Rosiglitazone Suppresses Renal Crystal Deposition by Ameliorating Tubular Injury Resulted from Oxidative Stress and Inflammatory Response via Promoting the Nrf2/HO-1 Pathway and Shifting Macrophage Polarization

Oxidative stress and inflammatory response are closely related to nephrolithiasis. This study is aimed at exploring whether rosiglitazone (ROSI), a regulator of macrophage (Mp) polarization, could reduce renal calcium oxalate (CaOx) deposition by ameliorating oxidative stress and inflammatory response. Male C57 mice were equally and randomly divided into 7 groups. Kidney sections were collected on day 5 or day 8 after treatment. Pizzolato staining and polarized light optical microscopy were used to detect crystal deposition. PAS staining and TUNEL assay were performed to assess the tubular injury and cell apoptosis, respectively. Gene expression was assessed by immunohistochemistry, immunofluorescence, ELISA, qRT-PCR, and Western blot. The reactive oxygen species (ROS) level was assessed using a fluorescence microplate and fluorescence microscope. Hydrogen peroxide (H2O2), malonaldehyde (MDA), and glutathione (GSH) were evaluated to determine oxidative stress. Lactic dehydrogenase (LDH) activity was examined to detect cell injury. Adhesion of CaOx monohydrate (COM) crystals to HK-2 cells was detected by crystal adhesion assay. HK-2 cell death or renal macrophage polarization was assessed by flow cytometry. In vivo, renal crystal deposition, tubular injury, crystal adhesion, cell apoptosis, oxidative stress, and inflammatory response were significantly increased in the 7-day glyoxylic acid- (Gly-) treated group but were decreased in the ROSI-treated groups, especially in the groups pretreated with ROSI. Moreover, ROSI significantly reduced renal Mp aggregation and M1Mp polarization but significantly enhanced renal M2Mp polarization. In vitro, ROSI significantly suppressed renal injury, apoptosis, and crystal adhesion of HK-2 cells and markedly shifted COM-stimulated M1Mps to M2Mps, presenting an anti-inflammatory effect. Furthermore, ROSI significantly suppressed oxidative stress by promoting the Nrf2/HO-1 pathway in HK-2 cells. These findings indicate that ROSI could ameliorate renal tubular injury that resulted from oxidative stress and inflammatory response by suppressing M1Mp polarization and promoting M2Mp polarization. Therefore, ROSI is a potential therapeutic and preventive drug for CaOx nephrolithiasis.


Introduction
Urolithiasis has a reported prevalence of 10% and 50% recurrence rates [1], potentially resulting in chronic kidney disease or even renal failure [2]. In addition, it has been estimated that, by 2030, the annual cost for managing urolithiasis in the United States is estimated to be about 4.6 billion dollars [3].
In the process of calcium oxalate (CaOx) crystal deposition, injury to the tubular epitheliums [4], which is frequently mediated by oxidative stress, plays a critical role during CaOx crystal deposition [5]. CaOx crystals upregulate NADPH oxidase p47phox and lead to overproduction of reactive oxygen species (ROS) in the tubular epitheliums. This is further enhanced by the proinflammatory effect of macrophages [6]. ROS are the leading mediators of oxidative stress injury, damaging the mitochondrial membrane and reducing the transmembrane potential [7,8]. In addition, damaged tubular cells can act as the adhesive site for crystals, increasing their adhesion [9]. Therefore, CaOx crystalinduced oxidative stress is a critical process for nephrolithiasis formation.
Inflammation is another crucial process in nephrolithiasis formation. Following exposure to crystals, renal epitheliums express increased levels of monocyte chemotactic protein-1 (MCP1), inducing substantial Mp recruitment [10]. Mps play pivotal roles in inflammation and display different functional roles when responding to diverse microenvironmental signals [11,12]. Two main phenotypes of Mps have been found: classically activated M1Mps, which are proinflammatory, and alternatively activated M2Mps, known to demonstrate an anti-inflammatory effect. CaOx crystals promoted M1Mp polarization which worsens the renal condition, resulting in fibrosis and chronic kidney disease [13,14]. M1Mps reportedly promote renal crystal deposition in the mouse model [5]. On the contrary, stimulating M2Mp polarization could reduce renal injury [15][16][17]. Moreover, Mps have plasticity, by which polarized M1Mps can shift to M2Mps by employing specific signals [18]. Consequently, the M1/M2 Mp phenotype shift plays a pivotal role in regulating inflammation.
Peroxisome proliferator-activated receptor γ (PPARγ) is a nuclear receptor and inflammatory regulator [22]. PPARγ activation plays an anti-inflammatory effect by repressing intranuclear signaling pathways, containing nuclear factor-κB (NF-κB) in Mps [23], and has an antioxidant effect [24]. Moreover, PPARγ activation is critical for the polarization and maintenance of M2Mps [25,26]. These characteristics indicate that PPARγ agonists might be a promising therapeutic option for nephrolithiasis.
However, data regarding the antioxidant and antiinflammatory activities of PPARγ agonist ROSI are greatly limited. This study revealed that ROSI could suppress crystal deposition in the kidney of hyperoxaluric mice by inhibiting oxidative stress and inflammatory response.

Materials and Methods
2.1. Reagents. Glyoxylic acid (Gly) was acquired from Sigma-Aldrich (St. Louis, MO). COM crystals were prepared by the chemical method as described previously [27] and used in in vitro experiments at a concentration of 300 μg/mL. PPARγ agonist ROSI and antagonist GW9662 were both procured from MedChemExpress (MCE, China).

Animal Experiments.
All animal experiments fulfilled the criteria of NIH and Guizhou Provincial People's Hospital for the humane treatment of laboratory animals and were approved by the Animal Care and Use Committee of Guizhou Provincial People's Hospital. Male C57BL/6 mice aged 6-8 weeks were acquired from the Experimental Animal Center and reared in SPF animal facilities at Guizhou Provincial People's Hospital.
In order to assess the effects of ROSI, 42 mice were equally assigned to 7 groups: the control group, 4 days  3 Oxidative Medicine and Cellular Longevity 80 mg/kg Gly (Gly 4d) group, 7 days 80 mg/kg Gly (Gly 7d) group, 7 days 80 mg/kg Gly plus 2.5 mg/kg or 5 mg/kg ROSI without pretreatment of ROSI (Gly+ROSI 7d(L)) group, Gly +ROSI 7d(H)) group, and 7 days 80 mg/kg Gly plus 2.5 mg/kg or 5 mg/kg ROSI with pretreatment of 3 days 2.5 mg/kg or 5 mg/kg ROSI (ROSI 3d+Gly+ROSI 7d(L), and ROSI 3d+Gly+ROSI 7d(H)) group. Mice were given a daily intraabdominal injection with Gly or vehicle (phosphate-buffered saline (PBS), Gibco) for 4 or 7 days. ROSI or vehicle (PBS) was administered via the gastric tube once daily. The kidneys were acquired on day 5 or day 8 after treatment to detect crystal deposition and for other related experiments.
2.3. Cell Culture and Treatment. THP-1 and HK-2 cells were purchased from the Cell Bank of the Chinese Academy of Science (Shanghai, China). THP-1 cells were cultured in RPMI-1640 (Gibco, USA) conditioned medium, and HK-2 cells were cultured in DMEM/F12 (Gibco, USA) conditioned medium containing 10% FBS and 1% penicillinstreptomycin in an incubator at 37°C and 5% CO 2 .
For differentiating into M0Mps, THP-1 cells were diluted into a density of 2 × 10 5 /mL and treated with 10 ng/mL phorbol-12-myristate-13-acetate (Sigma-Aldrich, USA) for 24 h. After removing the supernatant, the cells were washed thrice with PBS and cultured in RPMI-1640 conditioned medium for subsequent experiments.

Observation of the Deposition of Renal CaOx Crystal.
Crystal deposition in sections was examined through Pizzolato staining, as presented previously [28]. Crystal containing calcium was determined through polarized light optical microscopy (CX31P; Olympus, Japan) in unstained sections. Crystal deposition was assessed quantitatively by ImageJ (National Institute of Health, USA) to calculate the percentage of the crystal deposition area in the entire kidney section or corticomedullary border.
2.5. Immunohistochemistry (IHC) and Immunofluorescence (IF). Briefly, kidney samples were fixed for 24 h with 4% buffered formalin before embedding into paraffin. Subsequently, sections at 4 μm thick were acquired. For IHC, the slides were incubated within antibodies for PPARγ Immunoreactivity was examined by Histofine Simple Stain Kit for rabbit IgG following the protocol from the manufacturer. The Mp phenotypes were determined by IF staining for iNOS (1 : 100, 13120, CST) and Arg1 (1 : 100, 93668, CST). The slides of THP-1 cells were fixed for 15 minutes in 4% buffered formalin. After being rinsed with PBS and blocked with goat serum, the slides were incubated in primary antibodies at 4°C for over 12 h. Rinsed with PBS,   5 Oxidative Medicine and Cellular Longevity the slides were incubated in Alexa Fluor-conjugated CY3-or 488-secondary antibodies (1 : 5000, G-21234, Thermo Fisher) for 1 h at room temperature. Finally, after being rinsed with PBS, the slides were counterstained with the nuclear marker DAPI and wet mounted. All images were obtained through a fluorescence microscope (Nikon TE2000-U, Japan), and the fluorescence intensity was quantified using ImageJ.
2.6. Tubular Injury and Cell Apoptosis. PAS staining was used to detect the cellular injury of renal tubules. The percentage of injured tubules was determined in 10 random fields at ×400 magnification in each section. Additionally, TUNEL assay was conducted to examine apoptotic cells in the renal tissue using the In Situ Cell Death Detection Kit (Roche, Switzerland). Positive cells in the TUNEL assay were determined in 10 random fields at ×400 magnification for every section.

qRT-PCR.
Total RNA was acquired from THP-1 and HK-2 cells by TRIzol (Invitrogen, USA). Then, cDNAs were synthesized from 2 μg of total RNA by the PrimeScript RT Reagent Kit (TaKaRa). qPCR was conducted by SYBR green qPCR master mix (QIAGEN, Germany) via an ABI Prism 7300 system. All the reactions were triplicated. Target gene expression levels were quantified using the double-delta method (2 -ΔΔCt ) for 3 independent experiments with normalization. Primers (provided by TSINGKE) used in experiments are listed in Table 1. The membranes were then incubated in HRP-conjugated secondary antibodies (1 : 3000, 7074, CST) for 1 h at room temperature. We used the ECL Western Blot Kit (Thermo Scientific Pierce) to examine the bands and scanned them using a LAS4000 analyzer (GE Healthcare). The immunoblot density was examined by ImageJ and normalized by β-actin or GAPDH.
2.9. Enzyme-Linked Immunosorbent Assay (ELISA). Cell supernatants were obtained after centrifugation and were stored at −80°C until further use. Secretion levels of TNFα, IL-4, IL-6, and IL-10 were detected via specific ELISA kits, according to the instructions from the manufacturer (Dakewe, Shenzhen, China).      For quantifying the renal macrophage polarization in mouse kidneys, kidney tissues were minced into 1 mm 3 fragments and then digested in RPMI 1640 buffer containing 100 U/mL DNase I and 2 mg/mL collagenase type D for 60 min at 37°C and then passed through a 70 μm mesh to get single-cell suspension. Red blood cell lysis buffer (Sigma, USA) was used for lysing the red blood cells in the suspension. Mps were centrifuged and then resuspended in FACS buffer on ice. Incubated with 2.5 μg/mL Fc-blocking solu-tion, Mps were resuspended in FACS buffer. Then, 10 6 cells were stained with 3 fluorochrome-labeled antibodies: F4/80 (eBioscience)-PE, CD11c-FITC, and CD206-FITC. Finally, Mps were detected immediately on a FACS Canto II cytometer with DIVA software (Becton Dickinson). The data were analyzed by Cyflogic V.1.2.1 software.
2.14. Crystal Adhesion Assay. HK-2 cells were cultured to 100% confluency in a 6-well plate. Stimulated with COM crystal and ROSI or/and GW9662 treatment for 48 h, the plate was thoroughly washed 3 times using PBS to remove unbound crystals from cells. The crystal quantity was examined using a microscope. Images were randomly selected from 10 visual fields (magnification of ×400) and quantified using ImageJ Pro Plus software [29]. All experimental group data were normalized to the normal control group based on 3 independently repeated experiments. 2.16. Statistical Analysis. Data were presented as mean ± standard deviation (SD). A two-tailed t-test was conducted to identify statistical difference using GraphPad Prism 6.0    1(a) and 1(b)). Crystals in the entire kidney or renal corticomedullary borders were significantly larger and more in the Gly 7d group than in the Gly 4d group. Notably, crystals were significantly decreased in ROSI treatment groups than in the Gly 7d group in a dose-dependent way. Furthermore, there were significantly fewer and smaller crystals in the ROSI pretreatment groups than in the nonpretreatment groups (Figure 1(c)).
IHC results suggested that the expression levels of the crystal-related gene osteopontin (OPN) and crystal adhesion-related gene CD44 were markedly increased in

11
Oxidative Medicine and Cellular Longevity the Gly 4d group than in the control group, increased in the Gly 7d group than in the Gly 4d group, dose dependently decreased in the ROSI treatment groups than in the Gly 7d group, and decreased in the pretreatment groups than in the nonpretreatment groups (Figures 1(d) and 1(e)).

ROSI Decreased Renal
Cell Apoptosis, Tubular Injury, and Proinflammatory Response in the Mouse Model. For all groups, PAS staining showed that positively stained cells were mainly located in the renal tubules. Positive tubules were significantly fewer in the Gly 7d group and markedly more in the ROSI-treated groups than in the Gly 7d group in a dose-dependent way. In addition, positively stained tubules were substantially more in the pretreatment groups than in the nonpretreatment groups (Figures 2(a) and 2(c)).
TUNEL assay results revealed that positive cells were significantly more in the Gly 7d group, but fewer in the ROSI treatment groups than in the Gly 7d group, presenting a dose-dependent reduction. Moreover, positive cells were significantly fewer in the ROSI pretreatment groups than in the nonpretreatment groups (Figures 2(b) and 2(d)).
The expression of renal PPARγ was markedly and dose dependently increased in the ROSI treatment groups than in the Gly 7d group, and PPARγ expression was markedly increased in the ROSI pretreatment groups than in the nonpretreatment groups. The expression levels of the Mps-

12
Oxidative Medicine and Cellular Longevity 13 Oxidative Medicine and Cellular Longevity related molecule MCP1 and proinflammatory cytokine IL-1β were markedly increased in the Gly 4d and 7d groups; however, they were markedly decreased in the ROSI treatment groups in a dose-dependent manner. A significant decrease in expression levels of MCP1 and IL-1β was also noted in the pretreatment groups (vs. nonpretreatment groups) (Figures 2(e) and 2(f)).

ROSI Decreased Renal Oxidative Stress in the Mouse
Model. Our results demonstrated that the expression of Nrf2, HO-1, SOD1, and GSH markedly increased in the Gly 4d group than in the control group, decreased in the Gly 7d group than in the Gly 4d group, dose dependently increased in the ROSI treatment groups than in the Gly 7d group, and increased in the pretreatment groups than in the nonpretreatment groups. The expression of PPARγ was markedly and dose dependently increased in the ROSI treatment groups than in the Gly 7d group and was markedly increased in the ROSI pretreatment groups than in the nonpretreatment groups (Figure 3).
The generations of ROS, H 2 O 2 , and MDA markedly increased in the Gly 4d group than in the control group, increased in the Gly 7d group than in the Gly 4d group, dose dependently decreased in the ROSI treatment groups than in the Gly 7d group, and decreased in the pretreatment groups than in the nonpretreatment groups (Figure 3(c)).

ROSI Suppressed Oxidative Stress Injury via
Promoting the Nrf2/HO-1 Pathway in HK-2 Cell. For exploring the effect of ROSI on tubular epitheliums, we carried out cell experiments. The results revealed that LDH activity and death of HK-2 cells markedly increased in the COM group compared to the control group. Moreover, LDH activity and cell death markedly decreased after simultaneous ROSI treatment than in the COM group and increased in the GW9662 treatment group than in the ROSI group. In addition, Nrf2 silencing by siRNA also attenuated the effect of ROSI treatment (Figures 4(a) and 4(b)).
In HK-2 cells, expression of PPARγ markedly increased in the ROSI treatment group and markedly decreased in the GW9662-treated group, whereas it was not significantly affected by Nrf2-siRNA treatment. Genetic expression of Nrf2, HO-1, SOD1, and GSH markedly increased in the COM group than in the control group, increased in the ROSI group than in the COM group, and decreased in the GW9662 and Nrf2-siRNA treatment groups than in the ROSI group (Figures 4(c), 4(d), and 4(f)).
Furthermore, the productions of ROS, H 2 O 2 , and MDA in HK-2 cells markedly increased in the COM group than in the control group, decreased in the ROSI treatment group than in the COM group, and increased in the GW9662 and Nrf2-siRNA treatment groups than in the ROSI group (Figures 4(e) and 4(f)).

ROSI Decreased Macrophage Recruitment and
Polarization of M1Mps but Increased Polarization of M2Mps in the Mouse Model. IHC showed that expression levels of Mp marker F4/80, M1Mp marker iNOS, and M2Mp marker Arg1 were markedly increased in the Gly 4d and 7d groups. However, these levels were markedly decreased in the ROSI treatment groups in a dosedependent way, with a significant reduction observed in the ROSI pretreatment groups than in the nonpretreatment groups (Figures 5(a) and 5(b)).
Interestingly, the ratio of Arg1/F4/80 markedly decreased in the Gly 4d and 7d groups than in the control group, with a dose-dependent increase observed in the ROSI-treated groups. In addition, the ratio of Arg/F4/80 was higher in the ROSI pretreatment groups than in the nonpretreatment groups. Meanwhile, the ratio of iNOS/F4/80 demonstrated the opposite trend ( Figure 5(b)).
Furthermore, the flow cytometry results of renal macrophage polarization showed that the proportion of M1Mps markedly increased in the Gly 4d and 7d groups than in the control group, with a dose-dependent decrease in the ROSI treatment groups. Meanwhile, the proportion of    that ROSI significantly decreased the expression of PPARγ, M1Mps markers (iNOS and CD11c), proinflammatory cytokines (Ccl2, IL-6, and TNF-α), phosphorylated NF-κB and IL-1β, but increased the expression of M2Mp markers (Arg1 and CD206) and anti-inflammatory cytokines (IL-4 and IL-10). However, GW9662 diminished the effects of ROSI (Figures 6 and 7). Moreover, as for the COM crystal adhesion to HK-2 cells, ROSI significantly decreased crystal adhesion, whereas GW9662 diminished the effect of ROSI (Figures 6(d) and 6(e)).

Discussion
Crystal deposition may result in tubular injury [30], which in turn promotes crystal deposition [31]. In the above process, oxidative stress and inflammatory response play pivotal roles [32]. Being exposed to CaOx crystals, tubular epitheliums overproduce oxidative stress products, such as ROS [6]. As a leading mediator of oxidative stress, ROS leads to tubular injury, promoting the adhesion and deposition of crystals [9]. Nrf2 is a redox-sensitive transcription factor, playing critical roles in reducing intracellular oxidative stress and tissue injury [33]. Previous reports have revealed that the Nrf2/HO-1 pathway has potent antioxidant effects [21], whereby inhibiting the formation of CaOx-induced nephrolithiasis [34,35].
Deposition of THE CaOx crystal can pose migration and aggregation of Mps to the areas of crystal deposition [36].
Polarization of Mps is deeply involved in inflammation regulation, crystal phagocytosis, and crystal removal [37]. A previous study has revealed that CaOx crystals effectively induce M1Mp polarization, promoting renal injury and crystal deposition [38]. Furthermore, Taguchi et al. have observed that M2Mps have the ability of crystal phagocytosis and antiadherence [17]. As an agonist of PPARγ, ROSI takes a vital part in the intricate modulatory network of alternative activation of Mps [26]. Tabassum and Mahboob found that ROSI could reduce HFD-induced oxidative stress damage [39]. Deng et al. revealed that ROSI treatment inhibited renal macrophage infiltration and TGF-β and NF-κB pathway activation [40]. Liu et al. found that ROSI could regulate the TGF-β1 and HGF/c-Met pathways to play an antioxidant effect to reduce crystal deposition in the hyperoxaluria rat model [41]. In addition, previous studies have suggested that pioglitazone, another PPARγ agonist, can suppress crystal deposition, thus exhibiting prominent antiinflammatory ability [23,42,43].
In this study, we observed that ROSI could suppress oxidative stress injury by upregulating the Nrf2/HO-1 pathway. Moreover, we noted that ROSI could regulate inflammation by shifting COM-stimulated M1Mp polarization toward M2Mp polarization. Namely, the above effects might suppress tubular injury synergistically, consequently reducing the deposition of crystal.
As the above speculated, our results indicated that the deposited crystals were fewer and smaller and the renal injury was also milder in the ROSI-treated groups; these effects were more pronounced in the groups with ROSI pretreatment. Furthermore, the apoptosis of HK-2 cells was significantly reduced after treating with ROSI.
Moreover, our finding revealed that the generation of oxidative stress products, i.e., ROS, MDA, and H 2 O 2 , was significantly reduced. Conversely, expression levels of antioxidant products, i.e., SOD1 and GSH, were significantly upregulated after treating with ROSI. In addition, the effects of ROSI for suppressing oxidative stress and inflammation were found to be dose dependent both in vivo and in vitro.
Finally, in ROSI-treated hyperoxaluria mice, the expression of renal PPARγ was markedly increased. Subsequently, the MCP1 expression and Mp recruitment reduced with the decrease in M1Mps and the increase in M2Mps. The expression of iNOS was increased, but the expression of Arg1 was decreased in the Gly 7d group, which is similar to findings of a microarray analysis of renal papillary tissues from nephrolithiasis formers [13]. Interestingly, the renal Mps mainly were polarized into M1Mps after being stimulated by CaOx crystals. In contrast, the proportion of M2Mps increased in the ROSI-treated groups, reaching up to 70%. Accordingly, ROSI may pose a shift of Mp polarization. In vitro, ROSI also exhibited an anti-inflammatory ability due to shifting polarization of Mps.
We speculate that ROSI upregulates the expression of PPARγ, whereby it exerts antioxidant and antiinflammatory functions, which subsequently synergistically suppressed CaOx crystal deposition and renal injury. In terms of oxidative stress, ROSI potentially improves the imbalance between oxidative and antioxidant products by