Increases in cell proliferation are widely viewed as being of importance in carcinogenesis. We report that exposure of normal human lung fibroblasts to a low dose of α particles like those emitted by radon/radon progeny stimulates their proliferation in vitro, and this response also occurs when unirradiated cells are treated with supernatants from α-irradiated cells. We attribute the promitogenic response to superoxide dismutase- and catalase-inhibitable α particle-induced increases in the concentrations of transforming growth factor β1 (TGF-β1) in cell supernatants. TGF-β1 at concentrations commensurate with those in the supernatants capably induces increases in intracellular reactive oxygen species (ROS) in unirradiated cells. Furthermore, the addition of supernatants from α-irradiated cells to unirradiated cells decreases cellular levels of TP53 and CDKN1A and increases CDC2 and proliferating cell nuclear antigen in the latter. Like the increased intracellular ROS bystander effect, this “decreased TP53/CDKN1A response” can be mimicked in otherwise untreated cells by the addition of low concentrations of TGF-β1. Our results indicate that αparticle-associated increases in cell growth correlate with intracellular increases in ROS along with decreases in TP53 and CDKN1A,and that these cellular responses are mechanistically coupled. As well,the proliferating cell nuclear antigen and CDC2 increases that occur along with the decreased TP53/CDKN1A bystander effect also would expectedly favor enhanced cell growth. Such processes may account for cell hyperplastic responses in the conducting airways of the lower respiratory track that occur after inhalation exposure to radon/radon progeny, as well as, perhaps, other ROS-associated environmental stresses.
Direct nuclear hits byα3particles like those emitted by 222radon and radon progeny and several other radionuclides, e.g.,241americium and 238,239plutonium, can cause DNA damage, mutation,genomic instability, and oncogenesis (1, 2, 3, 4, 5, 6). Mounting evidence, however, indicates that some biological effects can occur in cells the nuclei of which do not experience nuclear or even whole-cell encounters with the α particles. Hickman et al.(7), for example, reported that α particles induced accumulations of the tumor suppressor TP53 (p53) protein in rat lung epithelial cells in a higher percentage of the exposed population than that calculated to receive a direct nuclear traversal. Such increases in TP53 in cells the nuclei of which were unirradiated provided early evidence consistent with an extranuclear mechanism for the induction of DNA damage in the form of strand breaks (8). The apparently large target size that is indicated by investigations of αparticle-induced genomic instability (9, 10, 11) also suggests a relationship with extranuclear and perhaps extracellular DNA damage-inducing processes as well.
We and others (12, 13, 14, 15) have found that α particles in the circa 1 cGy region can induce DNA damage, as indexed by increases in SCEs, in the absence of direct nuclear or even whole cell hits. This effect has been associated recently with the generation of ROS in culture medium and the induction of intracellular ROS, particularly superoxide anions and hydrogen peroxide (16). Both the increased SCE and ROS responses can occur in unirradiated, normal human cells via the actions of SOD-inhibitable transmissible factor(s)present in α-irradiated medium and in the supernatants ofα-irradiated cells (14, 15, 16, 17). These and other investigations (17) implicate an important role(s) for ROS in mediating at least some effects of α particles, and they demonstrate well that biological responses to α particles can be induced indirectly in unirradiated cells, i.e., as“bystander” effects, to the same extent as found with directlyα-irradiated cells (14, 15, 16).
Most recently, Azzam et al. (18) extended upon the findings of Hickman et al. (7) by demonstrating that exposure of human fibroblasts to low doses of αparticles causes unexpectedly high increases in TP53 protein and downstream CDKIN1A (p21Waf-1) protein in cell populations in which only a low percentage of the cells experienced a nuclear traversal. Moreover, these investigators showed that the excessive occurrence of increases in TP53 and CDKN1A may involve direct cell-cell communications in that disruption of intercellular gap junctions eliminated the effect.
In the present study, we set out to examine the possibility that the bystander responses reported by Hickman et al.(7) and Azzam et al. (18) may be mediated, at least in part, by fluid-phase, soluble factors generated in response to α-irradiation. We unexpectedly observed that the transfer of supernatants from α-irradiated cells to unirradiated cells consistently caused decreases in basal levels of TP53 and CDKN1A in the latter populations, not increases. We also found that these decreases were accompanied by increases in PCNA and CDC2. In follow-up experiments, we identified TGF-β1 as a mediator of both theα particle-associated “increased intracellular ROS bystander response” we have described elsewhere (16) and the newly discovered “decreased TP53/CDKN1A bystander effect.” Furthermore,we report that these bystander responses correlate with enhancements in cell proliferation in a manner that suggests they collectively explain the phenomenon of cell hyperplasia that occurs in response to the inhalation of radon/radon progeny (19) and, perhaps, cell hyperplastic responses to other ROS-associated environmental stresses as well (20, 21, 22, 23). Conceivably, such increases in cell proliferation may play a critical role(s) in radon-induced carcinogenesis.
MATERIALS AND METHODS
Cells and Exposures to α Particles
Normal human diploid lung fibroblasts (HFL1) initially obtained from a human fetus (CCL 153; American Type Culture Collection, Rockville, MD)were routinely cultured in 75-cm2 tissue culture flasks in α-MEM (Life Technologies, Inc., Grand Island, NY)supplemented with 10% fetal bovine serum (Hyclone Laboratories, Inc.,Logan, UT). All cell cultures were incubated at 37°C in humidified 5% CO2/95% air. Cells were harvested from the flasks by trypsinization and seeded in 1.5-μm-thick Mylar-bottomed,∼30-mm-diameter culture dishes (24) at an initial density of 1 × 104 to 2 × 105 cells/dish for 1–3 days prior to exposure to α particles or other treatments in experiments that involved exponentially growing cells. Higher initial seeding densities or longer growth periods were used in other experiments, depending on experimental requirements. Fresh medium was added to the cultures prior to all irradiation, unless indicated otherwise. Cells were exposed toα particles at doses ranging from 1.0 to 19 cGy at room temperature using a collimated 238Pu α particle exposure system (25). The average energy of the α particles at the cell-Mylar interface is ∼3.5 MeV delivered at a dose rate of 3.65 cGy · s−1. Control HFL1 cells were sham irradiated. In all experiments, precautions were taken to ensure that the pH of the media did not change. Supernatants were transferred quickly and efficiently, and no obvious changes in the media were observed.
Intracellular Hydrogen Peroxide
Previous studies showed that the intracellular ROS response to αparticles includes both superoxide anions and hydrogen peroxide(16). In this study, we flow cytometrically assessed for putative hydrogen peroxide to comparatively index the intracellular generation of ROS (16). Untreated cells were simultaneously incubated for 30 min with supernatants from sham- orα-irradiated cells and/or other reagents along with DCFH-DA (1μ m). The cells were then harvested, centrifuged, and resuspended in the medium in which they were initially treated(16). After cleavage of the acetate moieties by cell esterases, the oxidative potentials of hydrogen peroxide, along with peroxidases, are able to oxidize the trapped diacetate to form fluorogenically active DCF (26). Flow cytometric data were collected on a Becton Dickinson FACSCalibur flow cytometer (Becton Dickinson, San Francisco, CA) using standard computer, optics, and electronics. A 15-mW, air-cooled argon-ion laser provided excitation at 488 nm, and a 530/30 nm band pass optical filter was used for measuring DCF fluorescence. Cells were distinguished from background events by their forward angle and orthogonal light scattering characteristics,and the cellular events were evaluated after applying a bitmap (gate). Fluorescence signals collected in this gate were analyzed by the Cellquest data analysis software to determine fluorescence differences between control and test samples.
TGF-β1 concentrations were measured in cell supernatants by ELISA(Genzyme, Cambridge, MA). The assay used does not show cross-reactivity with TGF-β2 or TGF-β3, nor does it show cross-reactivity for other cytokines, including platelet-derived growth factor, epidermal growth factor, fibroblast growth factor, tumor necrosis factor-α, and tumor necrosis factor-β. This ELISA kit requires activation of TGF-β by acidification prior to its measurement. To measure biologically active TGF-β1 samples, standards and controls must be acidified using HCL for 1 h. It is unclear at this time whether the changes observed in TGF-β are attributable to the release of latent TGF-β from the cell surface or enhanced production and release of TGF-β from the irradiated cells. In some experiments, neutralizing antibodies to TGF-β1 (R&D Systems, Minneapolis, MN) were added to cell culture media to selectively inhibit its activity. In other experiments, human recombinant TGF-β1 (R&D Systems) was added to cell cultures to mimic responses observed with supernatants from α-irradiated cells.
Cells were seeded at 1 × 104 per Mylar dish for 24 h prior to sham- or α (1 cGy)-irradiation or other treatments, as indicated. In other experiments, confluent HFL1 cells were irradiated with 0 or 1 cGy of α particles, and immediately or 1 h thereafter, we transferred their supernatants onto other unirradiated, exponentially growing HFL1 cells. Cells were harvested 1–3 days thereafter for cell counts.
SDS-PAGE and Immunoblotting
Cells were harvested by trypsinization and washed twice with PBS. Cells were counted and lysed in NP40 lysis buffer and incubated on ice for 30 min. Samples were centrifuged at 12,500 × gfor 20 min at 4°C, and the supernatants were collected. For direct immunodetection, equal amounts of protein (measured by Bio-Rad Protein assay) were loaded and resolved on 10% SDS Ready Gels (Bio-Rad) in a mini-gel apparatus (Bio-Rad, Hercules, CA). Resolved proteins were transferred to a nitrocellulose membrane (Amersham, Arlington Heights,IL) using a wet mini-transfer unit (Bio-Rad). Membranes were stained with 0.1% Ponceau S to verify loading and transfer efficiency. Membranes were blocked by overnight incubation at 4°C in blocking buffer (5% dry nonfat milk in 10 mm Tris, 150 mm NaCl, pH 7.2). Membranes were then incubated with relevant antibodies in blocking buffer for 1 h at 25°C and washed extensively with Tris-buffered saline (TBST; 0.05% Tween 20, pH 8.0). Blots were incubated with peroxidase-linked anti-rabbit/mouse immunoglobulin (Amersham) in TBST for 1 h at 25°C and then washed extensively with TBST. Blots were placed in ECL enhanced chemiluminescence (ECL) reagents (Amersham) for 1 min, followed by exposure to autoradiographic film (ECL film; Amersham) for the required time, which was developed in an automated film processor (Kodak). Antibodies used were as follows: DO-1 p53 antibody (1:1000 dilution;Santa Cruz Biotechnology, Inc., Santa Cruz, CA),p21Waf-1 mouse monoclonal (1:1000; Santa Cruz Biotechnology), CDC2 (p34cdc2; 3:1000; Santa Cruz Biotechnology), PCNA (2.5:1000; Santa Cruz Biotechnology), and antimouse immunoglobulin horseradish peroxidase secondary antibody(Amersham).
Unless indicated otherwise, data shown from each series of experiments are representative of results obtained from three and often more replicate experiments.
Low-Dose α Particles and the Decreased TP53/CDKN1A Bystander Response
Similar to the experimental design followed by Azzam et al.(18), confluent HFL1 cells were exposed to a 1-cGy dose ofα particles, the cells were harvested 4 h later, and their TP53,CDKN1A, and CDC2 contents were assessed by Western analysis. Sham-irradiated cells served as controls. Similar to the study reported by Azzam et al. (18), exposure of the cells to only 1 cGy of α particles resulted in markedly elevated levels in TP53 and CDKN1A and a decrease in CDC2 (Fig. 1 A). On the basis of our previous morphometric analyses(13), ∼7% of the cells in this study actually received one or more nuclear hits by the α particles.
We next investigated the possibility that a transmissible factor(s) may be induced by α particles that could contribute to the greater than expected TP53 and CDKN1A elevations. The increased intracellular ROS bystander response we reported previously (16), for example, conceivably could cause DNA damage, including DNA strand breaks (27, 28, 29), and thereby induce increased accumulations of TP53 and downstream CDKN1A via transcriptional activation by TP53 (30, 31). We irradiated confluent HFL1 cells with 1 cGy of α particles; immediately and 1 h thereafter,we transferred their supernatants onto other unirradiated confluent HFL1 cultures. Four h later, the cells were harvested, and their TP53 and CDKN1A levels were analyzed by Western analysis. Control cell populations were treated with supernatants from sham-irradiated HFL1 cells. In repeated experiments, the supernatants from α-irradiated cells consistently caused decreases in the basal levels of TP53 and CDKN1A in the unirradiated cells (Fig. 1,B). More pronounced decreases in the cell cycle regulating proteins were obtained with supernatants that were harvested from cells 1 h after exposure than immediately after exposure (Fig. 1,B). Furthermore, the decrease in TP53 observed at 4 h was persistent at least up to 24 h after addition of the supernatants (Fig. 1 C). Hence, a soluble, transmissible factor(s) in the supernatants ofα-irradiated cells induces a decreased TP53/CDKN1A bystander response, and the decreases in TP53 are relatively persistent. Moreover, such a transmissible factor(s) evidently progressively increases shortly after exposure to α particles. In further analyses,we found that mdm-2 protein levels, which can be up-regulated by TP53-mediated transcriptional activation (32), were also decreased during the decreased TP53/CDKN1A bystander response (data not shown). Similar experiments were performed using irradiated cell-free media transferred 1 h after irradiation onto fresh unirradiated cells. No changes in TP53/CDKN1A were observed with media alone,excluding the possibility of any involvement of media constituents in the observed effects.
Cell Growth in Response to Low-Dose α Particles and Supernatants from α-Irradiated Cells
Although it is clear that exposure to high concentrations of ROS can induce DNA damage, cell cycle arrests, senescence, and cell death(33, 34, 35), numerous reports indicate that a relatively low-level exposure to ROS can stimulate cell growth (33, 36, 37, 38, 39) by pathways that are not well understood. The reported promitogenic effects of low levels of ROS suggested to us that low-dose exposure to α particles may induce a ROS-associated enhancement in cell proliferation. To test this possibility, cultures of exponentially growing HFL1 cells were irradiated with 1 cGy α particles, and cell counts were obtained over a 3-day period. On the first day after irradiation, cell counts were reduced relative to those obtained with sham-irradiated cultures in a manner that is consistent with αparticle-induced cell cycle arrests (40). By day 3,however, cell numbers in the irradiated cultures were significantly higher than control values (Fig. 2 A).
In follow-up experiments, we found that the above promitogenic effect can occur as an α particle-associated bystander effect in unirradiated HFL1 cells. In these experiments, confluent HFL1 cells were exposed to 0 or 1 cGy α particles and further incubated for 1 h at 37°C. Supernatants were then harvested from the cells and transferred onto nonconfluent, exponentially growing HFL1 cells. Over a 3-day period thereafter, cells were harvested and counted. As illustrated in Fig. 2 B, cell numbers from cultures that received the supernatants from the irradiated cells were ∼135 and∼150% higher that control cell numbers on assay days 2 and 3,respectively. Thus, the supernatants from α-irradiated cells, which have been shown previously to induce intracellular ROS(16) and that coincidentally reduce TP53 and CDKN1A protein levels, also increase cell growth.
PCNA and CDC2 in HFL1 Cells Treated with Supernatants fromα-Irradiated Cells
PCNA is an essential DNA replication protein that is increased in abundance in proliferating cells (41, 42, 43, 44). In that low levels of TP53 activate PCNA gene expression(45), we investigated the possibility that the levels of PCNA in HFL1 cells treated with the TP53-decreasing supernatants fromα-irradiated cells may become up-regulated. In these same experiments, we also assessed how cellular levels of CDC2, a protein kinase that complexes with mitotic cyclins and is required for mitotic entry (46, 47, 48), might also be altered in a manner that would favor an enhanced state of proliferation. As shown in Fig. 3,A, a 2-fold increase in PCNA levels was observed in HFL1 cells as early as 6 h and up to 24 h after treatment with the supernatants from α-irradiated cells relative to those of cells treated with sham-irradiated supernatants. Increases in CDC2 protein levels were also observed at these times in HFL1 that had been incubated with the supernatants from α-irradiated cells (Fig. 3 B).
α Particle-induced Increases in Extracellular TGF-β1
As mentioned earlier, supernatants from α-irradiated fibroblasts induce excessive SCE and intracellular ROS bystander responses when transferred to unirradiated cells (16). That the αparticle-induced factor(s) survives freeze-thawing and is heat labile(14, 15) suggested to us that it may be a cytokine(s). Because other investigators have reported that a low concentration of TGF-β can induce the production of ROS in human fibroblasts(14, 15, 49, 50, 51, 52) and because low concentrations of TGF-βcan suppress both p53 and p21 mRNA levels and CDKN1A protein levels in human lung fibroblasts (53), we especially focused on how this particular cytokine might contribute to the bystander responses described herein. In the next experiments, confluent HFL1 cells were irradiated with 0.4–19 cGy doses of α particles, their supernatants were harvested over a 24-h period thereafter, and TGF-β1 concentrations in the supernatants were measured. Exposure of cells to the α particles resulted in early dose-independent, step-like function increases in TGF-β1 (to a level of ∼1 ng/ml) at the earliest postexposure time point (30 min; Fig. 4 A). The early increase in TGF-β1 evidently was attributable to a posttranscriptional and posttranslational mechanism, inasmuch as the elevated TGF-β1 concentrations measured as of 30 min after exposure remained relatively stable for at least 8 h after their initial increase. More dose-dependent, yet still modest, increases in TGF-β1 were observed in the supernatants over the remaining 8–24-h period after irradiation; we have yet to further investigate this latter component of the TGF-β1 response to α particles. Even so, it should be noted that the levels of TGF-β1 ultimately measured in the supernatants rose from a background starting concentration of ∼0.5 ng/ml to no more than ∼2 ng/ml. Although such increases in TGF-β1 concentrations do not appear to be dramatic elevations over control levels, the addition of 1 ng/ml human recombinant TGF-β1 to HFL1 cells in culture medium that already contains nearly half of this concentration can induce an intracellular ROS response (to be described). Hence, even relatively small but abrupt increases in TGF-β1 can have functional consequences.
The question arises as to what mechanism may account for the early increases in TGF-β1 in the supernatants of α-irradiated cells. Conceivably, the TGF-β1 increases could be related to oxidative processes associated with the extracellular generation of relatively short-lived ROS we have found in α-irradiated medium(16). Of relevance to this possibility, Barcellos-Hoff and Dix (54) have reported that latent TGF-β can be activated by ionizing radiation and other oxidizing conditions, Ehrhart et al. (55) have observed increases in TGF-β in vivo with radiation doses as low as 10 cGy, and other investigators (56, 57) have found that the activation of TGF-β occurs concomitantly with the generation of ROS. On the basis of these findings, experiments were performed using the antioxidant enzymes SOD and catalase to determine how fluid phase O22 and/or H2O2 generated by αparticles (16) may contribute to the early increases in TGF-β1. As illustrated in Fig. 4,B, increases in TGF-β1 in response to α particles can be inhibited totally by adding SOD to cells prior to exposure to the α particles. As well, the addition of catalase to the cell cultures before they were irradiated also eliminated increases in TGF-β1 (Fig. 4 B). That the removal of either O2 or H2O2 prevents the increased extracellular availability of TGF-β1 in response to α-irradiation may be explained by a requirement for both species, with, perhaps, the ultimate mediator of the TGF-β1 release being the formation of the highly reactive hydroxyl radical (·OH) via the Haber-Weiss reaction:O2 + H2O2transition metalO2 + ·OH + OH−.
TGF-β1 as a Mediator of the Increased Intracellular ROS and Decreased TP53/CDKN1A Bystander Responses
A low concentration of TGF-β, i.e., 1 ng/ml, can induce the intracellular production of ROS in fibroblasts (51, 52), as noted previously. In subsequent experiments, we assessed the possibility that TGF-β1 at concentrations commensurate with those in the supernatants of α-irradiated cells can induce an intracellular ROS response in our cells. Clearly, the addition of recombinant TGF-β1 to HFL1 at a low concentration causes an increase in the intracellular production of ROS (Fig. 4 C), whereas much higher concentration TGF-β1 diminishes ROS production. TGF-β1 may also play a role in regulating basal levels of ROS production in HFL1 cells in that the addition of anti-TGF-β neutralizing antibody to the cells markedly decreases their ability to convert DCFH to DCF. Regardless, the above findings indicate that the prompt increases in TGF-β1 in the supernatants of α-irradiated cells can account for the induction of the increased intracellular ROS bystander response.
To further confirm the involvement of TGF-β1 in the α-irradiated bystander response, neutralizing antibodies to TGF-β1 were added to the supernatants from either sham or α-irradiated cells, and the supernatants were then transferred onto fresh, unirradiated cells. Control experiments were done in parallel wherein cells were either sham or α-irradiated, followed by the transfer of their supernatants to fresh cells. Treatment of supernatants from α-irradiated cells with neutralizing TGF-β antibodies prior to addition to fresh cells completely inhibited the ROS bystander response, whereas an increase in the DCF fluorescence was observed in the cells treated with the supernatants from α-irradiated cells without TGF-β antibody (Fig. 4 D).
That TGF-β1 may also be involved in mediating the decreased TP53/CDKN1A bystander response is suggested by the recent report that low concentrations of this cytokine can cause decreases in p53 and p21 mRNA levels and CDKN1A cellular protein levels (53). To evaluate this possibility, HLF1 cells were either sham-irradiated or irradiated with 1 cGy α particles, or we incubated HFL1 with varying concentrations of human recombinant TGF-β1, i.e., 1 ng/ml(40 pm), 5 ng/ml (200 pm),10 ng/ml (400 pm), or 50 ng/ml (2 nm) for 4 h. The cells were then harvested and processed for analysis of their TP53 contents by Western blotting. Cell cultures that received no TGF-β1 additions served as controls. Similar to the supernatants from α-irradiated cells, low concentrations of TGF-β1, i.e., 0.5, 1.0, 5, and 10 ng/ml,caused decreases in the cellular levels of TP53 (Fig. 5,A). Higher concentrations of TGF-β1, on the other hand,caused increases in TP53 protein levels (Fig. 5 A).
To more directly assess whether the increased yet low concentrations of TGF-β1 in the supernatants of α-irradiated cells underlie the abilities of the supernatants to decrease TP53 protein levels in HFL1,cells were irradiated with 1 cGy α particles, and their supernatants were harvested after 1 h. The supernatants were then incubated with 120 μg/ml of neutralizing TGF-β1 antibodies 30 min prior to being added to fresh cells. As shown in Fig. 5 B, TGF-β1 neutralization prevented decreases in cellular TP53 and CDKN1A protein levels that otherwise occur in HFL1 cells treated with the supernatants from α-irradiated cells. In addition to demonstrating how TGF-β1 can affect TP53 protein levels, the above collective evidence suggests a link between the increased intracellular ROS bystander response and the decreased TP53 bystander responses. As well, the above results suggest a possible inverse relationship between cellular oxidant levels and TP53 protein levels. That is, a low concentration TGF-β1-mediated increases in intracellular ROS levels are associated with decreases in TP53 like those observed in HFL1 cells treated with supernatants fromα-irradiated cells, whereas decreases in intracellular ROS levels mediated by higher concentrations of TGF-β1 are associated with increases in TP53 protein levels.
As shown by others (18, 58, 59), exposure of cells toγ-rays and α particles down-regulates CDC2 levels in a TP53/CDKN1A-dependent manner. We speculated that CDC2 decreases that occur when TP53 protein levels increase in response to ionizing radiation may not occur in cells that are treated with low concentrations of TGF-β1, but instead, such treatment may actually cause increases in CDC2. To examine this possibility, which would be expected to favor cell growth, we measured CDC2 protein levels in the cells that were treated with 1–50 ng of human recombinant TGF-β1 for 4 h. As shown in Fig. 5,C, treating HFL1 with a low concentration of TGF-β1 indeed resulted in increases in basal CDC2 protein levels. High concentrations of TGF-β1, i.e., 50 ng/ml, on the other hand, caused decreases in CDC2 levels (Fig. 5 C), as might be expected under conditions when TP53 protein is increased.
Relationships among ROS, TGF-β1, and the Decreased TP53/CDKN1A Bystander Response
Given the correlation between the ROS and the decreased TP53/CDKN1A bystander responses, we next examined how extracellular and intracellular ROS may be involved in TGF-β1-mediated decreases in TP53 protein. In the first experiments, HFL1 were incubated with SOD(100 units/ml) or catalase (10 units/ml) 30 min prior to the addition of recombinant TGF-β1. Catalase inhibited the TGF-β1-mediated TP53 decrease in HFL1 cells, as did SOD (Fig. 6,A). We also pretreated HFL1 populations with the flavoprotein inhibitor DPI (5 μm; refs. 16 and60). The cells were then treated with recombinant TGF-β1(1 ng and 5 ng/ml) and harvested 4 h later for analysis of their TP53 contents. DPI, which inhibits the α particle-induced intracellular ROS response (16), as well as the production of H2O2 by surface membrane NADH oxidase (51), completely inhibited the TGF-β1-mediated TP53 decrease (Fig. 6,A). Furthermore, the TGF-β1 induced increase in CDC2 was inhibitable by pretreating the cells with SOD, as shown in Fig. 6 B, implicating the involvement of ROS not only in the TGF-β1-induced decrease in p53 but also in the TGF-β1-stimulated increase in CDC2. The above collective results suggest that extracellular H2O2 and intracellular O2 may both play roles in the decreased TP53 protein bystander response. Apparently extracellular O2 contributes not only to the early TGF-β1 increases in the supernatants of α-irradiated cells but also to the decreased TP53 bystander response. Hence, both increases in extracellular and intracellular ROS apparently play roles in mediating the decreased TP53 protein bystander response.
Cell Growth in Response to TGF-β1
If TGF-β1 is an important mediator of the α particle-induced bystander enhanced growth effect as our previous results indicate, then treatment of cells with low concentrations of TGF-β1 should have a promitogenic effect and thereby mimic the effects of supernatants fromα-irradiated cells. In the final set of experiments, cells were treated with varying concentrations of TGF-β1, i.e., 1, 5,10, and 50 ng/ml. Cells were harvested on days 1, 2, and 3, and cell numbers were counted. As expected, treatment of cells with 1 and 5 ng/ml of TGF-β1 resulted in increase in cell numbers as early as day 1, with a significant increase becoming unequivocally evident on day 2 and thereafter (Fig. 7). However, only a moderate increase in cell numbers was observed at the 10 ng/ml concentration, and major decreases in cell numbers were found with cells treated with 50 ng/ml of TGF-β1 (Fig. 7). These results indicate that low concentrations of TGF-β1 are promitogenic, whereas high concentrations inhibit cell growth, which correlates with the inversely related TP53 and ROS dose response observed with TGF-β1, i.e., low levels of TGF-β1 decrease TP53 protein levels and increase ROS levels, whereas high concentrations of TGF-β1 induce elevations in TP53 protein and a decreased ROS response.
In addition to their effects on directly irradiated cells, both low- and high-linear energy transfer ionizing radiation can affect nontargeted or unirradiated cells via soluble, extracellular factors (61, 62, 63, 64, 65) and perhaps by direct cell-cell communications (18). These bystander effects, however,remain largely phenomenological and generally lack mechanistic explanations, and it remains unclear exactly how important radiation-induced bystander responses are in terms of their relevance to human health. Even so, several lines of evidence point to the likelihood that at least some radiation-associated bystander effects occur in vivo (64, 66). Also, reported in vitro observations that bystander effects can result in increases in SCEs (12, 13), increases in oxidative metabolism, and DNA oxidative adducts (16, 17), alterations in the expression of cell cycle regulating proteins (7, 18),increases in chromosomal breaks (64), and chromosomal instability (67) suggest that they may play a role in mediating carcinogenic responses to ionizing radiation in vivo, aside from the mutational effects that occur in directly irradiated, targeted cells. In this report, we show that exposure of cells to a low dose of α particles can result in enhancements in cell growth, and that this promitogenic effect can be transmitted to unirradiated cells by incubating them with supernatants harvested from the irradiated cells. Our earlier findings and the results from the present study, along with observations made by other investigators,suggest that these proliferative responses can be explained in the context of ROS playing a central role in mediating a cascade of events that ultimately lead to enhancements in cell growth.
Our results indicate that the pathway to increases in cell growth begins with the generation of extracellular ROS attributable to the interactions of α particles with extracellular constituents, as we have described elsewhere (68). The ROS cause a prompt increase in availability of extracellular TGF-β1, a process that we have shown to be SOD and catalase inhibitable in a manner that implies hydroxyl radicals as the primary oxidant species for its initiation. The increases in TGF-β1, which others have shown can occur in response to even low-dose ionizing radiation and other oxidizing conditions (54, 55, 56, 57), tentatively arise from the rapid secretion of TGF-β1 and/or the release of TGF-β1 from betaglycan receptors and/or possibly the extracellular matrix(69, 70, 71). Regardless, the essential step function-like increases in TGF-β1 activates cell membrane-associated NADPH oxidase that is present in human fibroblasts (72, 73), which, in turn, increases the intracellular production of superoxide anions and hydrogen peroxide. TGF-β1 may additionally activate cell surface membrane-associated NADH oxidase and the release of extracellular and cell-permeating H2O2 as a later and perhaps sustaining event (51, 74); as reported previously (16), the intracellular ROS increases in response to α-irradiation have been noted to persist for at least 24 h after exposure. That TGF-β1 increases in the supernatants of irradiated cells can indeed mediate the activation of the metabolic production of ROS is well illustrated by the fact that the increased intracellular ROS bystander response can be mimicked by the addition of recombinant TGF-β1 to unirradiated cells, and the ability of the cytokine to produce the response can be totally inhibited by anti-TGF-β1-neutralizing antibodies.
Existing evidence suggests that the production of intracellular ROS by the above mechanisms is a key step in mediating the promitogenic effects of low-dose α particles. Unlike with high level exposure to ROS, relatively low-level exposure to ROS or transient exposure to ROS can stimulate cell growth (36, 38, 39, 75), and the addition of antioxidants such as SOD, glutathione, catalase, inter alia, or decreases in endogenous cellular levels of oxidants all have antiproliferative effects (38, 39, 76). Moreover, some investigators have reported synergistic effects of oxidants and mitogens on cell growth (33). Although the mechanism(s) involved in ROS-stimulated cell proliferation are likely multifarious, it is evident that a cells redox status can play fundamental roles in cell growth-associated signal transduction pathways (77, 78), in the regulation of the expression of genes whose products contribute to cell growth (74, 79, 80), and in posttranslational modifications of cell cycle-regulating proteins (81, 82).
Increases in ROS alone, however, may be only one component of the promitogenic response to α particles, given the finding that supernatants from irradiated cells cause significant reductions in the levels of TP53 and CDKN1A proteins in unirradiated cells, with the latter reduction in CDKN1A perhaps being a result of initial decreases in its transcriptional activator, TP53, a possibility that is currently being investigated. Regardless, both of these proteins are well recognized as negative regulators of cell cycle progression (31, 83), perhaps to some extent even when expressed under normal basal conditions. At least in vitro, a variety of cell types, including fibroblasts, that are null for TP53 or CDKN1A show higher proliferation rates than their wild-type counterparts(e.g., Refs. 73, 84, and 85). As well, our findings that increases in PCNA and CDC2 protein levels during the increased ROS and decreased TP53/CDKN1A responses underscore how numerous regulators of cell proliferation may be contributing to the overall promitogenic response to α particles.
Although we have yet to elucidate the ultimate mechanism(s) that underlie the decreased TP53 bystander response, our results indicate that it is likely linked to the TGF-β1-induced intracellular ROS response in that it can be eliminated, like the intracellular ROS and promitogenic bystander responses, by the addition of anti-TGF-β1 neutralizing antibodies to the supernatants of α-irradiated cells and by inhibition of the intracellular ROS response. We currently hypothesize that the diminished TP53 protein levels that occur as a bystander effect result from ROS-associated alterations in one or more processes that down-regulate TP53 levels in cells, which potentially could include the transcription, translation, and/or degradation, with,perhaps, transcription being the most affected, given the recent report that low concentrations of TGF-β can suppress both p53 and p21 mRNA levels and CDKN1A protein levels in human lung fibroblasts(53).
A question that arises as to how our findings that a low dose of αparticles causes a promitogenic effect in directly irradiated and unirradiated cells can be reconciled with the investigation reported by Azzam et al. (18), in which low-dose αparticles can induce increases in TP53 and CDKN1A in a manner that expectedly would disfavor cell growth. Both sets of data may best be accommodated by the existence of two different processes. In the study by Azzam et al. (18), evidence was obtained that is consistent with the transmission of TP53/CDKN1A up-regulating signals from directly irradiated cells to nearby cells via gap junctional communications. In our studies, we also observed greater than expected increases in TP53 and CDKN1A in cells that were irradiated with a low dose of particles, and this was associated with evidence of decreases in cell growth as of 24 h after irradiation. Thereafter, however, we observed growth rates in the irradiated populations that, over time, exceeded control rates. Conceivably, the bystander effects observed in our study may serve to reduce the effect of the cell-cell communicated increased TP53/CDKN1A response, or perhaps, the enhanced growth we observed occurred in cells that did not experience the increased TP53/CDKN1A bystander effect.
Finally, ample evidence shows that the DNA-damaging effects of excessive ROS can activate cell cycle checkpoints, induce a senescence-like state, or cause cells to undergo apoptosis and even necrosis in vitro (34, 38, 86). As we have reported earlier, at least higher doses of α particles, i.e., 19 and 57 cGy, than those used in the present study cause both G1 and G2 cell cycle arrests in HFL1 cells (40). On the basis of these findings, one might expect that exposure to ROS stresses, if anything,would curtail cell proliferation in vivo as well. Paradoxically, however, cells along the conducting airways often undergo enhanced proliferation, i.e., hyperplasia, in vivo in a background of ROS-associated stimuli, including inhalation exposure to radon/radon progeny (19, 20, 21, 22, 23). In this regard, our results provide a mechanistic explanation for such hyperproliferative responses. Furthermore, excessive cell proliferation, especially in a concurrent background of DNA-damaging oxidative stress, expectedly could contribute to carcinogenic processes(87, 88, 89, 90, 91). How our overall results may extend to other important phenotypes, such as bronchial epithelial cells, remains to be determined.
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.
Supported in part by the Los Alamos National Flow Cytometry Resource (NIH Grant p41-RR01315), NIH Grant CA82598, and a grant from the Office of Biological and Energy Research, United States Department of Energy. The work was conducted under the auspices of the United States Department of Energy.
The abbreviations used are: α, high linear energy transfer α; SCE, sister chromatid exchange; ROS, reactive oxygen species; SOD, superoxide dismutase; PCNA, proliferating cell nuclear antigen; TGF, transforming growth factor; DCFH-DA,2′,7′-dichlorofluorescin diacetate; DCF, 2′,7′-dichlorofluorescein;DPI, diphenyleneiodonium.
We express our gratitude for the technical assistance provided by Yolanda Valdez, Rita Svensson, Agoyo Talachy, and Dr. P. K. Narayanan, SmithKline Beecham Pharmaceuticals, King of Prussia, PA.