Ectopically Expressed Human Tumor Biomarker MutS Homologue 2 Is a Novel Endogenous Ligand That Is Recognized by Human γδ T Cells to Induce Innate Anti-tumor/Virus Immunity*

Background: Ectopically expressed hMSH2 is a tumor biomarker; however, the mechanism of its recognition is unclear. Results: hMSH2 interacted with both TCRγδ and NKG2D on Vδ2 T cells, resulting in Vδ2 T cell activation. Its expression was up-regulated by EBV infection. Conclusion: hMSH2 is a ligand for TCRγδ and NKG2D. Significance: Recognition of hMSH2 by γδ T cells induces innate anti-tumor/virus immunity. Human (h) MutS homologue 2, a nuclear protein, is a critical element of the DNA mismatch repair system. Our previous studies suggest that hMSH2 might be a protein ligand for TCRγδ. Here, we show that hMSH2 is ectopically expressed on a large panel of epithelial tumor cells. We found that hMSH2 interacts with both TCRγδ and NKG2D and contributes to Vδ2 T cell-mediated cytolysis of tumor cells. Moreover, recombinant human MSH2 protein stimulates the proliferation and IFN-γ secretion of Vδ2 T cells in vitro. Finally, hMSH2 expression is induced on the cell surface of Epstein-Barr virus-transformed lymphoblastoid cell lines, and the induction increases the sensitivity of these lymphoblastoid cell lines to γδ T cell-mediated cytolysis. Our data suggest that hMSH2 functions as a tumor-associated or virus infection-related antigen recognized by both Vδ2 TCR and NKG2D, and it plays a role in eliciting the immune responses of γδ T cells against tumor- and virus-infected cells. The recognition of ectopic surface-expressing endogenous antigen by TCRγδ and NKG2D may be an important mechanism of innate immune response to carcinogenesis and viral infection.


Human (h) MutS homologue 2, a nuclear protein, is a critical element of the DNA mismatch repair system. Our previous studies suggest that hMSH2 might be a protein ligand for TCR␥␦. Here, we show that hMSH2 is ectopically expressed on a large panel of epithelial tumor cells. We found that hMSH2 interacts with both TCR␥␦ and NKG2D and contributes to V␦2 T cell-mediated cytolysis of tumor cells. Moreover, recombinant human MSH2 protein stimulates the proliferation and IFN-␥ secretion of V␦2 T cells in vitro. Finally, hMSH2 expression is induced on the cell surface of Epstein-Barr virus-transformed lymphoblastoid cell lines, and the induction increases the sensitivity of these lymphoblastoid cell lines to ␥␦ T cell-mediated cytolysis. Our data suggest that hMSH2 functions as a tumorassociated or virus infection-related antigen recognized by both V␦2 TCR and NKG2D, and it plays a role in eliciting the immune responses of ␥␦ T cells against tumor-and virus-infected cells. The recognition of ectopic surface-expressing endogenous antigen by TCR␥␦ and NKG2D may be an important mechanism of innate immune response to carcinogenesis and viral infection.
Human ␥␦ T cells account for a minor fraction of human peripheral T lymphocytes. The innate cell-like feature and MHC-independent antigen recognition define these cells as a special status in tumor surveillance and pathogen defense (1). However, the molecular mechanisms underlying tumor/infected cell recognition by ␥␦ T cells are poorly understood because of the limited number of antigens/ligands recognized by ␥␦ T cells that have been identified so far (2)(3)(4).
hMSH2 is a critical element of the highly conserved DNA mismatch repair system, normally located in the nucleus and dimerized with hMSH3 or hMSH6 to form complexes essential for the maintenance of genome integrity (19). Inherited and acquired defects in hMSH2 are closely related to the pathogenesis of hereditary nonpolyposis colon cancer as well as various sporadic cancers (20). Recent findings indicate that hMSH2 also plays important roles in DNA damage signaling, apoptosis, and antibody class switch recombination in B cells (21)(22)(23). Previously, with a CDR3␦ peptide-based affinity screening system, we identified hMSH2 as a putative tumor-associated protein ligand for human TCR␥␦ (24).
In this study, we further investigated the recognition of hMSH2 by human ␥␦ T cell receptors and the role of hMSH2 in ␥␦ T cell-mediated immune responses. We found that hMSH2, normally located in the nucleus, was ectopically expressed on the cell surface of different epithelial tumor cells, and its expression was induced by EBV transformation. To the best of our knowledge, our results provide the first evidence that cell surface-expressed hMSH2 is a ligand for both V␦2 TCR and NKG2D receptors and engages in anti-tumor and anti-virus immunity by enhancing the ␥␦ T cell-mediated cytolysis. * This work was supported by Grant 30930083 from the National Natural Science Foundation. 1
Normal B lymphocytes were isolated from PBMCs using a human B cell isolated kit II (Miltenyi Biotec). The purity of enriched B cells was evaluated by flow cytometry using a phycoerythrin-conjugated CD19 mAb (Miltenyi Biotec).
Flow Cytometry-For cell surface hMSH2 staining, 1 ϫ 10 6 cells resuspended in 50 l of PBS containing 1% BSA were incubated with 2 g of anti-MSH2 (H-300, Santa Cruz Biotechnology, Santa Cruz, CA) or rabbit IgG for 1 h at 4°C, followed by staining with FITC-conjugated goat anti-rabbit secondary Ab (The Jackson Laboratory) for 30 min shielded from the light. Suspension cells were pretreated with 2 l of FcR blocking reagent (Miltenyi Biotec) before adding primary antibody. Cells were fixed with 1% formaldehyde and analyzed on EPICS XL (Beckman Coulter, Bree) or C6 flow cytometer (Accuri Cytometers, Ann Arbor, MI). Data were analyzed with FCS Express Version 3.0 software (De Novo Software, Los Angeles). The percentage positive was calculated by Overton subtraction method. For one-or two-color staining of other cell surface antigens, cells were incubated with FITC/phycoerythrin-conjugated mAbs or isotype-matched control Abs for 30 min at 4°C.
Synthesis and Reverse Transfection of siRNA Duplexes-Three pairs of siRNA duplexes were synthesized from Invitrogen. Sequences targeting hMSH2 gene (NM 000251.1) were as follows: siRNA I, 5Ј-TAG GAC TGT GTG AAT TCC CTG ATA A-3Ј; siRNA II, 5Ј-CAG CTA GAT GCT GTT GTC AGC TTT G-3Ј; and siRNA III, 5Ј-CAA TTG AAA GGA GTC TCC ACG TTC A-3Ј. Stealth TM RNAi universal negative control med GC duplex (Invitrogen) was used as mock control. siRNA duplexes were reverse-transfected into tumor cells at a final concentration of 10 nM in a 6-well format according to the manufacturer's instructions. Briefly, 30 pmol of siRNA or negative control duplex was incubated with 5 l of Lipofectamine TM RNAiMAX (Invitrogen) in 500 l of Opti-MEM I medium (Invitrogen) for 10 -20 min at room temperature. 1.5-2.0 ϫ 10 5 HeLa, 803, or 3D5 cells were suspended in 2500 l of complete growth medium without antibiotics and added to each well with siRNA duplex-Lipofectamine TM RNAiMAX complexes. The delivery of siRNAs was measured by parallel transfection with BLOCK-iT TM Alexa Fluor red fluorescent oligonucleotide (Invitrogen) and analyzed by immunofluorescence microscopy at 6 -12 h after transfection. Efficiency of gene knockdown was measured by quantitative real time RT-PCR and Western blot as described below; the cell surface expression of hMSH2 was analyzed by flow cytometry and confocal microscopy at 48 or 72 h after the transfection.
Real Time PCR-Total RNA of target cells (2 ϫ 10 6 ) was isolated using TRIzol reagent (Promega, Madison, WI). cDNA was synthesized using oligo(dT) (Promega) and Moloney murine leukemia virus reverse transcriptase (Promega) in the reverse transcription reaction. Quantitative real time PCR was performed in a volume of 20 l containing cDNA template, oligonucleotide primers, and SYBR Green PCR master mix (Applied Biosystems, Warrington, UK) using a 7500 real time PCR system (Applied Biosystems). Primers used for the amplification of hMSH2 and ␤-actin (endogenous control) were as follows: hMSH2, 5Ј-TTCATGGCTGAAATGTTGGA (forward) and 5Ј-ATGCTAACCCAAATCCATCG (reverse); and ␤-actin, 5Ј-AGAAAATCTGGCACCACACC(forward) and 5Ј-TAGCACAGCCTGGATAGCAA (reverse). Cycling conditions were as follows: 95°C for 10 min, 40 cycles at 95°C for 15 s and at 55°C for 45 s. Data were analyzed by Sequence Detector Version 1.2 analysis software (Applied Biosystems).
Western Blot-Total protein from 2 ϫ 10 6 interference or mock control cells was extracted by Cytobuster protein extraction reagent (Merck) and measured by Micro BCA protein assay kit (Pierce). The protein samples (20 g) were separated on a 10% SDS-polyacrylamide gel and immunoblotted with mouse anti-hMSH2 mAb (65021-1-Ig, Protein Tech Group, Chicago) for 2 h and incubated with HRP-conjugated goat antimouse secondary Ab (The Jackson Laboratory) for 1 h at room temperature. Protein bands were visualized using Supersignal West Pico chemiluminescent substrate (Thermo Scientific). Anti-␤-actin mAb (C4, Santa Cruz Biotechnology) served as endogenous control.
Confocal Microscopy-Cells of appropriate number (2-3 ϫ 10 5 ) were grown on glass coverslips in 24-well plates overnight, fixed by PBS containing 4% paraformaldehyde for 10 -15 min, and blocked in PBS containing 0.5% BSA for 30 min at 4°C. Cells were then immunolabeled with anti-MSH2 (N-20, Santa Cruz Biotechnology) or rabbit IgG for 1 h and incubated with FITC-conjugated goat anti-rabbit secondary Ab for 30 min at 4°C. Nuclei were detected by DAPI (Sigma) staining for 5 min. Images were taken with a Leica DMIRE2 inverted microscope fitted with a Leica TCS SP2 SE confocal imager. The objective was ϫ40 with a numerical aperture of 1.25 (oil objective).
Surface Plasmon Resonance (SPR)-SPR experiments were performed using Biacore 3000 (GE Healthcare). Recombinant human NKG2D-Fc (R & D Systems) and IgG1 Fc (Sino Biological Inc, China) were coupled to the research-grade CM5 sensor chip using the standard amine coupling kit (GE Healthcare). The immobilization level was about 1300 resonance units. After buffer exchange to degassed 0.5% PBST (pH 7.4), affinity-purified soluble rhMSH2 (a six-histidine C-terminal fusion consisting of 323 amino acids and including the MutSI and MutSII domains) or BSA (as a control soluble protein) was injected at 30 l/min over the immobilized NKG2D-Fc and IgG1 Fc at 25°C. The binding response at each concentration of rhMSH2 (12.5-200 nM) or BSA (1 M) was calculated by subtracting the equilibrium response measured in the IgG1 Fc flow cell from the response in the NKG2D flow cell. BIAevaluation version 4.1 software (Biacore AB, Uppsala, Sweden) was used for data analysis. Curve fitting was performed with Sigmaplot 11.0 software (Systat Software, Inc.).
Selective Cell Surface Biotinylation and Immunoblotting-The Pierce cell surface protein isolation kit (Thermo Scientific) was used for selective cell surface biotinylation according to the manufacturer's protocol. Briefly, four T75-cm 2 flasks of 90 -95% confluent HK-2 (serving as negative control), HeLa, 803, and NCI-H520 cells were washed twice with cold PBS and gently agitated in 0.5 mg/ml EZLink sulfo-NHS-SS-biotin for 30 min at 4°C. 500 l of quenching solution was sequentially added to each flask to terminate the biotinylation reaction. Cells were harvested and lysed in 500 l of Lysis buffer in the presence of protease inhibitor mixture (Thermo Scientific). The cell lysates were clarified by centrifugation at 10,000 ϫ g at 4°C for 2 min. Labeled proteins were isolated by NeutrAvidin-agarose and eluted with 2ϫ SDS-PAGE sample buffer containing 50 mM DTT. The purified proteins were analyzed by Western blot with anti-hMSH2 mAb (65021-1 Ig) using tumor nuclear proteins as references. The nuclear proteins were prepared by NucBuster protein extraction kit (Novagen) following the manufacturer's protocol. 3 H-TdR Incorporation Assay-1 ϫ 10 5 resting ␥␦ T cells (V␦2 T cells Ͼ85%) were seeded in triplicate in 96-well plates and incubated for 60 h with rhMSH2 (20 g/well), irrelevant recombinant human His-tagged protein PKM2 (20 g/well, serving as negative control), or EP5 (100 g/well, serving as positive control) (24). Cells were cultured for an additional 12 h after adding [ 3 H]thymidine (0.5 Ci/well). The cellular fraction was suspended in scintillation fluid and counted in a beta counter to evaluate 3 H-TdR incorporation. The proliferation of V␦2 T cells was determined by counts/min.
Statistics-The statistical significance of differences between experimental groups was determined with Student's t test, with p values of less than 0.05 considered significant.

Ectopic Expression of hMSH2 on Epithelial Tumor Cell
Lines-Cell surface expression is a prerequisite for hMSH2 as a ligand for ␥␦ T cells. We examined the surface expression of hMSH2 on seven human epithelial tumor cell lines, including HeLa, SKOV3, HO8910, ES-2, 803, HR8348, and NCI-H520. HK-2, fibroblasts (derived from the normal tissues of an ovarian cancer patient), and ␥␦ T cells were used as normal cell controls. Flow cytometric analysis with specific anti-hMSH2 Ab (H-300) revealed that 10 -70% of the cell population among each of the tested tumor lines expressed hMSH2, whereas normal control cells rarely expressed hMSH2 on their surface (Ͻ2.2%) (Fig. 1), suggesting broadly ectopic surface expression of hMSH2 during carcinogenesis, which raises the possibility as a ligand for ␥␦ T cells.
Surface Expression of hMSH2 Enhanced ␥␦ T Cell-mediated Tumor Cytolysis in Vitro-To investigate the significance of ectopically expressed hMSH2 in ␥␦ T cell-mediated anti-tumor immunity, we chose HeLa, 803, and NCI-H520, three tumor cell lines expressing significant surface hMSH2, as targets for specific antibody blockade and target gene knockdown with siRNAs in cytolysis assays. We used ␥␦ T cells that were stimulated and expanded with immobilized anti-pan-TCR␥␦ mAb as effectors. The expanded ␥␦ T cells were mainly composed of V␦2 T subset (80 -95%) and expressed NKG2D at day 14 ( Fig.  2A). Anti-MSH2 Ab (N-20) significantly blocked ␥␦ T cell-mediated cytolysis of HeLa and 803 cells (Fig. 2B).
To further validate the hMSH2-mediated specific recognition by ␥␦ T cells, we used small RNA interference to knock down hMSH2 mRNA expression in HeLa and 803 cell lines (Fig.  3). Forty eight hours after specific siRNAs transfection, siRNA duplex I and II resulted in 80 and 66% reduction of hMSH2 mRNA expression in HeLa cells and 92 and 80% reduction in 803 cells, when compared with the mock control (Fig. 3A, top  panels). Reduced hMSH2 protein expression was confirmed by Western blot of total protein extracted from siRNA-treated cells or mock control cells (Fig. 3A, bottom panels). Decreased surface expression of hMSH2 on siRNA-treated HeLa and 803 cells was measured by flow cytometry and confocal microscopy ( Fig. 3B). The cytotoxicity of human ␥␦ T cells against siRNA-treated HeLa or 803 cells was compared with that of mock controls. At various E:T ratios, the decreased hMSH2 expression with siRNA I or II interference resulted in significantly reduced cytolysis of HeLa and 803 cells by ␥␦ T cells (Fig.  3C). Similar effects were observed in NCI-H520 cells (data not shown). The above results from N-20 Ab blocking or siRNA interference suggest that ectopically expressed hMSH2 mediates the recognition of carcinoma cells by human ␥␦ T cells and contributes to ␥␦ T cell-mediated tumor cytolysis.
Both TCR␥␦ and NKG2D Were Involved in Recognition of Surface hMSH2-␥␦ T cells express two main types of recognition receptors, TCR␥␦ and NKG2D. To determine the functional involvement of TCR␥␦ and NKG2D in recognizing hMSH2, the expanded ␥␦ T effector cells were pretreated with anti-TCR␥␦ (B1.1) and/or anti-NKG2D (149810) mAb before incubation with siRNA I-treated HeLa cells or mock control cells. The ectopically expressed hMSH2-mediated recognition by ␥␦ T cells was significantly inhibited by anti-TCR␥␦ (B1.1), anti-NKG2D (149810), or both mAbs (Fig. 4A, left panel). The combined use of anti-TCR␥␦/anti-NKG2D and siRNA I further decreased the cytotoxicity of ␥␦ T effectors when compared with use of anti-TCR␥␦/anti-NKG2D or siRNA I alone, suggesting that both TCR␥␦ and NKG2D may participate in hMSH2-mediated recognition and cytolysis of carcinoma cells by human ␥␦ T cells. Considering the significance of hMSH2 as a NKG2D ligand, we confirmed the recognition of hMSH2 by NKG2D using the NK-92 cell line as effectors. Combination of  anti-NKG2D (149810) and siRNA I treatment resulted in a notable decrease in the cytotoxicity of NK-92 cells compared with that using anti-NKG2D or siRNA I alone (Fig. 4A, right  panel).
We used affinity-purified rhMSH2, which was judged homogeneous by a single band on 10% SDS-polyacrylamide gel following Coomassie Blue staining (Fig. 4B), to confirm the specific binding of hMSH2 by TCR␥␦ or NKG2D receptors. Expanded ␥␦ T cells were incubated with rhMSH2, rPKM2 (negative control for both TCR␥␦ and NKG2D), anti-TCR␥␦ (positive control for TCR␥␦), or ULBP5 (positive control for NKG2D) before staining with FITC-conjugated anti-pan-TCR␥␦ or anti-NKG2D (1D11). As shown in Fig. 4C, decreased fluorescence intensity of TCR␥␦ and NKG2D was observed in the rhMSH2 preincubated group and the positive control groups, suggesting a specific binding of hMSH2 to both receptors. Furthermore, SPR analysis suggested a direct binding of rhMSH2 to NKG2D-Fc. The K d value derived from nonlinear curve fitting of the standard Langmuir binding isotherm was 132 nM for rhMSH2 (Fig. 4D). Preincubation with soluble rhMSH2 protein for 1 h at 4°C resulted in a significant decrease of killing activity of ␥␦ T cells and NK cells against HeLa cells when compared with rPKM2 blocking group and nontreated group (Fig. 4E). These results, together with our previous data (24,27), support that ␥␦ T cells recognize hMSH2 via both TCR␥␦ and NKG2D.
Surface hMSH2 Expressed on Epithelial Tumor Cells Was a Full-length Protein-To characterize the cell surface-expressed hMSH2, the membrane proteins of HeLa, 803, NCI-H520, and HK-2 cells were biotinylated and purified. Western blot analysis with anti-hMSH2 mAb (65021-1 Ig) demonstrated that the membrane-bound hMSH2 was a full-length protein with a mass of 105 kDa (Fig. 5A). No truncated form of hMSH2 protein was detected.
hMSH2 Preferentially Activates V␦2 Subset in Vitro-To determine the specific subset that recognizes the surface-expressed hMSH2 on tumor cells, the purified V␦1 and V␦2 subsets (purity 84 -95%) were used as effector cells in cytotoxicity assays (Fig. 5B). The V␦1 subset exhibited lower cytotoxicity against HeLa cells compared with the V␦2 subset. The cytotoxicity was not inhibited by siRNA I treatment (Fig. 5C, left panel). In contrast, V␦2 T cell-mediated cytotoxicity was significantly reduced in siRNA I-treated groups (Fig. 5C, right panel). These results indicate that hMSH2 is preferentially recognized by V␦2 T cells. This is consistent with the fact that hMSH2 was identified using a V␦2 CDR3 peptide (OT3) as the probe (24).
rhMSH2 was used to stimulate proliferation and IFN-␥ production of V␦2 T cells in vitro. Incubation of expanded V␦2 T cells (purity Ͼ85%) with immobilized rhMSH2 resulted in  marked proliferation (p Ͻ 0.05) (Fig. 5D, left panel) and a significant increase in IFN-␥ concentration (p Ͻ 0.05) (Fig. 5D, right panel) compared with irrelevant recombinant control protein PKM2, demonstrating the ability of hMSH2 to activate human V␦2 T cells in vitro.
Inducible Expression of hMSH2 on EBV-transformed Human PBMCs-␥␦ T cells have been implicated in immune responses against EBV infection (28). Some ␥␦ T cells ligands can be induced on EBV-infected B cells and facilitate ␥␦ T cell-mediated killing (7,16). To determine the role of hMSH2 in ␥␦ T cell-mediated anti-viral immunity, we infected PBMCs isolated from healthy donor with EBV and analyzed hMSH2 expression. We found that surface hMSH2 appeared on EBV-transformed LCLs but not on untreated PBMCs (Fig. 6A). hMSH2 mRNA expression was also elevated in LCLs compared with normal B cells, indicating EBV infection up-regulates hMSH2 expression (Fig. 6B). The sensitivity of freshly generated EBV-transformed B cells to autologous or allogenic ␥␦ T cell-mediated cytolysis was 2-3-fold greater than that of normal B cells (Fig. 6C). Treatment with N-20 Ab significantly reduced the specific cytolysis of these EBV-infected cells (Fig. 6C).
To further confirm that the induced hMSH2 facilitates the killing of EBV-transformed cells by ␥␦ T cells, specific siRNAs interference was done in LCL 3D5 (Fig. 6D). Consistent with that observed in carcinoma cells, specific cytolysis of siRNAs-treated 3D5 cells by ␥␦ T cells was significantly reduced, demonstrating that the surface expression of hMSH2 mediated the recognition of LCLs by ␥␦ T cells. Taken together, these results suggest that the cell surface-expressed hMSH2 is inducible and may mediate the specific clearance of EBV-infected cells by human ␥␦ T cells.

DISCUSSION
Most ␥␦ T cells have innate cell-like features and can be activated upon recognition of a diverse set of conserved yet poorly defined endogenous tumor-associated antigens/ligands through their TCR and/or NKG2D receptors (1). Understanding these natural ligands will be helpful not only for the elucidation of antigen recognition mechanism by which ␥␦ T cells discriminate healthy from malignant cells but also for the development of efficient tumor immunotherapeutic strategies (29). In this study, we provide evidence for the surface-expressed hMSH2 on carcinoma cells to serve as a tumor-derived recognition antigen for human peripheral ␥␦ T cells via both TCR␥␦ and NKG2D receptors. In addition, we show for the first time that hMSH2 is up-regulated on EBV-transformed B cells and involved in ␥␦ T cell-mediated anti-viral cytotoxicity.
Based on our previous study, we hypothesized that hMSH2 may be a tumor-associated ligand for TCR␥␦. To test this possibility, we first examined the cell surface expression of hMSH2 on cervical, gastric, ovarian, colorectal, and lung carcinoma cell lines. We observed a ubiquitous ectopic expression of hMSH2  at the surface of the tested tumor cell lines. Aberrant cell surface expression of hMSH2 provided a ligand for ␥␦ T cell receptors. The ectopically expressed hMSH2 was correlated with an enhanced ␥␦ T cell-cytolytic activity against carcinoma cells, suggesting an hMSH2-mediated recognition of carcinoma cells by ␥␦ T cells. Such ectopic expression of hMSH2 on the cell surface may reflect an alert mechanism to the innate immune system when they suffer stress or are in a process of carcinogenesis.
Despite the central role of TCR-dependent recognition in ␥␦ T cell activation, NKG2D takes a critical part in ␥␦ T cytolytic response to tumor cells by binding to stress-inducible ligands MHC class I-related chains and ULBPs (30,31). NKG2D has been described to express on essentially all V␦2 T cells and on the immobilized mAb expanded ␥␦ T cells. TCR-and NKG2Dmediated recognition of hMSH2 was supported by the results that hMSH2 silenced target cells became less sensitive to the cytotoxicity of ␥␦ T cells. Specific binding analyses with flow cytometry and SPR provided further evidence for the dual recognition. Taking into account the constitutive expression of NKG2D on V␦2 T cells, NK cells, and CD8 ϩ T cells (32), the NKG2D-hMSH2 interaction could be important in hMSH2associated tumor immune surveillance.
V␦2 T cells are the main subset of human peripheral ␥␦ T cells that recognize phospho-Ags and produce Th1 cytokines such as IFN-␥. They preferentially kill hematopoietic tumors (33,34). More recent studies reveal their ability to kill carcinoma cells pretreated by phospho-Ags (35), suggesting that tumor recognition of these cells strongly depends on isopentenyl pyrophosphate production. However, it is likely that some other molecules are involved in the recognition process but are as yet undefined. hMSH2 as appears to be such a molecule. Indeed, the V␦2 TCR-mediated recognition of hMSH2 was suggested by the specific binding of OT3-grafted V␦2 TCR to rhMSH2 proteins in our previous study (24,27). The functional significance of hMSH2 recognition by ␥␦ T cells is supported by our observation that the surface expression of hMSH2 on HeLa cells is correlated with the cytotoxicity of V␦2 T cells. hMSH2 readily activates V␦2 T cells to proliferate and produce IFN-␥. Thus, hMSH2-mediated recognition of carcinoma cells by V␦2 T cells opens new perspectives for V␦2 T cell-based tumor immunotherapy.
The human herpesvirus EBV is thought to act as a triggering agent for autoimmune diseases and cancer in genetically susceptible individuals (36). Proliferation of ␥␦ T cells and overexpression of HSPs and ULBPs in acute or chronic EBV infections have been reported previously (7,16,28,37). The mRNA transcription and cell-surface expression of hMSH2 were up-regulated by EBV infection, leading to a substantial enhancement of ␥␦ T cell cytotoxicity toward EBV-transformed cells. These results suggest that hMSH2 may be stress-inducible, a common feature of most ␥␦ T cell ligands, and potentially important for immunity in EBV infections. Interestingly, although previous reports studied V␦1 T cell response to EBV infection, our study suggests a role of V␦2 T cells in anti-EBV immunity.
Thus, we envision that when the host cells are in a cancerous or distressed condition, ectopic expression of hMSH2 on the cell surface will occur to serve as a danger signal to alert the innate immune system. The signal is sensed by TCR and NKG2D expressed on V␦2 T cells. TCR␦2 and NKG2D will work synergistically to initiate effective immune responses against tumor or virus-infected cells. However, it remains enigmatic how hMSH2 transports to the cell surface. As a nuclear protein, hMSH2 contains neither signal peptide sequence for membrane transport nor classic C-terminal KDEL sequence for endoplasmic reticulum retention. The re-location of this molecule to the membrane may be due to the change of membrane composition or the common alteration of other mismatch repair proteins (such as hMSH6) in malignant cells (38,39). Besides, it is not known whether cell surface-expressed hMSH2 molecules themselves act as immunogenic determinants for human ␥␦ T cells and NK cells, or whether they work as an Ag presentation molecule, just like CD1c and F 1 -ATPase do (40,41). A better understanding of these underlying mechanisms will help us to determine the precise role of hMSH2 in ␥␦ T cell-mediated immune surveillance and to optimize hMSH2based therapeutic strategies for malignancy.