Mild Oxidative Stress Reduces NRF2 SUMOylation to Promote Kras/Lkb1/Keap1 Mutant Lung Adenocarcinoma Cell Migration and Invasion

Nuclear factor erythroid 2-related factor 2 (NRF2) is a crucial transcription factor for cell adaptation and defense against oxidative stress. NRF2 activation confers Kras/Lkb1/Keap1 (KLK) mutant tumor cells with greater resistance to oxidative insults. We previously reported that SUMOylation at lysine residue 110 is important for the ability of NRF2 to promote reactive oxygen species (ROS) clearance in hepatocellular carcinoma. In this study, we investigated whether SUMOylation is necessary for the ability of NRF2 to inhibit KLK lung adenocarcinoma (LUAD) cell migration and invasion. Our experiments showed that mild oxidative stress reduced NRF2 SUMOylation, which promoted KLK LUAD cell migration and invasion. Mechanistically, NRF2 SUMOylation increased the antioxidant ability of NRF2 and reduced cellular ROS levels, mainly by transcriptionally activating Cat in KLK LUAD cells. With reduced NRF2 SUMOylation, increased ROS acted as signaling molecules to activate the JNK/c-Jun axis, which enhanced cell mobility and cell adhesion, to promote LUAD cell migration and invasion. Taken together, the results of this study reveal a novel signaling process in which reduced NRF2 SUMOylation permits increased KLK LUAD cell migration and invasion under mild oxidative stress.


Introduction
The role of reactive oxygen species (ROS) in cancer has remained controversial for decades, in part, because different levels of ROS confer different outcomes in cancer cells. High ROS levels are harmful to cell, but mild oxidative stress at sublethal levels activates signaling pathways to promote tumor growth and progression [1,2]. Cancer cell migration and invasion are the initial steps of tumor metastasis. During cell migration and invasion, members of the mitogenactivated protein kinase (MAPK) family of proteins are activated by ROS [3][4][5]. In lung adenocarcinoma cells (LUAD), H 2 O 2 activates epidermal growth factor (EGF) receptors [6]; hence, oxidization of receptor tyrosine kinases (RTKs) facilitates MAPK signaling activation and promotes migration and invasion [7].
Multiple studies have reported that NRF2 is a SUMOylated protein [21][22][23]. Our previous research revealed that SUMOylation of lysine residue 110 (K110) of NRF2 reduces ROS levels, promotes de novo serine synthesis, and maintains hepatocellular carcinoma tumorigenesis [23]. In the present study, we investigated the effects of NRF2 on KLK LUAD cell migration and invasion, and whether SUMOylation is critical for these effects. We studied the effect of mild oxidative stress on NRF2 SUMOylation and then investigated the underlying mechanism by which NFR2 influences KLK LUAD cell migration and invasion.

Cell
Culture and Construction of Stable Cell Lines. The human LUAD cell line A549 was cultured in F12K medium (Gibco), while H2122 and H23 cells were cultured in RPMI medium (HyClone). Each medium contained 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin. These cells were cultured in a humidified 37°C incubator with 5% CO 2 . Cells were infected by shNRF2 and/or PCDH-NRF2/NRF2 K110R lentivirus and then selected by Blasticidin S or Puromycin for 1 week.

Western
Blotting. Cells were washed with phosphatebuffered saline (PBS), lysed in radio immunoprecipitation assay (RIPA) buffer (150 mM NaCI, 50 mM Tris base, 0.1% SDS, 1% Triton-X-100, pH 7.4) on ice for 20 min, and then ultrasonicated until the solution became clear. Protease and phosphatase inhibitors were added to the RIPA buffer in advance. The cell lysates were centrifuged at 12,000 rpm for 15 min at 4°C, and the supernatant was collected for Western blotting. Total protein was resolved in 10% SDS/PAGE gels, followed by electrophoretic transfer to PVDF membranes in a Tris-glycine buffer. The membranes were blocked at room temperature for 1 h in 5% nonfat milk with TBS-Tween (TBS-T) on a shaker and then incubated with the primary antibodies overnight at 4°C. The membranes were washed in TBST at least 5 times (5 min each) and then incubated with HRP conjugated anti-rabbit or anti-mouse IgG at room temperature for 1 h with gentle shaking. The ECL substrate was added, and the results were visualized using Image-Quant LAS 4000 (GE). β-Actin was used as loading control.

Quantitative
Real-Time PCR. Total RNA was isolated using TRIzol universal reagent (TianGen). Then, 1 μg RNA was reverse transcribed into complementary DNA (cDNA) using the Fast King gDNA Dispelling RT SuperMix kit (Tian-Gen). Quantitative real-time PCR was performed on Light-Cycler 480 (Roche) using TB Green Premix (Takara). 18S rRNA was used as a control for normalization. The primers used in this study are listed in Supplementary Table S1.

Migration and Invasion
Assay. For the migration assay, cells were resuspended in serum-free medium and plated in the upper chamber of Transwells (Corning) in a 24-well plate. The cell numbers were 1 × 10 4 to 1 × 10 5 per well. For the invasion assay, Transwells were coated with Matrigel matrix (Corning). Medium with 10% fetal bovine serum (FBS) was added in the bottom chamber of the Transwells. The cells on the upper surface of the chamber were removed using a cotton swab, and the cells on the bottom of the chamber were fixed with 4% paraformaldehyde and stained with crystal violet 12-16 h later. For H 2 O 2 treatment, 0.5, 50, or 500 μM H 2 O 2 was added into the upper chamber. The migrating and invading cells were counted in micrographs taken of 5 random fields using Photoshop, and the data were analyzed using Graphpad Prism.
2.6. NRF2 SUMOylation by Ni 2+ -NTA Pull-down Assay. Cells were infected with PCDH-His-SUMO1 lentivirus and selected by puromycin for 1 week. Once the cells reached 70% confluency, they were treated with H 2 O 2 (0.5, 50, or 500 μM) for 12 h. For NAC (N-acety-L-cysteine) treatment, cells were pretreated with 1 mM NAC for 1 h, and then, H 2 O 2 was added for further incubation for 12 h. Ni 2+ -NTA pull-down assay was performed as previously reported [23].   3 Oxidative Medicine and Cellular Longevity measured using the Amplite Colorimetric Hydrogen Peroxide Assay Kit (AAT Bioquest) the next day. Briefly, 50 μl of H 2 O 2 cell working solution was added to each well of cells and H 2 O 2 standards to make the total H 2 O 2 assay volume of 100 μl/well. The reaction was incubated at room temperature for 10-60 min and protected from light. Then, the absorbance at 650 nm was measured by an absorbance plate reader.
2.9. RNA-Seq. RNA-Seq was performed by KangChen Biotech, Shanghai, China. Briefly, the RNA-seq library was prepared using Illumina kits. The sequencing was performed using Illumina Hiseq 4000. Sequencing was carried out by running 150 cycles. Principal component analysis (PCA), hierarchical clustering, correlation analysis, pathway analysis, and gene ontology (GO) were performed, and volcano plots and scatter plots were generated to identify the differentially expressed genes using R or Python environment for statistical computing and graphics.
2.10. Measurement of GSH and GSSG. Total glutathione (GSH+GSSG) and oxidized glutathione disulfide (GSSG) were measured using a GSH and GSSG Assay Kit (Beyotime) according to the manufacturer's instructions. The GSH/GSSG ratio was then calculated.
2.11. Chromatin Immunoprecipitation Assay. The chromatin immunoprecipitation (ChIP) assay was performed using the Simple ChIP Plus Sonication Chromatin IP Kit (Cell Signaling Technology) according to the manufacturer's protocol. Briefly, cells were fixed with formaldehyde, and the chromatin was sheared with sonication into 200-1000 bp DNA-protein fragments. Then, the NRF2 antibody (or IgG antibody as a control) was added. The complex was coprecipitated and captured by Protein G beads. The protein-DNA cross-links were reversed, and the DNA was purified. Then, the enrichment of the Cat promoter was detected by RT-PCR. The primers used in this assay are listed in Supplementary Table S2. 2.12. Statistical Analysis. All results were obtained from at least three independent experimental replicates and are presented as mean ± standard error of the mean (S.E.M.). Significance was determined by Student's t test. P < 0:05 was considered significant.    Figure S1A). Correspondingly, H23 cells showed a higher intracellular ROS level than A549 and H2122 cells did ( Figure S1B). Upon knockdown of NRF2, the ROS levels in A549 and H2122 cells were elevated, whereas the ROS level in H23 cells was not altered ( Figure S1C and S1D). Additionally, KLK LUAD cell migration and invasion were increased with NRF2 knockdown ( Figure S1E), revealing that NRF2 inhibits KLK LUAD cell migration and invasion. In contrast, NRF2 knockdown in H23 cells delayed cell migration and invasion ( Figure S1E), which is consistent with previous studies showing that downregulation of BACH1 inhibits lung cancer metastasis [19,20]. However, in the KLK LUAD cell lines, BACH1 expression was not affected ( Figure S1F), indicating that NRF2 regulates KLK LUAD migration and invasion independent of BACH1.

Oxidative Medicine and Cellular Longevity
To investigate the role of NRF2 and its SUMOylation in KLK cell migration and invasion, we restored NRF2 wild-type and NRF2 with SUMOylation site mutation (K110R) expression individually in A549 and H2122 cells (Figure 1(a)). Expression of NRF2 wild-type inhibited cell migration and invasion, which had been enhanced by NRF2 knockdown, but NRF2 K110R did not exhibit this inhibitory capability (Figures 1(b) and 1(c)). Although NRF2 SUMOylation promoted KLK LUAD tumorigenesis and did not affect KLK LUAD cell apoptosis ( Figure S2), as reported in HCC previously [23], the results in Figure 1 suggest that SUMOylation is critical for the inhibitory effect of NRF2 on KLK cell migration and invasion.

Mild Oxidative Stress Reduces NRF2 SUMOylation to
Induce KLK LUAD Cell Migration and Invasion. Hydrogen peroxide acts as a signaling molecule to induce cancer cell migration and invasion. In A549 cells, we found that mild oxidative stress generated by treatment with 0.5 μM H 2 O 2 induced cell migration and invasion, whereas moderate/severe oxidative stress (50 μM/500 μM H 2 O 2 ) did not induce and even reduced cell migration and invasion (Figure 2(a)). Consistent with previous reports, moderate/severe oxidative stress induced NRF2 expression as a cellular defense response, whereas mild oxidative stress had no impact on NRF2 expression (Figure 2(b)). Because endogenous NRF2 SUMOylation is difficult to detect experimentally, we constructed a His-SUMO1 overexpressing stable cell line named A549-His-SUMO1 and pulled down SUMO1-modified proteins with Ni 2+ -NTA. Surprisingly, we detected that mild oxidative stress (0.5 μM H 2 O 2 ) reduced NRF2 SUMOylation in A549-His-SUMO1 cells (Figure 2(c)), whereas SUMOylation of NRF2 was increased after treatment with 50 μM or 500 μM H 2 O 2 given that NRF2 expression was increased (Figure 2(d)). More importantly, the decrease in NRF2 SUMOylation caused by mild H 2 O 2 (0.5 μM) was reversed when antioxidant NAC (N-acety-Lcysteine) was added (Figure 2(c)). With a deficiency in NRF2 SUMOylation, as shown in shNRF2+K110R A549 cells, mild oxidative stress no longer induced cell migration and invasion (Figure 2(e)), indicating that mild oxidative stress in KLK LUAD cells induces cell migration and invasion by reducing NRF2 SUMOylation.

SUMOylation Is Critical for the Antioxidant Ability of NRF2 via Transcriptional Activation of Cat in KLK LUAD
Cells. We previously reported that NRF2 SUMOylation reduces the intracellular ROS level in HCC cells [23]. In A549 cells, NRF2 SUMOylation reduced both the total ROS and H 2 O 2 levels (Figure 3(a)). In addition, the reduced glutathione (GSH) to oxidized glutathione disulfide (GSSG) (GSH/GSSG) ratio was relatively higher in A549 cells in which NRF2 wild-type expression was rescued (Figure 3(a)). To explore the underlying mechanism, we performed RNA sequencing (RNA-Seq) analysis to compare the gene expression levels in A549 cells before (normal control, NC group) and after NRF2 knockdown (KD group), as well as in A549 cells expressing NRF2 wild-type (WT group) and NRF2

Focal adhesion
Cell and cell adhesion  K110R (KR group). Volcano plots showed that 1299 genes were significantly upregulated, and 687 genes were downregulated in the KD group versus NC group ( Figure S3A), and 99 genes were significantly upregulated, and 255 genes were downregulated in the KR group versus WT group ( Figure S3C). Gene ontology (GO) analyses showed that upon NRF2 knockdown, pathways involved in the response to oxidative stress and the response to hydrogen peroxide were downregulated ( Figure S3B). The heat map and trend graph revealed that in A549 cells, NRF2 SUMOylation promotes H 2 O 2 removal mainly by activating Cat transcription, which was validated by our quantitative real-time polymerase chain reaction (PCR) and Western blot results (Figures 3(b)-3(d)). Different from our previous report [23], NRF2 SUMOylation did not alter the protein expression of GPX2 in KLK cells ( Figure S3E). ChIP assay showed that NRF2 K110R had a significantly decreased ability to bind with the Cat gene promoter (Figure 3(e)). Moreover, upon challenge of A549 cells with oxidants (tert-butyl hydroperoxide [TBHP] and H 2 O 2 ), the antioxidant ability of NRF2 was reduced when its SUMOylation site was mutated (Figure 3(f)). These results collectively reveal that SUMOylation is critical for the antioxidant ability of NRF2, mainly via the transcriptional activation of Cat in KLK LUAD cells.

Increased ROS Level due to NRF2 SUMOylation Deficiency Promotes KLK LUAD Cell Migration and
Invasion via JNK/c-Jun Axis. Based on our finding that ROS and H 2 O 2 levels were increased with deficient NRF2 SUMOylation (Figure 3(a)), we then asked whether increased intracellular ROS act as signaling molecules to activate pathways that promote KLK LUAD cell migration and invasion. As shown in Figure 4(a), when NRF2 expression was inhibited and the intracellular ROS level was increased, the JNK/c-Jun axis was activated in KLK LUAD cells. Expression of NRF2 wild-type inhibited the activation of the JNK/c-Jun axis, whereas expression of NRF2 K110R did not (Figure 4(a)). Furthermore, the results of RNA-Seq and quantitative real-time PCR analyses revealed that through increasing the gene expression related with cell motility and cell adhesion, JNK/c-Jun axis activation promoted the migration and invasion in KLK LUAD cells ( Figure S3D and Figures 4(b)-4(d)). In addition, we demonstrated NRF2 SUMOylation inhibited cell migration and invasion by activating Cat transcription and inactivating the JNK/c-Jun axis by reducing the cellular ROS level, in KLK LUAD H2122 cells ( Figure S4).

Discussion
In KLK LUAD, KEAP1 mutation leads to NRF2 activation, which influences many of the hallmarks of cancer and confers "NRF2 addition" [24]. High NRF2 activity renders KLK LUAD more resistant to ROS accumulation compared with KL LUAD [14]. In contrast, to promote tumorigenesis, NRF2 SUMOylation inhibits the migration and invasion of KLK LUAD cells (Figure 1), further complicating the antiversus protumorigenic roles of ROS in cancer cells. The extent of ROS increase often determines the adaptive consequences of the cellular response to the oxidative insult. We revealed in the present study that, in comparison to severe oxidative stress, which inhibited cell migration and invasion, mild oxidative stress promoted the migration and invasion of KLK LUAD cells (Figure 2(a)). SUMOylation of vimentin (VIM), a type III intermediate filament protein involved in cytoskeleton organization and cell motility, favors cell motility and migration [25]. However, whether protein SUMOylation is involved in ROS-triggered cancer cell migration and invasion has remained largely unknown. The results of the present study demonstrate that the SUMOylation of NRF2, an essential antioxidant factor, plays a role in the inhibition of KLK LUAD migration and invasion.
NRF2 is stabilized upon hyperoxidation of cysteines on KEAP1 which disrupts the binding of the two proteins [15,17]. Here, we revealed that mild oxidative stress reduced NRF2 SUMOylation (Figure 2(c)), rather than increasing NRF2 expression as seen with moderate/severe oxidative stress (Figure 2(b)), suggesting a mechanism by which  NRF2 activity is regulated independently of KEAP1. It has been reported that SUMO-specific protease 3 (SENP3) is a redox sensor that is stabilized under mild oxidative stress and consequently de-SUMOyltates p300 and hypoxia inducible factor (HIF)-1α [26,27]. Recently, Zhou et al. reported that NRF2 activity is regulated by SENP3 in laryngeal carcinoma after cisplatin-induced ROS stress [28]. Because SENP1 is also reported de-SUMOylate NRF2 [22], whether mild oxidative stress decreases NRF2 SUMOylation by increasing SENP3 expression or by increasing the expression and/or activities of other SENPs, in KLK LUAD cells, needs to be explored in future experiments.
In the present study, we showed that NRF2 SUMOylation reduced the intracellular ROS level mainly via the transcriptional activation of Cat to promote H 2 O 2 removal (Figure 3). We also found that NRF2 SUMOylation reduced the intracellular ROS level by enhancing GPX2 protein expression in KLK LUAD cells ( Figure S3C). We found that the increase in the H 2 O 2 level caused by NRF2 SUMOylation deficiency in A549 cells was in the nanomolar range. Hence, at mildly increased level, H 2 O 2 acts as a signaling molecule to activate the JNK/c-Jun pathway, which enhances cell-cell adhesion as well as focal adhesion, to promote cell migration and invasion ( Figure 4). However, how the level of increased ROS brought about by NRF2 SUMOylation deficiency specifically regulates JNK/c-Jun axis activation in KLK LUAD cells remains to be determined in future research.
A recent report showed that ROS restriction by TIGAR supports premalignant tumor initiation while restricting metastasis in pancreatic ductal adenocarcinoma, indicating that the complexity of ROS regulation underpins full malignant progression [29]. The present study demonstrated that by reducing the ROS level, NRF2 SUMOylation can regulate KLK LUAD progression in a stage-specific manner. Thus, we conclude that NRF2 SUMOylation promotes KLK LUAD tumorigenesis. However, during the initiation stage of tumor metastasis, mildly increased oxidative stress resulting from detachment may reduce NRF2 SUMOylation to induce cell migration and invasion, a potential mechanism that should be investigated in vivo in future studies. Our findings herein are helpful in understanding the Kras-dependent pathways that promote LUAD progression and provide insights for the identification of potential therapeutic targets and the development of new treatments for KRAS mutant LUAD.

Conclusions
On the basis of our current findings, we depict a model that mild oxidative stress reduces NRF2 SUMOylation to promote KLK LUAD migration and invasion ( Figure 5). Mild oxidative stress reduces NRF2 SUMOylation, which promotes KLK LUAD cell migration and invasion. Mechanistically, NRF2 SUMOylation increases the antioxidant ability of NRF2 and reduces cellular ROS levels, mainly by transcriptionally activating Cat in KLK LUAD cells. With reduced NRF2 SUMOylation, increased ROS act as signaling molecules to activate the JNK/c-Jun axis, which enhances cell mobility and cell adhesion, to promote LUAD cell migration and invasion.