PIAS1 Alleviates Hepatic Ischemia-Reperfusion Injury in Mice through a Mechanism Involving NFATc1 SUMOylation

Recently, attentions have come to the alleviatory effect of protein inhibitor of activated STAT1 (PIAS1) in hepatic ischemia-reperfusion injury (HIRI), but the underlying molecular mechanistic actions remain largely unknown, which were illustrated in the present study. Microarray-based analysis predicted a possible regulatory mechanism involving the PIAS1/NFATc1/HDAC1/IRF-1/p38 MAPK signaling axis in HIRI. Then, growth dynamics of hypoxia/reoxygenation- (H/R-) exposed hepatocytes and liver injury of HIRI-like mice were delineated after the alteration of the PIAS1 expression. We validated that PIAS1 downregulation occurred in H/R-exposed hepatocytes and HIRI-like mice, while the expression of NFATc1, HDAC1, and IRF-1 and phosphorylation levels of p38 were increased. PIAS1 inactivated p38 MAPK signaling by inhibiting HDAC1-mediated IRF-1 through NFATc1 SUMOylation, thereby repressing the inflammatory response and apoptosis of hepatocytes in vitro, and alleviated liver injury in vivo. Collectively, the NFATc1/HDAC1/IRF-1/p38 MAPK signaling axis is highlighted as a promising therapeutic target for potentiating hepatoprotective effects of PIAS1 against HIRI.


Introduction
Hepatic ischemia-reperfusion injury (HIRI) is a common complication that may occur in various clinical situations, such as liver transplantation, liver resection, trauma, and vascular surgery [1,2]. Moreover, HIRI is one of the leading causes for postsurgery hepatic dysfunction and failure [3]. As knowledge accumulates on the molecular mechanisms underlying inflammation and apoptosis in the liver, the anticipated development and assessment of molecular targets may offer a promising perspective for HIRI [3][4][5].
Protein inhibitor of activated STAT1 (PIAS1), as a small ubiquitin-like modifier (SUMO) E3 ligase, was identified as a crucial player both in vivo and in vitro in numerous biological processes, such as inflammatory responses [6]. More specifically, a prior study has demonstrated the potential mechanistic action of PIAS1 as a protector against myocardial IRI by mediating PPARγ SUMOylation [7]. It should be noted that PIAS1 was a negative regulator of the transac-tivation activity of the isoforms of nuclear factor of activated T cells 1 (NFATc1) through SUMOylation [8]. NFATc1, also known as NFAT2, is a crucial member of the transcription factor NFAT family which is indispensable in the immune system and plays an important role in liver injury repair [9,10]. It has been reported that NFATc1 serves as a regulator of the transcriptional activity of histone deacetylase 1 (HDAC1) in its management of glioblastoma [11]. In addition, previous studies have confirmed that NFATc1 can regulate the expression of IL-2 and IL-10 in lymphocytes by interacting with HDAC1 [12,13]. HDAC1 is responsible for many processes including cell cycle, DNA repair, and apoptosis by deacetylating histones and nonhistone proteins [14]. The inhibition of HDAC1 was confirmed to delay the progression of IRI [15]. In addition, recent data unraveled that HDAC2-mediated histone modification played a role in the development of IRI by regulating interferon regulatory factor 1 (IRF-1) in mice [16]. Furthermore, IRF-1 was verified as a vital player in HIRI by promoting autophagic cell death through the p38 mitogen-activated protein kinase (MAPK) pathway [17]. Meanwhile, the activation of p38 MAPK signaling was confirmed to exert protective effects against apoptosis and inflammation during HIRI [18]. Accordingly, the above reports suggested a potential mechanism of PIAS1 on HIRI treatment involving the NFATc1/HDAC1/IRF-1/p38 MAPK signaling axis. This study was designed to investigate the interactions among them and thus to provide a novel target for alleviating HIRI.

Materials and Methods
2.1. Ethics Statement. The current study was performed with the approval of the Institutional Review Board of the Second Xiangya Hospital of Central South University. All animal experiments were approved by the Animal Care and Use Committee of the Second Xiangya Hospital of Central South University and performed in strict accordance with Guide for the Care and Use of Laboratory Animals.

In Silico Analysis.
Mouse HIRI-related microarray GSE10657 was obtained from the Gene Expression Omnibus (GEO) database. The microarray contains 6 sham-operated samples (sham) and 24 HIRI samples, which was annotated with the platform file GPL1261-56135. The differential analysis of HIRI-related genes was conducted using the "limma" package of R language, to single out differentially expressed genes (DEGs).
The correlation between DEGs and HIRI was analyzed using Phenolyzer. The protein-protein interaction (PPI) relationship network encoded by genes was analyzed by STRING (version: 11.0) with the threshold of interaction score ≧ 0:15 and max interactors = 300.
Reperfusion injury-related genes were searched in the CTD database, and the top 1000 genes ranked by the inference score were screened out. Then, jvenn tool was employed to determine candidate genes by taking an intersection between the interaction genes and HIRI-related genes. KEGG enrichment analysis of the candidate genes was conducted using the "clusterProfiler" package of R language. Pearson correlation between candidate genes was conducted using Chipbase (version = 2:0) based on GETX_ liver tissue expression data.
After the injection, the HIRI model was induced (SUPPLEMENTARY Figure 1) as previously described [20]. After made a midline incision by conducting laparotomy, a nontraumatic vascular clip was placed on the blood vessel blocking the blood supply of the median and left lateral lobe of the portal vein and hepatic artery for 1 h to cause ischemia. The clip was then gently removed and reperfused for 12 h, causing about 70% of liver ischemia/ reperfusion injury in mice. In the sham operation group, laparotomy and vascular separation were performed without occlusion. The liver tissues of mice were used for subsequent experiments.
2.11. Liver Injury Evaluation. The malondialdehyde (MDA) production [21] and myeloperoxidase (MPO) activity [22] in liver tissues were assessed as described. For MDA production detection, 200 μL of liver homogenate samples were mixed with the reagents provided in the KeyGEN MDA detection kit (KGT004, KeyGEN, Nanjing, China). The optical density (OD) value at 532 nm was measured using a microplate reader. In terms of MPO activity detection, 100 mg liver tissue was homogenized and centrifuged to collect pellet, which was resuspended in buffer solution containing 43.2 mmol/l KH 2 HPO 4 , 10 mmol/l EDTA, 6.5 mmol/ l Na 2 HPO 4 , and 0.5% hexadecyltrimethylammonium, followed by ultrasonication for 10 s. The supernatant was reacted with 3,3′3,5′-tetramethylbenzidine for a spectrophotometer observation to measure the OD value at 655 nm.
Liver necrosis was evaluated using the hematoxylin-eosin (H&E) staining. Liver tissue samples were fixed in 10% formalin, embedded in paraffin and sectioned (4-5 μm each), and then stained with hematoxylin and eosin, and damage degree was observed under a microscope (DM5000B; Leica).
2.12. TUNEL Staining. TUNEL staining was conducted using an In Situ Cell Death Detection Kit (11684795910, Roche, Basel, Switzerland). Then, sections were added with proteinase K working solution and incubated in a 37°C incubator for 30 min. After being immersed in blocking solution (3% H 2 O 2 ), samples were added with TUNEL test solution dropwise for incubation at 37°C in the dark for 60 min. TUNEL-positive cells were photographed using a fluorescence microscope (DMi8; Leica).
2.14. Statistical Analysis. The SPSS 21.0 statistical software (IBM Corp, Armonk, NY) was used for statistical analysis. Measurement data were expressed as mean ± standard deviation. Comparisons between two groups were analyzed by unpaired t-test. Comparisons among multiple groups were performed using one-way analysis of variance (ANOVA) with Tukey's post hoc test. A value of p < 0:05 was considered statistically significant.

Bioinformatics Prediction for Potential
Mechanism of PIAS1 in HIRI Pathogenesis. The top 20 DEGs with the smallest p value was screened out by analyzing the differential expression of the mouse HIRI-related microarray GSE10657 (Figure 1(a)). The correlation between the 20 DEGs and ischemia-reperfusion injury was determined using Phenolyzer, and it was revealed that PIAS1 presented the highest correlation score with ischemia-reperfusion injury ( Figure 1(b)). Also, analysis of microarray GSE10657 confirmed that PIAS1 was notably lowly expressed in HIRI (Figure 1(c)). Therefore, PIAS1 was selected as the target gene for follow-up research.
In order to further predict the downstream regulatory factors of PIAS1, 300 interaction factors of PIAS1 were obtained through STRING, and an intersection between the interaction factors and ischemia-reperfusion injuryrelated genes was taken using the jvenn tool to obtain 43 candidate genes (Figure 1(e)). Based on the results of KEGG enrichment analysis, the 43 candidate genes were mainly involved in the AGE-RAGE signaling pathway, FoxO signaling pathway, and MAPK signaling pathway (Figure 1(e)), among which 9 genes including AKT1, CHUK, JUN, MAPK14, MAPK3, NFATc1, NFKB1, RELA, and TP53 participated in the MAPK signaling pathway. We used STRING tool to analyze the interaction between NFATc1 and other candidate genes, and it was found that NFATc1 and HDAC1 are positively correlated in liver tissue (Pearson ' s r = 0:5578, p value = 2.44e-10) (Figures 1(f) and 1(g)).
Based on the above findings, it was inferred that the SUMO E3 ligase PIAS1 may be a key gene involved in the pathogenesis of HIRI by regulating NFATc1, HDAC1, and MAPK signaling pathway.

PIAS1 Inhibits Inflammatory Response and Apoptosis of
Hepatocytes through NFATc1 SUMOylation. We first went to validate the expression of PIAS1 and NFATc1 in the context of HIRI by RT-qPCR and western blot analysis. It was uncovered that the expression of PIAS1 was reduced, and the expression of NFATc1 was notably increased in the liver tissues of HIRI mice (Figures 2(a) and 2(b), SUPPLEMENTARY Figure 2A).
Next, it was observed whether NFATc1 is the SUMOylation target of PIAS1 in HIRI. The SUMOylation of NFATc1 in 293 T cells transduced with Flag-NFATc1 or HA-SUMO1 was determined by conducting Co-IP experiments, and the SUMO1-NFATc1 conjugate band was observed at -152 kDa (Figure 2(c)). In addition, the SUMOylation of endogenous NFATc1 in hepatocytes was also identified by conducting Co-IP experiments using the anti-NFATc1 antibody (Figure 2(d)).
The association of PIAS1 with the SUMOylation of NFATc1 was further illuminated in H/R-exposed hepatocytes. Co-IP assay suggested that the SUMOylation of NFATc1 was much lowered in H/R-exposed hepatocytes. The PIAS1 restoration enhanced the SUMOylation of NFATc1, while it was decreased following further treatment of SUMO inhibitor ML-792 (Figure 2(e)). The TUNEL staining indicated enhanced apoptosis of H/R-exposed hepatocytes. Besides, the PIAS1 overexpression repressed apoptosis of H/R-exposed hepatocytes, which was increased in the presence of additional SUMO inhibitor ML-792 treatment (Figure 2(f)).
Furthermore, the PIAS1 expression was elevated, and the NFATc1 expression was reduced in response to oe-PIAS1 + oe-NC transduction, while the NFATc1 levels were elevated in response to oe-PIAS1 + oe-NFATc1 compared with oe-PIAS1 alone (Figure 2(g)).
In addition, the restored expression of PIAS1 resulted in repressed expression of TNF-α, IL-1β, IL-6, Bax, and caspase-3 mRNA (cleaved caspase-3 protein) as well as cell apoptosis, while the expression of Bcl2 was increased in H/ R-exposed hepatocytes. Relative to oe-PIAS1 alone, simultaneous restoration of PIAS1 and NFATc1 led to the elevated expression of TNF-α, IL-1β, IL-6, Bax, and caspase-3 mRNA (cleaved caspase-3 protein) as well as cell apoptosis was increased, while the expression of Bcl2 was decreased in H/Rexposed hepatocytes (Figures 2(h)-2(j), SUPPLEMENTARY Figure 2B).
Taken together, PIAS1 reduced hepatocyte inflammatory response and apoptosis through the SUMOylation of NFATc1.

NFATc1 Induces Inflammatory Response and Apoptosis in Hepatocytes by Enhancing HDAC1 Transcriptional
Activity. The effect of HDAC1 in the hepatocytes after different treatments was further assessed. It was found that the expression of HDAC1 in the liver tissues of HIRI mice was upregulated (Figures 3(a) and 3 Type   Disease Markers SUPPLEMENTARY Figure 2C). In addition, compared with HDAC1-WT alone, the luciferase activity in the presence of oe-NFATc1 + HDAC1-WT was promoted, while there was no significant difference in luciferase activity of cells cotransduced with HDAC1-MUT, indicating that NFATc1 can bind to the promoter region of HDAC1 (Figure 3(c)).

Disease Markers
The levels of NFATc1 and HDAC1 expression were elevated in the oe-NFATc1 presence, while the HDAC1 expression was diminished in response to oe-NFATc1 + sh-HDAC1 as compared with oe-NFATc1 alone (Figure 3(f)).
Together, NFATc1 can induce hepatocyte inflammatory response and apoptosis by upregulating the transcriptional activity of HDAC1.

HDAC1 Induces Inflammatory Response and Apoptosis of Hepatocytes through Transcriptional Activation of IRF-1.
We next moved to clarify the effect of IRF-1 in hepatocytes after different treatments. It was observed that the expression of IRF-1 in the liver tissues of HIRI mice was notably increased (Figures 4(a) and 4(b), SUPPLEMENTARY Figure 2E). Compared with cells transduced with oe-NC + IRF-1-WT, the luciferase activity of cells after the transduction of oe-HDAC1 + IRF-1-WT was increased, while there was no significant difference in luciferase activity of those with IRF-1-MUT, indicating that HDAC1 can bind to the promoter region of IRF-1 (Figure 4(c)).
Further, the expression of IRF-1 in the presence of oe-HDAC1 was obviously increased, and the expression of IRF-1 was notably decreased in the presence of sh-HDAC1 (Figure 4(d)). Moreover, the decrease of IRF-1 in cells after the transduction of sh-IRF-1#1 was much more obvious compared with those with sh-IRF-1#2 (Figure 4(e)). So, sh-IRF-1#1 was selected for follow-up experiments.
HDAC1 and IRF-1 expression was elevated in the presence of oe-PIAS1, while the results were negated in response to additional oe-NFATc1 (Figure 4(f)).
Next, it was observed in hepatocytes that the expression of TNF-α, IL-1β, IL-6, Bax, and caspase-3 mRNA (cleaved caspase-3 protein) along with cell apoptosis was increased, while the expression of Bcl2 was decreased in response to oe-HDAC1, but the results were counteracted in the presence of oe-HDAC1 + sh-IRF-1 (Figures 4(g)-4(I), SUPPLEMENTARY Figure 2F).
To sum up, HDAC1 promoted hepatocyte inflammatory response and apoptosis through transcriptional activation of IRF-1.

IRF-1 Promotes Inflammatory Response and Apoptosis of
Hepatocytes by Activating the p38 MAPK Signaling Pathway. First, the phosphorylation levels of p38, JNK, and ERK1/2 in hepatocytes after different treatments were determined by western blot analysis. It was demonstrated that the

16
Disease Markers phosphorylation levels of p38, JNK, and ERK1/2 were notably increased in the liver tissues of HIRI mice ( Figure 5(a)). In order to confirm whether the role of IRF-1 in HIRI depends on the p38 MAPK signaling pathway, hepatocytes were transduced with oe-IRF-1, the phosphorylation level of p38 before and after the addition of the p38 MAPK inhibitor SB203580 was observed, and it was uncovered that the phosphorylation level of p38 increased obviously after adding the p38 MAPK inhibitor SB203580 (Figures 5(b) and 5(c)).
In short, inflammatory response and apoptosis of hepatocytes were enhanced by IRF-1 via activating the p38 MAPK signaling pathway.

PIAS1 Overexpression Plays a Protective Role in HIRI
Mice by Inhibiting the Activation of NFATc1/HDAC1/IRF-1/p38 MAPK Pathway. The in vitro findings on the molecular mechanism of PIAS1 were further substantiated in a HIRI mouse model. We also validated downregulated PIAS1 expression and upregulated expression of NFATc1, HDAC1, and IRF-1 and phosphorylation levels of p38 in the liver tissue of mice following HIRI versus mice receiving sham operation. Moreover, the expression of NFATc1, HDAC1, and IRF-1 and phosphorylation levels of p38 was decreased in the liver tissues of mice when the PIAS1 expression was restored in HIRI mice. Besides, additional anisomycin treatment in PIAS1-overexpressing HIRI mice, the phosphorylation levels of p38 were elevated in the liver tissue of the mice (Figure 6(a), SUPPLEMENTARY Figure 2(I)).
The activities of AST and ALT were promoted, liver injury was worsened, and the necrotic area, eosinophils, the MDA level, and MPO activity were increased in HIRI mice versus mice receiving sham operation. The restored PIAS1 expression suppressed activities of AST and ALT, alleviated liver injury, and diminished MDA level and MPO activity in HIRI mice. Further administration of anisomycin negated the effect of PIAS1 and aggravated liver injury (Figures 6(b)-6(f)).
According to the results of ELISA, immunofluorescence staining, and TUNEL staining, the serum level of TNF-α, IL-1β, and IL-6, the expression of neutrophil marker Gr-1 and macrophage marker CD68, and hepatocyte apoptosis were all promoted in the liver tissues of the HIRI mice, as compared with mice receiving sham operation. Meanwhile, the PIAS1 overexpression in HIRI mice repressed the levels of TNF-α, IL-1β, IL-6, Gr-1, CD68, and hepatocyte apoptosis and the effects of which was counterweighed following further treatment of anisomycin (Figures 6(g)-6(I)).
Taken together, the overexpression of PIAS1 inactivated the NFATc1/IRF-1/p38 MAPK pathway, thereby exerting a protective effect on HIRI mice.

Discussion
The obtained evidence suggested that PIAS1 reduced the inflammatory response and apoptosis of inactivating the p38 MAPK signaling by downregulating HDAC1-mediated IRF-1 through NFATc1 downregulation, thus ultimately alleviating HIRI-like symptoms.
A prior study has explored and confirmed the therapeutic effects of PIAS1 on protecting against myocardial IRI [23]. However, detailed studies about the downstream mechanism of PIAS1 in HIRI are still insufficient. The finding of our work suggested that PIAS1 reduced the inflammatory response and apoptosis of hepatocytes through the SUMOylation of NFATc1, evidenced by the increased Bcl2 expression and the decreased expression of TNF-α, IL-1β, IL-6, Bax, and caspase-3 in mice with HIRI-like symptom. SUMOylation was recognized as a posttranslational modification which is involved in various crucial cellular functions, such as cell cycle, DNA damage repair, and cell apoptosis [24]. Consistent with our finding, the negative regulation of PIAS1 on NFATc1 isoforms was also validated in a prior study [8]. Likewise, NFATc1 inhibition is responsible for the repressed inflammatory response [25]. More importantly, NFAT was illustrated to be positively correlated with inflammation and apoptosis and thus served as a crucial player for IRI [23]. Additionally, TNF-α, IL-1β, and IL-6 were wellknown proinflammatory cytokines, and their downregulation is a biomarker of the relieved inflammation in HIRI [8]. Likewise, Bax, caspase-3, and Bcl2 were apoptosisrelated protein markers, and the downregulation of Bax

20
Disease Markers and caspase-3 and upregulation of Bcl2 are also identified as a symptom of repressed apoptosis in HIRI [26]. In a word, PIAS1 was a promising inhibitor for the inflammatory response and apoptosis of hepatocytes and thus further a vital mediator of HIRI through the SUMOylation of NFATc1. It was then confirmed that NFATc1 elevated the expression of HDAC1 and subsequently activated the IRF-1mediated p38 MAPK signaling, which further promoted the inflammatory response and apoptosis of hepatocytes. Similar to our finding, the positive correlation between NFAT and HDACs was also certified by a recent study and was confirmed as key promoter in the regulation of inflammation during the preservation of islet function after islet transplantation [27]. Besides, inhibition of HDAC1 exerted a neuroprotective effect against cerebral IRI [28,29]. Furthermore, a prior study demonstrated that the increased expression of HDAC1 was partially responsible for the deterioration of mice with HIRI-like symptoms, and its repression acted as a promising therapeutic target for HIRI management [30,31]. The positive regulation between HDAC and IRF-1 was also demonstrated in another study [32]. IRF-1, a transcriptional regulator of IFNs and IFNinducible genes, is also a well-recognized vital player in response to inflammation [33]. Moreover, elevation of IRF-1 in hepatocytes was illustrated as a main cause of the enhanced apoptosis and thus further contributed to the pathogenesis of HIRI [34]. Besides, a prior study also identified the implication of IRF-1 in the progression of HIRI by positively mediating the p38 MAPK signaling pathway [17]. Additionally, activated p38 MAPK signaling pathway has been observed in HIRI [35]. In addition, inactivation of the p38 MAPK signaling represents a potential therapeutic target for HIRI [36].

Conclusion
In summary, SUMO E3 ligase PIAS1 alleviated HIRI by inhibiting the NFATc1/HDAC1/IRF-1/p38 MAPK pathway ( Figure 7). This study provides new insights for revealing the molecular mechanism of hepatoprotective PIAS1 in HIRI. Novel therapeutics based on PIAS1 upregulation may become potential target of personalized strategy in the future for enhancing therapeutic outcomes of HIRI.

Data Availability
The data and materials of the study can be obtained from the corresponding author upon request.

Conflicts of Interest
The authors declare no competing interest.