Prolyl hydroxylase 2 (PHD2) inhibition protects human renal epithelial cells and mice kidney from hypoxia injury

Prolyl hydroxylase domain protein 2 (PHD2) is a key oxygen sensor, setting low steady-state level of hypoxia-inducible factor-α (HIF-α). Here, we showed that treatment of cobalt chloride (CoCl2), a hypoxia mimic, in HK-2 tubular epithelial cells induced PHD2 and HIF-1/2α expression as well as cell apoptosis and autophagy activation. Three methyladenine (3-MA), the autophagy inhibitor, blocked autophagy and protected HK-2 cells from CoCl2. Significantly, siRNA knockdown of PHD2 also protected HK-2 cells from CoCl2, possibly via increasing HIF-1α expression. Reversely, HIF-1α siRNA knockdown almost abolished cytoprotection by PHD2 siRNA in CoCl2-treated HK-2 cells. In vivo, pretreatment with a PHD inhibitor L-mimosine remarkably attenuated mice renal ischemia-reperfusion injuries. Molecularly, L-mimosine inhibited apoptosis and inflammatory responses in injured mice kidneys. Together, our results suggest that PHD2 silence or inhibition protects human renal epithelial cells and mice kidney from hypoxia injuries.

Thus far three different PHDs have been identified, including PHD1, PHD2, and PHD3 [4,8]. PHDs have different but sometime overlapping functions. PHD2 is the critical oxygen sensor of HIF-1α in normoxia and hypoxia [8,9]. The characteristics of theses enzymes provide potential targets for small molecule manipulation, which could activate HIF in the presence of normoxia. For example, cobalt chloride (CoCl 2 ) and other iron chelators have been applied to activate HIFs [10,11]. Existing evidences have shown that enhancing HIF via PHD inhibition could protect against ischemic injuries [11,12].
Maintaining normal Kidney function requires a high blood flow and a high glomerular filtration rate (GFR) [13]. Additionally, the proximal tubules need large amount of energy in the process of electrolyte re-absorption [13]. The proximal tubules are thus particularly vulnerable to ischemia and/or hypoxia. Ischemic injury is often an integration of multiple complex pathological processes, including apoptosis and autophagy [13].
Autophagy occurs at basal level in most tissues and contributes to the routine turnover of cytoplasmic components [14,15]. However, autophagy can also be induced by nutrient deprivation or starvation [14,15]. In addition to turnover of cellular components, autophagy is also essential for cell homeostasis [16], as well as cell adaptation and defense to adverse environments [17,18]. Studies have highlighted the importance of autophagy Research Paper: Pathology in normal proximal tubule function and recovery from acute ischemic injuries [19][20][21]. Paradoxically, chronic activation of autophagy may also promote cell damage [19][20][21].
In the study, we employed the human renal proximal tubular epithelial cell line HK-2, as well as a mouse model of renal ischemia-reperfusion injury to explore the potential function of PHD2 in the process.

PHD2 and HIFα expressions are increased in HK-2 tubular epithelial cell after CoCl 2 treatment
CoCl 2 mimics the hypoxic condition and has been validated as a simple tool to examine the molecular mechanisms driven by hypoxia in vitro [10]. As shown in Figure 1, protein and mRNA expressions of HIF-1α and HIF-2α were low under normoxic conditions. After CoCl 2 treatment, there was a time-dependent increase of HIF-1/2α expression at the mRNA level ( Figure 1A). A similar pattern of protein expression was observed by Western blot assay ( Figure 1B). The protein expression of HIF-1α and HIF-2α both significantly increased, starting from 12h (P < 0.05 vs. "0h") ( Figure 1B). Whereas HIF-1α and HIF-2α mRNA expressions were induced at later periods (starting from 24h) ( Figure 1A). The protein and mRNA levels of HIF-1α reached the peak at 36h (P < 0.01 vs. "0h"), which then declined ( Figure 1A). The protein expression of PHD2 gradually increased after exposure to CoCl 2 , reaching the peak at 48h ( Figure 1B). While the PHD2 mRNA level reached its peak at 24h (P < 0.01 vs. "0h") ( Figure 1A). for applied time, mRNA and protein expressions of HIF-1α, HIF-2α and PHD2 were tested by RT-qPCR assay (A) and Western blot assay B., respectively. HIF-1α, HIF-2α and PHD2 mRNA expression was normalized to 18S (n = 12). HIF-1α, HIF-2α and PHD2 protein expression was normalized to β-actin (B, lower panels, n = 14). *P < 0.05 vs. 0h group; # P < 0.01 vs. 0h group. www.impactjournals.com/oncotarget CoCl 2 induces apoptosis and autophagy activation in HK-2 cells Next, we tested the expression of Bcl-xL, a known HIF-1α-regulaed gene [22], in CoCl 2 -treated HK-2 cells. As shown in Figure 2, the protein expression of Bcl-xL was increased after CoCl 2 treatment. It level was peaked at 12-24h and was then declined afterwards ( Figure 2). We also analyzed the apoptotic response of HK-2 cells to CoCl 2 . Western blot results in Figure 2 showed that CoCl 2 up-regulated the pro-apoptotic Bax [23,24] in HK-2 cells (Figure 2), suggesting apoptosis activation.
To test whether autophagy could be induced after hypoxia treatment, we investigated the expression of LC3-II, whose upregulation is considered as the characteristic marker of autophagy [25]. CoCl 2 induced a significant accumulation of LC3 in HK-2 cells. Upregulation of LC3-II was most significant at 12h (P < 0.01 vs. "0h") after CoCl 2 treatment ( Figure 2). And its level was then decreased thereafter ( Figure 2). We further examined CoCl 2 -induced autophagy in HK-2 cells by transmission electron microscope (TEM) ( Figure 3). As shown in the representative micrographs, autophagosomes, with characteristic features of double or multiple membranes, began to appear 6h after CoCl 2 treatment in the HK-2 cells (Figure 3). They were observed up to 48h after treatment of CoCl 2 ( Figure 3). These results indicate that CoCl 2 activates autophagy in HK-2 cells.

Autophagy inhibitor 3-MA inhibits HK-2 cell death and apoptosis by CoCl 2
To explore the role of autophagy in CoCl 2 -induced HK-2 cell death, pharmacological strategy was applied. One hour prior to the CoCl 2 treatment, we exposed cells to 3-methyladenine (3-MA), which is a selective inhibitor of the autophagy [26]. As shown in Figure 4A   The cells were then fixed and processed for electron microscopy. Autophagosomes (black arrows) with characteristic double or multiple membranes were identified. Experiments in this figure were repeated four times, and similar results were obtained. cell apoptosis, tested by ssDNA ELISA assay, was also attenuated by the pretreatment of 3-MA (Supplementary Figure 1A). Additionally, under electron microscopy, the number of autophagosomes decreased and the cell ultrastructure appeared relatively normal ( Figure 4C). These results suggested that the autophagic process may be harmful to human renal tubular epithelial cells during CoCl 2 -induced hypoxia.

L-mimosine administration attenuates renal ischemia-reperfusion injuries
To study the role of PHD2 in vivo, we applied a non-selective PHD inhibitor L-mimosine. Animals in the L-mimosine group (50mg/kg body, 6h before surgery) showed improved renal function compared to those in the vehicle group, and the serum level of creatinine was dramatically decreased in L-mimosine-treated mice ( Figure 7A). Histological examinations revealed significant tubular damage and a high injury score in the injured kidney 24h after surgery, which were largely inhibited by L-mimosine pretreatment. (Figure 7B).
For the signaling studies, we showed that L-mimosine administration augmented HIF-1/2α accumulation after ischemia ( Figure 7C). HIF-1α and HIF-2α protein expression in kidney increased by 35.4% (32.18±4.75 vs. 49.84±7.14, P <0.05) and 35.5% (37.02±8.06 vs. 57.42±9.44, P <0.05) respectively with L-mimosine treatment ( Figure 7C). Conversely, the expression of PHD2 was significantly decreased by 76.0% (61.46±5.89 vs. 14.78±3.79, P < 0.01) after L-mimosine treatment. pretreated with L-mimosine (L-Mim, 50 mg/kg, "IR+L-Mim") or vehicle control (10% NaHCO 3 , "IR+Veh"), were subjected to ischemia reperfusion operation. After 24h, serum creatinine (Scr) level was analyzed A; HE staining were performed on kidney slices, and History scores were calculated B; Expressions of listed proteins in fresh kidney slices were tested by Western blot C. and E; Level of TNF-α and IL-10 in the fresh kidney slices was tested by ELISA assay D. Sham-operated mice served as controls. Protein expression was normalized to β-actin (n = 10). *P < 0.05 vs. "IR+Veh" group (C-F); # P < 0.01 vs. "IR+Veh" group (C-F); www.impactjournals.com/oncotarget ELISA analysis results demonstrated that, with the L-mimosine pre-treatment, the level of inflammatory cytokine TNF-α was markedly decreased, yet the antiinflammatory cytokine IL-10 was up-regulated ( Figure  7D). The above apoptosis and autophagy markers along with inflammatory cytokine levels were also tested in vivo. The quantitative analysis revealed that expression of the anti-apoptotic protein Bcl-xL in kidney was upregulated, but the pro-apoptotic protein Bax was downregulated following L-mimosine administration ( Figure  7E). Further, conversion of LC3-I to LC3-II was upregulated by L-mimosine treatment, indicating autophagy augmentation ( Figure 7E). Collectively, we showed that pretreatment with L-mimosine markedly attenuated the deterioration of renal function and the severity of renal damage. Molecularly, L-mimosine inhibited apoptosis, facilitated autophagy and suppressed inflammatory responses in injured kidney.

DISCUSSION
The main findings of our study are as follows: PHD2 siRNA knockdown produced a renoprotective effect by attenuating apoptosis and autophagy in HK-2 cells. Further, L-mimosine-activated HIF offered significant protection by attenuating apoptosis, enhancing autophagy and suppressing inflammatory responses in mice.
We showed that treatment of CoCl 2 , a hypoxia mimic, in HK-2 tubular epithelial cells induced PHD2 and HIF-1/2α expression. siRNA knockdown of PHD2 protected HK-2 cells from CoCl 2 possibly though facilitating HIF-1α expression. Reversely, HIF-1α siRNA knockdown almost abolished cytoprotection by PHD2 siRNA in CoCl 2 -treated HK-2 cells. Therefore, target HIF-1α by modulating PHD2 could offer protection of tubular epithelial cells in hypoxia conditions. Autophagy could be cell detrimental or cytoprotective depending on the stimuli that provoked it. It is still largely unknown what mechanisms determine the final outcomes [14][15][16]. Here, we observed that CoCl 2 activated autophagy in HK-2 cells. On the other hand, 3-MA, the autophagy inhibitor, blocked autophagy and protected HK-2 cells from CoCl 2 . These results indicate that the autophagic pathway may act as a detrimental factor in the CoCl 2 hypoxic model at least in vitro, although more detailed studies are still needed.
Interestingly, our in vivo results demonstrated that autophagy induction was accompanied with improved kidney functions. These results indicated that autophagy may play a protective role against kidney ischemia reperfusion injuries. These in vivo results are consistent with other studies [28,29]. The mechanisms underlying the pro-survival effect of autophagy remain largely unknown [28,30]. In the process of renal ischemia reperfusion injury, autophagy activation was shown to eliminate damaged mitochondria and prevent accumulation of aggregate-prone proteins [28,29]. The rapid increased number of research publications on autophagy and hypoxia will lead to a better understanding the mechanisms involved.
Collectively, the results of our study suggest that activating HIF, i.e. via inhibiting PHD, may be a useful therapeutic strategy in ischemic renal injury.

Assay of cell viability
For cell viability assessment, 10 µL of Alamar Blue (Kaiji, Nanjing, China) was added to each well and the plates were incubated at 37°C and 5% CO 2 for 2h. The absorbance was measured by using a Synergy HT multidetection microplate reader (Bio-Tek, Winooski, VT) at a test wavelength of 570 nm.

Apoptosis assay by single strand DNA (ssDNA) ELISA
DNA denature is an important characteristic marker of cell apoptosis. Following the treatment of cells, denatured ssDNA was detected through a nucleosomal monoclonal antibody in an ELISA format. Briefly, cells (1×10 4 /well) were seeded onto 96-well plates. Cell apoptosis was analyzed by the ssDNA ELISA kit from Chemicon International (Temecula, CA, Catalog Number: 1217). OD value at 450 nm was utilized as a quantitative indicator of cell apoptosis.

Electron microscopy
Cells were grown in T25 flasks and fixed for 4h in 2.5% glutaraldehyde fixative in phosphate buffer at 4°C. After fixation, cell monolayers were placed in glutaraldehyde wash solution and then post-fixed in Millonigs Osmium Tetroxide fixative for 15-30 minutes. The cultures were dehydrated through a graded series of 10% to 100% ethanol, infiltrated, and then embedded in TAAB Premix Resin medium. Ultrathin sections were collected on copper grids and stained with uranyl acetate and lead citrate. Sections were viewed using a Hitachi H7600 transmission electron microscope (TEM).

Mouse renal ischemia reperfusion injury
Experiments were performed on 6-8 week male C57BL/6J mice weighing 20-22 g (from Animal Center of Fudan University, Shanghai, China). Animals were housed in temperature-and humidity-controlled cages, with free access to water and rodent food on a 12-h light/dark cycle. All protocols were approved by the Institutional Animal Care and Use Committee of Fudan University, and adhered strictly to the NIH Guide for the Care and Use of Laboratory Animals. All surgery was performed under sodium pentobarbital anesthesia, and efforts were made to minimize suffering. L-mimosine (L-Mim, Sigma, Dorset, UK) was resolved in 10% NaHCO 3 at a concentration of 15 mg/mL (adjusted pH to 7.4 by HCl plus PBS) and administered intraperitoneally at 50 mg/kg, 6h before renal ischemia reperfusion injury surgery.
Mice were anesthetized with intraperitoneal sodium pentobarbital (30mg/kg), and renal ischemia reperfusion injury was induced by bilateral renal pedicle clamping for 35 min. Sham-operated mice underwent the same surgical procedures but with no occlusion of the renal pedicle. Intra-rectal temperature of mice was maintained at 36.5°-37.0°C with a heating pad. Blood and kidney samples of all mouse groups were harvested at 24h after the surgery. Animals were divided into three groups (n = 4): Group 1 (Sham): Sham-operated animals without induction of ischemia reperfusion injury; Group 2 (Veh): Animals pretreated with NaHCO 3 (PH 7.4) 6h before of ischemia reperfusion injury; Group 3 L-mimosine (L-Mim, Sigma, M0253, Shanghai, China) treatment group: Animals pretreated with L-Mim 6h before of ischemia reperfusion injury;