Tumour growth-promoting effect of high-frequency electromagnetic fields

A pilot study conducted by the Fraunhofer Institute in 2010 found that high-frequency electromagnetic fields (RF-EMF) had a growth-promoting effect on developing lung and liver tumours but not on other tumours in mice treated with a proven carcinogen. RF-EMF alone did not cause cancer.
A follow-up study conducted at Jacobs University Bremen confirmed the results of the pilot study: HF-EMF was shown to have a tumour growth-promoting effect on cancer caused by a carcinogenic substance. Here, too, the tumour growth-promoting effect was observed only for lung and liver tumours and partly lymphomas but not for brain, kidney, and spleen. Furthermore, no increase in tumour rates with increasing exposure category (exposure-effect relationship) was found.
Further research into possible mechanisms of action showed that RF-EMFs do not promote the DNA damage triggered by the carcinogenic substance and are thus not involved in tumour formation. The tumour-promoting effect of HF-EMF unfolds at a later stage, when the cancer has already developed.

Evaluation by the BfS

The research results of the pilot and follow-up study consistently show a tumour growth-promoting effect of RF-EMF for certain tumours in the selected mouse strain in the presence of a carcinogenic substance. The HF-EMF exposure alone did not trigger tumours nor were RF-EMF involved in tumour development. RF-EMF accelerated tumour growth when the cancer had already developed. The tumour growth-promoting effect was shown mainly for lung and liver tumours in the mouse strain used. The mechanism of action is unclear. Because the exposure of people (to UMTS signals) results in completely different intra-body field distributions (e.g. the fields do not reach the lungs or liver), this result is not directly applicable to humans. Moreover, no tumour growth-promoting effect of RF-EMF was found in other animal models such as rats or another mouse strain. Considering these study results as a whole, the BfS therefore assumes that the growth of tumours is not promoted in humans. The BfS is commissioning further research in order to examine whether the observations in the one mouse strain are an animal-model-specific effect that comes into play only under special experimental conditions or whether a hitherto unknown and possibly generally relevant mechanism of action could underlie it.

Pilot study

Methods

In the pilot study of the Fraunhofer Institute^[1], the tumour initiator (carcinogenic substance) ethylnitrosourea (ENU) was administered to pregnant mice. In addition, they were exposed to a high-frequency electromagnetic field according to the UMTS standard. This was to test whether RF-EMF amplify the effect of the cancer-causing ENU (i.e. whether they have a cancer-promoting effect without themselves being able to trigger cancer).

Five groups of animals were investigated:

Control: Animals in the breeding room treated neither with ENU nor with HF-EMF
ENU: Animals in the breeding room, not exposed to RF-EMF, and ENU administered
Sham exposure: Animals in EMF exposure facility but not treated with either ENU or HF EMF
ENU + HF-EMF: Animals in the EMF exposure facility, low RF-EMF exposure level (4.8 W/m²), and ENU administered
HF-EMF: Animals in the exposure facility, high EMF exposure level (48 W/m²), and not ENU administered

The HF-EMF exposure began in utero and continued until 24 months of age.

A special mouse strain (B6C3F1) was chosen as an animal model. This is particularly well suited for experiments with ENU because it tolerates ENU comparatively well. Despite developing tumours, the animals of this mouse strain live quite long and can thus be exposed to EMF in the long term.

Results

Control animals, sham-exposed animals, and animals exposed exclusively to RF-EMF (groups 1, 3 and 5) showed comparable tumour incidences. Both groups treated with ENU (Group 2 and 4) showed increased tumour rates. ENU-induced tumour incidences in the brain, kidney, spleen, and lymph nodes were relatively low and were not further influenced by the additional RF-EMF exposure in Group 4.

However, a significantly increased tumour rate in lung and liver as well as an increase in the number of metastatic lung tumours in the ENU/HF-EMF group (Group 4) compared with the ENU group (Group 2) was described. Because the tumour incidences in the liver as well as in all groups not treated with ENU (Groups 1, 3 and 5) were unexpectedly high compared with historical controls and all groups were infected with Helicobacter, it cannot be excluded that this infection and not the fields led to the increased cancer rates in the liver.

Follow-up study sponsored by the BfS

In order to verify the results of the pilot study, the BfS funded the study "Tumour promotion by high-frequency electromagnetic fields in combination with carcinogenic substances – synergistic effects" at Jacobs University Bremen. The study was completed in 2015; the final report (only in German) is available in the DORIS on-line repository of the BfS.

In this animal study, it was investigated with a larger number of test animals and at more exposure levels than in the pilot study whether RF-EMF of the mobile communications standard UMTS (1.97 GHz) can promote the development and growth of tumours.

The results of the pilot study were essentially confirmed and expanded^[2].

Methods

Female B6C3F1 mice were chronically exposed to whole-body RF-EMF of the UMTS standard (SAR 0.04 W/kg, 0.4 W/kg or 2 W/kg) or sham exposed for life starting in utero.
The mother animals received an injection with the tumour initiator ethylnitrosourea (ENU), which is absorbed by the foetuses.
In addition, a cage control that was neither exposed, sham exposed, nor treated with ENU was included.
In a comparison of the RF-EMF-exposed groups (SAR 0.04, 0.4, or 2 W/kg) with the sham-exposed control (SAR 0 W/kg), it was tested whether the combination ENU + RF-EMF resulted in more tumours at the end of the study period than ENU alone.

The organs investigated were the brain, lungs, liver, kidney, spleen, and lymph nodes.

Results

Although the results of the two studies differ in detail, the main findings of the pilot study were confirmed.

As in the pilot study the incidence of tumours in the brain, kidneys, or spleen was generally low (less than 10 %) in all ENU groups and was not further increased by RF-EMF.
The mean survival time in both studies was below that of the cage control (without ENU) in all ENU groups but was not further influenced by the additional UMTS exposure.
As in the pilot study, ENU caused cancer rates of more than 10% in the liver and lungs of the sham-exposed animals (cage control below 5 %). The number of liver carcinomas and lung adenomas increased significantly in all ENU and HF-EMF treated groups compared with the sham-exposed and ENU-only treated animals.
Unlike in the pilot study in which only one EMF exposure level was investigated, a significant increase in the incidence of lymphoma was observed in one of the study groups (ENU + HF-EMF 0.4 W/kg). In the higher exposed group (ENU + RF-EMF 2 W/kg), the incidence of lymphoma did not increase.
For some of the liver and lung tumour types investigated, the tumour-promoting effects were significant in the group with the lowest SAR value of 0.04 W/kg.
A dose-response relationship was not evident; this aspect was not investigated in the pilot study.

Additional dosimetric investigations

Even though the results did not indicate a dose-response relationship, the final report of the project considered a thermally mediated effect as the cause for the observations. The hypothesis that tumours could use the electromagnetic energy absorbed and converted into heat to grow faster was formulated.

In order to test this thesis, the dosimetry of the study was refined: From high-resolution magnetic resonance images, anatomically correct CAD models of mice of different ages were developed. These are representative of the developmental stage of the animals in the different project phases. These models were used to determine first the SAR and then the resulting temperature increases in the mice at the organ level by means of computer-assisted simulation methods (FDTD methods). Because in the ENU + RF-EMF exposed groups, the tumour incidence was increased only in the lungs and liver (and lymphatic system) but not in the brain, kidneys, or spleen compared with the ENU-only treated groups, correspondingly differentiated temperature increases in the respective tissues could support the hypothesis of the research participants. A distinction between diseased and healthy tissue (e.g. on the basis of modelling individual tumours) was not part of the project. The report on additional investigations (only in German) is published in the Digital Online Repository and Information System of the BfS.

The investigations showed that the lungs indeed had the highest averaged absorption (SAR) of all the tissues distinguished in the body models but that the mean temperature rise in this organ was only average. Liver tissue was unremarkable in terms of both SAR and temperature increase compared with other tissues such as kidney and brain, neither of which were affected by altered tumour incidence.

Another finding of the study was the observation that, particularly in the early stages of the project when the animals were small, exposure may have been considerably higher than planned. The reason for this is a field elevation at the bottom of the cage; this was not taken into account in the original dosimetry. The importance of this finding for the tumour occurrence observed is unclear. However, it reduces the possibilities of transferring the findings to exposure scenarios that actually occur in everyday life (see also relevance for humans)

BfS-funded study on mechanisms of action

In order to clarify the mechanisms of action leading to the present results, the BfS funded the study "Synergistische Wirkungen hochfrequenter elektromagnetischer Felder in Kombination mit kanzerogenen Substanzen – Kokanzerogenität oder Tumorpromotion?" at Jacobs University Bremen from the beginning of 2016 to 2017.

It was investigated whether the uptake and/or the DNA-damaging effect of ENU is enhanced by RF-EMF in the initial stage of the experiment (co-carcinogenicity) or whether ENU initially causes the same number of tumours independently of RF-EMF and these are then later promoted in their growth and spread by RF-EMF (tumour promotion). The study was completed in 2017 ^[3]; the final report report is published in the DORIS on-line repository of the BfS.

Methods

Initially, the procedure was identical to the previous experimental animal study: pregnant mice were exposed to RF-EMF according to the UMTS standard or sham-exposed (SAR 0, 0.04 or 0.4 W/kg) and received an injection of ENU. In order to investigate the early effects of the combination of ENU and HF-EMF, which would indicate a co-carcinogenic effect (see above), some of the animals were euthanised 24, 36, and 72 h after the administration of ENU. The foetuses were removed, and the extent of DNA damage in the brain, lungs, and liver was investigated by immunohistochemical fluorescence staining.

Results

The least DNA damage was found in the brains; the most was found in the liver. The statistical analysis showed no significant differences between the different exposure groups, including the sham-exposed group. Thus, RF-EMF do not promote DNA damage by ENU; it is therefore not a co-carcinogenic effect. This means that the increased cancer rates in liver and lung observed in the mouse animal model are due to a tumour-promoting effect occurring later.

Discussion

The mechanism of action underlying the described tumour growth-promoting effect is unclear. There is evidence that the dielectric tissue properties of liver tumours and healthy liver tissue differ in humans^[4] and that RF-EMF are more strongly absorbed in tumours. Liver tissue is similar in all mammals; this finding thus probably also applies to mice. The additional energy absorbed could be used by the tumour tissue for faster growth.

The increase in tumour rates in individual organs hardly correlated with the organ-specific SAR values and the temperature increase. Differences in exposure are therefore probably not responsible for the organ-specificity of the tumour-promoting effect. It is striking that the increase in tumour rates after RF-EMF exposure occurred in the organs in which ENU caused a high number of DNA damages and in which, as a consequence, tumour rates were already relatively high without RF-EMF exposure. The tumour-promoting effect of exposure may not be organ-specific but rather become apparent only after many tumours have already been induced by ENU. The organ specificity of the effect of ENU depends on when and how ENU is administered and which animal species or strain was used.

It is also possible that the metabolism of the mice at the level of the whole organism is affected by the exposure to electromagnetic fields. Several studies show that the absorption of electromagnetic field energy can lead to an increase in body weight^[5] or a decrease in feed intake and metabolic turnover rate in rodents ^{[6, 7]}. These observations were made at whole-body exposures of 0.4–4 W/kg. If the metabolism of the mice at the level of the whole organism is affected by the exposure to electromagnetic fields and this had an influence on the tumour incidence, this would fit the low correlation of the tumour incidence with the organ-specific SAR and temperature values.

Relevance for humans

The results of animal experiments can be transferred to humans only to a limited extent. In the case at hand, the difficulty lies in particular in the fact that the cause of the findings may lie in the area of energy metabolism and the associated thermoregulation. These are physiological processes that depend strongly on body size and differ substantially between rodents and humans. The metabolic turnover rate of humans is lower; however, thermoregulation is much more efficient than in rodents. In addition, in small animals such as mice, the internal organs (e.g. liver and lungs) are reached more by the fields than is the case in humans. The reason for this is that the exposure of humans to mobile communications signals in the UMTS range results in completely different intra-body field distributions. Therefore, the question of the transferability of the results to humans cannot be answered for any of the studies.

It is difficult to transfer the results to other animal models. In rats, for example, it is possible to induce brain tumours with ENU; however, tumour promotion by RF-EMF (in this case according to the GSM standard) does not occur^[8]. Whether this is due to the animal species, the HF-EMF frequency, or other methodological differences is unclear. In another mouse strain (AKR mouse), which genetically develops more lymphomas, RF-EMF of the GSM or UMTS standard had no cancer-promoting influence^{[5, 9]}.

In the BfS-sponsored follow-up study, an exposure that was below the maximum recommended value for whole-body exposures in humans (0.08 W/kg) was chosen for the lowest-exposed group (0.04 W/kg). However, in the everyday life of the general population, such whole-body exposures do not occur. Relevant sources are mobile communications base stations. Typically, the limits applicable to such installations are exhausted by up to 1 % (and in a few scenarios up to 10 %). In the case of local exposures caused by the use of mobile communications end devices (e.g. smart phones), the maximum permissible value of 2 W/kg can be exhausted to a higher percentage. However, these are short-term exposures in which a possible slight local temperature increase is quickly offset. Human exposure during a cell phone call is not comparable to the permanent exposure of the mice here.

The results cannot be used to derive limit values. However, they support the recommendations of the BfS to minimise the exposure to HF-EMF.

Summary

In the overall view of the available study results, the BfS does not assume a tumour growth-promoting effect by high-frequency electromagnetic fields in humans. The BfS is commissioning further research in order to examine whether the observations in the one mouse strain are an animal-model-specific effect that comes into play only under special experimental conditions or whether a hitherto unknown and possibly generally relevant mechanism of action could underlie it.