IFN-β Down-Regulates the Expression of DNA Repair Gene MGMT and Sensitizes Resistant Glioma Cells to Temozolomide

Gliomas are the most common primary tumors of the central nervous system; they account for 30% of adult primary brain tumors. The prognosis for patients with the advanced glioma, glioblastoma multiforme, is very poor; the mean survival period is 8 to 10 months (1). Alkylating agents are among the most effective cytotoxic agents used for treating malignant gliomas, including glioblastoma multiforme, but responses remain very poor. The most frequent site of alkylation in DNA is the O⁶ position of guanine, which forms cross-links between adjacent strands of DNA, leading to cell death. A cellular DNA-repair protein, namely O⁶-methylguanine-DNA methyltransferase (MGMT) protein, reverses alkylation at the O⁶ position of guanine, thereby inhibiting the lethal cross-linking and bringing about resistance to alkylating agents (2, 3). A number of studies have suggested that MGMT deficiency is closely related to the sensitivity of brain tumors to alkylating agents (4–6). Furthermore, because MGMT protein loss may be a result of promoter hypermethylation, it was reported that methylation of the MGMT promoter in gliomas is a useful predictor of the responsiveness to alkylating agents (7). Temozolomide is a novel alkylating agent that has been currently approved for use in treatment of anaplastic astrocytoma in Europe and the United States. A phase II clinical trial has been organized in Japan. This drug is of significance because it can be administered orally; it readily crosses the blood-brain barrier and has minimal side effects (8). Before considering the treatment of malignant gliomas with temozolomide, a major obstacle may be the resistance pathway that occurs due to the actions of MGMT. Hence, efficient attenuation of the function of MGMT, which is expressed in ∼70% of gliomas (9), either by direct interaction with protein or by indirect means such as the transcriptional control, is required. It could be advantageous if MGMT depletion can be accomplished by a drug that also has antitumor activity, and, therefore, synergistic effects with the alkylating agent may occur. Type I IFNs, including IFN-α and IFN-β, a family of cytokines that elicit pleiotropic biological effects, are widely used either alone or in combination with other antitumor agents such as nitrosoureas in the treatment of malignant gliomas. Among the multiple functions of type I IFNs against human neoplasias, type I IFNs, particularly IFN-β, can act as a drug sensitizer enhancing toxicity against a variety of neoplasias when given in combination with nitrosoureas (10). Therefore, it is of interest to examine whether IFN-β can enhance chemosensitivity of malignant gliomas against temozolomide, the new alkylkating agent; to evaluate the mechanism; and to provide an experimental basis for the rational clinical use of such combinations.

Cell lines and reagents. Human glioma cell lines (T98, AO2, SKMG1, U251nu/nu, U251SP, and U251MG) derived from the Memorial Sloan-Kettering Cancer Institute (New York, NY) were maintained at 37°C in a humidified atmosphere of 5% CO₂ in air in Eagle's medium (Nissui, Tokyo, Japan). Medium was supplemented with 10% fetal bovine serum (FBS), 5 mmol/L l-glutamine, 2 mmol/L nonessential amino acids, and antibiotics (100 units/mL penicillin and 100 μg/mL streptomycin). Human IFN-β and temozolomide were kindly supplied by Toray, Co., Ltd. (Kamakura, Japan), and the Schering-Plough Research Institute (Kenilworth, NJ), respectively. Temozolomide readily discomposes in aqueous solution into DMSO.

RNA interference experiments. The target sequence in p53 and the control nonsilencing sequence are CGGCAUGAACCGGAGGCCCAU and AATTCTCCGAACGTGTCACGT, respectively. These synthetic sense and antisense oligonucleotides were obtained from Qiagen (Hilden, Germany). For the annealing of small interfering RNA (siRNA) oligonucleotides, sense and antisense oligonucleotides were incubated in siRNA Suspension Buffer (Qiagen) for 1 minute at 90°C, followed by 60 minutes at 37°C. siRNA oligonucleotides were mixed with Oligofectamine reagent (Invitrogen, Carlsbad, CA) in Opti-MEM (Life Technologies, Gaithersburg, MD). Cultured cells were washed with medium without serum and added to the siRNA-Oligofectamine mixture, of which the final concentration was 200 nmol/L. Medium with 10% FBS was added 4 hours later.

Determination of cell growth. To compare chemosensitivity of glioma cell lines to temozolomide, the agent (final concentration of 0-1,000 μmol/L) was added to the culture medium at 24 hours after aliquots of 2 × 10⁴ cells/well were placed in triplicate wells. Incubation was continued for 72 hours, and the number of viable cells was counted by trypan blue exclusion method. The number was expressed as a percentage of untreated control. The statistical significance of difference was determined by ANOVA using Bonferroni's correction for the multiple comparisons used. For sensitizing assay of IFN-β to temozolomide, IFN-β (100 IU/mL) and temozolomide (100 μmol/L) were added at 24 and 48 hours after cell inoculation, respectively. The number of viable cells was counted at 96 hours. To examine the effect of siRNA for p53 on the growth of T98 cells treated with IFN-β and temozolomide, either siRNA for p53 or nonsilencing siRNA was transfected as described above at 24 hours after aliquots of 4 × 10⁴ cells/well were placed in triplicate wells. IFN-β (100 IU/mL) and temozolomide (100 μmol/L) were added at 28 and 48 hours, respectively. The number of viable cells was counted at 96 hours.

Western blot analysis. Cell lysis and immunoblotting were carried out as described (11). Antibodies against the following proteins were purchased: p53 (DO-1: Santa Cruz Biotechnology, Santa Cruz, CA), p21 (EA10: Oncogene Research Products, San Diego CA), MGMT (MT3.1: Neomarkers, Fremont, CA), and β-actin (AC-15: Sigma-Aldrich, St. Louis, MO). Band intensities were quantified by densitometric scanning using the NIH IMAGE program.

Genomic DNA extraction and methylation-specific PCR. Genomic DNA was extracted using the QIAmp DNA Mini kit (Qiagen) following the manufacturer's instructions. DNA methylation patterns in the promoter region of the MGMT gene (Genbank accession no. X_61657) were determined by methylation-specific PCR as previously described (12). Primers for either methylated or unmethylated alleles were 5′-TTTCGACGTTCGTAGGTTTTCGC-3′ (sense) and 5′-GCACTCTTCCGAAAACGAAACG-3′ (antisense) or 5′-TTTGTGTTTTGATGTTTGTAGGTTTTTGT-3′ (sense) and 5′-AACTCCACACACTCTTCCAAAAACAAAACA-3′ (antisense). All PCRs were done with positive controls for methylated alleles and no DNA control. Human placental DNA was treated in vitro with excess SssI methyltransferase (New England Biolabs, Beverly, MA), generating DNA completely methylated at CpG sites, served as the positive control for methylated MGMT, and then the PCR conditions were determined. Each PCR product was loaded onto a 3% agarose gel, stained with ethidium bromide, and visualized under UV illumination.

Reverse transcriptase-PCR. To investigate MGMT mRNA expression, reverse transcription-PCR (RT-PCR) was done using the Superscript first-strand synthesis system for RT-PCR (Invitrogen) as previously described (13). β-actin–specific PCR products from the same RNA samples were amplified and served as internal controls.

Microarray analysis. T98 cells were incubated with 100 IU/mL of IFN-β for 48 hours, and RNA was isolated. Standard Trizol preparation protocol (Invitrogen) and reagents were used for total RNA isolation. RNA amplification and labeling were done using the Amino Allyl MessageAmp aRNA kit (Ambion, Austin, TX). Briefly, after reverse transcription reactions (2 μg total RNA/sample) were preformed, double-stranded cDNA was transcribed in vitro into amino allyl cRNA. The purified and concentrated cRNA (5 μg) was coupled with either Cy3 or Cy5 dyes (Amersham Biosciences). The dye-labeled aRNA was purified from uncoupled dye using Micro Bio-Spin P-30 Tris chromatography columns (Bio-Rad, Hercules, CA) and MicroconYM-30 centrifugal filter devices (Millipore, Billerica, MA). The cRNA was fragmented in fragmentation buffer [40 mmol/L Tris-acetate (pH 8.1), 100 mmol/L potassium acetate, 30 mmol/L magnesium acetate] at 94°C for 15 minutes and purified with Microcon YM-10 (Millipore). Microarrays were preblocked with 1% bovine serum albumin solution. Fragmented cRNA was added to microarrays in hybridization solution and hybridized at 42°C for 16 hours. After this, arrays were washed, scanned at 10 μm pixel size, gridded, and analyzed (GenePix 4000B; Axon Instruments, Union City, CA). Background was subtracted, and the median sum and median ration were calculated. Flagged spots and spots with sum intensity (CH1, CH2) <100 absorbance units were excluded. Data were normalized by trimmed mean at 10% to account for differences in the amounts of RNA labeled or labeling efficiencies.

Chromatin immunoprecipitation assay. The chromatin immunoprecipitation assay was done according to the protocol of the manufacturer (Upstate Biotechnology, Lake Placid, NY) with a slight modification. The specific antibodies used for immunoprecipitations were the anti-p53 antibody and a control antibody. After protein-DNA cross-links in the immunoprecipitates were reversed, the purified DNA was analyzed by PCR (35 cycles; 45 seconds at 95°C, 45 seconds at 55°C, 60 seconds at 72°C) with primers that detect the MGMT promoter sequence [5′-GCTCCAGGGAAGAGTGTCCTCTGCTCCCT-3′ (sense) and 5′-GGCCTGTGGTGGGCGATGCCGTCCAG-3′ (antisense)] and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) promoter (provided with the kit). The PCR products were visualized on an ethidium bromide gel.

Chemosensitivity to temozolomide, MGMT expression, and MGMT promoter hypermethylation of human glioma cell lines. First, we compared chemosensitivity of six human glioma cell lines to temozolomide; T98 and U251nu/nu cell lines were significantly resistant, whereas the other four (AO2, SKMG1, U251SP, and U251MG) showed similar degrees of sensitivity in a dose-dependent manner (Fig. 1A). MGMT expression was not detected by Western blot in the sensitive cell lines (Fig. 1B), and the hypermethylation of MGMT promoter was as assessed by methylation-specific PCR (Fig. 1C), but this was not observed in the T98 and U251nu/nu cell lines. Thus, the results showed that hypermethylation of MGMT promoter could prevent expression of this gene and be associated with chemosensitivity of glioma cells to temozolomide, which was consistent with the results of reported studies (7). Because ∼70% of gliomas express MGMT (9), it is essential to suppress the expression or function of MGMT in an efficient manner if treatment with the aid of temozolomide is to be considered.

IFN-β sensitizes resistant glioma cells to temozolomide and down-regulates MGMT expression. IFNs are a family of cytokines that possess pleiotropic biological effects mediated by a number of responsive genes. IFNs were the first human proteins to be recognized as being effective in cancer therapy (14). Although identified and named for their action to interfere with viral replication, IFNs have immunomodulatory, cell differentiative, antiangiogenic, and antiproliferative effects (15, 16). In the previous studies, we showed that IFN-β has multiple functions relevant to antitumor activity: (a) cytostatic effect on glioma cells, (b) supportive action on the differentiation of CTLs and augmentation of their antitumor immune responses, and (c) behavior as a drug or radiation sensitizer enhancing toxicity against gliomas (10).

To examine whether IFN-β sensitizes T98 and U251nu/nu cell lines that are resistant to temozolomide, the cells were incubated in a culture medium containing IFN-β (100 IU/mL) for 24 hours before the addition of temozolomide (final concentration of 100 μmol/L). As shown in Fig. 2A, whereas IFN-β or temozolomide alone did not suppress cell growth of both T98 and U251nu/nu cells significantly, a combination of IFN-β and temozolomide markedly inhibits the growth. On the other hand, U251SP cell lines that are sensitive to temozolomide, but are resistant to IFN-β did not show the synergistic cytotoxic effect of temozolomide and IFN-β. Consequently, we hypothesized that IFN-β might enhance chemosensitivity due to reduction of MGMT expression and did Western blotting and RT-PCR for MGMT (Fig. 2B and C). It was shown that IFN-β decreased both MGMT protein and mRNA levels of both T98 and U251nu/nu cells at 48 hours after treatment. Thus, IFN-β decreases MGMT transcription directly or indirectly and sensitizes resistant glioma cells to temozolomide.

Microarray analysis of IFN-β–regulated genes. Because limited information regarding the transcriptional regulation of IFN-β is available, we introduced microarray technology that enables the simultaneous examination of expression of a large number of genes in an experimental condition. This microarray contains 1,300 functionally well-characterized genes involved in various important cellular processes, including the IFN-related pathway, apoptosis, cell cycle, transcription, and immunology. The microarray experiments were repeated thrice about the T98 cell lines that showed the most drastic change of MGMT expression after IFN-β treatment. Data analysis identified 71 significantly induced genes and 54 repressed genes in several different categories in T98 cells treated with IFN-β compared with the parental cells. The representative results of these comparisons are reported in Table 1. Some of the alterations in gene expression following IFN-β treatment confirm the involvement of pathways known to be active in this process, and this was anticipated from prior evidence. Consistent with a report that the p53 gene is transcriptionally induced by IFN-β through ISGF3 activation (11), our microarray data showed that the p53 gene and its downstream genes, including bcl-2, FAS/Apo1, and Gadd45, are up-regulated in glioma cells treated with IFN-β, suggesting that p53 induction may be a key involved in antitumor action of IFN-β. Thus, we focused on IFN-β–mediated p53 induction. By Western blot, we confirmed that p53 expression level in both T98 cells treated with IFN-β was markedly higher than that in the untreated cells and the similar phenomenon was observed in the U251nu/nu cells as well (Fig. 3A).

IFN-β down-regulates MGMT expression via protein p53. For >10 years, p53 has been the focus of intensive research. This has led to a plethora of information regarding p53, its biological roles, and its relevance to cancer (17). p53 is primarily a sequence-specific transcriptional activator. It binds to cognate p53-responsive elements within the genome and activates the transcription of genes residing in the vicinity of these binding sites. The proteins encoded by the p53 target genes, whose number is probably in the hundreds, contribute in multiple ways to the biological effects of p53. The biological outcomes of p53 activity include apoptosis, inhibition of cell cycle progression, senescence, differentiation, and accelerated DNA repair. We first confirmed that siRNA for p53 specifically knocked down the protein p53 function by examining the diminished expression level of p21, a well-known target gene of p53 (Fig. 3B). As shown in Fig. 3B, knockdown of p53 by siRNA increased MGMT expression in T98 cells. Therefore, p53 can down-regulate transcription of MGMT. To further investigate whether protein p53 directly interacts with MGMT promoter and down-regulates transcription of MGMT, we carried out chromatin immunoprecipitation assay. We observed that MGMT promoter, but not GAPDH promoter, coprecipitated with p53. The irrelevant antibody specific to TFIIB, which is a transcription factor of GAPDH, did not coprecipitate with MGMT promoter (Fig. 3C). These findings are in agreement with the previous reports showing that p53 reduces the basal activity of the MGMT promoter, and adenoviral vector–mediated overexpression of p53 reduces MGMT expression (18, 19), although some papers suggested that inactivated MGMT may be linked to cytotoxic effect of alkylating agents and cell signaling events, but was independent of p53 status (20). The reasons for these differences are unclear and may be cell type dependent. Moreover, in cell growth experiments, knocking down of p53 before IFN-β treatment nullified the synergistic inhibitory effect of IFN-β and temozolomide on T98 cell growth (Fig. 3D). Thus, acting together, IFN-β down-regulates MGMT transcription via induction of the p53 expression and sensitizes resistant glioma cells to temozolomide. In conclusion, this report shows that IFN-β is able to decrease MGMT levels in glioma cells via the inhibition of MGMT gene transcription. Moreover, pretreatment of glioma cells with IFN-β markedly enhances chemosensitivity to temozolomide. This suggests that clinical efficacy of temozolomide might be improved by combination with IFN-β using appropriate doses and schedules of administration.

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.

The authors especially thank NGK Insulators, Ltd for making the custom made microarray, and thank Ms. Yasuko Yoshida and Mr. Saichi Yamada (NGK Insulators, Ltd) for technical assistance.