Insights into genotoxic effects of electromagnetic fields

The increasing use of appliances, which generate electromagnetic fields (EMFs), has
provoked public concern about their safety. Scientific research into possible health effects
however produced conflicting results. One of the open questions is whether or not EMF
exposure has genotoxic effects. Therefore, the main objective of my thesis was to
investigate DNA damage formation and repair, cell cycle progression, apoptosis and DNA
damage signalling in cultured human cells under EMF exposure. In particular, the nature of
possible genotoxic effects and the mechanisms underlying the cellular responses were to be
addressed.
As in the past, genotoxic effects of EMF exposure often could not be reproduced in
independent studies, I first aimed at the validation of results of previous studies [1-3]. In
these studies, different genotoxicity tests revealed increased DNA damage after exposure of
human fibroblast cells to EMFs in the low frequency range, as used in power lines, as well as
in the radiofrequency range, as applied by mobile phones and wireless technologies. I could
show that genotoxic effects of 50Hz EMFs can be reproduced independently. Effects of
radiofrequency EMF exposure, however, were detectable only in one particular cell line (HR-
1d), but not in the cell line used in the original study (ES-1). Because the visual scoring
method of DNA fragmentation analysis (comet assay) used in the previous studies was
criticized in the scientific community, I compared this method with an automated
computerized comet analysis. This established that increases of DNA fragmentation following EMF exposure are detectable in both types of analyses.
Expanding the study to other cell lines, I was able to show, that 50Hz EMF exposure in two
different fibroblast cell lines but not in the cancer cell line HeLa lead to comet assay effects.
Furthermore, I showed that DNA fragmentation is not found in G1 blocked cells, suggesting
replicating cells to be involved in EMF directed effects. This indicated, that the DNA
fragmentation detected following EMF exposure might not reflect direct induction of DNA
damage but rather an EMF dependent alteration of the S-phase population. Furthermore
apoptosis was suggested as confounder for comet assay effects before. Addressing this
question, I found decreased replication efficiency and an increased apoptotic fraction after
50Hz EMF exposure in the fibroblast cell line showing the higher comet assay effect.
Therefore, I conclude that these cells encounter problems in entering S-phase or progression through S-phase, which could lead to apoptosis and, hence, apoptotic DNA fragmentation, in
a subpopulation. These effects, however, cannot entirely explain the genotoxicity observed,
as the fraction of cells with increased DNA fragmentation was higher than the proportion of
apoptotic cells.
I then addressed the type of possible DNA damage generated by EMF exposure. An inhibitor
of the DNA single strand break (SSB) sensor poly-ADP-ribosylation polymerase was used to
examine an engagement of DNA single strand break repair following EMF exposure. The
results showed that the increase of DNA fragmentation did not change further by applying
both inhibitor and EMF exposure compared to inhibitor or EMF exposure alone. Therefore
the effects appear to be epistatic, indicating that EMF exposure may affect DNA SSB repair
rather than inducing DNA damage itself. To address the occurrence of DNA double strand
breaks or stalled replication forks, I made use of phosphorylated H2AX (γH2AX) as a marker
in EMF exposed cells. This revealed no difference between non-exposed and exposed cells,
suggesting that the increase in DNA fragmentation is unlikely due to such lesions.
Biological effects of EMF exposure were hypothesized to reflect an influence on the free
radical pool of cells and, thus oxidative stress. I examined the steady state levels of oxidative
DNA damage after EMF exposure and found no indications for increased generation of
indicator lesions. This result fails to support the hypothesis of EMF induced oxidative stress,
although I cannot completely rule out small changes of a sub-detectable level. DNA base
excision repair (BER) is the system specifically repairing small lesions including oxidative DNA
base damage. To examine, if this pathway is activated during EMF exposure, I examined the
formation and levels of nuclear XRCC1 foci. XRCC1 is a central component of the BER system
and can be seen to localize to sites of DNA damage and repair. However, immunostaining of
XRCC1 revealed no difference in numbers and distribution of foci following EMF exposure.
Adding a DNA Polymerase β (the BER polymerase) inhibitor, however, the subG1 fraction of
cells increased synergistically with ELF-EMF exposure. This could indicate, that either BER
protects cells from entering apoptosis following EMF exposure or that the DNA damage
generated by inhibiting DNA Polymerase β is less efficiently processed under EMF exposure.
Taken together, these results suggest, that the small increase in DNA fragmentation
observed in human fibroblasts exposed to 50Hz EMFs can be accounted for by a combination
of effects including impaired repair of endogenously arising DNA damage, disturbance of Sphase
progression and apoptosis in a small fraction of cells, rather than by directly induced DNA damage.
In a second part of my thesis, I used the highly sensitive comet assay, cell cycle analysis and
immunofluorescence staining technologies established for the EMF studies to contribute to
different projects addressing regulatory aspects of DNA BER. In a first study, we showed that
Thymine DNA Glycosylase (TDG) levels were cell cycle regulated and TDG is absent in Sphase
in biochemical assays. Regulation occurs at the protein level, as mRNA levels remain
constant throughout the cell cycle. The protein is ubiquitinated and degraded by the
proteasome. To provide biological evidence for such a regulation in vivo, I stained cells with
antibodies for TDG and the S-phase marker PCNA by immunofluorescence and counted cell
numbers of double and single stained cells. PCNA positive cells did not stain for TDG and vice
versa. As PCNA is a marker for S-phase, this shows, that TDG is absent in S-phase cells.
In a second study we provided evidence for a regulation of DNA Polymerase β (DNA Pol β) by
protein arginine methylation. This methylation has impact on its in vitro performance like
DNA binding and processivity, but an in vivo relevance of this modification remained to be
shown. I showed that DNA Pol β knock out cells complemented with a mutated form of DNA
Pol β, not able to be methylated, showed a higher level of DNA fragmentation upon induced
DNA damage than cell complemented with wild type DNA Pol β. Together with reduced
survival rates and an increased subG1 fraction in cells challenged with a DNA damaging
agent, this established the in vivo relevance of DNA Pol β methylation. Arginine methylation therefore might represent a novel regulatory protein modification in DNA BER.
In a third study, I contributed to the investigation of the toxicity mechanism of the
chemotherapeutic drug 5-fluorouracil (5-FU), which is not fully understood so far. An
involvement of the BER enzyme TDG was suggested by biochemical evidence, leading to the
question, if TDG wild type and knock out cells respond differently to 5-FU. TDG knock out
cells displayed hypersensitivity to 5-FU, which suggested a deleterious repair mechanism
through TDG, probably leading to the induction of DNA SSBs. I indeed found increased DNA
strand breaks in TDG wild type cells compared to knock out cells, while XRCC1, a marker for
BER, was more activated in knock out cells. In cell cycle analyses 5-FU induced accumulation
on S-phase of TDG deficient cells was less pronounced than in wild type cells. This suggests
that TDG contributes to 5-FU mediated cytotoxicity, probably by inducing DNA SSBs due to
its slow turnover rate and the resulting saturation of BER, leading then to checkpoint
activation and S-phase accumulation.