Adenine and formic acid were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Creatinine (batch No.: 100877-200901, Purity 99.8%) was obtained from the National Institutes for Food and Drug Control (Beijing, China). LC-grade methanol and acetonitrile were purchased from the Baker Company. Ultra-high purity water was prepared using a Milli-Q water purification system. Other chemicals were of analytical grade and purity was above 99.5%.

Sprague-Dawley rats (male, aged 6 weeks, 180–200 g) were obtained from the Experimental Animal Center of the Fourth Military Medical University (Xi'an, China). Rats were maintained at a constant humidity (~60%) and temperature (~23°C) with a light/dark cycle of 12 h. Experimental studies were approved by the Ethics Committee for Animal Experimentation of the Northwest University. The ethical approval reference number of the study is SYXK2010-004.

In January 2016, AO was collected from Fujian Province and was identified by Prof. M. F. Fang (School of Life Science, Northwest University, Xi'an, Shaanxi, China). A voucher specimen (Z150319) was deposited at the School of Life Science, Northwest University, Xi'an, Shaanxi. Crushed AO (2.5 kg) was extracted three times with 95% ethanol at room temperature for 5 days each time. The extraction rate of the ethanol extract of AO was 6.44%. The filtrates were concentrated under reduced pressure to give a crude reddish-brown extract, which was then dissolved in H₂O. The suspension was successively fractionated three times with petroleum ether, ethyl acetate, and n-BuOH, and concentrated under reduced pressure to yield four fractions. Previous studies have shown that the ethyl acetate fraction of AO contained the active compound (Feng et al., 2014). The ethyl acetate fraction of AO was orally administered to rats.

Male rats underwent an adaptation period of several days, during which they were fed commercial feed. Rats weighing 180 g to 200 g were randomly divided into three groups: (1) healthy control group (n = 8), (2) CKD model group (n = 8), and (3) AO-treated group with CKD (n = 8). Groups 2 and 3 were given 200 mg/kg body weight of adenine dissolved in 1% (w/v) gum acacia solution, administered by oral gavage once a day for three consecutive weeks (Zhao et al., 2013). Similarly, group 1 was given an equal volume of gum acacia solution. Group 3 received, 3 h after adenine gastric gavage, the ethyl acetate extract (130 mg/kg) by gastric irrigation during the 6-week study period.

Three weeks after the start of treatment, rats were anesthetized using 10% urethane, and blood samples were obtained by the carotid artery cannula. Blood was centrifuged at 3,000 rpm for 10 min, after which the serum was collected and stored at −80°C. After blood was collected, kidneys were harvested immediately and were washed with saline.

At the end of the experiment, rats were housed individually for 24 h in metabolic cages for urinary collection. Body weights were recorded after the 24 h time-period. Upon collection of their urine samples, rats were allowed free access to water. Urine samples were stored at −80°C prior to analysis, and kidneys were weighed to calculate organ indexes (organ index = organ weight/body weight ×100%). Serum levels of creatinine, urea, cholesterol, triglyceride, low density lipoprotein-cholesterol (LDL-C), and high density lipoprotein-cholesterol (HDL-C) were measured using an Olympus AU640 automatic analyzer.

A portion of each fresh kidney was immersed in 10% neutral, buffered formaldehyde solution, then dehydrated, embedded in paraffin, cut in 5 μm sections, and stained with Hematoxylin and Eosin (H&E) and Masson staining for histopathological examination. Immunohistochemical staining was performed using the published procedure (Chan et al., 2017). The antibodies used were directed against the following: nuclear factor kappa B p65 (NF-κB p65), cyclooxygenase-2 (COX-2), Monocyte Chemoattractant Protein-1 (MCP-1), inducible nitric oxide synthase (iNOS), tumor necrosis factor-alpha (TNF-α), Nrf2, heme oxygenase-1 (HO-1), transforming growth factor-β1 (TGF-β1), transforming growth factor-β receptor II (TGF-β RII), Smad2, Smad3, Smad4, Smad7, fibronectin, alpha-smooth muscle actin (a-SMA), plasminogen activator inhibitor-1 (PAI-1), a-SMA, collagen I, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Antibodies were purchased from Abcam (Cambridge, United Kingdom), Cell Signaling Technology (Danvers, MA), or Santa Cruz Biotechnology, Inc., (Santa Cruz, CA). Image analysis was performed by using Image-Pro Plus 6.0 software.

Proteins were extracted using radioimmunoprecipitation assay buffer, which contained a cocktail of proteinase inhibitors (Thermo Fisher Scientific Inc., Rockford, IL), and were quantified with a Bio-Rad protein assay. Equal amounts of protein were separated on 10% SDS-polyacrylamide gels in a Tris/HCl buffer system, transferred onto nitrocellulose membranes, and blotted according to standard procedures. Nonspecific proteins were blocked by incubating the membrane with 5% nonfat dried milk in TBS-T for 1 h at room temperature with agitation. Membranes were then incubated overnight at 4°C with the primary antibodies directed against PAI-1, a-SMA, collagen I, or GAPDH. Subsequently, membranes were washed and then incubated with goat anti-rabbit IgG or goat anti-mouse IgG secondary antibodies (Abcam, Cambridge, MA) for 1 h at room temperature (Pelletier et al., 2017). Specific bands indicating target proteins were analyzed using ImageJ software.

For the lipid profiling of blood samples, UPLC-HDMS was used. The extraction of total lipids by the Ostro 96-well plate system was performed as a single-step in-well extraction. Samples for lipidomics were prepared as previously described (Chen H. et al., 2017).

UPLC-HDMS was performed on a Waters Acquity Ultra Performance LC system, equipped with a Waters Xevo G2 QTof MS. UPLC analysis was performed using a HSS T3 column. The samples were characterized following our previously published methods with minor modifications (Chen H. et al., 2017). All acquisitions were operated using Waters MassLynx v4.1 software.

Precision and reproducibility were verified by previously described methods (Zhao et al., 2013). In brief, raw data were imported to Markerlynx XS (Waters Corporation, MA, USA) for peak detection and alignment. Data were normalized to the summed total ion intensity per chromatogram and the resultant data matrices were loaded into EZinfo 2.0 software (Umetrics Corporation, Sweden) for principal component analysis (PCA) (Zhao et al., 2013). Partial least-squares discriminant analysis (PLS-DA) was performed and potentially important variables were selected according to the variable importance in the projection (VIP) values, which reflected the contribution of each variable in the three groups. Metabolite peaks were assigned by MSE analysis or interpreted with available biochemical databases: HMDB (http://www.hmdb.ca/), KEGG (http://www.genome.jp/kegg/pathway.html) and Chemspider (http://www.chemspider.com) (Zhao et al., 2013). From the identified lipid species, heatmap, fold changes (CKD/control, CKD+AO/CKD, or CKD+AO/control), and receiver-operating characteristic (ROC) curves were analyzed by Metaboanalyst 3.0 and Medcalc 12.7. The one-way analysis of variance (ANOVA) or Mann–Whitney U test was used to calculate statistical significance with SPSS 22.0 software. The false discovery rate (FDR) correction was calculated to reduce the risk of a false positive value by the adjusted p < 0.05 based on the Benjamini Hochberg method. Values of p < 0.05 were considered significant.

In order to construct a metabolic network of identified lipids, we integrated interaction databases: Reactome, Database for Annotation, Visualization and Integrated Discovery (DAVID), Kyoto Encyclopedia of Genes and Genomes (KEGG), Biomolecular Interaction Network Database (BIND), Human Protein Reference Database (HPRD), BioGrid, Database of Interacting Proteins (DIP), Metabolite Set Enrichment Analysis (MSEA), and Quantitative Enrichment Analysis (QEA) (Xenarios et al., 2002; Bader et al., 2003; Huang da et al., 2009; Goel et al., 2012; Xu et al., 2017). These databases contained the most comprehensive publicly available repository of genes, proteins, and interaction of complexes for Homo sapiens. In these integrative databases, sets of lipid species from metabolite-related networks were mapped to their related molecular pathways and networks.

In Figure 1, body weight, urine volume, kidney weight index, creatinine, urea, cholesterol, triglyceride, LDL-C, and HDL-C are given for the three groups. Our data indicated that body weight, urinary volume, and kidney weight index were significantly decreased in the CKD group compared to the control group (p < 0.01). However, body weight, urinary volume, and kidney weight index were markedly increased in the AO-treated group compared to the CKD group. In addition, levels of serum creatinine, urea, cholesterol, triglyceride, and LDL-C were significantly increased in the CKD group compared to the control group (p < 0.01). Moreover, HDL-C levels were significantly decreased in the CKD group compared to the control group. These biochemical parameters were improved after treatment with AO.

Compared to control rats, adenine-induced CKD rats showed significant up-regulation in the nuclear translocation of p65 protein expression, indicating the activation of NF-κB signaling. In adenine-induced CKD rats, NF-κB activation was accompanied by a significant up-regulation of inflammatory proteins, such as COX-2, MCP-1, iNOS, and TNF-α, and down-regulation of the anti-oxidant system, including Nrf2, and HO-1. Compared with CKD rats, up-regulation of NF-κB p65, COX-2, MCP-1, iNOS, and TNF-α protein expression, and down-regulation of Nrf2 and HO-1 protein expression was attenuated by AO treatment (Figure 2). Thus, the possible mechanisms underlying the nephropathic effect of AO may involve the TGF-β1/Smad signaling pathway.

Activation of the TGF-β1/Smad signaling pathway contributes to renal interstitial fibrosis (Ma et al., 2010). Compared to control rats, protein expression of TGF-β1 and TGF-β RII were significantly up-regulated in adenine-induced CKD rats (Figure 3). This up-regulation was accompanied by significant up-regulation of Smad2, Smad3, Smad4, and Smad7 protein expression in adenine-induced CKD rats, indicating activation of the TGF-β1/Smad signaling pathway. Compared to CKD rats, activation of the TGF-β1/Smad signaling pathway was attenuated in AO-treated rats (Figure 3). Therefore, the possible mechanisms underlying the nephropathic effect of AO may involve the NF-κB/Nrf2 signaling pathway.

Figures 4, 5 demonstrate the H&E and Masson immunohistochemical staining results, and Western Blot analysis of kidney tissue derived from adenine-induced rats. The formation of foreign body granuloma in the renal tubules and renal interstitial fibrosis, as well as a significant degree of renal fibrosis was observed. Histological analysis demonstrated that in adenine-induced rats, typical pathological features of CKD were observed (Figures 4, 5). In contrast, these pathological abnormalities were alleviated in AO-treated rats. Adenine administration resulted in a significant increased protein expression of fibronectin, collagen I, PAI-1 and α-SMA, whereas AO administration reduced the expression of these molecules compared to CKD rats (Figure 5). In addition, in adenine-induced CKD rats, significant up-regulation of pro-fibrotic proteins was accompanied by activation of inflammatory and TGF-β1/Smad pathways. In conclusion, AO attenuated NF-κB/Nrf2, TGF-β1/Smad, and pro-fibrotic pathways in adenine-induced CKD rats.

For all analyses, reproducibility was confirmed by performing six replicated determinations of each plasma sample. For the method validation, eight ions, including 2.40_320.3081, 7.81_742.6677, 6.30_647.4589, 4.46_1194.8158, 5.02_662.4451, 8.03_784.7189, 5.02_722.5260, and 8.34_700.6142, were selected. The relative standard deviation values of the retention time and peak area were <0.73 and 2.95%, respectively. These results indicated sufficient reproducibility of these analyses.

In order to evaluate the ability of AO to ameliorate the lipid metabolic profile in adenine-induced CKD rats, the two-predictive component PLS-DA model [R2X(cum) = 0.961, Q2(cum) = 0.751] was used. In both the positive and negative ion modes, this model showed a satisfactory separation capacity by 384 variables among the three groups (Figure 6A). As shown in the clustering analysis, adenine-induced CKD rats can be separated from control and AO-treated rats, however AO-treated rats cannot be separated from the control rats (Figure 6B). Therefore, these data demonstrated the efficacy of AO in improving treatment of CKD rats.

To identify altered lipid species, variables were selected using VIP values based on the PLS-DA model: 383 variables in the positive ion mode and 58 variables in the negative ion mode had a VIP value of >1.0 (Figure 6C). Based on the one-way ANOVA analysis and Mann–Whitney U test (p < 0.05), 100 variables in the positive ion mode and 23 variables in the negative ion mode were selected for further analysis. After combining the FDR analysis (p < 0.05) and the ROC curves (AUC > 0.80), a total of 20 variables were selected for the identification of lipid metabolites in both the positive and negative ion modes. Xenobiotics and different fragment ions from the same lipid species were excluded, and 14 differential lipid species were identified (Table 1). These lipids included five diglycerides, four triglycerides, three fatty acids, one phosphatidylcholine, and one phosphatidylethanolamine. In AO-treated rats, changes in eight lipid species were completely restored when compared to CKD rats. In addition, five lipid species were restored to normal levels in the AO-treated group compared to the control group. The PCA score plot of the 14 lipid species in the AO-treated group was located between those of the CKD and control rats (Figure 7A), which was consistent with the results of the heatmap analysis (Figure 7C). Moreover, Figure 7B indicates the correlation coefficient analysis among lipid species and their corresponding groups. Lipid species situated in the upper panel positively correlated with the corresponding group, whereas those situated in the opposite panel negatively correlated with the corresponding group. The lipid species 8,9-epoxyeicosatrienoic acid (8,9-EET), DG (44:6), DG (40:9), TG (46:6), PE (33:3), DG (46:6), eicosapentaenoic acid (EPA), PC (34:6), TG (60:14), and docosahexaenoic acid (DHA) were positively correlated with control rats, whereas all other lipid species were negatively correlated with control rats. The correlation coefficient was significantly altered in adenine-induced CKD rats, indicating that the significant abnormalities of the lipid metabolic profile were caused by adenine. In AO-treated rats, nine lipid species showed similar altered tendencies observed in control rats. The results demonstrated that AO ameliorated abnormal lipid metabolism in CKD-afflicted rats using adenine, which coincides with our biochemical results.

PLS-DA-based ROC curves were performed to further identify potential lipid biomarkers that would indicate the therapeutic effects of AO on CKD. The area under the curve (AUC), 95% confidence interval (95%CI), sensitivities, and specificities of 14 lipid species are shown in Figure S5. Seven lipid species, including DG (44:6), DG (35:1), TG (46:6), DG (42:4), DG (46:6), TG (64:9), and DHA had a high AUC value (>0.80), sensitivity (>80%), and specificity (>80%). These lipid species included four diglycerides, two triglycerides, and one fatty acid and could be considered as potential lipid biomarkers of the therapeutic effects of AO on CKD. Figure 8 demonstrates the levels of the 14 lipid biomarkers among the three groups. In conclusion, these results indicate that diglycerides, triglycerides, and polyunsaturated fatty acids can serve as potential biomarkers in adenine-induced CKD rats, and that AO mitigates the abnormal lipid metabolism, and exhibits anti-fibrotic effects.

Adenine triggers an imbalance between lipid synthesis and degradation. To understand the functional role of the altered lipids, we employed the KEGG database using Metaboanalyst. We evaluated both a test for over-representation of altered lipid species within a pathway (hypergeometric tests), and the impact of the altered lipid species on the function of the metabolic pathway via alterations in critical junction points of the metabolic pathway (relative between centrality). The results obtained by the 82 mouse pathways from the KEGG database were plotted to highlight the most significant metabolic pathways according to the hypergeometric test p-values (Y-axis) and impact (X-axis) (Figure 9A). The top three pathways in CKD by p-value (top two) or impact (top one) were identified, including glycerophospholipid metabolism, arachidonic acid metabolism, and biosynthesis of unsaturated fatty acids (Table 2). As an example, detailed information of the glycerophospholipid metabolism is represented in Figure 9B. The remaining pathways are shown in Figures S1–S4. Alteration of these pathways in CKD indicates that the disturbance of certain central lipid species has an important impact on multiple metabolic lipid pathways that are interconnected. A pathway enrichment overview of altered lipids highlights the arachidonic acid metabolism, as well as γ-linolenic acid, and linolenic acid metabolism for its remarkable enrichment in quantitative lipidomics from adenine-induced CKD rats vs. control rats (data not shown).

To map the pathways overrepresented by the identified lipid species from adenine-induced CKD, identified lipid metabolic networks were constructed using various databases to determine the set that was most enriched by these lipid species. By analyzing known pathways, the image information outcome presented the biological pathway information associated with CKD. The pathway overrepresentation analysis of lipid species showed that glycerophospholipid metabolism, triacylglycerol degradation, fatty acid activity, phospholipid biosynthesis and metabolism, arachidonic acid metabolism, biosynthesis of unsaturated fatty acids, γ-linolenic acid and linolenic acid metabolism, and metabolism of lipids and lipoproteins were significantly overrepresented in adenine-induced CKD (Figure 9C).

In adenine-induced CKD rats, significantly decreased levels of three polyunsaturated fatty acids, including 8,9-EET, EPA, and DHA were observed. These changes were completely restored by treatment with AO. Cytochrome P-450 epoxygenases metabolizes arachidonic acid into four regioisomeric epoxyeicosatrienoic acids (EET), including 5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET by catalyzing the epoxidation of the olefinic bonds of arachidonic acid (Kim et al., 2014). Previous studies have indicated that EET has potential anti-inflammatory, antihypertensive, profibrinolytic, and anti-fibrotic effects (Imig et al., 2002; Spector and Norris, 2007; Kaspera and Totah, 2009). Soluble epoxide hydrolase converts EET to the less active dihydroxyeicosatrienoic acid. EET that has been increased by the inhibition of soluble epoxide hydrolase exerts nephropathic effects, such as hypertensive renal damage, diabetic nephropathy, and obstructive nephropathy (Kim et al., 2015). It has been shown that 5/6 nephrectomized rats contained significantly decreased serum levels of EET (Zhang K. et al., 2015). Kidney tissues increase EET levels by preventing their degradation to inactive dihydroxyeicosatrienoic acids, which exhibit antihypertensive, cardio- and nephropathic activities. In a recent study, it was shown that serum 8,9-EET levels were significantly decreased in patients with advanced CKD (Chen H. et al., 2017).

EPA and DHA are omega-3 polyunsaturated fatty acids (ω-3 PUFA). Several clinical studies have suggested that ω-3 PUFA exhibits beneficial effects on patients with end-stage renal disease (Svensson et al., 2006; Lok et al., 2012). Moreover, it has been reported that compared with healthy individuals, patients with CKD have reduced levels of plasma n-3 PUFA (Madsen et al., 2011). Furthermore, it was demonstrated in many studies that in hemodialysis patients, ω-3 PUFA is significantly decreased (Gharekhani et al., 2014; Hung et al., 2015; Umemoto et al., 2016). Highly prevalent systemic inflammation is related to a high mortality rate due to cardiovascular issues in CKD patients who are on dialysis. In several studies it was demonstrated that in the general population, long-chain EPA and DHA are cardioprotective (De Caterina, 2011). The anti-inflammatory effects of EPA and DHA are considered one of the most relevant mechanisms associated with CKD and the beneficial effects of ω-3 PUFA (Spencer et al., 2013). In one study, it was indicated that EPA decreased the production of cytokines in lipopolysaccharide-induced cytokine production (Saifullah et al., 2007; Spencer et al., 2013). Previous studies have shown a significant decrease in the levels of EPA and DHA in kidney tissues derived from CKD rats (Zhao et al., 2013). In the present study, we demonstrated that in adenine-induced CKD rats, a significant reduction in EPA and DHA levels is associated with an increase in systemic inflammation, particularly in nuclear factor NF-κB p65, and the chemokines COX-2 and iNOS. Additionally, it has been reported that in kidney tissues of CKD rats, a significant reduction in EPA and DHA levels is associated with up-regulation of the TGF-β1 level (Zhao et al., 2013). Accumulated evidence has demonstrated that adenine-induced CKD results in a significant reduction in serum levels of 8,9-EET, EPA, and DHA that are associated with elevated levels of inflammation and renal fibrosis in adenine-induced CKD rats. Therefore, significantly increased levels of 8,9-EET, EPA, and DHA caused by AO treatment may be associated with inflammation improvement in adenine-induced CKD rats.

Compared with control rats, significantly increased levels of TG (46:6), TG (64:9), TG (60:14), and TG (68:12) as well as significantly decreased levels of DG (44:6), DG (35:1), DG (40:9), DG (42:4), and DG (46:6) were observed in adenine-induced CKD rats. Patients with CKD demonstrated disturbed metabolisms of glycerolipids including monoglycerides, diglycerides, and triglycerides. Previous studies have indicated significant alterations in glycerolipid synthesis and catabolism in patients with CKD (Chen H. et al., 2017). Most previous studies have measured total TG levels in patients with CKD. It has been demonstrated that the triglyceride level of a low-density lipoprotein was significantly increased in CKD patients (Reis et al., 2015). However, increased levels of triglyceride were commonly accompanied by significantly reduced levels of HDL-C in CKD patients (Piperi et al., 2004). Hypertriglyceridemia in CKD was associated with impaired triglyceride clearance caused by down-regulation of lipoprotein lipase and a very low-density lipoprotein receptor (Vaziri et al., 2012). Moreover, it has been shown that a 3-month diglyceride intake lowered the level of abdominal fat and retarded serum lipid profiling in hemodialysis patients. Diglyceride intake significantly reduced serum levels of very low-density lipoproteins, altered monoglycerides, and elevated high-density lipoprotein levels at 3 months (Teramoto et al., 2004). Taken together, increased triglyceride and decreased diglyceride levels were consistent with clinical biochemical findings, as well as with previously published reports. The present study demonstrated that in the adenine-induced rats, both increased levels of triglyceride and decreased levels of diglyceride are restored by treatment with AO, indicating that AO may attenuate the perturbations of glycerolipids synthesis and catabolism.

Several studies have demonstrated that the degree of renal tubular interstitial injury positively correlates with an increase in oxidative stress and inflammation after renal injury (Djamali, 2007). Therefore, in this study we examined whether renal tubular interstitial injury was associated with increased oxidative stress. We found that renal tubular interstitial injury was associated with NF-κB activation, up-regulation of pro-oxidant and pro-inflammatory activities, and down-regulation of Nrf2 activity and its down-stream antioxidant and cytoprotective proteins. These findings were accompanied by activated canonical TGF-β1/Smad pathways (Zhao et al., 2013; Shang et al., 2016). Our results also revealed that renal tubular interstitial injury dramatically increased levels of TGF-β1, Smad2, Smad3, Smad4, and significantly down-regulated the activities of Smad7 in CKD rats. In summary, our findings demonstrated that AO treatment attenuated renal damage and that the mechanism may in part involve suppression of oxidative stress, inflammation and the TGF- β1/Smad signaling pathway.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.