Experimental Platform for Investigation of Low-Frequency Magnetic Field Effects on Cells

The workflow scheme (Fig. 1) outlines the design of a paired experiment suitable for the investigation of the effects of LF MF on microbial cultures. The workflow starts with a yeast strain from The Culture Collection of Yeasts (CCY) (Institute of Chemistry, Slovak Academy of Sciences, Slovakia) maintained on a solid medium in a Petri dish and used as inoculum for pre-cultivation.

Fig. 1.

The workflow diagram illustrates the experimental design, with the experimental platform serving as the central component.

After completing the pre-cultivation step, the yeast culture is diluted accordingly, distributed in vials, and placed in the cultivation chambers. The chambers are equipped with temperature regulation and temperature monitoring systems to ensure equal environmental conditions in both chambers. The integrated magnetic system enables the application of LF MF via two channels, both of which can be set to exposure or control mode. After cultivation, cell concentration measurements are performed, followed by subsequent data analysis with data control and filtration. Finally, selected data is subjected to statistical analysis.

YPD medium: 10 g yeast extract, 20 g peptone, 20 g glucose (MkB test a.s., Slovakia). YPD agar medium: 10 g yeast Extract (Biolife Italiana S.r.l., Italy), 20 g peptone (Peptone from meat, bacteriological, Sigma-Aldrich, Germany), 20 g D-glucose (Slavus s.r.o., Slovakia), agar 20 g (Agar Bios Special LL (Agar bacteriological), Biolife Italiana S.r.l, Italy) in 1 l medium. Trypan blue solution (Sigma-Aldrich, Germany).

The Saccharomyces cerevisiae strain CCY 21-4-99 was cultivated in YPD medium at 35 °C and stored long-term on YPD agar in a Petri dish. Prior to the experiments, a full inoculation loop of the culture was transferred from the Petri dish to 50 ml of liquid YPD medium in a 100 ml Erlenmeyer flask. The flask was shaken to ensure homogeneity and then incubated statically at 35 °C. After 24 hours, the culture was rapidly cooled on ice, stored at 7 °C, and used as a pre-culture for further experiments within 4 to 10 days. The viability of the pre-culture was determined using a Bürker counting chamber after staining the cells with Trypan blue, with the percentage of viable cells ranging from 98.3 % to 99 %. Microbiological purity was assessed at each step of the experiment to ensure that no bacterial or fungal contamination was present visually and/or microscopically. In the rare cases where contamination occurred, the results were not taken into consideration. All experimental vials used for yeast cultivation were designed with flat bottoms to ensure an even distribution of growing cells. The cell cultures were immediately placed in an ice bath upon removal from the incubator to halt yeast cell growth prior to measurement. The biomass concentration was determined using a Bürker counting chamber or by turbidity measurement. The relationship between turbidity and cell concentration was established at 4.4 × 10⁷ cells/ml, corresponding to 400 NTU, based on microscopic cell counting.

The magnetic system Fig. 2 for generating MF consists of two independent channels, each of which is responsible for supplying MF to one of the cultivation chambers. At its core is a signal generator (Owon, XDG3102, China), which is equipped with two channels, each of which generates alternating current signals. These signals are directed to a dedicated amplifier (NX3000D, Behringer, Germany), which has two separate inputs and outputs that allow independent amplification for each channel.

Fig. 2.

Schematic diagram of the MF generation system, illustrating the signal generator, amplifier, rheostats, Helmholtz coils in two chambers, and an oscilloscope with ammeters.

After amplification, the signals are routed through a rheostat, which – together with the 0.4 Ω resistance of the Helmholtz coils – can be set to a total resistance of 2 Ω, ensuring that the system impedance conforms to the amplifier. An oscilloscope (Owon, XDS3102, China) is connected to both channels, allowing real-time visualization of the waveforms and verification of frequency and amplitude. An ammeter is connected in series in each circuit to monitor the electrical current. This configuration ensures stable control of the MF throughout the exposure period. In addition, the MF system offers flexibility as MF activation can occur in either one or both chambers simultaneously, depending on the experimental requirements.

A quasi-homogeneous MF from Helmholtz coils was developed for the cell culture exposure. The coil parameters (coil diameter, wire diameter, number of turns, and number of wire layers) were optimized within the specified MF frequency range (Hz-kHz) and heat production of the coils (≤ 2 W) by modeling in MATLAB (version R2023b, MathWorks, Inc.) (Fig. 3(a)), with the aim of maximizing magnetic flux density in the center of the coils. Since the impedance and power dissipation increase with more turns, the amplifier cannot keep the current constant. We also had to enforce a strict power dissipation limit to prevent the chamber from overheating, resulting in the non-monotonic field curve shown in (Fig. 3(a)). The modeling led to the construction of two coils with a 110 mm diameter, made of 2.2 mm copper wire with 50 turns each (Fig. 3(b)). The inductance of the coils was 0.4 mH at 1 kHz. The coil support and other accessories were 3D printed.

Fig. 3.

(a) Modeling of the maximum magnetic flux density as a function of the number of coil turns and the wire diameter. (b) Details of the Helmholtz coils and the foil heating system on the chamber walls.

The arbitrary waveform generator provides an adjustable frequency and amplitude for the MF. The signal amplifier enhances AC currents up to 8 A (rms) to generate MF in the frequency range of 0–2 kHz, achieving a magnetic flux density of up to 5 mT. In the central part of the coils, where the biological experiment takes place, the homogeneity of MF is 95 %, as experimentally verified with a gaussmeter (model 5180, F.W. BELL, USA) at 15 equidistant points covering the entire experimental volume. Identical coils were placed in both chambers to ensure the same conditions. The measured maximum difference between the MFs was up to 1 %.

The customized thermal chambers were developed using a glass-door refrigerator (Candy CCV 150 EU, Italy). The housing consists of a steel sheet outer shell, polyurethane (PUR) foam insulation, and an inner ABS plastic lining. The dimensions of the appliance are 500 mm (width) × 560 mm (depth) × 840 mm (height), with an effective internal volume of 122 l. The chambers were heated with heating foils (Ultratherm Viv Mat 22, UK) attached to both side walls of the chamber (Fig. 3(b) and Fig. 4(a)). The measured contribution of the foils to the MF inside the coils was below the detection limit (0.005 mT); therefore, its impact on our experimental setup can be considered negligible. The temperature was controlled by a PID temperature controller (Rex C100V Pt100, rKc Instrument, Japan) equipped with a sensor that was immersed in the water bath holding the experimental vials. In addition, a temperature monitoring system was implemented using a microcontroller board (Arduino UNO Rev3, Italy) with four digital temperature sensors (DS18B20, Analog Devices, Inc., USA) and a sampling period of 10 s.

Fig. 4.

(a) Thermal image of the cultivation chamber. (b) Temperature stability test in the chamber with the PID controller set to 35 °C. T1 – top of the cultivation chamber, T2 – water bath in the beaker, T3 – bottom of the cultivation chamber.

A series of temperature distribution tests were carried out inside the chamber, with several sensors placed simultaneously at different locations. Here we show one such test performed with a PID controller set to 35 °C: one sensor was immersed in the water bath containing vials of water, while two other sensors were positioned at different heights within the chamber (Fig. 4(a)). Fig. 4(b) shows the inhomogeneous temperature distribution in the chamber. Initially, the system required time to stabilize, reaching different steady-state levels and exhibiting different thermal dynamics over time. Although temperature differences of more than 5 °C were observed between the different parts of the chamber, the submerged sensor in the water bath consistently recorded stable values. The depicted test scenario was more challenging as the coils were active and served as an additional heat source. Nevertheless, the target temperature of 35 °C was successfully maintained, which corresponded to the conditions observed when the coils were inactive. Under these control conditions, the temperature distribution within the chamber exhibited a more typical gradient, with temperatures gradually increasing from the bottom to the top of the box.

The temperature curves clearly show that it is necessary to place the temperature sensor used for regulation directly at the sample location – a practice that has not been consistently followed in similar studies. The initial overshoot of about 0.6 °C observed in the red curve was the result of the PID controller's response to opening the incubator and placing the preheated samples (almost 35 °C) in a water bath that had already stabilized at 35 °C. This transient temperature rise was unavoidable under the given conditions. More importantly, similar overshoots were recorded simultaneously in both chambers, regardless of whether they were under test or control conditions. As a result, not only was the mean sample temperature of 35 ± 0.3 °C achieved in both chambers, but also the mean temperature difference between the two chambers remained within ± 0.3 °C throughout the experiment.

Based on these tests, the “regulation” and monitoring sensors were placed close to each other and exclusively in the water bath of each cultivation chamber to monitor the temperature of the yeast samples.

For the optimal experimental setup, two interdependent parameters had to be determined: the initial cell concentration and the total cultivation time. As these parameters could not be determined simultaneously, they were determined iteratively. On the one hand, extending the duration of the experiment increases the probability of detecting the MF effect. On the other hand, practical considerations, such as limiting each run to a single day (24 h), result in a maximum cultivation window of approximately 20 hours, allowing at least 2 hours for preparation and post-experiment procedures. Knowing this approximate time frame, the next step was to look for suitable dilution levels that would allow the cells to remain in the desired growth phase during the available cultivation period.

In the search for the optimum initial concentration of Saccharomyces cerevisiae CCY 21-4-99, biomass growth was first assessed after 20 hours of stationary cultivation at 35 °C as a function of the initial cell density. The biomass growth, which was measured turbidimetrically (turbidimeter Lutron TU-2016, Taiwan), was expressed as a percentage relative to the 24-hour pre-culture (Fig. 5). The data show that only the sample containing 10 % of the pre-culture (i.e., a 1:10 dilution) achieved a biomass equivalent to 100 % of the pre-culture. In contrast, biomass growth of 50 % or less was observed at dilutions greater than 1:3.2×10⁴, indicating a significant decrease in growth. At a dilution of 1:10⁷, biomass growth was already negligible (2.2 %), illustrating the sensitivity of growth to lower initial cell density.

Fig. 5.

Cell concentration ( ) after 20 hours of cultivation as a function of the initial inoculum dilution.

Our target window for final biomass growth was between 20–60 %, as detecting differences in paired experiments becomes more challenging when final concentrations are either too low or too high (see explanation below and Fig. 6(b)). To further refine this range, we tested pre-culture dilutions from 5×10⁴ to 2×10⁵. Based on these experiments, the final pre-culture concentration corresponded to a 1.8×10⁵-fold dilution of the original pre-culture (results not shown), which corresponds to a calculated value of 306 cells/ml.

Fig. 6.

(a) Growth curve showing the increase in cell concentration of Saccharomyces cerevisiae CCY 21-4-99 at 35 °C over time (error bars ±SD). (b) Simulation of MF impact by delaying growth by 2 hours (blue curve). The vertical line at 18^th h indicates the moment of maximum difference between the two growth curves.

For the selected initial cell concentration (306 cells/ml), the culture was cultivated in parallel in two identical thermostatic chambers, and biomass growth was monitored over time. The yeast cell concentrations were measured turbidimetrically from the 12^th hour onwards, with a total of 14 measurement time points. At each time point, samples were taken from three vials per chamber. In this way, two growth curves were simultaneously obtained, one in each chamber. The data shown represent mean values of these two curves, fitted with a sixth-order polynomial, including an extrapolation to a single point at 36 hours (Fig. 6(a)).

The cell growth curve shows that the biomass levels remained below half of the fully developed biomass at 37 hours until the 19^th hour. In particular, the data show that the exponential growth phase continued until the 19^th hour, with the biomass levels remaining below 50 % of the stationary phase maximum (Fig. 6(a)). During this early growth period, the biomass doubled within approximately 2 hours or less. Such a growth curve is characteristic of stationary cultivation, where the cells grow at the bottom of the vials without mixing or agitation.

Next, it was necessary to determine a specific time point at which to terminate the experiment – a crucial step in optimizing the experimental design. The goal was to estimate a time range that would maximize the probability of detecting differences between the magnetic field and control conditions. Without this consideration, the most promising time window for observing potential effects could be missed, especially if the effect is subtle. Therefore, we attempted to estimate the optimal time point at which the differences are most pronounced and therefore most likely to be detectable. Otherwise, even if an effect were present, it could remain undetected if the measurements were taken at suboptimal time points.

To guide this decision, we used a simple modeling approach based on the assumption that any environmental influence, including MF, could shift the growth curve over time: either to the left (implying stimulation) or to the right (implying inhibition). The direction of the shift was not critical for this purpose. Rather, the goal was to determine the time point at which the difference between the shifted and unshifted curves is most pronounced. In Fig. 6(b), the blue curve illustrates the inhibition scenario, where the growth process is delayed by 2 hours compared to the original red curve.

The scenario illustrated clearly shows that for shorter experiment durations (t < ~13 h), the two curves do not diverge sufficiently to detect differences. Conversely, for longer durations (t > ~25 h), the curves begin to converge, and the differences become smaller. Thus, the largest gap occurs within an intermediate time window of 15–25 hours. Additional modeling with different time shifts – including minimal shifts on the order of 0.1 h – was performed to simulate subtle differences, as the goal was to develop a platform sensitive enough to capture small MF effects. Ultimately, the experiment duration was set at 18 hours under the given conditions, during which the yeast population underwent an estimated 16 consecutive doublings.

In this section, we present the proof of concept for the developed experimental platform. The paired experiment was conducted over a period of 18 hours, with five test sample vials placed in each chamber and operated in two modes: (1) both chambers under control conditions without MF, and (2) both chambers exposed to a 50 Hz MF of 3.5 mT, driven by an electric current of 4.5 A. The temperature in both cultivation chambers was maintained at 35 °C throughout the experiment.

For each experimental run, the environmental conditions were evaluated using three parameters: the mean temperature in chamber 1, the mean temperature in chamber 2, and the mean temperature difference between the two chambers. Only experimental runs that met the following three temperature criteria were included in further analysis: the mean temperature in chamber 1 and in chamber 2 had to remain within 35±0.3 °C and the mean temperature difference between the chambers had to be within ±0.3 °C.

In Fig. 7, the cell concentration difference between the cultures in the two chambers in a paired experimental setup is illustrated by boxplots for three data sets: simultaneous operation of both chambers in control mode with MF switched off (N = 5), both chambers in test mode with MF switched on (N = 5), and a combined set merging the previous two (N = 10). In all three cases, the mean difference in cell growth between the two chambers remained below 1 %. Statistical analysis using a one-sample t-test did not reject the hypothesis that the mean values of these sets were equal to zero (p-values = 0.56, 0.78, and 0.52, respectively). These results indicate that the two chambers provide sufficiently consistent conditions for cell growth within the given experimental settings.

Fig. 7.

Boxplots for cell concentration differences between cultivated cultures in two chambers.

C/C: both chambers in control mode without MF (N = 5).

T/T: both chambers in test mode with MF (N = 5).

Combined: merged data from the two previous sets (N = 10).