Among factors involved in cancer cells escaping from immune responses, an intrinsic defect in the IFNγ response is considered to be one of the major players allowing cancer cells to evade the host immunity. In this study, we investigated how tumor cells escape from the IFNγ-dependent immune response through the immunoediting process by analyzing originally established immune-escape variants of melanoma cells. We found that the immune-escape melanoma variants gained resistance to the IFNγ-induced oxidative stress response and identified glutathione-S-transferase-4 (GSTA4) as a critical molecule in this process. Furthermore, the immune escape melanoma variants acquired a greater metastatic ability by a GSTA4-dependent mechanism.

Implications:

Considering the importance of GSTA4 in controlling IFNγ responsiveness and the metastatic potential of other melanoma cells, our results highlight a novel mechanism whereby cancer cells escape from host immunity and gain metastatic ability by acquiring resistance to oxidative stress responses through the upregulation of GSTA4.

This article is featured in Highlights of This Issue, p. 1

Cancer immunosurveillance refers to the concept in which the immune system recognizes and eliminates cancerous cells (1); however, the immune system also selects (immunoedits) tumor cells displaying immunoevasive properties (2). Both cytolytic activity and IFNγ production by cytotoxic T lymphocytes (CTL) recognizing tumors are critical for cancer immunoediting through their antitumor function. Particularly, IFNs have been identified as critical molecules in cancer immunoediting processes (3). Also, it is known that IFNγ is involved in the protumorigenic process (4), including the promotion of genetic instability during the immunoediting process (5).

Cancer immunotherapy, such as immune checkpoint blockade (ICB), has been successful for many types of cancer including melanoma (6); however, unresponsiveness to ICB remains as a major obstacle to realizing further clinical benefit (7). In this context, the importance of the cancer-immunity cycle has been highlighted for successful immunotherapy, but in turn its disruption can compromise its efficacy (8). Among factors involved in cancer cells escaping from immune responses (9), an intrinsic defect in the IFNγ response is considered as one of the major players allowing cancer cells to evade host immunity (10, 11). Regarding the exact role of IFNγ in eliminating cancer cells, it has been reported that IFNγ induces cellular senescence in cancer cells by enhancing the reactive oxygen species (ROS) and DNA damage response (12). Furthermore, a recent study demonstrated that IFNγ released from CD8+ T cells promotes tumor cell lipid peroxidation and induces ferroptosis, which is an oxidative stress-related programmed cell death (13). These evidence suggest the relevance of the oxidative stress response in the IFNγ-dependent immune editing process of cancer cells.

In this study, we investigated how tumor cells escape from the IFNγ-dependent immune response through the immunoediting process by analyzing originally established immune-escape variants of melanoma cells. We found that the immune-escape melanoma variants gained resistance to the IFNγ-induced oxidative stress response, and glutathione-S-transferase-4 (GSTA4) was a critical molecule in this process. Furthermore, the immune-escape melanoma variants acquired higher metastatic ability by in vivo by a GSTA4-dependent mechanism. Considering the importance of GSTA4 in controlling the IFNγ responsiveness and metastatic potential of other melanoma cells, our results highlight a novel mechanism whereby cancer cells to escape from host immunity and gain metastatic ability by acquiring resistance to oxidative stress responses through the upregulation of GSTA4.

Reagents

4-Hydroxynonenal (4-HNE) was purchased from EMD Milipore (#393204). IFNγ was purchased from Pepro Tech. GSTA4 and β-actin primers were purchased from Invitrogen. Antibody against GSTA4 was obtained from EMD Milipore (#ABS1652), antibody against Flag-tag was obtained from Sigma-Aldrich (#F1804) and antibody against β-actin was obtained from Santa Cruz Biotechnology (#sc-47778). WST-8 was purchased from Nacalai Tesque (#07553–44). GSH/GSSG Quantification Kit was purchased from Dojindo (#G257). Cell ROX Deep Red Flow Cytometry Assay Kit was purchased from Invitrogen (#C10491). Pgl4.50 [luc2P/CMV-RE-Hygro] vector and D-luciferin were obtained from Promega. Lipofectamine 2000 was purchased from Invitrogen. Hygromycin B was obtained from Nacalai Tesque.

Cells

The murine B16 melanoma cell line expressing ovalbumin (MO4: B16OVA) was kindly provided by Dr. Shinichiro Fujii (RIKEN). To establish B16OVA cells stably expressing luciferase (B16OVAluc2), B16OVA cells were transfected with pGL4.50 vector and cloned by limiting dilution as described previously (14). B16 and B16F10 cells were cultured in Eagle's Minimum Essential Medium (EMEM). Malme3M, UACC62 (obtained from NCI), and MeWo (obtained from ATCC) were cultured in RPMI1640. All cell lines have not been authenticated and tested for mycoplasma since their original acquisition, and did not culture more than 1 month. All of the media were supplemented with 2 mmol/L L-glutamine, 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. The cells were maintained at 37°C in a humidified atmosphere with 5% CO2.

Mice

Six- to eight-week-old wild-type C57BL/6J (WT) mice were purchased from Japan SLC, Inc. IFNγ−/− (IFNγ KO) mice were kindly provided by Dr. Y. Iwakura (Tokyo University of Science) and maintained at Laboratory Animal Research Center, Institute of Medical Science, University of Tokyo. All experiments were approved (A2016INM-9, A2019INM-4) and performed according to the guidelines of Care and Use of Laboratory Animals of University of Toyama and the Animal Care and Use Committee of Institute of Medical Science of the University of Tokyo.

Establishment of immune-escape variants of B16OVA cells

To prepare cells with different in vivo immunological experience, B16OVAluc2 cells (105) were used to subcutaneously inoculate to WT untreated, WT vaccinated, or IFNγ KO vaccinated mice. Twelve (untreated WT or IFNγ KO vaccinated mice) or 19 (vaccinated WT mice) days after inoculation, approximately same-sized tumors (∼400 mm3) were aseptically harvested, dissected, and digested with 2 mg/mL collagenase (Roche Diagnostics GmbH) and 0.1 mg/mL DNase I (Roche Diagnostics GmbH) in serum‐free RPMI1640 for 1 hour at 37°C. Isolated tumor cells were named “NIMM” if prepared from tumors from WT untreated mouse, “IMM” if prepared from WT vaccinated mouse tumors, and “GKO-IMM” if prepared from IFNγ KO vaccinated mouse tumors. All newly established cell lines were cultured in vitro for at least 3 weeks before in vivo characterization.

Plasmid preparation and lentivirus production

The flag-tagged mouse Gsta4 sequence, which was synthesized by Thermo Fisher Scientific, was subcloned into the pENTR1A vector, and inserted into the pLenti CMV Hygro DEST vector (w117–1), which was a gift from Dr. Campeau E. and Dr. Kaufman P. (Addgene plasmid #17454; ref. 15). The pLenti CMV Hygro E2-Crimson vector was established previously (16). shRNAs targeting mouse GSTA4 (TRCN0000103431 and TRCN0000103434) or luciferase (SHC007) were from Dharmacon. Lentivirus particles were produced as described previously (17). Briefly, lentivirus was packaged in Lenti-X 293T cells per standard protocols. The amount of virus was titrated for near quantitative infection with <5% toxicity of nontemplate virus. For lentiviral delivery, 1 × 105 cells were plated on Day 1 in 6-well plates, infected on the following day, and cells were selected in 400 μg/mL hygromycin or 1 μg/mL puromycin.

DNA microarray analysis

Total RNA was extracted from cells using RNeasy Mini Kit (Qiagen). Gene expression was analyzed using a GeneChip system with GeneChip Mouse Gene 2.0 ST Array (Affymetrix), as described previously (18). In this study, a total of four arrays were used: one for parent cells and three for IMM cells (IMM 1, 2, and 8). Comparison of gene expression levels was conducted using Transcriptome Analysis Console software (Thermo Fisher Scientific). The microarray dataset was deposited to Gene Expression Omnibus (GEO) with accession number GSE199573. The top 10 genes and bottom 10 genes were picked up for Fig. 2A based on their values of fold change.

Cell viability assay

Cells were seeded at a final concentration of 5 × 103 cells/well (24 hours) or 103 cells/well (72 hours) in a 96-well plate. After 4-hour incubation, cells were treated with 4-HNE for 24 hours or IFNγ for 72 hours (37°C, 5% CO2). After treatment, 10 μL of WST-8 reagent was added and incubated for another 2 hours. The absorbance was measured in a microplate reader at 450/620 nm. Cell viability was determined from the absorbance of soluble formazan dye generated by the living cells.

qPCR

Cells were seeded at a final concentration of 105 cells/well in 6-well plates for 48 hours and incubated in a humidified atmosphere (37°C, 5% CO2). Total RNAs were prepared using the RNeasy Plus Mini Kit (Qiagen). Expression of GSTA4 was quantitatively determined by real-time PCR using an ABI Prism 7300 sequence detection system (Life Technologies Corporation). The expression level of GSTA4 mRNA was normalized to the β-actin gene. The primers used were: 5′-TGA TTG CCG TGG CTC CAT TTA-3′ (forward) and 5′-CAA CGA GAA AAG CCT CTC CGT-3′ (reverse) for GSTA4 mRNA and 5′-GGC TGT ATT CCC CTC CAT CG-3′ (forward) and 5′-CCA GTT GGT AAC AAT GCC ATG T-3′ (reverse) for β-actin mRNA. The expression level of antigen (OVA, gp100, TRP1, TRP2, MART1) mRNA was normalized to the GAPDH gene. he primers used were: 5′-CCT TGA GCA GCT TGA GAG TATA A-3′ (forward) and 5′-CCA TCT TCA TGC GAG GTA AGT-3′ (reverse) for OVA mRNA, 5′-AGC ACC TGG AAC CAC ATC TA-3′ (forward) and 5′-GTT CCA GAG GGC TGT GTA GT-3′ (reverse) for gp100 mRNA, 5′-TGG GGA TGT GGA TTT CTC TC-3′ (forward) and 5′-AGG GAG AAA GAA GGC TCC TG-3′ (reverse) for TRP1 mRNA, 5′-AGG TAC CAT CTG TTG TGG CTG GAA-3′ (forward) and 5′-AGT TCC GAC TAA TCA GCG TTG GGT-3′ (reverse) for TRP2 mRNA, 5′-ATT GCT CTG CTT ATC GGC TGC T-3′ (forward) and 5′-CAC CAT TCC TCC AAT ATC CCT CT-3′ (reverse) for MART1 mRNA, 5′-AAA TGG TGA AGG TCG GTG TG-3′ (forward) and 5′-TGA AGG GGT CGT TAG ATG C-3′ (reverse) for GAPDH mRNA.

Western blotting

Cells (105 cells/well in 6-well plates) were grown for 48 hours and collected, washed with PBS, and treated with lysis buffer (1 M DTT, 1 M sodium orthovanadate, 1 M β-glycerophosphate, 10 mg/mL aprotinin, 10 mg/mL leupeptin, 0.1 M PMSF). The cell lysates were separated (10% SDS-PAGE), and then transferred to PVDF membranes. After blocking (0.1% Tween 20 in PBS-5% BSA for 1.5 hours at room temperature), the membranes were incubated with primary antibodies (used at a dilution of 1:1,000) for overnight, and then with the secondary antibodies (used at a dilution of 1:2,000) for 1 hour.

Transwell invasion assay

The transwell invasion assay was performed accordingly (19). Transwell chambers (Costar) were attached with 8-μmol/L pore membrane filters (Whatman). The lower part of membrane filter was precoated with 25 μg/mL fibronectin (Fujifilm Wako) and the upper part of membrane filter was precoated with 10 μg/mL Matrigel (BD Biosciences). Cells (3 × 104 cells in serum-free media with 0.1% BSA) were added to the upper part of the chamber, and incubated for 6 hours. The invaded cells were stained (hematoxylin and eosin) and photographed (at 400× magnification, Biozero BZ-8000 microscope). Invading cells on the membrane were selected (five random visual fields) and the cells that invaded the filter were counted using ImageJ software.

ROS measurement

Cells were seeded at a final concentration of 5 × 104 cells/well in 6-well plate. After 4-hour incubation, cells were treated with IFNγ 20 U/mL for 72 hours and incubated in a humidified atmosphere (37°C, 5% CO2). In brief, the cells were harvested and stained with CellROX Deep Red reagent or medium for the unstained cells. After that, cells were incubated for another 1 hour in a humidified atmosphere (37°C, 5% CO2). Flow cytometric analysis was performed using FACSCanto II (BD Biosciences). FlowJo ver. 10 software (Tree Star) was used for quantification of intracellular oxidative stress.

In vivo tumor model

For the subcutaneous tumor model, cells were injected subcutaneously (105) and tumor growth was assessed by measuring the tumor diameter every 2 days. In some experiments, the group of mice received subcutaneous injections of 100 μg of chicken ovalbumin protein (OVA; Sigma) emulsified in complete (CFA; Sigma) or incomplete (IFA; Sigma) Freund's adjuvant 14 (OVA-CFA) or 7 (OVA-IFA) days prior to tumor inoculation. In some experiments, anti-PD-1 antibody (RMP1–14, 250 μg/mouse, BioXCell) was administered intraperitoneally on Days 3, 6, and 9 after tumor inoculation. The tumor volume was calculated by the formula (major axis) × (minor axis)2 × 0.52. Mice with no confirmed tumor were excluded from the data.

The experimental lung metastasis model was created as previously described (20) with some modifications. In brief, cells were harvested with 0.05% trypsin-EDTA (Gibco), washed once with PBS, and resuspended with appropriate concentrations in PBS. Cells were injected into the tail vein (i.v., 0.5–2.5 × 105). Mice were sacrificed 14 days after the injection, and then the lungs were removed and fixed in Bouin's solution for further analysis. The lung metastasis was evaluated by counting the number of metastasized tumor colonies on the surface of the lungs with the aid of a dissection microscope.

Statistical analysis

All the data are expressed as the mean ± SEM and are representative of at least two independent experiments, unless otherwise stated. The groups of 5 to 10 mice were used to perform all in vivo experiments. Significance was analyzed using Student t test for comparisons between two groups. One-way ANOVA with Bonferroni correction was used to compare 3 or more groups. P < 0.05 was considered significant. P < 0.05 was considered significant.

Data availability

The microarray dataset was deposited to GEO with accession number GSE199573. The GSTA4 expression in human melanoma in this study were obtained from Cancer Cell Line Encyclopedia. The human melanoma free survival rates data are publicly available in PRECOG database (GSE8401). The human melanoma respond to anti-PD-1 treatment in correlation with GSTA4 expression are publicly available in ROC plotter database. The human melanoma survival rates data in respond to anti-PD-1 treatment are publicly available in KM plotter database.

Establishment of immune-escape variants of B16OVA cells

To establish cancer cells escaping from host immunity, we created a murine B16 melanoma cell line expressing ovalbumin (B16OVA) and used it to inoculate to mice immunized with OVA (Supplementary Fig. S1A). We previously demonstrated that the growth of B16OVA cells initially progressed, but was then suppressed by a CD8+ T-cell-dependent immune response, and thereafter showed secondary progression by escaping from host immunity in OVA-immunized mice (14). Using this model, we established cell lines after in vivo passage through distinct immunologic conditions (Supplementary Fig. S1B). B16OVA tumors exposed to OVA-specific CD8+ T-cell immunity in OVA-immunized B6 mice were isolated, and we established five variants (IMM1, 2, 5, 6, and 8 cell lines). For comparison, B16OVA tumors from nonimmunized naïve B6 mice or OVA-immunized IFNγ-deficient mice were also isolated, and we established four variants, namely NIMM (1, 3, 4, 5) cell lines or GKO-IMM (1, 2, 3, 4) cell lines, respectively. To test the capacity of those different variants to provoke tumor antigen-specific immunity, IMM, NIMM, or GKO-IMM cell lines were rechallenged in OVA-immunized B6 mice. Contrary to NIMM and GKO-IMM cell lines, IMM cell lines showed progressive growth in OVA-immunized mice upon rechallenge. In addition, IMM cell lines, but not NIMM and GKO-IMM cell lines, specifically lost their OVA antigen expression without affecting other melanoma antigens (Supplementary Fig. S1C). These results suggest that IMM cell lines acquire the ability to escape from the OVA-specific antitumor immune response.

Resistance of immune-escape melanoma variants to IFNγ-induced cytostatic effect and ROS production

IFNγ is a critical effector molecule of tumor antigen-specific CD8+ T cells; thus, cancer cells often escape from antitumor immunity by losing their responsiveness to IFNγ. Therefore, we next examined the response of IMM variants to IFNγ in vitro. IFNγ treatment showed a dose-dependent cytostatic effect in parental B16OVA, NIMM3, or GKO-IMM1 cells; however, IMM2 cells showed significant resistance to in vitro IFNγ treatment in concert with the ability to evade the OVA-specific antitumor immune response in vivo (Fig. 1A). Such resistance to IFNγ treatment in IMM cell lines, but not in NIMM or GKO-IMM cell lines, was confirmed in all variants established (Fig. 1B). Importantly, IMM cell lines responded to in vitro IFNγ treatment by upregulating their expression of MHC class I (H-2Kd) or PD-L1 (Supplementary Fig. S2); therefore, such resistance to the IFNγ-induced cytostatic effect is not due to the lack of their responsiveness to IFNγ. Collectively, these results suggest that the ability of IMM variants to escape from antitumor immunity could be a result of acquiring resistance to the IFNγ-induced cytostatic effect, and not a defect of IFNγ-dependent signaling.

Figure 1.

Resistance of immune-escape melanoma variants to IFNγ-induced cytostatic effect and ROS production. The indicated cell lines were treated with different concentrations of IFNγ (A) or 20 U/mL of IFNγ (B and C) for 72 hours. Cell viability (A and B, % of untreated control) or the intracellular ROS level (C, fold change, untreated control = 1) was evaluated. Data are shown as mean ± SEM (**, P < 0.01; *, P < 0.05; ns, not significant).

Figure 1.

Resistance of immune-escape melanoma variants to IFNγ-induced cytostatic effect and ROS production. The indicated cell lines were treated with different concentrations of IFNγ (A) or 20 U/mL of IFNγ (B and C) for 72 hours. Cell viability (A and B, % of untreated control) or the intracellular ROS level (C, fold change, untreated control = 1) was evaluated. Data are shown as mean ± SEM (**, P < 0.01; *, P < 0.05; ns, not significant).

Close modal

Because it has been reported that IFNγ induces reactive oxygen species (ROS) and subsequent oxidative stress responses to exert its cytostatic effect (12), we next examined the oxidative stress response in IMM cell lines upon IFNγ treatment. IFNγ treatment induced ROS production in parental B16OVA cells, whereas such IFNγ-induced ROS production was weak in IMM cell lines (Fig. 1C). Therefore, we conclude that immune-resistant melanoma cell variants did not respond to the IFNγ-induced cytostatic effect because of acquiring resistance to the IFNγ-induced oxidative stress response.

Functional upregulation of GSTA4 in immune-resistant melanoma variants

To understand the molecular mechanism leading to IMM cell lines acquiring resistance to the IFNγ-induced oxidative stress response, we comprehensively analyzed gene expression of those cell lines using DNA microarray. The microarray dataset was deposited to GEO with accession number GSE199573. In Fig. 2A, 10 of the most up- and downregulated genes in IMM cell lines compared with parental B16OVA cells are shown. Among those, the most upregulated gene in IMM cell lines was a glutathione S-transferase alpha 4 (GSTA4, Fig. 2A). Such higher expression of GSTA4 in IMM cell lines was further confirmed by mRNA (Fig. 2B) and protein (Fig. 2C) expressions. Although we observed a minor increase of GSTA4 expression in other variants (NIMM and GKO-IMM cell lines, Fig. 2B), IMM cell lines, but not parental B16OVA cells or other variants, showed a relatively lower response to 4-HNE, which is an oxidative lipid mainly catalyzed and detoxified by GSTA4 (Fig. 2D and E). These results clearly indicate that immune-resistant melanoma variants specifically upregulated the expression of functional GSTA4.

Figure 2.

Functional upregulation of GSTA4 in immune-resistant melanoma variants. A, The top 10 genes up- and downregulated in immune-resistant melanoma variants (IMM) compared with parental B16OVA are listed. B, Relative mRNA expression (fold change, B16OVA = 1) and (C) protein expression of GSTA4 in the indicated cell lines are shown. The indicated cell lines treated with different concentrations of 4-HNE (D) or 40 μmol/L of 4-HNE (E) for 24 hours and cell viability were evaluated as relative to the untreated control. Data are shown as mean ± SEM (**, P < 0.01; *, P < 0.05).

Figure 2.

Functional upregulation of GSTA4 in immune-resistant melanoma variants. A, The top 10 genes up- and downregulated in immune-resistant melanoma variants (IMM) compared with parental B16OVA are listed. B, Relative mRNA expression (fold change, B16OVA = 1) and (C) protein expression of GSTA4 in the indicated cell lines are shown. The indicated cell lines treated with different concentrations of 4-HNE (D) or 40 μmol/L of 4-HNE (E) for 24 hours and cell viability were evaluated as relative to the untreated control. Data are shown as mean ± SEM (**, P < 0.01; *, P < 0.05).

Close modal

Critical involvement of GSTA4 to regulate IFNγ-induced oxidative stress response in melanoma cells

To investigate the functional role of GSTA4 in protecting immune-resistant cancer cells from the IFNγ-induced oxidative stress response, we established either B16OVA overexpressing GSTA4 (Supplementary Fig. S3A) or IMM2 with knockdown of GSTA4 (Supplementary Fig. S3B). First, B16OVA cells overexpressing GSTA4 (GSTA4 OE#1 and GSTA4 OE#2) became resistant to the IFNγ-induced cytostatic effect compared with the mock control cell B16OVA line (Fig. 3A). Such resistance to IFNγ partly corresponded to the reduction of ROS production (Fig. 3B). Secondary, IMM2 cells with knockdown of GSTA4 (IMM2-shGSTA4#1 and IMM2-shGSTA4#2) showed restored responsiveness to the IFNγ-induced cytostatic effect similar to their parental B16OVA cells (Fig. 4A). In response to IFNγ, IMM2-shGSTA4#2 cells showed increased intracellular ROS levels (Fig. 4B), similar to the parental B16OVA cells. Collectively, these results clearly indicate the critical involvement of GSTA4 to regulate the IFNγ-induced oxidative stress response in melanoma cells.

Figure 3.

Cytoprotective effect of GSTA4 overexpression in IFNγ-induced oxidative stress response. Control (Mock) or GSTA4 overexpressing (GSTA4 OE #1 and GSTA4 OE #2) B16OVA cell lines were treated with the indicated concentrations of IFNγ for 72 hours, and then the cell viability (A) and intracellular ROS level (B) were determined. Data are shown as mean ± SEM (**, P < 0.01).

Figure 3.

Cytoprotective effect of GSTA4 overexpression in IFNγ-induced oxidative stress response. Control (Mock) or GSTA4 overexpressing (GSTA4 OE #1 and GSTA4 OE #2) B16OVA cell lines were treated with the indicated concentrations of IFNγ for 72 hours, and then the cell viability (A) and intracellular ROS level (B) were determined. Data are shown as mean ± SEM (**, P < 0.01).

Close modal
Figure 4.

Critical involvement of GSTA4 in IFNγ-induced oxidative stress response of immune-resistant melanoma variant. IMM2 cells transduced with control shRNA (shCTRL) or GSTA4 shRNA (shGSTA4 #1 and shGSTA4 #2) were treated with the indicated concentrations of IFNγ for 72 hours, and then the cell viability (A) and intracellular ROS level (B) were determined. Data are shown as mean ± SEM (**, P < 0.01).

Figure 4.

Critical involvement of GSTA4 in IFNγ-induced oxidative stress response of immune-resistant melanoma variant. IMM2 cells transduced with control shRNA (shCTRL) or GSTA4 shRNA (shGSTA4 #1 and shGSTA4 #2) were treated with the indicated concentrations of IFNγ for 72 hours, and then the cell viability (A) and intracellular ROS level (B) were determined. Data are shown as mean ± SEM (**, P < 0.01).

Close modal

GSTA4 governs immune-resistance and metastatic ability of melanoma cells

To verify the importance of GSTA4 in IMM variants to escape from the tumor-specific immune response by acquiring resistance to the IFN-γ-induced oxidative stress response, we conducted two different in vivo experiments using either GSTA4 over-expressing cell line or GSTA4 knockdown cell line. Firstly, we inoculated parental B16OVA or B16OVA cells over-expressing GSTA4 (GSTA4 OE#2) to OVA-immunized B6 mice and monitored tumor growth. As shown in Fig. 5A, the growth of GSTA4 OE#2 cells was more aggressive than that of parental B16OVA cells. Secondary, mice bearing IMM2 cells or IMM2-shGSTA4#2 cells were treated with anti-PD-1 to initiate tumor-specific immunity in vivo. Contrary to IMM2 cells which did not show any response (Fig. 5B), GSTA4 knockdown in IMM2 cells clearly reinvigorated the responsiveness to anti-PD-1 treatment in vivo (Fig. 5C). Considering there were no difference in T cell infiltration into both IMM2-shCTRL and IMM2-shGSTA4 #2 tumors (Supplementary Fig. S4), these results strongly indicate that GSTA4 is a gene responsible for melanoma cells gaining resistance to anti-tumor immunity.

Figure 5.

GSTA4 determines melanoma cell responsiveness to antitumor immunity in vivo. A, OVA-immunized mice were subcutaneously injected (105 cells/mouse) with B16OVA or GSTA4 OE #2 cells. B and C, WT B6 mice were subcutaneously injected (105 cells/mouse) with IMM2 shCTRL cells (B) or IMM2 shGSTA4 #2 cells (C), and treated with saline (Control) or anti-PD-1 mAb (Anti-PD-1, 250 μg/mouse) on days 3, 6, and 9 after the injection. Tumor volumes were measured by calipers and the data are shown as mean ± SEM (*, P < 0.05; **, P < 0.01).

Figure 5.

GSTA4 determines melanoma cell responsiveness to antitumor immunity in vivo. A, OVA-immunized mice were subcutaneously injected (105 cells/mouse) with B16OVA or GSTA4 OE #2 cells. B and C, WT B6 mice were subcutaneously injected (105 cells/mouse) with IMM2 shCTRL cells (B) or IMM2 shGSTA4 #2 cells (C), and treated with saline (Control) or anti-PD-1 mAb (Anti-PD-1, 250 μg/mouse) on days 3, 6, and 9 after the injection. Tumor volumes were measured by calipers and the data are shown as mean ± SEM (*, P < 0.05; **, P < 0.01).

Close modal

In addition to the resistance to anti-tumor immunity, we also found that IMM cell lines gained metastatic ability in vivo compared with parental B16OVA cells in an experimental metastasis model (Fig. 6A). Importantly, the knockdown of GSTA4 in IMM2 cells reduced the in vivo metastatic ability (Fig. 6B). Furthermore, IMM2 cells showed higher invasiveness in vitro compared with parental B16OVA cells, and the knockdown of GSTA4 in IMM2 cells reduced it (Fig. 6C). These results strongly suggest that GSTA4 governs the metastatic ability of IMM2 cells in addition to resistance to IFN-γ. Along with IMM cell lines, B16F10 cells, well-known metastatic variants of B16 melanoma established by an in vivo selection method (21), also highly expressed GSTA4 (Supplementary Fig. S5A). To test the involvement of GSTA4 in regulating the IFN-γ-responsiveness and metastatic ability of B16F10 cells, we established GSTA4 knockdown B16F10 cells using shRNA similar to IMM2 cells (Supplementary Fig. S5B). As shown in Fig. 6D, GSTA4 knockdown also gained the responsiveness of B16F10 cells to IFNγ treatment in vitro, and further impaired their in vivo metastatic ability (Fig. 6E). Collectively, these data suggest the importance of GSTA4 for melanoma cells gaining metastatic ability as well as resistance to antitumor immunity.

Figure 6.

GSTA4 regulates metastatic ability of melanoma cells. A and B, The indicated melanoma cells were intravenously injected into B6 mice (0.5–2.5 × 105 cells/mouse) and lung metastases were quantified by counting the colonies on day 14. C, The indicated cells (3 × 104 cells/chamber) were seeded in transwell chambers and incubated for 6 hours. After incubation, invaded cells in five random visual fields were photographed with a microscope (400×) and counted using ImageJ software. D, Cell viability of B16F10, B16F10 transduced with control shRNA (shCTRL), or B16F10 transduced with GSTA4 shRNA (shGSTA4 #2) and treated with the indicated concentrations IFNγ for 72 hours. E, The indicated melanoma cells were intravenously injected into B6 mice (2.5 × 105 cells/mouse) and lung metastases were quantified by counting the colonies on day 14. Data are shown as mean ± SEM (*, P < 0.05; **, P < 0.01; ns, not significant).

Figure 6.

GSTA4 regulates metastatic ability of melanoma cells. A and B, The indicated melanoma cells were intravenously injected into B6 mice (0.5–2.5 × 105 cells/mouse) and lung metastases were quantified by counting the colonies on day 14. C, The indicated cells (3 × 104 cells/chamber) were seeded in transwell chambers and incubated for 6 hours. After incubation, invaded cells in five random visual fields were photographed with a microscope (400×) and counted using ImageJ software. D, Cell viability of B16F10, B16F10 transduced with control shRNA (shCTRL), or B16F10 transduced with GSTA4 shRNA (shGSTA4 #2) and treated with the indicated concentrations IFNγ for 72 hours. E, The indicated melanoma cells were intravenously injected into B6 mice (2.5 × 105 cells/mouse) and lung metastases were quantified by counting the colonies on day 14. Data are shown as mean ± SEM (*, P < 0.05; **, P < 0.01; ns, not significant).

Close modal

Relevance of GSTA4 expression in IFNγ responsiveness and metastatic progression of human melanoma

To determine the clinical relevance of our findings in murine melanoma cells, three different human melanoma cell lines with a distinct GSTA4 gene expression status were tested regarding their responsiveness to IFNγ treatment. According to the Cancer Cell Line Encyclopedia (22), mRNA expression of GSTA4 was high, medium, and low in Malme3M, UACC 62, and MeWo, respectively. Along with the expression of GSTA4, Malme3M was more resistant to in vitro IFNγ treatment compared with the other two cell lines (Fig. 7A). Furthermore, there was a significant correlation between the expression of GSTA4 and metastasis-free survival rate of human melanoma patients based on meta-Z analysis of the PRECOG database (GSE8401; ref. 23). As shown in Fig. 7B, patients with melanoma with a low expression status of GSTA4 showed a much better prognosis in terms of metastatic progression-free survival. Importantly, melanoma patients with low GSTA4 expression were better responders (Fig. 7C) and showed better progression free survival rate (Fig. 7D) to anti PD-1 therapy (samples were taken pre- and on-treatment; ref. 24, 25). These results strongly support the clinical relevance of our findings and suggest the importance of the GSTA4 expression status in human melanoma in responsiveness to immunotherapy and risk of distant metastasis.

Figure 7.

Relevance of GSTA4 expression in IFNγ responsiveness and metastatic progression of human melanoma. A, Human melanoma cell lines (Malme3M, UACC 62, and MeWo) with different GSTA4 expression levels (according to Cancer Cell Line Encyclopedia) were treated with the indicated concentrations of IFNγ for 72 hours, and then cell viability was evaluated. Data are shown as the mean ± SEM. B, Kaplan–Meier survival plot of patients with melanoma with high or low GSTA4 expression generated by meta-Z analysis of the PRECOG database (GSE8401) is shown. C, GSTA4 expression related to the respond of anti-PD-1 treatment of melanoma patients generated by ROC plotter database is shown. D, Kaplan–Meier survival plot of patients with melanoma treated with anti-PD-1 treatment related to the high or low GSTA4 expression generated by KM plotter database is shown.

Figure 7.

Relevance of GSTA4 expression in IFNγ responsiveness and metastatic progression of human melanoma. A, Human melanoma cell lines (Malme3M, UACC 62, and MeWo) with different GSTA4 expression levels (according to Cancer Cell Line Encyclopedia) were treated with the indicated concentrations of IFNγ for 72 hours, and then cell viability was evaluated. Data are shown as the mean ± SEM. B, Kaplan–Meier survival plot of patients with melanoma with high or low GSTA4 expression generated by meta-Z analysis of the PRECOG database (GSE8401) is shown. C, GSTA4 expression related to the respond of anti-PD-1 treatment of melanoma patients generated by ROC plotter database is shown. D, Kaplan–Meier survival plot of patients with melanoma treated with anti-PD-1 treatment related to the high or low GSTA4 expression generated by KM plotter database is shown.

Close modal

In this study, we demonstrated that melanoma cells acquired resistance to IFNγ-dependent host immunity and metastatic ability by upregulating GSTA4 expression to manage cellular oxidative stress responses. Therefore, manipulation of the oxidative stress response in cancer cells may be a new therapeutic target to overcome immune-resistance and regulate metastatic progression.

The GST are a family of phase II detoxification enzymes that catalyze the conjugation of glutathione (GSH) to a variety of endogenous and exogenous electrophilic compounds, and play an important role in cellular oxidative stress responses (26). In cancer, GSTs are considered as therapeutic targets because specific isozymes of GSTs are overexpressed in various tumors, and also involved in regulating oncogenic signals and anticancer drug resistance (27). As with other GSTs, GSTA4 is expressed in a variety of normal tissues, but its specific function is to catalyze the conjugation of reduced glutathione to 4-HNE (28); therefore, GSTA4 protects cells from the oxidative stress response. During tumor development, the diverse mechanisms are often seen in cancer cells to escape from immune surveillance (7), and it is likely that management of the oxidative stress response by a cellular redox system may play a substantial role. In this context, a recent study reported that increased tumor cell lipid peroxidation by immunotherapy-activated CD8+ T cells led to tumor ferroptosis promoted by IFNγ (13), suggesting the importance of the cellular redox system to regulate the tumor response to IFNγ.

Furthermore, it is known that the redox system also plays a critical role in metastasis progression of cancer cells. In this regard, GSTA4 is known to promote the metastatic progression of human hepatocellular carcinoma by activating AKT signaling (29). Our present results clearly show that management of the oxidative stress response through the upregulation of GSTA4 allows cancer cells to escape from the IFNγ-dependent host immune response and increase the metastatic ability; however, many other molecules are also known to be involved in the cellular redox system. For example, the overexpression of c-Met initiated a redox cascade involving hydrogen peroxide generation linked to metastasis in a murine model of invasive melanoma (30) and NRF2 and ATF4 are also known to increase glutathione and heme oxygenase 1, those are typical cellular redox molecules, to promote metastasis (31). Considering that GSTA4 overexpression cannot fully protect B16OVA cells from the oxidative stress response induced by IFNγ, GSTA4 cannot be solely responsible for protecting immune-escape variants from oxidative stress responses. In our gene expression analysis, we also identified the upregulation of AKR1B8 and NQO1 genes that are known to be involved in the cellular redox system (32, 33), and those genes may have some redundant role in collaboration with GSTA4 to fully protect immune-escape variants from IFNγ-induced oxidative stress responses.

There are several limitation remains to clear in this study. First, it is quite important to clarify whether those immune escape variants with high GSTA4 expression were generated a result of selection of preexisting population, or induced by a genetic evolution during immune-editing process. Although we could not directly answer to this question in this study, we previously reported the relevant findings that antigen-specific immunity by CTL and IFNγ increases of genomic instability to immunoedits cancer cells for escaping from immunity (5). Therefore, we presume that the bearing high GSTA4 expression in immune-escape variants should be a result of genetic modification during the immunoediting process. Secondary, our presented results are mostly concluded from different variants of B16 melanoma cell lines, and other human melanoma cell lines. Although the clinical relevance of our findings in patients with human melanoma were shown, we could not conclude a general importance of GSTA4 in immune-resistance and metastatic ability of other cancer types. Nevertheless, our results suggest a novel mechanism whereby melanoma cells escape from host immunity and gain metastatic ability by acquiring resistance to oxidative stress responses through the upregulation of GSTA4.

No disclosures were reported.

S. Ucche: Formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. S. Yokoyama: Resources, data curation, supervision, investigation, methodology, writing–review and editing. M. Mojic: Data curation, formal analysis, investigation, visualization, methodology, writing–review and editing. K. Oki: Data curation, formal analysis, validation, investigation. C. Ohshima: Data curation, formal analysis, validation, investigation. H. Tsuihiji: Data curation, formal analysis, validation, investigation. I. Takasaki: Resources, data curation, formal analysis, supervision, investigation, methodology, writing–review and editing. H. Tahara: Resources, supervision, writing–review and editing. Y. Hayakawa: Conceptualization, supervision, funding acquisition, visualization, methodology, writing–original draft, project administration, writing–review and editing.

We are grateful to Asuka Asami and Setsuko Nakayama for their technical assistance and all members of Hayakawa laboratory for their support. This study was partly supported by a Grant-in-Aid for Scientific Research on Innovative Areas (17H06398 and 21H02783), The Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan (to Y. Hayakawa), Yasuda Memorial Medical Founda-tion, and the Cooperative Research Project from the Institute of Natural Medicine, University of Toyama.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/).

1.
Dunn
GP
,
Bruce
AT
,
Ikeda
H
,
Old
LJ
,
Schreiber
RD
.
Cancer immunoediting: from immunosurveillance to tumor escape
.
Nat Immunol
2002
;
3
:
991
8
.
2.
Schreiber
RD
,
Old
LJ
,
Smyth
MJ
.
Cancer immunoediting: integrating immunity's roles in cancer suppression and promotion
.
Science
2011
;
331
:
1565
70
.
3.
Dunn
GP
,
Koebel
CM
,
Schreiber
RD
.
Interferons, immunity and cancer immunoediting
.
Nat Rev Immunol
2006
;
6
:
836
48
.
4.
Benci
JL
,
Xu
B
,
Qiu
Y
,
Wu
TJ
,
Dada
H
,
Twyman-Saint Victor
C
, et al
.
Tumor interferon signaling regulates a multigenic resistance program to immune checkpoint blockade
.
Cell
2016
;
167
:
1540
54
.
5.
Takeda
K
,
Nakayama
M
,
Hayakawa
Y
,
Kojima
Y
,
Ikeda
H
,
Imai
N
, et al
.
IFN-γ is required for cytotoxic T cell-dependent cancer genome immunoediting
.
Nat Commun
2017
;
8
:
1
12
.
6.
Sharma
P
,
Allison
JP
.
The future of immune checkpoint therapy
.
Science
2015
;
348
:
56
61
.
7.
Sharma
P
,
Hu-Lieskovan
S
,
Wargo
JA
,
Ribas
A
.
Primary, adaptive, and acquired resistance to cancer immunotherapy
.
Cell
2017
;
168
:
707
23
.
8.
Chen
DS
,
Mellman
I
.
Oncology meets immunology: the cancer-immunity cycle
.
Immunity
2013
;
39
:
1
10
.
9.
Kalbasi
A
,
Ribas
A
.
Tumour-intrinsic resistance to immune checkpoint blockade
.
Nat Rev Immunol
2020
;
20
:
25
39
.
10.
Dighe
AS
,
Richards
E
,
Old
LJ
,
Schreiber
RD
.
Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFNγ receptors
.
Immunity
1994
;
1
:
447
56
.
11.
Kaplan
DH
,
Shankaran
V
,
Dighe
AS
,
Stockert
E
,
Aguet
M
,
Old
LJ
, et al
.
Demonstration of an interferon γ-dependent tumor surveillance system in immunocompetent mice
.
Proc Natl Acad Sci U S A
1998
;
95
:
7556
61
.
12.
Hubackova
S
,
Kucerova
A
,
Michlits
G
,
Kyjacova
L
,
Reinis
M
,
Korolov
O
, et al
.
IFNγ induces oxidative stress, DNA damage and tumor cell senescence via TGFβ/SMAD signaling-dependent induction of Nox4 and suppression of ANT2
.
Oncogene
2016
;
35
:
1236
49
.
13.
Wang
W
,
Green
M
,
Choi
JE
,
Gijón
M
,
Kennedy
PD
,
Johnson
JK
, et al
.
CD8+ T cells regulate tumour ferroptosis during cancer immunotherapy
.
Nature
2019
;
569
:
270
4
.
14.
Mojic
M
,
Shitaoka
K
,
Ohshima
C
,
Ucche
S
,
Lyu
F
,
Hamana
H
, et al
.
NKG2D defines tumor-reacting effector CD8+ T cells within tumor microenvironment
.
Cancer Sci
2021
;
112
:
3484
90
.
15.
Campeau
E
,
Ruhl
VE
,
Rodier
F
,
Smith
CL
,
Rahmberg
BL
,
Fuss
JO
, et al
.
A versatile viral system for expression and depletion of proteins in mammalian cells
.
PLoS One
2009
;
4
:
e6529
.
16.
Xu
X
,
Eshima
S
,
Kato
S
,
Fisher
DE
,
Sakurai
H
,
Hayakawa
Y
, et al
.
Rational combination therapy for melanoma with dinaciclib by targeting BAK-dependent cell death
.
Mol Cancer Ther
2019
;
19
:
627
36
.
17.
Haq
R
,
Yokoyama
S
,
Hawryluk
EB
,
Jönsson
GB
,
Frederick
DT
,
McHenry
K
, et al
.
BCL2A1 is a lineage-specific antiapoptotic melanoma oncogene that confers resistance to BRAF inhibition
.
Proc Natl Acad Sci U S A
2013
;
110
:
4321
6
.
18.
Suzuki
S
,
Zhou
Y
,
Refaat
A
,
Takasaki
I
,
Koizumi
K
,
Yamaoka
S
, et al
.
Human T cell lymphotropic virus 1 manipulates interferon regulatory signals by controlling the TAK1-IRF3 and IRF4 pathways
.
J Biol Chem
2010
;
285
:
4441
6
.
19.
Kato
S
,
Hayakawa
Y
,
Sakurai
H
,
Saiki
I
,
Yokoyama
S
.
Mesenchymal-transitioned cancer cells instigate the invasion of epithelial cancer cells through secretion of WNT3 and WNT5B
.
Cancer Sci
2014
;
105
:
281
9
.
20.
Fujii
H
,
Inobe
M
,
Hayakawa
Y
,
Kimura
F
,
Murakami
M
,
Onishi
Y
, et al
.
Vaccination with B7–1+ tumor and anti-adhesion therapy with RGD pseudo- peptide (FC-336) efficiently induce anti-metastatic effect
.
Clin Exp Metastasis
1998
;
16
:
141
8
.
21.
Fidler
IJ
.
Selection of successive tumour lines for metastasis
.
Nat New Biol
1973
;
242
:
148
9
.
22.
Barretina
J
,
Caponigro
G
,
Stransky
N
,
Venkatesan
K
,
Margolin
AA
,
Kim
S
, et al
.
The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity
.
Nature
2012
;
483
:
603
7
.
23.
Gentles
AJ
,
Newman
AM
,
Liu
CL
,
Bratman
SV
,
Feng
W
,
Kim
D
, et al
.
The prognostic landscape of genes and infiltrating immune cells across human cancers
.
Nat Med
2015
;
21
:
938
45
.
24.
Tibor Fekete
J
,
Győrffy
B
.
A unified platform enabling biomarker ranking and validation for 1,562 drugs using transcriptomic data of 1,250 cancer cell lines
.
Comput Struct Biotechnol J
2022
;
20
:
2885
94
.
25.
Lánczky
A
,
Győrffy
B
.
Web-based survival analysis tool tailored for medical research (KMplot): development and implementation
.
J Med Internet Res
2021
;
23
:
e27633
.
26.
Sheehan
D
,
Meade
G
,
Foley
VM
,
Dowd
CA
.
Structure, function and evolution of glutathione transferases: implications for classification of non-mammalian members of an ancient enzyme superfamily
.
Biochem J
2001
;
360
:
1
16
.
27.
Townsend
DM
,
Tew
KD
.
The role of glutathione- S -transferase in anti-cancer drug resistance
.
Oncogene
2003
;
22
:
7369
75
.
28.
Balogh
LM
,
Atkins
WM
.
Interactions of glutathione transferases with 4-hydroxynonenal
.
Drug Metab Rev
2011
;
43
:
165
78
.
29.
Liu
CJ
,
Yang
JH
,
Huang
FZ
,
Nie
WP
,
Liu
CP
,
Mao
XH
, et al
.
Glutathione-s-transferase A 4 (GSTA4) suppresses tumor growth and metastasis of human hepatocellular carcinoma by targeting AKT pathway
.
Am J Transl Res
2017
;
9
:
301
15
.
30.
Ferraro
D
,
Corso
S
,
Fasano
E
,
Panieri
E
,
Santangelo
R
,
Borrello
S
, et al
.
Pro-metastatic signaling by c-Met through RAC-1 and reactive oxygen species (ROS)
.
Oncogene
2006
;
25
:
3689
98
.
31.
Dey
S
,
Sayers
CM
,
Verginadis
II
,
Lehman
SL
,
Cheng
Y
,
Cerniglia
GJ
, et al
.
ATF4-dependent induction of heme oxygenase 1 prevents anoikis and promotes metastasis
.
J Clin Invest
2015
;
125
:
2592
608
.
32.
Liu
Y
,
Zhang
J
,
Liu
H
,
Guan
G
,
Zhang
T
,
Wang
L
, et al
.
Compensatory upregulation of aldo-keto reductase 1B10 to protect hepatocytes against oxidative stress during hepatocarcinogenesis
.
Am J Cancer Res
2019
;
9
:
2730
48
.
33.
Ross
D
,
Siegel
D
.
Functions of {NQO}1 in cellular protection and {CoQ}10 metabolism and its potential role as a redox sensitive molecular switch
.
Front Physiol
2017
;
8
:
595
.