ARTICLE pubs.acs.org/jpr
Metabolomics Study on the Toxicity of Aconite Root and Its Processed Products Using Ultraperformance Liquid-Chromatography/ Electrospray-Ionization Synapt High-Definition Mass Spectrometry Coupled with Pattern Recognition Approach and Ingenuity Pathways Analysis Xijun Wang,* Huiyu Wang, Aihua Zhang, Xin Lu, Hui Sun,* Hui Dong, and Ping Wang National TCM Key Lab of Serum Pharmacochemistry, Heilongjiang University of Chinese Medicine, and Key Pharmacometabolomics Platform of Chinese Medicines, Heping Road 24, Harbin 150040, China
bS Supporting Information ABSTRACT: The mother and lateral root of Aconitum carmichaelii Debx, named “Chuanwu” (CW) and “Fuzi”, respectively, has been used to relieve joint pain and treat rheumatic diseases for over 2000 years. However, it has a very narrow therapeutic range, and the toxicological risk of its usage remains very high. The traditional Chinese processing approach, Paozhi (detoxifying measure),can decompose poisonous Aconitum alkaloids into less or nontoxic derivatives and plays an important role in detoxification. The difference in metabolomic characters among the crude and processed preparations is still unclear, limited by the lack of sensitive and reliable biomarkers. Therefore, this paper was designed to investigate comprehensive metabolomic characters of the crude and its processed products by UPLC-Q-TOFHDMS combined with pattern recognition methods and ingenuity pathway analysis (IPA). The significant difference in metabolic profiles and changes of metabolite biomarkers of interest between the crude and processed preparations were well observed. The underlying regulations of Paozhi-perturbed metabolic pathways are discussed according to the identified metabolites, and four metabolic pathways are identified using IPA. The present study demonstrates that metabolomic analysis could greatly facilitate and provide useful information to further comprehensively understand the pharmacological activity and potential toxicity of processed Aconite roots in the clinic. KEYWORDS: metabolomics, UPLC-Q-TOF-HDMS, Aconitum carmichaelii Debx, biomarkers, metabolites, pattern recognition analysis
’ INTRODUCTION Traditional Chinese medicine (TCM) has gained increasing acceptance worldwide in recent years and is generally considered as being natural and harmless.1 3 Aconite root, an herbal medicine in TCM, has been popularly used in herbal medicines in Asia for several thousand years and is widely distributed across Asia and North America. It has a wide range of pharmacological effects, although it has very narrow therapeutic ranges, commonly applied for various diseases, i.e., rheumatic fever, painful joints, bronchial asthma, gastroenteritis, collapse, syncope, diarrhea, edema, various tumors, and some endocrinal disorders.4,5 In Chinese Pharmacopoeia (CP) 2010, Aconitum carmichaeli Debx (Figure 1A) is recorded, extensively distributed in Sichuan Province of China. The mother root of Aconitum carmichaelii Debx is named “Chuanwu” (CW, Figure 1B), while the daughter or lateral root of Aconitum carmichaelii Debx is “Fuzi” (FZ, Figure 1C). They were first recorded in the earliest Chinese medicinal classic, Shennong’s Materia Medica. However, many cases of accidental and intentional r 2011 American Chemical Society
intoxication with this plant have been reported; some of these have been fatal.6,7 It is thus important to develop a more efficient method to lowering the toxicity of Aconite root. As we know, TCM herbal processing approaches, namely, “Paozhi”, by means of the transformation of secondary plant metabolites help to reduce the toxicity of the drug and might exert maximal therapeutic efficacy with minimal adverse effects.8 As a traditional processing form, Paozhi could remarkably reduce the toxicity of the CW and Fuzi by decomposing the diester diterpene alkaloids (DDAs) to the relatively less toxic monoester diterpene alkaloids (MDAs).5,9 11 According to the ways of further treated-processing form of CW and Fuzi, they are converted into Yanfuzi (YFZ, Figure 1E) and Zhichanwu (ZC, Figure 1F) preparations, whose forms are recorded in CP 2010. There are more than 20 commonly used proprietary herbal formulas from Received: September 24, 2011 Published: November 04, 2011 1284
dx.doi.org/10.1021/pr200963e | J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research
ARTICLE
Figure 1. (A) Aconitum carmichaeli Debx. (AB) Aconitum kusnezoffii Reichb. (C) The mother root of Aconitum carmichaelii Debx. (D) The daughter or lateral root of Aconitum carmichaelii Debx. (E) Zhichanwu. (F) Yanfuzi.
both historical literature and modern clinical reports containing processed medicine that include the processed YFZ or ZC as a main ingredient, such as “Yin Chen Si Ni Tang”, “Fuzi Lizhong Wan”, “Guifulizhong Wan”, “Jinguishenqi Wan”, and “Shenfu Injection”.12,13 The main indications of these products are muscular disorders, joint pain, and arthritis. Paozhi is an effective detoxifying measure to remove the poisonous Aconitum alkaloids. Several studies have shown that metabolomics have been used in the evaluation of differences in toxicological and pharmacological actions of aconite products when applying different TCM processing methods.14,15 However, the difference in metabolomic characters between the crude and their processed products is still unclear, restricting the further application of YFZ and ZC in the clinic. A clear understanding of the mechanism of Paozhi is essential to evaluate drug safety. However, the research is still limited to content changes of several alkaloids to determine its efficacy/ toxicity, while ignoring the tuber and its processed products as a complex group of plant metabolites.16,17 Moreover, little is known about the global change of the metabolome between crude and
processed medicines. Traditional markers of conventional clinical chemistry and histopathology methods are not regionspecific and only increase significantly after substantial disease injury. Therefore, more sensitive and reliable markers of disease are needed. The ideal biomarkers will identify disease early, resulting in increased safer drugs. Metabolite changes have been observed in diseased individuals as a primary indicator.18 Metabolomics is a comprehensive method for metabolite assessment that involves measuring the overall metabolic signature of biological samples.19 Metabolomics not only enables the parallel assessment of the levels of a broad range of metabolites but also has a great impact in investigation of physiological status, diagnosing diseases, discovering biomarkers, and identifying perturbed pathways due to disease or drug treatment.20 23 The small-molecule metabolites have an important role in biological systems and represent attractive candidates to understand toxicity phenotypes.24 They represent a diverse group of low-molecular-weight structures including lipids, amino acids, peptides, nucleic acids, and organic acids, vitamins, thiols, carbohydrates. Particularly, highly sensitive and specific biomarkers in biofluids are relatively 1285
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research more useful for the detection of toxicity.25,26 These biomarkers, which afford detection earlier than the traditional clinical chemistry and histopathology methods, could facilitate and improve the development of detection of toxicity, thus benefiting the public health. With the development of new analytical techniques, metabolomics could provide comprehensive and detailed evidence for further study on the efficacy/toxicity of processed medicine.27 LC MS profiling is an important approach for the identification and quantification of metabolites from complex biological samples. When MS is coupled with UPLC, high sensitivity, high resolution, wide dynamic range, coverage of a wide chemical diversity, robustness, and feasibility to elucidate the molecular weight and structure of unknown compounds can be achieved for the low abundance metabolites. UPLC MS has been used for observing the subtle metabolic changes in some diseases or treatment of diseases and has provided informative data for elucidating the biochemical basis of diseases and addressing the therapeutic effect of medicines.28,29 It is believed that metabolomics analysis will play an important role as an effective tool in terms of high-throughput elucidation of metabolic phenotypes of the crude and processed preparation. Thus, in this paper, a metabonomic approach based on the UPLC-HDMS technique with pattern recognition approach and IPA was employed to demonstrate the plasma metabolic characteristics, speculate on alternate metabolic pathways, and compare the metabolic difference between the crude and processed products.
2. EXPERIMENTAL SECTION 2.1. Chemicals and Reagents
Acetonitrile (HPLC grade) was purchased from Merck (Darmstadt, Germany). Distilled water was purchased from Watson’s Food & Beverage Co., Ltd. (Guangzhou, China), and formic acid (HPLC grade) was purchased from the Beijing Reagent Company (Beijing, China). Leucine enkephalin was supplied by Sigma-Aldrich (St. Louis, MO, USA). Other chemicals, except as noted, were analytical grade. The assay kits for alanine aminotransferase (ALT), aspartate amino transferase (ALT), creatine kinase (CK), lactate dehydrogenase (LDH), and triglyceride (TG) were obtained from the Biosino Biotechnology and Science Inc. (Beijing, China). The assay kits for NaK-ATP enzyme (Na-K-ATP) were purchased from the Nanjing Jiancheng Biotech Company (Nanjing, China). CW and Fuzi were collected in July 2009 from the same field of mianyang county in Sichuan province (South, China). They were authenticated by Prof. Xijun Wang, Department of Pharmacognosy of Heilongjiang University of Chinese Medicine (Harbin, China). The procedure for processed ZC and YFZ preparation was according to CP 2010. 2.2. Sample Preparation
The crude CW, Fuzi, and processed products (300 g) were immersed in deionized water two times (3 and 2.4 L each time) for 1 h and then decocted by boiling for 1.5 h. The extracted solution was collected by filtering through 8 layers of gauze, and finally the decoction was freeze-dried into a powder. 2.3. Animal Handling
Male Wistar rats (weighting 200 ( 20 g) were supplied by the GLP Center of Heilongjiang University of Chinese Medicine (Harbin, China). The room temperature was regulated at 25 ( 1 °C with 50 ( 5% humidity. A 12-h light/dark cycle was set, with free access to standard diet and water. All rats were randomly
ARTICLE
divided into 9 groups of 16 rats each as follows: control group, low-dose CW 1-month (LC-1M), low-dose Fuzi 1-month (LF1M), low-dose ZC 1-month (LZ-1M), low-dose YFZ 1-month (LY-1M), high-dose CW 6-month (HC-6M), high-dose Fuzi 6-month (HF-6M), high-dose ZC 6-month (HZ-6M), and highdose YFZ 6-month (HY-6M). All animals were allowed to acclimatize in metabolism cages for 1 week prior to treatment. The crude drug and processed products were dissolved in distilled water, and the dosages were 0.027 and 100 g/d and 0.108 and 100 g/d, respectively, orally dosed by gavage. The control group were orally administrated with an equivalent volume of distilled water. All experiments were performed in accordance with the approved animal protocols and guidelines established by Medicine Ethics Review Committee for animal experiments of Heilongjiang University of Chinese Medicine. 2.4. Animal Sample Collection and Handling
On the last day, rats were deeply anesthetized and then sacrificed. Blood was collected from the abdominal aorta, and plasma and serum were separated via centrifugation at 6000 rpm for 20 min at 4 °C. The plasma samples were collected and stored at 80 °C flash frozen in liquid nitrogen until metabolomics analyses were performed; the serum was used for biochemical assay according to the manufacturer’s instructions of commercial kits. The activities of AST, ALT, CK, and Na-K-ATP and the levels of LDH and TG were determined by using commercially available kits. All procedures completely complied with the manufacturer's guidelines. The fresh liver, spleen, lungs, and kidney were obtained and rapidly put into 10% formalin solution for histopathological examination. The heart (200 mg) was washed by cold normal saline immediately, and then 1 mm3 of myocardium was taken from the left ventricle to be used for electron microscopic examination. For histological evaluation, rats were deeply anesthetized with 10% chloral hydrate, and the heart was infused with physiological saline. After perfusion fixation with paraformaldehyde, the heart was removed, stored in fixative for 24 h, embedded in paraffin wax, and stained with hematoxylin and eosin. The histopathology analysis was done by the affiliated hospital of Heilongjiang University of Chinese Medicine. Fresh myocardial tissues (2 mm3) were excised from the myocardial tissues. Tissues were fixed with 3% gluteraldehyde and postfixed with 1% osmium tetroxide. The specimens were processed for ultrathin sections. The sections were stained with uranium acetate and lead citrate, and the ultrastructural changes of myocardial tissue were then evaluated by a transmission electron microscope. 2.5. Chromatography and Mass Spectrometry Conditions
Chromatographic analysis was performed in a Waters ACQUITY UPLC system controlled with Masslynx (V4.1, Waters Corporation, Milford, USA). An aliquot of 4 μL of sample solution was injected onto an ACQUITY UPLC HSS T3 C18 column (50 mm 2.1 mm, 1.8 μm, Waters Corporation, Milford, USA) held at 45 °C, and the flow rate was 0.5 mL/min. The optimal mobile phase consisted of a linear gradient system of (A) 0.1% formic acid in water and (B) 0.1% formic acid in acetonitrile: 0 3.0 min, B 1 52%; 3.0 6.2 min, B 52 74%; 6.2 7.7 min, B 74 80%; 7.7 8.2 min, B 80 90%; 8.2 9.5 min, B 90 100%; 9.5 12.0 min, B 100%; 12.0 13.0 min, B 100 1%; 13.0 16.0 min, B 1%. In addition, 10 plasma samples were randomly selected from each group and mixed together as the “quality control” (QC) sample. The QC sample was used to optimize the condition of UPLC-Q-TOF-HDMS, as it contained the most information of whole plasma samples. Every day, after the 1286
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research instrument was calibrated, the QC sample was first analyzed to test the stability of the instrument, making sure the instrument was in the same condition during the whole analytical procedure. The UPLC was directly interfaced with a Waters TOF mass spectrometer equipped with a dual electrospray ionization probe operating in positive mode (ESI+). Mass spectrometry detection was performed using a Waters Micromass Q-TOF micro Synapt High Definition Mass Spectrometer (Waters, Milford, USA) equipped with electrospray ionization in positive mode. The optimal conditions of analysis were as follows: the source temperature was set at 120 °C, desolvation gas temperature was 300 °C, cone gas flow was 50 L/h, desolvation gas flow was 650 L/h, capillary voltage was 2.5 kV, sampling cone voltage was 50.0 V, extraction cone voltage was 4.0 V. A lock mass calibrant of leucine-enkephalin (0.2 ng/mL) in water/acetonitrile (50:50, v/v) was continuously introduced in the mass spectrometer via the second ESI probe (Lock-Spray) at a flow rate of 100 μL/min, generating a reference ion for positive ion mode ([M + H]+ = 556.2771) to ensure accuracy during the MS analysis. Data were acquired between m/z 50 and 1000 Da with a 0.3 s scan time and a 0.1 s interscan delay and processed further in MassLynx 4.1 software (Waters). 2.6. Pattern Recognition Analysis and Data Processing
The mass data acquired were imported to Markerlynx within Masslynx software (version 4.1) for peak detection and alignment. The retention time and m/z data for each peak were determined by the software. The parameters of Markerlynx method were set as follows: mass tolerance 0.1 Da; noise elimination level 5; full scan mode was employed in the mass range of 100 1000 amu; the initial and final retention times were set as 0 and 12 min for data collection. All data were normalized to the summed total ion intensity per chromatogram, and the resultant data matrices were introduced to EZinfo 2.0 software for principal component analysis (PCA), partial least-squaresdiscriminant analysis (PLS-DA) and orthogonal projection to latent structures (OPLS) analysis. Prior to PCA, all variables obtained from LC MS data sets were mean-centered and scaled to Pareto variance. PCA is an unsupervised multivariate statistical approach. It is used for variable reduction and separation into classes. To maximize class discrimination and biomarkers, the data were further analyzed using the OPLS-DA method. S-plots were calculated to visualize the relationship between covariance and correlation within the OPLS-DA results. Variables that had significant contributions to discrimination between groups were considered as potential biomarkers and subjected to further identification of the molecular formula. Metabolite peaks were assigned by MS/MS analysis or interpreted with available biochemical databases, such as HMDB, http://www.hmdb.ca/; KEGG, http://www.genome.jp/kegg/; METLIN, http://metlin.scripps.edu/; and MassBank, http://www.massbank.jp/. IPA was performed with MetPA, which is a web-based tool for pathway analysis and visualization of metabolomics. Other statistical analyses used include one-way analysis of variance (ANOVA), least significant difference (LSD) test, and independent sample t test. They were performed with SPSS 17.0. Statistical differences are considered significant when the test p value is less than 0.05.
3. RESULTS AND DISCUSSION 3.1. Biochemical Analysis
To evaluate the toxicity effects of CW, the levels of Na+-K+ATP, CK, and LDH in myocardial homogenate and ALP, AST,
ARTICLE
ALT, and TG in serum were compared with those of control rats. The biochemistry parameters of the control group, the crude CW, Fuzi, and processing groups of low-dose and high-dose at 1 and 6 months are summarized in the Supporting Information (Tables S1 and S2), respectively. At the 1 month, the biochemical results observed in the applied TCM processing groups did not show significant difference compared to the control group (data not shown). The Na-K-ATP activity in the HC-1M group were obviously decreased (p < 0.05), while the levels of CK, LDH, and AST were increased, compared with those of control rats. However, comparing with the crude group, Na-K-ATP activity was increased in the LZ-1M group; the LDH and AST values were decreased in the high-dose processed CW group. So metabolic profiling in the following experiments only paid close attention to the plasma samples of high dose group rats. The Na-K-ATP activity in the Fuzi group were markedly decreased (p < 0.05), while the levels of CK, LDH, and AST were increased, compared with those of control rats. However, Na-K-ATP activity were increased and the LDH and AST values were decreased in the processed Fuzi group, compared with the crude group. These results suggested that heart and liver tissues from the processed group exhibited weak toxicity, while crude group showed serious toxicity at 6 months. The reduction of Na-K-ATP activity, which is often associated with energy metabolism, indicated the energy metabolization had been obstructed. The content of AST and ALT directly reflects the level of liver damage, and the increase of CK and LDH activity can be attributed to myocardial injury. 3.2. Histopathological Observations
Hematoxylin eosin staining was used to survey the histopathological changes after processed treatment. Histopathology of tissue (i.e., heart, liver, spleen, lungs, kidney) exposed to CW, Fuzi, and processed products was examined to further investigate the toxicity. Interestingly, only the heart from the HC-6M group produced more severe pathological damage; the other tissues had no variation (Figure 2). As demonstrated in Figure 2, histopathological examination of cardiac muscle tissue for the control, HC-6M, and HF-6M groups clearly depicted myonecrotic areas. Inflammatory cells and confluent areas were present in the cardiac muscle tissue from the crude CW and Fuzi groups, comparing with control group. Moreover, edema and rupture of striated muscle were also clearly observed. Furthermore, the pathological results of liver tissue showed that the structures of the hepatic lobule were unclear and the liver cells were developing fatty degeneration, empty lipocytes; balloon degeneration of cells was also shown in liver cells by ecletron microscop at 6 months. Additionally, hyperaemia, hepatic cord disorder, granular degeneration, and focal necrosis were clearly observed. Consequently, combined with the clinical chemistry results, we concluded that the presence of substantial heart damage after administration of crude CW and Fuzi could be confirmed at the sixth month. 3.3. Changes in Myocardial Ultrastructure
The ultrastructures of surrounding areas of the myocardial tissues from different groups were examined. Figure 3 presents representative electron micrographs of the myocardial tissues. The endothelium of capillaries in the control and processed groups at 6 months was preserved well, and both the capillary and surrounding tissue displayed normal ultrastructural features. The crude group led to an apparent swelling of the endothelium with a large number of caveolaes in endothelial cells and prominent edema in the vascular surroundings. In the control and processed 1287
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research
ARTICLE
Figure 2. Representative light photomicrographs of cardiac tissue sections stained with H&E; 200 magnification.
Figure 3. Typical transmission electron micrographs from male Wistar rats at 6 months following exposure to vehicle and various Acontium, Fuzi, and their processed products. (15,000 magnification).
groups, cardiac myofibrils stood regularly arranged with wellpreserved myofilaments, and mitochondria occupied the cytoplasm between myofibrils with densely packed cristae. The crude drugs challenge provoked a dramatic injury in cardiac muscle cells, as indicated by disrupted myofibrils and ruptured mitochondria. Therefore, the processing method attenuated the ultrastructural alterations induced by the crude groups.
3.4. Metabolomic Study
3.4.1. LC MS Analysis of Metabolic Profiling. Global metabolic profiling in both positive and negative ion modes were analyzed by UPLC-HDMS, and 4455 ions including ESI+ and ESI ions were obtained. Using the optimal UPLC-HDMS condition described above, the typical total ion current (TIC) chromatograms for the control, CW, and Fuzi samples in positive 1288
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research
ARTICLE
Figure 4. Typical TIC chromatogram of (A) control, (B) CW, and (C) Fuzi. (D) Chemical structure and mass fragment information of LPC (16:0) in positive ESI mode.
ion mode are presented in Figure 4A, B, and C, respectively. As Figure 4 reports, the low molecular mass metabolites could be separated well in 8 min. In order to better visualize the subtle similarities and differences among these complex data sets, multiple pattern recognition methods were employed to phenotype the urine metabonome of rats. Here, PCA, PLS, and OPLSDA were used to classify the metabolic phenotypes and identify the differentiating metabolites. In the PCA scores, each point represents an individual sample. Groupings, trends, and outliers can also be found. Another unsupervised multivariate analysis method, the PCA model, provides an overview of all observations or samples in a data set. The PCA results are displayed as score
plots indicating the scatter of the samples, which indicate similar metabolomic compositions when clustered together and compositionally different metabolomes when dispersed. The PCA scores plot could divide the different urine samples into different blocks, respectively, suggesting that the metabolic profiles have changed as a result of administration. The dispersion degree of urine samples of CW or Fuzi group was more significant than that of ZC group (i.e., Figures 5A, 5B, 6A, 7A, and 7B). Therefore, we conclude that the two groups had altered the urine metabolic profiles of normal rats, with the strongest perturbations occurring at the CW or Fuzi group. The supervised PLS-DA analysis reveal a much great metabolic signature difference of the urine 1289
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research
ARTICLE
Figure 5. Multivariate statistical analysis of LC MS metabolite profiles derived from various Acontium, Fuzi, and their processed products. (A) 3D scores plot of UPLC-HDMS spectra of HC-6M vs control group. PCA score plot represents separate clustering of MS profiles corresponding to HC-6M vs control group. (B) 2D scores plot of UPLC-HDMS spectra of plasma, HC-6M vs control group. (C) Corresponding loading plot of HC-1M vs HC6M. Corresponding loading plots of all samples show MS peaks that differ among samples. (D) OPLS-DA model results for HC-1M vs HC-6M group. (E) S-plot of OPLS-DA model for HC-1M vs HC-6M group. (F) Variable important plot (VIP) for HC-1M vs HC-6M group.
from CW- or Fuzi-treated groups, compared to the control group (i.e., Figure 5C). It is worth noting that this change is in a timeand dose-dependent manner. It can be concluded that Fuzi made the rats metabolite profiles deviate from the normal states, and farther extent in CW, suggesting metabolism turbulence and significant pathobiological changes. The mass spectrometry signals
responsible for this differentiation are characterized by a variable importance plot (VIP) from PLS-DA analysis (i.e., Figures 5E, 6C, 7D). As shown in Figure 5A, control and CW are appreciably separated in the 3D-PCA score plot. Although the PCA model provided an overview of all observations or samples, the details of differences in each cluster remained unclear. The supervised 1290
dx.doi.org/10.1021/pr200963e |J. Proteome Res. 2012, 11, 1284–1301
Journal of Proteome Research
ARTICLE
Figure 6. (A) PCA score plot derived from four representative HC-1M, HZ-1M, HC-6M, and HZ-6M groups. (B) VIP plot of HC-1M vs HK-1M. (C) Loading S-plot of HC-1M vs HK-1M. (D) Relative mean height intensity of different metabolites of HC-1M and control; 0 control, 9 HC-1M. (E) Comparison of biomarker intensity between HC-6M vs HC-1M; 0 HC-1M, 9 HC-6M. (F) Comparison of biomarker intensity between HC-6M vs HZ-1M; 0 HZ-1M, 9 HC-6M.
method, OPLS-DA, was then used to isolate the variables responsible for differences among the various group. The OPLS-DA score plots are shown in Figure 5D and Figure 7C and E. These results indicated that the OPLS-DA models were reliable. To highlight the metabolite differentials between the control and CW groups, feature selections were performed using OPLS-DA. The farther away from the origin, the higher the value of the ions in the VIP score plot. Potential markers were extracted from S-plots constructed following the OPLS analysis, and markers were chosen on the basis of their contribution to the variation and correlation within the data set. The OPLS-DA loading S-plot, a plot of the covariance versus the correlation in conjunction with the variable trend plots, allows easier visualization of the data. The variables that changed most significantly are plotted at the top or bottom of the S plot, and those that do not vary significantly are plotted in the middle. OPLS-DA S-plots for HC-1M vs HC-6M, HC-1M vs control, HF-1M vs HF-6M, and Fuzi vs YFZ are shown in Figures 5E, 6C, 7D and 7F, respectively.
3.5. Identification of the Endogenous Metabolites
All collected plasma samples were analyzed, and low molecular weight metabolites were represented as the chromatographic peaks in the total ion current (TIC) chromatograms. The robust UPLC-HDMS segregation analysis that was performed using the aforementioned protocol indicated the retention time and precise molecular mass and provided the MS/MS data that was necessary for the structural identification of the biomarkers. The precise molecular mass was determined within a reasonable degree of measurement error (
