Abstract—There is a substantial number of studies showing biological effects of microwaves of extremely high-frequency range [i.e., millimeter waves (MMWs)] at nonthermal intensities, but poor reproducibility was reported in few replication studies. One possible explanation could be the dependence of the MMW effects on some parameters, which were not controlled in replications. We studied MMW effects on chromatin conformation in Escherichia coli (E. coli) cells and rat thymocytes. Strong dependence of MMW effects on frequency and polarization was observed at nonthermal power densities. Several other factors were important, such as the genotype of a strain under study, growth stage of the bacterial cultures, and time between exposure to microwaves and recording of the effect. MMW effects were dependent on cell density during exposure. This finding suggested an interaction of microwaves with cell-to-cell communication. Such dependence on several genetic, physiological, and physical variables might be a reason why, in some studies, the authors failed to reproduce the original data of others.
Index Terms— Biological applications of electromagnetic radiation, biomedical effects of electromagnetic radiation, genetics, polarization.
MICROWAVES in the frequency range of 30–300 GHz are often called millimeter waves (MMWs) because the wavelength in vacuum belongs to the interval of 1–10 mm. The biological effects of MMWs have been studied for over 20 years starting with investigations of Webb , Vilenskaya et al. , Devyatkov , and Gründler et al. –. Several reviews were devoted to the effects of MMWs , –. The most recent review summarized more than 100 MMW investigations in bi- ology and medicine and indicated several problems in this field of research . One of them is the question about so-called nonthermal effects.
Due to the efficient absorption of MMWs in water solutions and biological tissues, significant variations in specific absorp-
Manuscript received November 10, 1999; revised May 3, 2000. This work was supported by the Swedish Council for Work Life Research and by the Swedish Radiation Protection Institute.
I. Y. Belyaev is with the Department of Molecular Genome Research, Stock- holm University, S-106 91 Stockholm, Sweden and also with the Department of Radiation Physics, Biophysics and Ecology, Moscow Engineering Physics Institute, Moscow 115409, Russia.
V. S. Shcheglov, E. D. Alipov, and V. D. Ushakov are with the Department of Radiation Physics, Biophysics and Ecology, Moscow Engineering Physics Institute, Moscow 115409, Russia.
tion rate (SAR) is observed through an irradiated sample. Khizh- nyak and Ziskin  found specific microoscillations of temper- ature in irradiated water solutions. Such phenomena were sup- posed to explain at least some bioeffects of MMWs. MMW ir- radiation of thin layers results in significant heating at power density (PD) above 1 mW/cm . MMW bioeffects at this and higher levels are usually attributed to induced heating. Never- theless, the observed MMW effects were not always explained by heating, even at the thermal levels of exposure .
The well-known example for nonthermal effects of MMWs is the study of Gründler et al. –. For over ten years, this group consistently reported the resonance effects of MMWs on the growth of yeast cells. Different exposure systems and analyt- ical facilities were used, leading to the same conclusions about resonance response of yeast cells to nonthermal MMWs. So- phisticated system for image processing recognition was used, which allowed a very precise analysis of the cell cycle in indi- vidual cells. The effects were observed at PD of 10 W/cm and could not be explained by heating .
Despite of a variety of reported MMW bioeffects, only a few independent replications were performed –. The apparent conclusion of these replications is that the original data on MMW effects are poorly reproduced in independent experi- ments.
Significant effects of nonthermal MMWs on the chromatin conformational state (CCS) in Escherichia coli (E. coli) cells and thymocytes of rats have been observed by our group –. MMW effects on CCS were dependent on several physical, physiological, and genetic parameters. The data suggested that a number of variables should be controlled in original experiments and in replication studies. In this paper, we describe the dependence of MMW effects on all these parameters based mainly on the data obtained by our group and in comparison with the recently published data of others.
II. AVTD TECHNIQUE
The main body of results analyzed in this paper was obtained with the method of anomalous viscosity time dependence (AVTD). This technique is based on the radial migration of high molecular weight DNA–protein complexes such as nu- cleoids in rotary viscometer . The physical model of AVTD was developed by Kryuchkov et al.  based on the theory of radial migration –. The changes in AVTD were observed in E. coli cells of several strains and rat thymocytes after exposure to microwaves in vitro –. The AVTD changes have been also observed upon treatment of cells with DNA/chromatin-specific chemicals such as ethidium bromide (EtBr) and etoposide VP-16 , , . Several experi- mental observations have suggested that an increase in AVTD in response to MMWs is caused by relaxation of DNA domains and, consequently, decrease in AVTD is caused by chromatin condensation. Single-cell gel electrophoresis confirmed this suggestion .
III. FREQUENCY AND POWER DEPENDENCIES OF MMW EFFECTS
Effects of low-intensity microwaves on repair of radiation-in- duced DNA breaks were studied by the AVTD method in E. coliK12 AB1157 . Significant suppression of repair was found when X-irradiated cells were exposed to microwaves within fre- quency ranges of 51.62–51.84 and 41.25–41.50 GHz. In both ranges, the effect had a pronounced resonance character with resonance frequencies of 51.76 and 41.32 GHz, respectively , . The effect of microwaves did not depend on the se- quence of cell exposure to X-rays and MMWs. The MMW ef- fect could not be explained by heating. First, statistically sig- nificant suppression of repair was observed at a very low PD of 1 W/cm . Second, no suppression of repair was observed upon heating of cell suspension by 5 C. Third, the PD averaged over the exposed surface did not depend on frequency within ob- served resonances.
It was established that the reduction of PD resulted in significant narrowing of the resonance response of E. coli cells to MMW exposure , . Ups to 15 frequencies were investigated inside each resonance range and all frequency dependencies obtained fitted well to Gaussian distribution . The experimental conditions allowed determination of the resonance frequency with an error of 1 MHz. Within this error, the resonance frequency of 51.755 GHz was stable with decreasing of PD from 3 10 to 10 W/cm . At the same time, the half-width of the resonance decreased from almost 100 to 3 MHz. The dependence of half-width of the 51.755-GHz resonance effect on PD had the steep decrease from 3 10 to 10 W/cm followed by slow decreasing from 10 to 10 W/cm . The question then arose: what happened in the frequency range of 51.65–51.85 GHz upon narrowing of the 51.755-GHz resonance? The cell response to MMWs at a PD of 10 W/cm was studied in this frequency range , . Three additional resonances were detected: 51.675 0.001, 51.805 0.002, and 51.835 0.005 GHz. The half-widths of all resonance including the main one, i.e., 51.755 0.001 GHz, were about 10 MHz at the PD of 10 W/cm . Therefore, sharp narrowing of the 51.755-GHz resonance in the PD range from 3 10 to 10 W/cm was followed by an emergence of new resonances. These data were interpreted as a splitting of the main resonance 51.755 GHz . Dependence of the MMW effect on PD was investigated at one of these resonances, i.e., 51.675 GHz . This dependence had the shape of a
“window” in the PD range from 10 to 10 W/cm . It is important that no MMW effect was observed at subthermal and thermal PDs. This type of PD dependence clearly indicated nonthermal mechanism of the MMW effects observed. The frequency dependencies were studied around 51.675 GHz at different PDs and this resonance frequency was shown to be stable within the range of 10 –10 W/cm . Along with disappearance of the 51.675-GHz resonance response at a higher PD of 10 –10 W/cm , a new resonance effect arose at 51.688 0.002 GHz . This resonance frequency was also stable within the studied PD range. Taken together, these data strongly suggested a sharp rearrangement of resonance spectra, which was induced by MMWs of the subthermal PD range. The half-widths of three studied resonances showed rather different dependencies on the PD, changing from 2 to 3 MHz to 16 to 17 MHz (51.675 and 51.668 GHz) or from 2 to 3 MHz to 100 MHz (51.755 GHz) , .
Significant narrowing in resonance response was found when studying the growth rate in yeast cells  and chromatin con- formation in thymocytes of rats . In the study of Gründler et al., the half-width decreased from 16 to 4 MHz as the PD was decreased within the range of 10 –10 W/cm . Based on these studies with different cell types, one may assume that narrowing of the resonance upon decease in the PD is one of the basic regularities in cell response to MMWs. On the other hand, different dependencies of a half-width on the PD may be expected for different resonance frequencies.
It was established that the dependence of the MMW effects on the PD had a linear section followed by a plateau . This type of PD dependence was observed in , and –. The data obtained in experiments with E coli cells and rat thymocytes provided new evidence for this type of PD dependence and indi- cated that PD dependencies might have the shape of a “window” , –. The summary of the data on PD dependencies is given in the Table I. The position of the window varied between different resonance frequencies and depended on cell density during exposure of cells . Nevertheless, window-like PD dependence was observed when studying MMW effect at dif- ferent resonances. The most striking window was observed at the resonance frequency of 51.755 GHz . When exposing the E. coli cells at the cell density of 4 10 cell/mL, the ef- fect reached saturation at the PD of 10 –10 W/cm and did not change up to PD of 10 W/cm . In these experiments, the direct measurements of PD below 10 W/cm were not available and lowest PDs were obtained using calibrated atten- uators. Osepchuk and Petersen  have suggested that MMW effects could be explained by the presence of temporal har- monics, but the body of our data did not support the hypoth- esis of Osepchuk and Petersen . The background MMW radiation has been estimated as 10 –10 W/m /Hz . Since the experimentally determined half-width of resonance was in the order of 1 MHz , background PD was estimated as 10 –10 W/cm within the resonance. The MMW effects were observed at these PD in experiments with E. coli cells , , , . The data suggested that the PD dependence of MMW effect might not have a threshold.