HSP27 phosphorylation increases after 45°C or 41°C heat shocks but not after non-thermal TDMA or GSM exposures

Abstract

Purpose: Experiments with cultured HeLa, S3 and E.A. Hy296 cells were performed to determine if exposure to acute (30 min at 45°C) or chronic (2 h at 41°C) heat shocks or to non-thermal exposures of radiofrequency radiation (RF) induce changes in HSP27 phosphorylation.

Materials and methods: The radiofrequency (RF) exposures used in this study were 847 MHz time division multiple access modulated (TDMA) at a specific absorption rate (SAR) of 5 W kg⁻¹ for 1, 2 or 24 h or 900 MHz GSM modulated (GSM) at a SAR of 3.7 W kg⁻¹ for 1, 2 or 5 h. HSP27 phosphorylation was evaluated by resolving the various phosphorylation forms using two-dimensional gel electrophoresis measuring the relative amount of each by densitometry. Alternatively, an antibody specific for phosphorylated HSP27 was used to detect changes in HSP27 phosphorylation levels. All heat shock and RF exposure conditions were analysed simultaneously along with a matched incubator control sample. Each experiment was repeated three times.

Results: Following heat shock, the degree of phosphorylation of HSP27 varied with the heat dose, with acute hyperthermia (45°C) having an increased proportion of higher phosphorylated forms. Exposure of HeLa S3 cells to 5 W kg⁻¹ TDMA for 1, 2 or 24 h did not induce significant differences in the levels of HSP27 phosphorylation compared to incubator control or sham. Exposure of E.A. Hy926 cells to 3.7 W kg⁻¹ 900 MHz GSM for 1, 2 or 5 h did not induce significant differences in the levels of HSP27 phosphorylation compared to sham exposed.

Conclusions: Acute and moderate hyperthermia significantly increase HSP27 phosphorylation, but there was no significant change in the levels of HSP27 following non-thermal exposure to TDMA and GSM modulated RF radiations.

• HSP27 phosphorylation
• acute heat shock
• moderate heat shock
• RF exposure

Introduction

Recent concern regarding the effects of radiofrequency (RF) radiation on biological systems is motivated by the large number of individuals (>10⁸ worldwide) exposed to the RF signals from cellular phones [1–3].1,2 The specific issue is whether there are any unknown, non-thermal, deleterious effects of RF exposure. Concerns would be magnified if RF radiation could conceivably contribute to a carcinogenic or other pathological process. Conversely, data demonstrating the lack of either non-thermal, carcinogenic or other potentially harmful effects would assure both the public and industry that non-thermal exposures to RF radiation have no or minimal risk. However, a recent report related an observed RF effect on HSP27 to a possible carcinogenic process [4], thus opening the possibility that stress responses, following RF exposure, should be addressed.

Operationally, the carcinogenic process is a multi-step process involving genetic and epigenetic changes. If RF radiation contributes to a carcinogenic process, it would have to result ultimately in gene structure or gene regulation changes or both. Most known carcinogens cause changes via genotoxic effects. Currently, there is no conclusive evidence that non-thermal RF exposures have genotoxic effects. Although potentially genotoxic effects have been reported [5–9] these findings have failed confirmation both in vivo and in vitro [10–19]. The situation is also controversial regarding potential epigenetic effects of RF radiation on changes in gene expression. Some workers have speculated that RF induced changes in HSP27 phosphorylation could lead to carcinogenic transformation [4]. If this speculation were true, the proposed carcinogenic process would have to involve an epigenic mechanism because most heat shock responses are not directly genotoxic. Many cellular stresses including heat shock induce epigenetic changes that are manifest in terms of gene expression changes. Thus, the reported effects of RF on gene expression should be considered. PC12 pheochromocytoma cells (rat neural origin) treated with nerve growth factor and exposed to time domain multiple access (TDMA) radiation under various combinations of slot average power densities (0.09, 0.9 and 9 mW cm⁻²) and durations (20, 40 and 60 min) showed no statistically significant RF effects on oncogene expression except for a 38% decrease in c-jun following cell exposure at 9 mW cm⁻² for 20 min [20]. Neither FDMA nor CDMA at a SAR of 0.6 W kg⁻¹ had any significant effect on proto-oncogene expression in C3H10T 1/2 during transit from quiescence to exponential growth. Moreover, those exposures did not affect the binding activity of transcription factors AP-1, AP-2 and NFkB during this transition or during the transition from exponential growth to plateau phase [21]. While the expression of c-jun and c-myc were not significantly affected during the transition from proliferation to plateau phases of growth, the levels of c-fos mRNA were reported to be 2-fold higher in the cells exposed FDMA and 1.45-fold higher in the cells exposed to CDMA than in the sham exposed cells [21]. Follow-up studies failed to confirm these changes in c-fos expression [22]. Further, there was no detectable activation of HSF-1 following exposure to 0.6 W Kg⁻¹ of either CDMA or FDMA [23]. Thus, it appeared that these RF signals did not induce a general stress response.

It has been reported by Leszcynski et al. [4], [24] that HSP27 phosphorylation increases following a 1 h exposure to 900 MHz GSM (3.7 W kg⁻¹). A similar study by Miyakoshi et al. [25] reported that 1950 MHz continuous wave RF radiation at 10 W kg⁻¹ for 1 or 2 h reduced the levels of phosphorylated HSP27 in MO54 cells. It is known that phosphorylation of HSP27 is an early step in the responses of cells to numerous stresses [26–33]. HSP27 plays a critical role in thermotolerance [34] and heat resistance [35], [36]. In addition, HSP27 contributes to the regulation of apoptosis [37], [38]. However, it was not clear, in these reports [4], [24], [25] that the reported RF induced changes in HSP27 phosphorylation activated any of the stress response pathways involving HSP27 phosphorylation. The possibility that HSP27 was phosphorylated without activating a known stress response pathway suggests that this phosphorylation occurred via an unknown pathway, for which there is no evidence except for the reported RF effect. Given that heat shock proteins, HSP70 and HSP27, are elevated in several human tumours, it is important that the possible role of RF exposure on HSP27 expression be addressed. Prior to searching for an unknown stress response pathway, it was decided to first confirm the observation that RF exposures induce HSP27 phosphorylation. Thus, the goal of the study presented herein is to determine if non-thermal exposures to radiofrequency radiation induce the phosphorylation of HSP27. In addition, this study compared the effects of moderate (41°C for 2 h) and acute (45°C for 30 min). Considering, for the purpose of experimental design that RF exposure could be inducing some stress response [4], [24], this approach would allow one to determine if there are any differences between non-thermal, moderate-chronic and acute short duration stress effects on the phosphorylation of HSP27.

Materials and methods

Cell lines maintenance and experimental procedures

HeLa S3 cells (ATCC, Manassas, VA) were maintained routinely in tissue culture flasks in α-MEM medium + 10% newborn calf serum supplemented with penicillin/streptomycin at 37°C in a 5% CO₂ incubator. Forty-eight hours prior to the experiment cells were exponentially grown in T75 tissue culture flasks. The cell cultures then were exposed to TDMA (5 W kg⁻¹; 1, 2 or 24 h) or heat shock (2 h at 41°C or 30 min at 45°C). After exposure, cells were harvested by washing the T-75 flasks three times with isotonic buffered sucrose (250 mM sucrose; 10 mM Tris pH 7.0). The cells then were lysed with 7 M Urea, 2 M thiourea, 2% CHAPS solution. The lysates were analysed for HSP27 levels and phosphorylation status as described below. HeLa cells were used because there is an extensive database on the effects of heat shock on these cells.

E.A. Hy926 endothelial cells (a gift from Dr Cora-Jean S. Edgell, University of North Carolina, Chapel Hill) [39] were grown in DMEM plus 10% foetal bovine serum and penicillin/streptomycin. Exponentially growing cultures were plated in T-75 flasks 48 h prior to exposure to GSM (3.7 W kg⁻¹; 1, 2 or 5 h) or heat shock. The cells were then harvested and samples prepared as described above for HeLa S3 cells. E.A. Hy926 cells were used to provide a direct comparison of this work with that of Leszcynski et al. [4].

RF exposure system and SAR

The radial transmission line (RTL) irradiator developed at Washington University in St Louis was used to expose cells in T-75 culture flasks to TDMA or GSM modulated RF radiation. A detailed description of this device and its environmental controls has been reported in several publications [40–43]. Briefly, the RTL is a waveguide formed by two horizontal parallel electrically conducting plates separated by 43 mm. The top plate is made of a lightweight composite hinged to a 6 mm thick bottom AL plate to facilitate access to the interior of the radiator. The bottom plate provides support and a highly thermally conductive surface that minimizes temperature gradients among flasks. A centrally located conical AL antenna emits transverse electromagnetic (TEM) waves, which are propagated radially outward through air and are terminated by an annulus of RF-absorbing foam material braced by a thin lamina of perforated aluminum. The lamina is located 533 mm from the antenna's centre. A total of 16 T-75 culture flasks can be distributed angularly around the antenna with their centres located at a radius of 292 mm. To boost and homogenize the SAR distribution, each flask was dielectrically loaded with a 3 mm thick shim of alumina (Al₂O₃) ceramic that fitted snuggly under the skirt of the flask [18], [43]. The nominal nearest and farthest distances of the cell layer from the antenna's centre were 247 and 338 mm, respectively. A sham RTL was identical but not connected to an RF source. The above power settings were selected to produce a nominal time-averaged cell-layer SAR of 5.0 ± 2.1 W kg⁻¹ (mean ± SD) for both signals because of variations in power density resulting from the angular position of the flask [40]. This SAR value is based on a series of comprehensive measurements and does not include the potential uncertainty in the measurement of power, which is usually estimated at ± 10%. The degree of SAR inhomogeneity reflected by the standard deviation value is typical for T-75 flasks exposed in a RTL as validated by independent finite differences time domain (FDTD) simulation studies [43], [44]. A computed histogram of the SAR distribution over the bottom surface of a flask exposed to FDMA in a RTL can be found in Vijayalaxmi et al. [18].

In this study, the powered and sham RTLs were each loaded with T-75 culture flasks containing 40 ml of media. The larger than usual amount of culture medium (i.e. 40 ml, medium height of 5 mm) was found to reduce SAR non-uniformities. The flasks for simultaneous sham exposures were handled in the same way as the flasks exposed to RF radiation. The temperature of the culture media during the 24 h exposure to RF radiation and sham exposure was maintained at 37 ± 0.3°C with a combination of temperature control systems for the room and for the RTLs [40].

Hyperthermia exposure

For acute heat shocks cell cultures were capped, then immersed in a temperature controlled waterbath (45 ± 0.5°C) for the indicated time (30 min). For chronic heat shocks cell cultures were placed in a CO₂ incubator set at 41°C. This incubator has been thermally mapped using Luxtron probes as described previously [45]. Cell cultures were maintained at 41 ± 0.5°C for the indicated time (2 h).

Analysis of HSP27 phosphorylation

Sample preparation; 75 µg total protein as measured by the Bradford assay (BioRad, Hercules, CA), was diluted to 125 µl with 7 M urea, 2 M thio urea, 2% CHAPS. Bio-Lyte 3/10 ampholytes (BioRad) and reducing agents were added. Resolution of proteins by 2D SDS-PAGE 7 cm, 4–7 IPG Ready Strips (BioRad) were used in the first dimension. 4–15% ready gels (BioRad) were used for the second dimension. Visualization of HSP27 by immunoblotting proteins were transferred to PVDF membranes by western blotting and then probed with anti-HSP27 antibodies, StressGen SPA-800, SPA523 for non-phosphorylated and phosphorylated HSP27 (StressGen, Ann Arbor, MI).

Quantification and statistical analysis

To quantify the HSP27 spots, immunoblots were digitized as TIFF files. The optical densities of the HSP27 spots were measured by ImageQuant software (Molecular Dynamics). The Optical density values from three experiments were averaged. Statistical significance (p ≤ 0.05) was determined using either the student's t-test or 95% confidence limits. The t-test was used to test the significance in Figures 2 and 4. The 95% confidence limits were used in Figure 1 due to the large frequency of zero value for several of the higher phosphorylated forms of HSP27.

Figure 1. Analysis of HSP 27 phosphorylation detected by two dimensional immunoblotting. (a) After the indicated experimental manipulations HeLa S3 cells were lysed. The proteins from whole cell extracts were resolved by 2 D polyacrylamide gel electrophoresis and probed with an antibody that recognizes all forms of HSP27. An equal protein was loaded on each gel. Results from a typical experiment are shown. (b) Western blots from three experiments like the one shown here were quantified and averaged. The average ± SD is plotted for each experimental condition. The indicated experimental conditions are: control, C; TDMA exposed for 2 or 24 h, T2 and T24, respectively; sham exposed for 2 or 24 h, S2 and S24, respectively; heated at 41°C for 2 h, 41; and heated at 45°C for 30 min, 45.

Results

Using 2D polyacrylamide gel electrophoresis, one was able to detect five isoforms of HSP27 in HeLa cells that had been heat shocked at 45°C for 30 min (Figures 1 and 3). One form consisted of the non-phosphorylated protein, which was present under all experimental conditions studied. The remaining four forms represented various phosphorylation states of HSP27 consistent with previous reports [29], [31], [32], [46], [47]. Two of the isoforms (3 and 4, in order of increasing charge) were difficult to resolve routinely. Thus, Figure 1(a) appears to resolve only four isoforms. The analysis of RF and heat effects was performed by quantifying the non-phosphorylated and three phosphorylated forms numbered 1–4 on Figure 1(b), with number one spot containing two unresolved forms (2 and 3). The unphosphorylated form and form 1 were detectable under all experimental conditions studied. While forms 2 and 3 appeared sporadically in some sham and control samples, form 4 was consistently present in extracts from cells heated at either 41 or 45°C. However, forms 2 and 3 were consistently present only in cells heated at 45°C. It should be noted that 2 h at 41°C is essentially a non-lethal heat shock [48], while 30 min at 45°C reduces cell survival to ∼ 25% [49]. Thus, heat shock clearly induced the phosphorylation of HSP27, with the acute heat shock inducing phosphorylation to a greater extent than the moderate heat shock. Exposure of HeLa cells to TDMA RF radiation for 2 or 24 h did not appear to induce any additional HSP27 phosphorylation above sham.

After both heat shocks, the fraction of unphosphorylated HSP27 decreased significantly (Figure 1(b)). Further, the fraction of forms 2 and 3 were significantly higher after 30 min at 45°C, than after 2 h at 41°C. However, the distribution of phosphorylated HSP27 among different phosphorylation states made comparing the amount of phosphorylated HSP27 to the amount of unphosphorylated difficult due to the fact that the distribution of HSP27 across phosphorylation states was changing with temperature. Therefore, to facilitate analysis of the fraction of total HSP27 that was phosphorylated, all of the phosphorylated forms were pooled into one total (Figure 2). For both TDMA exposures (2 h and 24 h) the extent of HSP27 phosphorylation was less than or equal to its corresponding sham. However, both heat shocks tested showed an extent of HSP27 phosphorylation greater than the incubator control. The incubator control and the 2 h sham showed the same extent of HSP27 phosphorylation. However, the 24 h sham showed a higher extent of HSP27 phosphorylation than the incubator control. This result appears to be due to experimental variation because the 2 h sham was intermediate between the two and not significant from either.

Figure 2. The effect of heating and TDMA exposure on the fraction of phosphorylated HSP27 in HeLa cells. The data shown was analysed for the fraction of Hsp27 that was phosphorylated by summing overall phosphorylated forms. The experimental conditions are the same as indicated in Figure 1. The results of t-tests show that none of the TDMA exposed samples were different from their repective shams. Both the 41 and 45°C samples were significant compared to control for both phosphorylated and unphosphorylated HSP27.

Since the above result appeared to be inconsistent with those reported by Leszczynski et al. [4], [24], these studies were repeated using a system and an approach that were similar to that reported by Leszczynski et al. [4]. Human endothelial cells, E.A. Hy926, in culture were exposed to GSM (3.7 W kg⁻¹) for 1, 2 or 5 h. In addition, it was modified to utilize an antibody specific to phosphorylated HSP27. The first step was to verify the specificity of the antibody. The results shown in Figure 3 demonstrate that the antibody for total HSP27 detected all five of the isoforms, while the antibody for phosphorylated HSP27 cross-reacted with the four most positively charged isoforms. Thus, the phosphorylated-HSP27-specific antibody did not cross-react with unphosphorylated HSP27.

Figure 3. Specificity of an antibody specific for phosphorylated HSP27. The HSP27 isoforms from HeLa cells that had been heated at 45°C for 30 min were resolved by 2D-polyacrlamide gel electrophoresis. The resolved proteins on three separate gels were probed with an antibody for total HSP27 and an antibody for phosphorylated HSP27, as indicated on the figure.

The two antibody technique was used to determine if exposure to 900 MHz GSM (3.7 W kg⁻¹) for 1, 2 or 5 h induced HSP27 phosphorylation. Positive controls, 2 h at 41°C and 30 min at 45°C were included to ensure that the methods used will detect known changes in HSP27 phosphorylation and to provide a comparison between the two methods for detecting HSP27 phosphorylation. The western blots from three independent experiments are shown in Figure 4(a) and the quantification and averaging of these blots is shown in Figure 4(b). Because actin, the total HSP27 and phosphorylated HSP27 are measured on separate western blots the amounts were normalized to their own control values. The actin levels did not change and the levels of total HSP27 were not significantly different for the various experimental treatments. This result shows that the loading for the various lanes on the western blots is equivalent. However, the levels of phosphorylated HSP27 were more than 4-fold elevated after the 2 h at 41°C heat shock and almost 6-fold elevated after the 30 min at 45°C heat shock. In contrast to the samples from the heat shocked cells, there were no significant changes in the levels of phosphorylated HSP27 between the control, the sham exposed and the GSM exposed for 1, 2 or 5 h. This result shows that GSM radiation at 3.7 W kg⁻¹ did not induce measurable changes in HSP27 phosphorylation.

Figure 4. Analysis of HSP27 phosphorylation detected by Western blotting. (a) E.A. Hy296 cells were either heat shocked or exposed to GSM modulated RF radiation as described in the text. Total HSP27, phosphorylated HSP27 and actin were detected in whole cell extracts by Western blotting. Western blots from three experiments are shown. (b) Actin, total HSP27 and phosphorylated HSP27 were quantified, averaged and normalized to their control values. The normalized, average values ± SD are plotted as follows: Control = C; GSM for 1, 2 & 5 h = G1, G2 & G5, respectively; Sham for 1, 2 & 5 h = S1, S2 & S5, respectively; 41°C for 2 h = 41; and 45°C for 30 min = 45.

Discussion

The goal of the present study presented was to determine if non-thermal exposures to RF radiation induce the phosphorylation of HSP27. Clearly, the phosphorylation of HSP27 increased after either an acute, short duration heat shock or a moderate, chronic heat shock. However, exposures to TDMA for 2 or 24 h or to GSM for 1, 2, or 5 h did not alter the levels of HSP27 phosphorylation. Thus, the work presented herein did not confirm, in part, the reports of Leszcynski et al. [4], [24] and of Miyakoshi et al. [25]. Although this study tried to duplicate the methods used in Leszcynski et al. as much as practically possible, there are differences that could account for the differences in the experimental outcomes. The main difference in methodology is that Leszcynski et al. [4] used ³²P-labelling to detect phosphorylated HSP27, which measures the HPS27 phosphorylated during the time that the label is available. In the Leszcynski et al. study cells were given ³²P during the exposure time. Thus, the phosphorylation of HSP27 that is measured by this method will depend upon the rate of HSP27 phosphorylation and the time required for the exogenous ³²P to equilibrate the intra-cellular phosphorus pools. By measuring the amount or fraction of phosphorylated HSP27 by specific immunoblotting or by resolving the phosphorylated HSP27 forms by 2D gel electrophoresis, these measurements and those of Miyakoshi et al. [25] are independent of phosphorylation rates and endogenous pools. One overall interpretation of these results and those of Leszcynski et al. [2], [24] is that the latter group could have observed a small, transient perturbation in the rate of HSP27 phosphorylation or in the rate of ³²P equilibration in the intra-cellular pools that did not affect significantly the ultimate level of HSP27 phosphorylation. The results reported in Leszcynski et al. [4] are consistent with the reduction of HSP27 phosphorylation observed after a 1 h RF exposure is a transient effect, because there was no observed effect of a 2 or 5 h RF exposure. Thus, any depression of the rate of HSP27 phosphorylation during the first hour of RF exposure was swamped out by the HSP27 phosphorylation during the second hour of RF exposure, thereby having little effect on the fraction of HSP27 that was phosphorylated. This argument is supported by Miyakoshi et al. [25] who reported that at 1 or 2 h after 1 or 2 W kg⁻¹ of 1950 MHz CW there was no difference in HSP27 phosphorylation between sham and exposed samples. However, differences between the reported inhibition of HSP27 phosphorylation at 1 and 2 h during a 10 W kg⁻¹ 1950 MHz CW exposure [25] was inconsistent with the lack of effect at 2 and 5 h during a 5 W kg⁻¹ GSM exposure. This difference could be due to the difference in cell lines and in a higher SAR. However, the level of reduction in phosphorylated HSP27 in the report by Miyakoshi et al. [25] was 20–25% of a low initial level suggestive of a small effect. In contrast, both acute and chronic heat shock did increase the level of HSP27 phosphorylation 4–6-fold. The magnitude of change in HSP27 phosphorylation observed after a known stress, i.e. heat shock, compared with the small changes reported after RF exposure, suggests that the small reduction in the rate of HSP27 phosphorylation, reported by Leszcynski et al. [4] or in levels after a high SAR reported by Miyakoshi et al. [25] are not biologically significant.

Several other studies suggest that non-thermal RF exposures do not induce a biologically significant stress response involving HSP70 [23], [50]. Responses of cells to environmental stresses include changes in the expression of stress responsive genes, in the levels of stress responsive proteins and/or in the phosphorylation status of signalling proteins. The results of gene expression studies of specific genes or groups of genes are mixed, with some suggesting changes [20], [21], [27], [51] while others find no effect [22], [23] following RF exposure. Further, the results of studies of using genome wide screening for expression changes are also mixed, with some reporting changes [52] while others found no changes [53] after RF exposure. Protein modifications and protein levels can not be detected directly as changes in gene expression. However, such changes are often parts of pathways that lead to changes in gene expression or occur in response to gene expression changes. Thus, one would expect that the results of proteomic studies would also find mixed results. This proved to be true for the studies of HSP27 phosphorylation by Leszcynski et al. [4] that followed changes detected by a proteomics screen [24], because the study by Miyakoshi et al. [25] and the present study found results that differ from those reported by Leszcynski et al. [4]. Thus, the results reported in this report, in the context of other studies, suggest that non-thermal RF do not induce a stress response.

In contrast to the lack of an effect of RF exposure on HSP27 phosphorylation both heat shocks studied caused a significant increase in the fraction of HSP27 that was phosphorylated. Interestingly there were differences due to heat shock in HSP27 phosphorylation, as seen in the ratio of phosphorylated to nonphosphorylated, the relative increase in phosphorylated and the relative fraction of total HSP27 represented by each of the various phosphorylated forms. The difference in the ratio of phosphorylated to non-phosphorylated was significantly different between sham and both 41 and 45°C (Figures 2 and 4(b)). However, the relative increase in total phosphorylated HSP27 was only marginally different between 41 and 45°C and not significant (Figures 2 and 4(b)). Since the thermal dose from the 2 h at 41°C heat shock is ∼1/8 that of the 30 min at 45°C heat shock, it is surprising that there was not more of a difference in the relative increase in phosphorylated HSP27 after these heat shocks. However, the relative fraction of total HSP27 represented by phosphorylated forms (2 and 3, Figure 1) is significantly different between the 41 and 45°C heat shocks. Thus, the pattern of HSP27 phosphorylation rather than the ratio of phosphorylated to non-phosphorylated HSP27 may be a marker for moderate vs acute heat shocks.

In summary, exposure of HeLa S3 cells to 5 W kg⁻¹ TDMA for 1, 2 or 24 h did not induce significant differences in the levels of HSP27 phosphorylation compared to incubator control or sham. Further exposure of E.A. Hy926 cells to 3.7 W kg⁻¹ 900 MHz GSM for 1, 2 or 5 h did not induce significant differences in the levels of HSP27 phosphorylation compared to sham exposed. Thus, in contrast to chronic and acute heat shocks RF exposure did not appear to alter the phosphorylation of HSP27 at a biologically significant level. However, the possibility that subtle changes in HSP27 phosphorylation rates or levels are produced by RF exposure remains open due to the mixed results in reported studies. It is interesting to note in view of the mixed results with RF effects on HSP27 phosphorylation that after 60 Hz electromagnetic field exposure (an area of study where mixed results are often found) there was no effect on HSP27 phosphorylation, localization or expression [54].

Acknowledgements

This work was supported in part by a research contract from the Mobile Manufacturers Forum (MMF) and in part by our Program Project grant (PO1 CA104457-02) from the National Cancer Institute of the US Department of Health and Human Services.

Notes

Notes

1. Health Effects from Radiofrequency Electromagnetic Fields, Documents of the NRPB, Volume 14 No. 2, National Radiological Protection Board, 2003. (http://www.nrpb.org/publications/documents_of_nrpb/abstracts/absd14-2.htm).

2. Cleveland RF, Ulcek JL. Questions and Answers about Biological Effects and Potential Hazards of Radiofrequency Electromagnetic Fields, OET Bulletin 56, 4th ed., FCC, Washington D.C., 20554, 1999. (http://www.fcc.gov/Bureaus/Engineering_Technology/Documents/bulletins/oet56/oet56e4.pdf).