Untargeted metabolite profiling of serum in rats exposed to pyrraline

Abstract

Pyrraline, one of advanced glycation end-products, is formed in advanced Maillard reactions. It was reported that the presence of pyrraline was tested to be associated with nephropathy and diabetes. Pyrraline might result in potential health risks because many modern diets are heat processed. In the study, an integrated metabolomics by ultra-high-performance liquid chromatography with mass spectrometry was used to evaluate the effects of pyrraline on metabolism in rats. Thirty-two metabolites were identified as differential metabolites. Linolenic acid metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, arachidonic acid metabolism, tyrosine metabolism and glycerophospholipid metabolism were the main perturbed networks in this pathological process. Differential metabolites and metabolic pathways we found give new insights into studying the toxic molecular mechanisms of pyrraline.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-023-01256-7.

Introduction

Advanced glycation end-products (AGEs) are formed in advanced Maillard reactions, including pyrraline, pentosidine, carboxymethyllysine, and others (Liu et al., 2017). AGEs play a crucial role in process of aging, chronic kidney function harm, diabetes mellitus, atherosclerosis, and Alzheimer’s disease (Li and Yu, 2018). It has been indicated that the level of pyrraline in serum of people with diabetes is higher than that of people with good health (Van Nguyen, 2006). It was reported that pyrraline was shown to be associated with renal failure and diabetes (Aso et al., 2004). As many modern foods are heat processed, they likely have high levels AGEs, which including pyrraline. It might result in potential health risks because of the daily diet (Sharma et al., 2015).

A global metabolomics study, which aims to find some changes in metabolites and pathways, is a powerful tool to study metabolic state of an organism (Chang et al., 2021; Holmes et al., 2008). Analysis of the metabolome has displayed remarkable potential in finding disordered metabolites and pathways, which caused by the endogenous or exogenous disturbance. It gives global insights on the hidden metabolic changes in organism, which can be used to understand living mechanisms better (Hu et al., 2014, 2019; Johnson et al., 2015). Mass spectrometry (MS) and nuclear magnetic resonance (NMR) are two methods to identity and quantify metabolites on a global scale (Goodman et al., 2020; He et al., 2022; Pareek et al., 2020). NMR metabolomics allows quick detection, small sample amounts and gives structural information (Li et al., 2020). However, the sensitivity of NMR is lower than MS.

To this day, there is no study on serum of rats exposed to pyrraline by metabolomics. In the study, integrated metabolomics by ultra-high-performance liquid chromatography with mass spectrometry (UHPLC-MS) was used to evaluate effects of pyrraline on metabolism in rats. The robust global metabolomics method was utilized to find more differential metabolites to comprehensively uncover underlying toxic molecular mechanisms of pyrraline.

Materials and methods

Chemicals and reagents

All standard substances (purity > 98%) for compound identification and 4-chloro-dl-phenylalanine were purchased from Sigma-Aldrich or Sigma (St. Louis, MO, USA). HPLC-grade methanol and formic acid were gotten from Tedia (Fairfield, OH, USA). Pyrraline (purity > 99%) was purchased from J&K scientific. Ultra high purity water was generated by a Milli-Q water system (Millipore Corp., Billerica, MA, USA). Other chemicals were all analytical grade.

Animals and treatments

Thirty male 5- to 6-week-old SD rats were bought from the Laboratory Animal Center of Peking University Health Science Center (Beijing, China). Animals were acclimated for one week before the experiment. They were kept in an environmentally controlled breeding room (temperature 22 ± 1 ℃, humidity 60 ± 5%, and 12-h dark/light cycle). They were offered with standard laboratory food and water. All experimental processes were conducted according to the European Community guidelines for using experimental animals. The study procedure was agreed by the Animal Care and Use Committee of Peking University Health Science Center. All rats were randomly assigned to three groups: high dose group (n = 10, HD) and low dose group (n = 10, LD) were injected 4.5 mg/kg and 0.5 mg/kg of pyrraline dissolved in normal saline each day by tail vein respectively, and the control group (n = 10, ND) received normal saline by the same method. All rats were fed and treated for three days. Body weight was noted every day. At the end of experiment, animals abstained from food for 12 h and then were sacrificed by collecting blood from eye veniplex. Blood was centrifuged at the speed of 3000×g for 10 min at 4 ℃. The supernatant was stored at − 80 ℃ until metabolomics analysis.

Sample processing

Serum samples were thawed at 4 ℃ before analysis. Added eight times methanol to serum and centrifuged at 13,000 rpm for 15 min at 4 ℃ to precipitate the protein. The supernatant was gathered and dried with nitrogen stream. The dried residue was redissolved in solvent, including 80% acetonitrile, and 20% water, and centrifuged at 14,000 rpm, 4 ℃ for 15 min for removing any particulates. After repeated centrifugation three times, the total supernatant was used for metabolomics analysis (Jiang et al., 2020). The quality control (QC) sample was mixed 20 μL of solution from each serum sample.

UHPLC-QTOF-MS analysis

Samples of serum were analyzed by UHPLC-QTOF-MS. UHPLC was performed on an Agilent 1290 UHPLC system (Agilent Corporation) equipped with an ACQUITY UHPLC BEH C18 column (2.1 × 100 mm,1.7 μm, Waters Corporation). The injection volume was 1 μL. The flow was 0.2 mL/min and the column was set at 40 ℃. The optimal mobile phase was made up of 0.1% formic acid in water (A) and methanol (B). The gradient elution program was: 0–5 min, 2–20%B; 5–8 min, 20–75%B; 8–18 min, 75–85%B; 18–33 min, 85–98%B; 33–34 min, 98%B; 34–35 min, 98–2%B. Mass spectrometry (MS) was performed on an Agilent 6530 QTOF-MS with an electrospray ionization (ESI) source in positive and negative modes. The ESI source conditions were as follows: capillary voltage, 3.5 kV (ESI+) and 3.0 kV (ESI−); cone voltage, 60 V; and ion source temperature, 350 ℃. Original data were collected between m/z 50 and 1200 in both modes (Li et al., 2017). QC samples were analyzed in every six samples throughout the whole analytical process to evaluate the stability of the system (Zeng et al., 2021).

Metabolomics data analysis and identification of differential metabolites

Sample data was extracted by Mass Hunter Profinder software (Agilent, USA) for analysis and comparison. Data were normalized and then entered into SIMCA 14.1 (Umetrics, Sweden) for multivariate data analysis. According to variable importance in the projection (VIP) values > 1.0 and p values < 0.05 gained from the Mann–Whitney U test, some differential metabolites were found. Metaboanalyst 5.0 was used to construct pathway. Differential metabolites in serum samples were identified according to accurate data of m/z values, retention time, and MS/MS fragments. All the accurate m/z values were searched in HMDB (http://www.hmdb.ca/) and metlin (https://metlin.scripps.edu/) databases (Zhao et al., 2020). The identified metabolites were verified with authentic standards whenever commercial standard substances were available.

Receiver operator characteristic (ROC) analysis

The ROC analysis was used for assessing the diagnostic ability of the differential metabolites to classify rats into a low or high pyrraline exposure. The area under ROC curve (AUC) from 0.5 to 1.0 showed diagnostic accuracy from no discrimination to good classification (Li et al., 2017). The ROC analysis was completed by SPSS 19.0 software.

Results and discussion

Metabolomics study and metabolite identification

Metabolites in serum samples were profiled by UHPLC-QTOF-MS. Score Plots generated by PCA and PLS-DA presented clear discrimination among them, as shown in Fig. 1.

Fig. 1.

PCA score plots of serum samples in the control group and pyrraline-exposed groups at different dosages by UHPLC-MS methods in positive (A) (R² = 77.9%, Q² = 61.3%) and negative (B) (R² = 83.4%, Q² = 73%) mode. PLS-DA score plots of serum samples in the control group and pyrraline exposed groups at different dosages by UHPLC-MS methods in positive (C) (R²X = 77.7%, R²Y = 87.7%, Q² = 85.5%) and negative (D) (R²X = 86.8%, R²Y = 77.3%, Q² = 67.9%) mode. The 200-permutation test demonstrated no overfitting in the PLS-DA model ((E) positive mode (R² = (0.0, 0.162), Q² = (0.0, − 0.349) (F) negative mode (R² = (0.0, 0.124), Q.² = (0.0, − 0.319))

The PCA score plots showed that the metabolic phenotypes of HD, LD and ND groups were different. As shown in Fig. 1, both positive and negative ion modes of three groups were completely separated in PLS-DA score plots. In the Fig. 1(A), we can find that HD group generally shifts to the left, and the LD group were between HD group and ND group. It indicated that the toxicity of exposing to pyrraline in rats showed a gradual aggravation with an increase of the dose. As shown in the Fig. 1(C) and (D), samples of quality control (QC) were tightly located in the PLS-DA model. It showed that the equipment was stable and analysis was reliable. It verified the repetitiveness of UHPLC-MS for metabolomics analysis. Quality of PLS-DA model was usually estimated by R²X, R²Y and Q². R²X and R²Y reveal the explanatory ability of the model, and Q² reveals the predictive power of the model. Figure 1 displayed good predictive ability of the two models. Validation plots in Fig. 1(E) and (F) were performed using a 200 times permutation test. The results indicated that the model was not over-fitted.

Based on the differential metabolites chosen from VIP values and p values in PLS-DA model, box plots (Fig. 2) were used to describe differences of metabolites among ND, LD and HD groups. The remarkable and dose-dependent decrease in levels of tyrosine was found. Compared with ND group, concentrations of LysoPC (18:3) and PS (22:1) in HD and LD groups increased. The importance of abnormal plasma phospholipid in kidney harm has been proved (Aksenov et al., 2017). In addition, contents of myristic acid were higher in HD group than those in ND and LD groups, showing dose-dependent changes.

Fig. 2.

Box plot of differential metabolites in the study groups. (A) LysoPC (18:3); (B) Tyrosine; (C) Myristic acid; (D) PS (22:1)

Metabolites with VIP values > 1.0 and p value < 0.05 were regarded as differential metabolites. A total of 32 differential metabolites was identified. Detailed information of identified metabolites, including accurate m/z and formula, is shown in Table 1.

Table 1.

Differential metabolites detected in serum of rats from different groups by UHPLC-QTOF-MS

Metabolite	m/z	Formula	FC (HD/N D)	VIP	p-value	FDR
Tyrosine	182.0814	C9H11NO3	0.53	1.31	6.46 × 10–3	2.91 × 10–2
Thymine	127.0368	C5H6N2O2	1.87	1.27	7.11 × 10–3	2.13 × 10–2
3-Indolepropionic acid	190.0858	C11H11NO2	0.23	1.30	4.67 × 10–6	8.41 × 10–5
Tetracosahexaenoic acid	357.2788	C24H36O2	3.04	1.17	2.83 × 10–2	3.91 × 10–2
γ-CEHC	279.1591	C16H22O4	0.64	1.27	4.11 × 10–2	4.93 × 10–2
Eicosapentae noic acid	303.2318	C20H30O2	2.47	1.32	2.51 × 10–2	3.77 × 10–2
LysoPC (18:2)	520.3398	C26H50NO7 P	6.23	1.41	1.74 × 10–3	1.57 × 10–2
LysoPC (20:4)	544.3398	C28H50NO7 P	3.43	1.41	2.50 × 10–2	4.49 × 10–2
LysoPC (16:0)	496.3398	C24H50NO7 P	2.18	1.24	3.03 × 10–2	4.96 × 10–2
LysoPC (18:3)	518.3241	C26H48NO7 P	2.09	1.40	7.11 × 10–3	2.56 × 10–2
LysoPC (18:1)	522.3554	C26H52NO7 P	2.87	1.35	2.43 × 10–2	4.85 × 10–2
LysoPC (17:0)	510.3554	C25H52NO7 P	2.16	1.25	4.57 × 10–2	4.84 × 10–2
LysoPC (18:0)	524.3711	C26H54NO7 P	1.67	1.27	1.53 × 10–2	3.93 × 10–2
LysoPC (20:3)	546.3554	C28H52NO7 P	1.57	1.25	4.84 × 10–2	4.84 × 10–2
Dodecanoic acid	336.2897	C21H34O2	3.68	1.37	1.81 × 10–2	4.07 × 10–2
7-Ketochol esterol	401.3414	C27H44O2	3.21	1.43	3.68 × 10–3	2.21 × 10–2
3-Deoxyvit amin	369.3516	C27H44	1.90	1.36	4.53 × 10–2	5.10 × 10–2
Octadecanamide	360.3261	C24H41NO	1.59	1.32	2.86 × 10–2	3.68 × 10–2
Cholic acid	407.2803	C24H40O5	4.40	1.18	7.96 × 10–5	1.11 × 10–3
5-HETE	319.2279	C20H32O3	2.39	1.29	2.65 × 10–3	9.28 × 10–3
Myristic acid	227.2017	C14H28O2	2.82	1.21	9.15 × 10–3	1.42 × 10–2
Linolenic acid	277.2173	C18H30O2	7.11	1.29	6.20 × 10–4	2.89 × 10–3
Hexadecenoic acid	253.2173	C16H30O2	4.03	1.23	5.43 × 10–4	3.80 × 10–3
Arachidonic acid	303.2330	C20H32O2	2.70	1.28	7.45 × 10–3	1.74 × 10–2
PS (22:1)	578.3463	C28H54NO9 P	1.67	1.21	7.84 × 10–3	1.57 × 10–2
2-Hexyldecanoic acid	255.2329	C16H32O2	1.67	1.18	2.23 × 10–2	2.40 × 10–2
PE (19:0)	554.3463	C24H50NO7 P	2.02	1.04	3.13 × 10–2	3.13 × 10–2
11-Octadecenoic acid	281.2486	C18H34O2	3.31	1.29	1.57 × 10–2	1.84 × 10–2
PS (22:0)	580.3620	C28H56NO9 P	1.81	1.22	1.51 × 10–2	1.92 × 10–2
Stearic acid	283.2643	C18H36O2	1.69	1.25	6.83 × 10–3	1.91 × 10–2
PC (18:0)	582.3776	C26H54NO7 P	4.40	1.19	8.73 × 10–3	1.53 × 10–2
MG (18:0)	417.3222	C21H42O4	0.52	1.14	1.18 × 10–2	1.65 × 10–2

FC fold change, γ-CEHC γ-carboxymethylhydroxychroman, 5-HETE 5-hydroxyeicosatetraenoic acid, PE phosphatidylethanolamine, PS phosphatidylserine, PC phosphatidylcholine, MG methylglyoxal, LysoPC lysophosphatidylcholine

ROC curve

Differential metabolites were subjected to ROC analysis. The ROC value is proportional to the dependability of the diagnostic effect. ROC analysis results with the MS quantitative data of thirty-two metabolites are shown in Table S1, of which metabolites exhibited good diagnostic (AUC > 0.7). These differential metabolites can distinguish between normal diet rats and relatively high exposure rats with pyrraline exposure.

Altered pathways related to pyrraline exposure levels

In order to visualize the disturbed metabolic pathways in rats fed with pyrraline, metabolomics pathway analysis was made with MetaboAnalyst 5.0 which was a powerful method to discover the significantly related pathways (Hu et al., 2021). Summary of pathway study with MetaboAnalyst 5.0 was in the Fig. 3 and listed in Table S2. In this study, the impact-value threshold gotten from pathway topology analysis was 0.10. Metabolic pathways in which impact-values were more than 0.1 were linolenic acid metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, arachidonic acid metabolism, tyrosine metabolism, glycerophospholipid metabolism. These target pathways showed marked perturbations in rats exposed to pyrraline. The metabolic pathway with the largest impact value is linolenic acid metabolism, as shown in Fig. S1. KEGG database was used to understand high level functions and utilities of biological system. Figure 4 was a general overview of the metabolites and metabolic pathway changed by pyrraline intervention according to the KEGG.

Fig. 3.

Disturbed pathways in response to pyrraline treatment (1) Linolenic acid metabolism. (2) Phenylalanine, tyrosine and tryptophan biosynthesis. (3) Arachidonic acid metabolism. (4) Tyrosine metabolism. (5) Glycerophospholipid metabolism

Fig. 4.

Potential metabolic pathways disturbed in rats induced by pyrraline. “↑”and “↓” indicate changes in compound levels compared to the normal group

Analysis of the metabolic function in serum of all rats

The remarkable decrease in tyrosine concentration may be related to the degradation of protein into amino acids to influence biological functions, including cell cycle progression and inflammatory responses (Drogan et al., 2015; Liang et al., 2016). 3-indolepropionic acid is formed by deamination of tryptophan. In the study, levels of 3-indolepropionic acid significantly decreased in HD group compared with those in ND group. It indicated that exposing to pyrraline can affect the amino acid metabolism, which is consistent with previous study (Wikoff et al., 2009). Hepatotoxicity is related to the enhancement of the endogenous oxygen free radical. 3-indolepropionic acid can protect the cells from oxidative stress and harm. Therefore, decrease in 3-indolepropionic acid level indicates that exposing to pyrraline disturbs the liver metabolism of rats (Wellner et al., 2011).

Levels of octadecanamide were to grow in the pyrraline-exposed group. Fatty acid amides are included in adjustment of different physiological and pathological process, such as inflammation (RoyChoudhury et al., 2016). Octadecanamide is a substrate of fatty acid amide hydrolase (FAAH). Hydrolysis relative rate is affected by the activity of FAAH. Octadecanamide could be a specific biological marker compound for diagnosis between alcohol and non-alcohol caused liver damage (Zhao et al., 2017). Fatty acids of ω-3 and ω-6 are polyunsaturated fatty acids (Guo et al., 2017). Eicosapentaenoic acid (EPA), linolenic acid and arachidonic acid are primary members of ω-6 and ω-3 fatty acids. ω-3 fatty acids activate lipopolysaccharide-stimulated TNF in macrophages by NF-B inhibition, and inflammatory mediators, by competing for metabolic enzymes with arachidonic acid. Arachidonic acid is regarded as the main source of proinflammatory lipid mediators. It is quickly changed into potent inflammatory mediators including leukotrienes, lipoxins and HETEs. Arachidonic acid pathway is linked to some complications of the metabolic syndrome, such as vascular disease (Kwon et al., 2016). Levels of ω-3 and ω-6 fatty acids were improved in pyrraline-exposed group. It is a signal of disturbed lipid metabolism. Fatty acids derivatives are involved in some biological processes, such as inflammation, differentiation and cellular proliferation. It was found that level of 5-HETE was elevated after the pyrraline exposure. 5-HETE gets involved in the formation of leukotrienes and other proinflammatory mediators (Hao and Breyer, 2007). In addition to a necessary structural component of cell membranes, cholesterol acts as a precursor for many important compounds including steroid hormones, oxysterols, and bile acids. Oxysterols are some cholesterol oxidation products. They can be obtained endogenously by autooxidation or by enzymatic reactions, and can be given by food. Oxysterols can pass lipophilic membranes. As a result, if their formation in organisms or their intake with dietary animal fat are too much, oxysterols are helpful for pathogenesis of some diseases (Poli et al., 2013). Oxidation of cholesterol sterol ring is almost finished by the nonenzymatic pathways, resulting in a generation of compounds like 7-ketocholesterol. 7-ketocholesterol has been showed to cause a distinct inflammatory phenotype in endothelial cells and the foam cell production. Thus, these results heavily display a disorder of cholesterol metabolism because of exposing to pyrraline.

In pyrraline-exposed groups, contents of phospholipids including PS (22:1), PE (19:0), PS (22:0), and PC (18:0) were elevated. Phospholipids are lipid compounds including phosphate in their molecular structures. Almost all biological cells have phospholipids. Phospholipids can be classified as glycerol phospholipids and sphingolipids. The glycerol phospholipids include phosphatidylcholine (PC), phosphatidyl inositol (PI), phospholipid acyl ethanol amine (PE), phosphatidylserine (PS), cardiolipin (DG), phosphatidylglycerol (PG), and acetal phospholipids. Phospholipids are irregularly located on the cell membrane. PS is the only one phospholipid that can adjust the function of key proteins in cell membranes. The main function of PS in signal transduction is to transfer the protein kinase C (PKC) to cell membrane by the specific connection of PS and PKC (Zhao et al., 2015). Moreover, PS can influence the activity of other enzymes in signal transduction. Therefore, it was revealed that because of exposing to pyrraline, the related protease activity is declined, signal transduction is barred, and phospholipid metabolism is disrupted. Phosphatidylcholines are structural constituents of cell membranes in animals. It has been reported that phosphatidylcholines are biological marker compounds for chronic glomerulonephritis and diabetic nephropathy. Earlier studies have indicated the prominent role of plasma phosphatidylcholine disorders in renal damage (Quansah et al., 2015), but few studies have been done on plasma phospholipids, such as lysophosphatidylcholine, which might be important factors in pathogenesis of kidney harms. Lysophosphatidylcholines are metabolites of phosphatidylcholines. Level of lysophosphatidylcholine can be a clinical diagnostic marker that shows changes of pathophysiology. Lysophosphatidylcholines have close ties with oxidative stress and immune inflammation (Igarashi and Kashiwagi, 2011). Levels of lysophosphatidylcholines in serum of rats exposed to pyrraline revealed remarkable difference. When body is suffering from some bad conditions, reactive nitrogen species and reactive oxygen species (ROS) are generated unusually, and the rate of production are faster than that of removing oxides from body. The balance between oxidation and antioxidant system is broken, which eventually leads to tissue harm. Unsaturated lysophosphatidylcholines are mainly generated from the hydrolysis of liver phosphatidylcholine by phospholipase A2, while saturated lysophosphatidylcholines is often transferring sn-2 fatty acid from phosphatidylcholine by lecithin cholesterol acyltransferase (Sas et al., 2015). This study revealed that up-regulated lysoPC (16:0), lysoPC (17:0), lysoPC (18:0), lysoPC (18:1), lysoPC (18:2), lysoPC (18:3) and lysoPE (22:4) were obviously found in rats exposed to pyrraline. Clearly, rats suffered high oxidative stress because of decrease in antioxidant defenses and increase in prooxidant factors. When oxidative stress happened, the production of free radical could activate the phospholipase A2, which can hydrolyse phosphatidylcholine to generate lysophosphatidylcholine. The increased lysophosphatidylcholine could generate excessive ROS which results in the imbalance of oxidation-antioxidation. Moreover, it has been reported that lysophosphatidylcholine is included in proinflammatory process of acute damage or chronic diseases (Bereketoglu and Pradhan, 2019).

As we know, this is the first time that UHPLC-QTOF-MS nontargeted metabolomics analysis has been done for study on serum of rats exposed to pyrraline. Thirty-two metabolites were identified as differential metabolites. Linolenic acid metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, arachidonic acid metabolism, tyrosine metabolism and glycerophospholipid metabolism were the main perturbed networks in this pathological process. The study highlights the evidence for the linkage of pyrraline exposure and adverse health risks. Differential metabolites and metabolic pathways we found give new insights into studying the toxic molecular mechanisms of pyrraline. The potential health risk associated with the exposure to pyrraline, should not be ignored.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

CH: Investigation, Methodology, Writing-original draft, Conceptualization. JW: Methodology, Formal analysis, Visualization. FQ: Conceptualization. YL: Conceptualization. FZ: Conceptualization. JW: Conceptualization, Supervision, Funding acquisition. BS: Conceptualization. All authors read and approved the final manuscript.

Declarations

Conflict of interest

The authors declare that they have no competing interests.

Ethical approval

The animal care and all experimental processes were conducted according to the European Community guidelines for using experimental animals. The study protocol was agreed by the Animal Care and Use Committee of Peking University Health Science Center. This article does not contain any studies with human participants performed by any of the author.