|Home | About | Journals | Submit | Contact Us | Français|
The radiation-induced bystander effect (RIBE) increases the probability of cellular response and therefore has important implications for cancer risk assessment following low-dose irradiation and for the likelihood of secondary cancers after radiotherapy. However, our knowledge of bystander signaling factors, especially those having long half-lives, is still limited. The present study found that, when a fraction of cells within a glioblastoma population were individually irradiated with helium ions from a particle microbeam, the yield of micronuclei (MN) in the nontargeted cells was increased, but these bystander MN were eliminated by treating the cells with either aminoguanidine (an inhibitor of inducible nitric oxide (NO) synthase) or anti-transforming growth factor β1 (anti-TGF-β1), indicating that NO and TGF-β1 are involved in the RIBE. Intracellular NO was detected in the bystander cells, and additional TGF-β1 was detected in the medium from irradiated T98G cells, but it was diminished by aminoguanidine. Consistent with this, an NO donor, diethylamine nitric oxide (DEANO), induced TGF-β1 generation in T98G cells. Conversely, treatment of cells with recombinant TGF-β1 could also induce NO and MN in T98G cells. Treatment of T98G cells with anti-TGF-β1 inhibited the NO production when only 1% of cells were targeted, but not when 100% of cells were targeted. Our results indicate that, downstream of radiation-induced NO, TGF-β1 can be released from targeted T98G cells and plays a key role as a signaling factor in the RIBE by further inducing free radicals and DNA damage in the nontargeted bystander cells.
The intense interest in radiation-induced bystander effect (RIBE) is a relative recent phenomenon dating back to the middle 1990s (Nagasawa and Little, 1992, 1999; Azzam et al., 1998; Mothersill and Seymour, 1998; Prise et al., 1998; Bishayee et al., 1999). The main interest has been on the impact of RIBE on low-dose risk assessment and radiation protection, especially in relation to radiation carcinogenesis induced by α-particles such as emitted by radon and its progeny. Mounting evidence has shown that TP53 expression (Azzam et al., 2001), chromosome damage (Deshpande et al., 1996), genomic instability (Seymour and Mothersill, 1997; Limoli et al., 2000; Morgan, 2003a,b), mutations (Little, 2000; Zhou et al., 2000) and malignant transformation (Lewis et al., 2001; Sawant et al., 2001) can be increased in nonirradiated normal cells that are in the vicinity of irradiated cells.
Most recently, bystander responses have also been observed in tumor cells. We (Shao et al., 2003a, b, 2004a) have found that irradiated neoplastic human salivary gland or glioblastoma T98G cells can induce multiple bystander effects including cell growth stimulation, micronucleus (MN) formation and apoptosis in nonirradiated tumor cells. Using microbeam approaches, it has been shown that, as well as irradiation of the nucleus, cytoplasmic irradiation also triggers bystander response in tumor cells (Shao et al., 2004b, Tartier et al., 2007). The medium from irradiated cells is another important source of RIBE, and this kind of bystander response is independent of radiation dose (Mothersill and Seymour, 1997). Moreover, normal and tumor cells can interact with each other via bystander signaling when one irradiated population is cocultured with the other nonirradiated population (Shao et al., 2005). These findings indicate a potential implication of the bystander effect for radiotherapy.
Studies investigating factors underlying RIBE indicate that both gap junctional intercellular communication (Azzam et al., 2001) and/or some soluble factors such as reactive oxygen species (ROS) (Narayanan et al., 1997) and transforming growth factor-β1 (TGF-β1) (Iyer and Lehnert, 2000) contribute to the bystander responses observed in normal fibroblast cells. We and others have found that nitric oxide (NO), an important signaling molecule, is involved in the conditioned medium (CM)-mediated bystander effect on neoplastic, lymphoma and glioblastoma cells (Matsumoto et al., 2001; Shao et al., 2003b, 2004a) and may be an early activator of DNA damage in bystander cells (Han et al., 2006). Since free radicals of NO and ROS have very short half-lives, it is unlikely that they are the key longer-term signaling factors present in the CM. It is largely unknown what other bystander signaling factors are involved and their relationship to radiation-induced free radicals, other than a report that TGF-β1 is a mediator of the increased intracellular ROS and decreased TP53/CDKN1A bystander responses in HFL1 fibroblast cells irradiated with low doses of α-particle (Iyer and Lehnert, 2000). In addition, calcium fluxes have been found to be an early response in RIBE leading to MN formation and apoptosis (Lyng et al., 2006; Shao et al., 2006). TGF-β1 is known to be a key extracellular sensor and signal of stress responses in irradiated tissues (for a review see Barcellos-Hoff, 2005).
In the present study, we investigated the role of NO and TGF-β1 in the bystander response of glioblastoma T98G cells targeted with a precise number of helium ions delivered by a microbeam, and found that TGF-β1 was a downstream product of irradiation-induced NO and it could further trigger the expression of NO in nonirradiated bystander cells.
As a result of chromosome damage, MN were produced in the irradiated T98G cells. When 100% of the cells in the population were individually irradiated with a single 3He2+ particle, the YMN was 0.305±0.035. When 1 or 10% of the cells in the population were individually irradiated with one 3He2+ particle, the YMN was less than that of 100% of cell irradiation and reduced to 0.176 and 0.225, respectively (Figure 1). According to the calculation method that we have described previously (Shao et al., 2003b), if there were no bystander effect, the yields of MN in the 1 and 10% fractionally irradiated population would be 0.140 and 0.155, respectively. These predicted values are significantly less than the above measured values, which indicates that significant numbers of nontargeted cells in the T98G population responded to their neighbors being irradiated.
Figure 1 also illustrates that when the cells were treated with either the inducible NO synthase (iNOS) inhibitor aminoguanidine (AG) or anti-TGF-β1, YMN of the irradiated population was reduced to the level predicted assuming no bystander effect. It has been reported that radiation-induced activity of iNOS will allow T98G cells to release NO (Matsumoto et al., 2000), therefore both NO and TGF-β1 are signaling factors involved in the targeted T98G-induced bystander responses. It is well known that the NO-free radical has a very short half-life and a very short diffusion distance, thus TGF-β1 is a strong candidate for a signaling factor that can stably induce further bystander effect in T98G cells.
To confirm the deduction that TGF-β1 was involved in the bystander responses, we measured TGF-β1 in the CM collected from irradiated T98G cells. Figure 2 illustrates that when 1, 10 and 100% of the cells in the population were individually irradiated with a single 3He2+ particle, the TGF-β1 level increased to about 1.2 times the control levels 24 h post-irradiation, but was independent of the fraction of irradiated cells. When the cells were treated with aminoguanidine so that iNOS was inhibited, the concentration of TGF-β1 in the CM was reduced to the control level, which indicates that the production of TGF-β1 may be regulated by NO. To verify this, we treated T98G cells with the NO donor diethylamine nitric oxide (DEANO) and measured TGF-β1 levels in the culture medium. Results showed that TGF-β1 was indeed induced by NO and its concentration varied with both the treatment time and the concentration of DEANO (Figure 3). With 4 h of NO treatment, the TGF-β1 level in the cell culture medium was increased by 100 μm DEANO, but was not significantly increased by 1 or 10 μm DEANO. However, with 24 h of NO treatment, it was 1 μm, but not 100 μm DEANO that enhanced the TGF-β1 production, and 10 μm DEANO also slightly increased the TGF-β1 level (P>0.05). Accordingly, as a bystander signaling factor, the production of TGF-β1 may be dependent on both the concentration of radiation-induced NO and the timescale of its production.
To investigate the effect of TGF-β1 on cellular damage and related pathways, we treated T98G cells with TGF-β1 for 24 h and measured cellular responses. It was found that, when cells were treated with TGF-β1 at concentrations as low as 1 and 5 ng/ml, the fluorescence intensity related to the intracellular NO level in the whole population was increased by about 20% (Figure 4). Under the same conditions, the yield of MN in the population was also increased, which is consistent with the data in Figure 1 that anti-TGF-β1 diminishes the bystander MN induction.
Figure 5 illustrates that 24 h after 1 or 100% of the cells in T98G population were individually irradiated with a single 3He2+ particle, the intracellular NO level increased to about 1.2 times of the control, but was independent of the fraction of irradiated cells, which is consistent with the result of irradiation-induced TGF-β1 as shown in Figure 2. It was found that when only 1% of cells were irradiated, the percentage of NO-positive cells increased by about 35% (data not shown). Clearly, most of the increased percentage was caused by the bystander response.
Interestingly, under the condition of 1% of the cells being irradiated, treatment of cells with anti-TGF-β1 inhibited the induction of bystander NO (see Figure 5). However, when 100% of the cells in the population were irradiated, the anti-TGF-β1 treatment did not influence the induction of NO. These data suggest that NO can be induced from two pathways, one is from direct irradiation, the other one from a bystander response. In general, TGF-β1 is a downstream product of radiation-induced NO, but it also regulates the induction of bystander NO in the nonirradiated cells.
We have previously reported that NO is a signaling factor contributing to the RIBE (Shao et al., 2003b), which is confirmed again by the result in Figure 1 that an iNOS inhibitor diminishes the bystander MN induction. The measured MN are probably a consequence of DNA damage, which have been induced in bystander cells. DNA double-strand break (DSB) have been detected by several groups in bystander cells using gH2AX as a marker (Sokolov et al., 2005; Yang et al., 2005) and are thought to occur via the accumulation of DNA damage in S phase leading to stalled replication forks (Burdak-Rothkamm et al., 2007). More interestingly, it is found that anti-TGF-β1 also can inhibit the bystander MN induction when a fraction of cells in the population are targeted individually. Thus TGF-β1 is a key signaling factor involved in the bystander response, which is supported by the result in Figure 2 demonstrating that TGF-β1 is released into the CM from irradiated cells. This result is in accord with another study showing that TGF-β1 could be released from irradiated cancer cells and its yield increased with time post-irradiation (Arnold et al., 1999).
It is found that when the fraction of irradiated cells increases from 1 to 100%, the relative level of TGF-β1 in the CM and the relative level of NO do not increase significantly. This finding is in agreement with our previous reports that the bystander MN induction did not increase with irradiation dose when the bystander cells were treated with the CM from irradiated cells, (Shao et al., 2003b) and that the increased fraction of irradiated T98G cells did not increase the bystander response in the neighboring nontargeted AG01522 cells (Shao et al., 2005). A mechanism of signal amplification may play a role in the phenomenon of dose-independent TGF-β1 production. Signaling factors released from a few directly targeted cells diffuse and react with adjacent nontargeted cells and then lead to new signaling molecules being produced. Therefore, the signaling level in the few cells being irradiated is amplified to a relatively high level compared to that in the population where 100% of cells are irradiated. Another possible reason for this phenomenon is that the concentration of TGF-β1 in the medium has no proportional relationship to the irradiation dose as reported by others who found that the relative mRNA amounts of TGF-β1 did not relate to either the radiation dose or the time postirradiation (Boerma et al., 2002).
We find that two bystander signaling factors, NO and TGF-β1, are not independent of each other but are closely related. Treatment of cells with AG, an iNOS inhibitor, reduces the concentration of TGF-β1 to control levels, which indicates that TGF-β1 is a product downstream of radiation-induced NO. In fact, TGF-β1 can be released from T98G cells when they are treated with an NO donor, that is TGF-β1 is regulated by NO (Ayache et al., 2002). Schmidt et al. (2003) reported that TGF-β1 could be induced by exogenous NO and then overexpressed in the culture medium, as well as intracellular and extracellular matrix components of human smooth muscle cells. In contrast, Vodovotz et al. (1999) found that when human lung adenocarcinoma cells were treated with DEANO, TGF-β1 was over-expressed and bound at or near the cell surface, but was not secreted to the culture medium. The present study also finds that TGF-β1 can be released from the DEANO-treated glioma cells, but the extent of this is dependent on both the concentration of DEANO and the treatment time (see Figure 3). The reason for this variable increment of TGF-β1 level is still unknown, but it may be related to the binding site of the TGF-β1 molecule. Once TGF-β1 is induced by NO, it can be bound inside of the cell, secreted outside of the cell and exist in the medium and bound to the extracellular matrix, so that the measured amount of TGF-β1 in the CM may only qualitatively reflect the level of NO-induced TGF-β1.
On the other hand, although most of studies prove that TGF-β1 is a downstream product of NO (Chesrown et al., 1994; Blanco et al., 1995), evidence suggests that TGF-β1 can either attenuate or augment iNOS expression, with the prevailing effect dependent on the experimental paradigm and the cell type (Hamby et al., 2006). The present work finds that, with the treatment of TGF-β1, the intracellular NO level in T98G cells is increased, that is TGF-β1 upregulates NO. Therefore, it is possible that NO has a biphasic effect on the expression of TGF-β1, involving downstream and/or upstream regulation, similar to that proposed by Vodovotz et al. (1999). If significant interplay with other ROS occurs, it is also possible that stabilization of latent TGF-β forms may be an important step (Jobling et al., 2006).
The expression of intracellular NO can be stimulated from two sources, direct irradiation and radiation-induced TGF-β1. When 100% of the cells are targeted, the increased intracellular NO results from the radiation-activated iNOS, and thus has no relationship with anti-TGF-β1 so that treatment of the T98G cell population with anti-TGF-β1 does not alter the production of radiation-induced NO. However, when only 1% of the cells in the population are targeted, the enhanced NO mostly results from the bystander response where TGF-β1 is involved; therefore, it can be reduced by the treatment of anti-TGF-β1.
Our recent study (Shao et al., 2006) has also shown that when T98G cells are treated with calcicludine, a highly potent Ca2+ channel blocker, induced bystander responses leading to both MN induction and NO generation are eliminated. This study showed that calcium flux is an early event leading to the formation of bystander MN induced by the CM from irradiated T98G cells, and the signaling factors contained in the CM are downstream of NO. The present study indicates that TGF-β1 downstream to NO could be one of those bystander signaling factors in the CM since it has been known that TGF-β1 at a wide range of concentrations transiently increases the level of intracellular calcium concentration (Gizatullina et al., 2003; Nesti et al., 2007). The induction of an intracellular calcium flux can activate some calcium-dependent NO synthases and induce the generation of NO (Newman et al., 2004). As a major product of NO oxidization, peroxynitrite can cause DNA strand breaks and induce MN formation (Dedon and Tannenbaum, 2004).
Taken together, both TGF-β1 and NO are involved in the targeted irradiation-induced bystander effect on glioblastoma T98G cells, and these two signaling factors are interdependent. Figure 6 illustrates a proposed model for radiation-induced bystander responses and the signaling pathway involved. When a fraction of cells within the population are individually targeted, NO and its downstream product TGF-β1 can be released from targeted cells. Although NO free radical has a very short half-life and can diffuse only few micrometers away, TGF-β1 is a relatively stable product and can diffuse freely in the medium. Once it interacts with nonirradiated bystander cells, TGF-β1 will induce the generation of intracellular NO as a secondary bystander signal, probably via a calcium-dependent pathway, and further cause MN formation in these bystander cells.
Human glioblastoma T98G cells (European Collection of Animal Cell Cultures, Porton Down, UK) were cultured in RPMI-1640 medium supplemented with 10% of fetal calf serum, 0.01% sodium pyruvate, 2 mm l-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. The culture was maintained at 37°C in an atmosphere of 95% air and 5% CO2.
One day before microbeam irradiation, approximately 1200 plateau phase T98G cells were seeded in the central area (5 mm in diameter) of a specially designed microbeam dish consisting of a 3 μm thick Mylar film base and cultured in RPMI-1640 medium. The regions prepared for cell seeding had been pretreated with 1.7 μg/cm2 Cell-Tak adhesive (Collaborative Biomedical Products, Bedford, MA, USA). The cell nuclei were stained with 0.2 μg/ml Hoechst-33342 1 h before irradiation, enabling individual nuclei to be identified by the microbeam system (Folkard et al., 1997b). The position of each Hoechst stained nucleus was found using a computerized imaging system, and its coordinates were stored in order to be revisited and irradiated automatically. Cell location and cell irradiation took about 15 min for single-cell irradiation and 40 min for 100% of cell irradiation, and during this time the cells were maintained in serum-free medium containing 10 mm 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES). The Gray Cancer Institute Microbeam system was used to deliver individual helium-3 (3He2+) ions to cells with high reproducibility (single ion delivered with >99% efficiency) and high accuracy (>99% within 2 μm) (Folkard et al., 1997a, b). A fraction of cells in the T98G population was irradiated through the center of the nucleus with a precise number of 100 keV/μm 3He2+ ions. Identical dishes of cells scanned but not irradiated were used as nonirradiated controls.
Immediately after irradiation, the medium was replaced with 2 ml of complete medium and cells were cultured for 24 h until further treatment for MN measurement. The CM from the irradiated population was also collected at this time for measurement of TGF-β1. In some experiments, cells were treated with either 20 μM AG during and after irradiation or 10 μg/ml of monoclonal anti-human TGF-β1 (Sigma, Poole, UK) after irradiation. AG inhibits the activity of intracellular iNOS.
To investigate the signals involved in the radiation-induced bystander responses, we treated the cells with DEANO (Molecular Probes Inc., Eugene, OR, USA) or human recombinant TGF-β1 (Sigma). DEANO can reliably generate NO in aqueous solution and its half-life and efficiency for NO release at 37°C is 2 min and 1.5, respectively (Keefer et al., 1996). In one arm of the experiment, 2 × 105 T98G cells per well seeded on a six-well plate were treated with DEANO of different concentrations for 4-24 h, then the CM was filtered and collected for measurement of TGF-β1 levels. In the other arm, 2 × 105 T98G cells were treated with human recombinant TGF-β1 of different concentrations for 24 h, and then MN formation and intracellular NO were measured.
The concentration of TGF-β1 in the CM generated from irradiated T98G cells and DEANO-treated T98G cells was assayed by using the human TGF-β1 Instant ELISA kit according to the protocol supplied (Bender MedSystems, Vienna, Austria).
The cytokinesis block technique was used to assay for MN in situ. Briefly, the cultures were treated with 1 μg/ml cytochalasin-B for 26 h and then fixed with methanol: acetic acid (9:1 (v/v)) for 20 min. Air-dried cells were stained with 10 μg/ml acridine orange. MN were scored in the binucleated cells and classified according to standard criteria (Albertini et al., 2000). The MN yield, YMN, was calculated as the ratio of the number of MN to the scored number of binucleated cells.
The intracellular NO, represented by its oxidization derivates, was assayed in situ by using 4-amino-5-methyl-amino-2′,7′-difluorofluorescein diacetate (DAF-FM, Molecular Probes Inc.). Briefly, cells were treated with 3.5 μm DAF-FM for 45 min at 37°C. After the excess probe was removed, cells were incubated for an additional 20 min to allow complete deesterification of the intracellular diacetates. The fluorescent images of 100–200 randomly selected cells per dish were captured using a 3CCD color camera (Photonic Science Ltd, East Sussex, UK) on a fluorescent microscope (Zeiss Axioskop) with manual UV-light shutter and filters. The exposure conditions were normalized to allow quantitative comparisons of the relative fluorescence intensity of the cells between groups.
Statistical analysis was done on the means of the data obtained from at least three independent experiments. Two replicates were counted for each experimental point in each experiment to determine MN. All results are presented as means±s.e.m. Significance was assessed using Student's t-test at P<0.05.
We are grateful to Stuart Gilchrist and Bob Sunderland for assistance with the microbeam irradiation. We acknowledge the support of Cancer Research UK (CUK) grant number C1513/A7047, the European Commission (NOTE, FI6R 036465) and the Gray Cancer Institute in these studies.