Reciprocal REGγ-Nrf2 Regulation Promotes Long Period ROS Scavenging in Oxidative Stress-Induced Cell Aging

Increased accumulation of reactive oxygen species (ROS) and decline of adaptive response of antioxidants to oxidative stimuli has been implicated in the aging process. Nuclear factor erythroid 2-related factor 2 (Nrf2) activation is a core event in attenuating oxidative stress-associated aging. The activity is modulated by a more complex regulatory network. In this study, we demonstrate the proteasome activator REGγ function as a new regulator of Nrf2 activity upon oxidative stress in cell aging model induced by hydrogen peroxide (H2O2). REGγ deficiency promotes cell senescence in primary MEF cells after H2O2 treatment. Accordingly, ROS scavenging is accelerated in WT cells but blunted in REGγ lacking cells during 12-hour recovery from a 1-hour H2O2 treatment, indicating long-lasting antioxidant buffering capacity of REGγ. Mechanistically, through GSK-3β inhibition, REGγ enhances the nuclear distribution and transcriptional activity of Nrf2, which is surveyed by induction of phase II enzymes including Ho1 and Nqo1. Meanwhile, Nrf2 mediates the transcriptional activation of REGγ upon H2O2 stimulation. More interestingly, short-term exposure to H2O2 leads to transiently upregulation and gradually descent of REGγ transcription, however sustained higher REGγ protein level even in the absence of H2O2 for 24 hours. Thus, our results establish a positive feedback loop between REGγ and Nrf2 and a new layer of adaptive response after oxidative stimulation that is the REGγ-GSK-3β-Nrf2 pathway.


Introduction
Aerobic creatures are constantly exposed to oxidants, which may originate from internal sources, such as mitochondrial dysfunction or from external sources, such as exposure to hydrogen peroxide (H 2 O 2 ) or other environmental pollutants [1,2]. Oxidative stress in the cell is caused by an equilibrium disruption between formation of free radicals and their removal by enzymatic and nonenzymatic antioxidant molecules, which detoxify harmful effects and are primary lines of defense and critical for maintaining various cell functions [3][4][5][6]. The cumulative and increasing oxidative damage to cells has been linked to a variety of age-related pathologies [7,8]. In addition, the decline of the repair systems, including the proteasomal degradation of damaged proteins and the adaptive response to oxidative stress is also associated with chronic oxidative state in aging. [9,10] Nuclear factor erythroid 2-related factor 2 (Nrf2), a major sensor of oxidative stress in the cell [11,12], belongs to the basic leucine zipper transcription factor family featuring a cap'n'collar motif [13,14]. Numerous studies have provided clear evidence that Nrf2 is a crucial molecule in the regulation of basal and induced expression of phase II genes through regulation of antioxidant response elements (AREs), also known as electrophile response element (EpRE) [15,16]. Depending on the cellular redox balance, the Nrf2-ARE signaling system is subjected to multiple layers of regulation [17]. The main regulation is provided by Kelchlike-ECH-associated protein 1 (Keap1), a cullin-3-(Cul3-) based E3 ubiquitin ligase substrate adaptor, allowing Nrf2 to be ubiquitylated by Cul3/Rbx1 [18]. Under normal homeostatic situations, Nrf2 is sequestered by forming a complex with numerous cytoplasmic proteins, including Keap1 and targeted to proteasomal degradation [19], resulting in a low baseline level of Nrf2. In addition, β-TrCP, another E3 ligase adaptor also facilitates Nrf2 ubiquitination and degradation through the Skp1-Cul1-Rbx1/Roc1 core E3 complex [20]. Under the conditions of oxidative stress, Nrf2-Keap1 interaction is disrupted [21,22], which leads Nrf2 translocation to the nucleus to increase the expression of protective genes and antioxidant enzymes including the phase II detoxifying enzymes heme oxygenase-1 (Ho1) and NAD(P)H quinone oxidoreductase 1 (Nqo1) [1,23]. Posttranslational modification such as phosphorylation and sumoylation is also another regulatory layer of Nrf2 intracellular distribution, activity, and stability [24].
Glycogen synthase kinase-3β (GSK-3β) is a serine/threonine kinase and originally reported as a key enzyme of glucose homeostasis through regulation of the rate of glycogen synthesis. It has subsequently been found to influence most cellular processes, including growth, differentiation and death, as part of its role in modulating response to hormonal, nutritional, and cellular stress stimuli [25,26]. Moreover, an increasing body of literature has indicated GSK-3β as a negative regulator of Nrf2. In the delayed/late response to oxidative stress, the Tyr216 of GSK-3β site is phosphorylated, and the activated GSK-3β acts upstream and phosphorylates Fyn kinase, which in turn translocates into nucleus and phosphorylates tyrosine 568 of Nrf2, leading to the nuclear export of Nrf2 and its turn over by proteasome [27]. Phosphorylation of Nrf2 Neh6 domain by GSK-3β also facilitates its interaction with β-TrCP and subsequent degradation. In contrary, chemicals or short interfering RNA mediated inhibition of GSK-3β results in nuclear accumulation of Nrf2 and transcriptional activation of the Nrf2 downstream gene Nqo1 and Ho1 [28,29]. Nevertheless, after transient stress stimulation, how Nrf2 activation is maintained for a prolonged period remains largely unknown.
In this study, we found that REGγ deficiency accelerates cell senescence in H 2 O 2 -induced cell aging model. H 2 O 2 activated Nrf2 to enhance REGγ expression, which maintained the nuclear location and activity of Nrf2 via inhibition of GSK-3β, resulting in decreased accumulation of reactive oxygen species (ROS) during 12-hour recovery from a 1hour H 2 O 2 treatment. Taken together, our results demonstrate a long-lasting proantioxidants function of REGγ in recovery from oxidative stress.

Materials and Methods
2.1. Animals. REGγ +/+ and REGγ -/mice were kindly provided by Dr. John Monaco at the University of Cincinnati College of Medicine [45]. Animal procedure was carried out in accordance with the American Association for the Accreditation of Laboratory Animal Care International's guidelines and was approved by the school's Animal Center.

Primary Fibroblast
Cultures from Mouse. The skin from mice were cut into small pieces approximately 2-3 cm and dissociated with dispase II rotated at 4°C overnight. The next day, the epidermis was peeled off and the dermis was cut into small pieces (1 mm) and transferred into the culture plate (6-wells or 3.5 cm dish). Then the cells were incubated at 37°C in 5% CO 2 , in DMEM (Invitrogen) supplemented with 10% FBS, and the medium was changed every day for 3-4 days.
2.4. Real-Time Quantitative RT-PCR (RT-qPCR) and Gel-Based PCR. Total RNA was isolated from cells using TRIzol reagent (Invitrogen). DNase treated RNA was reversetranscribed with a reverse transcription kit (Vazyme, China). PCR products were analyzed using 7500 Fast Real-Time PCR System (Applied Biosystems, USA), and relative transcript abundance was normalized to that of 18S mRNA. The oligonucleotide primers for PCR are as follows. 18S: 2.5. Western Blot Analysis. Whole cell proteins were extracted using radioimmunoprecipitation assay (RIPA) lysate (Beyotime, Shanghai, China). After the protein was separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), the target protein was transferred to the nitrocellulose filter (NC) membranes. After 5% of the skim milk powder was blocked for 2 h, it was 2 Oxidative Medicine and Cellular Longevity incubated with a specific primary antibody at 4°C overnight. The next day, after washing with Tris-Buffered Saline-Tween (TBST), fluorescent-labeled (Jackson ImmunoResearch) secondary antibodies were added and incubated for 1 hour on a shaker. Then, bands were detected and visualized by a LI-COR Odyssey Infrared Imaging System. Specific primary antibody Nrf2, Rabbit, 1 : 1000 (Abcam); Ho1 and Nqo1, Rabbit, 1 : 1000 (Abcam); anti-β-actin-mouse 1 : 1000 (Sigma).
To measure cellular senescence, cells were stained with the SA β-Gal staining kit (Beyotime Institute of Biotechnology, Nanjing, China) according to the manufacturer's instructions. Images containing >200 cells were taken using a bright-field microscopy and total, and blue-colored cells were counted. The percentage of SA β-Gal positive cells was represented by the ratio between the number of blue-colored cells and the number of total cells from at least three independent experiments. 2.9. REGγ Luciferase Reporter Constructs. PCR was used to amplify DNA fragments containing REGγ genomic sequences from 293 T cell genomic DNA, and primers were derived from human genomic REGγ and ligated into the kpn1/xhol sites of the promoter less pGL4-Basic (Promega, Madison, USA) vector and was named as pGL4-REGγ-luc.
2.10. Luciferase Assay. After transfection, cells were collected and washed three times with cold PBS, followed by lysis in a cell lysis buffer (Promega, Madison, WI, USA). After one freezing and thawing cycle, whole-cell lysates were centrifuged in a cold room (4°C) at 12,000 rpm for 15 min, and the supernatant obtained was collected in a fresh tube. Next, 20 μl supernatant was added to equal amounts of luciferase assay substrate, and luminescence was detected as relative light units by using the LUMIstar OPTIMA reader (BMG Labtech, Offenburg, Germany). Data were normalized with beta actin and obtained from three independent experiments. Fold change in values is represented as a mean of three experiments.
2.11. Immunofluorescence. The cells were fixed with 4% paraformaldehyde, and the goat serum was added drop wise for 1 h. The diluted antibody Nrf2 (1 : 500, Abcam, Cambridge, MA, USA) was added drop wise and placed in a refrigerator at 4°C overnight. The second antibody was added to the next day in the dark, and after 1 h of incubation, 4′,6-diamidino-2-phenylindole (DAPI) was stained and incubated in the dark. After 15 minutes, the images were taken by Leica laser confocal microscope.
2.12. Statistics. All data in the experimental analyses were generated using GraphPad Prism6 software. Significance among different groups was determined using two-tailed unpaired t-tests. All values were expressed as mean ± SD. Values of p < 0:05 were considered statistically different.  [47]. In order to clarify the long-term function of REGγ during recovering from acute oxidative stress, we applied a cell senescence model induced by H 2 O 2 treatment. SA-β-galactosidase activity assay was employed to monitor the cell aging. The results showed that SA-β-gal-positive cells were markedly increased in both WT and REGγ deficient primary MEF cells treated with H 2 O 2 compared to nontreated cells (Figures 1(a) and 1(b)). More importantly, the percentage of SA-β-gal positive staining cells was more pronounced in REGγ lacking cells compared to those REGγ normal cells at day 4 post 30 minutes treatment with H 2 O 2 , which was in agreement with previous work that REGγ KO mice developed phenotype characteristic of premature aging [46]. But, there was no big difference in nontreated control parts (Figure 1(b)). These results manifest the protective function of REGγ in the process of cell aging.

REGγ Accelerates ROS Scavenging during Recovery from
Short Exposure of H 2 O 2 . The link between the accumulation of reactive oxygen species (ROS) and cellular senescence has been well established [49,50]. Therefore, the intracellular ROS level was immediately measured in REGγ WT and KO MEF cells after exposure in H 2 O 2 for 2 hours. Surprisingly, we observed significantly higher ROS content in WT cells than REGγ knockout cells (Figures S1a and S1b) and the possible explanation will be discussed in the discussion part. We speculated that REGγ may have a long-lasting antioxidant effect, so we exposed cells to oxidative stress for a short period of time with H 2 O 2 and continued routing culturing for up to 24 hours after H 2 O 2 removal. The dynamic ROS level at different time point of recovery was tested. Interestingly, the results showed that the ROS level was gradually decreasing in REGγ WT MEF cells, however kept unaltered in REGγ deficient cells (Figure 2(a)). Starting 3 Oxidative Medicine and Cellular Longevity from around 4-hour recovery, the ROS accumulation in REGγ containing cells was evidently less than in REGγ lacking cells (Figures 2(a) and 2(b)), which was consistent with the cellular senescence phenotype. Taken together, these data declare that REGγ expedites ROS scavenging and maintains it at a lower concentration during recovery from oxidative stress, suggesting its long-term antioxidant and antiaging capacity.

REGγ Promotes the Nuclear Retention and Activity of
Nrf2 in Long-Term Recovery from Oxidative Stress. In searching for the potential mechanisms of ROS eliminating by REGγ, we first speculated that REGγ might regulate the activity of Nrf2, because Nrf2 is a critical redox sensor and activates the transcription of a number of antioxidant genes that is known for combating ROS [51]. As expected, we observed a notably higher expression of Nrf2 target genes, including Ho1 and Nqo1 after 2 hours of recovery from H 2 O 2 -induced WT cell conditions in compare with the REGγ -/cell conditions (Figures 3(a) and 3(b)). Given proteasome activator function of REGγ, we wondered whether REGγ influences Nrf2 stability. To answer this question, MEF WT/KO cells were treated with the protein synthesis inhibitor cycloheximide (CHX). Results showed that the Nrf2 protein level decreased faster in REGγ KO cells compared to WT cells, suggesting that Nrf2 is not a direct substrate of REGγ-20S proteasome ( Figure S2a and b). Next, immunofluorescence staining was executed to monitor the nuclear location of Nrf2 in primary mouse fibroblast cells from REGγ-deficient and REGγ wild type mice. There was no significant difference among the nonstressed control group (Figures 3(c) and 3(f)). However, H 2 O 2 induced more visible nuclear localization of Nrf2 in REGγ containing cells than REGγ lacking cells, especially in condition that H 2 O 2 had been removed for 8 hours (Figures 3(d)-3(f)). In summary, we draw the conclusion that REGγ regulates the Nrf2 activity by dominating its nuclear location.

REGγ Maintains the Activity of Nrf2 after Oxidative
Stress via GSK-3β. Inhibition of GSK-3β was reported to lead to Nrf2 nuclear accumulation and activation. Additionally, our previous study discovered that REGγ boosts GSK-3β decay in an ATP-and ubiquitin-independent manner [39]. Therefore, we hypothesized that REGγ promotes the nuclear retention and activation of Nrf2 through degrading GSK-3β. Faster degradation of GSK-3β in REGγ +/+ than REGγ -/cells was corroborated first (Figures 4(a) and 4(b)). Then, WT and REGγ deficient MEF cells were pretreated with 10 μM of GSK-3β inhibitor (CHIR99021) for 12 hrs followed by incubation with 50 μM of H 2 O 2 for 30 minutes. Ho1 and Nqo1 mRNA level were examined by real-time PCR and gel-based PCR after removing H 2 O 2 for 4 hours (Figures 4(c) and 4(d) and S2c). Western blot was employed to detect the protein level of Ho1 and Nqo1 after removing H 2 O 2 for 12 hours (Figures 4(e) and 4(f)). The results displayed that H 2 O 2 alone led in remarkable increase of Ho1 and Nqo1 mRNA and protein level in REGγ containing cells rather than in REGγ lacking cells (Figures 4(c)-4(f)). When pretreating cells with Gsk-3β inhibitor, less difference was observed between REGγ +/+ and REGγ -/group when compared with H 2 O 2 alone treatment group (Figures 4(c)-4(f )). These studies illustrate that REGγ augments oxidative  Oxidative Medicine and Cellular Longevity stress-induced activity of Nrf2 and expression of Ho1 and Nqo1 is at least partially mediated by a mechanism involving GSK-3β inhibition.

The Expression of REGγ Is Strengthened by Oxidative
Stress. To facilitate the removal of oxidatively damaged proteins, ubiquitin-/ATP-independent degradation by the 20S is prominently enhanced because of the increased amount of free 20S in cells, which is the result of oxidative stress triggered disassembly of the 26S proteasome [52]. Since the REGγ proteasome is part of the ubiquitin-/ATP-independent degradation system, we explored the expression of REGγ upon oxidative stress. Intriguingly, quantitative RT-PCR (qRT-PCR) revealed that REGγ mRNA level was obviously upregulated after short period exposure of H 2 O 2 in a time-dependent manner up to 8 hours, and then decreased steadily thereafter ( Figure 5(a)). Notably, even 24 hours after oxidative stress, the REGγ protein still kept at a markedly high level (Figures 5(b) and 5(c)). These results point out that REGγ may be upregulated and maintained to provide a long-lasting protection for oxidative stress challenged cells.

H 2 O 2 Induced REGγ Expression via Nrf2.
In order to test whether REGγ would be transcriptionally modulated by Nrf2, first we transiently transfected HaCaT cells with Nrf2 plasmid or Nrf2 specific siRNAs. Then, the mRNA level of REGγ was inspected by gel-based PCR. It turned up that REGγ was upregulated by overexpression of Nrf2 ( Figure S3a) and downregulated by knocking down of Nrf2 ( Figure S3b). Next, we analyzed the promoter of human REGγ genes for binding sequences of Nrf2. We focused on the regions from 1.7 kb upstream to 1 kb downstream of the transcription start site. Using the JASPAR database, two potential Nrf2 consensus binding sites were identified. Then, we generated luciferase reporter gene including wild type or mutated Nrf2 consensus binding sites (Figure 6(a)). Transient transfection of Nrf2 dramatically increased the activity of wild-typed luciferase reporter gene (Figure 6(b)). After we introduced site-specific mutagenesis in the potential binding site, the REGγ luciferase reporter gene became unresponsive (Figure 6(c)). Alternatively, specific Nrf2 inhibitor (ML385) was employed to blunt the activation of endogenous Nrf2. It emerged that H 2 O 2 substantially

Nqo1
Relative mRNA level   (Figure 6(d)). To sum up, these findings establish a positive feedback regulation loop between REGγ and Nrf2 to protect cells from oxidative stress driven cell senescence.

Discussion
A powerful antioxidant system has evolved to maintain redox homeostasis of aerobic creatures (including humans) to prevent disease and aging [53,54]. REGγ, as a member of the 11S proteasome activator family, degrades a variety of substrate proteins and participates in many important physiological and pathological processes [34]. In present study, we shed light on a REGγ-mediated unknown mechanism responsible for regulation of antioxidative ability and cell aging under oxidative stress and proposed the following working model. In REGγ normal cells, H 2 O 2 stimulation led to immediate Nrf2 activation and subsequent elevation of downstream target genes to clear out the oxygenic free radicle. Concurrently, REGγ, as a new identified target gene of Nrf2, was upregulated and maintained long period activity of Nrf2 via directing degradation of GSK-3β, contributing to long-term ROS removal and cell aging delay. However, in REGγ deficient cells, the negative regulation of Nrf2 by GSK-3β could not be antagonized, which resulted in increased ROS accumulation and cell aging (Figure 7). In conclusion, our research delineated a new regulatory layer of Nrf2 by REGγ and a positive feedback loop between Nrf2 and REGγ in promoting long-term ROS scavenging, hence delaying oxidative stress-induced cell aging.
Free radicals such as ROS can produce cumulative oxidative damage to macromolecules and lead to cellular senescence [55]. The 20S proteasome appears to be the main player in the process that recognizes and degrades oxidatively damaged proteins [56]. At first, redox modifications quickly boost the catalytic activity and proteolytic capacity of preexisting 20S proteasomes. Furthermore, short-term oxidative stress causes the 26S proteasome complex to disassemble into 20S and 19S particles [57]. Previously, our laboratory has depicted that the association of REGγ with 20S proteasome was enhanced by acute oxidative stress to promote proteasome-dependent proteolysis of oxidized protein substrates [47]. Our present study further announced that REGγ could act as antioxidant to resist oxidative stress and slow down cell aging in the adaptation stage after oxidative stress. It effectively maintained ROS concentration at a low level and provided long-lasting antioxidant effect, highlighting the link between ubiquitinindependent proteasome complex and the long-term adaption of cells to oxidative stress.
Previous study demonstrated that REGγ knockout or knockdown significantly elevates SOD2 expression in mouse heart tissues or in AC16 cell lines [58]. In accordance with this, we also observed that SOD2 level was higher in REGγ deficient MEF cells than in WT MEF cells in unstressed normal condition, however had no effect on the expression of SOD1 (data not shown), which may explain why ROS in REGγ knockout cells was temporarily less than in REGγ normal cells after exposure to H 2 O 2 (Figure 2 and S1). We prefer that this transiently lower ROS level in REGγ KO cells is due to a rescue or a compensation effect mediated by other factors, such as the SOD2, rather than the direct function of REGγ to regulate the ROS level in this particular REGγ lacking situation. Since a small amount of ROS act as signaling molecules and has health-promoting effects in the cells, REGγ deficient cells may be unable to produce enough ROS and fail to completely elicit and elongate the antioxidant response, as shown in Figure 3 in current data. After allowing the cells to recover for a while, REGγ lacking cells had considerably more ROS accumulation due to loss of prolonged transcriptional activity of Nrf2 by REGγ. These studies depict complicated regulation of oxidative stress response by REGγ. How cells sense the ROS level to switch REGγ-20S proteasome for the off/on regulation of Nrf2 activity under oxidative stress or normal conditions deserves further studies.

Oxidative Medicine and Cellular Longevity
It is well accepted that the equilibrium between antioxidant and oxidant is disrupted during aging [59]. However, the change of Nrf2 activity with age may depend on species, tissues, and cell types [9]. REGγ-deficient mice have been reported to develop premature aging phenotypes [46]. Here, our study evidenced that the deactivation of Nrf2 mediated by GSK-3β under prolonged exposure to oxidative stress was predominantly rescued by REGγ. Combined with our recent observation that REGγ level was declined with age (data not shown), we suggest that there is a bias towards deregulation of REGγ during aging, and a decay of REGγ can further accelerate aging process. Whether this regulation exists in divergent cell types, tissues or species deserve more systematic studies.

Conclusions
Overall, our findings demonstrate a scenario where REGγ, the 11S family of proteasome activator, prolongs the nuclear retention of Nrf2 and the expression of its downstream antioxidant enzymes Ho1 and Nqo1, to reduce the ROS accumulation, enhance the oxidant resistance, and eventually protect against oxidative stress-mediated cell senescence. These results make REGγ a promising drug target since pharmacological manipulation of REGγ could have potential benefits on oxidative damage-related disorders and aging.

Data Availability
The supporting information can be downloaded at: http:// www.xxx.xxx/xxx/. Figure S1: the level of ROS in REGγ WT and KO cells after short time exposure of H 2 O 2 . Figure S2: Nrf2 was not a substrate of REGγ. Figure S3: up-and downregulation of REGγ by overexpression and knockdown of Nrf2.

Conflicts of Interest
The authors declare no conflict of interests.

Authors' Contributions
Solomon Kibreab and Zimeng Wang contributed equally to this work.

Acknowledgments
This work was supported by the National Natural Science Foundation of China (82073483) and the Science and  Figure S1: the level of ROS in REGγ WT and KO cells after short time exposure of H 2 O 2 . Figure S2: Nrf2 was not a substrate of REGγ. Figure S3: up-and downregulation of REGγ by overexpression and knockdown of Nrf2. (Supplementary  Materials)