Ectopic Expression of Human MutS Homologue 2 on Renal Carcinoma Cells Is Induced by Oxidative Stress with Interleukin-18 Promotion via p38 Mitogen-activated Protein Kinase (MAPK) and c-Jun N-terminal Kinase (JNK) Signaling Pathways*

Background: Ectopically expressed MutS homologue 2 (hMSH2) is a ligand of γδ T cells. Results: Oxidative stress induces ectopic expression of hMSH2 on renal carcinoma cells with IL-18 promotion via p38 MAPK and JNK pathways, promoting γδ T cell-mediated lysis of renal tumor cells. Conclusion: Ectopically expressed hMSH2 is induced under stressful conditions. Significance: We reveal a mechanism of ectopic hMSH2 expression and its contribution to γδ T cell-mediated immunosurveillance in stress. Human MutS homologue 2 (hMSH2), a crucial element of the highly conserved DNA mismatch repair system, maintains genetic stability in the nucleus of normal cells. Our previous studies indicate that hMSH2 is ectopically expressed on the surface of epithelial tumor cells and recognized by both T cell receptor γδ (TCRγδ) and natural killer group 2 member D (NKG2D) on Vδ2 T cells. Ectopically expressed hMSH2 could trigger a γδ T cell-mediated cytolysis. In this study, we showed that oxidative stress induced ectopic expression of hMSH2 on human renal carcinoma cells. Under oxidative stress, both p38 mitogen-activated protein kinase (MAPK) and c-Jun N-terminal kinase (JNK) pathways have been confirmed to mediate the ectopic expression of hMSH2 through the apoptosis-signaling kinase 1 (ASK1) upstream and activating transcription factor 3 (ATF3) downstream of both pathways. Moreover, renal carcinoma cell-derived interleukin (IL)-18 in oxidative stress was a prominent stimulator for ectopically induced expression of hMSH2, which was promoted by interferon (IFN)-γ as well. Finally, oxidative stress or pretreatment with IL-18 and IFN-γ enhanced γδ T cell-mediated cytolysis of renal carcinoma cells. Our results not only establish a mechanism of ectopic hMSH2 expression in tumor cells but also find a biological linkage between ectopic expression of hMSH2 and activation of γδ T cells in stressful conditions. Because γδ T cells play an important role in the early stage of innate anti-tumor response, γδ T cell activation triggered by ectopically expressed hMSH2 may be an important event in immunosurveillance for carcinogenesis.

Human MutS homologue 2 (hMSH2), a crucial element of the highly conserved DNA mismatch repair system, maintains genetic stability in the nucleus of normal cells. Our previous studies indicate that hMSH2 is ectopically expressed on the surface of epithelial tumor cells and recognized by both T cell receptor ␥␦ (TCR␥␦) and natural killer group 2 member D (NKG2D) on V␦2 T cells. Ectopically expressed hMSH2 could trigger a ␥␦ T cell-mediated cytolysis. In this study, we showed that oxidative stress induced ectopic expression of hMSH2 on human renal carcinoma cells. Under oxidative stress, both p38 mitogen-activated protein kinase (MAPK) and c-Jun N-terminal kinase (JNK) pathways have been confirmed to mediate the ectopic expression of hMSH2 through the apoptosis-signaling kinase 1 (ASK1) upstream and activating transcription factor 3 (ATF3) downstream of both pathways. Moreover, renal carcinoma cell-derived interleukin (IL)-18 in oxidative stress was a prominent stimulator for ectopically induced expression of hMSH2, which was promoted by interferon (IFN)-␥ as well. Finally, oxidative stress or pretreatment with IL-18 and IFN-␥ enhanced ␥␦ T cell-mediated cytolysis of renal carcinoma cells. Our results not only establish a mechanism of ectopic hMSH2 expression in tumor cells but also find a biological linkage between ectopic expression of hMSH2 and activation of ␥␦ T cells in stressful conditions. Because ␥␦ T cells play an important role in the early stage of innate anti-tumor response, ␥␦ T cell activation triggered by ectopically expressed hMSH2 may be an important event in immunosurveillance for carcinogenesis.
The DNA MMR 2 system plays a central role in maintaining genetic stability by inhibiting base-base mismatches and insertion/deletion mispairs generated during DNA replication and recombination (1,2). Homologs of bacterial MMR proteins MutS and MutL in eukaryotes have received much attention because of many discoveries showing a linkage between human cancer and MMR defects (3). The critical role of the MMR system in carcinogenesis is exemplified by the fact that hereditary nonpolyposis colorectal cancer occurs primarily due to heterozygous germ line mutations in the MMR genes: hMLH1 (MutL homolog 1), hMSH2, hMSH6 (MutS homologue 6), and hPMS2 (MutL homologue) (4). The hMLH1 promoter hypermethylation has been observed in ovarian, endometrial, gastric, and colorectal carcinoma (5). The hMSH2 mutations have a higher incidence of extracolonic tumors (6). The defects of both MMR genes are involved in renal carcinogenesis and correlate with the occurrence of microsatellite instability (7).
Under physiological conditions, hMSH2 is mainly synthesized in cytoplasm and transported into the nucleus together with MutS homologue 3 (hMSH3) or hMSH6 to maintain the DNA replication fidelity (8,9). Increased expression of hMSH2 RNA and/or protein has been reported in various malignancies (10 -14). However, the underlying mechanism that causes increased levels of hMSH2 in malignancies is yet unknown. Previously, we found an unusual expression of hMSH2 on the surface of human ovarian carcinoma cells, and ectopically expressed hMSH2 promoted the lysis of ovarian carcinoma cells by ␥␦ T cells (15). We also identified that ectopically expressed hMSH2 could be recognized by ␥␦ T cells through the interaction with both TCR␥␦ (T cell receptor ␥␦) and NKG2D (natural killer group 2 member D) (77). However, the mechanism of ectopic expression of hMSH2 in tumor cells is unclear, and ectopic hMSH2 expression may be important for the recognition of potential damage-associated molecular patterns (DAMPs) by human ␥␦ T cells (16 -18).
Oxidative stress, defined as the imbalance between the generation of reactive oxygen species (ROS) and scavenging by antioxidant defenses of the cell, is involved in carcinogenesis because it induces DNA damage and causes oxidative modification (19). The tumorigenicity of oxidative stress has been observed among various carcinomas, especially in RCC cells, which possess low antioxidant capacities as a consequence of the absence of peroxisomes and the changed redox status (20,21). The defective antioxidant capacity also produces a prooxidant environment in RCC cells and promotes the process of renal tumor genesis (20,22).
Because hMSH2 or hMSH6 participates in removing complex DNA lesions produced in oxidative stress (23), we hypothesize that oxidative stress may be responsible for ectopic expression of hMSH2 on tumor cells. In the present study, we established a glucose oxidase (GO)-mediated oxidative stress model in RCC cells through oxidizing ␤-D-glucose and generating hydrogen peroxide (H 2 O 2 ) continuously at a relatively low level (24). We found that hMSH2 was constitutively expressed at a low level on the surface of RCC cells. However, ectopic expression of hMSH2 was increased in GO-mediated oxidative stress through p38 MAPK and JNK pathways in RCC cells. The regulating role of p38 MAPK and JNK pathways was mediated by a node kinase, ASK1, upstream and the stress-sensitive activating transcription factor 3 (ATF3) downstream of both pathways. Moreover, proinflammatory cytokine IL-18 produced by RCC cells was a crucial stimulator inducing ectopic expression of hMSH2 in oxidative stress. Meanwhile, exogenous IFN-␥ further promoted ectopic expression of hMSH2 on the surface of RCC cells, and ␥␦ T cells produced a high level of IFN-␥ during oxidative stress. We also showed that oxidative stress or IL-18 enhanced ␥␦ T cell-mediated lysis of RCC cells. Taken together, our finding provides a better understanding of oxidative stress-induced ectopic expression of hMSH2 on the surface of RCC cells. As a result, ectopically expressed hMSH2 may act as a candidate molecule of DAMPs recognized by ␥␦ T cells, which indicates that ␥␦ T cells may perform immunosurveillance of hMSH2-bearing tumor cells in an oxidative stress environment.

EXPERIMENTAL PROCEDURES
Chemicals-The glucose oxidase (G7141) catalyzes the oxidation of ␤-D-glucose into D-glucono-␤-lactone and H 2 O 2 , with molecular oxygen as an electron acceptor. N-Acetyl-L-cysteine (NAC) (A7250), a ROS scavenger, was selected to block the H 2 O 2 effect. Calyculin A from Discodermia calyx (C5552) was used as an inhibitor of protein phosphatases types 1 and 2A. The reagents above were purchased from Sigma-Aldrich. Pharmacological inhibitors of p38 MAPK (SB203580) and JNK (420119) were purchased from Merck, and the inhibitors of the MAPK/ERK kinase (MEK) and ERK pathway (U0126) were purchased from Selleck. The 1-methylprolyl 2-imidazolyl disulfide (PX-12) was purchased from TOCRIS.
Cell Culture and Treatment-Human renal carcinoma cell line G401, melanoma cell line A375, lymphoma cell line Jurkat, and leukemia cell line K562 were maintained in RPMI 1640 medium (Invitrogen) supplemented with 10% FCS (Invitrogen). Human renal carcinoma cell line A498 was maintained in minimum Eagle's medium supplemented with 10% FBS (Hyclone). Human renal proximal tubule epithelial cell line HK-2 was maintained in DMEM/F-12 medium supplemented with 10% FCS (Invitrogen). All cell lines were obtained from the Cell Center, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China). For constructing the oxidative stress model, the cells were activated by different concentrations of glucose oxidase (G7141; Sigma) in medium at 37°C, 5% CO 2 for prospective times (24).
␥␦ T Cell Expansion in Vitro-Peripheral blood mononuclear cells from healthy donors were isolated by Ficoll-Hypaque (TBD, Tianjin, China) centrifugation. ␥␦ T cells were expanded as described previously (25). Briefly, 24-well plates were coated with 500 l of purified anti-TCR␥␦ antibody (Ab) (IMMU510; 1 g/ml; Immunotech, Beckman Coulter) at 37°C, 5% CO 2 for 2 h. Peripheral blood mononuclear cells were transferred to the plate and cultured in RPMI 1640 medium supplemented with 10% FCS and recombinant human IL-2 (200 IU/ml; Beijing Read United Cross Pharmaceutical Co., Ltd., Beijing, China). After 2 weeks of culture, the purity of ␥␦ T cells was more than 90% as assessed by flow cytometry analysis and ready for further use.
Cytotoxicity Assay-Cell cytotoxicity was determined by the CytoTox 96 non-radioactive cytotoxicity assay kit (Promega). G401 cells were preincubated with the indicated amounts of GO, IL-18, or IL-18 plus IFN-␥. ␥␦ T cells of appropriate number were incubated with G401 cells (1 ϫ 10 5 /ml) in triplicates for 4 h at 37°C. The release of lactate dehydrogenase and percentage of cytotoxicity were determined by the absorbance of supernatants at 490 nm.
Enzyme-linked Immunosorbent Assay (ELISA) Analyses-The IL-18 in the oxidative stress-stimulated RCC, A375, Jurkat, or K562 cell culture supernatants and IFN-␥ or tumor necrosis factor-␣ (TNF-␣) in the oxidative stress-stimulated ␥␦ T cells culture supernatants were determined by ELISA kits (R&D Systems). The assay procedure was performed according to the manufacturer's instructions.
Western Blotting-Cell lysates were prepared by Cyto-Buster TM protein extraction reagent (71009, Novagen) for total protein and by the NucBuster protein extraction kit (71183-3, Novagen) for nucleoprotein. Halt Protease and Phosphatase Inhibitor Single-Use Mixture, EDTA-Free (Thermo) were used to prevent enzymatic protein degradation during extraction procedures. The protein concentration was quantified by the Pierce BCA protein assay kit (23227, Thermo). An equal amount of proteins per sample was separated by 12% SDS-PAGE and blotted onto nitrocellulose membranes. The blots were treated with PBS containing 5% skim milk or bovine serum albumin overnight and probed with primary Ab at room temperature for 2 h, followed by the relevant secondary Ab for 1 h. Protein bands were visualized using SuperSignal West Pico Chemiluminescent Substrate (NCI5079, Thermo). Monoclonal antibody (mAb) against phospho-JNK/phospho-p38/phosphoc-Jun (Cell Signling Technology), anti-human IL-18R␣ (R&D Systems), ATF3 (Abgent), hMSH2 (ProteinTech Group, Inc.), or ␤-actin (Signalchem) and polyclonal antibody to histone H3 (Abcam) were used as primary Abs.
Flow Cytometry-Cells were washed with PBS containing 1% BSA, stained with relevant Abs, and incubated at 4°C for 30 min. Then the cells were washed and resuspended in 300 l of PBS containing 1% formaldehyde. FITC-conjugated anti-human IL-18R␣ Ab (catalog no. 313810), APC-conjugated antihuman MICA/B Ab (catalog no. 320908), and the respective isotype control mAbs were purchased from Biolegend (San Diego, CA). Ab FITC-conjugated anti-TCR␥␦ (IMMU510) was purchased from Immunotech. When purified anti-MSH2 Ab (H-300) (sc-22771) and the isotype control polyclonal Abs (Santa Cruz Biotechnology, Inc.) were used as primary Abs, FITC-conjugated goat anti-rabbit secondary Ab (Zhongshan, China) was then added and incubated for 30 min at 4°C. For intracellular staining, the cells incubated with 20 ng/ml phorbol 12-myristate-13-acetate (Sigma), 0.5 g/ml ionomycin (Sigma), and 3 g/ml brefeldin A (eBioscience, San Diego, CA) acted as the positive control. After 4 h, the cells were collected and stained for surface markers. Then the cells were fixed, permeabilized, and incubated with intracellular Abs. All samples were analyzed on an Accuri C6 flow cytometer (BD Biosciences). Data analysis was carried out with FlowJo Software (version 7.6.4, Tree Star, Inc.).
Statistical Analysis-Data were expressed as the mean Ϯ S.D. and were representative of at least three independent experiments. One-way analysis of variance (SPSS version 16.0 software) was used to determine significant differences between groups. Differences were considered statistically significant at a p value of less than 0.05.

Oxidative Stress Induces Ectopic Expression of hMSH2 on
Surface of RCC Cells-To determine the ectopic expression of hMSH2 in stress, renal carcinoma cell lines G401 and A498 and a renal proximal tubular cell line, HK-2, were chosen because these cell lines from kidney exhibited a constitutive ectopic expression of hMSH2 on surfaces in normal cultures (Fig. 1A). In response to oxidative stress, ectopically expressed hMSH2 was induced on G401 cells treated by different concentrations of GO for 24 or 36 h, respectively (Fig. 1B). HK-2 and A498 cells showed similar results (data not shown). GO stimulation at 100 ng/ml for 36 h induced the strongest expression in all three cell lines (Fig. 1, B and C). A down-regulation of GO-induced hMSH2 ectopic expression was observed on G401, A498, and HK-2 cells by the addition of NAC, a ROS scavenger that strongly blocks the H 2 O 2 effect (Fig. 1, B and C). Because G401 cells exhibited the most obvious changes of ectopically expressed hMSH2 mediated by GO or NAC, we subsequently chose G401 cells as the cell model to investigate the mechanisms underlying ectopic expression of hMSH2 in oxidative stress. An increase of hMSH2 mRNA was observed as early as 1 h and lasted to 48 h by exposing G401 cells to 100 ng/ml GO-mediated stress (Fig. 1D). Accordingly, the total protein level of hMSH2 was up-regulated in G401 cells stimulated by GO (Fig. 1E). These results demonstrate that ectopic expression of hMSH2 on the surface of RCC cells is induced in response to oxidative stress.
We also found that oxidative stress-induced ectopic expression of hMSH2 showed a similar change pattern to surface expression of MICA/B, classical stress-sensitive ligands of NKG2D in oxidative stress (Fig. 1, F-I), and both of them could be reduced markedly by NAC processing (Fig. 1, F-I). Because expression of MICA/B is restricted or absent in normal tissues and up-regulated in stress and diseases (26,27), it further confirms the stress-inducible characteristic of ectopic expression of hMSH2 on RCC cells.

Both p38 MAPK and JNK Pathways Are Involved in Regulating Ectopic Expression of hMSH2 in RCC Cells in Oxidative
Stress-The MAPK family plays an important role in oxidative stress-induced cellular events (28). Among the MAPK family, the activation of ERKs usually causes cell survival or differenti-ation, whereas the activation of p38 MAPK and JNK pathways links to promote cell apoptosis, particularly under oxidative stress conditions (29,30). As shown in Fig. 2, A-C, GO-induced phosphorylation of p38 MAPK (Thr-180/Tyr-182), JNK (Thr-183/Tyr-185), and c-Jun was elevated, indicating that both p38 MAPK and JNK pathways were activated, and the MEK/ERK pathway might also present as an active state in G401 cells exposed to oxidative stress, consistent with previous studies   [31][32][33][34]. Quantitative RT-PCR detection showed that GO-induced hMSH2 mRNA was decreased in G401 cells when pretreated with either the specific inhibitor of p38 MAPK (SB203580) or JNK pathway (420119) (Fig. 2D). However, there was no significant changes of GO-induced hMSH2 mRNA pretreated with the MEK/ERK pathway inhibitor (Fig. 2D). Importantly, pretreatment with either specific inhibitor for the p38 MAPK or JNK pathway suppressed ectopic expression of hMSH2 on the surface of G401 cells induced by GO (Fig. 2, E  and F), whereas inhibition of the MEK/ERK pathway showed no obvious effects on ectopic expression of hMSH2 in oxidative stress (supplemental Fig. 1). The total expression of hMSH2 was also down-regulated when blocking the role of p38 MAPK or JNK pathway in GO-treated G401 cells (Fig. 2, G-J). The above data demonstrate the important roles of p38 MAPK and JNK pathways in regulating ectopic expression of hMSH2 during oxidative stress.
p38 MAPK and JNK Pathways Regulate Ectopic Expression of hMSH2 through ASK1, Node Kinase of Both Pathways-To further validate the functional involvement of the p38 MAPK and JNK pathways in regulating ectopic expression of hMSH2 in stress, we next examined the role of ASK1, a member of the mitogen-activated protein kinase kinase kinase (MAP3K) family that could activate both JNK and p38 MAPK pathways in response to oxidative stress (35). We used the combination of PX-12 and a low level of glucose (1 g/liter) to activate ASK1 by disrupting the interaction between thioredoxin and ASK1 (36 -38) and used NAC as an antioxidant to inhibit the PX-12-induced disassociation of thioredoxin from ASK1 and JNK activation (38). As shown in Fig. 2, K-M, both the total and ectopic expression of hMSH2 were enhanced in G401 cells when ASK1 was activated after treatment with PX-12. Meanwhile, this increase was suppressed substantially by inactivating ASK1 with NAC. To further confirm the role of activated ASK1 in enhancing ectopic expression of hMSH2, we used calyculin A, an inhibitor of protein phosphatase type 1 and 2A to inhibit the primary activation of ASK1 through inhibiting H 2 O 2 -induced dephosphorylation of ASK1 Ser-967 (39). It showed that PX-12-induced surface expression of hMSH2 was suppressed when dephosphorylation of ASK1 was inhibited by calyculin A (Fig. 2M). The effect of ASK1 on ectopic expression of hMSH2 further demonstrates the role of p38 MAPK and JNK pathways in regulating ectopic expression of hMSH2 in RCC cells.

ATF3 Acts as Executor Inducing Ectopic Expression of hMSH2 Regulated by p38 MAPK and JNK Pathways in Oxidative
Stress-Activities of the activator protein 1 (AP-1) family can be stimulated by many environmental and cellular factors, including oxidative stress (40,41). The synthesis of ATF3, a member of the AP-1 subfamily activating transcription factors, is quite robust relative to the levels of other oxidative stress-induced AP-1 proteins (42). We found that transcription factor binding sites for the AP-1 family presenting in the regulatory element of hMSH2 were matched to ATF3 transcription factor binding sites (43)(44)(45): AP-1, Ϫ611 to Ϫ605 (5Ј-TGAATCA-3Ј) and Ϫ297 to Ϫ291 (5Ј-TGAGTAA-3Ј); ATF/cAMP-response element, Ϫ1156 to Ϫ1149 (5Ј-TGACGTCA-3Ј). Therefore, we tested if ATF3 is involved in the regulation of ectopic expression of hMSH2 during oxidative stress. Interestingly, GO treatment enhanced the transcription of ATF3 significantly at 1-3 h compared with the level in control G401 cells (Fig. 3A). When the GO-mediated oxidative stress was blocked by NAC, the ATF3 transcription was down-regulated (Fig. 3A). Consistent with it, the total expression of ATF3 was also up-regulated in G401 cells exposed to oxidative stress (Fig. 3B). To investigate the potential role of ATF3 in regulating ectopic expression of hMSH2 in response to oxidative stress, ATF3 was knocked down in G401 cells by siRNAs (Fig. 3, C and D). In response to GO-mediated oxidative stress, knocking down ATF3 for 48 h blocked the most induction of ectopically expressed hMSH2 on G401 cells (Fig. 3, E and F). Similar ATF3-RNAi effects under GO-mediated oxidative stress were also observed after transfecting with ATF3 siRNAs for 24 or 36 h (supplemental Fig. 2). However, the surface expression of MICA/B induced by oxidative stress was not affected significantly by ATF3 knockdown (Fig. 3, G and H). Although hMSH2 and MICA/B are both stress-inducible proteins recognized by ␥␦ T cells, they show different responses of surface expression mediated by ATF3 in stress, and ATF3 appears to play a specific role in regulating ectopic expression of hMSH2 in oxidative stress.

Proinflammatory Cytokine IL-18 Produced by RCC Cells in Oxidative Stress Is Prominent Stimulator Inducing Ectopic
Expression of hMSH2-To further seek the crucial factors inducing ectopic expression or distribution of hMSH2 during oxidative stress, we noted that ROS could trigger the activation of inflammasome NLRP3/NALP3 and recruited the caspase-1 (46). The inflammasome/caspase-1 axis is essential for the maturation and release of proinflammatory cytokines, such as IL-1␤ and IL-18 (46,47). Importantly, IL-18 treatment enhances the expression of UL-16-binding protein 2 (ULBP2), a stress-inducible ligand for NKG2D in leukemia cells (48). Here, we examined whether IL-18 is a potential factor inducing ectopic expression of hMSH2 in oxidative stress. The present study showed that IL-18 was released increasingly into supernatants of GO-stimulated RCC cells (Fig. 4, A and B). IL-18R␣ was constitutively expressed on normal cultured G401 cells and increased in oxidative stress (Fig. 4C). Similar results were shown in A498 cells (data not shown). Exogenous IL-18 alone could up-regulate ectopic expression of hMSH2 in a dose-dependent manner (Fig. 4D). All these data indicate the potential ability of IL-18 in inducing ectopic expression of hMSH2 on RCC cells during oxidative stress. To assess the role of endogenous IL-18 in oxidative stress, siRNA-IL-18R␣ was used to knock down the expression of endogenous IL-18R␣ in G401 cells. Both levels of IL-18R␣ mRNA and protein were significantly reduced after transfection with siRNA-IL-18R␣ (Fig. 4, E-G). Of note, no significant up-regulation of hMSH2 transcription or surface expression on G401 cells was found in siRNA-IL-18R␣-treated cells (Fig. 4, H and I). The total protein level of hMSH2 in G401 cells treated with siRNA-IL-18R␣ also showed a down-regulated pattern after GO treatment (Fig. 4, J  and K). These results suggest that endogenous IL-18 released by RCC cells during oxidative stress has a direct effect on inducing ectopic expression of hMSH2.

Ectopic Expression of hMSH2 Is Promoted by IFN-␥ and Proinflammatory Factors Enhance Lysis of RCC Cells by ␥␦ T Cells in Oxidative Stress-Our
previous studies have demonstrated that hMSH2-bearing tumor cells were susceptible to V␦2 ␥␦ T cell-mediated cytolysis, in which IFN-␥ and TNF-␣ were primary factors produced by ␥␦ T cells (15,25,49). It led us to examine the level of these proinflammatory cytokines produced by GO-stimulated ␥␦ T cells and their effects on ectopic expression of hMSH2. When anti-pan-TCR␥␦ mAb-expanded ␥␦ T cells were treated with different concentrations of GO, an increased amount of IFN-␥ was released into supernatants compared with ␥␦ T cells in normal cultures (Fig. 5A). The largest amplitude of IFN-␥ in supernatant was increased from 205.78 (in normal culture) to 1027.09 pg/ml (treated with GO 50 ng/ml for 36 h), about a 5-fold increase (Fig. 5A). However, there was no obvious change of TNF-␣ upon GO treatment (Fig. 5B). Consistent results of the intracellular staining assay showed that the increased level of IFN-␥ was mainly produced by ␥␦ T cells in GO-mediated oxidative stress (Fig. 5C). Then exogenous IFN-␥ alone promoted GO-induced ectopic expression of hMSH2 on G401 cells, and this up-regulation was inhibited by the administration of NAC (Fig. 5, D and E). Moreover, the lysis of G401 cells mediated by ␥␦ T cells could be enhanced by treatment with GO, IL-18, or IL-18 plus IFN-␥ (Fig. 5F). These data suggest that IFN-␥ released by ␥␦ T cells in oxidative stress might promote ectopic expression of hMSH2, and the proinflammatory cytokine IL-18 or IFN-␥ from oxidative stress promotes the lysis of RCC cells by ␥␦ T cells. Here, the autocrine of IL-18 in normal or stressful conditions was detected in supernatants of A375 and RCC cell lines (Fig.  6A), and all of the above tumor cells exhibited a constitutive ectopic expression of hMSH2 on their surfaces (Fig. 6B). Importantly, oxidative stress further enhanced the ectopic expression of hMSH2 on A375, Jurkat, and K562 cells (Fig. 6B). Moreover, all of these tumor cells expressed IL-18R␣ in normal conditions (Fig. 6C), and exogenous IL-18 could enhance ectopic expression of hMSH2 on these tumor cells (Fig. 6D). These data suggest that ectopic expression of hMSH2 is broad in different tumor cells, and IL-18 induced by oxidative stress may be an important stimulator to promote ectopic hMSH2 expression on tumor cells.

DISCUSSION
Because hMSH2 participates in the removal of mispairs originating from damaged DNA in an oxidative environment (50), we have attempted to find a linkage between oxidative stress and ectopic expression of hMSH2 on tumor cells. To mimic the stressful status in patients with several types of tumors, we chose GO-mediated oxidative stress, which is characterized by forming a continuous H 2 O 2 streams into cultures at constant levels (24,(51)(52)(53). Our present findings demonstrate that expression of hMSH2 mRNA could be enhanced in G401 cells exposed to oxidative stress. Accordingly, ectopic expression of hMSH2 is induced by GO treatment in a dose-dependent manner and blocked by a ROS scavenger, NAC. It suggests that hMSH2 possesses the characteristic of inducible ectopic expression on the surface of RCC cells during oxidative stress. In addition, inducible ectopic expression of hMSH2 was also observed on other tumor cells, including melanoma A375, lymphoma Jurkat, and leukemia K562 cells (Fig. 6B). As a tumorassociated protein promoting V␦2 ␥␦ T cell-mediated lysis (15), the inducibility of ectopically expressed hMSH2 in tumor cells indicates that it may activate many more ␥␦ T cells in certain pathological conditions, such as the oxidative stress status in patients with renal tumors, and these activated ␥␦ T cells could further recognize other ligands presenting on the surface of tumor cells and display anti-tumor cytotoxic activity. Here, the enhanced anti-tumor immunity of ␥␦ T cells mediated by ectopic hMSH2 expression is consistent with our observation that the lysis of RCC cells by ␥␦ T cells is enhanced in oxidative stress or by proinflammatory cytokines, which could induce ectopic expression of hMSH2. Therefore, ectopically expressed hMSH2 might act as one of potential DAMPs recognized by ␥␦ T cells in stressful conditions. Stressful states, including oxidative stress, can activate MAPK pathways (54 -56). Exposure of retinal pigment epithelial cells to H 2 O 2 induces apoptosis through a Racl/JNK/p38 signaling cascade (57), and H 2 O 2 also causes apoptosis by upregulating both JNK and p38 in testicular germ cells (58). We showed in this study that transcription factor binding sites of the AP-1 family presented in the regulatory element of hMSH2, and p38 MAPK and JNK pathways were activated during oxidative stress. Phosphorylated c-Jun indicated that the MEK/ ERK pathway might also be active in GO-mediated oxidative stress. However, p38 MAPK and JNK but not MEK/ERK pathways indeed promoted surface expression of hMSH2 on RCC cells. The regulatory effect of p38 MAPK and JNK pathways was further confirmed by stimulating or inhibiting of ASK1, a kinase activating both MKK3/MKK6-p38 and MKK4/MKK7-JNK signaling cascades in environmental stresses (59,60).
On the other hand, we showed that ATF3, a stress-sensitive transcription factor locating downstream of the p38 MAPK and JNK pathways (42,61), played a crucial role in ectopic expression of hMSH2 in oxidative stress. When endogenous ATF3 is knocked down in G401 cells, there is no significant up-regulation of ectopically expressed hMSH2 even in oxidative stress, suggesting that ATF3 may act as an executor to promote ectopic expression of hMSH2 in an oxidative environment, although the specific mechanism needs to be clarified.
To further determine key factors regulating ectopic expression of hMSH2 on the surface of RCC cells in oxidative stress, we examined the role of IL-18, which could be produced largely by ROS-activated inflammasome NLRP3/NALP3 (46). Meanwhile, IL-18 shows the central role in promoting the inflammation and oxidative stress that lead to renal dysfunction and damage (62). Our data show that, accompanied by more IL-18 secretion, the expression of IL-18R␣ is up-regulated on the surface of GO-stimulated RCC cells, which leads these cells to be more sensitive to endogenous IL-18 during oxidative stress. Meanwhile, exogenous IL-18 alone effectively up-regulates surface expression of hMSH2, and GO-induced ectopic expression of hMSH2 could be inhibited by knockdown of IL-18R␣ in G401 cells. In order to identify the regulating effect of IL-18 on ectopic expression of hMSH2, we also tested the change of ectopically expressed hMSH2 on other tumor cell types, such as A375, Jurkat, and K562 cells, when treated with IL-18. Several reports have mentioned that these tumor cells could produce IL-18 under certain conditions (63)(64)(65)(66). Interestingly, although the autocrine of IL-18 is only observed in supernatants of normal or GO-stimulated G401 and A375 cells in our study, treatment with IL-18 could induce significantly ectopic expression of hMSH2 on all of the above tumor cells. Thus, we envision that abundant IL-18 in the supernatants of stress-stimulated RCC or melanoma cells is an important stimulator inducing ectopic expression of hMSH2 on multiple types of tumor cells, including RCC, lymphoma, and leukemia. Because most hematological cancers are sensitive to the anti-tumor cytotoxicity of ␥␦ T cells (67)(68)(69) and the production of IFN-␥, GM-CSF, and TNF-␣ by ␥␦ T cells is strongly enhanced when treated with IL-18 (70), it suggests that IL-18 released by tumor cells in oxidative stress may promote both recognition and clearance of tumor cells by ␥␦ T cells.
We also consider proinflammatory cytokines produced by ␥␦ T cells during lysis of tumor cells (25,49,71,72) and investigate their potential effects on regulating ectopic expression of hMSH2. Here, IFN-␥ but not TNF-␣ is highly produced by GO-treated ␥␦ T cells, and exogenous IFN-␥ indeed amplifies GO-induced ectopic expression of hMSH2 on G401 cells. The fact that proinflammatory cytokines IL-18 and IFN-␥ in combination or alone are involved in the production of ROS in various cell types (73)(74)(75)(76) indicates that a potential positive feedback by ROS and proinflammatory cytokines may regulate integrally ectopic expression of hMSH2 in oxidative stress. As shown in Fig. 7, oxidative stress induces ectopic expression of hMSH2 on RCC cells. Meanwhile, ROS existing in oxidative stress promote the production of IL-18 and IFN-␥ by tumor cells and ␥␦ T cells, respectively. IL-18 alone could induce ectopic expression of hMSH2 on RCC cells, and both proinflammatory cytokines in turn may maintain the high level of ROS, which subsequently induces highly ectopic expression of hMSH2 presenting on the surface of RCC cells.
In conclusion, ectopically expressed hMSH2 on tumor cells possesses the inducibility in oxidative environments. The p38 MAPK and JNK pathways mainly control ectopic expression of hMSH2 on RCC cells through ASK1 upstream and stress-sen- sitive ATF3 downstream of both pathways. IL-18 produced by RCC cells is a crucial stimulator inducing ectopic expression of hMSH2 in oxidative stress, and IFN-␥ produced by ␥␦T cells might promote the above effects. Furthermore, IL-18, IFN-␥ and ROS constitute a positive feedback, maintaining a high level of ectopically expressed hMSH2 on RCC cells, which promotes the lysis of RCC cells by ␥␦ T cells. All of these results suggest that inducible ectopic expression of hMSH2 on tumor cells may act as an alert signal to ␥␦ T cell-mediated immunosurveillance in stressful environments. FIGURE 7. A positive feedback model for regulating ectopic expression of hMSH2 in RCC cells during oxidative stress. A, high level of ROS in oxidative stress causes excessive production of proinflammatory cytokines IL-18 and IFN-␥ by RCC and ␥␦ T cells, respectively. IL-18 is shown as a stimulator inducing ectopic expression of hMSH2 in oxidative stress. IFN-␥ promotes the above trend presenting on the surface of RCC cells. The combination of both proinflammatory cytokines in turn maintains a high level of ROS, which may further induce ectopic expression of hMSH2. B, in oxidative stress-stimulated RCC cells, ectopic expression of hMSH2 is controlled by p38 MAPK and JNK pathways. A node kinase ASK1 upstream and stress-sensitive ATF3 downstream of both pathways participate in the intracellular regulation of ectopic expression of hMSH2. Because hMSH2-bearing RCC cells are susceptible to ␥␦ T cell-mediated lysis, this positive feedback indicates a potential contribution of ectopically expressed hMSH2 to ␥␦ T cell-mediated immunosurveillance of tumor cells in stressful conditions.