PGC-1α attenuates hydrogen peroxide-induced apoptotic cell death by upregulating Nrf-2 via GSK3β inactivation mediated by activated p38 in HK-2 Cells

Ischemia/reperfusion injury triggers acute kidney injury (AKI) by aggravating oxidative stress mediated mitochondria dysfunction. The peroxisome proliferator-activated receptor gamma coactivator 1α (PGC-1α) is a master player that regulates mitochondrial biogenesis and the antioxidant response. We postulated that PGC-1α functions as cytoprotective effector in renal cells and that its regulation mechanism is coordinated by nuclear factor erythroid 2-related factor 2 (Nrf-2). In this study, to understand the effect and molecular mechanisms of PGC-1α, we developed an empty vector or PGC-1α-overexpressing stable cell lines in HK-2 cells (Mock or PGC-1α stable cells). PGC-1α overexpression increased the viability of cells affected by H2O2 mediated injury, protected against H2O2-mediated apoptotic events and inhibited reactive oxygen species accumulation in the cytosol and mitochondria as compared to that in Mock cells. The cytoprotective effect of PGC-1α was related to Nrf-2 upregulation, which was counteracted by Nrf-2-specific knockdown. Using inhibitor of p38, we found that regulation of the p38/glycogen synthase kinase 3β (GSK3β)/Nrf-2 axis was involved in the protective effects of PGC-1α. Taken together, we suggest that PGC-1α protects human renal tubule cells from H2O2-mediated apoptotic injury by upregulating Nrf-2 via GSK3β inactivation mediated by activated p38.


PGC-1α was downregulated in the I/R-injured kidney, as well as in H 2 O 2 treated HK-2 cells.
To investigate the involvement of PGC-1α in I/R-induced AKI, we analyzed the PGC-1α expression pattern in I/R induced mouse model. Groups that were subjected to renal I/R (n = 4) showed marked deterioration of renal functional parameters along with significant increase in the plasma creatinine level (sCr) and blood urea nitrogen (BUN), as compared to the finding for the controls (n = 4; Fig. 1A). We then performed a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay to determine the degree of renal tubular apoptosis in I/R-injured kidney. We found an increased number of tubular epithelial cells with TUNEL-positive nuclei in I/R-injured kidney (Fig. 1B). The levels of apoptotic proteins, for example, the Bax to Bcl2 ratio and the cleaved caspase 3 to caspase 3 ratio, also increased in the I/R group (Fig. 1C). The mRNA level and protein level of PGC-1α were lower in the I/R-injured kidney group as compared to those in the control group (Fig. 1D,E). We assessed the effect of PGC-1α in HK-2 cells under oxidative stress condition. To mimic I/R stress in HK-2 cells, we treated them with H 2 O 2 , which is the main culprit in the pathogenesis of I/R injury 30 . The PGC-1α expression level in H 2 O 2 -treated HK-2 cells was gradually decreased under condition of concentrations (0.5 mM) and duration (3 h) of treatment with H 2 O 2 ( Fig. 2A,B). We further assessed whether H 2 O 2 -induced PGC-1α downregulation was dependent on ROS level by H 2 O 2 treatment and, if so, whether PGC-1α downregulation could be inhibited by removing ROS using the well-known chemical antioxidant N-acetyl-L-cysteine (NAC). The level of PGC-1α expression, which was downregulated by H 2 O 2 was restored by 20 mM of NAC (Fig. 2C).

PGC-1α overexpression protected cells from H 2 O 2 -mediated injury.
To examine the physiological effect of PGC-1α in proximal tubule cells, we stably transfected HK-2 cells with an empty vector (Mock) or a plasmid encoding human PGC-1α (PGC-1α) (Fig. 3A). Expression of c-terminal c-Myc tagged PGC-1α was assessed with anti-c-Myc antibody. Stable cells clone were selected via confirmation of expression of zeocine, which was present in the backbone plasmid, pCDNA4. Evaluation of Mock and PGC-1α stable cells following H 2 O 2 -mediated injury revealed increase in cell viability (Fig. 3B) and decrease in the number of Annexin-V-positive cells (Fig. 3C) or in apoptotic body formation (Fig. 3D) in PGC-1α stable cells. This result showed that PGC-1α overexpression alleviates H 2 O 2 -mediated apoptotic and necrotic cell death.

PGC-1α had an anti-apoptotic effect.
To verify the protective effects of PGC-1α in injured proximal tubule cells, the expression and activation of pro-apoptotic proteins was evaluated in H 2 O 2 -treated Mock and PGC-1α stable cells (Fig. 4). Although PGC-1α expression was lower in H 2 O 2 -treated PGC-1α stable cells than in H 2 O 2 -untreated PGC-1α stable cells, PGC-1α expression was higher in PGC-1α stable cells than in Mock cells (Fig. 4A). The phosphorylation of p53 at Ser 15, which is involved in stabilization and mitochondrial accumulation of p53, were lesser in H 2 O 2 -treated PGC-1α stable cells than in Mock cells (Fig. 4A,C). Further, the level of activated caspase 3, assessed using the cleaved caspase 3 to caspase 3 ratio, markedly reduced in H 2 O 2 -treated PGC-1α stable cells (Fig. 4A,D). Release of mitochondrial cytochrome C to the cytosol, which resulted in activation of caspase 3, was also lesser in H 2 O 2 -treated PGC-1α stable cells than in Mock cells (Fig. 4A,E).

PGC-1α had an anti-oxidative effect.
Many earlier studies showed that PGC-1α suppresses ROS production through the induction of ROS-detoxifying enzymes 31,32 . Therefore, we confirmed whether intracellular ROS in cytosol or in mitochondria could be regulated by PGC-1α overexpression in our system using a selective ROS probe, CM-H 2 DCF-DA or MitoTracker Red CM-H 2 XRos, respectively. In PGC-1α stable cells, ROS levels were noticeably lower in the mitochondria as well as the cytosol, as compared to levels in Mock cells (Fig. 5).
Upregulation of Nrf-2 expression by PGC-1α was involved in the cytoprotective effects of PGC-1α. To further analyze the protective mechanism of the PGC-1α-associated anti-apoptotic effect and anti-oxidative effect, we studied the expression of Nrf-2, which is involved in the coordinated induction of genes that encode many stress-responsive and cytoprotective enzymes and related proteins. The Nrf-2 mRNA levels increased in H 2 O 2 -treated PGC-1α cells (Fig. 6A,B). The Nrf-2 protein level also increased following PGC-1α overexpression both in the cytosol and the nucleus (Fig. 6C). Moreover, the expression of heme oxygenase-1 (HO-1), known as a downstream target molecules of Nrf-2, also increased in consistent to Nrf-2 upregulation in PGC-1α stable cells, its expression in PGC-1α stable cells decreased on Nrf-2 specific siRNA knockdown in PGC-1α stable cells (Fig. 6D).
We examined whether the anti-apoptotic and anti-oxidative effects on H 2 O 2 -treated PGC-1α cells were dependent on the Nrf-2 level. The level of activated caspase 3, which was reduced by PGC-1α overexpression, was partly restored by Nrf-2-specific reduction on H 2 O 2 -treated PGC-1α cells (Fig. 7A). In addition, the ROS level, which was reduced by PGC-1α overexpression, was also partly recovered by Nrf-2-specific reduction in H 2 O 2 -treated PGC-1α cells (Fig. 7B).
Regulation of the p38/GSK3β/Nrf-2 axis by PGC-1α was involved in the cytoprotective effects of PGC-1α. Three ubiquitin ligase systems that act as negative regulator of Nrf-2 have been reported, those are, the Keap1-, GSK3β-, and Hrd1-mediated systems 24,[33][34][35] . Nrf-2 binds with these negative regulators, and hence maintains Nrf-2 to basal level. Keap1 and GSK3β are present in cytoplasm, and Hrd1 in the endoplasmic reticulum (ER). We analyzed whether the molecular mechanisms underlying Nrf-2 upregulation in PGC-1α cells are Keap1-dependent or GSK3β-dependent or Hrd1-dependent. In cytosolic negative regulators of Nrf-2, The phosphor form of GSK3β at Ser9 was considerably increased in H 2 O 2-treated PGC-1α cells at 1, 3, and 6 h as compared to the levels in Mock cells, whereas the Keap1 levels did not show a significant difference between the two groups (Fig. 8A). Hrd1, as a negative regulator of Nrf-2 in the endoplasmic reticulum, was also reduced in H 2 O 2-treated PGC-1α cells as compared to the levels in Mock cells, although the difference in Hrd1 was lesser than that of GSK3β.
We investigated whether any of the other upstream molecules were involved in the GSK3β-inactivated Nrf-2 upregulation. Interestingly, p38, an upstream signaling molecule affected by GSK3β inactivation 36 , was specifically  (Fig. 8D). In contrast, use of an ERK inhibitor (PD98059, 50 μM), used as a non-effective control on PGC-1α effect, did not lead to significant difference.

Discussion
This study showed that PGC-1α is physiologically involved for the cytoprotective effects and one of its regulation mechanisms is the regulation of the p38/GSK3β/Nrf-2 axis by PGC-1α overexpression.
In the current study, we found decreased PGC-1α expression in I/R-induced AKI, which is associated with impaired renal function. The S3 segment of the proximal tubule and the thick ascending limb of Henle are highly susceptible to AKI, such as ischemic injury 37 . Moreover, an earlier study involving in situ hybridization for PGC-1α mRNA showed that PGC-1α is mainly expressed in proximal tubules and the thick ascending limb of Henle 16 . In addition, the PGC-1α protein level in H 2 O 2 -treated HK-2 cells was gradually decreased at high H 2 O 2 concentrations or following longer exposures to H 2 O 2 . These findings are consistent with previous observations 38,39 . And also, H 2 O 2 -induced PGC-1α downergulation was inhibited by NAC pre-treatment in H 2 O 2 -treated HK-2 cells. It has been recently reported that NAC plays a role as a mitochondrial enhancer as well as an antioxidant precursor to glutathione (GSH) 40 . In psychiatry and related neurodegenerative diseases, NAC used to increase mitochondrial resilience and prevent allostatic load by inhibiting mechanism of oxidative stress and inflammation 41,42 . Given the prominent role of PGC-1α in mitochondrial biology, it is not surprising that PGC-1α is involved in the cellular response to ischemia. These findings suggest that PGC-1α could be a potential target to improve renal recovery following I/R-induced kidney injury. In stable cells, PGC-1α overexpression attenuated H 2 O 2 -induced cellular toxicity via anti-apoptotic and anti-oxidative effects. Mitochondria are the central executer of apoptosis 43 , and ROS generation has been suggested to be a major inducer of mitochondrial dysfunction and to play an important role in apoptosis regulation 44 . Our results suggest that a defect in PGC-1α is one of the major causes of H 2 O 2 -induced renal tubule cell apoptosis, and provides a novel strategy for preventing ROS-induced kidney tubule injury.
Nrf-2 serves as a master player of mitochondrial redox homeostasis by regulating the expression of diverse cytoprotective proteins that allow for cellular adaptation and survival under stress conditions 45,46 . The findings of the current study suggest that the PGC-1α upregulated Nrf-2 expression and sequentially induced phase 2 detoxifying enzymes and related proteins, such as HO-1, leading to a cytoprotective effect against ROS-mediated injury. Consistent with our data, it has recently been reported that Nrf-2 knockout cells have higher mitochondrial ROS levels than wild-type cells, suggesting that Nrf-2 regulates mitochondrial ROS production 47 . In addition, HO-1-knockout mice were found to be markedly sensitized to diverse forms of AKI 48 . In this study, we didn't check whether or not Nrf-2 mediated HO-1 expression in PGC-1α stable cells is specific in mitochondria. The several papers were reported about its mitochondrial function 49 . HO-1 overexpression appears to protect the heart from oxidative injury by regulating mitochondrial quality control 50 . In addition to HO-1, it has been reported that PGC-1α regulates the mitochondrial antioxidants, such as MnSOD, Prx5, and Prx3 in vascular endothelial cells 51 . In agreement with previous data, we also identified upregulation of their expression in PGC-1α stable cells (data not shown). But, whether upregulation of their expression in PGC-1α stable cells was dependent on the expression of Nrf-2 would be further elucidated.
In this study, we first demonstrated that regulation of the p38/GSK3β/Nrf-2 axis by PGC-1α is one of mechanisms protecting renal tubule cells against ROS-mediated cellular toxicity. Under basal condition, Nrf-2 is a short-lived protein that is subjected to continuous ubiquitination and proteasomal degradation. Under stress condition, Keap1 is inactivated by oxidation of the reactive cysteine residue or down-regulated by epigenetic silencing 52 ; then, GSK3β is inactivated by Ser9 phosphorylation. Nrf-2/Keap1 or Nrf-2/GSK3β complex is disrupted by conformational change. Consequently, Nrf-2 stabilizes and is then translocated to the nucleus for Nrf-2 mediated gene expression. Among cytosolic negative regulators of Nrf-2, we showed that GSK3β, in particular, changed to the inactive form on phosphorylation at Ser9 in PGC-1α cells at the increased time point of H 2 O 2 treatment, while Keap1 did not. Consistent with our data, mice with renal-proximal-tubule-specific GSK-3β knockout and chemical inhibition of GSK3 were found to have better survival and renal function than wild-type mice in an HgCl 2 -induced model of AKI in a previous study 53 . In our study, the protein expression of Nrf-2 was increased both in cytosol and in nucleus by the overexpression of PGC-1α. The degree of expression was greater in cytosolic fractionation than in nuclear fraction. It is speculated that Nrf-2 is more regulated by cytosolic events, such as inactivation of GSK3β. However, detailed mechanisms of Hrd1-dependent Nrf-2 upregulation would be further elucidated. Furthermore, our data showed that p38 was specifically activated in PGC-1α stable cells and that p38-specific inactivation in PGC-1α cells inversely affected the sequential activation and expression of GSK3β/Nrf-2/HO-1 as its downstream targets. 15d-PGJ(2), a potent endogenous ligand for peroxisome proliferators-activated receptor gamma, induces HO-1 expression using p38 MAP kinase and Nrf-2 pathway in ROS-mediated VSMCs; an inhibitor of p38 MAP kinase was found to abolish 5d-PGJ(2)-induced HO-1 expression 54 . Although we didn't check the regulation mechanism of p38/GSK3β/Nrf-2 axis by PGC-1α in I/R-injured kidney, several studies reported that the reduction of Nrf-2 and its downstream molecule, HO-1, in I/R-injured kidney resulted in I/R-induced ROS generation and inflammation, followed by acute kidney injury 55,56 . Renal ischemia-reperfusion injury (IRI) is a major cause of AKI, which has common pathophysiological features including renal tubular apoptosis/necrosis, ROS generation, mitochondria dysfuncrion and inflammatory cell infiltration. Therefore, it is reasonable to explore and to understand the therapeutic mechanism to treat common pathophysiology features of IRI. Our studies imply that PGC-1α may be one of therapeutic targets against AKI. In conclusion, PGC-1α protects human renal tubule cells from H 2 O 2 -mediated apoptotic injury by upregulating Nrf-2 via GSK3β inactivation mediated by activated p38. Regulation of the p38/GSK3β/Nrf-2 axis by PGC-1α could be a viable target for ameliorating mitochondrial dysfunction following AKI.
All experimental protocols were approved by the Animal Care Regulations (ACR) Committee of Chonnam National University Medical School (CNU IACUC-H-2016-26). Eight-week-old male C57BL6 mice were purchased by Samtako (Korea). Mice were anesthetized with 2% isofurane and 100% oxygen and placed on a temperature-regulated table (38.5 °C) to maintain body temperature. Renal ischemia was induced by clamping both renal pedicles with micro clamp (ROBOZ, Gaithersburg, USA) for 30 min. I/R group (n = 4) was sacrificed after 1 day of reperfusion. Control group (n = 4) underwent the same procedure, except that the clamp was not applied. Blood samples were then collected from the heart, and the left kidney was rapidly removed and processed for western blotting or fixed in 4% paraformaldehyde solution for immunohistochemistry (IHC). The right kidney was frozen at −80 °C for real-time PCR.   Tunel staining in kidney tissue. Apoptosis was determined using the ApopTag Plus Peroxidase In Situ Apoptosis Detection Kit (Chemicon International; Temecula, CA, USA) according to the manufacturer's protocol. The sections were counterstained with hematoxylin and examined by light microscopy. Image was magnified at x100. siRNA knockdown. RNA interference of Nrf-2 was performed using an Nrf-2-specific siRNA from Santa Cruz (Cat# sc-37030). Briefly, cells were transfected with indicated concentration of siRNA (30 nM and 50 nM) using DhamaFECT 1 transfection reagent according to the manufacturer's protocol. Cells transfected with control siRNA (Santa Cruz, Cat# sc-37007) were used as controls for direct comparison.
Cell viability. The MTT assay was applied to test cell viability. In brief, stable HK-2 cells (Mock or PGC-1α) were seeded into plates at 1 × 10 4 cells per 96 wells. After 1 day, cells were incubated in 100 μl of 0.5 mM H 2 O 2 diluted in HBSS for the indicated time at 37 °C and with 5% CO 2 . Subsequently, 10 μl of MTT reagent (5 mg/ml) was added to yield a final concentration of 0.5 mg/ml. After 2 h of additional incubation, all solution was removed, and 100 μl of DMSO was directly added to the cells to dissolve water-insoluble MTT-formazan. Absorption at 590 nm was determined with an ELISA reader (BioTek, Winooski, VT, USA).

Apoptosis assay. The number of apoptotic cells was quantified using the Ezway Annexin V-FITC Apoptosis
Detection Kit (KOMA BIOTECH, Seoul, Korea) according to the manufacturer's protocol. Cells were sequentially probed with Annexin V-FITC and propidium iodide (PI) dye. Fluorescent intensity was measured by a FACSCalibur TM flow cytometry (BD Biosciences, San Jones, CA, USA). DAPI staining for apoptosis analysis. The apoptotic effect was analyzed by using florescent nuclear dye 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI). The cells were seeded onto four well-cell culture slides (5 × 10 4 /well) and treated as mentioned previous. Cells were then washed with PBS and fixed in 4% paraformaldehyde for 10 min. Subsequently the cells were permeablized with equilibration buffer (1% BSA and 0.5% Triton X-100 in PBS) and stained with DAPI dye. After staining, the images were captured using a confocal microscope (LSM 510; Carl Zeiss). Image was magnified at x800, Bar = 20 μm. 58,59 were assessed using 5,6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H 2 DCFDA; Invitrogen, Carlsbad, CA, USA) or Rosamine-based MitoTracker probes (MitoTracker Red CM-H 2 XROS, Invitrogen, Carlsbad, CA, USA), respectively. Labeling with both probes was conducted on lived cells, but not fixed cells. After cells were treated with 0.5 mM H 2 O 2 for 30 min, and then were loaded with 10 μM CM-H 2 DCFH-DA or 0.2 μM CM-H 2 XROS for 30 min at 37 °C. Images were immediately visualized using confocal microscopy on a laser scanning microscope (LSM 510; Carl Zeiss), and analyzed using imageJ software. Image was magnified at x200 or x400.

Measurement of ROS generation. Level of intracellular and mitochondrial ROS
Preparation of nuclear and cytoplasmic extracts. HK-2 cells were lysed using NE-PER ® nuclear extraction reagent (NER) (Pierce Biotechnology, Rockford, IL, USA) according to the manufacturer's protocol. Briefly, 100 µL of ice-cold cytoplasmic extraction reagent (CER) I was added to the harvested cells. After incubated on ice for 10 min, ice-cold CER II was added to the tube. The tube was centrifuged at 16,000 × g for 5 min and the supernatant fraction was saved as cytosolic protein. The remained pellet was suspended in 50 µL of ice-cold NER. After centrifuging the tube at 16,000 × g for 10 min, the supernatant (nuclear protein) fraction was transferred to a clean tube.

Statistical analysis.
All experiments were conducted in triplicate. The results were expressed as mean ± standard deviation (S.D). We used Student's t test for the comparison between two gorups, and used One-way ANOVA when we compared more than two groups. Differences with values of p < 0.05 were considered significant.