B

TtplnTmtdT

RD9

Ic

A

Int. J. Radiation Oncology Biol. Phys., Vol. 68, No. 1, pp. 187–195, 2007Copyright © 2007 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/07/$–see front matter

doi:10.1016/j.ijrobp.2006.12.057

IOLOGY CONTRIBUTION

INHIBITION OF TRANSFORMING GROWTH FACTOR-� SIGNALING INNORMAL LUNG EPITHELIAL CELLS CONFERS RESISTANCE TO IONIZING

RADIATION

ANNA REEVES, PH.D.,* MARIANNA ZAGUROVSKAYA, M.D.,* SEEMA GUPTA, PH.D.,*MOHAMMED M. SHAREEF, PH.D.,* MOHAMMED MOHIUDDIN, M.D.,† AND

MANSOOR M. AHMED, PH.D.*‡

*Weis Center for Research, Geisinger Clinic, Danville, PA; †Geisinger-Fox Chase Cancer Center, Geisinger Clinic, Wilkes-Barre,PA; and ‡Department of Radiation Oncology, Fox Chase Cancer Center, Philadelphia, PA

Purpose: To address the functional role of radiation-induced transforming growth factor-� (TGF-�) signaling ina normal epithelial background, we selected a spontaneously immortalized lung epithelial cell line derived fromthe normal lung tissue of a dominant-negative mutant of the TGF-� RII (�RII) transgenic mouse thatconditionally expressed �RII under the control of the metallothionein promoter (MT-1), and assessed this cellline’s response to radiation.Methods and Materials: A spontaneously immortalized lung epithelial cell culture (SILECC) was established andall analyses were performed within 50 passages. Colony-forming and terminal transferase dUPT nick endlabeling (TUNEL) assays were used to assess clonogenic inhibition and apoptosis, respectively. Western-blotanalysis was performed to assess the kinetics of p21, bax, and RII proteins. Transforming growth factor-�-responsive promoter activity was measured using dual-luciferase reporter assay.Results: Exposure to ZnSO4 inhibited TGF-� signaling induced either by recombinant TGF-�1 or ionizingradiation. The SILECC, treated with either ZnSO4 or neutralizing antibody against TGF-�, showed a significantincrease in radio-resistance compared to untreated cells. Furthermore, the expression of �RII inhibited theradiation-induced up-regulation of the TGF-� effector gene p21waf1/cip1.Conclusions: Our findings imply that inhibition of radiation-induced TGF-� signaling via abrogation of the RIIfunction enhances the radio-resistance of normal lung epithelial cells, and this can be directly attributed to theloss of TGF-� signaling function. © 2007 Elsevier Inc.

TGF-�1 signaling, Normal lung cells, Ionizing radiation, Radiation resistance, p21waf1/cip1.

ct

uoit(ptil

fiNteGt

A

INTRODUCTION

he transforming growth factor-� (TGF-�) family of mul-ifunctional cytokines has a wide range of physiologic andathologic effects on cellular processes ranging from pro-iferation and differentiation to apoptosis (1, 2). Originallyamed for their transforming activities in in vitro assays,GF-� isoforms now unequivocally demonstrate both tu-or-suppressor and oncogenic activities. Dysregulation of

hese processes can result in various fibrotic and malignantiseases. In normal tissues, the suppressor activities ofGF-� dominate, whereas during tumorigenesis, the

Reprint requests to: Mansoor M. Ahmed, Ph.D., Weis Center foresearch, Geisinger Clinic, Office 121A, 100 N. Academy Ave.,anville, PA 17822-2616. Tel: (570) 214-3972; Fax: (570) 214-861; E-mail: mmahmed@geisinger.eduThis work was supported by grants from the National Cancer

nstitute (NCI) (R01CA86937) and from the Kentucky Lung Can-er Program to M.M.A.

Conflict of interest: none.

cknowledgments—The authors are grateful to Dr. L. M. Wake-

187

hanges in TGF-� expression-associated cellular responsesilt the balance in favor of its oncogenic activities (3).

Cellular proliferation is complex, and involves both stim-latory and inhibitory signals. The abnormal proliferationbserved in cancer cells is caused by mutations that eitherncrease positive signals or decrease negative growth-con-rol signals, or both. Transforming growth factor-� isoformsTGF-�1, TGF-�2, and TGF-�3) are 25-kD homodimerolypeptides that regulate cell proliferation and differentia-ion (4). Transforming growth factor-� molecules are potentnhibitors of growth in a variety of cells, including epithe-ial, endothelial, and lymphoid cells (5). Transforming

eld of the Laboratory of Cell Regulation and Carcinogenesis,ational Cancer Institute, Bethesda, MD, for providing AM3

ransgenic DNR mice breeders. The authors gratefully acknowl-dge Dr. Craig C. Wood of the Center for Health Research of theeisinger Clinic, Wilkes-Barre, PA, for his technical assistance in

he statistical analysis of the experimental data.Received Sept 29, 2006, and in revised form Dec 27, 2006.

ccepted for publication Dec 27, 2006.

gtpfnGtbt

rbjsw(pdctctre

puided

irsftorigsp

R2msrtlrta

ii

tiapmfi(Tsliel(ctlce

Gt

wMrUSsfdWisAfem(vawcs

I

nuAl

I

ww

188 I. J. Radiation Oncology ● Biology ● Physics Volume 68, Number 1, 2007

rowth factor-� is expressed in growth-arrested cells and inhe G1 phase of the cell cycle, and contributes to an orderlyrogression through the cell cycle (6). Transforming growthactor-� signals by binding to transmembrane serine-threo-ine kinases, which are termed receptor I (RI) and RII (7).enetic evidence shows that both receptors are required for

he signaling of TGF-� (7). A third receptor (RIII) is notelieved to be directly involved in TGF-� signaling, but actso present TGF-� to RII (7).

Upon TGF-� treatment, RI is recruited to RII to form aeceptor-ligand complex on the cell surface, and is activatedy type II receptor kinase upon phosphorylation of itsuxtamembrane GS domain. This activates RI, which thenerves as a docking site for Smad-2 or Smad-3 proteins,hich are associated with the chaperone protein SARA

Smad anchor for receptor activation) (8). Following phos-horylation by RI, the Smad 2 and 3 protein complexissociates from RI, associates with Smad 4, and translo-ates into the nucleus. The Smad complex then activates theranscription of target genes (such as p21waf1/cip1 and otheryclin-dependent kinase [CDK] inhibitors) through an in-ermediate transcription factor or by binding to DNA di-ectly. Thus, the translocated Smads are thought to be keyffectors for TGF-�-mediated growth inhibition (7).

Although Smad signaling has a central role in the TGF-�athway, Smad-independent TGF-� signaling was also doc-mented (9, 10). For instance, p21-activated kinase (PAK2)s activated by TGF-� in a Smad 2- and Smad 3-indepen-ent manner and is specific to fibroblast cells, but not topithelial cells, indicating that the response to TGF-� isifferentially regulated in different cell types (8).It is well-known that radiation induces the TGF-�1 isoform

n various cell types (11–13). Interestingly, radiation down-egulates the TGF-�3 isoform, and does not alter the expres-ion of the TGF-�2 isoform (12, 13). Transforming growthactor-�1 is implicated primarily for negative growth regula-ion (14) and cell death (15, 16). Based on exclusive inductionf TGF-�1 by radiation and its implication on negative growthegulation and apoptosis, we hypothesized that radiation wouldnduce an endogenous TGF-� protein that will exert clono-enic inhibition and cause apoptosis in cells with intact TGF-�ignaling components. This hypothesis was demonstrated andreviously proved by our group (17, 18).

Lung tumors are known to harbor aberrant expression ofII, and to lack or contain mutations in Smad-4 expression (19,0). The finding that RII and Smad-4 expression is low inany lung-cancer cell lines raises the possibility that an ab-

ence of RII and Smad-4 expression may cause a lack ofesponse to TGF-�, and this may play a critical role in resis-ance to chemotherapy and radiation. In various cell typesacking RII expression, ectopic re-expression of RII led to aestoration of TGF-� signaling (21). Zhao and Young reportedhat RII is indispensable for mediating both the mitogenic andntiproliferative effects of TGF-� in lung fibroblasts (22).

We previously reported that restoration of TGF-� signal-ng enhances sensitivity to ionizing radiation in a pancreat-

c-cancer background (17). In the same study, we reported G

hat in the pancreatic cancer cell line MIA PaCa-2, radiationnduces TGF-� signaling, up-regulates p21waf1/cip1 and Bax,nd activates caspases (17). Using a loss-of-function ap-roach with a dominant-negative mutant for RII in normalouse embryonic fibroblasts, we demonstrated that thesebroblast cells were resistant to radiation-induced apoptosis17). To address the functional role of radiation-inducedGF-� signaling in a normal epithelial background, weelected a spontaneously immortalized lung epithelial celline and disrupted the RII receptor gene, and assessed itsmpact on radio-sensitivity. To achieve this, we isolated andstablished a mouse lung cell line derived from the normalung tissue of a dominant-negative mutant of a TGF-� RII�RII) transgenic mouse that conditionally expressed �RII,ontrolled by the zinc-regulated promoter. Here, we reporthat the RII-defective, spontaneously immortalized epithe-ial normal lung cells are significantly more radio-resistantompared to the primary lung cells that express functionalndogenous RII with intact TGF-� signaling.

METHODS AND MATERIALS

eneration of primary mouse cell culture from �RIIransgenic mouse

The AM3 transgenic DNR mice breeders of the strain FVB/Nere kindly provided by Dr. Lalage Wakefield (NCI, Bethesda,D). The mice were housed in a pathogen-free environment, and

eceived care according to guidelines in the Guide for the Care andse of Laboratory Animals (Department of Health and Humanervices publication NIH-85-23). They were crossbred to obtain aizeable �RII mouse population. Primary lung cells were isolatedrom DNR transgenic mice, and cultured according to protocolescribed previously (23, 24). Briefly, cells were resuspended inaymouth medium containing 2.5% fetal bovine serum, antibiot-

cs, 50 mg/mL amphotericin B, and soybean trypsin inhibitor, andeeded into chamber slides coated with air-dried rat-tail collagen.fter 2 hours of recovery and attachment, hormones and growth

actors were added to the medium. After the epithelial cells reached anxponential growth phase with no fibroblast cells present (approxi-ately five passages), cultures were grown in regular DMEM

GIBCO BRL, Invitrogen, Carlsbad, CA) containing 10% fetal bo-ine serum (GIBCO BRL) and 1% penicillin/streptomycin at 37°Cnd 5% CO2. Since the lung primary cultures exceeded 10 passages,e named these cells the “spontaneously immortalized lung epithelial

ell culture” (SILECC). All experiments were begun after 10 pas-ages, and were completed within 50 passages.

mmunocytochemistryCytokeratin expression was determined in SILECC by immu-

ocytochemical analysis, as we described previously (17, 25),sing pan-cytokeratin antibody (Santa Cruz, CA) and the EliteBC kit (Vector Laboratories, Burlingame, CA). We used A549

ung adenocarcinoma cells as a positive control for cytokeratin.

rradiationA 100-kV industrial X-ray machine (Phillips, New York, NY)

as used to irradiate cultures at room temperature. The dose rateith a 2-mm aluminum plus 1-mm beryllium filter was 3.85

y/min, at a focus–surface distance of 20 cm.

Sr

aBcamFf2att(f�Ba5�pDs

W

stwasuo(VCiwuP

D

mlkdMfatGZinbiSet

C

tlAcTnpXw(b

Q

lwdcronwTdcffiepm

S

sebZu

Fbfmg7fgts

189TGF-� RII inhibition-mediated radio-resistance ● A. REEVES et al.

emiquantitative reverse transcription-polymerase chaineaction

Total RNA was isolated from exponentially growing untreatednd ZnSO4-treated SILECC, using a TRIzol reagent (GIBCO-RL). One microgram of total RNA was reverse-transcribed intoomplementary DNA (cDNA) by using random hexamer primersnd reverse transcriptase in a 20-�L reaction mix according to theanufacturer’s recommendations (Perkin-Elmer, Waltham, MA).ive microliters of this cDNA were used as the starting templateor the primary polymerase chain reaction (PCR) (35 cycles). The5 �L of PCR mix (Perkin-Elmer) contained 0.5 U Taq polymer-se, reaction buffer, 1.5 mM MgCl2, 200 �M deoxynucleotideriphosphates, and 1.0 �M of forward and reverse primers flankinghe cDNA sequence of the RII gene, as we reported previously26). The primers were designed to amplify a 123-base pair (bp)ragment of the human RII sequence between nucleotides �7 and573, encoding extracellular and transmembrane domains (Gene-ank accession no. M85079) (Fig. 1). The primer sequences weres follows: 5=-GTCTGCCATGGGTCGGGGGC-3= (sense) and=-GTGACTATCATGTCGTTATT-3= (antisense). The control-actin gene was amplified from the same pool of cDNA. The PCRroducts were resolved on ethidium bromide-stained agarose gel.ensitometry was performed to obtain a semiquantitative expres-

ion of RII as the ratio between RII and �-actin.

estern blot analysisThree types of treatment were performed. The SILECC was

ubjected to 5 Gy irradiation alone, ZnSO4 alone, and ZnSO4

reatment immediately followed by 5 Gy radiation. Control cellsere left untreated. Protein extracts were prepared at 1, 3, 6, 12,

nd 24 h after treatment. Total protein extracts from SILECC wereubjected to Western-blot analysis as described previously (25),sing anti-TGF-� RII antibody (sc-400, Santa Cruz Biotechnol-gy, Santa Cruz, CA) and polyclonal anti-human TGF-� RIIresidues 1–28) antibody (Upstate Biotechnology, Charlottesville,A), or p21waf1/cip1 (sc-6246, Santa Cruz Biotechnology, Santaruz, CA), or Bax (6A7, BD-Pharmingen, San Jose, CA) antibod-

es. Anti-�-actin antibody (Sigma Chemical Co., St. Louis, MO)as used as internal loading control. These proteins were detectedsing an enhanced chemiluminescence method (GE Healthcare,

ig. 1. Schematic representation of �RII transgene. A 19.1-kilo-ase (kb) SalI DNA fragment containing �RII (DNR) was releasedrom pMT-LCR-�RII and used for microinjections. MT-1, mouseetallothionein I promoter; hGH, poly(A) region of human

rowth-hormone gene; 10-kb 5 locus control region (LCR) and-kb 3 LCR, 10-kb EcoRI fragment of the 5 LCR and 7-kb EcoRIragment of the 3 LCR of the mouse metallothionein (MT) I and IIene locus (28). Location of primers for PCR is shown. Nru 1 ishe name of the restriction enzyme that cleaves the sequence in theite indicated.

iscataway, NJ). v

ual-luciferase reporter assayTransforming growth factor-�-responsive promoter activity waseasured using 6XSBE-Luc, a Smad 2/3- and Smad 4-dependent

uciferase reporter kindly provided by Dr. Scott Kern (Johns Hop-ins University, Baltimore, MD). A 6XSBE-Luc (1.6 �g/mL me-ia) and a pRL-TK expression plasmid (0.4 �g/mL, Promega,adison, WI) were cotransfected into lung cells, using Lipo-

ectamine Plus transfection reagent (Invitrogen, Carlsbad, CA)ccording to the manufacturer’s protocol. Twenty-four hours afterransfection, cells were treated with a combination of radiation (5y) or human recombinant TGF-� (10 ng/mL), with or withoutnSO4 (0.12 �M). To block radiation-induced TGF-�, a neutral-

zing antibody against TGF-�1 (AF-101-NA, R&D Systems, Min-eapolis, MN) was added (500 ng/mL media) to the cells 1 hefore irradiation. Firefly luciferase activity was assayed 24 h afterrradiation, using a TD-20/20 Luminometer (Turner Designs,unnyvale, CA), and was compared to the activity obtained for thempty vector. All results were normalized against the activity ofhe cotransfected Renilla luciferase vector pRL-TK.

olony-forming assayFor clonogenic cell-survival studies, two different cell concen-

rations in quadruplet sets were used for each radiation dose. Cellines were left untreated or exposed to a 1 6 Gy dose of radiation.fter incubation for 10 or more days, colonies were stained with

rystal violet, and the colonies containing �50 cells were counted.he surviving fraction (SF) was calculated as a ratio between theumber of colonies formed and the product of the number of cellslated with the plating efficiency. The curve was plotted using–Y log scatter (Delta Graph 4.0, SPSS, Inc., Chicago, IL), andas fitted by using the formula for the single-hit multitarget

SHMT) model to obtain the D0 a dose required to reduce survivaly 37% and survival fraction (SF2) (17).

uantitation of apoptosisApoptosis was quantified by terminal transferase dUTP nick end

abeling (TUNEL) staining (Roche Diagnostics, Indianapolis, IN),hich detects DNA strand breaks by terminal transferase-mediatedUTP-digoxigenin end-labeling. Briefly, SILECC was seeded inhamber slides, and after 24 h, cells were treated with 5 Gy ofadiation or human recombinant TGF-� (10 ng/mL) with or with-ut ZnSO4 (0.12 �M). To block radiation-induced TGF-�, aeutralizing antibody to TGF-�1 (AF-101-NA, R&D Systems)as added (500 ng/mL media) to the cultures before irradiation.wenty-four hours after treatment, the DNA was then tailed withigoxigenin-dUTP and conjugated with anti-digoxigenin fluores-ein. Cells were counterstained with propidium iodide and anti-ade. The stained specimen were observed in a triple band-passlter, using a Nikon-Microphot (Donsanto Corp., Natick, MA)pifluorescence microscope. To determine the percentage of apo-totic cells, four independent experiments were performed, and ainimum of 1,000 cells was scored in each experiment.

tatistical analysisDescriptive statistics, such as the calculation of means and

tandard deviations, were used to examine the data from separatexperiments. To examine the difference in survival curves by doseetween the “radiation alone” group and the “radiation plusnSO4” treatment group, a multiple linear regression model wassed. The logarithm of the surviving fraction was the outcome

ariable, and dependent variables included group, dose, and inter-

aWaw

Et

tombtopr

lwfseRp3ds

lewlpifi�s

S

ZaalecepS

Rsu

s

190 I. J. Radiation Oncology ● Biology ● Physics Volume 68, Number 1, 2007

ction between group and dose. The nonparametric Kruskall-allis test was used to determine if there were differences in

poptotic cell frequency. All tests were two-sided, and p � 0.05as considered significant.

RESULTS

xpression of dominant-negative mutant for TGF-� RIIransgene (�RII) is regulated by ZnSO4

To determine whether ZnSO4 regulates the expression ofhe transgene in SILECC, the RNA and protein expressionf the dominant-negative mutant of TGF-� RII was deter-ined using reverse transcriptase (RT)-PCR and Western

lotting, respectively. The RT-PCR results demonstratedhe induction of expression of �RII at either 0.5 or 2.5 �Mf ZnSO4 treatment (Fig. 2a), with no presence of theroduct in the untreated lane, suggesting that ZnSO4 tightlyegulates transgene �RII expression in SILECC.

To further confirm the expression of �RII at the proteinevel and to establish a dose–response relationship, SILECCas treated with different concentrations of ZnSO4 ranging

rom 0.2 to 1 �M, and total proteins were extracted andeparated by sodium dodecyl sulphate-polyacrylamide gellectrophoresis (SDS-PAGE). Using both antibodies againstII (sc-400 and polyclonal anti-RII antibody), a 32-kD �RIIrotein was detected at all concentrations of ZnSO4 tested at

h, 6 h, and 24 h after treatment, but no protein wasetected in untreated cells. Mouse �-actin was used on the

Fig. 2. Regulation of the expression of �RII by ZnSO4 trreaction and Western-blot analysis. (a) Reverse transcripof �RII mRNA from SILECC cells in the presence (laneTotal mRNA was isolated from SILECC that was treatereaction (35 cycles) was performed using the reverse-trwith ethidium bromide, and intensities of bands were qexpression normalized to �-actin. (b) Western-blot analyexpression. Cells were left untreated (UT) or treated witsc-400 anti-RII antibody (Santa Cruz), anti-RII antibBiotechnology), and �-actin as loading control.

ame blot as a loading control (Fig. 2b). Interestingly, the p

evels of endogenous RII (at 70 kD) were found to belevated in ZnSO4-treated SILECC. This is in accordanceith a previous study in which the presence of �RII protein

ed to an increase in endogenous receptor RI, RII, and RIIIroteins, and this increase does not interfere with the dom-nant-negative function of �RII protein (27, 28). Thesendings demonstrate that both RNA and protein levels ofRII in SILECC can be directly regulated by ZnSO4 expo-

ure.

ILECC expresses cytokeratinHaving demonstrated the regulation of transgene �RII by

nSO4, we next determined the expression of cytokeratin toscertain that the SILECC was of epithelial origin. Evenfter 50 passages, the SILECC derived from mouse normalung epithelial tissue of �RII transgenic mice maintained anpithelial morphology (Fig. 3a), since these cells expressedytokeratin (Fig. 3b). Thus, the presence of cytokeratinxpression, together with morphologic data showing theolygonal shapes typical of epithelial cells, suggests thatILECC behaves as normal epithelial lung cells.

ecombinant TGF-� or radiation failed to induce TGF-�ignaling in SILECC treated with ZnSO4 but not inntreated SILECCTo assess the effect of �RII protein expression on TGF-�

ignaling in SILECC cells, TGF-�-responsive SBE-Luc re-

t, as assessed by reverse transcriptase-polymerase chainolymerase chain reaction was used to detect expressionor absence (lane 7) of ZnSO4. Lane 8, positive control.ZnSO4 for time intervals indicated. Polymerase chain

tion reaction and appropriate primers. Gel was staineded by densitometry. In graph, columns represent �RIIs performed to detect �RII and endogenous RII protein

4 for time intervals indicated. Blots were probed usingrgeting the amino terminus of RII protein (Upstate

eatmentase-ps 1–6)d with

anscripuantifisis wah ZnSOody ta

orter activity was analyzed by a transient transfection

asmerSbsps

pstoc

t�ssZ

Tei

iTsiai

as pos

191TGF-� RII inhibition-mediated radio-resistance ● A. REEVES et al.

pproach (Fig. 4a). Transforming growth factor-�-respon-ive promoter activity (as assessed by Smad-binding ele-ent activity) was expressed as the ratio of Firefly lucif-

rase activity and Renilla luciferase activity. Humanecombinant TGF-�1 up-regulated the TGF-�-responsiveBE-promoter activity, while treatment with ZnSO4 in com-ination with recombinant TGF-� abrogated TGF-�-re-ponsive SBE-promoter activity, suggesting that the �RIIrotein induced by ZnSO4 effectively blocked the TGF-�ignaling function.

Based on the abrogation of TGF-� function by �RIIrotein, we hypothesized that radiation-induced the TGF-�ignaling function would be abrogated by �RII protein. Toest this hypothesis, SILECC was treated with a 5 Gy dosef ionizing radiation, with or without ZnSO4. Irradiationaused induction of TGF-�-responsive SBE-promoter ac-

Fig. 3. Cytokeratin expression in SILECC. (a) MicropMagnification, �150. (b) The SILECC was stained withadenocarcinoma cells) was stained with pan-cytokeratin

Fig. 4. Inhibition of TGF-� signaling by ZnSO4 in SILEC6XSBE luciferase constructs, and treated for 24 h withserum. (b) The SILECC cells were transfected with the swith 5 Gy in the presence of serum. Transforming groexpressed as the ratio of firefly luciferase activity to

independent experiments. Error bars represent standard error.

ivity, and this activity was abrogated by the expression ofRII induced by ZnSO4 (Fig. 4b). These results strongly

uggest that both TGF-� and radiation activate the TGF-�ignaling pathway, and this is potentially abrogated by thenSO4-induced expression of �RII protein.

he transgene �RII protein protects normal lungpithelial cells from radiation-induced clonogenicnhibition and apoptosis

We previously reported that restoration of TGF-� signal-ng led to enhanced sensitivity to ionizing radiation inGF-�-defective pancreatic cancer cells (17). In the presenttudy, we addressed the functional role of TGF-� signalingn radiation-inducible clonogenic inhibition (Fig. 5) andpoptosis in a normal cell background. The SILECC wasrradiated with or without ZnSO4 to assess the response to

aph of SILECC cells in standard growing conditions.toxylin and eosin and pan-cytokeratin; and A549 (lungitive control.

The SILECC cells were cotransfected with pRL-TK andand/or recombinant human TGF-�1 in the absence ofnstructs, treated for 24 h with ZnSO4, and/or irradiated

ctor-�-responsive promoter activity (6XSBE-Luc) wasa luciferase activity. The data are the mean of three

hotogrhema

C. (a)ZnSO4

ame cowth faRenill

rtciGZsenatsetwap0LI

TSfici

RZ

icgttwaastTWbFadaspbetpificctU

FCea

192 I. J. Radiation Oncology ● Biology ● Physics Volume 68, Number 1, 2007

adiation by a colony-forming assay for clonogenic inhibi-ion, and a TUNEL assay for apoptosis (Fig. 6). In theolony-forming assay, the SF2 value of exponentially grow-ng, irradiated SILECC was 0.355, with a D0 value of 1.71y (Fig. 5). On the other hand, the SF2 for SILECC withnSO4 treatment was 0.515, with a D0 value of 2.78 Gy,uggesting that compared to SILECC with no �RII proteinxpression, the SILECC expressing �RII protein was sig-ificantly more resistant to ionizing radiation. Statisticalnalysis, using a multiple-regression model, indicated thathe differences between the two slopes were statisticallyignificant (p � 0.0001). Interestingly, both curves failed toxhibit a shoulder, since both have n � 1.1, suggesting thathere was no accumulation of sublethal damage. Consistentith the observation of clonogenic inhibition, the TUNEL

nalysis (Fig. 6) indicated that the SILECC with �RIIrotein expression was significantly more resistant (p �.0001) to ionizing radiation-inducible apoptosis than SI-ECC with no ZnSO4 treatment (29), at 24 and 48 h.

nterestingly, treatment with neutralizing antibodies against

Fig. 6. Transgene �RII protects SILECC from radiation-indicated, and terminal transferase dUTP nick end labetreatment. Columns show percentages of TUNEL-positiv

ig. 5. Expression of �RII increases radio-resistance of SILECC.lonogenic survival of SILECC exposed to radiation in the pres-nce or absence of ZnSO4 was assessed by colony-forming assaynd analyzed by single-hit multitarget model curve fit.

represent standard error.

GF-� caused a reversal of radiation-induced apoptosis inILECC (p � 0.0001), at 24 and 48 h. Together, thesendings demonstrated that radiation resistance in normalells can be directly attributed to the loss of TGF-� signal-ng function.

adiation-induced p21 up-regulation is inhibited bynSO4 treatmentThe p21waf/cip1 is a p53 response gene, whose product

nhibits the function of CDKs during the G1 phase of theell cycle (30, 31). The SILECC harbors a wild-type p53ene function. It is also known that TGF-� signalinghrough the Smad complex activates the transcription ofarget genes such as p21waf1/cip1 and other CDK inhibitors,hich are speculated to be key effectors for TGF-�-medi-

ted growth inhibition (32). Further, we reported that radi-tion-induced TGF-� induces the Bax protein (17). Wepeculated that the increased radiation resistance in ZnSO4

reated-SILECC might be mediated by abrogation of theGF-�-induced expression of p21waf1/cip1 and Bax by �RII.estern-blot analysis showed a basal level of p21waf1/cip1 in

oth untreated and ZnSO4-treated SILECC. As shown inig. 7, the levels of p21waf1/cip1 were up-regulated after 1nd 3 h in response to radiation plus ZnSO4 treatment, wereown-regulated at 6 and 12 h, and were completely absentt 24 h. However, with radiation alone, significant andustained induction of p21waf1/cip1 was observed at all timeoints. Further, no change in Bax levels was observedetween treated and untreated groups. In this normal lungpithelial cell line, induction of Bax may not be transcrip-ionally regulated by radiation. This is in accordance with arevious report that no induction of Bax in response toonizing radiation was detected in normal human diploidbroblasts (33) or in normal prostate epithelial and stromalells (34). Rather, radiation may cause activation of Bax viaonformational changes by exposing the NH2-terminus andhe carboxy terminal (COOH)-terminus to mitochondria.nder such conformational activation of Bax, no changes

d apoptosis. The SILECC was exposed to the conditionsUNEL) assay was performed after 24 h and 48 h of

. Data represent a mean of four experiments. Error bars

induceling (Te cells

mlpwptps

tretao(fcwwsst3Rmeii

mittfipial

iatf

ttgrrttttidrgicpmtticiswireintigodtn

ol.

193TGF-� RII inhibition-mediated radio-resistance ● A. REEVES et al.

ay be observed at either the RNA or protein expressionevel. These data suggest that the down-regulation of21waf1/cip1 protein in irradiated and ZnSO4-treated SILECCas due to the abrogation of TGF-� signaling by the ex-ression of �RII, since endogenous wild-type p53 transac-ivation function will be disrupted because of reduced phos-horylation of serine 18 caused by disruption of TGF-�ignaling (35).

DISCUSSION

In the present study, we adopted a novel ex vivo approacho investigate the functional role of TGF-� signaling inadiation-induced apoptosis and clonogenic inhibition in anstablished, spontaneously immortalized primary lung epi-helial cell line derived from normal lung epithelial tissue oftransgenic mouse expressing a dominant-negative mutant

f RII under the control of the metallothionein promoter28). The advantage of this approach is that unlike trans-ormed tumor cells, these cells are normal except that theyontain a silent �RII that could be induced, when necessary,ith zinc. With this approach, no other factors associatedith the transformed tumor-cell phenotype or paracrine

tromal–epithelial interactions would interfere with TGF-�ignaling. In particular, the paracrine stromal–epithelial in-eraction is a well-established event in tumorigenesis (36,7). A recent report indicates that specific inactivation ofII in fibroblasts, but not in epithelial cells, in transgenicice brought about an increase in proliferation of adjacent

pithelia (36). Hence, this novel approach establishes anmportant model for understanding the role of radiation-nduced TGF-� in normal lung epithelium.

We confirmed that SILECC maintains normal epithelialorphology. These cells showed the polygonal shapes typ-

cal of epithelial cells. No altered cell morphology or cy-oskeletal organization was observed. It is well-documentedhat cytokeratin polypeptides constitute the intermediatelament cytoskeleton of epithelial cells (38). Further, theatterns of cytokeratin expression represent markers of ep-thelial differentiation status. We found that SILECC had anbundant level of cytokeratin, suggesting that this normal

Fig. 7. Radiation-induced p21waf1/cip1 up-regulation is iwere used in Western blotting of total lysates of SILECC5 Gy alone, or treated with ZnSO4 exposure followed bp21waf1/cip1 and Bax. �-actin was used as loading contr

ung-cell culture is of epithelial origin. The abundant stain- i

ng of cytokeratin found in SILECC using a pan-cytokeratinntibody can also be attributed to the fact that normal lungissue has been known to harbor the expression of variousorms of cytokeratin polypeptides (38).

We previously showed that RII expression is decreased inhe highly tumorigenic pancreatic cell line MIA PaCa-2, andhe expression of functional RII restores TGF-�-mediatedrowth inhibition caused by both recombinant TGF-� andadiation (17). In this present study, using a novel system toegulate TGF-� signaling in a normal epithelial backgroundhrough controlled expression of �RII, we demonstratedhat the abrogation of endogenous TGF-� function causeshe following responses: (a) increased proliferative poten-ial, as assessed by growth curves (data not shown); (b)nhibition of radiation-induced or recombinant TGF-�-in-uced Smad signaling, as assessed by a TGF-�-responsiveeporter assay; (c) resistance to radiation-induced clono-enic inhibition; and (d) resistance to recombinant TGF-�-nduced or radiation-induced apoptosis. Our findings areonsistent with those of Bottinger et al. (28), where primaryancreatic acinar cell cultures from control mice showedarked inhibition, as judged by [3H]thymidine incorpora-

ion when treated with all isoforms of TGF-�. At the sameime, acinar cells from the transgenic mouse line AM3 werensensitive to all isoforms of TGF-� but not activin, indi-ating that DNR expression only inactivates TGF-� signal-ng, but not signaling by other members of the TGF-�uperfamily (28). Further, these findings are in agreementith our previous report on the abrogation of radiation-

nduced TGF-�-responsive promoter activity with enhancedesistance to radiation-induced apoptosis in primary mousembryonic fibroblasts (MEF) transfected with an RII-dom-nant negative mutant (17). It is of note that treatment witheutralizing antibodies against TGF-� also reversed radia-ion-induced apoptosis, indicating that these changes werendeed dependent on TGF-� signaling in SILECC. To-ether, these observations correspond well with our previ-us reports, in which the expression of �RII significantlyiminished radiation-induced apoptosis in MEF cells, andhe treatment of pancreatic MIA PaCa-2/RII cells witheutralizing anti-TGF-� antibodies caused these cells to

d by treatment with ZnSO4. Antibodies to p21waf1/cip1

s were left untreated (UT), treated with ZnSO4 alone orof radiation. Blots were probed with antibody against

nhibite. Cell

y 5 Gy

ncrease their radiation resistance (17).

pbtMreoiscotTfwadt

mdncafdwfbtb

srots

1

1

1

1

194 I. J. Radiation Oncology ● Biology ● Physics Volume 68, Number 1, 2007

In our previous reports (17, 18), radiation failed to induce21waf/cip1 protein expression in MIA PaCa cells, probablyecause the p53 and TGF-� pathways are not functional inhese cells. However, restoration of TGF-� signaling in

IA PaCa cells led to p21waf/cip1 protein elevation afteradiation. This induction was abrogated when the cells werexposed to anti-TGF-�-neutralizing antibody. Similarly, inur present study, we found that the expression of �RIInhibited radiation-induced p21waf/cip1 up-regulation, whichuggests that enhanced radio-resistance of SILECC may beonferred by the abrogation of TGF-�-induced expressionf p21waf/cip1. Since p21waf/cip1 is regulated by p53, it standso reason that there is a crosstalk between p53 and theGF-� signaling pathways. It was documented that p53

unction is required for TGF-� response by cooperatingith Smads (39). Also, it was reported that transcriptional

ctivation of p21waf/cip1 by TGF-� requires p53. Ewan et al.emonstrated that abrogation of TGF-� function could lead

o decreased p53 serine-18 phosphorylation in irradiated t

REFEREN

Res 1996;146:561–568.

1

1

1

1

1

1

2

2

2

2

2

ammary epithelial tissue (35). In previous reports, weocumented the induction of p21waf/cip1 when TGF-� sig-aling was restored in mutant p53 MIA PaCa-2 pancreatic-ancer cells (17). In addition, we demonstrated that thebrogation of endogenous TGF-� function by �RII in p53unctional MEFs caused enhanced resistance to radiation-in-uced apoptosis (17). These previous observations, togetherith the present study, demonstrate that abrogation of TGF-�

unction may impair p53-mediated p21waf/cip1 up-regulation,ecause of either impaired p53 transactivation function (owingo reduced levels of phospho-p53), or an impaired Smad:p53inding ratio to the p21waf/cip1 promoter for its activation.

In conclusion, the findings of this study underscore theignificance of TGF-� function, because targeting the dys-egulation of TGF-� function with simultaneous retentionf function of all other pivotal pathways in a normal epi-helial cell background can impact significant cellular re-ponse, leading to enhanced cell proliferation and resistance

o the killing effects of radiation.

CES

1. Massague J. TGF-beta signal transduction. Annu Rev Biochem1998;67:753–791.

2. Whitman M. Smads and early developmental signaling by theTG-beta superfamily. Genes Dev 1998;12:2445–2462.

3. de Caestecker MP, Piek E, Roberts AB. Role of transforminggrowth factor-beta signaling in cancer. J Natl Cancer Inst2000;92:1388–1402.

4. Roberts AB, Sporn MB. The transforming growth-betas. In:Sporn MB, Roberts AB, editors. Peptide growth factors andtheir receptors. Heidelberg: Springer-Verlag; 1991.

5. Lyons RM, Moses HL. Transforming growth factors and theregulation of cell proliferation. Eur J Biochem 1990;187:467–473.

6. Howell GM, Sun L, Ziober BL, et al. The role of growthregulatory aberrations in progression of human colon carci-noma. Cancer Metastasis Rev 1993;12:275–286.

7. Wrana JL, Attisano L, Wieser R, et al. Mechanism of activa-tion of the TGF-beta receptor. Nature 1994;370:341–347.

8. Wilkes MC, Murphy SJ, Garamszegi N, et al. Cell-type-specific activation of PAK2 by transforming growth factorbeta independent of Smad2 and Smad3. Mol Cell Biol 2003;23:8878–8889.

9. Engel ME, McDonnell MA, Law BK, et al. InterdependentSMAD and JNK signaling in transforming growth factor-beta-mediated transcription. J Biol Chem 1999;274:37413–37420.

0. Hocevar BA, Brown TL, Howe PH. TGF-beta induces fi-bronectin synthesis through a c-Jun N-terminal kinase-depen-dent, Smad4-independent pathway. EMBO J 1999;18:1345–1356.

1. Suardet L, Li C, Little JB. Radio-induced modulation oftransforming growth factor beta1 sensitivity in a p53 wild-typehuman colorectal-cancer cell line. Int J Cancer 1996;68:126–131.

2. O’Malley Y, Zhao W, Barcellos-Hoff MH, et al. Radiation-induced alterations in rat mesangial cell Tgfb1 and Tgfb3 geneexpression are not associated with altered secretion of activeTgfb isoforms. Radiat Res 1999;152:622–628.

3. Wang J, Robbins ME. Radiation-induced alteration of ratmesangial cell transforming growth factor-beta and expressionof the genes associated with the extracellular matrix. Radiat

4. Zimmerman CM, Padgett RW. Transforming growth factorbeta signaling mediators and modulators. Gene 2000;249:17–30.

5. Shima Y, Nakao K, Nakashima T, et al. Activation ofcaspase-8 in transforming growth factor-beta-induced apopto-sis of human hepatoma cells. Hepatology 1999;30:1215–1222.

6. Motyl T, Gajkowska B, Ploszaj T, et al. Expression andsubcellular redistribution of Bax during TGF-beta1-inducedprogrammed cell death of HC11 mouse mammary epithelialcells. Cell Mol Biol (Noisy-le-Grand) 2000;46:175–185.

7. Ahmed MM, Alcock RA, Chendil D, et al. Restoration oftransforming growth factor-beta signaling enhances radiosen-sitivity by altering the Bcl-2/Bax ratio in the p53 mutantpancreatic cancer cell line MIA PaCa-2. J Biol Chem 2002;277:2234–2246.

8. Alcock RA, Dey S, Chendil D, et al. Farnesyltransferaseinhibitor (L-744,832) restores TGF-beta type II receptor ex-pression and enhances radiation sensitivity in K-ras mutantpancreatic cancer cell line MIA PaCa-2. Oncogene 2002;21:7883–7890.

9. Hougaard S, Norgaard P, Abrahamsen N, et al. Inactivation ofthe transforming growth factor beta type II receptor in humansmall cell lung cancer cell lines. Br J Cancer 1999;79:1005–1011.

0. Jakowlew SB, Moody TW, You L, et al. Reduction in trans-forming growth factor-beta type II receptor in mouse lungcarcinogenesis. Mol Carcinog 1998;22:46–56.

1. Sun L, Wu G, Willson JK, et al. Expression of transforminggrowth factor beta type II receptor leads to reduced malig-nancy in human breast cancer MCF-7 cells. J Biol Chem1994;269:26449–26455.

2. Zhao Y, Young SL. Requirement of transforming growthfactor-beta (TGF-beta) type II receptor for TGF-beta-inducedproliferation and growth inhibition. J Biol Chem 1996;271:2369–2372.

3. Logsdon CD. Pancreatic duct cell cultures: there is more toducts than salty water. Gastroenterology 1995;109:1005–1009.

4. Logsdon CD, Williams JA. Pancreatic acinar cells in mono-layer culture: direct trophic effects of caerulein in vitro. Am J

Physiol 1986;250:G440–G447.

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

195TGF-� RII inhibition-mediated radio-resistance ● A. REEVES et al.

5. Fralix KD, Ahmed MM, Mattingly C, et al. Characterizationof a newly established human pancreatic carcinoma cell line,UK Pan-1. Cancer 2000;88:2010–2021.

6. Venkatasubbarao K, Ahmed MM, Mohiuddin M, et al. Dif-ferential expression of transforming growth factor beta recep-tors in human pancreatic adenocarcinoma. Anticancer Res2000;20:43–51.

7. Tang B, de Castro K, Barnes HE, et al. Loss of responsivenessto transforming growth factor beta induces malignant trans-formation of nontumorigenic rat prostate epithelial cells. Can-cer Res 1999;59:4834–4842.

8. Bottinger EP, Jakubczak JL, Roberts IS, et al. Expression of adominant-negative mutant TGF-beta type II receptor in trans-genic mice reveals essential roles for TGF-beta in regulationof growth and differentiation in the exocrine pancreas. EMBOJ 1997;16:2621–2633.

9. Abe T, Ouyang H, Migita T, et al. The somatic mutationfrequency of the transforming growth factor beta receptor typeII gene varies widely among different cancers with microsat-ellite instability. Eur J Surg Oncol 1996;22:474–477.

0. el-Deiry WS, Tokino T, Velculescu VE, et al. WAF1, apotential mediator of p53 tumor suppression. Cell 1993;75:817–825.

1. Hu YX, Watanabe H, Ohtsubo K, et al. Infrequent expressionof p21 is related to altered p53 protein in pancreatic carci-

noma. Clin Cancer Res 1998;4:1147–1152.

2. Hu PP, Datto MB, Wang XF. Molecular mechanisms oftransforming growth factor-beta signaling. Endocr Rev 1998;19:349–363.

3. Sun Y, Tran BN, Worley LA, et al. Functional analysis of thep53 pathway in response to ionizing radiation in uveal mela-noma. Invest Ophthalmol Vis Sci 2005;46:1561–1564.

4. Bromfield GP, Meng A, Warde P, et al. Cell death in irradi-ated prostate epithelial cells: role of apoptotic and clonogeniccell kill. Prostate Cancer Prostatic Dis 2003;6:73–85.

5. Ewan KB, Henshall-Powell RL, Ravani SA, et al. Transform-ing growth factor-beta1 mediates cellular response to DNAdamage in situ. Cancer Res 2002;62:5627–5631.

6. Bhowmick NA, Chytil A, Plieth D, et al. TGF-beta signalingin fibroblasts modulates the oncogenic potential of adjacentepithelia. Science 2004;303:848–851.

7. Bhowmick NA, Moses HL. Tumor-stroma interactions. CurrOpin Genet Dev 2005;15:97–101.

8. Schlage WK, Bulles H, Friedrichs D, et al. Cytokeratin ex-pression patterns in the rat respiratory tract as markers ofepithelial differentiation in inhalation toxicology. I. Determi-nation of normal cytokeratin expression patterns in nose,larynx, trachea, and lung. Toxicol Pathol 1998;26:324–343.

9. Cordenonsi M, Dupont S, Maretto S, et al. Links betweentumor suppressors: p53 is required for TGF-beta gene re-

sponses by cooperating with Smads. Cell 2003;113:301–314.