Metabolomics reveals the renoprotective effect of n-butanol extract and amygdalin extract from Amygdalus mongolica in rats with renal fibrosis

Abstract

Renal fibrosis (RF) is a pathological process of progression from chronic kidney disease to end-stage renal disease. Amygdalus mongolica is a traditional Chinese medicine, and our previous studies demonstrated that the n-butanol extract (BUT) and amygdalin extract (AMY) from its seeds can prevent RF. However, the underlying mechanism remains unclear. The present study investigated the exact mechanism of the protective effect of A. mongolica on RF. A renal fibrosis rat model was induced with unilateral ureteral obstruction. Biochemical indicators were measured and combined with histopathology of renal tissue to evaluate the anti-RF effects. A serum metabonomic method was used to clarify the changes in the metabolic profile. The tubulointerstitial damage and fibrosis were significantly improved and metabolic perturbations were restored after treatment with BUT and AMY. Thirty-eight metabolites associated with RF progression and related to the regulation of arginine and proline metabolism, nicotinate and nicotinamide metabolism, and histidine metabolism were identified. They were restored to levels similar to those in controls after treatment. Moreover, no significant differences in efficacy were observed between the BUT and AMY groups. This study reveals and compares the potential mechanisms of the renoprotective effects after treatment with BUT and AMY from a metabolomic perspective.

Keywords:

• Serum metabolomics
• renal fibrosis
• Amygdalus mongolica
• protective effect
• n-butanol extract
• amygdalin extract

Introduction

Renal fibrosis (RF) is a pathological injury process that can lead to continuous deterioration in various chronic renal diseases until the gradual loss of kidney function, due to wound infection, inflammation, blood circulation obstacles, and various pathogenic factors, that stimulates the immune response. RF progressively induces cell damage, develops late in life, results in marked collagen deposition and accumulation, and causes sclerosis of the renal parenchyma and scarring until the kidneys completely lose visceral function. Moreover, RF causes cellular hardening and induces abnormal extracellular matrix (ECM) deposition [1]. It has been reported that the regulation of non-coding RNAs and multiple signalling pathways play a central role in the onset and progression of renal fibrosis, such as TGF-β signalling pathways, miR-21, etc. [2]. However, there is still a lack of effective therapeutic drugs for renal fibrosis in the clinic. Inhibition of the renin-angiotensin-aldosterone-system (RAAS) is one of the therapeutic strategies to slow the progression of renal fibrosis. The major drugs currently utilized for RAAS blockade primarily include angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin II receptor inhibitors (ARB) [3].

Amygdalus mongolica (Maximowicz) Ricker from the Rosaceae family is a special medicinal material in Inner Mongolia. The seeds of A. mongolica are used as a traditional herbal medicine [4,5]. The amygdalin monomer was obtained from the n-butanol extract (BUT) of A. mongolica. Amygdalin has many bioactivities, including anti-tumor, analgesia, and cough relief properties [6]. In a previous study by our group, we were the first to isolate and identify amygdalin from A. mongolica [7]. We identified amygdalin as the main bioactive compound and found that it accounted for 47.72% of the BUT contents. The molecular formula of amygdalin is C₂₀H₂₇NO₁₁. Because it is an organic compound containing nitrogen atoms, amygdalin is also an alkaloid compound [8]. We detected the amygdalin content in A. mongolica from different producing areas in China [9]. In addition to amygdalin, total flavonoids, alkaloids, and quercetin were found in BUT [10]. In a previous study, we demonstrated that A. mongolica extract has protective effects on liver fibrosis induced by carbon tetrachloride in rats [5]. Moreover, the n-butanol extracts of A. mongolica have blood lipid-lowering properties and produced a good protective effect on pulmonary fibrosis induced by bleomycin in rats [7,11]. Emodin and amygdalin are the main biologically active components in A. mongolica [8]. Other studies have shown that amygdalin exhibits antifibrogenic effects in the liver and kidney [5,8,12–15]. However, because of the complex interactions among the active ingredients, little is known about the exact mechanism by which A. mongolica prevents RF. Therefore, we focussed on the effects of the amygdalin monomer and the n-butanol extract in the process of RF. The present study aimed to observe the therapeutic effect and mechanism of A. mongolica extract on rats with RF and to provide a scientific basis for finding effective drugs against kidney fibrosis.

New methods are urgently needed to elucidate these complex mechanisms to identify effective anti-RF drugs and to comprehensively evaluate the systematic clinical efficacy of A. mongolica. Metabolomics uses animal fluids, tissues, and cells as research objects and monitors the dynamic changes in metabolite composition, quality, and quantity induced by physiological, pathological, or drug stimulation. Metabolomics analyzes the whole spectrum of low molecular weight compounds, rather than focussing on individual metabolites. Moreover, metabolomics approaches produce a generic framework that provides information on the functional integrity of the entire organism in a stage of a disease or after a given intervention [16]. Therefore, metabolomics, as an analytical platform, can be used to discover novel biomarkers involved in different disease processes [17] and to evaluate drug efficacy [18] and safety [19,20]. Among metabolomics analysis tools, UPLC-QTOF-MS has become useful because of its good separation efficiency and detection sensitivity [13]. We used this method to explore the possible antifibrotic mechanisms of A. mongolica. The experiment was carried out according to the flow design shown in Figure 1.

Figure 1. Scheme of the study. BUT: n-butanol extract; AMY: amygdalin extract; CON group: control group; MOD group: model group.

Materials and methods

Chemicals

Amygdalus mongolica seeds were obtained from Alashan in Inner Mongolia. Professor Songli Shi of Inner Mongolia University of Science and Technology Medical College of Baotou identified the dried mature seeds. Carboxymethylcellulose sodium and pentobarbital sodium were purchased from Tianjin Kaitong Chemical Reagent Co., Ltd., and Merck Germany Ltd., respectively. The purity of these chemicals (analytical grade) was above 99.5% (Tianjin, China). An MDA kit, superoxide dismutase (SOD) kit, hydroxyproline (HYP) kit, quantitative determination kit for total protein, and Masson staining kit were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). HPLC-grade methanol, acetonitrile, and formic acid were purchased from Honeywell (USA).

Preparation and analysis of A. mongolica extracts

Amygdalus mongolica seeds were peeled to remove the seed shells and crushed. Solutions of 95 and 70% ethanol were used as the extraction solvents. The conditions of the extraction were as follows: temperature, 70 °C; solid-liquid ratio, 1:10; and time, 2 h. The combined extracts were concentrated under pressure to obtain a brown ethanol extract. The ethanol extract was partitioned with water and three organic solvents with different polarities [petroleum ether (PE) < ethyl acetate (EA) < n-butanol (BU)] to obtain the PE-extracted solution, EA-extracted solution, and BU-extracted solution, respectively. After concentration and drying, the BU extract (BUT) was obtained. The isolation and identification of amygdalin from BUT of A. mongolica are described in previous studies [7]. Amygdalin was derived from BUT and accounted for 47.72% of the BUT content.

Analysis of but and isolation and identification of amygdalin

TLC spray reagents containing bismuth potassium iodide indicated the presence of alkaloids. The BUT of A. mongolica was added to a D101 macroporous resin column by the wet method and eluted with 10, 30, 50, 70, and 90% ethanol. The flow component was detected by thin-layer silica gel chromatography. The developing solvent was carbon tetrachloride:methanol (4:1). The 10 and 30% ethanol eluents were added to an ODS column by the wet method and eluted with methanol solution. The developing solvent was ethyl acetate:formic acid:water (20:3:1). According to the TLC results, the liquid components with the same chromatographic characteristics were combined, the solvent was recovered, and a colourless powder was obtained. The powder was then identified by UPLC-DAD-ESI-MS and nuclear magnetic resonance analysis. Chromatographic conditions: The chromatographic column was an Agilent ZORBAX Eclipse Plus C₁₈ (2.1 × 50 mm, particle size, 1.8 μm). Mobile phase A was 0.1% formic acid solution, mobile phase B was methanol, and the flow rate was 0.4 ml/min. The elution gradient was as follows: 0–3 min, 10–30% B; 4–5 min, 30–100% B; and 5–6 min, 100% B. With D₂O as the solvent, the NMR spectra of the isolated compounds were obtained at a frequency of 500 MHz The results showed that the compound was amygdalin (Figure S1).

Animals and sample collection

The animal experiments were approved by the Ethical Committee of Baotou Medical College (Approval Number: 20190314). Male Sprague-Dawley rats were purchased from Animal Experiment Centre of Peking University Health Science Centre (animal licence number: SCXK 2017-0005). The rats underwent adaptive feeding for a week, with free food and water intake. The SD rats were randomly divided into the control (CON) group (n = 8), model (MOD) group (n = 8), BUT group (n = 8), and amygdalin extract (AMY) group (n = 8). Rats in the MOD, BUT, and AMY groups underwent unilateral ureteral obstruction (UUO) to establish a model of RF. Unilateral ureteral ligation is widely used in animal models for pharmacology studies. Ligating the renal drainage system in UUO can increase urinary pressure, decrease renal blood flow, and promote venous blockage, macrophage infiltration, and fibre cell proliferation; thus, UUO gradually develops into renal interstitial fibrosis [21]. The UUO methods are described in the Supplementary Information. Only the left ureter was separated but not ligated, and the abdominal cavity was closed in the sham operation group. The BUT and AMY groups were gastrically administered BUT (1.25 g/kg/d) and AMY (0.02 g/kg/d), respectively, for 3 weeks. An equal volume (4 ml/kg/d) of saline was administered to the rats in the CON and MOD groups. The general condition of the rats was observed, the number of deaths was recorded, and rats were allowed to drink and eat freely. The laboratory temperature was 21–27 °C and the relative humidity was 30–50%.

Blood was obtained from the abdominal aorta of rats after anaesthetization with pentobarbital (35 mg/kg of body weight) (Merck Ltd., Germany). Serum was obtained from blood samples by centrifugation and stored at −80 °C. Tissues from the left kidney were fixed with 10% paraformaldehyde, and pathological changes were observed in paraffin sections.

Histological changes

The paraffin-embedded kidney tissues were cut into 5-µm sections, and haematoxylin & eosin (H&E) and Masson staining were performed to observe pathological changes [22]. The fibrosis injury score was determined according to the description by Zhi-Hao Zhang [23,24] in ten randomly selected non-overlapping fields per rat.

Biochemical indicator measurements

Blood was centrifuged at 3000 rpm for 15 min, and serum was removed and detected by an automatic biochemical instrument. Biochemical parameters were measured using an Olympus AU640 automatic analyzer (Olympus, Japan) [24].

Serum sample preparation and UPLC-QTOF-MS assay

All serum samples were thawed in an ice water bath and vortex-mixed before analysis. The analysis of six random serum samples in each group by UPLC-QTOF-MS is detailed in the Supplementary Information [20].

Serum metabolomics study and data analysis

The blood metabolic profiles of the rats (six rats were randomly selected from each group) were imported into the MarkerLynx software (v4.1) for data reduction and to acquire mass spectrometry matrix information. The EZinfo software (3.0) module was then used to conduct principal component analysis (PCA) with the data in each group, and the differences in the metabolic profiles of each group were identified.

To more effectively identify differences in metabolites, the blood metabolic profile of each group of rats was further analyzed by partial least square discriminant analysis (PLS-DA) [18], and a VIP plot was obtained that showed the metabolic differences among the groups. P-values were obtained from the VIP plot and two-tailed Student's t-test. HMDB (http://www.hmdb.ca/), KEGG (https://www.kegg.jp/kegg/pathway.html), MetPA (https://www.metaboanalyst.ca/) and other data platforms were used to retrieve potential biomarkers [20,25]. The data were analyzed and visualized using R Version 3.4.1 and packages (pheatmap_1.0.12, corrplot_0.84, ggplot2_3.3.3).

Results

Histological changes

Micrographs of renal tissue with H&E staining and Masson staining showed that the renal tubules in the CON group remained intact and that the interstitium did not significantly change (Figure 2(a,e)). In the MOD group, inflammatory cells infiltrated, renal tubules atrophied, and interstitial fibrosis was observed. Masson staining showed vacuolar changes in renal tubular epithelial cells and fibrotic changes (Figure 2(b,f)). Inflammation and fibrosis were improved in the MOD group after the administration of AMY and BUT (Figure 2(c,d,g,h)). The histopathological score showed the extent of damage to the tissues (Figure 2i). The results showed that BUT and AMY from A. mongolica had specific protective effects on RF in rats.

Figure 2. Pathological renal sections and determination of biochemical indicators in the CON, MOD, AMY, and BUT groups. (a–d) Renal tissues were stained using H&E staining (×100). (e–h) Renal tissues were stained using Masson staining (×100). Arrows indicate the lesion sites. (i) The histopathological score results of Masson staining. (j) Determination of BUN, Scr, ALB, MDA, SOD, and HYP. All values are presented as the mean ± SD. Statistical significance was calculated with ANOVA. *p < .05 and **p < .01 compared to the MOD group. #p < .05 and ##p < .01 compared to the CON group. The degree of severity was assigned the following scores according to the extent of injured cortex areas: (0) normal cortex; (1) <25% of the cortex area was injured; (2) 26–50% of the cortex area was injured; (3) 51–75% of the cortex area was injured; and (4) >75% of the cortex area was injured. CON: control group; MOD: model group; BUT: n-butanol extract group; AMY: amygdalin extract group.

Determination of clinical biomarkers of renal function

Biochemical markers associated with renal function in clinical settings, including blood diabetic nitrogen (BUN), serum creatinine (Scr), HYP (routine markers of fibrosis), and antioxidant index, were detected in the MOD, CON, BUT, and AMY groups. The levels of Scr, HYP, albumin (ALB), BUN, and malondialdehyde (MDA) were increased in the model group compared with the CON group (p < .05 or p < .01). The level of SOD was decreased (p < .05 or p < .01) in the MOD group compared with the CON group (Figure 2j and Table S1). After treatment with BUT and AMY, biochemical indicators related to renal function tended to be restored to normal levels. The anti-RF effects of A. mongolica were shown through the changes in the biochemical markers above.

Comparison of different groups using PCA and PLS-DA

The metabolic profiles of serum samples from different groups were determined by UPLC-QTOF-MS in positive and negative electrospray ionization (ESI) mode. According to the PCA results, the CON and MOD groups were divided into two areas, as shown in the PCA score plot (Figure 3(a,b)). This result suggested that the metabolic statuses of the CON and MOD groups were significantly different. Furthermore, the AMY and BUT groups were clearly separated from the MOD group. PLS-DA was used to maximize differences in the metabolic profiles (Figure 3(e,f)), which were verified by class permutation tests (Figure 3(g,h)). The R² and Q² values were 0.82 and 0.28, respectively, in positive mode and 0.79 and 0.36, respectively, in negative mode. The values indicate that these models have good fitness and prediction.

Figure 3. PCA and PLS-DA score plots of the serum samples from each group and ROC curve. (a,b) PCA score plots of the CON and MOD groups in positive and negative mode, respectively. (c,d) V-plot of the CON and MOD groups in positive and negative mode. (e,f) PLS-DA score plots in positive and negative mode, respectively. (g,h) Verification of the PLS-DA models by class permutation tests in positive and negative mode, respectively. (i–l) ROC analysis of 38 potential biomarkers. CON: control group; MOD: model group; PLS-DA: partial least square discriminant analysis; ROC: receiver operating characteristic curve.

Screening and identification of metabolic differences

We identified thirty-eight differentially expressed metabolites between the CON and MOD groups based on VIP > 2 and p < .05, which provided more rigorous screening than VIP > 1 and p < .05 in the volcano plots (V-plots) (Figure 3(c,d)). To verify the diagnostic potential of the 38 potential biomarkers, ROC analysis was performed, and the results showed that the AUC values of all 38 potential biomarkers were above 0.9 (Figure 3(i–l)), indicating that these metabolites had good stability. Among these differentially expressed metabolites, 22 metabolites were identified in positive ion mode, 10 that were upregulated and 12 that were downregulated. In the negative ion mode, 16 metabolites were identified, eight that were upregulated and eight that were downregulated (as listed in Table 1). The thirty-eight metabolites can be divided into seven classes, including two benzenoids, four lipids, and lipid-like molecules, four nucleosides, nucleotides, and analogs, 13 organic acids and derivatives, eight organic oxygen compounds, six organoheterocyclic compounds, and one organic nitrogen compound (Figure 4a). These results are shown in a heat map showing relative increases (red) or decreases (green) in the MOD group compared with the CON group (Figure 4c). Significant differences were found for the comparison MOD vs. CON, BUT vs. MOD, AMY vs. MOD, the results were visualized by p-value heatmap (Figure 4d). Twenty and 13 of these metabolites were reversed after treatment with BUT and AMY, respectively, including tyrosyl-Tyrosine, aminoparathion, nicotinicacid mononucleotide, L-beta-aspartyl-L-serine, acetyl-N-formyl-5-methoxykynurenamine, tetradecanoylcarnitine, 4-Hydroxy-4-(3-pyridyl)-butanoic acid, N-Acetylarylamine, diketogulonic acid.

Figure 4. Analysis of the differentially expressed metabolites. (a) The classification of the 38 identified potential biomarkers in serum samples. (b) Venn diagram of all impacted variables (VIP > 2 and p < .05) in the comparisons of AMY and BUT with MOD. (c) Heat map of the differential abundance of metabolites in each group. Rows, samples; columns, metabolites. The degree of colour saturation indicates the metabolite expression value, with blue representing the lowest expression and red representing the highest expression. (d) p-Value heatmap of differential metabolites in serum of rats in each group. BUT: n-butanol extract; AMY: amygdalin extract; MOD: model group.

Table 1. Thirty-eight differentially expressed metabolites were selected and identified between the MOD and CON groups.

No.	Metabolites	m/z	Retention time [s]	MS1hmdbID	Ion mode	VIP^a	Regulated (M/C)	Regulated (B/M)	Regulated (A/M)
1	4E,15Z-BilirubinIXa	585.27	454.18	HMDB0000488	+	3.40	↑#
2	Ornithine	133.09	44.77	HMDB0000214	+	3.23	↑#
3	Succinyladenosine	384.11	128.99	HMDB0000912	+	3.33	↑#	↓*	↓*
4	Diketogulonic acid	215.01	64.67	HMDB0005971	+	2.29	↑#	↓*
5	Tetracosahexaenoic acid	357.28	187.37	HMDB0002007	+	2.40	↑#
6	Tauroursodeoxycholic acid	544.27	157.92	HMDB0000874	+	2.40	↑#
7	o-Tyrosine	182.08	124.39	HMDB0006050	+	2.04	↑#	↓*
8	N-Acetylarylamine	136.07	68.05	HMDB0001250	+	2.19	↑#
9	4-Hydroxy-4-(3-pyridyl)-butanoic acid	182.06	68.16	HMDB0001119	+	2.19	↑#
10	Tetradecanoylcarnitine	372.31	186.85	HMDB0005066	+	2.34	↑#
11	dIMP	371.01	171.06	HMDB0006555	+	2.34	↓#	↑*	↑*
12	5,6-Dihydrouridine	247.09	162.91	HMDB0000497	+	2.65	↓#		↑*
13	4-Imidazolone-5-propionic acid	157.06	170.90	HMDB0001014	+	2.70	↓#	↑*	↑*
14	3′-O-Methylguanosine	315.40	177.01	HMDB0006038	+	2.38	↓#
15	12a-Hydroxy-3-oxocholadienic acid	387.25	181.51	HMDB0000385	+	2.40	↓#
16	Phenylalanyl-Threonine	267.13	170.43	HMDB0029005	+	3.37	↓#	↑*	↑*
17	Methionyl-Threonine	268.13	170.23	HMDB0028983	+	3.42	↓#	↑*	↑*
18	Homoanserine	255.14	209.48	HMDB0005767	+	3.37	↓#	↑*	↑*
19	3-Hydroxypicolinic acid	157.05	142.71	HMDB0013188	+	3.98	↓#	↑*	↑*
20	Hydroxyprolyl-Isoleucine	245.14	170.40	HMDB0028866	+	3.65	↓#	↑*	↑*
21	9′-Carboxy-alpha-chromanol	391.28	174.05	HMDB0012866	+	3.02	↓#
22	Acetyl-N-formyl-5-methoxykynurenamine	265.11	169.93	HMDB0004259	+	3.88	↓#	↑*	↑*
23	Cysteinyl-Proline	217.06	142.52	HMDB0028783	−	3.01	↑#
24	Nicotinic acid mononucleotide	336.04	90.57	HMDB0001132	−	2.92	↑#	↓*	↓*
25	L-beta-aspartyl-L-serine	236.09	149.77	HMDB0011168	−	2.78	↑#	↓*
26	Inodxyl glucuronide	32.10	289.65	HMDB0010319	−	2.43	↑#		↓*
27	S-Adenosylmethioninamine	354.14	147.61	HMDB0000988	−	2.30	↑#
28	Neuraminic acid	283.11	148.81	HMDB0000830	−	2.56	↑#	↓*	↓*
29	Methionyl-Tryptophan	351.14	108.26	HMDB0028984	−	2.19	↑#
30	Dopamine glucuronide	345.13	108.22	HMDB0010329	−	2.05	↑#		↓*
31	Tetragastrin	333.09	93.24	HMDB0005775	−	2.30	↓#
32	N-Acetyl-D-Glucosamine 6-Phosphate	317.07	145.47	HMDB0001062	−	2.11	↓#
33	SR 49498	505.26	209.38	HMDB0013845	−	2.14	↓#	↑*
34	APGPR Enterostatin	533.21	186.36	HMDB0006117	−	2.08	↓#
35	Aminoparathion	277.07	142.68	HMDB0001504	−	2.66	↓#	↑*	↑*
36	Propionylcholine	159.12	163.19	HMDB0013305	−	3.17	↓#	↑*	↑*
37	Tyrosyl-Tyrosine	360.15	209.08	HMDB0029117	−	2.33	↓#		↑*
38	Tetrahydrofolyl-[Glu](2)	573.20	142.53	HMDB0006825	−	4.27	↓#	↑*	↑*

^aVIP: variable important in projection, VIP values were obtained from the PLS-DA model. # compared with the CON group; * compared with the MOD group. ↑: content increased; ↓: content decreased; M, MOD: model group; A, AMY: amygdalin extract group; B, BUT: n-butanol extract group; m/z: mass to charge ratio; s: second.

Metabolic pathway and protective effects of A. mongolica analysis

To directly measure the correlations between the 38 metabolites, Pearson rank correlation analysis (Figure 5a) was performed. The Pearson correlation coefficients of the 32 metabolites were r > 0.8 and r ≤ −0.8. The results showed a good correlation with each other. These significantly expressed metabolites were subjected to MetPA pathway enrichment and functional analysis, and the results showed that the metabolites are involved in metabolic pathways (Figure 5b), including nicotinate and nicotinamide metabolism, pentose and glucuronate interconversions, arginine and proline metabolism, arginine biosynthesis, histidine metabolism, cysteine and methionine metabolism, and amino sugar and nucleotide sugar metabolism. These pathways may be related to the occurrence of RF.

Figure 5. Analysis of potential biomarkers and related metabolic pathways. (a) Pearson rank correlation analysis between the 38 potential biomarkers. The red and blue colour saturation represents the positive and negative correlation coefficients, respectively, between the markers. (b) Summary of the altered metabolism pathways determined with MetPA. The size and colour of each circle are based on the pathway impact value and p-value, respectively. (c) Network of the identified key biomarkers and pathways of BUT and AMY treatment according to the KEGG pathway database. The metabolites coloured green represent common metabolites in “BUT” and “AMY.” TCA: tricarboxylic acid cycle; BUT: n-butanol extract group; AMY: amygdalin extract group.

In addition, in the AMY group compared with the MOD group, 48 metabolites were selected and identified (VIP > 2 and p < .05), of which the concentrations of 47 metabolites were reversed or returned to normal levels (Figure S2(a,b), Table S2). In the BUT group compared with the MOD group, 53 metabolites were identified (VIP > 2 and p < .05), of which the concentrations of 48 were reversed (Figure S2(c,d), Table S3). BUT and AMY from A. mongolica may reverse the contents of serum metabolites in rats with renal fibrosis to normal levels.

To reveal changes in the metabolites and the mechanism of fibrosis treatment with BUT and AMY, the differential metabolites were analyzed by KEGG enrichment. Among the 38 differentially expressed metabolites between the CON and MOD groups, 18 metabolites were acted on by either BUT or AMY, and 14 metabolites were co-acted on by BUT and AMY (Figure 4b, Table 1). The KEGG enrichment results showed that the nicotinate and nicotinamide metabolism, histidine metabolism, and purine metabolism metabolic pathways were significantly changed in the BUT and AMY groups. These pathways are associated with oxidative stress, the release of inflammatory cytokines, and pro-fibrogenic factors, which can promote fibrinolysis. In addition, AMY can also act on pentose and glucuronate interconversions, which were significantly enriched and are associated with improving body energy metabolism (Figure 5c).

Discussion

RF is usually related to the insidious progression of chronic kidney disease (CKD), and the initial symptoms are not obvious. Biomarkers can be used not only as specific diagnostic tools but also as therapeutic targets for diseases. In this study, through metabolomics analysis, nicotinate and nicotinamide metabolism, pentose and glucuronate interconversions, arginine and proline metabolism, arginine biosynthesis, histidine metabolism, cysteine and methionine metabolism, and amino sugar and nucleotide sugar metabolism were found to be related to RF. Therefore, because of their involvement in the development of RF, these metabolic pathways and metabolites could serve as potential diagnostic indicators for early RF and as potential targets to reverse CKD [26].

Among the 38 differentially expressed metabolites between the CON and MOD groups, 18 metabolites were acted on by either BUT or AMY, while 14 metabolites were co-acted on by BUT and AMY. Four metabolites were acted on by BUT alone. For example, diketogulonic acid (DKG) may be involved in damage to the kidney caused by oxalate. DKG is a vitamin C degradation metabolite. The vitamin C degradation pathway produces L-erythrulose and oxalate as final products, and the oxalate formed in this way may contribute to the formation of kidney stones in susceptible individuals [27]. AMY acted alone on four metabolites, which increased the excretion of toxic substances by increasing glucuronic acid conjugation. Glucuronidation assists the excretion of toxic substances, drugs, and other substances that cannot be used as an energy source.

Ornithine can improve athletic performance through anabolic and wound-healing effects, and it can enhance immune function in cells. L-Ornithine is located in the mitochondria and cytoplasm [28]. Moreover, ornithine is associated with cystinuria, hyperdibasic aminoaciduria, and lysinuric protein intolerance, which are genetic metabolic defects. Ornithine is produced in the urea cycle through the cleavage of urea from arginine [26]. L-Ornithine is also a precursor of proline, citrulline, and arginine. Proline is the elementary element of collagen tissue and can promote collagen synthesis and deposition at the lesion site. Ornithine metabolism is involved in the arginine biosynthesis pathway. Arginine biosynthesis may reduce renal tubular interstitial fibrosis and ameliorate renal function [29].

Hydroxyproline is one of the main components of collagen and is unique to collagen fibres. The hydroxyproline content in renal tissue is closely related to the severity of renal interstitial fibrosis [30]. Therefore, hydroxyproline content can be an indicator to judge the severity of renal interstitial fibrosis.

S-adenosine homocysteine (SAH) is the demethylation product of S-adenosylmethionine (SAM), and SAH is hydrolyzed into adenosine (Ado). Ado primarily regulates most internal organ function. SAM is a major methyl donor and increasing concentrations of SAM indicate RF and subsequent impaired kidney function [31]. S-Adenosylmethioninamine is involved in the arginine and proline metabolism and cysteine and methionine metabolism pathways. Hyperhomocysteinemia (HHcy) is a risk factor for glomerular fibrosis [32].

Homocysteine (Hcy) is a sulfur-containing amino acid and is an important intermediate product of methionine metabolism [20,33]. Hcy leads to impaired renal function, which is mainly related to oxidative stress, endoplasmic reticulum stress, and four other aspects [34]. Meyrier et al. [35] found that increased Hcy levels can lead to glomerular sclerosis, renal tubular atrophy, interstitial fibrosis, and a decreased glomerular filtration rate. Hiromichi et al. [36] and Li et al. [37] found arterial and arteriolar wall thickening and focal tubular interstitial fibrosis in the kidneys of rats with high Hcy, suggesting that increased plasma Hcy may be an important pathogenic factor for glomerular injury in hypertension.

We also compared the changes in metabolomics induced by BUT and AMY in the treatment of RF. Nicotinate and nicotinamide metabolism, histidine metabolism, and purine metabolism were influenced by both BUT and AMY. We found that AMY also regulates the disruption of pentose and glucuronate interconversions, nicotinate, and bile secretion [38]. Thus, these pathways were probably the main metabolic pathways involved in the therapeutic effects of the active ingredients in BUT [39].

The main difference in metabolomics between the BUT and AMY groups is that pentose and glucuronate interconversions were more significantly impacted by AMY, which suggests that amygdalin from A. mongolica improves metabolic disorders caused by fibrosis mainly by acting on both pathways. Notably, amygdalin has a significant regulatory effect on SAM and indoxyl glucuronide. Based on the above speculation, we will conduct more in-depth verification and provide additional discussion in future research.

Conclusions

In summary, through this serum metabolomics study based on UPLC-QTOF-MS, we found that BUT and AMY from A. mongolica could improve the changes in the metabolic profile of blood caused by fibrosis. Both extracts had protective effects on rats with renal fibrosis. In addition, amygdalin regulates disrupted pentose and glucuronate interconversions, consequently blocking the progression of fibrosis. In the present study, some potential biomarkers in the development of RF were identified, and the mechanism of BUT and AMY in the treatment of RF was revealed from the perspective of metabolic pathways. In the future, we will verify the metabolic characteristics determined in this study using human samples.

Ethical approval

The animal experiments were approved by the Ethical Committee of Baotou Medical College (Approval Number: 20190314).

Author contributions

Wan-fu Bai carried out the experiments, participated in the study design, and wrote the paper. Chen Gao and Ying-chun Bai completed the model via unilateral ureteral obstruction. Qing Liu performed the extractions and analyzed the extracts. Hong Chang performed the pathological section and analysis. Hong-bing Zhou and Quan-li Liu were involved in the anti-oxidant activity tests in rats. Song-li Shi guided all experiments and revised the article.

Abbreviations
BUT	=	n-butanol extract
AMY	=	amygdalin extract
RF	=	renal fibrosis
PCA	=	principal component analysis
PLS-DA	=	partial least square discriminant analysis
UPLC-QTOF-MS	=	ultra-performance liquid chromatography-quadrupole time of flight-mass spectrometry
ECM	=	extracellular matrix
SOD	=	superoxide dismutase
HYP	=	hydroxyproline
PE	=	petroleum ether
EA	=	ethyl acetate
BU	=	n-butanol
MDA	=	malondialdehyde
HPLC	=	high performance liquid chromatography
CON	=	control
MOD	=	model
H&E	=	haematoxylin & eosin
HMDB	=	human metabolome database
KEGG	=	Kyoto Encyclopaedia of Genes and Genomes
BUN	=	blood diabetic nitrogen
Scr	=	serum creatinine
ALB	=	albumin
CKD	=	chronic kidney disease
DKG	=	diketogulonic acid
SAH	=	S-adenosine homocysteine
SAM	=	S-adenosylmethioninamine
Ado	=	adenosine
Hcy	=	Homocysteine
UUO	=	unilateral ureteral obstruction
TCA	=	tricarboxylic acid cycle

Acknowledgements

The authors would like to thank LC-Bio Technologies (Hangzhou) Co., Ltd., for their help.

Disclosure statement

The authors declare that they have no competing interests.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Additional information

Funding

This research was financially supported by the National Natural Science Foundation of China (NSFC) [No. 81641137 and 81760782]; the Natural Science Foundation of Inner Mongolia Autonomous Region of China [No. 2018LH03027, 2021LHMS08013 and 2019MS08189]; a Scientific Research Fund Project of Baotou Medical College [No. BYJJ-YF 201706]; a Baotou Medical College Doctoral Scientific Research Foundation Project [No. BSJJ201814]; and a Grassland Talent Project [No. Q2017046].