Thymoquinone Induces Telomere Shortening, DNA Damage and Apoptosis in Human Glioblastoma Cells

Background A major concern of cancer chemotherapy is the side effects caused by the non-specific targeting of both normal and cancerous cells by therapeutic drugs. Much emphasis has been placed on discovering new compounds that target tumour cells more efficiently and selectively with minimal toxic effects on normal cells. Methodology/Principal Findings The cytotoxic effect of thymoquinone, a component derived from the plant Nigella sativa, was tested on human glioblastoma and normal cells. Our findings demonstrated that glioblastoma cells were more sensitive to thymoquinone-induced antiproliferative effects. Thymoquinone induced DNA damage, cell cycle arrest and apoptosis in the glioblastoma cells. It was also observed that thymoquinone facilitated telomere attrition by inhibiting the activity of telomerase. In addition to these, we investigated the role of DNA-PKcs on thymoquinone mediated changes in telomere length. Telomeres in glioblastoma cells with DNA-PKcs were more sensitive to thymoquinone mediated effects as compared to those cells deficient in DNA-PKcs. Conclusions/Significance Our results indicate that thymoquinone induces DNA damage, telomere attrition by inhibiting telomerase and cell death in glioblastoma cells. Telomere shortening was found to be dependent on the status of DNA-PKcs. Collectively, these data suggest that thymoquinone could be useful as a potential chemotherapeutic agent in the management for brain tumours.


Introduction
A major concern of cancer chemotherapy is the side effects induced by the non-specific targeting of both normal and cancerous cells by chemotherapeutic drugs. Much emphasis has been placed on discovering new compounds that specifically target cancer cells with minimal toxicity to normal cells. Recently various natural compounds have been shown to be promising chemotherapeutic agents with lesser cytotoxicity to the normal cells [1]. Thymoquinone (TQ) is the main active constituent found in the crude extracts of the seeds of Nigella sativa Linn. and studies have shown that TQ possesses antineoplastic properties [2,3,4]. Decrease in cell survival with pro-apoptotic properties have also been identified in various cancer cell lines including the canine osteosarcoma cells, human colon carcinoma, breast adenocarcinoma cells, and the ovarian adenocarcinoma cells [5]. TQ has been demonstrated as a cytotoxic agent in several multi-drug resistant human tumour cell lines [6]. However, the underlying molecular mechanism of its anticancer properties is not well understood.
Telomerase activity is generally absent in human somatic cells whereas in 90% of tumour cells, activation of telomerase prevents telomere shortening and thus allows for unlimited replicative capability [7]. Telomerase inhibition therefore presents an attractive target for cancer therapeutics [8,9]. Early events in the development of glioblastoma include the re-activation of human telomerase reverse transcriptase (hTERT) [10]. In addition, a p53 mutation also contributes to the progression to malignancy in gliomas [11]. Studies have shown that human glioblastoma cells deficient in DNA-PK activity are able to recruit a slow, error prone repair process that can result in the increased formation of chromosome aberrations [12]. DNA-dependent protein kinase is a nuclear, serine/threonine protein kinase consisting of a 470-kDa catalytic subunit (DNA-PKcs) and a heterodimeric regulatory complex ku70/80.This enzyme is essential for the repair of DNA double-strand breaks and it mediates repair via phosphorylation of downstream DNA binding proteins such as p53, DNA ligase IV and transcription factors such as Fos, Jun, myc, Oct1, NF-kappa B, and RNA polymerase H [13,14].
Brain tumours have low prognosis with survival rate of only few months and traditional therapy methods such as surgery, radiotherapy and chemotherapy have low efficacy. Furthermore, brain tumour cells are resistant to numerous chemotherapeutic agents. Inhibition of DNA repair proteins has been suggested to increase the sensitivity of DNA damaging chemotherapeutic agents. In this study, we sought to determine the anti-cancer effects of TQ in a defined DNA repair proficient and DNA repair deficient cells and its impact on the telomerase-telomere status in DNA-PKcs proficient and deficient human brain cancer cells. We performed in vitro studies to determine the level of cytotoxicity induced by TQ in human glioblastoma cells. In addition, the effects of TQ on both the telomerase activity and telomere length in cancer cells with different DNA-PK status were also investigated. Our findings demonstrate that at selective dose of TQ, glioblastoma cells were more sensitive to TQ-induced damage as compared to normal cells as indicated by the higher levels of apoptosis and reduced cell viability. Increased expression of pro-apoptotic proteins Bax and cytochrome c were also observed in glioblastoma cells. In addition, DNA-PKcs proficient glioblastoma cells were more sensitive to TQ as compared to DNA-PKcs deficient glioblastoma cells.

Cells, Cell Culture and Drug Treatment
Two human glioblastoma cell lines, M059K (CRL-2365) and M059J (CRL-2366) (American Type Culture Collection, USA) were used in the study. M059J cells lack DNA-PKcs activity while M059K cells express normal levels of DNA-PKcs. Glioblastoma cells were cultured in Dulbecco's Modified Eagles Medium (DMEM) supplemented with 10% heat inactivated foetal bovine serum (Hyclone, USA) and 100 U/ml of penicillin/streptomycin (Gibco, USA). hTERT immortalised human foreskin fibroblasts (hTERT-BJ1; Clontech, USA) were cultured in 4:1 ratio of DMEM and Medium 199, supplemented with 10% foetal bovine serum, 100 U/ml of penicillin/streptomycin, 1% sodium pyruvate and 2% L-glutamine (Gibco, USA). Normal human lung fibroblasts (IMR-90; Coriell Cell Repositories, USA) were cultured in Minimal Essential Medium (Gibco, Invitrogen, USA) supplemented with 15% foetal bovine serum, 100 U/ml of penicillin/ streptomycin, 1% vitamins and 1% essential and non-essential amino acids. All cells were maintained in a humidified 5% CO 2 incubator at 37uC. Stock solution of thymoquinone (TQ) (Sigma, USA) and a DNA-PKcs inhibitor NU7026 (Calbiochem, USA) [15] were prepared in dimethyl sulfoxide (DMSO) and suitable working concentrations were made from the stock using complete medium. Exponentially growing cells were treated with 0 to 200 mM TQ for 24 hours following which 25 and 50 mM doses were selected for subsequent studies.

Assay for cell viability
Following TQ treatment, attached cells were washed once with phosphate buffered saline (PBS). Crystal violet solution (0.75% crystal violet in 50% ethanol: distilled water with 1.75% formaldehyde and 0.25% NaCl), which stains DNA by binding electrostatically to nuclear proteins was added to the wells and incubated for 20 minutes at room temperature. Following PBS washes, to remove excess crystal violet solution, wells were air dried. Sodium dodecyl sulphate: PBS solution was added to lyse the cells and solubilise the dye. We measured the amount of crystal violet taken up by cells at 595 nm absorbance using an automated ELISA reader.

Cell cycle analysis
Following TQ treatment, cells were harvested, washed in 0.1% BSA: PBS, fixed in 70% ethanol: 1 6 PBS, and stained with propidium iodide (Sigma, USA): RNase A (Roche, USA) (2 mg propidium iodide and 2 mg RNaseA/100 mL 0.1% BSA in 1 6 PBS). Samples were analysed by flow cytometry (FACSCalibur TM , Becton Dickinson, USA) at 488 nm excitation l and 610 nm emission l. A total of 10,000 events were captured. Data obtained was analysed using WINMDI software.

Analysis for apoptosis and necrosis
Untreated and TQ-treated glioblastoma cells were stained with an Annexin V antibody and propidium iodide using Annexin-V-FITC staining kit (Sigma, USA). Samples were then analysed by flow cytometry. A total of 10,000 events per sample was obtained and the data was analysed using WINMDI software.

Protein estimation by Western blot analysis
Total cellular proteins were isolated using RIPA (radioimmunoprecipitation assay) buffer (1% nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M sodium phosphate, 2 mM EDTA, 50 mM sodium fluoride, 0.2 mM sodium vanadate and 100 U/ml aprotinin, pH 7.2) from control and treated cells. The whole cell lysate was recovered by centrifugation at 14,000 rpm for 10 minutes. Protein concentration was determined by the bicinchoninic acid method using an assay kit (Pierce Biotechnology, USA) with bovine serum albumin as a standard. Western blot analyses of cell cycle regulatory proteins (p53, p21), pro-apoptotic factors (Bax, Cytochrome c), hTERT and b-actin (Santa Cruz Biotechnology, USA) were performed using specific antibodies.

Assessment of DNA damage
TQ induced DNA damage was evaluated using alkaline single cell gel electrophoresis (Comet) assay. Following treatment with TQ for 24 hours, the extent of total DNA damage was evaluated as described previously [16]. Cells were harvested and resuspended in Hank's Balanced Salt Solution (Sigma, USA) with 10% DMSO and 0.5 M EDTA. The cell suspension was then suspended in 0.7% low melting agarose at 37uC (Conda, Spain), and layered on to comet slides (Trevigen, USA). The cells were then lysed in lysis solution containing 2.5 M NaCl, 100 mM pH 8.0 EDTA, 10 mM Tris-HCL, 1% Triton -X at 4uC for 1 hour. Denaturation was carried out for 40 minutes, in chilled alkaline electrophoresis buffer (pH 13.0-13.7). Electrophoresis was subsequently carried out for 20 minutes. Slides were immersed in neutralization buffer (500 Mm Tris-HCL, pH 7.4), dehydrated, dried and stained with SYBR Green dye (Trevigen) and scored with Comet Analysis Software (Metasystems, Germany). The images were captured using Zeiss Axioplan 2 imaging fluorescence microscope (Carl Zeiss, Germany) equipped with triple band filter. Fifty comets per sample were randomly selected and analysed. The extent of DNA damage was expressed as tail moment, which corresponded to the fraction of the DNA in the tail of the comet.

Measurement of Telomerase Activity
Telomerase activity was assessed using Telomeric Repeat Amplification Protocol (TRAP) using TRAPezeH XL Telomerase Detection Kit (Chemicon International, USA). All steps were done according to the manufacturer's instructions with some modifications. Briefly, total protein was extracted using CHAPS lysis buffer provided and 1.5 mg protein was treated with 1 ml/ml RNase inhibitor to perform PCR reaction. The PCR was initiated by using the telomerase mediated elongation products (i.e. the telomeric DNA repeats) as template. Subsequently normal PCR cycle was performed using primer pairs with quenched fluorescein to amplify the telomeric DNA repeats. Fluorescence signals were generated by unquenching the fluorescein on PCR primers, and the fluorescence signals were measured by fluorescence multi-well plate reader. Fluorescence signals of PCR products were measured using fluorescence plate reader TECAN SpectraFluor Plus.

Telomere length analysis
Telomere Restriction Fragment (TRF) Length Analysis.
DNA extraction was performed according to the manufacturers' protocol using DNeasy Tissue Kit (Qiagen, USA). The telomere restriction fragment length analysis assay was performed using Telo-TAGGG Length Assay Kit (Roche Applied Science, USA). Two micrograms of pure genomic DNA was digested with Hinf1 and Rsa1 restriction enzymes for 2 hours at 37uC. These restriction enzymes digest the whole genomic DNA except the sub-telomeric and telomeric regions. Digested DNA fragments were separated by gel electrophoresis in 0.8% agarose gel at 60 V for 3 hours, transferred via overnight Southern blot onto a nylon membrane and cross-linked onto the membrane using a UV cross-linker (Stratagene, USA). Telomere restriction fragments were hybridised with telomere-specific digoxigenin-labelled probe and incubated with Anti-DIG alkaline phosphatase and tetramethylbenzidine according to manufacturer's protocol. The chemiluminescence signals were scanned by the Kodak Gel imaging system and analyzed by the Kodak imaging software to calculate the quantitative measurements of the mean TRF length.

Quantitative Fluorescence in situ Hybridisation (qFISH)
Analysis. Cells were arrested at mitosis by treatment with colcemid (0.1 mg/ml). Cells were subsequently incubated with a hypotonic solution of potassium chloride at 37uC for 15 minutes followed by fixation in Carnoy's fixative. Quantitative fluorescence in situ hybridisation (qFISH) was performed using telomere sequence-specific peptide nucleic acid (PNA) probe labelled with Cy3 as described [17,18]. Metaphase spreads for different samples were hybridised simultaneously. To avoid selection bias, good and well spread metaphases were randomly chosen for analysis. Images were acquired on the same day (within four hours) for all the samples using the Zeiss Axioplan 2 imaging fluorescence microscope. Fluorescence intensity of telomere signals was measured in 10-15 metaphases using the in situ imaging software (Metasystems, Germany).

Gene expression analysis
The total RNA was extracted from M059K and M059J cells using QIAmp RNA Blood Mini Kit (Qiagen, Hilden, Germany). The extracted RNA was quantified using NanoDrop 1000 (Thermo Scientific, USA). RNA integrity was checked using Bio-Analyzer (Agilent Technologies, Inc., USA). Five hundred nanograms of extracted RNA from each sample were used for gene expression study. TotalPrep RNA Amplification Kit (Ambion Inc., TX, USA) was used for cRNA amplification process. The biotinylated amplified RNA thus generated was used for hybridization with HumanRef8 V3.0, Human Whole-Genome Expression BeadChips (Illumina Inc., USA) for 16 hours at 58uC. After the incubation period, the arrays were washed and stained with Streptavidin-Cy3 (GE Healthcare, Bio-Sciences, UK). Illumina Bead Array Reader was used to scan the arrays. The array data thus obtained after scanning was imported and analysed using PartekH Genomics Suite TM (Partek GS) (Partek Incorporated, MO, USA).

Chromosome analysis by multicolour fluorescence in situ hybridisation (mFISH)
The experimental procedure is outlined in our earlier publications [19,20]. Chromosome paints were obtained from MetaSystems GmbH, Germany. After pepsin treatment (1 g/ 50 ml; Sigma) for about 2 minutes at 37uC the chromosomes were post-fixed in 1% formaldehyde in 1XPBS with 50 mM MgCl 2 for 10 minutes at room temperature. The chromosomes are stabilised in SSC prior to the denaturation and were subsequently denatured with a basal solution (0.07 N NaOH). Afterwards, the chromosomes were rinsed in SSC buffer again to stop the denaturation process and to stabilize their structure. The probe cocktail was denatured 75uC for 5 minutes. The probe was allowed to prehybridise for half an hour at 37uC to reduce unspecific binding of short or repetitive DNA pieces. The denatured and prehybridised probe cocktail was applied onto the denatured chromosome preparation, overlaid with a coverslip and sealed with rubber cement. The slides were then incubated at 37uC in a humidified chamber for 72 hours. Post-hybridisation washing was done in 1xSSC at 75uC for 5 minutes after removing the coverslips. Slides were incubated in 4xSSCT for 5 minutes and blocked with blocking reagent at 37uC for 10 minutes. Subsequently, slides were incubated with detection 1+3 reagent to detect Cy5 which is indirectly labelled. Slides were then counterstained with DAPI (Metasystems). Microscopic analysis was performed using an Axioplan II imaging microscope (Carl Zeiss) with an HBO-103 mercury lamp and filter sets for FITC, Cy3.5, Texas Red, Cy5, Aqua, and DAPI. Images were captured, processed, and analyzed using ISIS mFISH imaging software (MetaSystems). In the mFISH technique, all 23 chromosomes (1-22 and X) are each painted in a different colour, using combinatorial labelling, so that any interchromosomal translocations are observed as colour junctions on individual chromosomes.

Statistical Analysis
Statistical significance in the data sets was assessed by Student's t-test using Microsoft Excel 2003 (Microsoft Corporation, USA) and two-way ANOVA using Graphpad Prism. The difference was considered to be statistically significant when p,0.05.

Results
Thymoquinone reduced the viability of the human brain cancer cells more effectively than non-cancerous cells Exposure to TQ resulted in a dose dependent decrease in cell survival in all the cell lines from 0 to 50 mM of TQ ( Figure 1A). The glioblastoma cells displayed greater TQ induced cell death compared to the normal cells tested. The greatest difference in survival between the normal cells and glioblastoma cells following TQ treatment occurred at concentration of 50 mM. The death induced in glioblastoma cells was significantly higher than normal (IMR90) cells ( Figure 1A). Based on the viability trend obtained, 50 mM was chosen for subsequent assays. Furthermore at 50 mM, DNA-PKcs proficient M059K cells were more sensitive TQinduced cell death as compared to DNA-PKcs deficient (M059J) cells.
To determine whether the growth inhibition by TQ involves cell death or growth arrest, cells stained with PI were subjected to FACS analysis. Increase in sub-G1 population, an indication of apoptotic death, was observed only in telomerase positive cells ( Figure 1B). At 50 mM TQ, the glioblastoma cells exhibited higher percentage of apoptotic cell death (sub-G1 population) than the normal IMR90 cells ( Figure 1B). There was a two fold increase in glioblastoma cell death following TQ treatment compared to untreated controls. In addition, M059J cells exhibited G2/M arrest (about 10%). In contrast, a three-fold increase in S-phase population in normal cells treated with TQ was observed. However, in telomerase positive (hTERT-BJ1) cells, there was a 5 fold increase sub-G1 population following TQ treatment. Impairment of growth in normal cells following TQ treatment was accompanied by the increase in p53 and p21 proteins ( Figure 1C). However, hTERT-BJ1 cells showed an increase in p53 protein level followed by a decrease in the expression of cyclin dependent kinase inhibitor protein, p21.

Thymoquinone induces apoptotic death in glioblastoma cells
As shown in Figure 2A, all tested cells treated with 50 mM TQ exhibited change in morphology compared to their respective controls. However, more floating and rounded cells which suggest dying cells were observed in glioblastoma cells. Expression of proapoptotic Bax protein level increased only in glioblastoma cells (M059K and M059J) following 50 mM TQ ( Figure 2B). To investigate the apoptotic cascades involved in the TQ-induced cell death in glioblastoma cells, total and cytosolic cytochrome c protein level was measured. The level of cytosolic cytochrome c, which is critical for the initiation of apoptosis, was significantly increased in M059K cells ( Figure 2B). Cell death following TQ treatment in glioblastoma cells was predominantly due to apoptosis as demonstrated by the increase in cell population showing positive staining for Annexin V ( Figure 2C).
Thymoquinone produces significant DNA damage in a dose dependent manner Single-cell gel electrophoresis (known as comet assay) under alkaline condition (pH.13) was used to determine the DNA damage induced by TQ (Figures 3A and B). Different types of DNA damage such as double-strand breaks, single-strand breaks and alkali labile sites were measured using comet assay. Representative images of comet analysis were shown in Figure 3A. Following TQ treatment, all cells showed a dose dependent increase in DNA damage compared to respective controls ( Figure 3B). At 50 mM TQ, M059J cells were less susceptible to TQ induced DNA damage compared to M059K cells. Interestingly, DNA damage was lower at 25 mM TQ in normal fibroblast cells (IMR90 and hTERT-BJ1) than in glioblastoma cells. On the contrary, TQ treated M059J cells showed significant DNA damage unlike TQ treated M059K cells that displayed minimal increase when compared to respective untreated control cells at 25 mM TQ ( Figure 3B). The DNA damage data obtained by comet assay were supported by higher micronuclei detected in the TQ treated cells (data not shown) which are suggestive of both DNA damage and resulting chromosome instability.

Thymoquinone induces telomere attrition
Our findings from telomerase positive hTERT-BJ1 fibroblasts demonstrate increased sensitivity to TQ induced anti-proliferative effect as compared to normal cells. To determine whether the decrease in cell proliferation was linked to effect of TQ on telomerase activity, we evaluated the level of telomerase activity in cells with and without TQ treatment. Firstly, the basal level of telomerase activity in the different cells was investigated. As expected, telomerase activity was negligible in normal lung fibroblasts (IMR90), while the hTERT-BJ1 and glioblastoma cells showed positive telomerase activity ( Figure 4A). M059J cells displayed lower telomerase activity compared to M059K cells. At 24 hours post-TQ treatment (50 mM), a significant decrease in telomerase activity was detected in hTERT-BJ1 and M059K but not in M059J cells ( Figure 4B). Furthermore, we have observed that there is a decrease in the expression of hTERT following TQ treatment ( Figure 4C) demonstrating a novel effect of TQ on telomere-telomerase complex.
Since, telomerase is pivotal for the maintenance of telomere length, we monitored changes in telomere length, if any, following TQ treatment using TRF assay. Lower dose (25 mM) of TQ was used for an extended time of 15 days in glioblastoma cells and telomere length was measured at 4 and 15 days of treatment. Only glioblastoma cells were evaluated for telomere length analysis because our findings have shown that the level of TQ-induced DNA damage ( Figure 3) and anti-proliferative effect (Figure 1) of TQ were more pronounced in glioblastoma cells as compared to normal cells. Short term treatment for 4 days with 25 mM TQ did not show detectable change in telomere length ( Figure 4D). Following TQ (25 mM) treatment for 15 days, M059K cells showed greater telomere shortening compared to M059J cells ( Figure 4E). Thus the data suggest that TQ disrupts telomere length maintenance by inhibiting the activity of telomerase over time in cancer cells. However, M059J cells deficient in DNA-PKcs were less sensitive to TQ mediated effects on telomere length maintenance. Therefore, we investigated the role of DNA-PKcs in TQ mediated effects on telomeres.

Role of DNA-PKcs in thymoquinone mediated telomere attrition
In order to understand the influence of DNA-PKcs on the TQ mediated effects on telomeres in the glioblastoma cells, a two pronged approach was taken. Firstly, we profiled M059K and M059J cells for inherent similarities and differences by gene expression profiling and karyotyping. Secondly, we validated the role of DNA-PKcs by the using its specific pharmacological inhibitor, NU7026 [15].
It is reported that M059J cells were derived from M059K cells and were deficient in DNA-PKcs [21,22]. A detailed characterisation and gene expression profiles of these cells have not been reported so far. mFISH was performed to karyotype M059K and M059J cells. Modal chromosome number for M059K cell type is to be 75 (with a range of 65 to 79 (www.atcc.org) with a polyploidy rate of 22% whereas M059J cells reported to be aneuploid (www. atcc.org). Representative mFISH images of M059K and M059J metaphase spreads are shown in Figure 5A. Our analysis has revealed that average chromosome number for M059K is 87 and that of M059J is 68. The karyotypes clearly indicate complex chromosomal rearrangements like reciprocal and non-reciprocal translocations in both the cell types. Recurrent chromosomal translocations such as t(8;21;5), t(1;21;6), t(5;21) and t(4;17;22;21;9) were detected in M059K cells. Likewise, translocations such as t(7;16;14); t(11;12); t(8;6); 2 copies of t(9;21;3), and two copies of t(9;21;19) were detected in M059J cells. Translocations such as t(16;17), t(16;10) and multiples copies of t(9;21) were commonly seen in both the cell types. At this moment, it is not possible for us to conclude that the translocations detected in M059J cells are result of defective DNA-PKcs. Following microarray analysis to determine the gene expression profiles of M059K and M059J cells, we found that a total of 587 genes (approximately 2% of the genes analysed) were differentially expressed based on the set criteria  (p,0.05; fold difference of 2) as indicated in Figure 5B. This might suggest that genetic differences exist, but minimal, among the two different cell types (data not shown).
As shown in Figures 4B and D, cells deficient in DNA-PKcs (M059J) exhibited relatively lesser TQ-mediated telomerase inhibition and telomere attrition. To further understand the role of DNA-PKcs in the differential effect of TQ on telomeretelomerase homeostasis, we used NU7026, a DNA-PKcs specific pharmacological inhibitor along with TQ. We observed that TQ induced telomerase inhibition in M059K cells which were pretreated with NU7026 was 10% lesser than TQ alone treatment (TQ -23% inhibition and NU7026 + TQ -13% inhibition) ( Figure 6A). Similar trend was seen with regards to TQ mediated telomere attrition and NU7026 was found to slow this process down in M059K cells (Figures 6B-D). These results were comparable to those seen in M059J cells (Figures 4A-E) highlighting the role of DNA-PKcs in TQ mediated telomere shortening in glioblastoma cells. We hypothesised that telomere attrition resulted from TQ treatment in glioblastoma cells may produce dysfunctional telomeres. Treatment with drugs may generate aberrant telomeres [23] due to dysfunctional telomeres and replication stress. To determine whether telomerase inhibition and telomere shortening in TQ treated M059K cells has resulted in telomere dysfunction, we analysed fragile telomeres in M059K cells following TQ exposure with or without NU7026. In metaphase spreads, fragile or aberrant telomeres are characterised by the presence of undetectable telomeres at chromosome ends, telomeric doublets and different signal intensities on sister chromatids. Representative metaphase images with fragile telomeres are shown in Figure 6E. Number of aberrant telomeres per cell displayed in Figure 6F corresponded to telomere attrition and telomerase inhibition (Figures 6A-D). TQ alone has resulted in elevated number of dysfunctional telomeres, while combined treatment produced a lower number of dysfunctional telomeres compared to TQ single treatment but higher than the NU7026 alone treatment ( Figure 6F).

Discussion
There has been an increasing interest in natural plant products as potential chemotherapy in human tumours due to their lower cytotoxic effects on the normal cellular system [1]. Limited number of reports is available in the literature on the growth inhibitory and pro-apoptotic effects of TQ in tumour cells [5,24]. Moreover, effect of TQ on telomerase activity and telomere length in cancer cells has not yet been explored. We demonstrate in this study that TQ exhibits specific cytotoxicity in cancer cells at selective doses. Glioblastoma cells were found to be more sensitive to TQ-induced anti-proliferative effect as compared to normal cells (IMR90 and hTERT-BJ1). Reduction in cell viability was attributed to the induction of apoptosis by TQ in glioblastoma cells as evidenced by higher percentage of cells with Annexin V staining and more cells in sub-G1 phase of the cell cycle following treatment.
Majority of the cancer chemotherapy drugs function as DNA damaging agents that subsequently trigger cell death. In this study, we show that TQ induces DNA damage and cell death in glioblastoma cells. When DNA damages are irreparable, the cells undergo apoptosis. Induction of DNA damage in normal cells led to p53 mediated p21 triggered growth inhibitory effects which We believe that the apoptotic effect of TQ was modulated by increased expression of Bax and cytochrome c proteins. Apoptosis via mitochondrial pathway involves the release of the intermembrane mitochondrial proteins, cytochrome c into cytosol and subsequently triggering a cascade of proteolytic events [25]. Based on our findings, we speculate that TQ induces apoptosis in glioblastoma cells proficient in DNA-PKcs by the abovementioned mechanism, which was evidenced by an increase in the cytosolic cytochrome c level. Studies have shown that TQ induces apoptosis by both p53-dependent [24] and p53-independent pathways [26]. Our data agrees with the latter wherein cell death triggered by TQ in glioblastoma was p53 independent as shown by the minimal change in expression of p53.
Tumour cells develop resistance to chemotherapy or radiotherapy over time via genetic alteration resulting in less effective treatment. DNA-PKcs is essential for DNA double strand breaks repair and maintenance of genomic integrity [27]. Studies have shown that DNA double strand break repair in cells deficient in DNA-PKcs usually results in high frequencies of mis-repair, chromosome aberrations and complex changes [12]. In our study, we show that glioblastoma cells with different DNA-PKcs status respond differently to TQ exposure. DNA-PKcs deficient glioblastoma cells displayed significant cell cycle arrest at G 2 /M phase in addition to cell death. Significant level of TQ induced DNA damage was also observed in these cells at lower dose of TQ. This observation correlates with earlier findings that only DNA-PKcs deficient glioblastoma cells showed G 2 /M accumulation upon treatment with DNA double strand break inducing agents [28]. Moreover, the frequency of binucleated cells with micronuclei, indicative of chromosome instability, was higher in DNA-PKcs deficient glioblastoma cells as compared to DNA-PKcs proficient glioblastoma cells (data not shown). This points out that although DNA-PKcs deficient glioblastoma cells were less sensitive to TQ induced cytotoxicity as compared to DNA-PKcs proficient glioblastoma cells, greater extent of genomic instability could be detected in DNA-PKcs deficient glioblastoma which might be due to the inefficient non-homologous end joining DNA repair processes. The increased cell death shown in DNA-PKcs proficient glioblastoma cells supports the view that TQ targets DNA-PKcs proficient glioblastoma cells and that DNA-PKcs activation may play a role in apoptosis. This finding corroborates an earlier study, which showed a greater resistance of DNA-PKcs deficient glioblastoma cells to cell death as compared to DNA-PKcs proficient glioblastoma [29].
Interestingly, telomerase positive human foreskin fibroblast cells (hTERT-BJ1) were also more sensitive to growth inhibitory effect of TQ as compared to normal lung fibroblasts (IMR90). Telomerase activity was evaluated following TQ treatment to hTERT-BJ1 cells to determine whether increased sensitivity was due to telomerase inhibition. Telomerase is essential in regulating telomere length in immortal cells as well as maintaining the proliferative capacity of cells [8,30,31,32]. Telomerase may play a central role in cellular resistance to apoptosis of cancer cells and inhibition of telomerase may induce apoptosis [33,34]. Telomere dysfunction, whether due to replicative shortening or experimental disruption of telomeric DNA-protein complex structure, leads to rapid DNA damage response which can ultimately induce senescence and/or apoptosis [35]. Attempts are being made to exploit this aspect in cancer therapy. In our study, for the first time, we found that TQ significantly reduces telomerase activity and induces telomere attrition. A significant reduction in telomerase activity and expression of hTERT protein was observed in hTERT-BJ1 and M059K cells and not M059J cells following TQ treatment. Among the cancer cells, we observed that the level of telomerase activity decreased significantly only in DNA-PKcs proficient glioblastoma cells with respect to its controls. Similar results were obtained with regards to TQ mediated telomere attrition which can be attributed to the selective inhibition of telomerase activity in DNA-PKcs proficient cells.
The differential response of M059J and M059K gliobloastoma cells to TQ treatment prompted us to characterise these cells to determine whether defective DNA-PKcs status has resulted in any specific changes in the basal genetic traits. Earlier, we had shown that DNA-PKcs is critical for telomere capping [36,37] apart from its function in responding to DNA damage. DNA-PKcs deficient mouse embryonic fibroblasts displayed higher telomere fusions and chromosome instability [36]. Considering the role of DNA-PKcs at telomeres, we went on to validate our observation on the role of DNA-PKcs in TQ mediated telomere attrition. Subtle differences in the karyotypes of M059K and M059J cells were identified and these cells harbour several recurrent complex chromosomal rearrangements. In order to further the understanding if the above findings, gene expression profiling was done in these cell lines. A total 587 genes (,2% of the total genes profiled) were differentially expressed based on the set analysis criteria. However, more work needs to be done to validate the functional relevance of these genes and whether they are dependent on the status of DNA-PKcs. All in all, there are differences in the chromosome rearrangements and gene expression profiles of two glioblastoma cell types used in the study but the gross genetic changes might be minimal in the two glioblastoma cell types used in the study. Telomere-mediated chromosome instability was commonly detected in cells deficient in DNA-PKcs [36,38,39,40]. Therefore, we believe that defective DNA-PKcs might be the cause of differential response to TQ in our study.
In order to validate and substantiate our data on DNA-PKcs deficient M059J cells, we used NU7026 to inhibit the kinase activity of DNA-PKcs in M059K cells to study the telomere attrition following TQ treatment. Pre-treatment with NU7026 in M059K cells produced a similar response to that of M059J cells for TQ exposure. Therefore, it is clear that TQ mediated effects on telomerase and telomere length were significantly reduced in the absence of DNA-PKcs achieved either by using a DNA-PKcs deficient cell type or by functional inhibition of its kinase activity in DNA-PKcs proficient cells. We therefore, suggest the DNA-PKcs activity is rather relevant to mediate the above observed TQ effects. It was recently reported that the absence of DNA-PKcs increases the spontaneous telomere attrition rate in telomerase knockout (mTERC 2/2 ) mice [41]. Based on our results, we can measurements using TRF analysis (B and D) and qFISH (C) in glioblastoma cells following 15 day-treatment with 25 mM TQ and NU7026 (10 mM). Changes in telomere length by TRF are expressed as percentage with respect to its controls (D). Mean 6 standard error from two independent experiments are shown. * p,0.05. (C) Telomere fluorescence intensity (TFI) values from qFISH analysis are displayed for different samples. (E-F) Telomere dysfunction induced by 15-day treatment with TQ (25 mM) in DNA-PKcs inhibited (NU7026, 10 mM) M059K cells. (E) Partial metaphase spreads from M059K cells displaying chromosomes stained for telomeres (red) and centromeres (green) with telomeric doublets (1), different signal intensities on sister chromatids (2)  cautiously speculate that DNA-PKcs activity is relevant to specifically facilitate TQ effects. Further investigation is warranted to find out the factors downstream of DNA-PKcs which are required for external agent induced telomere attrition in human cells. In conclusion, we demonstrate that TQ induces higher induction of apoptosis in glioblastoma cells. Furthermore, we could detect telomerase inhibition, telomere attrition, increased DNA damage and apoptosis by TQ in DNA-PKcs proficient glioblastoma cells as compared to DNA-PKcs deficient glioblastoma cells. In addition, TQ also mediates apoptosis independent of telomerase attrition as observed in DNA-PKcs deficient glioblastoma cells. Thus, this agent can be a potential candidate as chemotherapeutic agent for the treatment of brain cancers. However, detailed studies are required to profile the genome wide effects of TQ to exploit its therapeutic potential more effectively.