Glioblastoma (WHO grade IV) is a devastating disease with a median survival of less than 1 yr after diagnosis (Scott et al. 1998). Glioblastomas are defined pathologically by histopathologic features of cellular atypia, mitotic figures, necrotic foci with peripheral cellular pseudopalisading, and microvascular hyperplasia (Louis et al. 2007). Glioblastomas consist of heterogeneous populations of cancer cells. The tumor is regarded as an aberrant organ containing a minor fraction of cancer stem cells and a major fraction of non-stem cells (Singh et al. 2004). Cancer stem cells represent a population of drug-resistant cells that can survive treatment and repopulate the tumor. “Side population” (SP) in the tumor is enriched in cancer stem cells (Zhang and Rosen 2006). SP cells express ATP-binding cassette (ABC) transporters, especially ABCG2, pumping out the fluorescent nuclear dye Hoechst 33342 and many other small-molecule substances. Therefore, they were not stained by Hoechst 33342 and could be isolated by fluorescent-activated cell sorting (FACS; Zhou et al. 2001).

SP cells are resistant to chemotherapy (Hambardzumyan et al. 2008). Basically, the chemoresistance is due to ABCG2-mediated efflux of small-molecule drugs such as topotecan (Smith et al. 2009) and mitoxantrone (Bleau et al. 2009). Temozolomide, which is the standard chemotherapy for glioblastoma, is not a substrate for ABCG2. However, SP cells are also resistant to temozolomide because SP cells express higher level of O6-methylguanine DNA methyltransferase (MGMT) than non-SP (Bleau et al. 2009), and increased expression and elevated activity of MGMT has been shown to be a key mechanism involved in the resistance of gliomas to temozolomide (Kitange et al. 2009).

Gene therapy holds a great promise for malignant tumors. One of the best-characterized strategies is the herpes simplex virus thymidine kinase/ganciclovir (HSV-TK/GCV) suicide gene therapy system. HSV-TK/GCV has succeeded in vitro with a number of tumor cell lines and has produced strong tumor growth delay and some complete tumor regressions in animal models. The HSV-TK gene which is transferred into the tumor cells can sensitize the cells to the normally non-toxic antiviral drug, ganciclovir (Moolten 1986). This effect is based on the selective conversion of GCV to its monophosphate by HSV-tk and further phosphorylation by catalyzed cellular kinases to its cytotoxic triphosphate derivative which competes with endogenous dGTP for incorporation into DNA, interfering with cellular DNA synthesis, thereby inducing apoptosis (Balzarini et al. 1993; Ilsley et al. 1995).

Many studies have been done trying to treat glioblastoma with HSV-TK/GCV gene therapy (Rainov et al. 2000, 2001; Moriuchi et al. 2005). Now that HSV-TK/GCV system involves a small-molecule drug GCV, we wonder if glioblastoma SP cells are able to “pump out” GCV and thus are resistant to this suicide gene therapy. In this study, we compare the sensitivity of SP and non-SP glioblastoma cells to HSV-TK/GCV and then try to make the mechanism clear.

Primary glioblastoma sample CLB05 was obtained from the pathology specimen of a glioblastoma multiforme patient undergoing resection in accordance with a protocol approved by the Zhejiang University School of Medicine Institutional Review Board and with prior informed consent from the patient. The CLB05 primary culture was at early passage (less than 15 passages) in this study. U251 glioblastoma cell line and HEK 293 packaging cell line were provided by the Shanghai Institute of Biochemistry and Cell Biology, Shanghai, China. Cells were cultured in serum-containing medium at a density of 1 × 105 cells per milliliter. The serum-containing medium was composed of Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Grand Island, NY) and 10% fetal bovine serum (FBS; Gibco). Cells were incubated at 37°C with 95% air, 5% CO2, and 100% humidity.

We used the BD Adeno-X™ Expression System (Clontech, Mountain View, CA) to produce recombinant adenovirus which carried HSV-TK gene. Our lab keeps the HSV-TK gene which is cloned into the pUCm-T vector (Sangon, Shanghai, China) in reverse direction. HSV-TK gene was then cloned into pShuttle2 using XbaI and NotI restriction sites. The expression cassette was excised from recombinant pShuttle2-HSV-TK plasmid DNA by digesting with I-CeuI and PI-SceI restriction enzymes. The expression cassette was ligated to BD Adeno-X Viral DNA which was then digested with SwaI and transferred into Escherichia coli cells. The ampicillin-resistant transformants were selected and expanded. Recombinant adenoviral DNA was isolated from the E. coli cells and confirmed by restriction analysis and PCR analysis according to the manual. The recombinant adenoviral plasmid was digested with PacI to expose the inverted terminal repeats located at either end of the genome. HEK 293 cells were transfected with PacI-digested adenoviral DNA using Lipofectamine 2000 Transfection Reagent (Invitrogen, Carlsbad, CA). Recombinant adenoviral particles were harvested after cytopathic effect appeared and purified with Improved Adeno-XTM Virus Purification Kits (Clontech). Adenoviral titer was determined using the protocols in the manual. Afterwards, U251 and CLB cells were infected with recombinant adenovirus to express HSV-tk. RT-PCR was performed to confirm the gene expression. The primers for HSV-tk were 5′-TGACTTACTGGCGGGTGCTG (forward) and 5′-CCATTGTTATCTGGGCGCTTG (reverse). The primers for housekeeping gene GAPDH were 5′-GCACCGTCAAGGCTGAGAAC (forward) and 5′-ATGGTGGTGAAGACGCCAGT (reverse). The amplification conditions were: one cycle of 5 min at 95°C, followed by 35 cycles of 15 s at 95°C, 30 s at 60°C, and 30 s at 72°C, and one cycle of 5 min at 72°C.

In order to do cell flow, U251 and CLB05 cells were dissociated by 0.25% trypsin to prepare cell suspension. The cell suspension was filtered through a cell strainer to get rid of cell clusters. Count the total cell number. The single-cell suspension was centrifuged at 600 rpm for 1–2 min at 4°C. Pipette off the supernatant. The cell pallet was resuspended with 5 ml DMEM + 5%FBS. Hoechst33342 (Sigma, St. Louis, MO) was added to cells at a final concentration of 5 µg/ml. If we needed to block the ABC transporter, verapamil (100 µM, Sigma) was added 15 min prior to Hoechst33342. Mix well and incubate for 90 min at 37°C. After staining, the cells were washed twice in Hanks’ balanced salt solution with 2% FBS. Cells were collected by centrifugation and resuspended with 1–2 ml phosphate-buffered saline. Then, the samples were ready for flow cytometry or FACS. Cells were analyzed on a MoFlo high-speed cell sorter (DakoCytomation, Fort Collins, CO) using 350-nm excitation. In cells showing reduced fluorescence of both blue (670 nm) and red (450 nm), the “side population” was counted or collected.

Western blot was performed to detect HSV-tk protein expression in SP, non-SP, and unsorted cells. Proteins were extracted with RIPA buffer (Sigma) plus Protease Inhibitor Cocktail (Sigma). After quantification, proteins were loaded on the 4–12% Pre-Cast NuPAGE Novex Bis-Tris Gel (Invitrogen). Run the gel at 80–90 V for approximately 3 h. Then, the proteins were transferred to the polyvinylidene fluoride membrane. The membrane was blocked in a solution of 5% non-fat milk in TBST for 1 h at room temperature and incubated with primary antibodies 1:100 HSV-TK (Santa Cruz Biotechnology, Santa Cruz, CA) or 1:500 Beta-Actin (Santa Cruz Biotechnology) overnight at 4°C on shaker. After thorough washing, the membrane was incubated with secondary antibodies 1:2,000 donkey anti-goat IgG-HRP (for HSV-TK, Santa Cruz Biotechnology) or donkey anti-mouse IgG-HRP (for beta-actin, Santa Cruz Biotechnology) for 1 h at room temperature on shaker. After washing, the protein bands were visualized with enhanced chemiluminescence and exposed to X-ray film for 15 s–30 min.

For the purpose of comparing the sensitivity to HSV-tk gene therapy of SP and non-SP, we have five groups: (1) Control: unsorted U251 or CLB05 cells with HSV-tk expression but without ganciclovir treatment; (2) NSP-GCV: sorted U251 or CLB05 non-SP cells with HSV-tk expression, with ganciclovir treatment (10 μg/ml, Sigma; Ahn et al. 2009); (3) SP-GCV: sorted U251 or CLB05 SP cells with HSV-tk expression, with ganciclovir treatment; (4) SP-GCV-gefitinib: this group is the SP-GCV group plus gefitinib treatment (5 μmol/L, AstraZeneca, Macclesfield, UK), which is an epidermal growth factor receptor tyrosine kinase inhibitor but also an ABCG2 inhibitor and able to reverse ABCG2-mediated drug resistance (Nakamura et al. 2005); (5) SP-GCV-Fumitremorgin C: this group is the SP-GCV group plus Fumitremorgin C treatment (5 μM. Sigma), which is also a specific ABCG2 inhibitor (Rabindran et al. 2000). For this part, cells were seeded in triplicates in 96-well plates at 1,000 cells/100 μl per well. Fresh media and drug were changed everyday. (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt (MTS) assay was performed at 5–6 d when the control group cells reached almost 100% confluence. Add 100 µl phenazine methosulfate (PMS), 0.92 mg/ml (Promega, Madison, WI) to 2 ml MTS (2 mg/ml, Promega). Mix well by pipetting. Add 20 μl MTS/PMS mix reagent to 100 μl media per well (96-flat bottom plate). Incubate at 37°C for 1–2 h depending on the confluence. Then measure the optical density (O.D.) at 490 nm and reference at 650 nm. Do three flashes per plate.

In order to determine if GCV is a substrate for ABCG2, we performed a competition assay. As mentioned above, ABCG2 is mainly responsible for the side population phenotype (Zhou et al. 2001). This SP profile disappears in the presence of verapamil or reserpine which blocks ABCG2. Thus, we deduce that if GCV is the substrate for ABCG2, it may compete with Hoechst33342 and result in the reduction of side population. So we detected the side population of CLB05 cells (without HSV-TK transfection) with or without co-incubation with overdose ganciclovir (100 μg/ml).

Data were expressed as the mean ± standard deviation. Student’s t tests were used to compare the mean. A values of P <0.05 was accepted as statistically significant.

U251 and CLB05 cells were infected with 1 × 107 pfu/ml recombinant adenovirus for 6 h. RNA was extracted 24 h later, and RT-PCR indicated that both U251 and CLB05 cells expressed HSV-tk (Fig. 1 A).

Figure 1.
figure 1

Preparing SP and non-SP glioma cells with HSV-tk expression. A RT-PCR. Twenty-four hours after recombinant adenovirus-mediated gene transfection, both U251 and CLB05 (primary glioblastoma culture) cells expressed HSV-TK gene. B, C Sorting SP and non-SP cells from HSV-TK (+) U251 (B) and CLB05 (C). D Sorted SP and non-SP cells expressed HSV-tk protein at similar level.

We sorted SP and non-SP cells from U251 (Fig. 1 B) and CLB05 (Fig. 1 C) cells that had been confirmed to be HSV-tk (+). Proteins were extracted from SP, non-SP, and also unsorted cells. Western blot revealed that SP and non-SP glioblastoma cells expressed HSV-tk at similar level (Fig. 1 D), which excluded the possibility that different HSV-tk protein expression levels may affect the sensitivity of SP and non-SP. We noted that the unsorted cells seemed to show a little bit more expression, implying that the cell sorting process may affect the cell viability slightly.

By MTS assay, we found that HSV-TK/GCV gene therapy killed both SP and non-SP glioma cells to some extent. However, SP cells were much more resistant than non-SP (Fig. 2). This drug resistance can be reversed by both gefitinib and Fumitremorgin C (Fig. 2), implying that ABCG2 may play an important role in the resistance to HSV-TK/GCV of SP cells.

Figure 2.
figure 2

MTS assay revealed that SP are more resistant to HSV-TK/GCV gene therapy (optical density (O.D.) U251 SP 0.883 ± 0.09 vs. NSP 0.45 ± 0.04, P < 0.01. CLB05 SP 0.904 ± 0.08 vs. NSP 0.43 ± 0.04, P < 0.01), and this resistance effect can be reversed by gefitinib and Fumitremorgin C.

The competition assay supported our hypothesis that ganciclovir is a substrate for ABCG2. Flow cytometry results revealed that SP of CLB05 (Fig. 3 A) can be completely removed by verapamil (Fig. 3 B). As expected, overdose of ganciclovir can reduce the side population to approximately one third (Fig. 3 C), indicating that ganciclovir was pumped by ABCG2 and competing with Hoechst33342.

Figure 3.
figure 3

Competition assay. A, SP of CLB05 glioblastoma cells. B, SP was completely removed by verapamil. C, SP was reduced by overdose GCV treatment.

Our in vitro study suggests that SP glioma cells are more resistant to HSV-TK/GCV suicide gene therapy, which is not due to different HSV-tk expression level but to the ABCG2 efflux capacity.

In this study, we designed and successfully applied the competition assay to determine the interaction of ABCG2 and small-molecule drug GCV. However, we are unable to detect if phosphorylated GCV is a substrate for ABCG2 because phosphorylated GCV are very toxic to cells.

A unique advantage of the HSV-TK/GCV system is the so-called bystander effect which is due to the transfer of the toxic GCV metabolites from HSV-tk-transduced cells to non-transduced cells through gap junctional intercellular communication (Fick et al. 1995; Mesnil and Yamasaki 2000). The bystander effect can enable to complete tumor eradication in animal models when only 10–20% of the tumor cells carry the HSV-tk gene (Culver et al. 1992). In our study, the bystander effect has not been evaluated. However, considering that SP cells are few and diffusedly distributed among non-SP cells, they may be easily killed in vivo through bystander effect.

Our data also implied that the ABCG2 inhibitors, especially gefitinib, may have a good promise in improving the effect of HSV-TK/GCV gene therapy.