The Ginsenoside Rg1 Rescues Mitochondrial Disorders in Aristolochic Acid-Induced Nephropathic Mice

Chronic exposure to aristolochic acid (AA) leads to renal interstitial fibrosis and nephropathy. In this study, we aimed to investigate the renoprotective effects of Panax ginseng extract (GE) and ginsenoside saponin (GS) on AA-induced nephropathy (AAN) in mice. Eighty female C3H/He mice were randomly divided into eight groups, including normal; AA (3 μg/mL for 56 days); AA with GE (125, 250, or 500 mg/kg/d for 14 days); and AA with important GE ingredients, Rg1, Rb1, or Rd (5 mg/kg/d for 14 days). Compared with the AA group, renal injuries were significantly decreased in the GE (250 mg/kg/d), Rb1, and Rg1 treatment groups. Rg1 exhibited the best renoprotection among all GS-treated groups. There were 24 peaks significantly altered among normal, AA, and AA + Rg1 groups, and four mitochondrial proteins were identified, including acyl-CoA synthetase medium-chain family member 2, upregulated during skeletal muscle growth 5 (Usmg5), mitochondrial aconitase 2 (ACO2), and cytochrome c oxidase subunit Va preprotein (COX5a). We demonstrated for the first time that the AAN mechanism and renoprotective effects of Rg1 are associated with expression of mitochondrial proteins, especially ACO2, Usmg5, and COX5a.


Introduction
In the 20th century, rapidly progressive tubulointerstitial nephritis was reported in female patients who followed slimming regimens containing aristolochic acid (AA). According to National Health Insurance (NIH) data, about one-third of Taiwanese people were prescribed with AA-containing products between 1997 and 2003 [1]. Although AA-containing products are prohibited in various countries, AA-induced nephropathy (AAN) is still frequently reported all over the world, especially in Asian populations [2]. Aristolochic acid I (AAI) and aristolochic acid II (AAII) are the major components of the AA mixture in the plant extracts. However, AAI is more toxic than AAII due to the substitution of the methoxyl group. Intraperitoneal injection of AA induces acute kidney injury (AKI), which leads to progressive chronic kidney disease (CKD) in mice [3]. In AA-induced AKI mice, oxidative stress was associated with inflammation [4]. The proximal renal tubules are the primary cellular target of AA, especially in the proximal tubular S3 segment [5,6].
Recent studies have confirmed that AA can induce severe mitochondrial damage and condense mitochondrial membrane density through increasing the levels of superoxide anion, ferrous ions, and reactive oxygen species (ROS); lowering mitochondrial membrane potential (MMP); decreasing mitochondrial DNA (mtDNA) copy number; reducing adenosine triphosphate (ATP) production; and activating the renal mitochondrial apoptosis

Preparation of the Ginseng Extract
The extraction of Panax ginseng was performed following the protocol published by Kim et al. with a minor modification [25]. Six-year-old Ginseng Radix Rubra (300 g) was boiled for 6 h in 50% ethanol (1:10, w/v), and we collected the filtrate in triplicate. The filtrate was concentrated using rotary evaporation and was lyophilized. The overall yield of GE was about 39% of the dry weight of Ginseng Radix Rubra.

Animal Experiments and Sample Collection
Six-week-old female C3H/e mice were purchased from the National Laboratory Animal Center (Taipei, Taiwan) and maintained in the Laboratory Animal Center of Taipei Medical University. All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee or Panel (permit number: LAC-101-0318) to minimize pain and discomfort. Mice were given feed and water ad libitum. The mice were randomly divided into eight groups (n = 10/each), including normal; AA; AA with GE (125, 250 m or 500 mg/kg/d); and Rg 1 , Rb 1 , or Rd 5 mg/kg/d treatment. In the AA group, AA sodium salt was dissolved in distilled drinking water (3 µg/mL, 0.5 mg/kg/d) for 56 days [26]. After 56 days, drinking water was replaced by distilled water for another 14 days. In treatment groups, after AA induction for 56 days, 0.1 mL of GE or GS was given orally for the next 14 days. The normal group was given distilled water during the experiment.
Twelve-hour urine was collected at day 70. At the end of the experiment, mice were sacrificed. Blood and kidney tissues were collected. Blood was collected from the tail vein and left at 25 • C for clotting. Afterwards, blood was centrifuged at 4 • C, 1006× g for 15 min, and then the supernatant was collected as the sera. The kidneys of mice were removed, rinsed with 0.9% normal saline, decapsulated, and fixed in 10% buffered neutral formalin, followed by dehydration in a gradient ethanol solution, and xylene clearing. Subsequently, kidneys were embedded in paraffin.
Urinary protein level was determined using Bio-Rad protein assay kits (Bio-Rad, Hercules, CA, USA) according to the Bradford method [31]. Bovine serum albumin (Bio-Rad Inc., Herecules, CA, USA) was used as a standard solution, and the protein content was positively correlated with the O.D. at 590 nm.
2.5. Determination of Rg 1 , Rb 1 , and Rd in GE Rg 1 , Rb 1 , and Rd were quantified in GE through HPLC coupled with a UV-visible detector (Hitachi, Tokyo, Japan). The GE sample was dissolved in ddH 2 O (100 mg/mL, w/v) and filtrated through a Millipore 0.45 µm syringe filter (Millipore, Darmstadt, Germany). The resultant filtrate (30 µL) was directly injected to the HPLC system, followed by separation on an Intertsil ODS-2 column (4.6 × 150 mm I.D., 5 µm) (GL Sciences Inc., Tokyo, Japan) at 25 • C. The mobile phases of Rg 1 , and Rb 1 and Rd were 20% or 30% aqueous ACN, respectively. The flow rate was set as 1 mL/min. The wavelength was 203 nm. All HPLC equipment was purchased from Hitachi (Tokyo, Japan). All HPLC methods were validated, in terms of the precision of intra-assay, inter-assay (CV < 15%), and accuracy (85-115%).

Fluorogenic Derivatization-Liquid Chromatography-Tandem Mass Spectrometry (FD-LC/MS/MS)-Based Renal Proteomics
The FD-LC system, consisting of an L-2130 Intelligent Pump, an L-2200 Intelligent Autosampler, and an L-2485 Fluorescence Detector (FD, Hitachi, Tokyo, Japan), was used to separate the mixture protein in a WX-RP column (4.6 × 250 mm I.D., 3 µm) (Imtakt Co., Kyoto, Japan). Twenty microliters of the fluorogenic sample was injected into the HPLC column at 0. 55  The gradient was set with the following elution: 5% B and 1% C held for 10 min; 30% B and 35% C for 5 min and held for 15 min; 35% B and C for 10 min; 38% B and 35% C for 20 min; 44% B and 55% C for 30 min and held for 50 min; 47% B and 53% C for 10 min; 48% B and 52% C for 45 min; 51% B and 49% C for 25 min; 60% B and 40% C for 200 min; 70% B and 30% C for 80 min; 90% B and 10% C for 30 min and held for 30 min; 100% B.
The significantly altered peaks from FD-LC were collected. The eluates were concentrated and dried using Savant Speed Van (Model SPD111V, Savant Instruments, Inc., Holbrook, NY, USA), and then digested with 2.5 µL of sequencing grade trypsin, 10 mM CaCl 2 , and 20 µL 50 mM NH 4 HCO 3 at 37 • C for 2 h. The digested samples were dried and concentrated as before, and then refilled with 10 µL of mobile phase (A) for protein identification. The MS system that consisted of API 4000Q TRAP, Aglient 1100, and 1200 (Agilent, Santa Clara, Germany) was used to identify proteins. After biological samples were injected into the MS system, these were condensed in a guard column (5 × 0.3 mm I.D., particle size 5 µm) (ZORBAX 300SB-C18, Agilent Santa Clara, Germany) and then separated with a C 18 column (75 µm × 150 mm I.D., particle size 5 µm) (CVC Micro-Tech Scientific Inc., Fontana, CA, USA). Subsequently, MASCOT searching engine with NCBInr database was applied for protein identification.

Isolation of Renal Mitochondria for Western Blot Analyses
Mitochondria were extracted using a previously published protocol [32]. The renal homogenates were centrifuged at 600× g for 10 min, and the supernatant (S1) was collected. S1 was centrifuged at 9000× g for 10 min, and the supernatant was removed. The pellet (P1) that contained mitochondria was resuspended in the mitochondrial buffer (200 µL of 250 mM sucrose, 10 mM Tris-HCl, pH 7.6, and 1 mM Na 2 EDTA) and centrifuged at 9000× g for 10 min. The pellet (P2) was resuspended in the mitochondrial buffer (200 µL). Protein concentrations were determined using the Pierce™ BCA protein assay kit (Thermo Scientific, Waltham" IL, USA).

Multivariate and Univariate Statistical Analysis
Multivariate statistical analysis (Metaboanalyst 4.0) was used to evaluate the renal differential proteomics [33]. Partial least squares discriminant analysis (PLS-DA) was one of the supervised methods that were used to predict the significant observation. Due to PLS-DA having an overfitting problem, the validation of the PLS-DA model was necessary. In leave-one-out cross-validation (LOOCV), the difference between R 2 and Q 2 was less than 0.3, meaning the PLS-DA model was suitable for the discovery of the potential feature in omics. The normality of distribution was analysis with Metaboanalsyt online. The distribution normalization to sample median is shown in box plots in the Supplementary Material ( Figure S1).
All experimental data are expressed as mean ± standard deviation. Two-tailed Mann-Whitney U tests were used to analyze the significant differences between pairs of groups. A p-value of less than 0.05 meant a significant difference. SPSS version 20.0 was used (SPSS Inc., Chicago, IL, USA) for analysis.

Clinical Chemistry and Histological Examination of Renal Tissues
Compared with the normal group, the AA group had higher urinary protein, NAG, BUN, and serum creatinine levels. After co-treatment with GE 125, 250, or 500 mg/kg/d, urinary protein, NAG, and BUN levels were significantly decreased, and the serum creatinine level was significantly decreased only in AA and GE 250 mg/kg treatment groups compared with the AA group (Table S1). After co-treatment with Rg 1 , Rd, or Rb 1 , urinary protein, NAG and BUN levels were significantly decreased, and the serum creatinine level was significantly decreased in Rb 1 and Rg 1 groups compared with the AA group (Table 1). In the AA group, there was tubular atrophy, inflammatory cell infiltration, and interstitial fibrosis in renal tissues. After co-treatment, renal damage was alleviated by the GE treatment, especially at a dose of 250 mg/kg ( Figure S2), and in Rg 1 (Figure 1).

Quantification of GS Contents in GE
The quantification of GS content in GE was validated, and the results met the criterion (CV < 15%) ( Table S2). The results showed that GS accounted for 2.64 to 8.10% of GE (Table S3).

Renal Differential Proteomics Coupled with Multivariate Analysis
We chose normal, AA, and AA with Rg 1 (5 mg/kg/d) groups to conduct renal differential proteomics with multivariate analysis ( Figure S3). The renal homogenates were analyzed using FD-LC-MS/MS. After separation, 24 peaks were significantly different between these groups ( Figure 2). In the PLS-DA model, three groups were significantly separated in the score plot ( Figure 3A), and several observations contributed to this model in the loading plot ( Figure 3B). The LOOCV revealed that the PLS-DA model was not overfitting ( Figure S4) and was suitable for the discovery of the potential biomarkers. All peaks of significance were further identified using the Mascot search engine, and the results are shown in Table 2.

Western Blotting Analysis
According to the coefficient score of PLS-DA, six observations were greater than 40, which were regarded as important features ( Figure 4A), especially ACO2. Western blotting analysis showed that ACO2 was significantly upregulated in the AA group compared with the normal group. After Rg 1 (5 mg/kg) co-treatment, ACO2 expression slightly decreased ( Figure 4B,C). The trends of Western blotting analysis and FD-LC/MS/MS results were similar. Therefore, FD-LC/MS/MS was suitable for the discovery of the potential biomarkers.

Discussion
This study reveals that ginsenoside Rg 1 can rescue mitochondrial disease in mice with aristolochic acid-induced nephropathy. Our results showed that administration of AA caused renal interstitial fibrosis, inflammatory cell infiltration, and proximal tubular damage. Liu et al. demonstrated that administration of AA leads to renal interstitial fibrosis accompanied by mitochondrial fragmentation in chronic AAN rats [34]. AA promotes mitochondrial DNA (mtDNA) damage; decreases mtDNA copy number, mitochondrial protein expression, and ATP content; and increases oxidative stress in cultured podocytes [35]. Mitochondrial damage leads to the release of cytochrome c into cytoplasm, which results in apoptosis [36]. Therefore, mitochondria-related proteins might have a crucial role in AAN mice.
It is well known that GE (100 or 250 mg/kg/d) attenuates renal dysfunction by alleviating advanced glycation end product-mediated with diabetic nephropathy [17]. Yokozawa et al. demonstrated that administration of GE decreases BUN levels in nephrectomized rats [37]. According to our results, administration of 250 mg/kg GE had better renoprotective effects than other doses (125 and 500 mg/kg) on chronic AAN in mice. In order to investigate the renoprotective effects of ginseng on AAN mice, we selected major components of GE, including Rg 1 , Rd, and Rb 1 , as potential candidates. All ginsenosides ameliorated AA-induced renal injury. Notably, Rg 1 provided the best renoprotection, revealed using histopathology and clinical chemistry. Rg 1 prevents 1-methyl-4-phenylpyridinium ion-induced apoptosis by inhibiting the production of ROS and activating the JNK pathway in SHSY5Y cells [38]. Early studies have shown that Rg 1 (20 to 50 mg/kg) reduces the damage to the glomerular structure in spontaneously hypertensive rats by inhibiting oxidative stress, and inhibits renal interstitial fibrosis in rats with unilateral ureteral obstruction [19,39]. Rg 1 inhibits renal interstitial fibrosis in rats with unilateral ureteral obstruction via suppressing both active tumor growth factor-β1 and phosphorylated Smad2 [40], and improves anti-glomerular basement membrane-induced nephritis in rats [41]. Therefore, we suggested that Rg 1 had a renoprotective effect against AAN in mice.
Proteomics is a powerful tool for systematic investigation of potential biomarkers and relative mechanisms. Renal differential proteomics is a powerful tool for systematic investigation of altered proteins, used to discover a variety of potential biomarkers. FD-LC-MS/MS is one of the HPLC methods that was well-established by Masuda et al. [42] and verified by various applications [10,43]. Several biofluids have been used in differential proteomics studies, including serum, urine, and tissue homogenate. Using these, renal differential proteomics is the most suitable and direct method for the discovery of the alteration proteins in renal injury models [10,44,45]. Several groups were successful in identifying potential biomarkers in several diseases using FD-LC-MS/MS [46,47]. Using untargeted LC-MS-based metabonomics can elucidate the mechanism of AA-induced testicular toxicity [48].
In our PLS-DA analysis, six observations were most important, including peaks 2, 4, 21, 23, 24, and 26. Among these, we chose mitochondrial ACO2 (peak 21) for further Western blot analysis to validate the FD-LC/MS/MS method. According to our results, ACO2 expression was significantly increased in the AA group and decreased in the AA with Rg 1 group. The patterns were consistent with the results of FD-LC-MS/MS. Our previous study also showed the same with Western blot analysis and FD-LC/MS/MS [10]. Therefore, FD-LC/MS/MS is suitable for the discovery of the potential biomarkers and mechanisms in the field of proteomics.
AA leads to renal glomerular defects through the activation of antioxidant enzymes, increasing oxidative stress, and causing mitochondrial dysfunction [6,35]. Mitochondrial dysfunction may be involved in AA-induced apoptosis in proximal tubular epithelial cells [8]. A recent study found that AA also can induce liver injury in rats through increasing the expression levels of apoptotic proteins caspase-9 and caspase-3, and can cause severe mitochondria damage [49]. Mitochondrial ACO2 is an enzyme that catalyzes the interconversion of citrate to isocitrate via cis-aconitate in the second step of the TCA cycle. Proteomic analysis of mitochondria from animal models, such as those of sepsis, diabetes, and aging, revealed that ACO2 is inactivated and nitrated [50][51][52]. One study showed that inactivation of ACO2 might prevent H 2 O 2 and superoxide formation via the mitochondrial respiratory chain [53]. Pozdzick et al. showed that AA tubulotoxicity leads to the downregulation of antioxidative enzymes and mitochondrial damage in rats [6]. Furthermore, Li et al. also demonstrated that the antioxidative capacity is significantly decreased in AAN mice compared with the normal mice [54]. Antioxidative compounds reducing AA-generated ROS resulted in attenuation of AA-induced cytotoxicity. Mitochondrial iron overload-mediated Nrf2-HO-1/GPX4 antioxidative system inhibition would assist AA-induced ferroptosis in renal tubular epithelial cells [7].
However, there are opposite opinions showing that ACO2 is upregulated during the early stage of type 1 diabetes in the rat [55] and downregulated in 1,2-(dichlorovinyl)-Lcysteine-induced renal injury in LLC-PK1 cells [56]. The gel-based proteomics showed that the expression of ACO2 increases at first and then decreases in the unilateral ureteral obstruction model [57]. These results showed that mitochondrial aconitase might be a primary target of oxidative stress in vivo. In the AA with Rg 1 group, the expression of ACO2 slightly decreased compared with the AA group. A study demonstrated that co-treatment with Rg 1 (2.5 µM) recovers the inactivation of mitochondrial ACO2 activity in oxidative stress-mediated neurotoxicity in vitro. Unfortunately, the expression of ACO2 was not determined [58]. To the best of our knowledge, our study is the first one to demonstrate that AA-induced renal injury is positively correlated with the upregulation of ACO2 in kidney tissues; the detailed mechanism of ACO2 in AAN still needs to be elucidated.
AA-induced oxidative stress in mice was associated with increases in NADPH oxidase 2 (NOX2) and CYP2E1 expression, and decreases in catalase, superoxide dismutase, and glutathione synthetase expression [59]. Additionally, AA can cause mtDNA depletion, respiratory chain defects, and lower ATP content, leading to impaired respiratory complex I activity [60]. AA can also cause a depolarization of mitochondrial membrane, release of cytochrome c, and an increase of caspase-3 activity in proximal tubular epithelial cells [61]. Rapamycin protects against AA induced nephropathy by activating the mTOR autophagy axis [11].The Usmg5, known as diabetes-associated protein in insulin-sensitive tissues (DAPIT), is a part of the mitochondrial ATP synthase [62]. In human embryonic kidney 293T cells, the over-expression of Usmg5 causes mitochondrial dysfunction and activates hypoxia-inducible factor 1α and Wnt/β-catenin signaling, which results in a shift of aerobic metabolism toward a more glycolytic direction, and epithelial to mesenchymal transition [63]. In other cases, suppression of Usmg5 in HeLa cells causes loss of ATP synthase in mitochondria [64]. Therefore, we speculated that Usmg5 might play an important role in mitochondrial ATP synthesis and might correlate with renal interstitial fibrosis in AAN mice.
Cytochrome c oxidase subunit Va (COX5a) is one of the subunits of mitochondrial respiratory chain complex IV. The downregulation of COX5a is observed in type 2 diabetic mice [64]. However, gel-based proteomics showed that upregulation of COX5a occurs in renal cell carcinomas [65]. In neonatal rat kidney, COX5a is upregulated after partial unilateral ureteral obstruction [66]. Another study showed that the expression of COX5a is upregulated in breast cancer tissues compared with healthy tissues [67]. In tumor cells, Bcl-2 interacts with COX5a, which results in the alteration of mitochondrial respiration under oxidative stress [68]. Although our study did not show any renal tumors, the upregulation of COX5a might correlate with the carcinogenic activity of AA.
The present study had some limitations. First, study limitations concerned design, inherent risk of bias, and small sample sizes. Second, using GE or GS in mice with AA -induced nephropathy should be verified to ensure the rationality of the experimental grouping, which is also a limitation that should be noted for future experiment group design. Third, the experiment lacked a mitochondrial function assay.

Conclusions
In this study, we demonstrated, for the first time, that mitochondria-related proteins (Usmg5, COX5a, and ACO2) are upregulated in renal tissues in AAN mice and downregulated following Rg 1 treatment. We speculate that the renoprotection of Rg 1 might be ascribable to the suppression of interstitial fibrosis (Usmg5), mitochondrial respiratory chain-related protein (COX5a), and oxidative stress (ACO2). However, the possible mechanisms of these proteins remain to be elucidated in the future.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/life11101018/s1. Figure S1: Box plots and kernel density plots before and after normalization. Figure S2: Light microscopy findings of renal tissue of normal group (a), AA group (b), and AA plus GE groups (125 mg/kg, 250 mg/kg, 500 mg/kg) (c, d, e). Tubulointerstitial histological scores (f). Figure S3: Quantification of Rg 1 , Rb 1 , and Rd in GE using HPLC method. Figure S4: Leave-one-out cross validation of PLS-DA model. Table S1: The effect of GE on clinical chemistry in chronic AAN mice. Table S2: The method validation for analyzing Rg 1 , Rb 1 , and Rd.   Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.