50-Hertz magnetic field and calcium transients in Jurkat cells: results of a research and public information dissemination (RAPID) program study.

An effect on intracellular calcium continues to be proposed as a biochemical pathway for the mediation of biologic effects of electrical-power-frequency magnetic fields (MF). However, reproducible results among laboratories are difficult to attain and the characteristics of magnetic field effects on intracellular free calcium ([Ca(2+)](i)) are not well understood. We attempted to repeat the studies of Lindström et al. [Intracellular Calcium Oscillations in a T-Cell Line by a Weak 50 Hz Magnetic Field. J Cell Physiol 156:395-398 (1993)] by investigating the effect of a 1.5-G 50-Hz MF on [Ca(2+)](i) in the Jurkat lymphocyte T-cell line. Changes in [Ca(2+)](i) were determined using microscopic imaging of fura-2 loaded Jurkat cells on poly-l-lysine-coated glass coverslips. The MF was generated by a single coil constructed with bifilar wire and located in the same plane as the cells. Cells were randomly exposed for 8 min to MF, sham field (SF), or no field (NF) conditions. The exposure condition remained coded until data analysis was complete. Each exposure period was preceded by an 8-min data collection to establish a baseline for [Ca(2+)](i). After each exposure condition, cells were exposed to anti-CD3 antibody that induced a rapid increase in [Ca(2+)](i) in responsive cells; this provided a positive control. [Ca(2+)](i) was analyzed for individual cells as spatially-averaged background-corrected 340/380 nm ratios, and a [Ca(2+)](i) transient was considered significant for positive deviations from baseline of 3 [multiple] an estimate of noise in the baseline. Typically, 25-50 cells/field were viewed and approximately 50% had no [Ca(2+)](i) transients in the baseline period and also responded to positive control. Only cells responding to positive control and lacking changes in [Ca(2+)](i) during the baseline period were considered qualified for assessment during the exposure period. The incidences of [Ca(2+)](i) transients during the exposure period for two experiments (40 [multiple] objective) were 16.5, 14.6, and 14.2% for MF, SF, and NF, respectively, and were not statistically significantly different. Previous studies by Lindström et al. [Intracellular Calcium Oscillations in a T-Cell Line after Exposure to Extremely-Low-Frequency Magnetic Fields with Variable Frequencies and Flux Densities. Bioelectromagnetics 16:41-47 (1995)] showed a high response rate (92%) for exposure to 1. 5-G 50-Hz MF when individual cells were preselected for investigation. We found no such effect when examining many cells simultaneously in a random and blind fashion. These results do not preclude an effect of MF on [Ca(2+)](i), but suggest that responsive cells, if they exist, were not identified using the approaches that we used in this study. ImagesFigure 1Figure 2Figure 3

MF was generated by a single coil constructed with bifilar wire and located in the same plane as the cells. Cells were randomly exposed for 8 mi to MF, sham field (SF), or no field (NF) conditions. The exposure condition remained coded until data analysis was complete. Each exposure period was preceded by an 8-min data collection to establish a baseline for [Ca2+]1. After each exposure condition, cells were exposed to anti-CD3 antibody that induced a rapid increase in -.(1995)V showed a high response rate (92%) for exposure to 1.5-G 50-Hz MF when individual cdls were preselected for investigation. We found no such effect when examining many cells simultaneously-in a random and blind fishion. These results do not predude an effec of MF on [Ca2+], but suggest that responsive cells, if they exist, were not identified using the approaches that we used in this study. Key wordk: fira 2, intracellular calcium, Jurkat cells, lymphocytes, magneic fields. Environ Healb Pepect 108: 135-140 (2000). [Online 7 January 2000] hap:/llehpntl nies.nib.gov/docs/2000/108p135-140weylabstract.btml The reported association between exposure to magnetic fields (MF) and a variety of adverse health effects (1,2) has led to repeated attempts to define possible mechanisms (3). Calcium biochemistry has been a central facus of many of the studies because calcium is a ubiquitous second messenger in cell processes and is, in many cases, regulated at cell membranes, a purported target of MF. Over the past two decades, research using diverse in vitro models has demonstrated the ability of MF exposure to modify calcium metabolism, presumably by affecting calcium flux across membranes (4)(5)(6)(7)(8)(9)(10). However, several investigations have not detected an effect of MF on calcium metabolism or flux in a variety of cell types (11)(12)(13)(14)(15). This apparent lack of consistency among investigating groups on the effects of MF on cellular calcium mobilization has been reflected throughout the field of research into the biologic effects of MF (16).
Explanations for the conflicting results have invoked physical complexities related to MF exposure, biologic complexities related to heterogeneity of biologic response, or the need to establish a responsive biologic state in cells (6,12,(17)(18)(19)(20)(21)(22)(23). Physical complexities of MF exposure are demonstrated by reports of alternating current (AC)-direct current (DC) MF interactions, frequency windows, and the dependence of effects on both frequency (AC fields) and flux density. In the case of biologic complexity associated with calcium metabolism, Walleczek and Liburdy (24) reported that calcium uptake in rat lymphocytes was increased by brief exposures to 60-Hz MF only in cells stimulated with the mitogen concanavalin A (ConA). In this study, those lymphocytes least responsive to ConA exhibited the greatest response to MF. In a follow-up study, Liburdy (22) reported that the age of the rats from which the cells were harvested also contributed to a variation in calcium response to MF.

Materials and Methods
Cell culture. We obtained the Jurkat cell line (clone E6-1), derived from a human T-cell leukemia, from E. Lindstrdm (Umea University, Umea, Sweden). The cells were shipped as a cell suspension in RPMI-1640 (Life Technologies, Grand Island, NY) containing 10% fetal bovine serum (Hyclone, Logan, UT) in a sealed flask. The cell suspension was aliquoted into flasks at 30 mL suspension/75 cm2 growing surface. The viability was initially low (approximately 30% by trypan blue exclusion), but reached 98'% after 1 week in culture. The cells were grown in RPMI-1640 containing 10% fetal bovine serum (hereafter called complete medium), and maintained in a Forma incubator (Forma Scientific, Inc., Marietta, OH) at 37°C with a humidified atmosphere of 95% air and 5% CO2. In the incubator with mu metal shielding (Amuneal Manufacturing Corp., Philadelphia, PA), the maximum AC and DC magnetic flux densities were 1.4 and 7 mG, respectively. These flux densities were measured with a MultiWave II monitor (Electric Research and Management, Inc., State College, PA). The culture was diluted to 1 x 105 cells/mL complete medium every 2-3 days to maintain cell density below 2 x 106 cells/mL. Treatment for mycoplasma contamination was initially performed by adding 5 lig/mL 4-oxo-quinoline-3-carboxylic acid derivative (ICN, Costa Mesa, CA) to complete medium for 7 days. Tests for mycoplasma contamination were routinely performed using an immunocytochemical test kit (ICN). Stocks of low passage cells (a passage is defined as a dilution; thus, there are three passages/week) were frozen in liquid nitrogen and used for all experiments.
Intracellular calcium measurements. We determined intracellular calcium through changes in fluorescence of the ratiometric dye fura-2 using a fluorescence microscopy imaging system. Jurkat cells were simultaneously loaded with fura-2 and immobilized on poly-L-lysine (PLL)-coated glass coverslips (25 mm Fisherbrand, catalog no. 12-545-102; Fisher Scientific, Inc., Pittsburgh, PA). One side of the glass coverslips was coated with PLL by placing them on the surface of a 0.01% solution of PLL in sterile water for 30 sec. Excess liquid was drained by touching the edge of the coverslip with tissue. Coverslips were allowed to air dry before being placed individually, coated-side up, in 35-mm tissue culture dishes. For plating on coverslips, Jurkat cell suspension in complete medium was diluted 1:1 with Krebs-Hepes buffer containing 1% bovine serum albumin (KHB) and an aliquot of fura-2/AM (Molecular Probes, Eugene, OR) added to a final concentration of 2 pM, and a 200-pL aliquot was placed on top of the PLL-coated glass coverslip. The coverslip cultures were incubated at 37°C in a humidified atmosphere of 95% air and 5% CO2 for 60 min. KHB (2 mL) was then carefully added to the culture dish containing the Jurkat coverslip culture. The coverslip was removed with forceps, dipped in KHB, mounted in a holder, and the culture covered with KHB (0.5 mL, 37°C). The coverslip holder containing the culture was immediately mounted in the exposure chamber on the microscope stage and suffusion with KHB (37°C) at 0.6 mL/min was started. The cells were illuminated briefly as the fluorescent image was brought into focus and the intensifier gain adjusted such that the pixel values for cell fluorescence were in the middle of the dynamic range for both 380 and 340 nm excitation. The process of mounting the culture caused [Ca2+]i transients, which subsided after 20 min. Thus, cells were kept in the dark for approximately 20 min before the start of an experiment.
Instrument control and data acquisition were facilitated by a personal computer-based image analysis system (MetaFluor; Universal Imaging Corp., West Chester, PA). The excitation source was a 75-W xenon lamp. Wavelength selection was accomplished using bandpass filters (340HT15 or 380HT15 from Omega Optical, Brattleboro, VT) mounted in a 10-position filter wheel (Sutter Instrument Company, Novato, CA). Neutral density filters were mounted together with the bandpass filters, as needed, to balance light intensity at 340 and 380 nm excitation. The position and movement of the filter wheel were controlled by the imaging software, and were integrated with data collection. All excitation light also passed through an additional neutral density filter to reduce light intensity, and through an infrared absorption filter to minimize localized sample heating. The xenon lamp and filter wheel were mounted on a Nikon Diaphot inverted epifluorescence microscope (Nikon, Inc., Melville, NY) equipped with a variety of low ultraviolet (UV)-absorbing objectives (20x, 40x, and 100x) and a filter cube containing a 400-nm dichroic mirror and a 515nm barrier filter. Fluorescence was detected by an intensified charge-coupled device camera (DAGE/MTI CCD72; DAGE-MTI, Inc., Michigan City, IN) and subsequently digitized by the image analysis system.
Fluorescence data were collected and saved as image sets every 2-2.5 sec and subsequently analyzed to determine [Ca2+]i ( Figure 1). Each image set contained 340 and 380 nm excitation data. A circular target area was established around all qualified cells using the first image of an experiment and was applied to every image set in an experiment. Qualified cells included all cells that did not exhibit [Ca2+]i transients in the initial baseline period and responded to the positive control (anti-CD3 antibody). Targets were only created around fully visible cells that were not in contact with other cells. Small circular targets were also created in background space adjacent to targeted cells for the collection of background fluorescence. At each image set the average pixel value for a background target was subtracted from the average pixel value for the corresponding cell target. We were only interested in detecting the presence of [Ca2+]i transients as defined by a significant change in [Ca2+]i above a stable baseline. Therefore, a calibration procedure was not performed, and the ratio of fluorescence at 340 to 380 nm excitation was used as a measure of [Ca2+] .
Magnetic field exposure. The MF exposure chamber was custom machined from an aluminum block. The exposure chamber was open in the center and secured the coverslip holder in place over the microscope objective. A commercially available coverslip holder (Leiden open perfusion chamber; Medical Systems Corp., Greenvale, NY) was modified by the replacement of the stainless steel base with an aluminum base. The top of the coverslip holder was made of Teflon with rubber 0-rings. A plexiglass cover over the top of the coverslip holder secured entry and exit tubing for the suffusion of cells with buffers. A thin 3.2-cm coil was located in a circular groove surrounding the coverslip holder and was in the same plane as the coverslip. Current flow was produced by a Hewlett-Packard 8904A signal generator (Hewlett-Packard, Palo Alto, CA) and a Techron 5515 power amplifier (Techron, Elkhart, IN). The coil operated at 50 Hz and produced a 1.5-G field over an exposure volume of 1 cm3. The coil contained bifilar magnet wire custom-wound to allow either active field cancellation (sham exposure) or applied field generation. A cross-over switch with positions labeled "A" and "B" was used to set MF and sham field (SF) conditions. In addition, a second switch labeled "C" and "D" was used for no current flow (no field; NF) or current flow for MF or SF conditions. In the Volume 108, Number 2, February 2000 * Environmental Health Perspectives NF condition, current flowed only through a dummy coil that was remotely located from the exposure volume. This arrangement ensured that the experimenter was blind to the field conditions. MF conditions were measured with a Multiwave II monitor. In addition, we measured the DC and AC MF perturbation (orientation, strength) from our objectives with a gauss meter (MNA-1904-VH and MNT-4E04-VH probes and 450 meter; Lake Shore Cryotonics, Inc., Westerville, OH). At the cell location, the maximum DC perturbation was 23, 26, and 22 mG (with no changes in field direction) for the 20x, 40x, and IOOx objectives, respectively. The DC magnetic flux density was 491 mG in the absence of objectives. The AC magnetic flux densities were perturbed 1.7, 0.2, and 0.6% by the 20x, 40x, and 1OOx objectives, respectively.
Temperature control. The temperature of the buffer in direct contact with the cells was maintained within a small range. For any given experiment, the mean temperature of the buffer in contact with cells was between 36.5 and 37.0°C with a standard deviation around the mean not exceeding 0.08°C. A plexiglass box (Nikon) enclosed the microscope stage, objective, condenser, and filter cube. Air temperature within the box was controlled by an air heater-controller (Air-Therm; World Precision Instruments, Sarasota, FL) with a thermocouple placed in a slot in the aluminum exposure chamber. This arrangement maintained the aluminum exposure chamber on the microscope stage at a set temperature of 370C. The suffusate buffer was warmed by running though tubing contained in a water jacket. The temperature of the water jacket was maintained by a circulating water bath (Lauda M3, Lauda-Konigshofen, Germany). The temperature of the buffer in the cell holder was continuously monitored with a YSI451 temperature probe (Yellow Springs Instruments, Yellow Springs, OH) and meter (model H-08502-16; Cole-Palmer Instrument Co., Vernon Hills, IL).
Experimentalprocedure and data analysis.
All experiments were blind. Two switches were used to set the field condition, as described in "Magnetic Field Exposure." Experiments were performed in a randomized block design, and all exposure conditions were performed in random order once each day. In their first publication, Lindstrom et al. (25) collected fluorescence data from experiments that lasted 16 min. We followed the same design with the modification that data were collected for an additional 6 min to include exposure to a positive control (anti-CD3 antibody). For each experiment, we obtained a baseline in the first 8 min, followed by MF, SF, or NF conditions for an additional 8 min. Finally, cells were exposed to anti-CD3 antibody (0.5 pg/mL) for 6 min.
Lindstrom et al. (25)(26)(27) used a photomultiplier tube for fluorescence detection. Use of the tube allowed data to be collected from only one cell at a time. This approach also necessitated the use of selection criteria for choosing a cell for study before MF exposure. Primary criteria were the general morphology of the cell, the absence of spontaneous calcium transients within a 5-min observation period, the absence of cell-cell contact, and a low basal calcium level. The use of fluorescence imaging enabled data collection from a field of cells simultaneously and eliminated the need to impose subjective selection criteria before MF exposure. The response of cells during each period was determined as yes (Y) or no (N) for the pres-  Figure  2). The incidence of responders in the presence of MF was compared to the incidence of responders in the absence of MF (SF and NF groups) by analysis of variance using a randomized block design. To Figure 2B, C. Three experiments were performed: two with a 40x objective and one with a 20x objective (Table 1). Of the experiments with a 40x objective, the first included eight sets of MF, SF, and NF conditions, and the second included seven sets of MF, SF, and NF conditions. For the experiment with the 20x objective, five sets of MF, SF, and NF conditions were performed (Table 1). There were no differences between any of the exposure conditions in the percentage of cells disqualified because of a preexposure response or a lack of response to anti-CD3 (Table 1). However, the preexposure response of a cell preparation was a predictor of the incidence of [Ca]i transients during the exposure period. This is shown in Figure 3 as a positive correlation between the preexposure response and the response during the exposure period, and could be interpreted as an influence of "culture excitability" on the occurrence of [Ca]i transients during the exposure period. The influence of preexposure response appeared to only be an issue for experiments using the 40x objective because, as shown in Figure 3, the incidence of spontaneous [Ca2ii transients was consistently lower for experiments using the 20x as compared to the 40x objective. To account for this relationship in the data analysis of experiments 1 and 2 (40x), the preexposure response was included in the analysis of variance model as a covariate. For all three experiments, the incidence of  Environmental Health Perspectives * Volume 108, Number  The incidence of spontaneous [Ca2+] transients was consistently lower for experiments using the 20x as compared to the 40x objective. To address whether this was a significant difference and also if there was a difference between a 40x and a IOOx objective, which was used by Lindstrom et al. (25), an additional series of experiments were conducted with objectives with different magnification (20x, 40x, and 1OOx). Experiments were performed using all objectives on the same day. These experiments only assessed the level of spontaneous [Ca2+]i transients during no field conditions. The 20x objective resulted in a significantly lower incidence of spontaneous [Ca2+]i transients than the 40x or 100x objectives ( Table 2). There was no significant difference between the 40x and lOOx objectives.

Discussion
The objective of the present investigation was to replicate the results reported by Lindstrom et al. (25). They examined MF modulation of [Ca2+]i in Jurkat cells in real time. This provided the opportunity for each cell to serve as its own control and to monitor cells before, during, and after exposure to MF. These studies demonstrated that in 80-85% of Jurkat cells examined, [Ca2+1] increased shortly after the introduction of an MF and returned to preexposure levels when the field was turned off (25-27. A key consideration in the identification  Figure 2. 'Data for experiments 1 and 2 were analyzed together by ANOVA with preexposure response as a covariate. There were no statistically significant differences (p < 0.05) between any of the exposure conditions. Experiment 3 was analyzed separately by ANOVA, and there were no significant differences (p < 0.05) between any of the exposure conditions. ip < 0.06 for the comparison of NF to MF in experiments 1 and 2 combined (ANOVA with preexposure response as a covariate). [Ca2+]i transients, only cells free from contact with other cells were included in the final analysis. These procedures are consistent with those reported by Lindstrom et al. (25)(26)(27). However, Lindstrbm et al. (25)(26)(27) included one additional criterion: a subjective morphologic assessment in the selection of a single cell within a microscopic field that would be monitored during MF exposure. This morphologic assessment was accomplished with phase observation through a lOOx objective, but was not feasible with our experimental setup and design.
In the present investigation, we exposed Jurkat cells to 50-Hz, 1.5-G MF in a attempt to replicate the findings of Lindstrom et al. (26) that this MF frequency and flux density induced an 85% incidence of cells responding with a elevation in [Ca2+]i. Although it was our intention to duplicate, to the extent possible, the protocol of Lindstrom et al. (25), early in our investigation we realized that the preselection of a subpopulation of acceptable cells based on morphologic considerations and personal judgment could not be duplicated in the absence of the original investigators. We also realized that this process could introduce a selection bias by either including or excludin, cells with a capacity for spontaneous [Ca +] transients. Therefore, we chose to use image analysis and initially include all of the cells within a microscopic field. This increased the number of cells we were able to examine by > 1 order of magnitude. We also administered field conditions in random fashion and data analyses were conducted without the knowledge of the experimental field conditions. This approach avoided any potential bias that could result from the nonrandom selection of cells and exposure conditions. Using this approach, we did not find any difference in the incidence of [Ca2+]i transients in cells exposed to a 50-Hz 1.5-G MF as compared to NF or to active field cancellation (SF).
The results are consistent with a recent report by Lyle et al. (15), who used the calcium probe Fluo-3 and flow cytometry to detect MF-induced transients in Jurkats exposed to 60 Hz at approximately 0.1 mT (1 G). Our data do not replicate the high response rate of cells to MF exposure as reported by Lindstrom et al. (26) (25). Consistent with present observations, Galvanovskis et al. (29) found that 50-Hz MF did not stimulate a change in [Ca2+]i in Jurkat cells with stable low-level [Ca2+]1i. However, these authors also examined the effects of MF on cells exhibiting sustained [Ca2+]i oscillations. The authors attributed the sustained oscillations to an "inhomogeneous" response to PLL, which was used, as in the present study, to secure Jurkats to cover slips. In the present investigation, cells exhibiting sustained lations were significantly reduced during exposure to 50 Hz MF as compared to fieldoff conditions in the same cells. Although the results reported by Galvanovskis et al. (29) are intriguing, the limited number of cells evaluated requires that follow-up studies be performed for verification. There is the potential for the present data set to be evaluated for this phenomenon, but because it consists of [Ca2+]i measurements on > 4,000 Jurkat cells, it will have to await an independent assessment. Another potentially important difference between the present results and those of Lindstrom et al. (25) pertains to the type of microscope objective. We conducted our studies with both 20x and 40x objectives, whereas Lindstrom et al. (25,26) used a IOOx objective. Too few cells were contained within the field of view at 100x (approximately [8][9][10][11][12] for the use of the statistical approach in the present study. On the other hand, a 100x objective provides superior assessment of morphology and therefore was the most useful method for the approach taken by Lindstrom et al. (25). Microscope objectives can perturb the MF (30), and the influence that this might have on the interaction with cells is not known. At the location of the cells, the maximum DC perturbation was 23, 26, and 22 mG (with no changes in field direction) for the  (32) reported that double flashes of 334 and 380 nm induced a [Ca2+]i transient in 18% of mouse neuroblastoma cells versus 0% in cells not exposed to the double flashes. The illumination intensity for epifluorescence microscopes is proportional to the square of the numerical aperture of the objective, and generally, higher magnifications have brighter illumination fields. Thus, a 40x objective would be expected to give a larger illumination intensity than a 20x objective. By this same reasoning, a 100x objective should provide more light throughput than a 40x objective. However, we did not observe a significant difference in the incidence of spontaneous [Ca2+>] transients for the latter two objectives. This could be due to other factors, such as number of lenses and types of coatings, that may further modify the light throughput.
Although it is virtually impossible to repeat every detail of an experiment, especially when investigator judgment is a factor, we set out to replicate the results of Lindstrom et al. (25). We selected an MF with the frequency (50 Hz) and flux density (1.5 G) that produced maximum results. We attempted to eliminate selection bias by including all cells that qualified for assessment based on the minimum requirements of Lindstrom et al. (25). We chose a technique that allowed us to evaluate [Ca2>]i transients in hundreds of individual cells. Finally, we replicated our own experiments several times using different microscope objectives. In the end, we found no effect of MF on [Ca2+]1 transients in Jurkat cells nor did we arrive at a satisfactory explanation for why we were unable to replicate the results of Lindstrom et al. (25). Perhaps there is a subpopulation from what we classified as qualified cells that Lindstrom et al. (25)  sure. It would also not include the large proportion of cells in our study that were nonresponders during MF exposure. Until this subpopulation is defined in a way that leads to reproducible identification, we conclude, based on our present results, that a 50-Hz, 1.5-G magnetic field does not effect [Ca2+]i in Jurkat cells.