Bioelectromagnetics 25:216^227 (2004)

Bone Morphogenetic Protein Expressionin Newborn Rat Kidneys After PrenatalExposure to Radiofrequency Radiation

Athina Pyrpasopoulou,1 Vassiliki Kotoula,1 Angeliki Cheva,1 Prodromos Hytiroglou,1

Eleni Nikolakaki,2 Ioannis N. Magras,3 Thomas D. Xenos,4

Theodoros D.Tsiboukis,4 and Georgios Karkavelas1*1Laboratory of Pathology, Department of Medicine, School of Health Sciences,

Aristotle University of Thessaloniki,Thessaloniki, Greece2Laboratory of Biochemistry, Department of Chemistry, School of Natural Sciences,

Aristotle University of Thessaloniki,Thessaloniki, Greece3Laboratory ofAnatomy, Histology and Embryology, Department of VeterinaryMedicine,School of Health Sciences, Aristotle University of Thessaloniki,Thessaloniki, Greece

4Department of Electrical and Computer Engineering, School of Engineering,Aristotle University of Thessaloniki,Thessaloniki, Greece

Effects of nonthermal radiofrequency radiation (RFR) of the global system of mobile communication(GSM) cellular phones have been as yet mostly studied at the molecular level in the context of cellularstress and proliferation, as well as neurotransmitter production and localization. In this study, asimulation model was designed for the exposure of pregnant rats to pulsed GSM-like RFR (9.4 GHz),based on the different resonant frequencies of man and rat. The power density applied was 5 mW/cm2,in order to avoid thermal electromagnetic effects as much as possible. Pregnant rats were exposed toRFR during days 1–3 postcoitum (p.c.) (embryogenesis, pre-implantation) and days 4–7 p.c. (earlyorganogenesis, peri-implantation). Relative expression and localization of bone morphogeneticproteins (BMP) and their receptors (BMPR), members of a molecular family currently considered asmajor endocrine and autocrine morphogens and known to be involved in renal development, wereinvestigated in newborn kidneys from RFR exposed and sham irradiated (control) rats. Semi-quantitative duplex RT-PCR for BMP-4, -7, BMPR-IA, -IB, and -II showed increased BMP-4 andBMPR-IA, and decreased BMPR-II relative expression in newborn kidneys. These changes werestatistically significant for BMP-4, BMPR-IA, and -II after exposure on days 1–3 p.c. (P<.001 each),and for BMP-4 and BMPR-IA after exposure on days 4–7 p.c. (P<.001 and P¼.005, respectively).Immunohistochemistry and in situ hybridization (ISH) showed aberrant expression and localization ofthese molecules at the histological level. Our findings suggest that GSM-like RFR interferes with geneexpression during early gestation and results in aberrations of BMP expression in the newborn. Thesemolecular changes do not appear to affect renal organogenesis and may reflect a delay in thedevelopment of this organ. The differences of relative BMP expression after different time periods ofexposure indicate the importance of timing for GSM-like RFR effects on embryonic development.Bioelectromagnetics 25:216–227, 2004. � 2004 Wiley-Liss, Inc.

Key words: RFR; nonthermal effects; GSM; BMP; molecular effect; relative expression

INTRODUCTION

Possible health hazards from exposure to radio-frequencies in the global system for mobile commu-nication (GSM) cellular phone range (800–3000 MHz)has been an issue of concern and debate during the pastseveral years. Several studies have been dealing withinvestigation of the effects of such radiofrequencyradiation (RFR), using various frequencies and powerintensities. According to these, there are no direct toxic,genotoxic, or mutagenic effects from RFR and, in thecase of adverse effects described on organisms, these

�2004Wiley-Liss, Inc.

——————A. Pyrpasopoulou and V. Kotoula contributed equally to this work.

Grant sponsor: Greek General Secretariat of Research andTechnology; Grant number: ED 325/PENED 99.

*Correspondence to: Dr. Georgios Karkavelas, Laboratory ofPathology, Department of Medicine, School of Health Sciences,Aristotle University of Thessaloniki, 54006 Thessaloniki, Greece.E-mail: gkarkav@med.auth.gr

Received for review 27 February 2003; Final revision received2 September 2003

DOI 10.1002/bem.10185Published online in Wiley InterScience (www.interscience.wiley.com).

result mainly from hyperthermia related to the powerdensity of applied RFR [reviewed by Brusick et al.,1998] or prolonged exposure [Tice et al., 2002].

However, RFR effects at the molecular level havebeen described under very low power conditions, whichare not considered to induce hyperthermia (nonthermalRFR specific effects). These include gene expressionalteration of heat shock proteins, such as hsp70[Di Carlo et al., 2002] and hsp27/p38MAPK[Leszczynski et al., 2002], early response genes likec-fos [Fritze et al., 1997; Goswami et al., 1999], as wellas neurotransmitter production, concentration and loca-lization [Mausset et al., 2001; Testylier et al., 2002].Based on the above, RFR specific effects were sug-gested to induce or promote the development of cancerand other diseases [French et al., 2001; Leszczynskiet al., 2002], a hypothesis not proven for any type ofcancer [Heikkinen et al., 2001; Bartsch et al., 2002] andnot proven for the proliferation of glioma cells [Stagget al., 1997; Higashikubo et al., 2001].

In a previous study [Magras and Xenos, 1997], wereported a progressive, irreversible infertility effect onmice exposed under open air conditions to GSM-typeRFR at very low power (0.2–1 mW/cm2), with anincreased growth effect (size and birth weight) on thefew newborns obtained. In the present study, we soughtto investigate nonthermal RFR effects at the molecularlevel under environmentally controlled conditions onrat newborns exposed during prenatal development.For this purpose, a simulation model was designed forGSM-like RFR exposure of pregnant dams; the new-borns delivered were checked for gross malformations,histology, and alterations in the expression of fivemembers of the bone morphogenetic protein (BMP)family of molecules that are considered as key mor-phogenetic factors.

BMPs are members of an evolutionary conservedfamily of secreted signaling molecules implicated inembryonic patterning events, as well as in the devel-opment of most organ systems via modulation ofepithelial–mesenchymal interactions. BMPs were ori-ginally purified and characterized as osteogenesis-inducing factors contained in bone extracts [Wang et al.,1988; Celeste et al., 1990], and their involvement in theregulation of chondrogenesis and osteogenesis duringbone repair has been extensively studied [Sandberget al., 1993]. Electromagnetic field irradiation at variouspulse frequencies is known to induce BMP-2 and -4upregulation in rat and human osteoblastic cell linesin vitro and in chick embryonic calvaria in vivo [Nagaiand Ota, 1994; Bodamyali et al., 1998].

Beyond their important role in skeletal develop-ment and bone formation, BMPs -2, -4, and -7 are cur-rently considered as universal endocrine and autocrine

morphogens [Reddi, 2000]. BMP-2 is involved inblastocyst implantation and cavity formation duringgastrulation [Coucouvanis and Martin, 1999; Pariaet al., 2001]. BMP-4 heterozygous null mutant miceexhibit abnormalities in the kidney/ureter morphogen-esis, including among others, hypo/dysplasia of thekidneys, hydroureters etc. [Miyazaki et al., 2000].BMP-7 is considered to be crucial for metanephrickidney development [Lipschutz, 1998], while BMP-7deficient mice demonstrate abnormal glomeruli forma-tion [Karsenty et al., 1996]. BMP signaling is mediatedby specific receptors that act as heterodimer/tetramersof two different transmembrane serine/threonine kinasesubunits. In the mouse, a single type II subunit (BMPR-II) has been identified, while at least three type Isubunits (Alk2-BMPR-IA, Alk3-BMPR-IB, and Alk6)are also known [Liu et al., 1995; Massague, 1996].

In this study, we present data on the relativeexpression and localization of BMP-4, -7, BMPR-IA,-IB, and -II in the newborn rat kidney after exposure toGSM-like RFR at two distinct periods during the firstweek of embryonal development.

MATERIALS AND METHODS

Experimental Model

Groups of female adult Wistar rats, pre-viously mated on day 0, were continuously exposed toGSM-like RFR during days 1–3 postcoitum (p.c.)(embryogenesis, preimplantation period), and days 4–7 p.c. (early organogenesis, peri-implantation period).The rats were obtained from ‘‘Theagenion AnticancerInstitute of Thessaloniki.’’ The use of these experi-mental animals was approved by the Veterinary Ser-vice of the Municipality of Thessaloniki, according tothe provisions of the laws 1191/81 and 2015/92 and thePresidential Decree 160/91 of the Greek Democracy.

Upon arrival, all rats were quarantined for 2 weeksto discover any disease and to allow them to acclimatizein the laboratory. All the rats were healthy and showedno signs of illness during the course of this study. Tapwater and certified feed (Greek Sugar Factory) werefreely available. The rats were maintained under thestandard laboratory conditions, i.e., 22 8C tempera-ture and 60% relative humidity, lighting during daytimeand darkness during the night. Low power electricalventilators provided four volume changes of air perhour through air ducts. Four independent sets of experi-ments were performed, two for each gestation stagedescribed above.

The experiments were conducted in two iden-tical neighboring rooms. The background electromag-netic noise was measured to be the same, and the normal

RFREffects on Kidney BMPs 217

for seasonal light–dark cycle was maintained. Theexperimental setup consisted of transmitting facilitiesand monitoring instruments.

The transmitting facilities consisted of twoidentical transmitting systems per experimental setup.Each consisted of one X band microwave generatorwith a maximum power output of þ25 dBm, a 27 cmhorn-fed parabolic reflector antenna providing a 26 dBgain at 9.35 GHz, an X band waveguide networkconnecting the generator to the antenna and a �20 dBdirectional coupler for continuously monitoring themicrowave signal. The antennas were placed 10 mapart, in line and facing each other.

Six plastic cages with the experimental animals(four dams per cage) were positioned in two hori-zontal rows by three vertical columns, 5 m away fromthe antennas, in the middle of the two antennas dis-tance corresponding to the antenna maximum. Thecage dimensions were about 50� 40� 25 cm andthe frequency chosen was 9.4 GHz, which gives afree-space wavelength of 3.19 cm; consequently thedistance between two consecutive minima was about1.59 cm. Therefore, if standing waves existed,25 minima and 25 maxima would be detected alongeach cage. Yet, since the rats could, and in fact did,freely move in the cages and their crown-rump lengthwas of the order of 15 cm, the existence of standingwaves could hardly affect the experiment. Therewere four freely moving rats per cage, therefore, theymight sometimes electromagnetically shade each other.However, this probability was statistically the same forall animals.

Pulsed microwaves, with a pulse length of 20 msand a pulse repetition rate of 50 Hz, at a frequency of9.4 GHz were applied. This frequency was chosenbecause the resonant frequency of a rat is in the order of700 MHz, whereas the resonant frequency of an averageman is about 70 MHz [Thuery, 1992]. Consequently, ifthe ratio 1:10 for a European GSM cellular telephonewas to be maintained, a frequency in the order of 9 GHzshould be applied. The average penetration depth at thehead of the rat at 9.4 GHz, if the dielectric character-istics of the various tissues are to be considered, forexample, the Brooks Air Base Model [2003], is in theorder of 5 mm; in other parts it can be much lower.In comparison, the penetration depth in the case ofEuropean GSM cell telephony (frequencies 915–945 MHz) is approximately 55 mm. For the EuropeanDCS cell telephony (frequency 1.8 GHz), it is 26 mm;and for the most commonly used WLAN (2.5 GHz),19 mm.

In order to study possible nonthermal electro-magnetic effects, a microwave power density of5 mW/cm2 was applied. This corresponded to a maxi-

mum SAR of 0.5 mW/kg, calculated by means of anumerical method using Finite Elements [Xenos andMagras, 2003]. The applied microwave power densitywas continuously monitored by a spectrum analyzer viaa standard gain (16 dB) horn antenna, whereas theelectromagnetic power density, applied and back-ground, was cross-checked by a Narda 8100 surveymeter and probe, respectively. All measurements wereperformed according to the relevant IEEE standard[IEEE Std C95.3, 1991]; and all cables, connectors, andwaveguides used were standard.

Animals and Tissues

For this study, 20 newborns were collected fromsham-exposed mother animals (controls not exposedto RFR), 25 newborns from mothers exposed duringdays 1–3 p.c., and 26 from mothers exposed duringdays 4–7 p.c. After delivery, the newly born rats weresacrificed by euthanasia and macroscopically examinedfor possible malformations (Table 1). Kidneys wereplaced into formalin for paraffin embedding, and weresnap-frozen in liquid nitrogen for RNA extraction.

Histological and Molecular Analysis

Three micron thick paraffin sections were cutfrom all the samples collected, stained with haematox-ylin–eosin (H&E) and examined histologically.

Immunohistochemistry

To investigate BMP expression in tissue samplesfrom control and exposed animals, 3 mm thick paraffinsections were incubated overnight with polyclonal goatantimouse antibodies purchased from Santa Cruz Bio-technology (Heidelburg, Germany) for BMP-4 (N-16,cat # sc-6896) and BMP-7 (N-19, cat # sc-6899),respectively, after appropriate antigen retrieval andblocking, as suggested by the manufacturer. Biotiny-lated antigoat immunoglobulins were used as secondaryantibodies (Dako, Glostrup, Denmark). The sectionswere subsequently incubated with biotin conjugatedstreptavidin label and finally stained using diamino-benzidine as chromogen. BMP immunostaining was

TABLE 1. Rat Newborns Included in This Study

n Histology IHC ISH RT-PCR

Controls 20 10 10 5 10RFR exposed

Days 1–3 p.c. 25 17 10 5 8Days 4–7 p.c. 26 18 14 5 8

Controls, sham exposed; n, number of newborns; IHC, immuno-histochemistry; ISH, in situ hybridization; RT-PCR, reverse trans-cription-PCR; RFR, radiofrequency radiation; p.c., postcoitum.

218 Pyrpasopoulou et al.

evaluated qualitatively by three independent observers(VK, PH, and GK). The intensity of immunostaining inthe cortical mesenchyme of control kidneys was gradedas positive (þ). Specimens were considered as stronglypositive (þþ) when more intense staining wasobserved, and as negative (�) when no clearly positivecells could be detected.

Reverse Transcription-PCR (RT-PCR)

Frozen intact newborn kidneys were homoge-nized in Kontes tubes. RNA was extracted with Trizol(Gibco/Invitrogen, Paisley, UK) and was reversetranscribed into cDNA using random hexamers andSuperscript II (Gibco/Invitrogen) according to themanufacturer’s instructions. For each sample, expres-sion of BMP-4, -7, and their receptors BMPR-II,BMPR-IA, and BMPR-IB was analyzed by PCR usingthe primers described by Shimazaki et al. [1999]. Thelength of PCR products was 464 bp for BMP-4, 386 bpfor BMP-7, 455 bp for BMPR-IA, 457 bp for BMPR-IB,and 380 bp for BMPR-II.

To approach the relative expression of transcriptsin the tissue, the ratio of the corresponding PCR pro-ducts versus that of a housekeeping gene (b-actin) wassemi-quantified with the EDAS software (Kodak,Rochester, NY), according to Pizzonia [2001]. Fromthe b-actin mRNA, a 287 nucleotide region was ampli-fied with primers: 50-CCG CCC TAG GCA CCA GGGTG-30 (forward) and 50-GGC TGG GGT GTT GAAGGT CTC AAA-30 (reverse). For EDAS semi-quanti-fication, duplex reaction mixtures were used at an 80 mlreaction volume (3 ml of cDNA). The mix contained1.2� PCR buffer, 2 mM MgCl2, 0.125 mM dNTPs,25 pmoles of each BMP-primer, 10 pmoles for b-actinprimers, and 2 U Taq polymerase (Gibco/Invitrogen).The reaction conditions (primer concentration, numberof cycles, mono- and bi-valent ion concentration) wereoptimized in simple and duplex reactions with controlcDNAs in order to have both products reaching theplateau phase of amplification in the last two cycles ofthe reaction. PCR conditions were: denaturation at 94 8C for 3 min, 35 cycles at 94 8 C for 30 s/58 8C (BMP-7and BMPR-IB) for 30 s/72 8C for 30 s, and finalelongation at 72 8C for 10 min. For BMP-4, BMPR-IA,and -II annealing was performed at 63 8C.

Electrophoresis of the PCR products was per-formed on 2.2% agarose gels at 120 V. PCR productsfrom the standardization runs were verified by se-quencing. The ratio of BMP related versus b-actinproducts was evaluated on digitally obtained photo-graphs (inverted phase) and calculated according to theformula: (mean intensity of BMP related productband�mean background)/(mean intensity of b-actinband�mean background). Except for the kidneys of

two control animals that were used to optimize the PCRprotocols, all other samples were processed blindly.Three runs of duplex reactions were carried out.

Construction of DIG-Labeled Probesand In Situ Hybridization (ISH)

BMPR-IA and -II PCR products were ligated intopGEM-T plasmids (Promega, France) and cloned intocompetent cells. Cloned inserts were sequenced andsubsequently used to create PCR products containingboth SP6 and T7 sites with PUC-primers 50-GTT TTCCCA GTC ACG AC-30 (forward) and 50-CAG GAAACA GCT ATG AC-30 (reverse). These PCR productswere submitted to in vitro transcription with paralleldigoxigenin labeling for the construction of anti-sense and sense riboprobes (RNA DIG labeling kit,SP6/T7 [Roche, Mannheim, Germany]). Probes werechecked on 1.5% agarose gels.

ISH on paraffin sections was performed accordingto Braissant and Wahli [1998], with some modifica-tions. After deparaffinization, sections were incubat-ed in DEPC–PBS, and processed for proteinolysis(0.02 mg/ml proteinase K in prewarmed TE buffer for30 min at 37 8C). The sections were washed againin DEPC–PBS, acetylated in triethanolamine–MgCl2(Sigma, Munich, Germany) containing 0.25% v/v aceticanhydride, equalized in 5� SSC buffer for 15 min andsubsequently prehybridized for 2 h at 48 8C. Thehybridization buffer contained 50% deionized forma-mide, 5� SSC, 40 ml RNAse inhibitor (Gibco/Invitro-gen), and 100 mg/ml freshly denatured salmon spermDNA (Gibco/Invitrogen). For hybridization, 500 ng/mlof antisense or sense probe were added into the hybri-dization solution, and 50 ml were added on each section.The sections were covered with parafilm and hybridizedat 45 8C for 16 h. After hybridization, sections werewashed in 2� SSC (20 min, RT), 1� SSC (2� 15 min,RT), and 0.2� SSC (15 min, RT, and 15 min, 55 8C).For hybridization signal detection, an antidigoxigeninantibody coupled to alkaline phosphatase (Roche) wasused, and the signal was developed with nitro-blue-tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate(NBT/BCIP [Roche]). Sections were counterstainedwith Nuclear Fast Red. Hybridization results wereassessed qualitatively, as positive or negative.

Statistics

Statistics were performed with the SPSS software.Nonparametric tests (Mann–Whitney, Kolmogorov–Smirnov, significance <0.05), as well as independentsample t-tests (95% confidence interval) were used forthe evaluation of relative expression values in eachstudy group. Nominal values, such as semi-quantitative

RFREffects on Kidney BMPs 219

results from BMP-4 immunohistochemistry were com-pared with Pearson’s w2- and Cramer’s V-test.

RESULTS

No malformations were observed in the newbornswhich had been exposed to RFR irradiation duringpregnancy. At the histological level, kidney sections ofexposed newborn rats occasionally exhibited slighthydropic swelling in the tubular cells when compared tocorresponding material from control (sham exposed)animals (data not shown). No changes were observed inthe kidneys of the dams after RFR.

Immunohistochemical stains showed notable dif-ferences in the intensity and localization of BMP-4between the kidneys of the prenatally exposed new-borns and those of the control animals. In the latter,positive (þ) staining for BMP-4 was observed incortical mesenchymal cells located among glomeruliand convoluted tubules. In the exposed groups, theintensity of staining in the mesenchyme was stronger(more oftenþþ thanþ), while BMP-4 was additionallypresent in epithelial cells of the collecting tubules(Fig. 1). These immunohistochemical findings are sum-marized in Table 2. Statistical tests for nominal values(Pearson’s w2, Cramer’s V) showed significant dif-ferences between control kidneys and those of ex-posed newborns (Pearson’sP< 0.001 andP¼ 0.001 forthe groups exposed during days 1–3 and 4–7 p.c.,respectively). Immunohistochemical stains for BMP-7were interpreted as unsatisfactory. We further analyzedthe absolute or relative expression of these growthfactors and their receptors in kidney tissue extracts byPCR (Fig. 2). The semi-quantitative results obtainedafter analysis with the EDAS software are summarizedin Table 3 and presented in Figure 3. Overall, therelative expression values obtained for each group ofanimals and for each parameter tested were quitehomogeneous, with minimal variations among the threeruns for each sample. The mean values and standarderrors shown in Table 3 derive from all calculationsperformed during the three individual runs for eachsample.

Statistically significant relative expression changeswere observed in the exposed groups for BMP-4,BMPR-IA, and -II, more pronounced in the animalsexposed to RFR during days 1–3 p.c. The changes inBMP-4 expression were consistent with those observedat the morphological level with immunohistochemistryfor the corresponding protein. The relative expressionvalues for BMP-4 were significantly increased afterRFR during both periods of exposure (P for both groupsvs. controls <0.001), although the values in the groupexposed during days 4–7 p.c. were lower than those in

the group exposed during days 1–3 p.c. (P¼ 0.001).The same pattern of relative expression aberrations wasobserved for BMPR-IA, although less pronounced(days 1–3 p.c. vs. control P< 0.001, days 4–7 p.c. vs.control P¼ 0.005, and 1–3 vs. 4–7 group P¼ 0.001).By contrast, changes in the relative expression ofBMPR-II in the irradiated newborns showed an inversepattern with a significant decrease in the group expos-ed during days 1–3 p.c. (P vs. control <0.001) andan increase in the group exposed during days 4–7 p.c.(P vs. control ¼0.067 (nonsignificant), P betweenirradiated groups ¼0.001). These patterns are shown inFigure 2D.

The relative expression ratios for BMP-7 werealways around 1 in the control animals, in agreementwith already published data about the correspondingprotein being the most abundantly expressed memberof the BMP family in the kidney. Nevertheless, theincrease observed for this transcript in the group expos-ed during days 4–7 p.c. was only marginally significant(P¼ 0.058). No significant changes were found forBMPR-IB.

Finally, ISH for BMPR-IA mRNA showed thismolecule to be upregulated in newborns exposed toRFR. In control tissues, BMPR-IA mRNA was localiz-ed mainly in the convoluted tubules and glomeruli,whereas after RFR it was additionally detected in thecollective tubular epithelium (Fig. 4A,B). These resultsare in agreement with those obtained with the EDASsoftware on BMPR-IA upregulation in the kidney ofRF-exposed animals. ISH for BMPR-II mRNA showedthat this molecule is expressed throughout the kidneyparenchyma in the control newborns. After exposure toRFR during days 1–3 p.c., no or very weak hybridiza-tion signals were observed for the BMPR-II mRNA,while the picture in the newborn kidneys after exposureduring days 4–7 p.c. did not differ from controls(Fig. 4D–F). The almost negative appearance of therenal tissue after RFR exposure during days 1–3 p.c. incomparison to the respective relative expression ratiosis not a discrepancy and can be explained by thedifferent sensitivity of the two methods. In fact, thisfinding reflects the downregulation of BMPR-II mRNAobserved with the duplex PCR protocols.

DISCUSSION

GSM-like RFR effects in vivo have been studiedin experimental animals, mostly by applying frequen-cies and conditions considered to affect humans[Brusick et al., 1998 {rev}; Heikkinen et al., 2001;Mausset et al., 2001; Bartsch et al., 2002; Braune et al.,2002; Di Carlo et al., 2002; Testylier et al., 2002]. Basedon the different resonant frequencies of humans and rats

220 Pyrpasopoulou et al.

Fig. 1. Bonemorphogenetic protein-4 (BMP-4) protein localization in the kidneys of control-shamexposed (A,C,E) and radiofrequency radiation (RFR) exposed (B,D,F) rat newborns, as shownbyimmunohistochemistry. In control animals BMP-4 was observed in the mesenchyme surroundingmainly the convoluted tubules in the cortex, while in irradiated animals BMP-4 was additionallyfound in the epithelial cells of collecting tubules in the medulla. m, medulla; c, cortex. Collectingtubulesarenegative for BMP-4 in controlanimalsandpositive in RFRexposedrat newborns.Origi-nalmagnifications:A,B:�25; C,D:�100; E,F:�400. [Color figure canbeviewedin theonline issue,whichisavailableatwww.interscience.wiley.com.]

RFREffects on Kidney BMPs 221

(1 vs. 10) [Thuery, 1992], the application of fieldsbetween 0.8 and 3 GHz on rat experimental modelsreflects the exposure of human organisms to 80 and300 MHz, which is far below GSM-like RFR. With thisin mind, the simulation model applied in our study wasdesigned in an attempt to adapt GSM-like RFRexposure on this specific experimental system, i.e.,the rat organism, taking into account that completeextrapolations of results are not possible. In addition, anextremely low power intensity and resulting SAR wasapplied (5 mW/cm2/0.5 mW/kg), in order to avoid any

thermal microwave effects from RFR; the relevant limitaccording to the EN50166-2 is 80 mW/kg.

Thermal effects are known to cause adverseeffects in vivo and in cell culture [reviewed by Brusicket al., 1998], induce heat shock responses at the cellularlevel [French et al., 2001] and interfere at different

TABLE 2. Semi-Quantitative Assessment of BMP-4 Altera-tions in the Kidneys of Sham (Control) and RFR ExposedNewborns by Immunohistochemistry

Cortex(mesenchyme)

Medulla(collecting tubules)

Negative þ þþ Negative þ

Controls 4 6 0 10 0Days 1–3 p.c. 0 3 7 0 10Days 4–7 p.c. 0 5 9 1 13

BMP, bone morphogenetic proteins.Signs: þ, positive; þþ, strongly positive.

Fig. 2. Representative results fromBMP-4 and -7 and BMPreceptoranalysisbyduplex RT-PCRinthekidneysamples fromcontrol (shamexposed) andRFRexposedrat newborns.Theelectrophor-esispictureshavebeeninverted for EDASrelativequantitationanalysis.Neg, negative PCR control(H2O).L,100 bpDNAladder.

TABLE 3. Relative BMP-4 and -7 and BMP ReceptorExpression in Whole Kidney Extracts in Comparison tob-Actin

Controls

RFR exposed

Days 1–3 p.c. Days 4–7 p.c.

BMP-4 0.24 (�0.04)a 1.41 (�0.05)b 1.05 (�0.02)b,c

BMP-7 1.09 (�0.02) 1.10 (�0.04) 1.16 (�0.04)BMPR-IA 0.34 (�0.02) 0.75 (�0.04)b 0.47 (�0.03)b,c

BMPR-IB 0.16 (�0.03) 0.16 (�0.05) 0.22 (�0.06)BMPR-II 0.54 (�0.05) 0.30 (�0.03)b 0.67 (�0.04)c

aMean (�standard error).bStatistically significant differences between RFR exposed versussham exposed (control) groups and between the two exposedgroups, respectively.cStatistically significant differences between RFR exposed versussham exposed (control) groups and between the two exposedgroups, respectively.

222 Pyrpasopoulou et al.

intensities with genomic DNA conditions [Lai andSingh, 1997; Tice et al., 2002]. Considered as non-thermal GSM–RFR effects, yet with power absorptionsof >1 W/Kg, are influences on human skin fibroblastmorphology and cell cycle related gene expression[Pacini et al., 2002], on human endothelial cell stressresponses [Leszczynski et al., 2002], as well as onchicken embryo hsp70 expression and cytoprotectionagainst hypoxia [Di Carlo et al., 2002].

At the frequency and power density tested in thisstudy, GSM-like RFR seems to have influenced theregulation of the expression of critical morphogeneticfactors, such as the BMP related genes, in newborn ratkidneys after exposure to this type of radiation duringthe early stages of embryonal development. Resultsfrom, respectively, low intensity RFR (at the order of<1 W/kg) on experimental models in vivo are currentlynot available in the literature. Comparable studies oncells in culture have shown that nonthermal GSM-likeRFR may induce increased c-fos expression [Goswamiet al., 1999], without significantly interfering with cellproliferation [Higashikubo et al., 2001]. Our data onaberrant BMP related gene expression and localizationin the rat newborn kidney provide evidence of persistent

changes at the molecular level that can be ascribed toGSM-like RFR.

How the altered profile in BMP expression iscaused by GSM-like RFR and whether these changesare of importance to the health of the newborn organismcannot be concluded from this study. However, twopoints should be discussed here: (i) the statisticallysignificant differences in the relative expression ofBMP-4 and type IA and II receptors that were observedafter prenatal exposure at different time periods of earlygestation and (ii) the significance of altered BMPexpression observed in newborn kidneys.

According to the whole body radiation model, weused [Brooks Air Base Model 2003], the averagepenetration depth, i.e., the depth up to which RFReffects should be considered with certainty, in the headof the pregnant dam was 5 mm, while in other parts ofthe body it was much lower. Overall, the deeper anorgan is located inside the animal, the less it would beaffected by RFR. Considering the head of the exposeddams, the cortex and pineal gland certainly receivedfully the emitted RFR, whereas the pituitary andhypothalamic area of the brain received lesser amountof RFR. It is obvious that whatever the effects of RFR on

Fig. 3. A,B,C:Scattergrampresentationof relative expressionresults for BMP-4,BMPR-IA, and -IIin the kidneys. In (D), the differences in the relative expression values (here: mean values for eachgroup) are shown collectively. D.p.c., days postcoitum. Note the more pronounced deviation fromcontrolvaluesafter RFRexposureduring1^3 d.p.c.

RFREffects on Kidney BMPs 223

the aforementioned areas in the brain of the dams, thesewere exerted during each period of exposure (days 1–3 p.c. vs. days 4–7 p.c.) to the same extent. RFR at verylow power intensity, but still higher than in our case, hasbeen shown to interfere with neurotransmitter levels ofthe hippocampal cholinergic system [Testylier et al.,2002], as well as of the Purkinje cells in the cerebellum[Mausset et al., 2001]. In man, GSM–RFR has beenshown to cause a transient decrease in thyrotropin levels[de Seze et al., 1998].

Concerning RFR effects on the developingembryo, we need to consider (a) the stages of embryonaldevelopment at the different periods of exposure and(b) the topography of the reproductive system in the rat.

In terms of embryonal development, days 1–3 p.c.in rodents correspond to embryogenesis, i.e., the stage

where exponential mitotic proliferation of the fertilizedovum takes place to form the blastocyst, which becomescavitated late on day 3. During this period the embryosare still surrounded by the zona pellucida whichprevents direct contact of the embryonal cells with thematernal environment. Implantation has not yet takenplace, and the embryos move through the oviductinto the respective horn of the uterus. The maternalenvironment (ovary, oviduct, and uterine mucosa)undergoes several hormonally guided changes, classi-cally considered to be the result of the hypothalamo–pituitary–gonadal axis, in order to receive the embryos.During the second period of our exposure experiments(days 4–7 p.c.), implantation of the embryos takesplace; this process starts on day 4 and is completedwithin the next 2 days. For the developing embryo, this

Fig. 4. In situ hybridization (ISH) for BMPR-IA (A^C) and BMPR-II (D^F) mRNAs on newborn kid-neys. Medullary areas with collecting tubules and cortical areas with glomeruli are shown forBMPR-IAandBMPR-II, respectively.A:Shamexposed, antisenseprobe,BMPR-IA.B:RFRexposedduringdays1^3p.c., sameprobe.C:Shamexposed, senseprobe,BMPR-IA.D:Shamexposed,anti-sense probe, BMPR-II. E: RFR exposed during days1^3 p.c., same probe. F: RFR exposed duringdays 4^7 p.c., sameprobe.Originalmagnifications:A^C,�100; D^F,�400.

224 Pyrpasopoulou et al.

period is characterized as early organogenesis; fromthe maternal point of view, more changes take place,as described above, and the maternal–embryonal cellinteractions, which start at the transit of day 3–4, arenow fully developed.

Concerning the topography of the rat reproductivesystem in terms of RFR accessibility in our experi-ments, the ovaries and oviducts are located very closeto the dorsal body wall. Both these structures, in-cluding the descending embryos contained therein,were (marginally) amenable to RFR. The two hornsof the uterus form a close V-shape toward the midlineof the animal, making it unlikely that RF waveswould fully, if at all, reach these areas. From all theabove it can be concluded that RFR exposure duringdays 1–3 p.c. (embryogenesis, pre-implantation) mayhave affected the maternal response and the developingembryo per se, while exposure during days 4–7 p.c.(early organogenesis, peri-implantation) mostly (orexclusively) affected the maternal response. Our datashow that RFR influences gene expression duringembryonal development, obviously in a time dependentmanner. It is possible, that the more pronounced effectson the relative expression of BMP-4, BMPR-IA, andBMPR-II after exposure to RFR during days 1–3 p.c.versus days 4–7 p.c. derive from the increased RFRabsorption during the first period for the reasonsmentioned above.

Again, whatever changes were induced at thelevel of regulation of gene expression during the firstweek of intrauterine development seem to be persistentor affect development beyond the removal of the RFRemitting source. There are currently no data on theinfluence of GSM-like RFR on the regulation of geneexpression during intrauterine rat development and anyattempt for such interpretations would be speculative.Nevertheless, increased c-fos expression as a GSM-likeRFR effect has been shown by two independent studieson different models, either transiently in vivo in therat brain [Fritze et al., 1997], or in mouse fibroblastsin transit from the exponential to the plateau phase[Goswami et al., 1999]. The immediate early responsegene c-fos, with c-jun as a heteromer, is important forthe expression and autoregulation of BMP-4 duringearly organogenesis in Xenopus [Knochel et al., 2000].It is also important for successful rat embryo implanta-tion that c-fos levels decrease around day 5 in the uterusbut remain stable in the brain [Verneaux et al., 1992].Checking for c-fos upregulation during and after in-trauterine RFR exposure, along with related pathways,might shed light on this hypothesis.

In terms of what might be the significance ofaltered BMP expression in the newborn kidney, someaspects about the development of this organ and the

related role of BMPs should be considered. Kidneysdevelop somewhat later than other organs, with meta-nephri appearing on embryonal day 13, glomeruli notpresent before day 18 and nephrons completely devel-oping only after birth [Wilson and Warkany, 1949].Kidneys depend on the BMP signaling pathway for theirdevelopment and function [Godin et al., 1999; Schedland Hastie, 2000; Martinez et al., 2001]. Among theever growing family of BMP growth factors, BMP-4and BMP-7 are the ones mostly involved in kidneydevelopment. BMP-7 seems to control the formation ofmetanephric mesenchyme and nephrons [Dudley et al.,1995, 1999; Luo et al., 1995; Karsenty et al., 1996; Jenaet al., 1997], while it remains expressed in the adultkidney [Bosukonda et al., 2000]. BMP 4 is an earlymorphogenetic factor in the kidney acting mainly onureteric buds to promote the development of the ureter[Miyazaki et al., 2000; Raatikainen-Ahokas et al.,2000], while high concentrations of this protein in vitroinhibit branching of the ureteric epithelium and nephro-genesis [Martinez et al., 2002].

The effects of these BMPs are mediated like allother effects of these morphogenetic agents via re-gulation of mesenchymal–epithelial interactions andare substantiated through BMP specific receptors,BMPR-IA and BMPR-IB for BMP-2 and -4 [Martinezet al., 2001], and specifically through BMPR-II forBMP-7 [Bosukonda et al., 2000]. Expression regulationof both these ligands and receptors follows a certaintiming during development; for example, although thepeak for type I receptors occurs on embryonal day 18,lowering thereafter [Ikeda et al., 1996], BMP type IIreceptors are continuously expressed into adult life inrat kidneys [Vukicevic et al., 1998; Bosukonda et al.,2000].

GSM-like RFR appears to influence BMP expres-sion in the kidney, because changes in the relativeexpression values for BMP transcripts, as well as aber-rant localization at the tissue and cellular level could beobserved days after exposure to RFR was ceased.However, the final effect of these alterations appears tobe of minor significance for the development of theorgan itself, since no major histological changes werefound. The expressed pattern of BMP-4 and BMPR-IAin the kidneys of exposed newborns and their aberrantlocalization in the collecting tubule system wouldrather be expected in earlier stages during histogenesisof this organ; overall, the observed effects from GSM-like RFR on BMPs may reflect a delay in kidneydevelopment.

In summary, aberrant expression of BMP-4 and itsreceptors BMPR-IA and BMPR-II were observed in thekidney of newborn rats whose mother animals had beenexposed to GSM-like RFR during early pregnancy. It

RFREffects on Kidney BMPs 225

seems that these persistent or delayed effects of GSM-like RFR are more pronounced when exposure takesplace during embryogenesis rather than early organo-genesis, suggesting that timing plays an important rolefor RFR effects on the developing organism.

REFERENCES

Bartsch H, Bartsch C, Seebald E, Deerberg F, Dietz K, Vollrath L,Mecke D. 2002. Chronic exposure to a GSM-like signal(mobile phone) does not stimulate the development ofDMBA-induced mammary tumors in rats: Results of threeconsecutive studies. Radiat Res 157:183–190.

Bodamyali T, Bhatt B, Hughes FJ, Winrow VR, Kanczler JM, SimonB, Abbott J, Blake DR, Stevens CR. 1998. Pulsed electro-magnetic fields simultaneously induce osteogenesis andupregulate transcription of bone morphogenetic proteins 2and 4 in rat osteoblasts in vitro. Biochem Biophys ResCommun 250:458–461.

Bosukonda D, Shih MS, Sampath KT, Vukicevic S. 2000.Characterization of receptors for osteogenic protein-1/bonemorphogenetic protein-7 (OP-1/BMP-7) in rat kidneys.Kidney Int 58:1902–1911.

Braissant O, Wahli W. 1998. Differential expression of peroxisomeproliferator-activated receptor-alpha, -beta, and -gammaduring rat embryonic development. Endocrinology 139:2748–2754.

Braune S, Riedel A, Schulte-Monting J, Raczek J. 2002. Influence ofa radiofrequency electromagnetic field on cardiovascular andhormonal parameters of the autonomic nervous system inhealthy individuals. Radiat Res 158:352–356.

Brooks Air Base Model. 2003. http://www.brooks.af.mil/AFRL/HED/hedr/reports/dielectric

Brusick D, Albertini R, McRee D, Peterson D, Williams G,Hanawalt P, Preston J. 1998. Genotoxicity of radiofrequencyradiation. DNA/Genetox Expert Panel. Environ Mol Muta-gen 32:1–16.

Celeste AJ, Iannazzi JA, Taylor RC, Hewick RM, Rosen V, WangEA, Wozney JM. 1990. Identification of transforming growthfactor beta family members present in bone-inductive proteinpurified from bovine bone. Proc Natl Acad Sci USA 87:9843–9847.

Coucouvanis E, Martin GR. 1999. BMP signaling plays a role invisceral endoderm differentiation and cavitation in the earlymouse embryo. Development 126:535–546.

de Seze R, Fabbro-Peray P, Miro L. 1998. GSM radiocellulartelephones do not disturb the secretion of antepituitaryhormones in humans. Bioelectromagnetics 19:271–278.

Di Carlo A, White N, Guo F, Garrett P, Litovitz T. 2002. Chronicelectromagnetic field exposure decreases HSP70 levels andlowers cytoprotection. J Cell Biochem 84:447–454.

Dudley AT, Lyons KM, Robertson EJ. 1995. A requirement for bonemorphogenetic protein-7 during development of the mam-malian kidney and eye. Genes Dev 9:2795–2807.

Dudley AT, Godin RE, Robertson EJ. 1999. Interaction betweenFGF and BMP signaling pathways regulate development ofmetanephric mesenchyme. Genes Dev 13:1601–1613.

EN 50166-2. 1995. Human exposure to electromagnetic fields—high frequency (10 kHz to 300 GHz).

French PW, Penny R, Laurence JA, McKenzie DR. 2001. Mobilephones, heat shock proteins, and cancer. Differentiation67:93–97.

Fritze K, Wiessner C, Kuster N, Sommer C, Gass P, Hermann DM,Kiessling M, Hossmann KA. 1997. Effect of global systemfor mobile communication microwave exposure on the geno-mic response of the rat brain. Neuroscience 81:627–639.

Godin RE, Robertson EJ, Dudley AT. 1999. Role of BMP familymembers during kidney development. Int J Dev Biol 43:405–411.

Goswami PC, Albee LD, Parsian AJ, Baty JD, Moros EG, PickardWF, Roti Roti JL, Hunt CR. 1999. Proto-oncogene mRNAlevels and activities of multiple transcription factors in C3H10T 1/2 murine embryonic fibroblasts exposed to 835.62 and847.74 MHz cellular phone communication frequencyradiation. Radiat Res 151:300–309.

Heikkinen P, Kosma VM, Hongisto T, Huuskonen H, Hyysalo P,Komulainen H, Kumlin T, Lahtinen T, Lang S, Puranen L,Juutilainen J. 2001. Effects of mobile phone radiation onX-ray-induced tumorigenesis in mice. Radiat Res 156:775–785.

Higashikubo R, Ragouzis M, Moros EG, Straube WL, Roti Roti JL.2001. Radiofrequency electromagnetic fields do not alter thecell cycle progression of C3H 10T and U87MG cells. RadiatRes 156:786–795.

IEEE Std C95.3. 1991. IEEE recommended practice for themeasurement of potentially hazardous electromagneticfields—RF and microwave.

Ikeda T, Takahashi H, Suzuki A, Ueno N, Yokose S, Yamaguchi A,Yoshiki S. 1996. Cloning of rat type I receptor cDNA for bonemorphogenetic protein-2 and bone morphogenetic protein-4,and the localization compared with that of the ligands. DevDyn 206:318–329.

Jena N, Martin-Seisdedos C, McCue P, Croce CM. 1997. BMP7 nullmutation in mice: Developmental defects in skeleton, kidney,and eye. Exp Cell Res 230:28–37.

Karsenty G, Luo G, Hofmann C, Bradley A. 1996. BMP 7 is requiredfor nephrogenesis, eye development, and skeletal patterning.Ann NY Acad Sci 785:98–107.

Knochel S, Schuler-Metz A, Knochel W. 2000. c-Jun (AP-1)activates BMP-4 transcription in Xenopus embryos. MechDev 98:29–36.

Lai H, Singh NP. 1997. Melatonin and a spin-trap compoundblock radiofrequency electromagnetic radiation-inducedDNA strand breaks in rat brain cells. Bioelectromagnetics18:446–454.

Leszczynski D, Joenvaara S, Reivinen J, Kuokka R. 2002. Non-thermal activation of the hsp27/p38MAPK stress pathway bymobile phone radiation in human endothelial cells: Mole-cular mechanism for cancer- and blood-brain barrier-relatedeffects. Differentiation 70:120–129.

Lipschutz JH. 1998. Molecular development of the kidney: A reviewof the results of gene disruption studies. Am J Kidney Dis 31:383–397.

Liu F, Ventura F, Doody J, Massague J. 1995. Human type II receptorfor bone morphogenic proteins (BMPs): Extension of thetwo-kinase receptor model to the BMPs. Mol Cell Biol 15:3479–3486.

Luo G, Hofmann C, Bronckers AL, Sohocki M, Bradley A, KarsentyG. 1995. BMP-7 is an inducer of nephrogenesis, and is alsorequired for development and skeletal patterning. Genes Dev9:2808–2820.

Magras IN, Xenos TD. 1997. RF radiation-induced changes in theprenatal development of mice. Bioelectromagnetics 18:455–461.

Martinez G, Loveland KL, Clark AT, Dziadek M, Bertram JF.2001. Expression of bone morphogenetic protein receptors

226 Pyrpasopoulou et al.

in the developing mouse metanephros. Exp Nephrol 9:372–379.

Martinez G, Mishina Y, Bertram JF. 2002. BMPs and BMP receptorsin mouse metanephric development: In vivo and in vitrostudies. Int J Dev Biol 46:525–533.

Massague J. 1996. TGF beta signaling: Receptors, transducers, andMad proteins. Cell 86:947–950.

Mausset AL, de Seze R, Montpeyroux F, Privat A. 2001. Effects ofradiofrequency exposure on the GABAergic system in the ratcerebellum: Clues from semi-quantitative immuno-histochemistry. Brain Res 912:33–46.

Miyazaki Y, Oshima K, Fogo A, Hogan BL, Ichikawa I. 2000.Bone morphogenetic protein 4 regulates the budding siteand elongation of the mouse ureter. J Clin Invest 105:863–873.

Nagai M, Ota M. 1994. Pulsating electromagnetic field stimulatesmRNA expression of bone morphogenetic protein-2 and -4.J Dent Res 73:1601–1605.

Pacini S, Ruggiero M, Sardi I, Aterini S, Gulisano F, Gulisano M.2002. Exposure to global system for mobile communication(GSM) cellular phone radiofrequency alters gene expression,proliferation, and morphology of human skin fibroblasts.Oncol Res 13:19–24.

Paria BC, Ma W, Tan J, Raja S, Das SK, Dey SK, Hogan BL. 2001.Cellular and molecular responses of the uterus to embryoimplantation can be elicited by locally applied growthfactors. Proc Natl Acad Sci USA 98:1047–1052.

Pizzonia J. 2001. Electrophoresis gel image processing and analysisusing the KODAK 1D software. Biotechniques 30:1316–1320.

Raatikainen-Ahokas A, Hytonen M, Tenhunen A, Sainio K, SaridaH. 2000. BMP-4 affects the differentiation of metanephricmesenchyme and reveals an early anterior–posterior axis ofthe embryonic kidney. Dev Dyn 217:146–158.

Reddi AH. 2000. Bone morphogenetic proteins and skeletaldevelopment: The kidney-bone connection. Pediatr Nephrol14:598–601.

Sandberg MM, Aro HT, Vuorio EI. 1993. Gene expression duringbone repair. Clin Orthop 289:292–312.

Schedl A, Hastie ND. 2000. Cross-talk in kidney development. CurrOpin Genet Dev 10:543–549.

Shimazaki S, Zacow RJ, Li D, Kim H, Iemura S, Ueno N, SampathK, Chang J, Erickson GF. 1999. A functional bone morpho-genetic protein system in the ovary. Proc Natl Acad Sci USA96:7282–7287.

Stagg RB, Thomas WJ, Jones RA, Adey WR. 1997. DNA synthesisand cell proliferation in C6 glioma and primary glial cellsexposed to a 836.55 MHz modulated radiofrequency field.Bioelectromagnetics 18:230–236.

Testylier G, Tonduli L, Malabiau R, Debouzy JC. 2002. Effects ofexposure to low level radiofrequency fields on acetylcholinerelease in hippocampus of freely moving rats. Bioelectro-magnetics 23:249–255.

Thuery J. 1992. Microwaves: Industrial and medical applications.London: Artech House. p 418.

Tice RR, Hook GG, Donner M, McRee DI, Guy AW. 2002.Genotoxicity of radiofrequency signals. I. Investigation ofDNA damage and micronuclei induction in cultured humanblood cells. Bioelectromagnetics 23:113–126.

Verneaux V, Jouvenot M, Compagnone P, Propper AY, Adessi GL.1992. Expression of c-fos and c-myc genes in uterus andnonreproductive organs during the implantation period in therat. Biochem Cell Biol 70:142–148.

Vukicevic S, Basic V, Rogic D, Basic N, Shih MS, Shepard A,Jin D, Dattatreyamurty B, Jones W, Dorai H, Ryan S,Griffiths D, Maliakal J, Jelic M, Pastorcic M, StavljenicA, Sampath TK. 1998. Osteogenic protein-1 (bone mor-phogenetic protein-7) reduces severity of injury afterischemic acute renal failure in rat. J Clin Invest 102:202–214.

Wang EA, Rosen V, Cordes P, Hewick RM, Kriz MJ, Luxenberg DP,Sibley BS, Wozney JM. 1988. Purification and characteriza-tion of other distinct bone-inducing factors. Proc Natl AcadSci USA 85:9484–9488.

Wilson JG, Warkany J. 1949. Aortic arch and cardiac anomalies inthe offspring of vitamin A deficient rats. Am J Anat 85:113–155.

Xenos ThD, Magras IN. 2003. Low power density RF-radiationeffects on experimental animal embryos and foetuses.Chapter 5.6. In: Stavroulakis P, editor. Biological effectsof electromagnetic fields. Berlin: Springer-Verlag. pp 579–602.

RFREffects on Kidney BMPs 227