P<0.01, P<0.01, P<0.05); while alanine and methionine acid were negatively correlated to brain natriuretic peptide (BNP) and heart rate (P<0.01, P<0.01). Methionine was negatively correlated to glucose (P <0.05). Age. body mass index (BMI), serum total cholesterol, triglycerides and uric acid didn’t have detectable correlation to the metabolites that mentioned above; the analysis of the receiver operating characteristic curve (ROC) indicated that AUC of brain natriuretic peptide, acetone,3-hydroxybutyrate, succinate was0.98,0.92,0.90and0.86respectively. Diagnosis of chronic heart failure using acetone as the biomarker achieved89.13%sensitivity and80%specificity, while3-hydroxybutyrate as a biomarker achieved89.13%sensitivity and93.33%specificity. Conclusion: Serum metabolites compositions are clearly different among various MEE groups, most significantly between the low-MEE group and the high-MEE group, and can be efficiently distinguished by serum metabolomics.8metabolites were selected from screenings as the biomarkers of myocardial energy metabolism. According to their corresponding characteristics, these metabolites were categorized into3groups:1. metabolites that were positively correlated to myocardial energy consumption, including3-hydroxybutyrate, acetone and succinate;2. Mebabolites that were negatively correlated to myocardial energy consumption, including valine, alanine, glutamine, and Methionine;3. Metabolites that were positively correlated to myocardial energy consumption in moderate MEE but negatively correlated to myocardial energy consumption in high MEE, including creatine; AUC of both acetone and3-hydroxybutyrate have no statistical differences from AUC of brain natriuretic peptide, suggesting that acetone and3-hydroxybutyrate can be used as small molecule biomarkers to diagnose chronic heart failure. Chapter Ⅲ Metabolomics studies based on GC/MS for the serum of heart failure rat modelsData collected from clinical patients with chronic heart failure may be perturbed by difference in individuals and subjected to influence by activities such as diet and emotion. To provide additional experimental evidence to our results, we studied the metabolomics differences in rats between the heart failure groups and the control group in this chapter, by using GC/MS-based techniques and coronary artery ligation to simulate ischemic heart failure.MATERIALS AND METHODS:51SPF SD male rats with body weight of200-300g were selected as the subjects. Among them,35were randomly selected for sham operation, with the other16for coronary ligation. Chronic ischemic heart failure model was prepared by coronary artery ligation. After four weeks of heart failure establishment, rats were operated under anesthesia. Blood sample was quickly taken from the inferior vena cava. The sample was kept at room temperature for10minutes before it was centrifuged to extract plasma for serum metabolomics analysis. The metabolomics instrument used in this study was Agilent7890A GC/5975C MS system, and the capillary column was the Agilent J&W Scientific HP-5MS (30m×250um id). To analyze the serum sample,120μl chromatography grade methanol was added in the sample tube. The sample tube was then centrifuged.120μl supernatant was taken into a high recovery rate vial and was dried under mild nitrogen flow.80μl of15mg/ml Methoxamine pyridine solution and80μl of BSTFA reagent (containing1%trimethylsilyl chloride (TMSCl)) were added into this vial. After60minutes of reaction at70℃, the reaction mixture was analyzed by GC/MS. In this study, Mass spectrometry detection was processed at full scan with a range of50-600(m/z). The raw data was extracted from the R software platform and was imported to TagFinder software for spectrum fragments retention time correction, peak alignment and deconvolution analysis. Finally, it was post-edited in EXCEL software. Simca-P software (version11.5) was applied for data principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and orthogonal partial least squares discriminant analysis (OPLS-DA). Results:1. In the Sham operation group,14out of16rats survived, with a survival rate of87.5%. In the Coronary artery ligation group,18out of35rats survived, with a survival rate of51.4%. These results indicated that the left ventricular ejection fraction (LVEF) of surgery group was significantly less than the sham group(47%vs76%, P=0.001).2. PCA score plot showed that the difference between severe heart failure group with the control group or the mild heart failure group was significant, while the difference between mild heart failure group and control group was not obvious.3. A stable and reliable OPLS-DA model can be established, with44metabolites selected from the metabolomics study of different heart failure groups. Conclusions:1. The serum metabolomics of severe heart failure group changed significantly compared to that of the control group, while the serum metabolomics of mild heart failure group had no detectable changes compared to that of the control group. These results were consistent to our study on heart failure patients.2.44metabolites were selected from the metabolomics study of different heart failure groups, which indicated that1H-NMR and GC/MS can be complementary in the metabolomics evaluations of heart failure.3. The main different metabolites between the severe heart failure and control groups were free fatty acids, lactic acid, glucose, branched-chain amino acids and other substances. The concentration of free fatty acids, glucose, lactic acid was significantly increased in severe heart failure group, which is consistent to the data of human. On contrary, the change in concentration of the branched chain amino acids was different from human. These discoveries provided valuable information and clues for future metabolism treatment in heart failure. |
PDF Full Text Request