The widespread deployment of 5G cellular networks alongside existing GSM technologies underscores the urgent need to investigate the potential biological effects of exposure to multiple radiofrequency electromagnetic fields (RF-EMFs) simultaneously. A new French study examined how simultaneous exposure to 5G-modulated 3.5 GHz and GSM-modulated 1.8 GHz signals affects the electrical activity of neurons, mitochondrial reactive oxygen species (ROS) production, and cellular stress protein responses in neurons and skin fibroblasts. At carrier frequencies between 1 and 4 GHz, which are typical for GSM and low-band 5G, the penetration depth into biological tissue is approximately 1 cm. Therefore, skin cells and the central nervous system, particularly neurons in the cerebral cortex, are of primary interest. Previous studies on sequential exposure suggest that the biological effects of multifrequency exposure may be more pronounced than those of single-frequency exposure. These studies underscore the need to understand how RF-EMFs interact at multiple frequencies.
Primary cortical neurons and immortalized human skin fibroblasts were exposed to RF-EMFs at specific absorption rates (SAR) of 1 and 4 W/kg for 15 minutes and 24 hours, respectively. Neuronal activity was analyzed using multielectrode arrays, and mitochondrial production of reactive oxygen species (ROS) was measured using MitoSOX Red. The authors investigated the activity of stress proteins using bioluminescence resonance energy transfer (BRET) assays that target RAS, PML, and HSF1 proteins. BRET is a cell-based technique that enables the real-time observation of protein-protein interactions and conformational changes in proteins. HSF1 (heat shock factor 1) is a stress-responsive transcription factor that activates the expression of heat shock proteins, such as HSP70 and HSP90, in response to protein denaturation or damage. RAS proteins act as molecular switches that control cell proliferation and division. PML regulates apoptosis, senescence, and antiviral responses.
The EMF exposure setups were based on two systems: a cell culture incubator that was converted into a Hall chamber exposure system for skin fibroblasts and an open transverse electromagnetic (TEM) cell exposure system for neurons. Signals from two generators were combined, amplified, and emitted into the Hall chamber and TEM cell. The 3.5-GHz signal used in this study is a genuine 5G New Radio (NR) signal, not just a pure 3.5-GHz sine wave. The same applies to the GSM signal.
Simultaneous exposure at 4 W/kg (2 W/kg per signal) did not significantly impact the burst and firing rates of cortical neurons. These results are consistent with previous findings by the same research group, which revealed that 5G-modulated 3.5 GHz signals with SAR values of 1 and 3 W/kg did not alter neuronal firing or burst rates in vitro [1]. However, a previous study demonstrated considerable inhibition of neuronal electrical activity at significantly higher SAR levels (28 W/kg). The underlying molecular mechanisms remain to be elucidated.
The present study found no effects on mitochondrial ROS production in fibroblasts at 1 or 4 W/kg. In BRET assays, minor effects were observed at 1 and 4 W/kg on basal activity of RAS and PML, as well as the maximal efficacy of PMA- and As₂O₃-induced RAS and PML activation (only at 1 W/kg for RAS and at both 1 and 4 W/kg for PML).
Combined exposure at 1 and 4 W/kg decreased the RAS basal activity according to BRET by 13% and 26% compared to the maximum potency of the specific toxin PMA on the RAS BRET probe. Similarly, exposure at 1 and 4 W/kg triggered slight increases in the PML BRET probe value, corresponding to approximately 16% and 13% of the maximum potency of the arsenic trioxide toxin. However, the maximum potency of PMA and arsenic trioxide in activating RAS and PML was slightly reduced (both are toxins that maximally activate these stress proteins).
No effects of combined 5G and GSM exposure on the basal activity of HSF1 were observed.
According to the abstract, there is no conclusive evidence of significant biological effects from simultaneous exposure to 5G and GSM. Any observed deviations were small and likely within the range of experimental variability. However, in their discussion, the authors present a more nuanced picture. They provide a detailed comparison of numerous similar studies, including some from their own French research group. They point out that many studies have detected subtle effects that often contradict earlier findings. The authors view these contradictory results as evidence of the highly complex interactions between RF-EMFs and biological systems. In studies that detect effects, it remains unclear how these effects arise.
Regarding ROS production in mitochondria, the present study found no effects at either 1 W/kg or 4 W/kg. This contrasts sharply with earlier findings that exposure to 5G-modulated RF-EMFs alone at 1 W/kg decreased mitochondrial ROS levels in fibroblasts [2], while exposure to 4 W/kg had no effect. Overall, these observations suggest that the combination of 5G and GSM signals modulates mitochondrial ROS production and stress protein signaling differently than exposure to a single frequency does. However, the heterogeneity of the results makes it difficult to draw clear conclusions, as the observed effects appear to depend not only on the type of signal (5G alone, GSM alone, or 5G+GSM), but also the specific SAR value, the cell type studied, and possibly other unidentified factors.
Editor’s note:
This is a well-planned and well-conducted study that identifies low-amplitude effects (compared to the effects of potent toxins that fully activate the investigated stress response pathways). Although the authors downplay this aspect somewhat, their results may reflect reality. It is possible that the effects of 5G+GSM exposure on isolated cells are only slightly harmful. But is it wise to ignore weak toxins? Why do many in vivo studies show stronger harmful effects than cell culture studies [3, 4]? One potential limitation of this study is its focus on pure neuron cultures. Numerous studies have demonstrated that neurons interact with astrocytes and other brain cells, such as microglia. Often, isolated neurons react differently than actual brain tissue. Tri-cultures would better simulate brain tissue.
The study authors provide a nuanced overview and summarize many earlier experimental findings. This illustrates why the field of electromagnetic fields has thus far been characterized by complexity, confusion, and insufficient causal relationships. These questions will likely only be resolved once the precise mechanisms by which electromagnetic fields influence biology are better understood. (AT)
1. Patrignoni L, Hurtier A, Orlacchio R, Joushomme A, Poulletier de Gannes F, Lévêque P et al. (2024). Evaluation of mitochondrial stress following ultraviolet radiation and 5G radiofrequency field exposure in human skin cells. Bioelectromagnetics, 45(3), 110–29. https://doi.org/10.1002/bem.22495
2. Canovi A, Orlacchio R, Poulletier de Gannes F, Lévêque P, Arnaud-Cormos D, Lagroye I et al. (2023). In vitro exposure of neuronal networks to the 5G-3.5 GHz signal. Frontiers in Public Health, 11, 1231360. https://doi.org/10.3389/fpubh.2023.1231360
3. Souchelnytskyi S (2025). Human cells response to electromagnetic waves of radio and microwave frequencies. The Ukrainian Biochemical Journal, 97(6), 5–22. https://doi.org/10.15407/ubj97.06.005
4. Weller SG, McCredden JE, Leach V, Chu C, Lam AK (2025). A scoping review and evidence map of radiofrequency field exposure and genotoxicity: Assessing in vivo, in vitro, and epidemiological data. Frontiers in Public Health, 13, 1613353. https://doi.org/10.3389/fpubh.2025.1613353