Rodent cell transformation and immediate early gene expression following 60-Hz magnetic field exposure.

Some epidemiological studies suggest that exposure to power frequency magnetic fields (MFs) may be associated with an elevated risk of human cancer, but the experimental database remains limited and controversial. We investigated the hypothesis that 60-Hz MF action at the cellular level produces changes in gene expression that can result in neoplastic transformation. Twenty-four hour 200 microT continuous MF exposure produced negative results in two standard transformation systems (Syrian hamster embryo cells and C3H/10T1/2 murine fibroblasts) with or without postexposure to a chemical promoter. This prompted a reexamination of previously reported MF-induced changes in gene expression in human HL60 cells. Extensive testing using both coded and uncoded analyses was negative for an MF effect. Using the same exposure conditions as in the transformation studies, no MF-induced changes in ornithine decarboxylase expression were observed in C3H/10T1/2 cells, casting doubt on a promotional role of MF for the tested cells and experimental conditions. ImagesFigure 1.Figure 2. AFigure 2. BFigure 2. CFigure 2. DFigure 3. AFigure 3. BFigure 4.Figure 5. AFigure 5. BFigure 5. CFigure 5. DFigure 5. EFigure 6. AFigure 6. BFigure 6. CFigure 6. DFigure 6. EFigure 7.Figure 8. AFigure 8. BFigure 8. CFigure 9.Figure 10. AFigure 10. B

Adverse health effects, induding cancer, have been positively associated with human exposures to power frequency (50 or 60 Hz) magnetic fields (MF) in numerous epidemiological studies. Childhood leukemia, brain cancer, breast cancer, and certain neurological and reproductive effects among residentially exposed children and among occupationally exposed adults were reported in some studies; other studies were negative for these effects (1,2). The possible scientific foundation for these effects is uncertain. Results of animal and cellular studies that investigated MFs for possible cancer-related activity were either contradictory or not replicated, as were results concerning effects on cell growth and differentiation and modulation of gene transcription and/or translation (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14).
In the experiments reported here, we tested the hypothesis that changes in gene expression occurring under MF exposure may result in neoplastic transformation. In vitro studies of neoplastic transformation are very relevant to the purposes of carcinogenic risk assessment (15)(16)(17). They permit investigation of the mechanisms of cancer initiation, promotion, and progression without interference from other in vivo factors at the tissue and organ-system levels. Available in vitro assays are more practical and precise than are studies on exposed animals. Most importantly, in vitro neoplastic transformation correlates well with cancer induction in humans in the sense that few, if any, tested agents have yielded false positive results (1$).
For the purposes of the current study, we used two well-characterized embryonic cell models, based on murine C3H/1OT1/2 cells or Syrian hamster embryo (SHE) cells (19)(20)(21)(22), to examine the transforming potential of 60-Hz MFs. The C3H/10Ti/2 system is suitable for initiation-promotion protocols (16,17,(23)(24)(25)(26)(27)(28); we previously described twostage protocol for investigating 12-atetradecanoylphorbol-13-acetate (TPA) promotion of 2.45-GHz microwave-and X rayinduced transformation (24)(25)(26)(27). In addition to our long established experience with the C3H/10T1/2 assay, this assay has been chosen in particular because it is widely used and exhibits a low spontaneous background (19) and no sensitivity to heat (29)(30)(31); in the case ofother endpoints, such as enzymatic or gene activity or proliferation, thermal effects might confuse a possible MF effect. The SHE system is suitable for studying diverse agents (chemical, viral, hormonal, and physical irritants), including weak and epigenetic carcinogens (15)(16)(17)(18)(19)(20)(21)(22). It should be noted that there are no analogous assays based on human cell lines. Our transformation studies were performed with long continuous exposures at 200 jiT (equal to 2 G), a field level in the typical upper range of power line and household exposure (11). To our knowledge, only one other study of neoplastic transformation in vitro has been reported, and this mainly concerns the effects of intermittent coadministration ofTPA and MF on preinitiated C3H/1OTi/2 cells (8).
Because our working hypothesis is based on changes in gene expression, we needed to address contradictory reports concerning the ability of MFs to increase levels of a number of gene transcripts (32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43). Our second series of studies was designed to confirm with solid data whether MFs induce significant changes in transcription of seven genes. Studies of c-myc, cfos, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 1-actin, and two ribosomal genes were performed using HL60 cells (44,45), short exposure times, and several flux densities matching the exposure conditions of Goodman et al. (34). Three replication experiments have recently been reported (12)(13)(14), but they were more limited in terms of both the number of genes and MF conditions studied. Our ornithine decarboxylase (ODC) transcription studies using C3H/10T1/2 cells were performed at 200 jiT to elaborate the C3H/10T1/2 transformation results.
The c-myc, c-fos, GAPDH, 1-actin, ODC, and ribosomal genes play an important role in normal cell physiology, including metabolism, protein synthesis, growth, apoptosis, and differentiation. Aberrant activity of c-myc, c-fos, and ODC is associated with neoplastic transformation. For example, cell lines stably expressing transfected c-myc, c-fos, and ODC readily convert to a tumorigenic phenotype (46)(47)(48)(49)(50). Another common characteristic of c-myc, cfos, and ODC is their ability to be induced by potent tumor promoters such as TPA Articles * Magnetic field effect on cell transformation and gene expression (51)(52)(53)(54)(55)(56). Furthermore, both c-fos and cmyc have been implicated in the regulation of ODC transcription induced by TPA (56)(57)(58)(59). In the context of cancer promotion by TPA, the induction of these genes is seen as an indicator of a global stress response that leads to genetic instability (60)(61)(62). Based specifically on Goodman's data on gene transcription (32)(33)(34)(35), it has recently been proposed that the deleterious effects of MFs are also connected with similar TPA-like indirect genotoxic effects and epigenetic changes (63). These implications add to the need to confirm the original findings on gene expression.
Materials and Methods 60-Hz magnetic field exposure system. Cells were exposed in tissue culture flasks placed in the central planes of long solenoids. The growing cells formed a thin layer (HL60 cells in suspension) or a monolayer (attached SHE or C3H/1OTi/2 cells) parallel to the field lines. The exposure system, as described previously (28), was recently modified and characterized as required by the NIEHS Electric and Magnetic Fields Research and Public Information Dissemination (EMF RAPID) program (64). The solenoids (38.7 cm long, 12 cm in diameter, 199 turns, 1.59 mH, 1.76 Q) were wound on thin-walled plexiglas cylinders; the internal diameter was sufficient to accommodate 75-cm2 tissue culture flasks. A sham-control solenoid was bifilar-wound and energized at the same level as the exposure solenoids. Consequently, control cells experienced zero magnetic field, but were subject to the same level of possible disturbances, including local acoustic noise and convective heating as field-exposed cells. A system schematic is shown in Figure 1. Power to the coil was line derived, with added surge protection. To eliminate the geomagnetic and other background DC/AC fields inside the solenoids, they were enclosed in individual magnetic shield cylinders with a diameter of 22.9 cm, fabricated from 0.13cm thick Amumetal (Amuneal Manu-facturing Corp., Philadelphia, PA). These cylinders were constructed with two removable friction-fit end caps. Both the cylinders and end caps were fabricated with 4-mm holes for water circulation, with the holes covering 10% of the surface area of the bottom halves of their circumference.
The control and up to two exposure solenoid assemblies were placed parallel to each other in a circulating temperature-controlled water bath (1.2 m long, 0.7 m wide, 11 cm deep). The total water volume of approximately 100 liters was circulated through an external reservoir with heaters and pumps, with temperature control referenced to thermistors immersed in the exposure bath. This  arrangement was used to minimize extraneous fields in the exposure bath, notwithstanding the effective shielding already provided for the solenoids. The exposure bath was insulated with styrofoam to reduce heat loss, and the temperature of the bath was maintained at 37.0 ± 0.1°C. This temperature was realized over the central region of the bath, induding the location of the solenoids. Although several hours were required to equilibrate this large thermal mass, the target temperature was then very stable over long-duration exposures. Exposure flasks were placed horizontally on plexiglas platforms in the solenoid so that they were half submerged in water. Flask contact with circulating liquid ensured good heat transfer and clamped the cell temperature at the desired level. According to direct measurement of culture medium temperature within flasks, temperature equilibrium was reached within 3 min for flasks with 15 ml of medium or within 20 min for medium-filled flasks. Magnetic field measurements. Measurements of the fields within the exposure coils were made with a three-axis detector (Bartington model MAG-03MC; Bartington Instruments, Oxford, U.K) and agreed within 5% to calculated solenoid fields values as well as with measurements by National Institute of Science and Technology (NIST) personnel and equipment during an EMF RAPID-sponsored site visit.
NIST measurements also indicated that the field in the energized sham exposure coil did not exceed 0.02 pT. Field uniformity was ± 2% along the 18-cm-long central plane inside exposure coils where exposure flasks were positioned. The voltage across each exposure coil was monitored continuously during experimental runs to document temporal stability. Mean, minimum, and maximum levels were logged and indicated stability to ± 3% over 24 hr. Occasional transient power losses were noted in a few runs. No distortion of the 60-Hz magnetic field wave form was evident in oscilloscope displays; measurements of the power spectrum (Hewlett-Packard model 3562A; Hewlett-Packard, Palo Alto, CA) of exposure and background fields revealed a third-harmonic component 20 to 40 dB down from the 60-Hz fundamental, according to spot measurements on three separate occasions.
The presence of the expected magnetic fields in each coil containing treatment flasks was verified before each experiment using a small field sensor (Bell model 4060; F.W. Bell, Orlando, FL). Exposure flasks were manually introduced into energized solenoids and removed prior to turning them off in order to minimize transient fields.
Concerning possible accessory field exposure, the maximum 60-Hz background field in our exposure facility obtained by activating all the power devices was 0.1 pT, but typical levels were less than 0.01 JT. In previous 2.45-GHz microwave exposures of C3H/10T1/2 cells at the same location (26-28) extensive controls gave Articles * Balcer-Kubiczek et al.
A 0.5 m x 0.5 m x 0.25 m doublewalled container fabricated from 0.13-cm Amumetal was used to carry cells to and from the MF exposure facility. This journey was approximately 200 m, over which the 60-Hz background did not exceed 0.2 pT before shielding, according to spotcheck measurements.
In processing and maintaining cells, SHE and C3H/10Ti/2 cells were not centrifuged before or after exposure (transformation experiments or RNA extractions). HL60 cells (except positive control treatments as noted below) were centrifuged after exposure (1000 rpm for 8 min) for RNA extraction. During this time, cells from all groups experienced MF levels of 1-3 pT. Cells were maintained in a facility with six similar incubators (Forma model 3158; Forma Scientific, Marietta, OH). Four of these incubators had MF levels in their chambers ranging between 0.1 to 1 pT, while the other two had much higher field levels, apparently emanating from the fan motor; motor replacement has solved this problem.
Positive control treatments. We defined positive controls as agents other than MF previously known to induce neoplastic transformation or to modify the abundance of a specific mRNA. Sham-exposed cells were subjected to appropriate positive control treatments immediately following MF exposure.
We used 1.5 Gy of X rays to elicit a positive transformation response in SHE or C3H/ 1 0T1/2 cells (16,20,(24)(25)(26)(64)(65)(66)(67)(68) and as a negative control for c-fos expression in HL60 cells (55). Hyperthermia at 45.5°C for 9 min (53) and TPA at 0.3 pg/ml for 30 min or 50 ng/ml for 3 hr (51-55) were used as treatments modifying the expression of cmyc and c-fos in HL60 cells and the expression of ODC in C3H/IOT1/2 cells. TPA was obtained from Chemsyn Science Laboratories (Lenexa, KS). The TPA solvent acetone (purity >99.9 %) was obtained from Sigma Chemical Company (St. Louis, MO). For heat treatments, glass test tubes, each containing 15 ml of HL60 cell suspension at 5 x 105 cells/ml, were submerged in a 45.5 ± 0.1°C temperature-controlled water bath for 9 min (70). TPA treatment and X irradiation of C3H/10Ti/2 cells were provided as previously described (24)(25)(26). assay (65); our quality control ofexperiments with C3H/1OTi/2 cells is also described in these reports. A single lot of serum [selected according to the criteria described previously (65,69)] was used in all the experiments reported here. Actively growing cells in passage 10 were used for experimentation. Each experiment included one 200-pT MF group and two control groups. Immediately following a 24-hr exposure to MF, cells were transported in a p-metal container to our X-ray exposure facility and one sham group was exposed to 1.5 Gy of X rays. Then, the groups were blinded using an alphanumeric code according to a computer-generated sequence of random numbers and cells plated into correspondingly coded 100-mm dishes containing culture medium with 0.1 pg TPA/ml or TPA solvent (acetone at 0.05%) for cell survival and transformation determination (24)(25)(26). As in our previous experiments, the growth medium was renewed at weekly intervals (2 weeks for survival assay, 8 weeks for transformation assay) (24)(25)(26)67,69). Cell survival and plating efficiency were determined by colony formation, while neoplastically transformed foci (Fig. 2) were identified according to published criteria (19,65). The endpoint of transformation per surviving cell was calculated by the null method (66), with uncertainties determined according to our analysis (62). SHE (pH 6.7) transformation assay.
Several aspects of the SHE transformation assay (16,20) have recently been modified (21,22). This now-standardized assay was used in our present study. Cryopreserved stocks of SHE cells, culture media samples, detailed laboratory protocols, and references of morphological phenotypes were kindly provided by G.A. Kerckaert. Cells were cultured in Dulbecco's LeBoeuf Modified Eagle's medium, pH 6.7 (Quality Biological, Gaithersburg, MD), containing 20% fetal bovine serum (HyClone) at 37°C in 10% CO2 in air (21,22). Following thawing of frozen stock SHE cells and culturing cells for 2 days, feeder layers were prepared by plating 5 x 104 Xirradiated (60 Gy) SHE cells in 2 ml of full medium into up to 80 60-mm tissue culture dishes. One day prior to feeder layer preparation, cells from another vial of frozen stock were used to establish cultures for exposure to MF by plating 5 x 104 cells into several 25-cm2 flasks containing 5 ml of medium. Flasks containing 1-day-old SHE cell cultures were filled with medium as described previously for experiments with C3H/ 1OTi/2 cells (24)(25)(26)(27)67,69). Following 24hr MF exposure, cells were plated onto the feeder layer at 80 cells per plate in 3 ml of fresh medium for survival and transformation assessment (21,22). The cultures were grown for 8 days without medium change. Other methods used were as in our present experiments with C3H/10Ti/2 described above. Morphologically normal and transformed colonies in Figure 3 were identified using published criteria (16,(20)(21)(22). Transformation frequency is the number of Articles * Magnetic field effect on cell transformation and gene expression I, -JJ Figure 3. Morphological criteria of neoplastic transformation in SHE cells cultured at formed colony with disorganized cellular growth pattern including multilayered, criss-cr ution; (B) colony of normal cells that retained the ability to divide after treatment, exhib nized streaming pattern with uniform cellular separation. Figure 4. Ethidium bromide-stained coded RNA samples in a nondenaturing 1% agaro RNA extracted from control or field-exposed HL60 cells. Cells were lysed immediately ( after 10, 20, or 40-min exposures; there were two flux density groups: in the first run, 0, 5. in the second run 0, 57, and 570 pT. All the samples were obtained in a single experime and 6 for Northern analysis results performed using these RNA samples. positive colonies divided by the total number of colonies; uncertainty was calculated by assuming that transformants are Poissondistributed with the mean equal to the number ofpositive colonies (68).
To determine the levels of ODC mRNA, C3H/1OTi/2 cells at passage 10 were plated into several 75-cm2 flasks at 2000 cells/cm2 3 days prior to MF exposure (200 pT, 24 hr) and cultured as for transformation experiments. Following MF exposure, the spent medium was removed and flasks were refilled with 15 ml of fresh medium containing 0.3 pg TPA/ml; a final acetone concentration in the culture medium was 0.015%. Following the 30-min treatment, medium was removed and cells were washed three times with phosphate buffered saline. The cells were immediately lysed in situ using 4 ml of TRIzol (Gibco BRL) and stored at -70°C. Three independent experiments were performed, each consisting of the four groups described above. were then filled with 20 ml a um and placed inside a C overnight. This medium was seeding flasks for exposure. Each of three experimen three MF flux densities (5.; pT), three MF exposure dur and 40 min), and corresponc no-field controls, as describe et al. (34). However, in c experiments, exposures wert two consecutive runs, each n nine field/time combinations sity groups, one sham-exposed control group, three exposure times), for a total of 18 groups per run [instead of one sham-< ts .-. exposed control, one field, and three exposure durations per experiment in experiments described in (34)]. Immediately following MF treatments, cells were concentrated by centrifugation --t t (1000 rpm, 8 min at 4°C) and lysed using 1 ml of ice-cold TRIzol. Cell lysates were transferred into new coded 1.5-ml microfuge pH 6.7. (A) Transtubes and refrigerated at -700C. For interlabossed cell distrib-oratory comparison, cell lysates were divided iting typical orgainto two equal parts before freezing; two collaborating groups in the Department of Radiation Oncology and the Department of Medicine, University of Maryland School of Medicine, were involved in this study. Similar methods were used in processing HL60 cells subjected to positive treatments * 28S rRNA with X irradiation or hyperthermia. Modification of these methods was necessary  (71). Other methods used lls were seeded were from our previously published proceks at 5 x 105 dures (72,73). RNA concentrations were ;periment, those determined spectrophotometrically. The ed and distrib-A260/A280 was greater than 1.8. In addilber of 25-cm2 tion, the integrity of the RNA samples was 15 ml of fresh determined by examination of the ethidi-H stability durum bromide-stained 1% agarose gels; are flasks were Figure 4 shows one representative example with CO2; they of the RNA samples analyzed in Figures 5 )f growth medi-and 6. Both the sharpness of the bands and 0°2 incubator the relative intensity of the 28S, 18S, and removed before 5S rRNA were assessed. Samples were considered to be degraded if the 28S band was its consisted of not more intense than the 18S band or if 7, 57, and 570 the fluorescence suggested low-molecular rations (10, 20, weight products. If any sample in a series ling concurrent was degraded, the experiment was repeated. d by Goodman RNA samples (10-20 pg) were size-fraciur replication tionated by electrophoresis (35 V  Conditions for hybridization and washing of Northern blots were as described previously (72,73). After washing, membranes were exposed to X-Omat Kodak Scientific Imaging film (Eastman Kodak, Rochester, NY) at -700C for 2 hr (ribosomal genes) to 2 days. To remove hybridized probes for reprobing, membranes were washed in 0.5X Denhardt's solution (Quality Biologicals), 0.1% sodium dodecyl sulfate, 25 mM Tris, pH 7.5, at 90°C for 1 hr, followed by another 30-min wash at 85°C.
Hybridization probes. The ribosomal cDNA probes for the 5-kb 28S rRNA and 1.9-kb 18S rRNA were generated from total RNA using SuperScript RNase-H Reverse Transcriptase (Gibco BRL). Other probes used in Northern analysis were the clone pODC 821 containing a 7.78-kb DNA fragment of the rat ODC inserted into the pUC13 vector (American Type Culture Collection, Rockville, MD) for the 2.2-kb and 2.7-kb ODC transcripts; the 1.4-kb h-uman c-myc (3rd exon) fragment (Oncor, Inc., Gaithersburg, MD) for the 2.3-kb cmyc transcript; the clone pc-fs-1 containing a 9-kb fragment of the human c-fos inserted into the pBR322 vector (American Type Culture Collection for the 2.2-kb c-fs transcript); and fragments of human j-actin and GAPDH for the 2.2-kb j-actin transcript and the 1.1-kb GAPDH transcript was a gift from Kathleen Daher (University of California, Los Angeles, CA).
To determine the amounts of hybrids formed between the transcript of interest and an appropriate probe, autoradiograms were analyzed by densitometry using a ScanJet 4C (Hewlett Packard) and the Intelligent Qualifier, version 2.1 (BioImage, Ann Arbor, MI).
The coded autoradiographic data for 18S rRNA, 28S rRNA, P-actin, GAPDH, and c-myc, exemplified in Figure 5 C-E, were analyzed as previously described (74,75). Accordingly, principal component analysis was performed by transforming the raw densitometry data into their Z-scores (obtained by subtracting the mean and dividing by the standard deviation). Z-score distributions provide an objective measure of the variability of the transcripts that is independent from the mean and from the scale of the optical densities. Therefore, this method is well suited for comparing the coded data from different autoradiograms. The Z-scores from three independent experiments were pooled for each of five genes (54 Z-scores for each gene) and organized as a 5 x 54 analysis of variance (ANOVA) table. Statistical variation within groups and among five gene groups was tested by two-way ANOVA. In addition, Zscores in each of the five gene groups were analyzed by the box-whisker plots. In Figure  7, p is the one-sample s-test two-tailed probability that the group mean is zero.
The uncoded radiographic data for 18S rRNA, 28S rRNA, P-actin, GAPDH, and cmyc, as shown in Figure 6, and the data for ODC (Fig. 8)   Articles Magnetic field effect on cell transformation and gene expression the standard deviation/mean) were calculated. The coefficients of variations and visual examinations ofautoradiograms provide similar information about heterogeneity within the data subset because both methods are sensitive to the mean optical density and the standard deviations of the mean. To minimize this source of error, the differences in the mean basal abundance among various transcripts were partially compensated by adjusting the amount of RNA to be examined and by differential autoradiography, so that in terms of optical density, the bands for shamor magnetic field-exposed HL60 cells. After the coded samples in Figure 5 were analyzed, the group identities were revealed and the RNA samples were grouped according to magnetic field exposure level/duration patterns. (A) Ethidium bromide-stained RNA samples (from Fig. 4) (21,22). Determination of transformation frequency in this study involved morphological evaluation of nearly 20,000 colonies from nine experiments. Eight transformed colonies (Fig. 3) were found in the 1.5-Gy group and no transformed colonies were found in sham or MF groups. Transformation frequency at a moderately cytotoxic X-ray dose of 1.5 Gy was 2.86 x 10-3 per clonogenic cell (Table 1). This estimate is about five times the frequency reported previously for SHE cells grown in conventional media, pH 7.1-7.3, (20); no other X-ray data are available for comparison, but we note that enhanced carcinogeninduced morphological transformation and low spontaneous background were expected for SHE cells grown at pH 6.7 (21,22).
Neoplastic transformation of C3HI 1OT1/2 cell. Transformation frequency and survival of MF-exposed C3H/1OT1/2 cells with and without the promoter TPA, together with data for X irradiation, are shown in Table 2. More than 500 dishes per group were accumulated to establish the effects of MF with or without TPA with some degree of statistical certainty.
Under the present experimental conditions, there were no significant differences in plating efficiency among MF-exposed and sham-exposed cells, as expected. Positive control results with X rays are in agreement with observations from related experiments using the C3H/1OT1/2 assay in this and other laboratories (23)(24)(25)(26)(65)(66)(67)(68)(69). The effect of TPA on sham-exposed cells was not significant (sham-irradiated control Table 1. The frequency of neoplastic transformation of Syrian hamster embryo cells exposed to 60-Hz magnetic field or X rays and cultured at pH 6.7 Total PE or Total positive TR ± SE Treatmenta SF ± SEb colonies colonies (10-3)C vs. sham-irradiated control + TPA; p = 0.35), whereas it produced the expected enhancement of transformation induced by X rays (X irradiated vs. X irradiated + TPA; p < 10-6) (25)(26)(27)(28). Consistent with the results obtained using the SHE assay, MF and 0 ,ug TPA/ml post-treatment produced no significant effect on transformation (sham-irradiated control vs. MF; p = 0.70). Similarly, the treatment MF and 0.1 ,Ig TPA/ml post-treatment did not increase the transformation frequency above the corresponding control level (sham-irradiated control + TPA vs. MF + TPA; p = 0.59).
An approximately twofold effect of TPA in MF-exposed cells was not significant (MF vs. MF + TPA;p= 0.13).
Epression ofODC in O3HIOT/2 cells. As shown in Figure 8, Northern blot analysis of ODC mRNA demonstrated the presence of two distinct ODC mRNA transcripts, approximately 2.2 and 2.7 kb in size. Two ODC mRNA transcripts have been demonstrated in a number of different tissues; the 2.2-kb band is typically weaker than the 2.7kb band in fibroblasts (76). Compared to basal levels, we observed an increased expression of 2.2-and 2.7-kb ODC mRNA transcripts immediately following TPA application for 30 min; no similar increases in ODC mRNA were seen in MF-or sham-exposed cells without TPA. Quantitative densitometry indicated that the ODC mRNA induced by TPA in cells previously exposed for 24 hr to MF was similar to the abundance of ODC mRNA in TPA-treated control cells. Therefore, we confirmed ODC as a TPAinducible but not an MF-inducible gene. According to these data, it seems unlikely that MFs could play a role in carcinogenesis by mimicking the activity of TPA in rodent cell systems, as supposed by others (3,4,8,42,63).
Expression ofc-myc, 0-actin, GAPDH, and ribosomal genes in HL60 cells (coded data). Figures 4-7 demonstrate the sequential steps of processing and analysis of MF and sham-irradiated samples of RNA. Onemicroliter volumes of RNA samples, separated on a nondenaturing 1% agarose gel directly following RNA extraction from a flask of cells (each representing one experimental group), but before the RNA sample concentrations were measured and equalized, are shown in Figure 4; this demonstrates that approximately equal numbers of cells were collected in each group.
Representative Northern blot analyses from our present studies of the expression of c-myc mRNA, GAPDH mRNA, and ,Bactin mRNA in coded samples are displayed in Figure 5. In this example, Figure 5A shows equal amounts of ethidium bromidestained RNA samples separated on a denaturing formaldehyde-agarose gel. Figure 5B represents RNAs shown in Figure 5A after transfer on the nylon membrane. The membrane in Figure 5B was hybridized with two 32P-labeled DNA probes for human c-myc (exon 3) and GAPDH (Fig.  5C) together and sequentially rehybridized with the DNA probe for f-actin (Fig. 5D) and cDNA probe from reverse transcribed total RNA as described in Materials and Methods (Fig. 5E).
The example of the data in Figure 5 was obtained in one day from a single experiment designed as a series of two runs, each containing three exposure durations of 10, 20, and 40 min. In the first run, the magnetic flux densities were 0, 5.7, and 570 pT. In the second, the magnetic flux densities were 0, 57, and 570 pT. Disagreement of our data in Figure 5 with the data reported by Goodman et al. (34) is apparent, even from a visual comparison of band densities for c-myc and P-actin in the same group: the optical densities of bands resulting from hybridization with c-myc do not track the optical densities of bands resulting from hybridization with ,-actin, contrary to observations reported for c-myc and ,-actin by these authors (34). Table 2. Effect of post-irradiation treatment with the tumor promoter on the frequency of neoplastic transformation of C3H/lOT1/2 cells exposed to 60-Hz magnetic field or X rays aExperimental groups were defined as follows: 0, Sham-irradiated control; MF, 24-hr exposure to 60-Hz magnetic field at 0.2 mT; X, X irradiation at 1.5 Gy. After treatment, cells were plated in medium containing acetone or 0.1 pg 12-0-tetra-decanoylphorbol-13-accetate/ml in acetone (TPA). bPlating efficiency (PE) or surviving fraction (SF) ± standard error (SE) from pooling the results from six independent experiments on survival and transformation.
cTransformation frequency (TR) per surviving cell (48) ± SE derived from total number of dishes and total number of positive dishes (49).
Volume 104, Number 11, November 1996 * Environmental Health Perspectives Articles -Magnetic field effect on cell transformation and aene expression Z-score distributions of the coded image data for c-myc, 3-actin, GAPDH, and ribosomal genes pooled from several similar experiments (induding independent interlaboratory comparisons of Northern analysis of the RNA samples from the same experiment) are shown in Figure 7. Analysis of the entire data set by ANOVA yielded p = 0.833 (F= 0.364; df= 4) for comparison among gene groups and p = 0.217 (F= 1.302; df= 53) for comparison within groups, confirming that the group means were equal (similarly large p-values were obtained for the data subsets for individual experiments). Box-whisker plots of Z-score distributions of individual groups are shown in Figure 7. One sample t-test performed on the group data provided the following two-tail probabilities that the group mean is zero: c-myc group, p = 0.707; [B-actin group, p = 0.279; GAPDH group, p = 0.324; 18S rRNA, p = 0.984; and 28S rRNA, p = 0.662 (53 dfin each data subset). Taken together, these analyses show no statistically significant departures from random variability. Therefore, our data contradict previous conclusions for c-myc and [B-actin (32)(33)(34).
Expression ofc-myc, [-actin, GAPDH, and ribosomalgenes in HL60 cels (uncoded data). The data-handling methods used in the original publications (34) as well as in the recent papers reporting the replication results (12)(13)(14) can only be used if the group identities are known. Figure 6 shows independent Northern blots of the RNA samples in Figure 4 after the group identities were revealed.
MF-group means of replicate ratios of optical densities of a band from an MFexposure group to the appropriate control group band, along with the coefficients of variation, are summarized in Table 3. In terms of the coefficients of variation, the normalized autoradiographic data for the 18S ribosomal transcript and a metabolic transcript, GAPDH, were the least heterogeneous, perhaps justifying their use as reference genes, as reported in the recent MFrelated work (12,14). Normalized data sets for the c-myc, [-actin, and 28S rRNA gene transcripts were similar to each other but were more heterogeneous than the other two. Less than 20% of variations in the relative levels of 28S ribosomal gene, c-myc and [B-actin transcripts among nine MFgroups in Table 3 have no recognizable biological consequences (77)(78)(79).
In terms of the band density ratios shown in Table 3, the variation of five transcripts, 18S rRNA, 28S rRNA, c-myc, [B-actin, and GAPDH was not associated with MF exposure parameters and remained close to basal levels. Thus, the results from the present experiments offer no support for the previous findings that MFs affect transcription of c-myc and [Bactin (32)(33)(34). Expression of c-fos in HL60 cells. In control treatments, TPA and hyperthermia significantly increased c-fos mRNA abundance above basal levels in control cells (Fig. 9); these results are consistent with the data for these agents in the literature (51)(52)(53)(54)(55). The negative result for X irradiation at 1.5 Gy was expected; the threshold X-ray dose for visualization of the c-fos transcript in HL60 cells is 5 Gy (55). The phenotypic consequences of expression of c-fos were observed in the course of the present study: hyperthermia induced apoptosis as demonstrated by the DNA fragmentation assay, and TPA treatments caused dramatic population-wide differentiation as evidenced by changes in HL60 cell morphology and adhesion of cells to a plastic substrate (data not shown).
The responses of HL60 cells to positive control treatments and two MF flux densities are shown in Figure 10. In Figure lOB,  1.07 (0.05) aHybridization signal of the RNA sample extracted from HL60 cells exposed to 60-Hz MF, which was normalized to hybridization signal in a corresponding control group. bValues shown as mean and coefficient of variation (in parenthesis). cNumber of independent estimates of the effect of 60-Hz MF from three independent experiments at the field levels and exposure durations shown; in each of three experiments, one flux density group was replicated, providing a within-experiment variation in two groups of samples independently exposed to the same flux density (see text and Fig. 6).
we compared the c-fos hybridization signals in the RNA positive control samples with the c-fos hybridization signals in the RNA samples from sham-exposed and MFexposed HL60 cells. The high levels of responsiveness to TPA and hyperthermia were not observed with MFs. In fact, the cfos transcription was undetectable in the MF samples. These negative results for the c-fos mRNA induction by our MF exposures of HL60 cells were contrary to the reports by others (32,33,35,38).

Discussion
In designing these experiments, we were forced to limit our choices of MF exposure parameters, as mentioned in the Introduction. Overall, we used MF parameters shown by others to be critical in influencing biological responses. In the case of transformation experiments, we considered animal studies that suggest a tendency toward a dose response in terms of MF flux density and exposure duration (3)(4)(5)(6), so that higher-level, longer exposures would increase the likelihood of seeing a positive effect. Since the cell cycle time of C3H/10Ti/2 cells is approximately 20 hr (19,66,69), our 24-hr exposures at 200 pT could be taken as emulating a lifetime MF exposure of mice scaled to a mouse single cell. Surprisingly, previous studies of gene expression in cultured cells failed to identify critical MF exposure parameters. For example, positive results on the transcription of c-myc and histone (35), as well as on ODC enzymatic activity (42,43), have indicated that a brief exposure to a magnetic field is enough to evoke the same level of effect as a prolonged exposure. In other studies, no differences in gene responses were observed in several extremely low frequencies (up to 100 Hz) and MF flux densities (32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43). For our present work involving gene expression in HL60 cells, we selected nine 60-Hz MF exposure parameters tested by Goodman et al. (34). Table 4 shows the results with MFs for the series as a whole. Our findings, presented in this paper, did not provide any significant   Figure 10. Expression of c-fos mRNA sham-exposed or 60-Hz magnetic field (MF)-exposed Hi Cells were exposed to MF for 10, 20, or 40 min at 0, 5.7, or 570 pT; 50 ng 12-0-tetra-decanoylph acetate (TPA)/ml for 30 min; or 45.  We report here for the first time the MF response data obtained using two classic transformation systems. We found no statistically significant evidence of carcinogenic or initiating activities of MF in these in vitro models for the tested cells and expe conditions. This can be compared conclusion from the promotion s Cain et al. (8) using a variant ass preinitiated C3H/1OTi/2 cells; inter these authors reported a 1.9-fold TPA-treated MF-exposed groups, w also not statistically significant. I these in vitro results corroborate results from animal studies (5,7,9). not study MF as a promoter in vitro er, considering the lack of MF inflL ODC mRNA in C3H/10Ti/2 cells, promotional activity of MF seems unlikely, at least 4 28S at the MF-time combination studied. Moreover, even when positive results for ODC have been reported with MF, the effects were transient (42,43). In contrast, only the sustained overexpression of ODC is associated with neoplastic transformation -18S (46) and cancer promotion in vivo (56,57). We observed no effects of MFs on transcription ofseveral genes performing a variety of functions within the cell. Our  blots are unsuitable for a quantitative analysis of transcript levels because the equal amounts and integrity of RNA samples cannot be verivel expo-fied. In fact, Goodman et al. (34) stated that they could not confirm the dot blot data by ie Northern blot analysis of the selected samples. The presence or absence of geomagnetic ct field in the Goodman study or ours had no apparent influence on the results. Both positive and negative results for P-actin and/or cmyc were obtained in previous replication studies that used shielded (13) and unshieldtct ed (12,14) MF exposure systems. Ict Our agreement with other recent results (12,13) should not obscure the fact that ct differences exist between our study and ct theirs. Both studies (12,13)  Consequently, a many-fold increase of c-fos mRNA is required for c-fos transcript visualization after an inducing treatment. For instance, the most effective treatment in our experiments, heat shock (45.50C for 9 min) 1 hr posttreatment (Fig. 9), represents an estimated 1000-fold increase in c-fos transcription (53). In view of these results, reported increases of approximately 20-30% in c-fos transcription (relative presumably to basal levels) by MF (20 min exposure at 8 to 80 JT) in HL60 cells, most recently claimed by Henderson (38), must in reality be considered undetectable. Even more puzzling is the finding by these authors that c-fos transcription was decreased by certain combined treatments with MFs (37,38) since basal levels are so low to begin with.
In summary, the negative data on neoplastic transformation and gene expression are inconsistent with the proposed hypothesis in the range of MF exposure parameters studied. Studies of neoplastic transformation in vitro deserve to be continued at various MF field/time combinations to determine whether MFs represent a real health risk. The significance of our negative results on gene expression is twofold. First, the findings presented in this paper complement the data already available for c-myc and f-actin (12)(13)(14) and extend these observations to all the MF exposures studied by the Goodman group (34). In addition, we presented new data for several genes, including ODC and c-fos, that have not been investigated in previous replication experiments. Our control treatments with TPA, hyperthermia, and X irradiation produced the expected cellular and gene responses expected from the literature data. Thus, our experiments were sufficiently sensitive to detect the reported increases in gene expression after MF exposures. Second, positive molecular data, exemplified by Goodman's data on c-myc induction by MFs, provide a mechanistic background for several models of MF carcinogenesis. In view of our present results, the genetic effects of MF remain unconfirmed, thus diminishing the biological plausibility of a causal relationship between MF and cancer.