Differentiation (2002) 70:120–129

Blackwell Verlag 2002

ORIGINAL ARTICLE

Dariusz Leszczynski

Sakari JoenvääräJukka Reivinen

Reetta Kuokka

Non-thermal activation of the hsp27/p38MAPK stress pathway bymobile phone radiation in human endothelial cells: Molecularmechanism for cancer- and blood-brain barrier-related effects

Accepted in revised form: 8 January 2002

Abstract

We have examined whether non-thermal ex-posures of cultures of the human endothelial cell lineEA.hy926 to 900 MHz GSM mobile phone microwaveradiation could activate stress response. Results obtaineddemonstrate that 1-hour non-thermal exposure of EA.hy926 cells changes the phosphorylation status of numerous, yet largely unidentified, proteins. One of theaffected proteins was identified as heat shock protein-27 (hsp27). Mobile phone exposure caused a transientincrease in phosphorylation of hsp27, an effect whichwas prevented by SB203580, a specific inhibitor of p38mitogen-activated protein kinase (p38MAPK). Also,mobile phone exposure caused transient changes in theprotein expression levels of hsp27 and p38MAPK. Allthese changes were non-thermal effects because, as de-termined using temperature probes, irradiation did notalter the temperature of cell cultures, which remainedthroughout the irradiation period at 37

∫

0.3

C.Changes in the overall pattern of protein phosphoryla-tion suggest that mobile phone radiation activates a var-iety of cellular signal transduction pathways, amongthem the hsp27/p38MAPK stress response pathway.Based on the known functions of hsp27, we put forwardthe hypothesis that mobile phone radiation-induced acti-vation of hsp27 may (i) facilitate the development of brain cancer by inhibiting the cytochrome c/caspase-3apoptotic pathway and (ii) cause an increase in blood-brain barrier permeability through stabilization of endo-

D. Leszczynski (

)

S. Joenväärä*

J. Reivinen

R. KuokkaBio-NIR Research Group, Radiobiology Laboratory,Department of Research and Environmental Surveillance,STUK – Radiation and Nuclear Safety Authority, Laippatie 4,FIN-00880, Helsinki, Finlande-mail: dariusz.leszczynski

stuk.ﬁ

*Present address

: MediCel Oy, Haartmaninkatu 8, FIN-00290Helsinki, Finland

U.S. Copyright Clearance Center Code Statement:

0301–4681/2002/7002–120 $15.00/0

thelial cell stress ﬁbers. We postulate that these events,when occurring repeatedly over a long period of time,might become a health hazard because of the possibleaccumulation of brain tissue damage. Furthermore, ourhypothesis suggests that other brain damaging factorsmay co-participate in mobile phone radiation-inducedeffects.

Key words

mobile phone

RF-EMF

non-thermal

phosphorylation

hsp27

p38MAPK

Introduction

The rapid increase in the use of mobile phones hasbrought about an urgent need to determine whethertheir emitted microwave radiation (radio-frequency-modulated electromagnetic ﬁelds; RF-EMF) couldcause health hazards. The possibility of such effects re-mains controversial. However, if proven, it would necess-arily lead to re-evaluation of the present RF-EMF radi-ation safety limits. Especially, the possibilities of induc-tion of brain cancer, disturbances in functioning of blood-brain barrier, and the long-term exposure effectsare considered by scientists and the general public to beof great importance (Jokela et al., 1999; The RoyalSociety of Canada Report, 1999; Stewart Report, 2000;Hyland, 2000; Zmirou Report, 2001).The present mobile phone radiation safety standardsare set so that the RF-EMF radiation absorbed by theliving tissue does not cause heating and thus does notcause detrimental health effects (for review see Jokela etal., 1999; The Royal Society of Canada Report, 1999;Stewart Report, 2000; Zmirou Report, 2001). On theother hand, there appears to be sufﬁcient evidence thatRF-EMF, although not causing tissue heating, can causenon-thermal biological effects (for review see Jokela etal., 1999; The Royal Society of Canada Report, 1999;

121

Stewart Report, 2000; Zmirou Report, 2001). Whetherthese non-thermally-induced biological effects arestrong enough to alter physiological processes is a mat-ter of controversy. One of the causes of this uncertaintyis that the possible biophysical mechanisms for RF-EMF interaction with living cells remain unknown. Theenergy deposited in tissue by a 900 MHz GSM mobilephone (4

eV) or by a 1800 MHz GSM mobilephone (7

eV) is far lower than the energy neededto break a chemical bond (1 eV) (Stewart Report, 2000).Therefore, some consider it questionable whether thislow energy would be able to induce biological effects atall.The early signs of cell response to a stress factor (e.g.chemical, radiation, heat, etc.) are changes in proteinphosphorylation and activation of stress response (Onoand Han, 2000). There are several published studiesexamining effects of electromagnetic ﬁelds (60 Hz) onprotein phosphorylation and activity of stress proteinsand kinases (Pipkin et al., 1999; Jin et al., 2000; More-house and Owen, 2000; Woods et al., 2000; Wetzel et al.,2001). However, results of these studies are not directlyapplicable for the clariﬁcation of the possible biologicaland health effects of the RF-EMF radiation emitted bymobile phones. The effects of mobile phone radiationon cellular stress proteins were examined in only a fewstudies (Cleary et al., 1997; Fritze et al., 1997b; Daniellset al., 1998; de Pomerai et al., 2000; Kwee et al., 2001).Together, the results obtained in these studies suggestthat, indeed, mobile phone radiation causes activationof cellular stress response and, possibly, related changesin cell physiology.The aim of the present study was to determinewhether non-thermal exposure of cells to 900 MHzGSM mobile phone radiation activates signal transduc-tion pathways and induces cellular stress response in ahuman model. Using cultures of human endothelial cellline EA.hy926, we examined the overall change in pro-tein phosphorylation, which is indicative for the acti-vation of cellular signal transduction pathways. The ef-fect on cellular stress response was examined by deter-mining phosphorylation and expression status of one of the members of stress protein family – heat shock pro-tein-27 (hsp27). Results obtained suggest that mobilephone radiation activates cellular signal transductionand stress response pathways, and therefore, it might beable to alter cell physiology and potentially cause ahealth hazard.

Methods

Cell cultureEA.hy926 cells (a gift from Dr. Cora-Jean S. Edgell, NorthCarolina University at Chapel Hill, NC, USA) (Edgell et al., 1983)were grown in Dulbecco’s MEM, supplemented with antibiotics,10% fetal bovine serum, L-glutamine, and HAT-supplement. Forexperiments, cells were removed from culture ﬂasks with trypsine,washed, and seeded at a density of 1.2

cells per 55 mm-diam-eter glass petri dish (DURAN, Germany). After overnight culture,semi-conﬂuent monolayers of EA.hy926 cells (Fig. 1A) were ex-posed to sham or RF-EMF radiation. Cell cultures for sham andirradiation were prepared in the same kind of glass dishes, derivedfrom the same batch of cells, seeded at the same cell density, andgrown for the same period of time before the experiment. The onlydifference between irradiated and sham samples was that the ir-radiated dishes resided for 1 hour in the incubation chamber withRF-EMF radiation turned on, whereas sham dishes resided in theirradiation chamber for the same period of time but with ir-radiation turned off.RF-EMF exposure system and exposure protocolCells were irradiated with a simulated mobile phone microwaveradiation in a specially constructed exposure system, based on theuse of a high Q waveguide resonator operating in TE

mode. Theirradiation chamber (Fig. 1B) was placed vertically inside a cell-culture incubator with two 55 mm-diameter glass petri-dishesplaced so that the E-field vector was parallel to the plane of theculture medium. Temperature-controlled water was circulatedthrough a thin (9 mm) rectangular glass-fiber-molded waterbedunderneath the petri dishes. In all experiments reported here, cellswere exposed for 1 hour to a 900 MHz GSM signal at an averageSAR of 2 W/kg. SAR values ranged from 1.8 W/kg to 2.5 W/kgdepending on the area of the dish, caused by the non-uniform dis-tribution of the RF-EMF radiation field. The average SAR levelof 2 W/kg was selected because it is the safety limit for the mobilephone microwave radiation emission as defined by ICNIRP (Inter-national Commission on Non-Ionizing Radiation Protection). TheRF-EMF signal was generated with an EDSG-1240 signal gener-ator and modulated with a pulse duration of 0.577 ms and a rep-etition rate of 4.615 ms to match the GSM signal modulationscheme. The signal was amplified with RF Power Labs R720Famplifier and fed to the exposure waveguide via a monopole typefeed post. The SAR distribution in the cell culture (Fig. 1C) andthe E-field above the cell culture were determined using computersimulations (FDTD method). The simulations were done with acommercial XFDTD code (Remcom, USA) with simulation gridsize of 3

3 mm

in the main grid and 1

1 mm

in thesub grid, consisting of the culture dishes and part of the waterbed.The maximum SAR was obtained in the center of the petri dish,decreasing to about 6 dB at the edges of dish. Simulation resultswere verified with measurements. The electric field in the air abovethe cell cultures was measured with a calibrated miniature Narda8021B E-field probe. The measured E-field values differed less than15% from the corresponding simulated E-fields. The SAR distri-bution was measured with small, calibrated temperature probes(Luxtron and Vitek) directly from the culture medium. The meas-urements were done at room temperature outside the incubatorwith increased culture medium height, in order to reduce the meas-urement uncertainty at the air-medium boundary. The temperaturewas measured (Vitek probe) for 10 sec in order to limit the effectof heat convection and conduction (Moros and Pickard, 1999). TheLuxtron probe has a lower temperature resolution (0.1

C) com-pared to the Vitek probe (0.001

C), and thus, the temperature hadto be measured for 1 min to achieve a sufﬁcient rise in temperature(1

C). Due to these short measurement times, the power fed to thechamber was increased up to 25 W, and the resulting SAR valuewas afterwards scaled down to 1 W of input power. The measuredSAR values at the center of the culture medium (3-mm depth) were2.5 W/kg (Luxtron) and 5.0 W/kg (Vitek). These values can becompared to the simulated value of 2 W/kg and 3.6 W/kg, respec-tively, with simulation parameters changed to correspond with themeasurement situation. The measured values can be considered asthe upper and lower limits of SAR, due to measurement uncertain-ties described above, and thus, they validate the simulations. The

122

Fig. 1

(

) Phase-contrast im-age showing growth patternand density of cell cultures onthe day of sham or RF-EMFtreatment. (

) Cross-sectionscheme of the RF-EMF ex-posure chamber. (

) Degree of uniformity of RF-EMF distri-bution in culture dishes simu-lated with XFDTD method.Culture medium height was 6mm and maximum speciﬁc ab-sorption rate (SAR) at the me-dium was 3.6 W/kg for 1 Winput power.

Dark gray color

corresponds with the areas of highest SAR and

white

withthe lowest SAR.waveguide resonator’s water-cooling system was tested by long-term temperature measurements using a Luxtron probe. The tem-perature was recorded twice a minute over the normal 1-hour ex-posure period at 2 W/kg. The temperature remained at 37

∫

0.3

Cduring the whole measurement time. Therefore, the reported bio-logical effects are of non-thermal nature.Analysis of protein expression and phosphorylationImmediately after the sham or RF-EMF exposure, culture disheswere placed on ice, washed with ice-cold phosphate-buffered saline(PBS) with 4 mM orthovanadate, and cells were detached by scrap-ing and pelleted. Pellets were lysed in buffer consisting of 9.5 Murea, 2% CHAPS, 0.8% Pharmalyte pH 3–10 (Pharmacia,Sweden), 1% dithiothreitol, 1 mM phenylmethylsulphonyl-ﬂuorideand 4 mM orthovanadate. Non-soluble cellular debris was removedfrom lysates by centrifugation for 1 hour at 42000

at 15

C.Protein content was determined with the Bradford method.Samples containing 125

g of protein were loaded into the 11 cmIPG strips pH 3–10 by overnight re-swelling. The isoelectro-focusing (DryStrip Kit with Multiphor II apparatus; Pharmacia,Sweden) was carried out at 20

C for 84 000 Volt-hours. SDS-PAGEof proteins in IPG strips was performed using a 20 cm long 8%gel in a ProteanIIxi Multicell apparatus (Bio-Rad, UK). Proteinswere visualized in 2D-gels using standard Merril’s silver-stainingmethod with Morrissey’s modiﬁcation. Silver-stained gels werescanned while wet with a Bio-Rad GS-710 densitometer. In phos-phorylation experiments, prior to RF-EMF or sham exposure, cellmonolayers were washed with pre-warmed phosphate-free DMEM,supplemented with 10% dialyzed FBS, antibiotics, L-glutamine,and HAT. Just before the exposure, a 1 mCi/ml (5 ml/dish) of

P-orthophosphate (NEN, Belgium) was added to the cells. To detectthe

P-phosphoproteins, the gels were dried and X-ray ﬁlms wereexposed at

C for up to 3 days. The computerized analyses of the protein distribution pattern in silver-stained 2D-gels and thedistribution pattern of

P-phosphoproteins in 2D-autoradiogramswere performed using PDQuest 6.1.0 software (Bio-Rad, UK).Detection of hsp27 expression in cell culturesHsp27 was detected using the standard indirect immunofluor-escence method. Cells were fixed in 3.7% buffered paraformalde-hyde on ice for 10 min, and their membranes permeabilized with0.5% Triton X-100 in PBS for 10 min at room temp. Non-specificbinding sites were blocked with 5% bovine serum albumin (BSA)in PBS (PBS/BSA) for 30 min. To visualize, hsp27 cells were incu-bated for 1 h with anti-hsp27 antibody (ImmunoCruz, CA, USA),diluted 1:200 with PBS/BSA, then washed, and incubated for 30min with FITC-conjugated goat anti-mouse Ig (DAKO, Denmark),diluted 1:50 with PBS/BSA, then washed. Specimens were observedusing a Leitz fluorescence microscope, ERGOLUX AMC. Imageswere acquired using a digital camera IMAC-CCD S30 and Metasy-stems ISIS In Situ Imaging System (ISIS, Germany).Detection of hsp27 phosphorylationTo determine changes in protein phosphorylation,

P-ortophosph-ate was present in the cultures during the 1-hour sham/RF-EMFexposure period. In experiments where the time-course of hsp27phosphorylation was determined,

P-orthphosphate was presentin cell cultures during the whole post-exposure incubation period(2 or 5 hours). In some experiments, p38MAPK inhibitorSB203580 (2 uM

Ω

) was present in cell cultures during the1-hour sham/RF-EMF exposure.Western blotsBlots of the 2D-gels and 1D-gels (standard SDS-PAGE) were pre-pared immediately after electrophoresis. Proteins were blotted onPVDF membranes (Bio-Rad, UK) using transfer buffer (25 mMTris, 192 mM glycine, 20% MeOH, and 0.1% SDS) and APBiotechNovablot semi-dry blotting apparatus (0.8 mA/cm

for 45 min).Membranes were blocked o/n at

C in 0.1 M Tris-HCl pH 7.4containing 5% non-fat dry milk. Hsp27 and p38MAPK were de-tected in the membranes using appropriate ImmunoCruz goatpolyclonal antibodies (dilution 1:50) and alkaline phosphatase conjugated ImmunoCruz anti-Goat IgG (dilution 1:1000) and en-hanced chemiluminescence (ECL) staining kit (Pierce, UK).ImmunoprecipitationCells were disrupted in ice-cold RIPA buffer. The lysates were cen-trifuged 10 000

for 10 min at

C to remove cellular debris.230 ug of protein (Lowry-Ciocalteau method) from sham and ir-radiated lysates was placed in Eppendorf tubes and pre-cleared