Non-Thermal Effects of EMF Upon the Mammalian Brain: The Lund Experience

Leif G. Salford Æ Henrietta Nittby Æ Arne Brun Æ Gustav Grafstro ̈m Æ Jacob L. Eberhardt Æ
Lars Malmgren Æ Bertil R. R. Persson

Non-thermal effects of EMF upon the mammalian brain: the Lund experience

Leif G. Salford Æ Henrietta Nittby Æ Arne Brun Æ Gustav Grafstro ̈m Æ Jacob L. Eberhardt Æ
Lars Malmgren Æ Bertil R. R. Persson

Published online: 25 July 2007
Ó Springer Science+Business Media, LLC 2007

Abstract The environment in which biology exists has dramatically changed during the last decades. Life was formed during billions of years, exposed to, and shaped by the original physical forces such as gravitation, cosmic irradiation and the terrestrial magnetism. The existing organisms are created to function in harmony with these forces. However, in the late 19th century mankind intro- duced the use of electricity and during the very last decades, microwaves of the modern communication society spread around the world. Today one third of the world’s population is owner of the microwave-producing mobile phones. The question is: to what extent are living organisms affected by these ubiquitous radio frequency fields? Since 1988 our group has studied the effects upon the mammalian blood-brain barrier (BBB) by non-thermal radio frequency electromagnetic fields (RF-EMF). These have been revealed to cause significantly increased leak- age of albumin through the BBB of exposed rats as compared to non-exposed animals—in a total series of about two thousand animals. One remarkable observation is the fact that the lowest energy levels give rise to the most pronounced albumin leakage. If mobile communication, even at extremely low energy levels, causes the users’ own albumin to leak out through the BBB, also other unwanted and toxic molecules in the blood, may leak into the brain tissue and concentrate in and damage the neurons and glial cells of the brain. In later studies we have shown that a 2-h exposure to GSM 915 MHz at non- thermal levels, gives rise to significant neuronal damage, seen 28 and 50 days after the exposure. In our continued research, the non-thermal effects (histology, memory functions) of long-term exposure for 13 months are studied as well as the effects of short term GSM 1,800 MHz upon gene expression. Most of our findings support that living organisms are affected by the non- thermal radio frequency fields. Studies from other labo- ratories in some cases find effects, while in other cases effects are not seen. Our conclusion is that all researchers involved in this field have the obligation to intensify this research in order to reduce, or avoid, the possible nega- tive effects of the man made microwaves!

1 Introduction

The world’s largest biological experiment ever? So it has been called, the introduction of microwaves to the modern communication society where, mobile phones held in close vicinity of the brain and base stations, spread everywhere around us, send their energy into our bodies. Is this only of good? Or might it impose effects upon biology. Effects that we must anticipate and evaluate as far as possible, and if needed, reduce or avoid!

The questions might seem easily answered—in spite of the fact that today one third of the worlds population are owners of the microwave-producing mobile phones, and even more, live in a milieu filled with microwave-emitting equipment such as base stations and other systems now introduced—there exists little evidence that the human organism is definitively damaged. However, during recent years, several scientific reports in respected journals have shown significant, but often weak, effects upon cells in vitro, experimental animals and also humans (for reference see Hyland 2000).

Since 1988 our group has studied the effects of RF electromagnetic fields upon the blood-brain barrier (BBB) and upon tumour growth in the mammalian brain and we have collected an extensive experimental experience in this field. While our studies on the effects of continued and pulsed modulated microwaves at 915 MHz upon brain tu- mour growth have not disclosed any growth-promoting effects in our rodent models (Salford et al. 1997), the same RF electromagnetic fields have been revealed to cause significantly increased leakage of albumin through the

BBB of exposed rats as compared to non-exposed ani- mals—in a total series of about two thousand animals. We have exposed rats to various magnetic and electromagnetic fields as well 915 MHz continuous wave (CW) and pulse- modulated at various repetition rates (50–200 pulses per s), and we have confirmed these findings in our laboratory in follow-up studies with real GSM-900 and GSM-1800 exposures (Persson et al. 1997; Salford et al. 1992, 1993, 1994, 2001 and 2003). Recently we have also examined the effects of long term exposure—55 weeks upon brain morphology and cognitive functions. The effects of GSM RF upon gene expression have been studied and 3G exposure studies are under way.

The mammalian brain is protected from exposure to potentially harmful compounds in the blood by the blood- brain barrier (BBB). The BBB is a hydrophobic barrier formed by the vascular endothelial cells of the capillaries in the brain with tight junctions between the endothelial cells. This results in a highly restricted passage of blood components through the endothelial lining. Astrocytes are surrounding the outer surface of the endothelial cells with protrusions, called end feet, and are implicated in the maintenance, functional regulation and repair of the blood- brain barrier. The astrocytes form a connection between the endothelium and the neurons but are not directly involved in the tightness of the barrier (Fig. 1).

Other periendothelial accessory structures of the BBB include pericytes and a bi-layer basal membrane, which surrounds the endothelial cells and pericytes. The basement membrane (basal lamina) supports the ablumenal surface of the endothelium and may act as a barrier to passage of macromolecules. The pericytes are a type of macrophages, expressing macrophage markers with capacity for phago- cytosis, but also for antigen presentation. In fact, the pericytes, which cover about 25% of the capillary surface (Frank et al. 1987), seem to be in a position to significantly

contribute to central nervous system (CNS) immune mechanisms (Thomas 1999).

Also physiologically, the microvasculature of the central nervous system (CNS) differs from that of peripheral or- gans. It is characterized not only by its tight junctions, which seal cell-to-cell contacts between adjacent endo- thelial cells, but also by the low number of pinocytotic vesicles for nutrient transport through the endothelial cytoplasm; and its lack of fenestrations; and the five-fold higher number of mitochondria in BBB endothelial cells compared to muscular endothelia in rat (Oldendorf et al. 1977). All this speaks in favour of an energy-dependent transcapillary transport.

The blood-brain barrier functions mainly by membrane properties controlling the bidirectional exchange of mole- cules between the general circulation and the central ner- vous system, but there is also an enzymatic barrier in the cerebral endothelium, capable of metabolizing drugs and nutrients (Ghersi-Egea et al. 1988).

In summary, the BBB serves as a regulatory system that stabilises and optimises the fluid environment of the brain’s intracellular compartment (Oldendorf 1975; Rapoport 1976; Salford et al. 2001).

The intact BBB protects the brain from damage, whereas a dys-functioning BBB allows influx of normally excluded hydrophilic molecules into the brain tissue. This might lead to cerebral oedema, increased intracranial pressure and in the worst case, irreversible brain damage. The normal selective permeability of the BBB can be altered in several pathological conditions such as epileptic seizures (Miha`lyandBozo`ky1984a,b)orextremehypertension (Sokrab et al. 1988) and also transient openings of the BBB might lead to permanent tissue damage (Sokrab et al. 1988). Considering the ensuing leakage of substances from the blood circulation into the brain tissue, harmful substances might disrupt the cellular balance in the brain tissue. Thus, opening of the BBB can have detrimental effects and since it has been shown for a few decades, that EMF has the potency to increase the permeability of this barrier, our laboratory has studied the phenomenon in a rat model.

It has become evident that with high-intensity EMF exposure resulting in tissue heating, the BBB permeability is temperature dependent (Williams et al. 1984). Thus, the importance of differentiating between thermal and non- thermal effects on the integrity of the BBB was realised. Non-thermal effects of EMFs, are the only levels studied and now reported by us.

The effects of magnetic resonance imaging (MRI) upon BBB permeability were investigated by Shivers et al. (1987) who observed that exposure to a short (23 min) standard clinical MRI procedure at 0.15 tesla (T) temporarily increased the permeability of the BBB in anaesthetised rats

to horseradish peroxidase (HRP), due to an amplified vesi- cle-mediated transport of HRP across the microvessel endothelium to the abluminal basal lamina and extracellular compartment of the brain parenchyma. The findings were confirmed in later studies by the same group (Prato et al.1990, 1994).

Our research group at Lund University started its work on effects of MRI on rat brain in 1988 and found, by the use of Evans Blue, the same increased permeability over BBB for albumin as the Shivers group (Salford et al. 1992). We continued by separating the constituents of the MRI field: RF, undulant magnetic field and static magnetic field and found that the most efficient component of the MRI field is the RF. Therefore, we focused all our fol- lowing studies on the RF effects. We also utilised endog- enous substances such as albumin and fibrinogen, which occur naturally in the blood circulation, for the detection of BBB leakage, which is identified by anti rat albumin rabbit antibodies and rabbit anti-human fibrinogen.

In the majority of our studies, EMF exposure of the animals has been performed in transverse electromagnetic transmission line chambers (TEM-cells) (Martens et al. 1993; Persson et al. 1997; Salford et al. 1992, 1993, 1994, 2001, 2003; Van Hese et al. 1991) (Fig. 2). These TEM- cells are known to generate uniform electromagnetic fields for standard measurements. In each TEM-cell, two animals can be placed, one in an upper compartment and one in a lower compartment. It is important to point out that the position of the animals in upper or lower compartments does not effect the magnitude of observed albumin leakage. Also, we have concluded, with our total series of more than 2,000 exposed animals, that there is no difference in the sensitivity to EMF exposure between male and female animals as far as albumin leakage is concerned.

We started our RF experiments with the frequency modulation 16 Hz and its harmonics 4, 8, 16 and also 50 Hz, which was felt relevant as it is the standard line frequency of the European power system, with a carrier wave of 915 MHz. At an early stage also 217 Hz modu- lation was added as this was the frequency of the then planned GSM system. This work was published in 1994 (Salford et al. 1994) and 1997 (Persson et al. 1997) and comprised sham or 915 MHz exposure for in most cases 2 h but in a minority of the experiments lasting between 2 and 960 min (both continuous and pulsed modulated waves). These results based on 246 rats 1994 and more than 1,000 rats 1997 (the majority EMF exposed and about 1/3 sham-exposed) concluded that there was a significant difference between the albumin extravasation from brain capillaries into the brain tissue between the differently exposed groups and the controls.

Repetitions of our initial findings of albumin leakage have been made in similar studies by other groups (Fritze et al. 1997; To ̈re et al. 2001, 2002).

It is important to point out that even though all animals in the 1997 series (and basically all of our experiments) are performed in inbred Fischer 344 rats, only at the most 50% of the identically exposed animals display albumin extravasation in CW animals and somewhat less in the other exposed animals. Also the sham-exposed animals had some albumin leakage though only in 17% as a mean of all controls. The leakage observed in unexposed animals presumably is due to our very sensitive immunohistologi- cal methods. The peculiar fact that at the most only every second exposed inbred animal displays leakage, is difficult to explain.

We have now performed a statistical re-evaluation of our material published in 1997 (Persson et al. 1997) where only exposed rats with a matched unexposed control rat are included (see Table 1). For the most interesting modulation frequency 217 Hz, i.e. that of GSM, the following figures are found (Wilcoxon ́s Rank Test, 2-sided p-value).

The most remarkable observation was that exposure with whole-body average power densities below10 mW/kg gave rise to a more pronounced albumin leakage than higher power densities, all at non-thermal levels. If the reversed situation were at hand, we feel that the risk of cellular telephones, base-stations and other RF emitting sources could be managed by reduction of their emitted energy. The SAR value of around 1 mW/kg exists at a distance of more than 1 m away from the mobile phone antenna and at a distance of about 150–200 m from a base station.

Table 1 Difference in albumin extravasation between exposed and control animals at different SAR values

The experimental model used in our studies allows the animals, which are un-anaesthetized during the whole exposure, to move and turn around in the exposure cham- ber, thus minimising the effects of stress induced immo- bilization described by Stagg et al. (2001).

Another remarkable observation in our studies is the fact that a significant (p < 0.002) neuronal damage is seen in rat brains 50 days after a 2 h exposure to GSM at SAR values 2, 20 and 200 mW/kg (Salford et al. 2003). We have fol- lowed up this observation in a study where 96 animals were sacrificed 14 and 28 days respectively after an exposure for 2 h to GSM mobile phone electromagnetic fields at SAR values 0 (controls), 2, 20, 200 and now also 0.2 mW/kg. Significant neuronal damage is seen after 28 days and albumin leakage after 14. Our findings may support the hypothesis that albumin leakage into the brain is the cause for the neuronal damage observed after 28 and 50 days. (Eberhardt et al. 2007).

The albumin diffuses in the neuropil, and is taken up by some astrocytes and neurons, some with, and others with- out, signs of damage, while yet other seemingly intact neurons are surrounded by interstitial albumin.

A group in Turkey performed similar experiments. However, also the presumed protective effects of the anti- oxidant Ginko biloba (Gb) were examined (Ilhan et al. 2004). About 22 female Wistar rats were exposed to a 900 MHz electromagnetic GSM near-field signal for 1 h a day for 7 days. In the GSM only group, the pathological examination revealed scattered and grouped dark neurons in all locations, but especially in the cortex, hippocampus and basal ganglia, mixed in among normal neurons. A combined non-parametric test for the four groups revealed that the distributions of scores differed significantly between the control and the GSM only exposure group (p < 0.01).

Other attempts to repeat our results with dark neurons have been performed but without success in a collaborative effort between Patrick Masons group in San Antonio (personal communication), Chiyoji Okhubos group in To- kyo (personal communication) and Bernard Veyrets group in Bordeaux (Poulletier de Gannes et al. 2006). (The latter group studied the situation at 14 and 50 days after a 2 h exposure for GSM-900 radiation at average brain SAR values of 0.14 and 2.0 W/kg). However, we have suggested that these groups and the Lund group shall exchange blindly, unstained sections from exposed and control ani- mals for handling by the neuropathologists at the other involved centres in an attempt to clarify the differences and possibly find the reason for the different results.

Most of our experiments have been acute, with survival of the animals only for hours after the exposure. It is however of greatest importance to mimic the real human situation with, as it seems, lifelong exposure. On this

SAR (mW/kg)

0.2–4 25–50

Number of animals, exposed + controls

48 + 48 22 + 22

Difference, significance

p < 0.001 Ns

123

Environmentalist (2007) 27:493–500

497

ground we have studied male and female Fischer 344 rats, exposed to GSM and sham exposed in our TEM-cells once a week for 13 months. With average whole-body SAR- levels of 0.6 and 60 mW/kg no thermal effects are induced (Yamaguchi et al. 2003; ICNIRP 1998). After 13 months the animals were studied for cognitive functions and compared to cage controls. Finally, all animals were sac- rificed and are now examined for albumin leakage, neu- ronal and glial damage and other signs of pathology. This material was first presented at the 4th International Workshop, 16–20 October 2006, Crete, Greece (Eberhardt et al. 2006). Significant effects of exposure upon memory functions will be presented (Nittby et al. 2007). We have also spared one testis per animal in formaldehyde and in freezer at –80°C: testis, ovary, kidney, spleen, liver, bone marrow and blood for future analyses and possible col- laborations.

At the moment the Lund group also has other ongoing studies: in a series of six animals exposed to GSM and two controls, the brains are under examination with electron microscopy in an attempt to demonstrate the mechanism of albumin transport across the capillary wall.

We have also undertaken an examination of the brain specimens from our animals that were exposed for 2 h GSM at 200, 20 and 2 mW/kg and allowed to survive for 50 days thereafter, in order to find out whether apoptosis was involved in the production of damaged or dying neu- rons, observed in up to 2% of the neurons in cortex and hippocampus.

Alterations of BBB permeability may also be revealed by genetic investigations. In collaboration with Belyaev and his group we have exposed rats for 6 h to GSM-900 RFs at SARs of 0.4 mW/kg (Belyaev et al. 2006) and investigated the genetic expression from cerebellar tissue. Alterations of genes encoding proteins for BBB functions were observed.

These findings are now followed up by our group in a study where 4 animals have been exposed to GSM for 6 h in the TEM-cells and another four, sham exposed simul- taneously. Micro-array analysis of the expression of all the rats’ genes in cortex and hippocampus, respectively, has shown interesting differences between exposed ani- mals and controls. This material was first presented at the 4th International Workshop, 16–20 October 2006, Crete, Greece (Salford et al. 2006). Preliminary evaluation shows that genes of interest for membrane transport show highly significant differences. This may be of importance in conjunction with our earlier findings of albumin leakage into neurons around capillaries in exposed animals (the full paper is to be published).

Finally we could mention experiments performed in our laboratory where an increase of the Ca2+-efflux over plasma membranes has been observed in plasma vesicles from spinach exposed to extremely low frequency (ELF) electromagnetic fields (Baure ́us Koch et al. 2003). We could show that suitable combinations of static and time varying magnetic fields directly interact with the Ca2+- channel protein in the cell membrane, and we could quantitatively confirm the model proposed by Blanchard and Blackman (1994). Possibly, this could be a mechanism for alterations of the BBB permeability, but no final con- clusions can so far be drawn.

2 Is the albumin leakage into the brain parenchyma a risk to the brain?

It has been suggested that BBB leakage is the major reason for nerve cell injury such as that seen in dark neurons (Fredriksson et al. 1988).

In a series of experimental situations, neuronal degen- eration has been observed in areas with BBB disruption; after intracarotid infusion of hyperosmolar solutions in rats (Salahuddin et al. 1988); in the stroke-prone hypertensive rat (Fredriksson et al 1988); after acute hypertension by aortic compression in rats (Sokrab et al. 1988). Further, epileptic seizures cause extravasation of plasma into brain parenchyma (Miha`ly and Bozo`ky 1984a, b). The cerebellar Purkinje cells are heavily exposed to plasma constituents and degenerate in epileptic patients (Sokrab et al. 1990). This effect may, however, as well be attributed to hypoxia.

It has been postulated that albumin is the most likely neurotoxin in serum (Eimerl and Schramm 1991).

Hassel et al. (1994) have demonstrated that injection of albumin into the brain parenchyma of rats, gives rise to neuronal damage. When 25 ll of rat albumin is infused into rat neostriatum, 10 and 30, but not 3 mg/ml albumin causes neuronal cell death and axonal severe damage. It also causes leakage of endogenous albumin in and around the area of neuronal damage. Albumin in the dose 10 mg/ml is approximately equivalent to 25% of the serum concentration. However, it is still unclear whether the albumin leakage demonstrated in our experiments locally reaches such concentrations.

It is striking, though, that we see areas in hippocampus and cortex of exposed animals where the cytoplasm of neurons are filled with autologous albumin, while neigh- bouring neurons display the shrunk and dark state of a ‘‘dark neuron’’, which is a very sick or dying neuron. It may be so that the leakage of albumin out in the neuropil starts a deleterious process whereby more albumin leaks through the endothelium and finally becomes a too heavy burden for the struck neurons (cf Hassel et al. 1994).

The word leakage might not be the most adequate to describe what actually happens in the BBB when albumin passes from the capillary lumen out into the brain tissue. It seems more probable that it is an active transport as suggested by several authors (Albert and Kerns 1981; Neubauer et al. 1990; Shivers et al. 1987). The fact that the CNS capillaries have a five-fold higher number of mitochondria in their endothelial cells compared to mus- cular endothelia in rat (Oldendorf et al. 1977) speaks in favour of an energy-dependent transcapillary transport of albumin.

Another important factor to bear in mind is that when the albumin molecule with its 67 kD passes the BBB, many other smaller molecules might do the same. Both this and the fact that albumin is a transporter of a variety of sub- stances in the blood which may follow out through the BBB into the neuropil adds further weight to the risks of the studied problem.

Our answer to the posed question: Is the albumin leak- age into the brain parenchyma a risk to the brain? is a probable yes!

In this connection could be mentioned, that the only molecule other than albumin we have studied is fibrinogen and we have seen leakage of this molecule in only one single animal through all the years.

3 How comes that the very low SAR values seem to be more efficient in the opening of the BBB?

One remarkable observation in our studies is the fact that SAR values around 1 mW/kg give rise to a more pronounced albumin leakage than higher SAR values—all at non-thermal levels (Persson et al. 1997). If the reversed situation were at hand, we feel that the risk of cellular telephones, base-sta- tions and other RF emitting sources could be managed by reduction of their emitted energy. The situation that the weakest fields, according to our findings, are the biologically most effective, poses a major problem. The most pronounced BBB-opening effect of the cellular telephone may not be in the most superficial layers of the brain, but several centi- metres deep in central cerebral structures! It cannot be ex- cluded that non-users in the vicinity of the cellular phone users, may be influenced by these weak effects. The SAR value of around 1 mW/kg is produced in air at a distance of more than one meter away from the mobile phone antenna and it can be calculated that this energy level exists centrally in the human brain or even in the contralateral hemisphere when the mobile phone is held at the ear. Concerning the emission from base stations for mobile communication, the SAR value 1 mW/kg exists at a distance of about 150–200 m from the station. The reason for the higher effects of the lowest examined SAR values is not readily revealed.

Our experiments showing that controlled frequency and amplitude of ELF EM fields upon spinach plasma vesicles can steer transport over the membrane (Baure ́us Koch et al. 2003) may be a first proof of a resonance phe- nomenon where appropriate levels of frequency and amplitude in the right combination have the potency to communicate with the biology of membranes and trans- port systems. Ross Adey was convinced that these bio- logical windows play an important role. He used to take an example from the auditory apparatus: if you scream too loudly to a person he will not be able to collect your information, neither if you speak too weakly. The biology of your cochlea demands an appropriate window for the physical stimulation in order to perceive the communi- cation (personal communication).

The mechanisms by which the EMFs may alter BBB permeability are not well understood. At low field strengths, the effects on body temperature are negligible and thus heating effects are not involved. The Shivers- Prato group suggested in their initial study of MRI induced BBB leakage (1987) that MRI stimuli might modify the physicochemical characteristics of membranes.

Some concepts that should be observed in the continued search for the answer should be:

• A quantum mechanical model for interaction with protein-bound ions (Baure ́us Koch et al. 2003).

• Microwaves have effect directly on the protein confor- mation (vibration energy levels) (De Pomerai et al. 2003).

• Autooxidative processes which lead to oxidation in the cells (Ilhan et al. 2004).

• Or possibly interaction microwaves –> water molecules bound in biologically active molecules.

  If mobile communication, even at extremely low SAR values, causes the users’ own albumin to leak out through the BBB, which is meant to protect the brain, also other unwanted and toxic molecules in the blood, may leak into the brain tissue and concentrate in and damage the neurons and glial cells of the brain. It can not be excluded that this, (especially after many years of intense use) may promote the development of autoim- mune and neuro-degenerative diseases, and we conclude that the suppliers of mobile communication—and our politicians—have an extensive responsibility to support the exploration of these possible risks for the users and the society. This holds true also for the new and hitherto barely examined 3G technique which sends microwaves of a different character, and it is quite possible that the biological effects of 3G differ from that of NMT and GSM.

References

Albert, E. N., & Kerns, J. M. (1981). Reversible microwave effects on the blood-brain barrier. Brain Research, 230, 153–164.

Baure ́us Koch, C. L. M., Sommarin, M., Persson, B. R. R., Salford, L. G., & Eberhardt, J. L. (2003). Interaction between weak low frequency magnetic fields and cell membranes. Bioelectromag- netics, 24, 395–402.

Belyaev, I. Y., Baure ́us Koch, C., Terenius, O., Roxstrom-Lindquist, K., Malmgren, L. O. G., Sommer, W. H., Salford, L. G., & Persson, B. R. R. (2006). Exposure of rat brain to 915 MHz GSM microwaves induces changes in gene expression but not double stranded DNA breaks or effects on chromatin conforma- tion. Bioelectromagnetics, 27, 295–306.

Blanchard, J. P., & Blackman, C. F. (1994). Clarification and application of an ion parametric resonance model for magnetic field interactions with biological systems. Bioelectromagnetics, 15, 217–238.

De Pomerai, D. I., Smith, B., Dawe, A., North K., Smith, T., Archer, D. B., Duce, I. R., Jones, D., & Candido, E. P. (2003). Microwave radiation can alter protein conformation without bulk heating. FEBS Letters 543, 93–97.

Eberhardt, J. L., Persson, B. R. R., Brun, A. E., Malmgren, L. O., Grafstro ̈m, G., & Salford, L. G. (2006). Long term effects of microwaes from GSM mobile phones on the rat brain. Abstract to the 4th International Workshop 16–20 Oct, Crete Greece.

Eberhardt, J. L., Persson, B. R. R., Malmgren, L. O., Brun, A. E., & Salford, L. G. (2007). Blood-brain barrier permeability and nerve cell damage in rat brain 14 and 28 days after exposure to microwaves from GSM mobile phones. Submitted for publication.

Eimerl, S., & Schramm, M. (1991). Acute glutamate toxicity and its potentiation by serum albumin are determined by the Ca2+ concentration. Neuroscience Letters, 130, 125–127.

Frank, R. N., Dutta S., & Mancini, M. A. (1987). Pericyte coverage is greater in the retinal than in the cerebral capillaries of the rat. Investigative Ophthalmology & Visual Science, 28, 1086–1091.

Fredriksson, K., Kalimo, H., Nordborg, C., Johansson, B. B., & Olsson, Y. (1988). Nerve cell injury in the brain of stroke-prone spontaneously hypertensive rats. Acta Neuropathologica (Berl), 76, 227–237.

Fritze, K., Sommer, C., Schmitz, B., Mies, G., Hossmann, K.-A., Kiessling, M., & Wiessner, C. (1997). Effect of global system for mobile communication (GSM) microwave exposure on blood-brain barrier permeability in rat. Acta Neuropathologica, 94, 465–470.

Ghersi-Egea J. F., Minn A., Siest G. (1988). A new aspect of the protective functions of the blood-brain barrier: Activities of four drug-metabolizing enzymes in isolated rat brain microvessels. Life Sciences, 42, 2515–2523.

Hassel, B., Iversen, E. G., & Fonnum, F. (1994). Neurotoxicity of albumin in vivo. Neuroscience Letters, 167, 29–32.

Hyland, G. (2000). Physics and biology of mobile telephony. The Lancet, 356, 1833–1836.

ICNIRP. (1998). Guidelines for limiting exposure to time-varying electric, magnetic and electromagnetic fields (up to 300 GHz). Health Physics, 74, 494–522.

Ilhan, A., Gurel, A., Armutcu, F., Kamisili, S., Iraz, M., Akyol, O., & Ozen, S. (2004). Ginkgo biloba prevents mobile phoneinduced oxidative stress in rat brain. Clinica Chimica Acta, 340, 153–162.

Martens, L., Van Hese, J., De Sutter, D., De Wagter, C., & Malmgren, L. O. G. (1993). Electromagnetic field calculations used for exposure experiments on small animals in TEM-cells. Bioelect- rochemistry Bioenergetics 30, 73–81.

Miha`ly, A., & Bozo`ky, B. (1984a). Immunohistochemical localization of serum proteins in the hippocampus of human subjects with partial and generalized epilepsy and epileptiform convulsions. Acta Neuropathology, 127, 251–267.

Miha` ly, A., & Bozo` ky, B. (1984b). Immunohistochemical localiza- tion of extravasated serum albumin in the hippocampus of human subjects with partial and generalized and epileptiform convulsions. Acta Neuropathology, 65, 471–477.

Neubauer, C., Phelan, A. M., Kues, H., & Lange, D. G. (1990). Microwave irradiation of rats at 2.45 GHz activates pinocytotic- like uptake of tracer by capillary endothelial cells of cerebral cortex. Bioelectromagnetics, 11, 261–268.

Nittby, H., Grafstro ̈m, G., Dong Ping, T., Brun, A., Persson, B., Salford, L. G., & Eberhardt, J. L. (2007). Cognitive impairments in rats after long-term exposure ot GSM-900 radiation. Submit- ted for publication.

Oldendorf, W. H. (1975). Permeability of the blood-brain barrier. In D. Tower (Ed.), The Nervous System. (pp. 229–289). New York: Raven Press.

Oldendorf, W. H., Cornford, M. E., & Brown, W. J. (1977). The large apparent work capability of the blood-brain barrier: A study of the mitochondrial content of capillary endothelial cells in brain and other tissues of the rat. Annals of Neurology, 1, 409–417.

Persson, B. R. R., Salford, L. G., & Brun, A. (1997). Blood-brain barrier permeability in rats exposed to electromagnetic fields used in wireless communication. Wireless Networks, 3, 455–461.

Poulletier de Gannes, F., Haro, E., Taxile, M., Ladevze, E., Mayer, L., Lascau, M., Leveˆque, P., Ruffie, G., Billaudel, B., Lagroye, I., & Veyret, B. (2006). Do GSM-900 signals affect blood-brain barrier permeability and neuron viability? Abstract at the 28th Annual meeting of the Bioelectromagnetics Society (pp. 164– 165). Cancun, Mexico.

Prato, F. S., Frappier, R. H., Shivers, R. R., & Kavaliers, M. (1990). Magnetic resonance imaging increases the blood-brain barrier permeability to 153-gadolinium diethylenetriaminepentaacetic acid in rats. Brain Research, 523, 301–304.

Prato, F. S., Wills, J. M., Roger, J., Frappier, H., Drost, D. J., Lee, T. Y., Shivers, R. R., & Zabel, P. (1994). Blood-brain barrier permeability in rats is altered by exposure to magnetic fields associated with magnetic resonance imaging at 1.5 T. Micros- copy Research Technique 27, 528–534.

Rapoport, S. I. (1976). Blood-brain barrier in physiology and medicine. New York: Raven Press.

Salahuddin, T. S., Kalimo, H., Johansson, B. B., & Olsson, Y. (1988). Observations on exsudation of fibronectin, fibrinogen and albumin in the brain after carotid infusion of hyperosm- olar solutions. An immunohistochemical study in the rat indicating longlasting changes in the brain microenvironment and multifocal nerve cell injuries. Acta Neuropathologica (Berl) 76, 1–10.

Salford, L. G., Brun, A., Eberhardt, J., Malmgren, L., & Persson, B. (1992). Electromagnetic field-induced permeability of the blood- brain barrier shown by immunohistochemical methods. In B. Norde ́n, & C. Ramel (Eds.), Interaction mechanism of low-level electromagnetic fields in living systems. (pp. 251–258). Oxford, UK: Oxford University Press.

Salford, L. G., Brun, A., Eberhardt, J. L., & Persson, B. R. R. (1993). Permeability of the blood-brain-barrier induced by 915 MHz elec- tromagnetic-radiation, continuous wave and modulated at 8, 16, 50 and 200 Hz. Bioelectrochemistry Bioenergitics, 30, 293–301.

Salford, L. G., Brun, A., Sturesson, K., Eberhardt, J. L., & Persson, B. R. R. (1994). Permeability of the blood-brain-barrier induced by 915 MHz electromagnetic-radiation, continuous wave and mod- ulated at 8, 16, 50 and 200 Hz. Microscopy Research Technique 27, 535–542.

Salford, L. G., Brun, A., & Persson, B. R. R. (1997). Brain tumour development in rats exposed to electromagnetic fields used in wireless cellular communication. Wireless Networks 3, 463–469.

123

500

Environmentalist (2007) 27:493–500

Salford, L. G., Krogh, M., Grafsto ̈ m, G., Nittby, H., Rehn, G., Berlin, H., Eberhardt, J. L., Malmgren, L., Persson, R. B. R., & Widegren, B. (2006). GSM exposure changes gene expression in rat hippocampus and cortex. Abstract for the 4th International Workshop: ‘‘Biolo- gical Effects of Electromagnetic Fields’’ 16–20 Oct. 2006, Crete, Greece.

Salford, L. G., Persson, B., Malmgren, L., & Brun, A. (2001). Te ́ le ́ phonie Mobile et Barrie` re Sang-Cerveau. In: Pietteur Marco (Ed.), Te ́le ́phonie mobile—effects potentiels sur la sante ́ des ondes e ́lectromagne ́tiques de haute fre ́quence (pp. 141–152). Belgium: Emburg.

Salford, L. G., Brun, A. E., Eberhardt, J. L., Malmgren, L., & Persson, B. R. R. (2003). Nerve cell damage in mammalian brain after exposure to microwaves from GSM mobile phones. Environ- mental Health Perspectives 111, 881–883.

Shivers, R. R., Kavaliers, M., Teskey, G. C., Prato, E. S., Pelletier, R. M. (1987). Magnetic resonance imaging temporarily alters blood-brain barrier permeability in the rat. Neuroscience Letters 76, 25–31.

Sokrab, T. E. O., Johansson, B. B., Kalimo, H., & Olsson, Y. (1988). A transient hypertensive opening of the blood-brain barrier can lead to brain damage. Acta Neuropathology, 75, 557–565.

Sokrab, T. E., Kalimo, H., & Johansson, B. B. (1990). Parenchymal changes related to plasma protein extravasation in experimental seizures. Epilepsia 31, 1–8.

Stagg, R. B., Havel, L. H. III, Pastorian, K., Cain, C., Adey, W. R., & Byus, C. V. (2001). Effect of immobilization and concurrent exposure to a pulse-modulated microwave field on core body temperature, plasma ACTH and corticosteroid, and brain

ornithine decarboxylase, Fos and Jun mRNA. Radiation

Research, 155, 584–92.
Thomas, W. E. (1999). Brain macrophages: On the role of pericytes

and perivascular cells. Brain Research. Brain Research Reviews,

31, 42–57.
To ̈re, F., Dulou, P. E., Haro, E., Veyret, B., & Aubineau, P. (2001).

Two-hour exposure to 2-W/kg, 900-MHZ GSM microwaves induces plasma protein extravasation in rat brain and dura mater. Proceedings of the 5th International congress of the EBEA (pp. 43–45). Helsinki, Finland.

To ̈re, F., Dulou, P. E., Haro, E., Veyret, B., & Aubineau, P. (2002). Effect of 2 h GSM-900 microwave exposures at 2.0, 0.5 and 0.12 W/kg on plasma protein extravasation in rat brain and dura mater. Proceedings of the 24th annual meeting of the BEMS (pp. 61–62).

Van Hese, J., Martens, L., De Zutter, D., De Wagter, C., Malmgren, L., Persson, B. R. R, & Salford, L. G. (1991). Simulations of the effect of inhomogenities in TEM transmission cells using the FDTD-method. IEEE Transactions on Electromagnetic Com- patibility, 34,292–298.

Williams, W. M., Lu, S. T., del Cerro, M., & Michaelson, S. M. (1984). Effect of 2,450 MHz microwave energy on the blood- brain barrier to hydrophilic molecules. D. Brain temperature and blood-brain barrier permeability to hydrophilic tracers. Brain Research, 319, 191–212.

Yamaguchi, H., Tsurita, G., Ueno, S., Watanabe, S., Wake, K., Taki, M., Nagawa, H. (2003). 1,439 MHz pulsed TDMA fields affect performance of rats in a T-maze only when body temperature is elevated. Bioelectromagnetics 24, 223–230.