IFN-γ plays a critical role in tumor immunosurveillance by affecting either immune cells or tumor cells; however, IFN-mediated effects on tumor elimination are largely unknown. In this study, we showed that IFN regulatory factors (IRF) modulated by IFNs up- and downregulated Noxa expression, a prodeath BH3 protein, in various cancer cells. Inhibition of Noxa expression using short hairpin RNA in tumor cells leads to resistance against lipopolysaccharide (LPS)-induced tumor elimination, in which IFN-γ is known as a critical effecter in mice. Chromatin immunoprecipitation analysis in both CT26 cells and SP2/0 cells, sensitive and resistant to LPS-induced tumor elimination, respectively, revealed that the responsiveness of IRF1, 3, 4, and 7 in the Noxa promoter region in response to IFN-γ might be crucial in LPS-induced tumor elimination. IRF1, 3, and 7 were upregulated by IFN-γ and activated Noxa expression, leading to the death of Noxa wild-type baby mouse kidney (BMK) cells but not of Noxa-deficient BMK cells. In contrast, IRF4 acts as a repressor for Noxa expression and inhibits cell death induced by IRF1, 3, or 7. Therefore, although IFN-γ alone are not able to induce cell death in tumor cells in vitro, Noxa induction by IFN-γ, which is regulated by the balance between its activators (IRF1, 3, and 7) and its repressor (IRF4), is crucial to increasing the susceptibility of tumor cells to immune cell-mediated cytotoxicity. Mol Cancer Res; 9(10); 1356–65. ©2011 AACR.

This article is featured in Highlights of This Issue, p. 1267

The immune system can recognize and eliminate primary transformed tumor cells through a complicated process known as tumor immunosurveillance (1, 2). Various cytokines, including IFN-γ (3–5), perforin (6), granulocyte macrophage colony stimulating factor (GM-CSF; ref. 7), TNF-related apoptosis-inducing ligand (8), and interleukin 12 (9) have been implicated in the process of tumor cell elimination by immune cells. The inhibitory effect of IFN-γ on tumor cell growth was first identified when mice transplanted with tumor cells treated with a neutralizing monoclonal antibody to IFN-γ showed resistance to lipopolysaccharide (LPS)-induced tumor elimination (3). In addition, subsequent studies using mice lacking IFN-γ, IFN-γ/GM-CSF, or IFN-γ/perforin showed a high incidence of chemically induced or spontaneous tumor formation, showing the role of IFN-γ on the host immune cells involved in tumor elimination (7, 9–11). IFN-γ–unresponsive Stat1-deficient or IFN-γ receptor (IFNGR)-deficient mice that failed to reject immunogenic tumors further established the role of IFN-γ in the host immune system (5, 12). Mice lacking γδ T cells have been shown to exhibit a higher incidence of tumor formation and impaired endogenous IFN-γ production upon induction by either methylcholanthrene injection or inoculation of B16 melanoma cells. This suggests that γδ T cells are a primary cellular source of IFN-γ for tumor cell elimination (13). Natural killer (NK) cells have also been suggested as another major source of IFN-γ (14, 15).

In addition to the role of IFN-γ on immune cells, IFN-γ also has a critical and direct effect on tumor cells. The first indication about the role of IFN-γ in tumor cells was obtained by the observation that tumors generated with IFN-γ–insensitive Meth A.mgRΔIC tumor cells (Meth A transfectants with truncated murine IFN-γ receptor, a dominant-negative IFNGR1 mutant) were resistant to LPS-mediated tumor regression (3). In addition, RAD.gR28 tumor cells derived from IFNGR1-deficient mice are highly tumorigenic when transplanted into syngenic wild-type mice whereas RAD.gR28.mgR tumor cells reconstituted with wild-type IFNGR1 fail to form tumors in wild-type mice (5), indicating that the responsiveness to IFN-γ is a critical factor for tumor cell elimination. Furthermore, a downstream effect of IFN-γ on tumor cells has been shown to recover tumor cell immunogenicity by inducing the expression of MHC class I antigen processing components such as TAP1 or LMP2. Restoration of TAP1 expression in antigen presentation pathway-deficient lung carcinoma CMT.64 cells increases tumor-specific immune responses and animal survival (16–19).

Although it has been shown that IFN-γ plays a critical role in tumor surveillance, the effects of IFN-γ on cell death machinery proteins such as BH3 proteins in conjunction with tumor cell elimination have not been well established. The purpose of this study was to extend these findings and establish the role of IFN-γ on tumor cells by examining changes of BH3 profiles in response to IFN-γ. We have found that IFN-γ alters BH3 profiles by upregulating and downregulating prodeath and antideath BH3 proteins, respectively, and that Noxa induction by IFN-γ, which is regulated by the balance between its activators [IFN regulatory factors (IRF)1, 3, and 7] and its repressor (IRF4), is crucial to increasing the susceptibility of tumor cells to immune cell-mediated cytotoxicity.

Cell cultures and reagents

HCT116 parental cells, HCT116 p53-deficient cells (provided by Dr. Vogelstein, Johns Hopkins University, Baltimore, MD), CT26 cells [purchased from Korean cell line bank (KCLB)], baby mouse kidney (BMK) WT and Noxa-deficient cells (provided by Dr. Hiscott, McGill University, Canada; ref. 20), Jurkat cells (purchased from KCLB), HeLa cells (purchased from KCLB), and SP2/0 cells (purchased from KCLB) were cultured in McCoy's 5A media, Dulbecco's modified Eagle's medium, or RPMI with 10% FBS, supplemented with 2 mmol/L l-glutamate, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37°C under 5% CO2 in humidified incubator. Recombinant human IFN-α, IFN-β, and human/murine IFN-γ were purchased from Cytolab/Perotech Asia. 5-Flurouracil (5-FU) and LPS derived from Escherichia coli were purchased from Aldrich-Sigma. Velcade was obtained from Janssen Pharmaceutica N.V. Anti-actin antibody from Chemicon International, anti-Bad, anti–Bcl-2, anti–Bcl-XL, anti-Bim, anti-IRF1, anti-IRF3 antibodies, and Flag-tagged antibody from Cell signaling Technology, anti-Bak antibody from Upstate Biotechnology, anti-Bax, anti-IRF4, and anti-GFP antibodies from Santa Cruz Biotechnology Inc., anti-Bid and anti-IRF7 antibodies from Prosci Incorporated, anti–Mcl-1 antibody from BD Transduction Laboratories, anti-Noxa, anti-p53 antibody from Oncogene Research Products, and horseradish peroxidase (HRP)-conjugated goat anti-rabbit/-mouse IgGs from Jackson ImmunoReasearch Lab were purchased.

Plasmid construction of Noxa shRNA, scNoxa shRNA, and IRFs

Vector pC6A:GFP:H1 was derived from pcDNA6A plasmid (Invitrogen) to express the green fluorescent protein (GFP) and short hairpin RNA (shRNA) under the CMV promoter and H1 promoter, respectively. Oligonucleotides for Noxa shRNA and scrambled Noxa shRNA were synthesized and cloned into the pC6A:GFP:H1 vectors under the H1 promoter named pSUPER Noxa shRNA and pSUPER scNoxa shRNA, respectively. The sequences for Noxa shRNA and scNoxa shRNA are listed in Supplementary Table S1. pSUPER Noxa shRNA or pSUPER scNoxa shRNA plasmid was transfected into CT26 cells by Effectene (Qiagen), and the transfected CT26 cells were selected with blasticidin (10 μg/mL) for 48 hours. The selected cells were used for tumor generation in BALB/c mice or treated with IFN-γ (10 ng/mL) at indicated time points. The full-length cDNA fragments of human IRF1, 3, 4, and 7 amplified by PCR were cloned into N-terminal flag-tagged PCR8/GW/Topo entry vector, and confirmed by DNA sequencing analysis. The open reading frame region of IRFs in entry vector was subcloned into N-terminal GFP-tagged destination vector following standard protocol of Gateway Technology (Invitrogen).

Syngeneic tumor model

Eight week-old male BALB/c mice were purchased from the Samtako (Samtako Bio-Korea). The protocol used for animal experiments was approved by Committee of Chosun University Animal Experiment Centre, and all animals were treated as humanely as possible. Ten BALB/c mice were used for each experimental group. The subcutaneous tumor was induced by injection of 1.5 × 105 cells of CT26 WT or CT26 cells expressing Noxa shRNA in 100 μL PBS into the subcutaneous tissue of the flank of the mice. Tumor cells were grown for 12 days, then tumor-bearing mice were challenged by intraperitoneal injection of LPS (3 mg/kg) or PBS. The tumor size was measured with Vernier caliper and the volume of the tumor was calculated as length × width2 × 0.5.

Histologic examination of tumor

Mice were sacrificed at 4 days after the LPS/PBS injection before complete regression of tumor, and the tumor tissues and internal organs such as liver, heart, spleen, lung, and kidney were excised from each animal, fixed on 10% neutral buffered formalin solution, and embedded in paraffin. Sections of fixed tissue were stained with hematoxylin and eosin (H&E).

Reverse transcriptase PCR

Total RNA was isolated from cells using TRIzol (Invitrogen) following the standard protocol. Two micrograms of total RNA was reverse transcribed with Improm-II reverse transcriptase (Promega) following standard manufacturer protocol. About 200 ng of the reverse transcribed cDNA was analyzed by semiquantitative reverse transcriptase PCR (RT-PCR). Twenty-five cycles of amplification following denaturation at 94°C for 30 seconds, annealing at 60°C for 40 seconds, and extension at 72°C for 30 seconds were done. The primer sets for Noxa (5′-AGCTACCACCTGAGTTCGCA-3′ and 5′-TCACTTTGTCTCCAATCCTCCG-3′) were used.

Western blotting

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mmol/L Tris-Cl pH 7.5, and 150 mmol/L NaCl) in the presence of 1 mmol/L phenylmethylsulfonylfluoride, 2 μg/mL aprotinin, 2 μg/mL leupeptin, 1 μg/mL pepstatin A, 2 mmol/L sodium fluoride, and 1 mmol/L sodium orthovanadate. Soluble lysates were subjected to SDS-PAGE and transferred to a polyvinylidene fluoride membrane (Bio-Rad). The membrane was then probed with specific antibodies, followed by HRP-conjugated anti-IgG antibody. Immobilon Western chemiluminescent HRP substrate (ECL; Millipore) was used for the visualization.

ChIP analysis

CT26 wild-type (2 × 108) cells were seeded on 100-mm dish (Nunc) and treated with IFN-γ (10 ng/mL) for the indicated time points. Bound proteins were cross-linked to chromatin by adding media containing 1% formaldehyde (Sigma-Aldrich) dropwise, directly on dish and rotated gently at room temperature for 10 minutes. After sonication, cross-linked protein–DNA complexes were immunoprecipitated from nuclear extracts by using the indicated anti-IRFs antibody and then captured on protein A sepharose beads (GE, Healthcare). After immunoprecipitation and elution from the beads, DNA–protein cross-linking was reversed by addition of proteinase K and incubation at 65°C overnight. The DNA was then purified with phenol extraction followed by ethanol precipitation. PCR amplification of Noxa promoter was done using specific primers detailed in Supplementary Table S2.

Statistical analysis

Values are presented as mean ± SD. The paired student's t test was applied to determine whether difference between values were significant. A statistical probability of P < 0.05 was considered as significant.

To search for the internal effector molecule(s) of the death machinery in tumors, which are crucial mediators of IFN-γ in LPS-induced tumor regression, a BALB/C mouse syngeneic tumor model was adapted. Mouse colon carcinoma CT26 cells and mouse myeloma SP2/0 cells–2 tumor generating cell lines derived from BALB/C mice–were subcutaneously injected into the flanks of BALB/C mice. When the tumors grew to about 200 mm3, LPS (3 mg/kg) or PBS was intraperitoneally (i.p.) injected into the mice. Tumors generated by CT26 cells were completely eradicated by LPS injection, whereas the PBS-injected tumors remained unaffected (Fig. 1A). The tumors generated with SP2/0 cells, however, showed no reduction in response to either LPS or PBS (data not shown).

Figure 1.

LPS-induced tumor regression and BH3 profiles in tumors treated with LPS, and Noxa is the key mediator of LPS-induced tumor regression in mice. A, tumors were developed by subcutaneous injection of CT26 cells (1.5 × 105) into the flanks of BALB/c mice. Following tumor development, the mice were i.p. challenged with LPS (3 mg/kg) or PBS at day 12 (n = 10). Tumor volume was calculated as the longest diameter × width2 × 0.5. B, tumors were developed by subcutaneous injection of CT26 cells (1.5 × 105) or SP2/0 cells (6 × 105) into the flanks of mice. Tumor masses were excised at 4 days after LPS or PBS injection and lysed with RIPA buffer. Western blot analyses were carried out using the indicated antibodies. C and D, mice were subcutaneously injected with 1.5 × 105 cells of either CT26 parental cells, pSUPER scNoxa shRNA or pSUPER Noxa shRNA. At day 12, LPS (3 mg/kg) or PBS (control) was administered i.p. Tumor volume was calculated as described in Figure 1A. The data were obtained from 3 separate experiments and is presented as tumor size (mean ± SD) of 5 mice per group. D, mice were sacrificed on day 4 and tumors were obtained from mice challenged with LPS or PBS and stained with H&E. Images are 400× magnification.

Figure 1.

LPS-induced tumor regression and BH3 profiles in tumors treated with LPS, and Noxa is the key mediator of LPS-induced tumor regression in mice. A, tumors were developed by subcutaneous injection of CT26 cells (1.5 × 105) into the flanks of BALB/c mice. Following tumor development, the mice were i.p. challenged with LPS (3 mg/kg) or PBS at day 12 (n = 10). Tumor volume was calculated as the longest diameter × width2 × 0.5. B, tumors were developed by subcutaneous injection of CT26 cells (1.5 × 105) or SP2/0 cells (6 × 105) into the flanks of mice. Tumor masses were excised at 4 days after LPS or PBS injection and lysed with RIPA buffer. Western blot analyses were carried out using the indicated antibodies. C and D, mice were subcutaneously injected with 1.5 × 105 cells of either CT26 parental cells, pSUPER scNoxa shRNA or pSUPER Noxa shRNA. At day 12, LPS (3 mg/kg) or PBS (control) was administered i.p. Tumor volume was calculated as described in Figure 1A. The data were obtained from 3 separate experiments and is presented as tumor size (mean ± SD) of 5 mice per group. D, mice were sacrificed on day 4 and tumors were obtained from mice challenged with LPS or PBS and stained with H&E. Images are 400× magnification.

Close modal

To investigate the differential LPS responsiveness of the cell types in detail, the profiles of BH3 domain containing proteins (hereafter referred to as BH3 profiles) were determined in LPS-treated tumor tissues bearing CT26 and SP2/0 cells. LPS treatment significantly increased the expression levels of the prodeath BH3 proteins like Noxa, Bim, Bad, Bax, and Bak in tumors bearing CT26 cells, whereas the expressions of the antideath BH3 proteins like Bcl-xL and Bcl-2 were decreased (Fig. 1B, CT26). Interestingly, the expression of the Mcl-1 protein, a protein selectively inhibited by Noxa (21, 22), was relatively sustained in response to LPS in tumors bearing CT26 cells (Fig. 1B, CT26). However, tumors bearing SP2/0 cells showed significantly increased expression levels of Mcl-1, Bcl-xL, and Bcl-2 but exhibited reduced or unchanged expression levels of Bim, Noxa, and Bad in response to LPS (Fig. 1B, SP2/0). It is worth noting that LPS itself showed little or no death-inducing activity in CT26 or SP2/0 cells. These results indicate that tumors bearing CT26 cells can be eliminated by LPS due to LPS-induced increased expression of prodeath BH3 proteins and decreased expression of antideath BH3 proteins; however, it follows from our previous observations that tumors bearing SP2/0 cells cannot be eliminated by LPS due to the increased expression levels of antideath BH3 proteins and the absence of any induction of the prodeath BH3 proteins Noxa, Bim, and Bad. Taken together, our results suggest that alteration of the BH3 profile in tumor cells may be a crucial factor in determining LPS susceptibility of tumor cells.

As depicted in Supplementary Figure S1A, Bad is a suppressor of Bcl-2 and Bcl-xL, and Noxa is a suppressor of Mcl-1 (22). CT26 cells sustained the expression of Mcl-1 but decreased the expression of Bcl-2 and Bcl-xL in tumors injected with LPS. Because Noxa-induced suppression takes significance over Bad-induced suppression in antideath activity of Mcl-1, we were compelled to investigate the role of Noxa in LPS-induced tumor regression using CT26 cells.

To further investigate whether Noxa induction is a necessary event in LPS-induced tumor elimination, tumors were generated in BALB/c mice by subcutaneous injection of CT26 parental cells, CT26 cells expressing pSUPER:scrambled (sc) Noxa shRNA (CT26:scNoxa shRNA), or CT26 cells expressing pSUPER:Noxa shRNA (CT26:Noxa shRNA). It was confirmed that Noxa shRNA (but not scNoxa shRNA) significantly inhibited Noxa expression as shown in Supplementary Figure S1B. Tumors generated with CT26 parental cells or CT26:scNoxa shRNA cells were completely regressed by LPS injection, whereas tumors generated with CT26:Noxa shRNA cells were resistant to LPS-induced tumor regression (Fig. 1C). These results were confirmed by H&E staining of LPS-treated tumor tissues generated by CT26 parental, CT26:scNoxa shRNA, and CT26:Noxa shRNA cells. Massive cell death was observed in tumor tissues from CT26 parental cells and CT26:scNoxa shRNA cells but not in tumor tissues from CT26:Noxa shRNA cells (Fig. 1D). Noxa induction in tumor tissues, as well as in other normal organs such as the heart and liver, was confirmed by RT-PCR (Supplementary Fig. S1C). It should be noted, however, that severe damage in normal tissues was not observed by microscopic H&E analysis (data not shown). Collectively, our results indicate that Noxa induction in tumor cells is an obligate event for LPS-induced tumor regression.

It was previously reported that IFN-γ is a key mediator of LPS-induced tumor regression (3). As we have shown in Figure 1B, BH3 profiles in tumors bearing CT26 cells are altered by LPS in our animal model. Therefore, we hypothesized that IFN-γ may influence the expression of BH3 proteins in tumor cells, which, in turn, may increase the susceptibility of tumor cell death. The effect of IFN-γ on BH3 profiles was analyzed in both CT26 cells sensitive to LPS-induced tumor elimination and SP2/0 cells resistant to LPS-induced tumor elimination, and we found that in CT26 cells, IFN-γ significantly increased the expression levels of Noxa, Bad, Bax, Bak, and Bim although sustaining or reducing the expression levels of Mcl-1, Bcl-2, and Bcl-xL (Fig. 2A); conversely, in SP2/0 cells, IFN-γ failed to increase the expression levels of Noxa, Bim, Bad, Bax, and Bak, whereas it increased the expression levels of Mcl-1, Bcl-2, and Bcl-xL (Fig. 2A). BH3 profiles of CT26 and SP2/0 cells in response to IFN-α, IFN-β, and IFN-γ showed similarities with those observed in tumors bearing CT26 cells and SP2/0 cells injected with LPS (Supplementary Fig. S2A, top panel and summarized in Supplementary Fig. S2A, bottom panel). Furthermore, the changes of BH3 profiles by IFN-α, IFN-β, and IFN-γ (IFNs) were independent of p53 status observed using HCT116 parental cells and HCT116 p53-deficient cells (Supplementary Fig. S2B). These results indicate that BH3 profiles of tumor cells in vivo in response to LPS may reflect the effects of IFNs on tumor cells.

Figure 2.

Noxa induction by IFNs is p53-independent. A, CT26 cells or SP2/0 cells were treated with IFN-γ (10 ng/mL) at the indicated times. Whole-cell lysates were prepared using RIPA buffer and subjected to Western blot analysis using the indicated antibodies. Actin was used as the loading control. Data are representative of 3 independent experiments. B, HeLa, Jurkat, and HCT116 cells were treated with IFN-α, IFN-β, and IFN-γ (10 ng/mL) for 0, 6, 12, or 24 hours. Western blot analyses were carried out using anti-Noxa antibody. C, HCT116 parental cells (left) and HCT116 p53-/- cells (right) were treated with IFN-α, IFN-β, and IFN-γ for 0, 6, and 12 hours. After treatment, whole-cell lysates were analyzed with the indicated antibodies.

Figure 2.

Noxa induction by IFNs is p53-independent. A, CT26 cells or SP2/0 cells were treated with IFN-γ (10 ng/mL) at the indicated times. Whole-cell lysates were prepared using RIPA buffer and subjected to Western blot analysis using the indicated antibodies. Actin was used as the loading control. Data are representative of 3 independent experiments. B, HeLa, Jurkat, and HCT116 cells were treated with IFN-α, IFN-β, and IFN-γ (10 ng/mL) for 0, 6, 12, or 24 hours. Western blot analyses were carried out using anti-Noxa antibody. C, HCT116 parental cells (left) and HCT116 p53-/- cells (right) were treated with IFN-α, IFN-β, and IFN-γ for 0, 6, and 12 hours. After treatment, whole-cell lysates were analyzed with the indicated antibodies.

Close modal

As shown in Figure 1, Noxa induction in tumor cells bearing CT26 cells in response to LPS is crucial for LPS-induced tumor regression; thus, we further examined the expression levels of Noxa in HeLa, Jurkat, and HCT116 cells. In these cells, IFNs strongly increased the expression of Noxa (Fig. 2B). Noxa was originally identified as a p53 target gene (23) and subsequent studies showed that velcade, IFN-α, and IFN-β induced Noxa in a p53-independent manner (24–27). In consistence with these reports, our results have also indicated that velcade, IFN-α, IFN-β, and IFN-γ strongly induced Noxa in a p53-independent manner in both HCT116 parental cells and HCT116 p53-deficient cells (Fig. 2C and Supplementary Fig. S2A). In contrast, the DNA damaging agent 5′-FU induced Noxa in a p53-dependent manner (Supplementary Fig. S2C). Together, these results show that IFNs induced Noxa expression is not closely associated with p53.

To understand the molecular mechanism by which IFNs can transactivate Noxa in a p53-independent manner, transcription factor–binding elements in the mouse Noxa promoter region were identified by a computer analysis using MatInspector in Genomatix (http://www.genomatix.de/). We found many putative binding elements for IRFs 1, 3, 4, and 7 located within 8 kb upstream of the translational start site (Supplementary Fig. S3). It was previously reported that IRF, a well-known transcriptional factor activated by IFNs and toll-like receptors (TLR), is involved in feedback regulation of IFNs (28, 29), immune cell differentiation (30, 31), and tumor suppression and proliferation (25). In our study, levels of IRF1, 3, and 7 were increased by IFNs in CT26 cells (Fig. 3A), HCT116 parental cells, and HCT116 p53-deficient cells (Supplementary Fig. S4), indicating that the presence of p53 is not necessary for induction of IRF1, 3, and 7 by IFNs. In CT26-derived tumors challenged with LPS, levels of IRF1, 3, and 7 were also significantly increased; however, in SP2/0 cells, IRF1 and 3 showed no changes, whereas IRF7 was increased. Interestingly, IRF4, which can act as a negative regulator of TLR signaling (31), were decreased in tumors derived from CT26 but was increased in those derived from SP2/0 cells (Fig. 3B).

Figure 3.

IFNs upregulate the expression levels of IRF1, 3, and 7. A, CT26 cells were treated with IFN-α, IFN-β, and IFN-γ (10 ng/mL) for 0, 6, 12, and 24 hours. The whole-cell lysates were analyzed with indicated antibodies. B, tumor tissues generated with CT26 cells or SP2/0 cells were obtained from LPS- or PBS-treated mice 4 days after treatment. The lysates were analyzed with the indicated antibodies.

Figure 3.

IFNs upregulate the expression levels of IRF1, 3, and 7. A, CT26 cells were treated with IFN-α, IFN-β, and IFN-γ (10 ng/mL) for 0, 6, 12, and 24 hours. The whole-cell lysates were analyzed with indicated antibodies. B, tumor tissues generated with CT26 cells or SP2/0 cells were obtained from LPS- or PBS-treated mice 4 days after treatment. The lysates were analyzed with the indicated antibodies.

Close modal

To investigate whether or not these IRFs participate in the regulation of Noxa gene expression, we conducted chromatin immunoprecipitation (ChIP) analysis using IFN-γ–treated CT26 and SP2/0 cells. IRF1, 3, and 7 bound to several putative binding sites in CT26 cells treated with IFN-γ, although remaining unbound in the untreated CT26 cells (Fig. 4A). One site (−3454∼−3223) for IRF1, 4 sites (−6259∼−5957; −5113∼−4750) for IRF3, and 4 sites (−5295∼−5055; −3812∼−3633; −2208 ∼ 2057) for IRF7 were responsible for Noxa transactivation in response to IFN-γ in CT26 cells (Fig. 4A and summarized in Fig. 4B, some primer sets cover 2 IRF-binding sites), implicating IRF1, 3, and 7 as activators for Noxa expression. On the other hand, IRF4 seemed to be a repressor for Noxa expression because its binding sites were occupied only in untreated CT26 cells and not in IFN-γ–treated CT26 cells (Fig. 4A). In SP2/0 cells, IRF1 and IRF7 bound to many putative binding sites (−3454∼−3223; −5295∼−5055; −3812∼−3633; −2208 ∼ 2057; −641∼−417) in cells treated with IFN-γ but not in IFN-γ–untreated cells, showing a binding site pattern similar to that observed in CT26 cells. On the other hand, IRF3 failed to bind to its binding sites (−6259∼−5957; −5113∼−4750) in SP2/0 cells treated with IFN-γ. Moreover, IRF4-binding sites (−6583∼−6420; −4755∼−4589) continued to be occupied by IRF4 in IFN-γ–treated SP2/0 cells, indicating that in SP2/0 cells, IFN-γ does not detach IRF4 proteins from IRF4-binding sites (−6583∼−6420; −4755∼−4589) in the Noxa promoter region (Fig. 4C). These results suggest that IRF1, 3, and 7 are activators for Noxa expression in response to IFN-γ, whereas IRF4 is a repressor for Noxa expression.

Figure 4.

Chromatin precipitation analysis of mouse Noxa promoter region. A, CT26 cells or (C) SP2/0 cells for 12 hours were treated with IFN-γ and were subjected to ChIP analysis using the indicated IRF antibodies and primer sets. The input DNAs from cell lysates before immunoprecipitation was used as the positive control. Elutes from beads without antibodies were used as the negative controls. The data are representative of 3 separate experiments. The data for each IRFs are presented in the separate panels. B, schematic diagram of functional IRFs binding elements in CT26 cells in response to IFN-γ is presented.

Figure 4.

Chromatin precipitation analysis of mouse Noxa promoter region. A, CT26 cells or (C) SP2/0 cells for 12 hours were treated with IFN-γ and were subjected to ChIP analysis using the indicated IRF antibodies and primer sets. The input DNAs from cell lysates before immunoprecipitation was used as the positive control. Elutes from beads without antibodies were used as the negative controls. The data are representative of 3 separate experiments. The data for each IRFs are presented in the separate panels. B, schematic diagram of functional IRFs binding elements in CT26 cells in response to IFN-γ is presented.

Close modal

To elucidate the roles of IRFs in Noxa expression as an activator or repressor, we conducted transient transfection experiments using individual IRF expression vectors. We found that Noxa was significantly increased by transfection of individual IRF1, 3 or 7 expression vectors, resulting in a reduction of cell viability (Fig. 5A). To further confirm that Noxa induction is an essential event in IRF1-, 3-, or 7-induced cell death, immortalized BMK cells from wild-type or Noxa-deficient mice were infected with adenovirus expressing either GFP (as a control) or one of IRFs 1, 3, or 7. Noxa expression was increased by IRF1, 3, and 7 in wild-type BMK cells but not in Noxa-deficient BMK cells (Fig. 5B). Also, IRF1, 3, and 7 induced cell death in wild-type BMK cells but not in Noxa-deficient BMK cells (Fig. 5B). These results show that Noxa is a key factor in cell death induced by IRFs 1, 3, and 7, suggesting that Noxa induction in tumor cells by IRFs 1, 3, and 7 in response to IFN-γ may be an essential event in LPS-induced tumor regression.

Figure 5.

IRF1, 3, and 7 are the activators for Noxa expression, and IRF4 is a repressor for Noxa expression. A, the expression level of Noxa was estimated using Flag-tagged IRF1-, Flag-IRF3-, and Flag-IRF7–transfected HCT116 cells. HCT116 parental cells and Flag-tagged empty vector–transfected HCT116 cells (Flag-XB) served as controls. At 24 hours after transfection, the percentage of cell death was measured by means of trypan blue exclusion assay. B, BMK WT and BMK Noxa-deficient cells were infected with adenovirus expressing GFP (Ad-GFP), Ad-IRF1, Ad-IRF3, and Ad-IRF7 [all at a multiplicity of infection (MOI) of 50], and the cell lysates harvested 24 hours after infection were used for Western blot determination of IRF1, 3, and 7 expression. Noxa expression level was assayed by RT-PCR. BMK WT and Noxa-deficient cells were infected with 25 MOI of Ad-GFP, Ad-IRF1, Ad-IRF3, or Ad-IRF7. Forty-eight hours after infection, the percentage of cell death was calculated on the basis of trypan blue exclusion.

Figure 5.

IRF1, 3, and 7 are the activators for Noxa expression, and IRF4 is a repressor for Noxa expression. A, the expression level of Noxa was estimated using Flag-tagged IRF1-, Flag-IRF3-, and Flag-IRF7–transfected HCT116 cells. HCT116 parental cells and Flag-tagged empty vector–transfected HCT116 cells (Flag-XB) served as controls. At 24 hours after transfection, the percentage of cell death was measured by means of trypan blue exclusion assay. B, BMK WT and BMK Noxa-deficient cells were infected with adenovirus expressing GFP (Ad-GFP), Ad-IRF1, Ad-IRF3, and Ad-IRF7 [all at a multiplicity of infection (MOI) of 50], and the cell lysates harvested 24 hours after infection were used for Western blot determination of IRF1, 3, and 7 expression. Noxa expression level was assayed by RT-PCR. BMK WT and Noxa-deficient cells were infected with 25 MOI of Ad-GFP, Ad-IRF1, Ad-IRF3, or Ad-IRF7. Forty-eight hours after infection, the percentage of cell death was calculated on the basis of trypan blue exclusion.

Close modal

As shown in Figure 4A, the binding sites of IRFs 1, 3, and 7 were occupied only in IFN-γ–treated CT26 cells but not in IFN-γ–untreated CT26 cells; we thus hypothesized that IRF4 is a repressor for Noxa induction mediated by IRF1, 3, or 7. Thus, we measured Noxa expression levels and the percentage of cell death when IRF4 was coexpressed with IRF1, 3, or 7 by using individual expression vectors. IRF4 repressed the expression of Noxa upregulated by IRFs 1, 3, and 7 and recovered the level of cell viability back to the control (Fig. 6A). Our results show that IRF4 is a repressor for Noxa expression by IRF 1, 3, and 7, however, it could not repress the basal level of Noxa. To further investigate whether or not IRF4 repression of Noxa expression can be derepressed by IRFs 1, 3, and 7, HCT116 cells were cotransfected with the IRF4 expression vector (constant amounts) and IRF1, 3, and 7 expression vectors (variable amounts). Although IRF1 only weakly derepressed the repression activity of IRF4 on Noxa expression, IRFs 3 and 7 significantly derepressed the repression activity of IRF4 on Noxa expression, resulting in increased protein levels of Noxa in a concentration-dependent manner (Fig. 6B). It would also be of interest to note that IRF4 expression levels were decreased by some unknown mechanism as the expression levels of IRFs 1, 3, and 7 were increased (Fig. 6B). Taken together, these results indicate that IRF4 represses IRF1-, 3-, and 7-induced Noxa expression and that this repression can, in turn, be derepressed by IRFs 1, 3, and 7.

Figure 6.

Derepression of IRF4 by IRF1, 3, or 7. A, HCT116 parental cells were transfected with Flag-tagged vector or GFP vector and were cotransfected with Flag-IRF1, Flag-IRF3, Flag-IRF7, and GFP-IRF4 vectors. The expression levels of Noxa and IRFs were analyzed by Western blot using the indicated antibodies. The percentage of cell death was measured by means of trypan blue exclusion assay. B, HCT116 cells were transfected with indicated amounts of GFP-IRF4 expression vector (0 or 500 ng) and expression vectors of Flag-IRF1, Flag-IRF3, or Flag-IRF7 (100, 200, 300, 400, 500, or 500 ng). Whole-cell lysates were prepared 24 hours after transfection and were subjected to Western blot analysis using anti-Noxa, anti-GFP (for IRF4), and anti-Flag (for IRF1, 3, and 7) antibodies.

Figure 6.

Derepression of IRF4 by IRF1, 3, or 7. A, HCT116 parental cells were transfected with Flag-tagged vector or GFP vector and were cotransfected with Flag-IRF1, Flag-IRF3, Flag-IRF7, and GFP-IRF4 vectors. The expression levels of Noxa and IRFs were analyzed by Western blot using the indicated antibodies. The percentage of cell death was measured by means of trypan blue exclusion assay. B, HCT116 cells were transfected with indicated amounts of GFP-IRF4 expression vector (0 or 500 ng) and expression vectors of Flag-IRF1, Flag-IRF3, or Flag-IRF7 (100, 200, 300, 400, 500, or 500 ng). Whole-cell lysates were prepared 24 hours after transfection and were subjected to Western blot analysis using anti-Noxa, anti-GFP (for IRF4), and anti-Flag (for IRF1, 3, and 7) antibodies.

Close modal

The results of our study reveal the underlying mechanism surrounding LPS-induced tumor regression. IFN-γ modulates the expressions of BH3 proteins (BH3 profiles) in tumors. In particular, Noxa is upregulated by the activator proteins IRF1, 3, and 7, and is downregulated by the repressor protein IRF4. Modulation of BH3 profiles by IFNs is likely to be a determining factor for LPS-induced tumor regression because blocking of prodeath BH3 protein Noxa expression using Noxa shRNA inhibits tumor regression by LPS in BALB/C mice. This view is further supported by the BH3 profiles in SP2/0 cells that fail to upregulate or downregulate the pro- or antideath BH3 proteins, respectively, resulting in complete resistance to LPS-induced tumor regression. However, the modulation of BH3 profiles by IFNs in tumor cells in itself may not be enough to eliminate tumors in vivo because IFNs themselves can minimally induce tumor cell death in vitro. Thus, it is more likely that IFNs sensitize tumor cells in vivo by modulating the cellular BH3 profiles so that immune cells or cytokines activated by LPS or IFNs are able to stimulate the death of tumor cells in vivo.

It has been well documented that TLRs recognize the patterns of pathogen-associated molecules such as LPS and viral nucleic acids and transmit signals to IRFs for induction of type I IFNs, cytokines, and chemokines through adaptor proteins (such as MyD88, TRIF, TIRAP, and TRAM) and kinases (such as TBK1, IRAKs, and IKK; refs. 28, 32, 33). Consistent with previous reports showing that IFNs transactivate Noxa in endothelial cells, primary macrophages, and various tumor cells through activation of IRF3 and 7 (20, 24, 34, 35), our data show that IFNs upregulate Noxa expression by increasing the expression levels of IRF1, 3, and 7 in a p53-independent manner and that IRF4 can be substantially reduced by IFN-γ possibly acting as a counter partner of IRF1, 3, and 7. It is also worth noting that IRF4, an addiction protein of multiple myeloma cells induced by c-myc (36), may allow tumor cells to survive by repressing Noxa expression. In addition to the effects of IFNs on Noxa expression, our data also show that IFNs can upregulate Bax, Bak, Bim, and Bad. Thus, it is possible that the IRF family of proteins may modulate the BH3 profiles in tumor cells to provide either susceptibility or resistance to immune cells. This possibility can be supported by the recent observation that IRF1 induces cell death by upregulating PUMA transcription in gastric colon cancer (37). Taken together, the data suggest that the IRF family of proteins may be major regulators in the process of tumor elimination by determining the susceptibility of tumor cells through altered BH3 profiles in the tumor cells.

No potential conflicts of interest were disclosed.

We thank Dr. Bill Gurley, University of Arkansas for Medical Sciences, for comments and advice in writing the manuscript.

This work was supported by Basic Science Research Program through the National Research Foundation (NRF) of Korea funded by the Ministry of Education, Science and Technology (2009-0072161) and by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, KRF-211-2005-1-E00007).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Burnet
FM
. 
An immunological approach to ageing
.
Lancet
1970
;
2
:
358
60
.
2.
Burnet
FM
. 
The concept of immunological surveillance
.
Prog Exp Tumor Res
1970
;
13
:
1
27
.
3.
Dighe
AS
,
Richards
E
,
Old
LJ
,
Schreiber
RD
. 
Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN gamma receptors
.
Immunity
1994
;
1
:
447
56
.
4.
Shankaran
V
,
Ikeda
H
,
Bruce
AT
,
White
JM
,
Swanson
PE
,
Old
LJ
, et al
IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity
.
Nature
2001
;
410
:
1107
11
.
5.
Kaplan
DH
,
Shankaran
V
,
Dighe
AS
,
Stockert
E
,
Aguet
M
,
Old
LJ
, et al
Demonstration of an interferon gamma-dependent tumor surveillance system in immunocompetent mice
.
Proc Natl Acad Sci U S A
1998
;
95
:
7556
61
.
6.
Smyth
MJ
,
Thia
KY
,
Street
SE
,
MacGregor
D
,
Godfrey
DI
,
Trapani
JA
. 
Perforin-mediated cytotoxicity is critical for surveillance of spontaneous lymphoma
.
J Exp Med
2000
;
192
:
755
60
.
7.
Enzler
T
,
Gillessen
S
,
Manis
JP
,
Ferguson
D
,
Fleming
J
,
Alt
FW
, et al
Deficiencies of GM-CSF and interferon gamma link inflammation and cancer
.
J Exp Med
2003
;
197
:
1213
9
.
8.
Takeda
K
,
Hayakawa
Y
,
Smyth
MJ
,
Kayagaki
N
,
Yamaguchi
N
,
Kakuta
S
, et al
Involvement of tumor necrosis factor-related apoptosis-inducing ligand in surveillance of tumor metastasis by liver natural killer cells
.
Nat Med
2001
;
7
:
94
100
.
9.
Smyth
MJ
,
Thia
KY
,
Street
SE
,
Cretney
E
,
Trapani
JA
,
Taniguchi
M
, et al
Differential tumor surveillance by natural killer (NK) and NKT cells
.
J Exp Med
2000
;
191
:
661
8
.
10.
Street
SE
,
Hayakawa
Y
,
Zhan
Y
,
Lew
AM
,
MacGregor
D
,
Jamieson
AM
, et al
Innate immune surveillance of spontaneous B cell lymphomas by natural killer cells and gammadelta T cells
.
J Exp Med
2004
;
199
:
879
84
.
11.
Street
SE
,
Trapani
JA
,
MacGregor
D
,
Smyth
MJ
. 
Suppression of lymphoma and epithelial malignancies effected by interferon gamma
.
J Exp Med
2002
;
196
:
129
34
.
12.
Fallarino
F
,
Gajewski
TF
. 
Cutting edge: differentiation of antitumor CTL in vivo requires host expression of Stat1
.
J Immunol
1999
;
163
:
4109
13
.
13.
Gao
Y
,
Yang
W
,
Pan
M
,
Scully
E
,
Girardi
M
,
Augenlicht
LH
, et al
Gamma delta T cells provide an early source of interferon gamma in tumor immunity
.
J Exp Med
2003
;
198
:
433
42
.
14.
Caligiuri
MA
. 
Human natural killer cells
.
Blood
2008
;
112
:
461
9
.
15.
Elboim
M
,
Gazit
R
,
Gur
C
,
Ghadially
H
,
Betser-Cohen
G
,
Mandelboim
O
. 
Tumor immunoediting by NKp46
.
J Immunol
2010
;
184
:
5637
44
.
16.
Lou
Y
,
Vitalis
TZ
,
Basha
G
,
Cai
B
,
Chen
SS
,
Choi
KB
, et al
Restoration of the expression of transporters associated with antigen processing in lung carcinoma increases tumor-specific immune responses and survival
.
Cancer Res
2005
;
65
:
7926
33
.
17.
Seliger
B
,
Hohne
A
,
Knuth
A
,
Bernhard
H
,
Ehring
B
,
Tampe
R
, et al
Reduced membrane major histocompatibility complex class I density and stability in a subset of human renal cell carcinomas with low TAP and LMP expression
.
Clin Cancer Res
1996
;
2
:
1427
33
.
18.
Dovhey
SE
,
Ghosh
NS
,
Wright
KL
. 
Loss of interferon-gamma inducibility of TAP1 and LMP2 in a renal cell carcinoma cell line
.
Cancer Res
2000
;
60
:
5789
96
.
19.
Jefferies
WA
,
Kolaitis
G
,
Gabathuler
R
. 
IFN-gamma-induced recognition of the antigen-processing variant CMT.64 by cytolytic T cells can be replaced by sequential addition of beta 2 microglobulin and antigenic peptides
.
J Immunol
1993
;
151
:
2974
85
.
20.
Goubau
D
,
Romieu-Mourez
R
,
Solis
M
,
Hernandez
E
,
Mesplede
T
,
Lin
R
, et al
Transcriptional re-programming of primary macrophages reveals distinct apoptotic and anti-tumoral functions of IRF-3 and IRF-7
.
Eur J Immunol
2009
;
39
:
527
40
.
21.
Kuwana
T
,
Bouchier-Hayes
L
,
Chipuk
JE
,
Bonzon
C
,
Sullivan
BA
,
Green
DR
, et al
BH3 domains of BH3-only proteins differentially regulate Bax-mediated mitochondrial membrane permeabilization both directly and indirectly
.
Mol Cell
2005
;
17
:
525
35
.
22.
Chen
L
,
Willis
SN
,
Wei
A
,
Smith
BJ
,
Fletcher
JI
,
Hinds
MG
, et al
Differential targeting of prosurvival Bcl-2 proteins by their BH3-only ligands allows complementary apoptotic function
.
Mol Cell
2005
;
17
:
393
403
.
23.
Oda
E
,
Ohki
R
,
Murasawa
H
,
Nemoto
J
,
Shibue
T
,
Yamashita
T
, et al
Noxa, a BH3-only member of the Bcl-2 family and candidate mediator of p53-induced apoptosis
.
Science
2000
;
288
:
1053
8
.
24.
Sun
Y
,
Leaman
DW
. 
Involvement of Noxa in cellular apoptotic responses to interferon, double-stranded RNA, and virus infection
.
J Biol Chem
2005
;
280
:
15561
8
.
25.
Porta
C
,
Hadj-Slimane
R
,
Nejmeddine
M
,
Pampin
M
,
Tovey
MG
,
Espert
L
, et al
Interferons alpha and gamma induce p53-dependent and p53-independent apoptosis, respectively
.
Oncogene
2005
;
24
:
605
15
.
26.
Qin
JZ
,
Ziffra
J
,
Stennett
L
,
Bodner
B
,
Bonish
BK
,
Chaturvedi
V
, et al
Proteasome inhibitors trigger NOXA-mediated apoptosis in melanoma and myeloma cells
.
Cancer Res
2005
;
65
:
6282
93
.
27.
Perez-Galan
P
,
Roue
G
,
Villamor
N
,
Montserrat
E
,
Campo
E
,
Colomer
D
. 
The proteasome inhibitor bortezomib induces apoptosis in mantle-cell lymphoma through generation of ROS and Noxa activation independent of p53 status
.
Blood
2006
;
107
:
257
64
.
28.
Tailor
P
,
Tamura
T
,
Ozato
K
. 
IRF family proteins and type I interferon induction in dendritic cells
.
Cell Res
2006
;
16
:
134
40
.
29.
Honda
K
,
Yanai
H
,
Negishi
H
,
Asagiri
M
,
Sato
M
,
Mizutani
T
, et al
IRF-7 is the master regulator of type-I interferon-dependent immune responses
.
Nature
2005
;
434
:
772
7
.
30.
Lehtonen
A
,
Veckman
V
,
Nikula
T
,
Lahesmaa
R
,
Kinnunen
L
,
Matikainen
S
, et al
Differential expression of IFN regulatory factor 4 gene in human monocyte-derived dendritic cells and macrophages
.
J Immunol
2005
;
175
:
6570
9
.
31.
Lu
R
. 
Interferon regulatory factor 4 and 8 in B-cell development
.
Trends Immunol
2008
;
29
:
487
92
.
32.
Freudenberg
MA
,
Tchaptchet
S
,
Keck
S
,
Fejer
G
,
Huber
M
,
Schutze
N
, et al
Lipopolysaccharide sensing an important factor in the innate immune response to Gram-negative bacterial infections: benefits and hazards of LPS hypersensitivity
.
Immunobiology
2008
;
213
:
193
203
.
33.
O'Neill
LA
,
Bowie
AG
. 
The family of five: TIR-domain-containing adaptors in toll-like receptor signalling
.
Nat Rev Immunol
2007
;
7
:
353
64
.
34.
Lallemand
C
,
Blanchard
B
,
Palmieri
M
,
Lebon
P
,
May
E
,
Tovey
MG
. 
Single-stranded RNA viruses inactivate the transcriptional activity of p53 but induce NOXA-dependent apoptosis via post-translational modifications of IRF-1, IRF-3 and CREB
.
Oncogene
2007
;
26
:
328
38
.
35.
Bai
Y
,
Ahmad
U
,
Wang
Y
,
Li
JH
,
Choy
JC
,
Kim
RW
, et al
Interferon-gamma induces X-linked inhibitor of apoptosis-associated factor-1 and Noxa expression and potentiates human vascular smooth muscle cell apoptosis by STAT3 activation
.
J Biol Chem
2008
;
283
:
6832
42
.
36.
Shaffer
AL
,
Emre
NC
,
Lamy
L
,
Ngo
VN
,
Wright
G
,
Xiao
W
, et al
IRF4 addiction in multiple myeloma
.
Nature
2008
;
454
:
226
31
.
37.
Gao
J
,
Senthil
M
,
Ren
B
,
Yan
J
,
Xing
Q
,
Yu
J
, et al
IRF-1 transcriptionally upregulates PUMA, which mediates the mitochondrial apoptotic pathway in IRF-1-induced apoptosis in cancer cells
.
Cell Death Differ
2010
;
17
:
699
709
.