Human GBM cell lines (T98G, U87, LN229, U373 and U251) were obtained from the China Academia Sinica Cell Repository, Shanghai, China. Human cells U373 and U87 were exposed to TMZ (Temodal, Schering-Plough, Whitehouse Station, NJ, USA) until stable TMZ-resistant subclones (U373-R and U87-R) were derived from the parental TMZ-sensitive cell lines. Both cells were exposed to stepwise increasing concentrations of TMZ (2–100 μM) over a period of 6 months.

Antibodies against MGMT, p65, Ki-67, caspase-3 and MMP2/9 were obtained from Cell Signaling Technology (Danvers, MA, USA). SFN (St. Louis, MO, USA) was prepared in DMSO at the stock solution of 100 mM and further diluted to appropriate concentration with cell culture medium immediately before use.

Cell proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride membranes (Millipore). Protein binding to the MGMT, Ki-67, MMP2/9, p65 and caspase-3 antibody (Santa Cruz Biotechnology, Inc.) was assessed by enhanced chemiluminescence and exposed to a chemiluminescent film.

Total RNA was isolated from GBM cells by the TRIzol method. Reverse transcription was performed using a reverse transcription kit (Qiagen), and the primers were synthesized by GenePharma (Shanghai, China) (28).

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma) was used to determine the cell viability and the 50% growth inhibitory concentration (IC₅₀) of TMZ or SFN in both parental and TMZ-resistant cell lines. Cells were cultured in 96-well plate at a density of 5×10³ cells/well in medium containing 10% fetal bovine serum (FBS) and were incubated for 72 h. Subsequently, TMZ was added to culture medium at final concentrations of 0, 500, 1000, 2000, 3000, 4000, 5000 or 6000 μM, SFN was added at a final concentration of 0, 5, 10, 20, 30, 40 or 50 μM. After incubation for 48 h, cells were detached by trypsinization and the viable cell population was determined using MTT. All experiments were done in triplicate.

U87-R, U373-R and T98G cells were seeded into 6-well plates and allowed to attach overnight. Cells were then treated with different concentration of SFN (0, 10 or 25 μM) and colony formation assay was done as described (29). Only colonies with 50 or more cells were counted as surviving colonies under an inverted microscope.

Cells were treated with different concentrations of SFN (10, 20 or 30 μM), and after 8 h TMZ (250 μM) was added. After 48 h, the invasion ability of GBM cells was determined using Transwell plates (BD Biosciences, Bedford, MA, USA) according to the manufacturer's protocol.

Cells cultured in 6-well plates for 24 h were treated with SFN (10, 20 or 30 μM) and TMZ (250 μM, after 8 h) and cultured for additional 48 h. For detection of apoptosis, the FITC Annexin V Apoptosis Detection kit (BD Pharmingen, San Jose, CA, USA) was used according to the manufacturer's protocol, and quantification of apoptosis was determined through measurement of caspase-3/7 activation using caspase-3/7 Detection in Living Cells kit (Biotium Inc., Hayward, CA, USA).

The Luciferase reporter (Stratagene, La Jolla, CA, USA) contains the NF-κB enhancer consensus sequences [(TGGGGACTTTCCGC) ×5] and NF-κB-dependent firefly luciferase gene. Cells were seeded in 24-wells plate, and then transiently transfected with pNF-κB-luc plasmids using Lipofectamine 2000 after 24 h. Luciferase activity was measured with the Dual-Luciferase assay system kit (Promega Corp.).

All animal experiment procedures were carried out according to the regulations and internal biosafety and bioethics guidelines of Tianjin Hospital and Tianjin Huanhu Hospital Medical Ethics Committee. Nude mouse models were established with U373-R GBM cells inoculated subcutaneously into the right flanks of 4–6-week-old female mice. Twenty-four mice were randomized in 4 groups (control, TMZ, SFN alone and the combination of TMZ and SFN) when tumors reached a mass of ~200 mm³ in size. TMZ was given at 50 mg/kg/day for 5 days/week for 4 cycles. SFN was injected subcutaneously at 50 mg/kg/day for 28 consecutive days. Mice were observed daily and euthanized after 5 days of treatment. At the end, all the mice were sacrificed and tumors were collected for further experiments.

The apoptosis in the tumor specimens of mouse models from the in vivo study was examined by TUNEL (Roche, Indianapolis, IN, USA) according to the manufacturer's protocol.

All of the statistical analyses were evaluated by commercially available software SPSS version 13.0. Differences between the groups were analyzed by using Student's t-test and statistical significance was determined as P<0.05 or P<0.01.

To explore the different conditions of MGMT in various types of GBM cells, we firstly detected expression of MGMT mRNA and protein in U251, T98G, U87, LN229 and U373 cells. The expression of MGMT mRNA was high in T98G cells, and very low or hardly detected in other GBM cells (Fig. 1A). Furthermore, the detection of MGMT protein was consistent with mRNA status and much higher in T98G cells (Fig. 1B). As reported, MGMT overexpression was able to resist cell death induced by TMZ. Cellular sensitivity to TMZ in five cell lines was also evaluated after 48 h of treatment. IC₅₀ values varied between the different cell lines, and the IC₅₀ of T98G was much high than other cell lines (Table I). Based on the above data, T98G cell line was considered as resistant to TMZ and chosen to carry out the following experiments.

We then successfully established TMZ-resistant GBM cell models to search the methods of TMZ sensitization. Two TMZ-resistant human glioma cell sublines, U373-R and U87-R, were generated by a stepwise exposure to increasing TMZ concentrations for 6 months. The two established variants showed effective resistance to further TMZ treatment. The IC₅₀ of U373-R and U87-R showed more than 5-fold increase when compared with their parental cell lines (2529.3 vs. 483.5 μM and 3657.2 vs. 702.4 μM) (Fig. 2A). Previous it was reported that most DNA-damaging chemotherapeutic agents including TMZ could activate the transcription of NF-κB (30). We then detected the activity of NF-κB in TMZ-inducing cells by transiently transfecting with pNF-κB-luc reporter plasmids. When compared with control counterparts, NF-κB transcription was significant activated in U373-R and U87-R cells (Fig. 2B). At the end of this section, we investigated whether the increase in TMZ resistance in GBM cells was associated with high MGMT expression. PCR and western blot analyses revealed that MGMT content increased significantly at both mRNA and protein levels in the TMZ-resistant cells (Fig. 2C and D). Taken together, these results suggested a link between the onset of TMZ resistance and activity of NF-κB signaling pathway.

SFN has been indicated for the prevention and suppression of tumorigenesis in various solid tumors. Our previous study demonstrated that SFN could effectively enhance TMZ-induced apoptosis by inhibiting miR-21 via Wnt/β-catenin signaling in GBM cells (31). Herein, we aimed to examine the inhibition of growth of SFN in TMZ-resistant GBM cells. The role of SFN in TMZ-resistant cell proliferation was investigated by conducting MTT and colony formation assays. We treated U373-R, U87-R and T98G cells with 5–50 μM for 48 h and assessed the number of viable cells. The MTT assay showed that SFN had a growth inhibitory effect on all the tumor cell lines in a dose-dependent manner (Fig. 3A). Next, colony formation assay suggested that cells treated with SFN (10 and 25 μM) for 48 h formed significantly less colonies than control cells (P<0.01) (Fig. 3B), suggesting an inhibitory effect of SFN on anchorage-dependent growth of TMZ-resistant cells.

To choose the correct dose for the following in vitro combination experiments, the IC₅₀ value of SFN was investigated in these three cell lines. The IC₅₀ of SFN for U373-R, U87-R and T98G was 41.7, 44.8 and 41.5 μM, respectively (Fig. 4). Based on the above results, concentrations lower than the respective IC₅₀ value of SFN was identified for further study in TMZ-resistant cells. To determine whether SFN as a chemosensitizer in TMZ-resistant GBM was mediated by decreasing the expression of MGMT through inhibiting NF-κB, we examined the effects of SFN on NF-κB transcription activity. The U373-R, U87-R and T98G cells were transfected with the NF-κB reporter plasmids. A range of concentrations of SFN was added to the transfected cells, as observed in Fig 5A, the transcriptional activity of NF-κB was significantly attenuated by SFN in a concentration-dependent manner. A previous study provided evidence that MGMT is a target gene for NF-κB (16). We then analyzed the expression of MGMT following SFN treatment in TMZ-resistant GBM cells. The PCR and western blot analyses revealed that SFN resulted in decreased expression of MGMT in a dose-dependent manner (Fig. 5B).

In order to examine whether SFN can enhance the cytotoxity of TMZ, a series of experiments were conducted, and the dose of 250 μM TMZ was selected, which has been evidenced as no growth inhibitory effects on TMZ-resistant cells. Initially, nonlinear regression of a sigmoid dose response model and combination index (CI) approaches were performed to test drug interaction between SFN and TMZ. The results revealed antagonistic effects (CI>1) when cells were simultaneously treated with TMZ and each concentration of SFN for T98G, U87-R and U373-R. Nevertheless, the response to concomitant treatment was drastically reversed when the administration order was changed. Synergistic effects (CI<1) were observed when different concentrations of SFN were used as pretreatment in TMZ-resistant cells for 8 h before exposure to TMZ. Sequential exposure also resulted in high- dose reduction index (DRI) values suggesting that TMZ doses could be signicantly reduced to achieve comparable cytotoxicity (Table II). The three cell lines were then treated in schedule dependency with the previous (8 h) exposure to SFN. To identify the synergistic effects of SFN and TMZ on the invasive ability, Transwell Matrigel invasion assay was used in T98G, U87-R and U373-R cell lines. The results suggested that TMZ alone could not inhibit cell invasion, but cells pretreated with different concentrations of SFN combined with TMZ significantly reduced GBM cell invasive capacity (Fig. 6A). We further investigated the effect of SFN on TMZ-inducing apoptosis, TMZ alone did not demonstrate apoptosis induction in the cell lines compared with control, however, the sequential therapy of the SFN and TMZ caused a significant increase of apoptotic death even at lower doses of SFN, suggesting that a synergistic induction of apoptosis developed in the cells co-treated with SFN and TMZ (Fig. 6B). Moreover, caspase-3/7 activity was also considerably elevated after synergism with SFN in TMZ-resistant cells (Fig. 6C). Collectively, these observations provide strong evidence for the role of pretreatment of SFN in reversing TMZ-resistant GBM cells to TMZ.

To facilitate the in vitro experiments, stable chemo-resistant cell line U373-R was used to establish subcutaneous tumors. We subsequently investigated the synergistic effects of a combination of SFN and TMZ in glioma xenograft nude mouse models. In xenograft growth assay, TMZ alone did not effectively suppress growth of the U373-R tumor (P>0.05). However, SFN alone suppressed tumor growth more significantly than control groups from day 28 (P<0.01), and the combination of SFN and TMZ decreased tumor growth to a more significant extent, moreover, the differences in tumor mass of the combined groups were still marked compared with control and TMZ alone group at the termination of the study (P<0.01) (Fig. 7A). Therefore, the SFN plus TMZ treatment inhibited tumor growth significantly in the TMZ-resistant glioma xenografts compared with the TMZ alone. To extend these observations, western blot analysis and TUNEL analysis were used to examine solid tumors after 40-day observation. As demonstrated in Fig. 7B, the expression of NF-κB p65 and MGMT protein were significant decreased in combined groups. The protein levels of Ki-67, MMP2/9 and caspase-3 related with the cell proliferation, invasion, and apoptosis in combined group were prominently changed which is similar to results obtained from the in vitro study. The TUNEL assay analysis verified the synergistic pro-apoptotic effects of the combination of SFN and TMZ in vivo (Fig. 7C). In conclusion, these data demonstrated that SFN could effectively inhibit TMZ-mediated cell growth and enhance TMZ-mediated cell death in chemo-resistant xenograft nude mouse models.