The Role of TLR4 on PGC-1α-Mediated Oxidative Stress in Tubular Cell in Diabetic Kidney Disease

The role and precise mechanism of TLR4 in mitochondria-related oxidative damage and apoptosis of renal tubules in diabetic kidney disease (DKD) remain unclear. We examined the expression of TLR4 in renal biopsy tissues. Db/db diabetic mice and HK-2 cells cultured under high glucose (HG) were used as in vivo and vitro models. Real-time RT-PCR, Western blot, and immunohistochemistry were performed to examine the mRNA and protein levels of TLR4, NF-κΒ, PGC-1α, cytochrome C, and cleaved caspase-3. ATP level, activity of electron transport chain complex III, and antioxidant enzymes were investigated for mitochondrial function. Electron microscopy (EM) and MitoTracker Red CMXRos were used for mitochondrial morphology alteration. DHE staining and TUNEL assay were detected for ROS accumulation and apoptosis. PGC-1α plasmids were used for the overexpression of PGC-1α in HK-2. TAK242 and parthenolide were used as TLR4 and NF-κB blockers, respectively. Results showed that TLR4 was extensively expressed in the renal tubules of DKD patients and db/db diabetic mice, which was positively related to the tubular interstitial damage score and urinary β-NAG levels. In diabetic mice, inhibition of TLR4 could reverse the decreased expression of PGC-1α, increased expression of cytochrome C and cleaved caspase-3, mitochondrial dysfunction and deformation, increased accumulation of ROS, and activation of tubular cell apoptosis. In vitro, inhibition of TLR4 or NF-κB showed consistent results. PGC-1α overexpression could reverse the mitochondrial dysfunction, increased cleaved caspase-3, and apoptosis in HK-2 cells treated with HG. Data indicated that the TLR4/NF-κB signaling pathway might be the upstream pathway of PGC-1α and promote the tubular damage of DKD by modulating the mitochondria-related oxidative damage and apoptosis.


Introduction
Toll-like receptors (TLRs) are pattern recognition receptors and play a fundamental role in the activation of innate and adaptive immune responses [1,2]. Among the 11 human TLRs, TLR4 has been implicated in the pathogenesis of acute and chronic renal disorders such as acute kidney injury (AKI), renal fibrosis, and DKD [3,4]. Further researches have reported that TLR4 knockout diabetic mice have reduced the expression of MyD88 and TRIF and decreased NF-κB activity and the release of inflammatory cytokines and renal fibrosis [5,6]. In addition, HG was proved to induce the overexpression of TLR4 through NF-κB-dependent signaling and lead to the accumulation of ROS in podocytes [7], indicating the important role of TLR4 NF-κB signaling in the mechanism of DKD progression. According to our previous study, renal tubular oxidative stress injury and apoptosis play a key role in the progression of DKD in high glucose (HG) conditions [8]. These let us speculate that activation of the TLR4-NF-κB signaling pathway might be involved in mitochondrial dysfunction and mitochondriarelated oxidative damage of renal tubular epithelial cell (RTEC) in hyperglycemia, which is gradually to have an extremely important effect in the progression of DKD.
PGC-1α is proved to stimulate mitochondrial biogenesis and respiration through the induction of uncoupling protein 2 (UCP-2) and the regulation of nuclear respiratory factors (NRFs) [10]. In addition, our previous study has also confirmed that, by adjusting transcription factors such as NRFs, PGC-1α could protect mitochondrial respiratory chain function and antioxidant enzymes, so as to maintain the stability of the mitochondrial structure and function [8]. Moreover, in cardiac cells, researchers found that NF-κB p65 represses PGC-1α activity leading to metabolic dysregulation that underlies heart dysfunction and failure [11]. However, the protective effect of PGC-1α on mitochondria and its relationship with TLR4/NF-κB signaling path in DKD are not fully clear and need to be further investigated.
Our aim of this study is to explore the function of the TLR4/NF-κB pathway in mitochondria-related oxidative damage and apoptosis of RTEC in hyperglycemia and to investigate the role of PGC-1α in the TLR4/NF-κB pathway in DKD.

TLR4 Expression Was Upregulated in Renal Biopsy
Specimens of DKD Patients. The clinical characteristics of the DKD patients and N-DKD as controls in this study are shown in Table 1. PASM and PAS staining showed morphological changes in both glomerular and tubulointerstitial areas, including mesangial area expansion (Figure 1(a), B, arrow), focal tubular atrophy, and interstitial fibrosis (Figure 1(a), D, arrow) in DKD patients, compared with those in non-DKD patients. Significantly, in the N-DKD group, mitochondria with an elongated cylindrical shape with organized cristae were shown by electron microscopy (Figure 1(a), G). However, diffused fragmented mitochondria were observed in DKD patients (Figure 1(a), H). The mitochondrial changes in EM were quantified (Figure 1(c)): mitochondrial length was measured in tubular cells to determine the percentage of cells that showed filamentous mitochondria less than 1% long (>2 μm). * P < 0 05 compared with the N-DKD group. An observably enhanced TLR4 expression was demonstrated by IHC staining in the renal tubules of DKD patients (Figures 1(a), F, and 1(b)). Correlation analysis showed that TLR4 expression was positively correlated with the interstitial fibrosis and tubular atrophy (IFTA) scores and urinary β-NAG level as a tubular injury marker (Figures 1(d) and 1(e)).

Inhibition of TLR4 Protects Tubular Cell by Regulating
Mitochondria-Related Proteins in Diabetic dbdb Mice. The levels of blood urea nitrogen (BUN), serum creatinine (Cr), urine protein (Upro), and urinary albumin : creatinine ratio (ACR) were significantly increased in the db/db group; * P < 0 05 compared with the db/m group. However, BUN, Cr, and Upro were significantly attenuated following treatment with TAK242 (Table 2), * * P < 0 05 compared with the db/ db mice group. These results suggested that TAK242 administration could preserve the renal function of db/db mice to a certain extent.
Loss of brush border and early tubular atrophy were observed compared with the control group by HE staining (Figure 2(a), A-C), which were ameliorated by the injection of TLR4 inhibitor TAK242. The urinary excretion of β-NAG, which is a marker of tubular damage, was reflected by a significant increase in diabetic dbdb mice, while it was substantially reduced by intrarenal injection of TAK242 ( Figure 2(b), B2).

Inhibition of TLR4 Protects Tubular Cell from Mitochondrial-Dependent
Apoptosis by Regulating Mitochondrial Structure and Function in Diabetic dbdb Mice. ROS production was stained with red fluorescence by ROS-sensitive vital dye DHE and increased notably in the tubules of diabetic dbdb mice. Under the inhibition of TLR4 expression, ROS generation was significantly reduced (Figure 3(a), A1, A-C, A2). In addition, the inhibition of TLR4 expression dramatically reduced the degree of apoptosis in the tubular cells of diabetic dbdb mice by TUNEL assay (Figure 3(a), A1, D-F, A3). Tubular cells show elongated mitochondria with organized cristae in dbm mice (Figure 3(a), A1, G) (marked by asterisks); however, in the dbdb group, most mitochondria exhibited spherical shapes and had cristolysis (Figure 3(a), A1, H), which was partly attenuated following treatment with TAK242 ( Figure 3(a), A1, I, A4).
The level of ATP production (Figure 3(b)), activity of electron transport chain complex III (Figure 3(c)), and activity of antioxidant enzymes: catalase (CAT) and manganese superoxide dismutase (MnSOD) (Figures 3(d) and 3(e)) were significantly decreased in the dbdb mouse group, which could be reversed following the injection of TAK242. These indicated that inhibition of TLR4 could attenuate the depression of mitochondrial functions in diabetic mice.

HG Increased TLR4 Expression and Activated NF-κB p65
Phosphorylation in HK-2 Cells under HG Ambience. As shown in Figure 4, the protein level of TLR4 increased in a Values are means ± SE; * P < 0 05, compared with N-DKD.
concentration-and time-dependent manner in HK-2 cells, and beta-actin served as a loading control (Figures 4(a) and 4(b)). In addition, the expression of phospho-NF-κB p65 increased significantly in HK-2 cells treated with 30 mM HG for 2 h compared to the control (Figure 4(c)), while TLR4 inhibitor (TAK242, 5 μM) [12] reversed HG-induced NF-κB p65 phosphorylation (Figure 4(e)) for 2 h and for 24 h (Figure 4(f)), which indicates that NF-κB is the downstream signal molecule of TLR4. Consistent with the protein expression, the mRNA level of NF-κB p65 increased in HK-2 cells treated with 30 mM HG for 2 h compared to the control, while TAK242 reversed HG-induced activation of NF-κB p65 phosphorylation (Figure 4(d)), which indicates that HG activated the TLR4/NF-κB signaling pathway. These data suggested that the TLR4/NF-κB pathway might be involved in HG-induced mitochondria-related ROS accumulation and apoptosis. In addition, the expression of PGC-1α increased in TLR4/NF-κB blocked groups suggesting that PGC-1α might be a downstream protein of the TLR4/NF-KB signaling pathway in HG-induced changes in HK-2 cells.  30 mM concentration of D-glucose, release of cytochrome C from mitochondria to cytoplasm was increased in HK-2 cells by Western blot (Figure 6(a), A1-A3), suggesting a remarkable HG-induced mitochondrial malfunction. With the intervention of the TLR4 inhibitor (TAK242) and NF-κB blocker (parthenolide), the translocation of cytochrome C was ameliorated ( Figure 6(a), A1-A3). The level of ATP production ( Figure 6(b)), activity of electron transport chain complex III (Figure 6(c)), and activity of CAT and MnSOD (Figures 6(d) and 6(e)) were significantly decreased in the HG group, which could be reversed in the TAK242 + 30 Glu group and parthenolide + 30 Glu group. These indicated that HG activated TLR4/NF-κB signaling to influence mitochondrial function.
As shown in Figure 6(f), mitochondria, which were stained by MitoTracker with red fluorescence, were filamentous with a thread-like appearance and were often interconnected to form a network in the control group ( Figure 6  Values are means ± SE; * P < 0 05, compared with the db/m group; * * P < 0 05, compared with the db/db group; urinary albumin : creatinine ratio (ACR). In addition, Hoechst 33258 staining shows that HG (30 mM) increased karyorrhexis, which was a hint of early apoptosis, while TAK242 and parthenolide treatment reversed this trend (Figure 6(g), arrow shown). These data demonstrated that HG induced mitochondrial malfunction and aggravated nuclear fragmentation of early apoptotic cells by activating the TLR4/NF-κB signaling pathway. Furthermore, the results also demonstrated that phospho-NF-κB p65 increased in HK-2 cells with 30 mM HG for 2 h (Figure 7(c)) and 24 h (Figure 7(d)), while it did not decrease in PGC-1α-overexpressed cells, which further proved that PGC-1α might be the downstream protein of the TLR4/NF-κB signaling pathway in HG-induced changes in HK-2 cells.

Overexpression of PGC-1α Restored HG-Induced
Mitochondrial Membrane Potential (△Ψ m ) Alteration and Apoptosis in HK-2 Cells under the Inhibition of the TLR4/ NF-κB Signaling Pathway. By flow cytometry analysis, a loss of mitochondrial △Ψ m detected by TMRM staining was observed under 30 mM HG ambience in cells undergoing early apoptosis, which was recovered to baseline with the overexpression of PGC-1α (Figures 8(a), A). In addition, the apoptotic rate was only 11.92% with 5 mM D-glucose in HK-2 cells, but peaked at 68.1% in the 30 Glu group (Figures 8(b), B), which was agreed with the reported literature that high glucose led to increased cell apoptosis. However, the apoptotic rate was, respectively, 52.27% in the "PGC-1α + 30 Glu group," 59.86% in the "PGC-1α + TAK242 + 30 Glu group," and 47.12% in the "PGC-1α + parthenolide + 30 Glu group," which were significantly decreased when compared with that of the 30 Glu group. Statistics indicated that PGC-1α overexpression inhibited HG-induced loss of mitochondrial △Ψ m and apoptosis in HK-2 cells.

Discussion
Recently, researches have demonstrated that mitochondriarelated oxidative damage and apoptosis play a key role among the multifactorial pathogenesis of DKD patients [13,14]. In addition, TLR4 has been previously described involving in hyperglycemia-induced inflammatory state of renal tubulus in vitro and vivo [15,16], but its role in oxidative damage and apoptosis in renal tubular cells in DKD remains unclear. This study describes a cascade event that links TLR4/ NF-κB activation to mitochondria-related oxidative damage and apoptosis through downregulation of PGC-1α in renal tubular cells under HG condition.
There are increasing evidences showing the significance of the TLR4 pathway in the development of DKD [17]. However, most of them were focusing on its effect related to the tubulointerstitial inflammatory response [4,12]. In this study, we found that TLR4 was extensively expressed in tubular cells in the kidney of patients with DKD and was coexistent with fragmented mitochondria. Further analysis revealed a positive correlation between TLR4 expression and tubular injury. These results indicated that TLR4 might play an important role in mitochondriarelated tubular oxidative damage in DKD, besides its activation effect of inflammatory response. This speculation was further identified by a diabetic dbdb mouse model in our study.
Many studies have reported TLR4 to be critical for the activation of NF-κB and subsequent production of proinflammatory cytokines implicated in diseases [18,19]. Molecular silencing of TLR4 in tubular cells with siRNA attenuated HG-induced IκB/NF-κB activation, indicating that the TLR4-NF-κB signal pathway plays an important role in diabetic nephropathy [20,21]. In this study, we have also proved that the TLR4 inhibitor could effectively decrease the expression of phospho-NF-κB p65 increased in HK-2 cells under HG conditions, indicating that NF-κB is the downstream signal molecule of TLR4 in tubular cells.
Mitochondrial dysfunction has proved to be a contributing factor in the pathogenesis of DKD [22]. Mitochondria-related oxidative damage could result in the release of mitochondrial cytochrome c and activation of caspase-3, leading ultimately to apoptosis [23]. In order to investigate the mechanism for the effect of TLR4/NF-κB on the mitochondria-related tubular oxidative damage in hyperglycemia, we inhibited the expression of TLR4 or NF-κB in vivo and in vitro. We observed a partial rescue of HG-induced deformation of mitochondria, inhibition of ATP production, depressed mitochondrial respiration and activity of antioxidant enzymes, and increased caspase-3 and cytoplasm cytochrome c, which indicated that inhibition of the TLR4/ NF-κB signaling pathway could protect mitochondrial Researches have identified mitochondrial fragmentation as a novel mechanism contributing to mitochondrial damage and apoptosis in vivo in mouse models of disease [24]. In tubular cells, a significant portion of mitochondria line up perpendicular to the basement membrane, making an excellent model for studying mitochondrial fragmentation in vivo. Cross sections of tubules normally show 10%-20% longitudinally sectioned long mitochondria, while they appear as short rods or spherical fragments in pathological apoptosis in     disease models. Fragmented mitochondria were found to refuse if the injurious stress is removed before permanent injury happens to trigger apoptosis [25]. In this study, we observed that most of the mitochondria within the injured tubules in renal biopsy tissues and diabetic mice were fragmented and had a randomly disorganized cellular distribution, which could be reversed by inhibition of the TLR4/ NF-κB inhibitor, indicating that the TLR4/NF-κB signaling pathway induced mitochondrial fragmentation, which contributes to subsequent tubular cell apoptosis. As a main regulator of mitochondrial function, PGC-1α exerts a rescuing effect in substance metabolism and oxidative metabolism by regulating mitochondrial biogenesis and ROS scavenging enzymes [26]. An increased PGC-1α expression may protect cells from oxidative stress and oxidative stress-mediated cell apoptosis [27,28]. In cardiac myocytes, Schilling et al. showed that TLR4 activation could lead to the phosphorylation and nuclear translocation of NF-κB, triggering cardiac energy metabolic reprogramming by repressing genes encoding PGC-1α, and the absence of TLR4 abolished LPS-induced downregulation of PGC-1α, indicating that the suppression of PGC-1α was shown to occur through a TLR4-and NF-κB-dependent mechanism [29]. In this study, we observed that PGC-1α was decreased in a diabetic mouse model and was reversed after the inhibition of TLR4 in vivo. In HK-2 cells, overexpression of PGC-1α inhibited a series of HG-induced changes including downregulation of mitochondrial membrane potential, upregulation of apoptosis-related protein cleaved caspase-3, and cell apoptosis directly, verifying the protective function of PGC-1α to mitochondrial function and cell survival in HG condition. Moreover, PGC-1α overexpression did not decrease HG-induced TLR4 activation and NF-κB p65 phosphorylation, further confirming that PGC-1α is a downstream protein of TLR4/NF-κB, which protected renal tubular cells from HG-induced mitochondria-related oxidation and apoptosis.
In conclusion, TLR4 plays a significant role in HGinduced mitochondrial dysfunction, mitochondria-related oxidation, and apoptosis by regulating downstream protein PGC-1α in RTEC, hoping to provide a better understanding and a more effective therapeutic approach for the prevention and treatment of DKD.

4.2.
Morphological Analysis of Kidney. 12 patients with DKD and 12 non-DKD controls (normal kidney tissue) were recruited for this study. Human renal biopsy tissues from the 24 cases were studied by staining of PAS and PASM. A semiquantitative scoring system was used to evaluate the tubulointerstitial lesion index, and tubular damage was also scored [30,31]. All procedures were carried out in accordance with the approved guidelines. All patients did not use adrenal cortical hormones or immunosuppression. The institutional review board and the administrators of the Department of Nephrology in The Second Xiangya Hospital approved the protocol for this study. An informed consent was obtained from all the participants.
The kidney tissue of mice was routinely processed, embedded in paraffin, and sectioned at 2-3 mm thickness, deparaffinized, and rehydrated using standard techniques. Mouse sections were stained with hematoxylin-eosin stain (HE).

Examination of Mitochondrial Fragmentation in Renal
Tissue and HK-2 Cells. The alterations of mitochondria in renal tubules were gauged by electron microscopy (EM). Mitochondria having a length < 1 μm and spherical configuration were identified as fragmented. We determined the percentage of cells that had less than 1% long filamentous mitochondria to indicate the degree of mitochondrial fragmentation in patients and mice [24].
MitoTracker Red CMXRos was used for the evaluation of mitochondrial morphology in HK-2 cells. The mitochondria within a cell were often either filamentous or fragmented. In cases of mixed mitochondrial morphology, we classified the cells based on the majority (>70%) of mitochondria, according to earlier studies. For each sample, several random fields of cells (≥100 cells per dish) were evaluated [32,33].

Animal Experimental Design.
A total of 10 male dbm mice and 20 adult male dbdb mice at 16 weeks of age (body weight 32-40 g) were divided into three groups of 10 animals each. The first group was male dbm mice, which served as a control. The second group of dbdb mice received an intraperitoneal injection with vehicle alone (dbdb group). The third group included dbdb mice which received an intraperitoneal injection of TLR4 inhibitor TAK242 (3 mg/kg for 7 days). All animals were killed at 17 weeks following administration. The Institutional Animal Experimentation Ethics Committee as described above approved the animal experimental protocols.
4.6. Immunohistochemistry (IHC). Renal tissue sections from human and mice for immunostaining were deparaffinized and rehydrated. Immunohistochemistry was performed using anti-TLR4 antibody (1 : 100, Abcam, USA), caspase-3 (1 : 100, CST, Boston, USA), and cytochrome c (1 : 100, CST, Boston, USA) antibody as a primary antibody followed by a secondary antibody. Then, slides were visualized by using a DAB detection kit according to the manufacturer's instructions, and the tissue specimens were examined by light microscopy.
For human and mouse tissue sections, the average intensity from at least 20 randomly selected fields was measured using ImageJ software (National Institutes of Health, Bethesda, MD). 4.7. ATP Assay. ATP levels were determined in cell lysates obtained from mouse renal tissue and HK-2 cells using an ATP Assay Kit (Genmed, Shanghai, China). The assays were run according to the manufacturer's instructions. The values were expressed as fold change compared to the control.

Respiratory
Chain Complex III Activity. Mitochondrial respiratory chain complex III activity was measured by using a spectrophotometer (Genmed, Shanghai, China) according to the manufacturer's instruction. The values were expressed as fold change compared to vehicle control. 4.9. Mitochondrial Enzyme Activities. The Superoxide Dismutase Activity Assay Kit (Alexis Biochemicals) and Catalase Activity Colorimetric/Fluorometric Assay Kit (Bio-Vision Inc.) were used for the detection of Mn-SOD and catalase activity, following the manufacturer's guidelines. Enzyme activities were displayed as fold change compared to the control.

Measurements of Superoxide Generation and Apoptosis.
DHE was used for the detection of mitochondrial superoxide generation in vivo. A specific mitochondrial superoxide indicator MitoSOX red (Molecular Probes) was used for the detection of ROS production in HK-2 cells in vitro. 20 randomly selected fields were photographed for tissue sections, and the mean fluorescence intensity was calculated by using NIH ImageJ software and was expressed relative to the control (set as 1).
The TUNEL procedure and Hoechst 33258 staining were used to detect apoptosis following the manufacturer's instructions. 10 random fields of cells (approximately 100 cells per group) were counted to determine the percentage of cells undergoing apoptosis.
4.14. Assessment of Mitochondrial Membrane Potential (△ Ψ m ). Mitochondrial △Ψ m of cells was assessed by TMRM Detection Kit (Genmed Scientifics Inc., USA). HK-2 cells plated on 24-well culture dishes were harvested after trypsinization when they were seeded at 70% confluence, and then the cells were stained with TMRM dyeing liquid according to the manufacturer's instruction. For 20 min incubation in the dark at 37°C, mitochondrial △Ψ m of cells was examined by a FACSCalibur flow cytometer (BD Biosciences, San Jose, USA). The fluorescence intensity of TMRM was monitored at 575 nm.

Data Analysis.
Statistical analysis was carried out using the SPSS 20 software. Results were expressed as mean value s ± standard error of the mean (SEM). Statistical differences among groups were analyzed by one-way ANOVA, and two-tailed P values are reported. P values less than 0.05 were considered statistically significant.

Disclosure
The authors alone are responsible for the content and writing of the article.

Conflicts of Interest
The authors declare no competing financial interests.