ICANCER RESEARCH 58. 2588-2593. June 15. 1998]

Preferential Cytotoxicity of Cells Transduced with Cytosine Deaminase Comparedto Bystander Cells after Treatment with 5-Flucytosine1

Theodore S. Lawrence,2 Alnawaz Rehemtulla, Emily Y. Ng, Melinda Wilson, James E. Trosko, and Philip L. Stetson

Departments of Radiation Oncology ¡T.S. L, A. R., E. Y. N.¡and Pharmacology ¡P.L S./, University of Michigan Medical Center, Ann Arbor, Michigan 48/09-0010: and

Department of Pediatrics and Human Genetics, Michigan Slate University, East Lansing, Michigan 48224 [M. W., J. E. T.}

ABSTRACT

In vitro experiments from our laboratory and others have suggestedthat herpes simplex virus thymidine kinase (HSV-TK)/ganciclovir (GCV)

gene therapy depends on gap junctional intercellular communication(GJIC) to produce a strong bystander effect. Furthermore, we have shownthat cells transduced with HSV-TK can be protected from GCV-mediated

toxicity by GJIC with bystander cells. We wished to determine whetherGJIC affected either the bystander or protective effect of the cytosinedeaminase (CD)/5-flucytosine (5-FC) gene therapy approach, in which CDconverts 5-FC to 5-fluorouracil (5-FU). To test this, we designed a cocul-ture system using communication-competent WB rat hepatocytes and a

noncommunicating subclone (aBl), which were transduced with CD andwith antibiotic resistance genes so that we could independently determinethe survival of the CD-containing or bystander cells. We found that,compared to the HSV-TK/GCV strategy, bystander killing resulting fromtreatment with CD/5-FC does not depend on GJIC. However, our moststriking finding was that both communication-competent and -incompetent CD-transduced cells were preferentially killed, by a factor of up to500, compared to bystander cells. The lesser dependence of the CD/5-FC

system on GJIC, combined with the finding that most cancer cells lack thecapacity for GJIC, suggest that the CD/5-FC system may be superior tothe HSV-TK/GCV approach for gene therapy. However, the prematuredeath of the CD-transduced 5-FU "factory" suggests that other strategies

may be necessary to produce a sufficient quantity of 5-FU for a duration

long enough to produce permanent tumor regression.

INTRODUCTION

Two of the most widely studied methods of enzyme prodrug genetherapy involve the use of the viral enzyme HSV-TK3 with the

prodrug GCV or the bacterial enzyme CD with the prodrug 5-FC(1-3). In the former approach, GCV is a far better substrate forHSV-TK than for mammalian thymidine kinase, leading to preferen

tial production of GCV monophosphate and, ultimately, cytotoxicity.In the latter strategy, CD can convert the prodrug 5-FC to 5-FU, which

has both cytotoxic and radiosensitizing properties (4). A recent comparison of these two approaches in cell culture (5) and implantedtumors (6) suggested that the CD/5-FC system was more effective,

although the mechanism underlying the improved efficacy of CD/5-FC is not known.

One possible reason that the CD/5-FC system might be superior tothe HSV-TK/GCV approach is that the former may depend less on

GJIC for bystander killing. Because gene therapy approaches appearto transduce no more than 1-5% of tumor cells, it is crucial that the

transduced cells effectively kill surrounding cells which have not been

Received 12/22/97; accepted 4/16/98.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

'Supported by N1H Grants POI CA42761 (to T. S. L.) and ROI CA21104 (to

J. E.T.).2 To whom requests for reprints should be addressed, at Department of Radiation

Oncology. University of Michigan. 1500 East Medical Center Drive, UH-B2C490, AnnArbor. MI 48109-0010. Phone: (313) 647-9955; Fax: (313) 763-7371; E-mail:

tsl@umich.edu.' The abbreviations used are: HSV-TK, herpes simplex virus thymidine kinase; GCV.

ganciclovir; CD, cytosine deaminase; 5-FC, 5-flucytosine; GJIC, gap junctional intercellular communication; FdUMP, fluorodeoxyuridine monophosphate; 5-FU, 5-fluorouracil;

FRAP, fluorescence recovery after photobleaching.

transduced (the "bystander effect"). Although there exist a number ofpossible mechanisms for a bystander effect (see "Discussion"), the

most direct mechanism involves the transfer of toxic metabolites fromone cell to another. In the case of the strategy using HSV-TK and

GCV, the toxic metabolite, phosphorylated GCV, cannot transit themembrane. GJIC permits the transfer of molecules up to Mr ~ 1000(which includes phosphorylated GCV; Refs. 7-9) directly from one

cell to another. This is consistent with the finding that communicationcompetent cells demonstrate a greater bystander effect than do incompetent cells (10, 11). In the case of the CD/5-FC approach, althoughthe toxic phosphorylated metabolites of 5-FU, such as FdUMP,

would, like phosphorylated GCV, require GJIC to affect a bystandercell, 5-FU itself can diffuse across the cell membrane to neighboringcells. Therefore, although the CD/5-FC strategy might have some

dependence on GJIC, the influence of communication status should besubstantially less than that observed for the HSV-TK/GCV approach.

This difference assumes greater importance in light of the fact that thecapacity for GJIC has been lost in most cancer cells (12-15) and was

not found on ultrastructural examination of the tumors in the studydescribed above (6).

Therefore, we wished to determine the role of GJIC in these twoenzyme prodrug strategies. To carry out these studies, we used communication-competent WB rat hepatocytes and a communication-

incompetent subclone (aBl). We chose experimental designs thatpermitted us to quantify separately the effects of drug treatment onboth the transduced cell and the bystander cell. Using this approach,we have recently confirmed that the in vitro bystander effect forHSV-TK-induced killing is strongly dependent on GJIC, and we madethe novel observation that communication-competent HSV-TK-trans-

duced cells could be protected by communication competent neighbors (the "Good Samaritan" effect; Ref. 11). We next wished to carry

out analogous studies to assess the role of GJIC on the efficacy of theCD/5-FC strategy. In confirmation of our hypothesis, we found thattreatment of cocultures of CD-transduced cells produced bystander

killing that was independent of GJIC. Similarly, although bystandercells protected CD-transduced cells, this protection showed little, if

any, dependence on GJIC.However, we also discovered that the transduced cells were much

more effectively killed by 5-FC than were the bystander cells, almostsurely because of the high intracellular concentrations of 5-FU that are

generated. Although successful gene therapy will require that thetransduced cells ultimately be killed, preferential killing of the transduced cells compared to the bystander cells might be disadvantageousbecause the transduced cell (the 5-FU "factory") would die before

producing sufficient 5-FU to cause tumor regression. This suggests

that strategies that produce more uniform killing of the transduced andbystander cells need to be explored.

MATERIALS AND METHODS

Cell Culture. Cells were cultured under standard conditions using DMEMcontaining 10% bovine calf serum, penicillin-streptomycin, and 2 mM gluta-

mine with a final thymidine concentration of 0.2 ¿J.M.All cells were checkedfor mycoplasma every 3 months.

2588

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PREFERENTIAL CYTOTOXICITY OF CD-TRANSDl'fKD CELLS

Cell Lines. Communication-competent WB rat hepatocytes and the communication-incompetent aBl subclone have been described (16). To constructWB-CD and aBl-CD cells, the CD coding sequence was PCR-amplified using

the DNA of Kohara phages 137 and 138 as template (kindly provided by F.Nienhardt, University of Michigan, Ann Arbor, MI; Ref. 17). The 5' primer

(GTCGACGAAGCAATCATGTCGAATAACGCTTTA) was designed to introduce a Kozak sequence and an initiator ATG (in place of GTG) formammalian cell expression, as well as a Sail site. The 3' primer (TA-

CAAACGTTGAACGACTGGAATTCTCTAGA) introduces an EcoRl and anXbal site after the stop codon. The amplified fragment was cloned into themammalian expression vector pZ (kindly provided by Genetics Institute,Cambridge, MA) and sequenced to confirm the sequence. We transfected WBand aBl cells with this plasmid (pZ-CDl) using LipofectAMINE (Life Tech

nologies, Gaithersburg, MD), and 48 h after transfection, cells were split10-fold in the presence of 600 fig/ml G418. The resulting neomycin-resistant

colonies were isolated and screened for the presence of the CD protein byWestern blot (Fig. 1) using CD-specific antibodies (described below). Puro-mycin-resistant WB and aB 1 cells were constructed similarly using the pPURplasmid and selected in 2 /Lig/ml puromycin. MCF-7 cells were obtained from

American Type Culture Collection.Western Analysis. Logarithmically growing cells were trypsinized,

counted, and lysed with NP40 lysis buffer [1% Nonidet -40, 150 mM NaCl.

50 mM Tris (8.0), and 0.05% SDS] containing a cocktail of protease inhibitors(BMB, Indianapolis. IN). Aliquots of the lysates were resolved using standardSDS-PAGE techniques and transferred onto nitrocellulose membrane. CD wasdetected using a CD-specific rabbit antiserum (kindly provided by Dr. BrianHuber, Glaxo-Wellcome, Raleigh, NC) and processed for horseradish perox-idase-induced chemiluminescense, as recommended by the manufacturer

(Pierce, Rockford, IL).Assessing GJIC by FRAP. The communication status of mixed cultures

was determined essentially as described previously (18) using a MeridianACAS 570 (Meridian Instruments, Inc. Okemos, MI), coupled with a computerworkstation. Briefly, one cell type was incubated with fluorescent beads(Fluoresbrite Carboxylate, Llo-jim diameter; Polysciences, Inc., Warrington,

PA), which it phagocytosed, thus permitting it to be distinguished from theother cell types in a coculture. Cells were cocultured for 12-16 h and loadedwith carboxyfluorescein. Single bead-loaded cells were then photobleached

and observed for recovery of fluorescence by laser rescanning. The totalfluorescence recovery over 3 min after photobleaching was determined for12-50 selected cell sites for each coculture condition.

o9

WB-CD oraBl-CD

_WB-puro oraBl-puro

m ODm

CytosineDeaminase

Fig. 1. Characterization of CD-transduced cells. WB and aBl cells were transfectedwith pZ-CD, and the resulting clones (WB-CD and aB 1-CD) were assessed for expressionof CD. The parental cells and CD transfected cells were assessed for CD by Westernanalysis, as described in "Materials and Methods."

high density

Trypsinize, piati1 at clonal density

WB-CD oraBl-CD

WB-puro oraBl-puro

Fig. 2. Schematic for experimental design. See text tor details.

Assessing GJIC by Scrape Loading. Cells were grown to near confluencein 35-mm culture dishes. The medium was removed, and the cultures wererinsed three times with prewarmed calcium magnesium-free PBS. One ml of a0.05% solution of Lucifer yellow in CMF-PBS was added to each dish. Theedge of a razor blade was pressed into the monolayer to form the "scrape line,"

along which the dye enters the cells. After 1 min, the dye was removed, andthe cultures were rinsed three times with prewarmed PBS. After 15 min at37°C, the PBS was removed, and the cells were fixed with 3.7% (v/v)

formaldehyde in PBS for 15 min. after which the dishes were washed threetimes with PBS. The culture dishes were examined under blue excitation(450-490 nm) using a Leitz Laborlux D microscope equipped with an epif-

lourescence unit. Fields were examined for cells beyond those along the scrapeline, which had taken up the dye (19).

Cell Survival Assay. Cocultures using different ratios of transduced tobystander cells were constructed so that the total cell density was constant.Separate cultures were plated for each coculture condition to control for thenumber of CD-transduced cells. The actual ratios of cells at the time of 5-FCaddition were 4-7% and 45-60% transduced when the nominal ratios were5:95 and 50:50. respectively. Cocultures were treated with 5-FC for 48 h. Cellsurvival of the neomycin-resistant CD-expressing cells (aBl-CD and WB-CD)and the puromycin-resistant bystander cells (aBl-puro and WB-puro) was

determined by plating cells into selective medium (Fig. 2) and assessed usinga standard clonogenic assay, as described previously (20). The plating efficiencies of WB-puro and aBl-puro cells in puromycin were 0.48 ±0.08 and0.49 ±0.05, respectively. The plating efficiencies of WB-CD and aBl-CD

cells in G418 was 0.50 ±0.03 and 0.69 ±0.02, respectively. The survivingfractions of the puro clones in G418 and the CD clones in puromycin were all<10~5. Cocultures were —90% confluent at the time of processing for cell

survival.Determination of 5-FU. Conversion from 5-FC to 5-FU by tissue culture

cells was measured using a gas chromatographic/mass spectrometric assay, asdescribed previously (20). Briefly, conditioned medium was collected andderivatized, and quantification of the derivatized products was performed usinga Hewlett-Packard 5987A gas chromatograph/mass spectrograph in selectedion-monitoring mode.

Apoptosis. Cells were trypsinized and washed once with PBS. They werethen fixed by incubation in 4% paraformaldehyde at a concentration of I0h

cells/ml for 30 min at room temperature and then washed again with PBS. Thefixed cells were resuspended in PBS at IO4 cells/ml. Thirty-five /¿Iof the cell

suspension were mixed with 5 /il of 100 mg/ml acridine orange solution and10 p\ of Vectashield (Vector Laboratories. Burlingame. CA). The mixture wasexamined using a Leitz Laborlux S microscope equipped with a 1-LambdaPleomopak incident light fluorescence illuminator (450-490-nm excitationwavelength with 520-nm barrier filter). Cells were scored as apoptotic if they

exhibited both chromatin condensation and cellular shrinkage. Data are presented as the mean percentages of apoptotic cells.

Statistical Analysis. Data are presented as means ±SE. Student's t test

and the F test were used to compare two means and multiple means, respectively. Statistical significance was defined at the level of P < 0.05 (two-tailed).

2589

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

l'KIIIKIMIAI ( YIOKIXICITY OF CD-TKANSDll ID CELLS

Prebleach Bleached Recovery

a)

Fig. 3. Determination of the communication status of WB-CD andaBl-CD cells by FRAP. u, WB:WB-CD coculture. The "pre-hleached" image shows a WB cell (containing latex beads) against abackground of WB-CD cells. The "bleached" image was obtained

immediately alter photobleaching of the WB cell and a WB-CD cell.The "recovery" image was obtained 3 min after pholobleaching. Both

cells showed recovery of fluorescence, indicating efficient GJIC. h.aBl-CD:WB coculture. The prebleached image shows an aBl-CD

cell (containing latex beads) against a background of WB cells. Thebleached and recovery images were obtained as described in <;.TheaBl-CD cell showed minimal recovery of fluorescence, demonstrating lack of GJIC. whereas the WB cell shows recovery (positivecontrol), f. summary of FRAP data. X axis, cell pairs: *. beaded cell.

*. fluorescence recovery from a single cell pair. Twelve lo 50 cellswere assessed for each condition.

b)

c)20

~ 18

I 16£fr 14

I "S 10uoCouSo

00

î$III

Co-culture group

RESULTS

To test our hypothesis that bystander cell killing resulting fromtreatment of cocultures of cells transduced with CD was relativelyindependent of GJIC. it was first necessary carry out two types ofcontrol experiments. First, we needed to confirm the communicationstatus of these cells. It has been previously demonstrated that WBcells are communication competent and that the aBl clone derivedfrom WB cells does not show detectable GJIC (16). We began with adetailed assessment of WB-CD and aBl-CD cells using FRAP (Fig.3). As described in "Materials and Methods," we cocultured WB cells

(marked by latex beads) with WB-CD cells and photobleached eithera WB cell or a WB-CD cell. We found that there was efficient transfer

of dye from the surrounding cells into the bleached cell, demonstrating communication competence (Fig. 3, a and c). In contrast, when anaBl-CD cell (marked by beads) was photobleached, dye transfer fromsurrounding cells did not occur, even when aBl-CD cells were cocul

tured with WB cells (communication competent; Fig. 3, b and c).Scrape loading, as described in "Materials and Methods," confirmed

these results (data not shown) and demonstrated that WB-puro andaBl-puro cells also maintained the communication status of the parental cell (Fig. 4). Second, we needed to determine the 5-FU sensitivity of WB-puro, aBl-puro, WB-CD, and aBl-CD cells. We foundthat there were no significant differences in 5-FU sensitivity amongthe WB-puro, aBl-puro, and aBl-CD cells, although WB-CD cellswere more resistant to 10 JAM5-FU than the other cell lines (Fig. 5).

We could then determine whether the communication status of cellsinfluenced the bystander effect. For these experiments, WB-CD (communication-competent) cells were cocultured with either (communication-competent) WB-puro cells or (communication-incompetent)aBl-puro cells. Cocultures were constructed using a ratio of transduced cells to bystander cells of either 50:50 or 5:95, exposed to 5-FC,

and assessed for survival of the bystander cells according the methoddescribed in Fig. 2. Whereas 1 mM 5-FC had no effect on WB-purocells or aBl-puro cells cultured alone, there was substantial WB-puroand aB 1-puro cell killing (a strong bystander effect) at the higher ratio

and a weaker one at the lower ratio. However, bystander killing did2590

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PREFERENTIAL CYTOTOXICITY OF CD-TRANSDUCED CELLS

Fig. 4. Determination of the communicationstatus of WB-puro and aB J-puro cells by scrapeloading. WB-puro and aBl-puro cells were cul

tured and assessed for GJIC by scrape loading asdescribed in "Materials and Methods." A, WB-puro

cells. B, aBl-puro cells. Cells that are capable ofGJIC (WB-puro) exhibit penetration of the dye

through several cell layers beyond the scrape line.In communication-incompetent cells (aBl-puro).

Lucifer yellow is present only in those cells alongthe scrape line.

not depend on the communication competence of the cells (Fig. 6A).Similar experiments were carried out using aBl-CD cells coculturedwith either WB-puro cells or aBl-puro cells, which confirmed thatcommunication status had no influence on the ability of CD-trans

duced cells to kill bystander cells (Fig. 6B).We next wished to determine whether communication status af

fected the survival of the transduced cells. We conducted these experiments because, as described in the "Introduction," we had found

that communication-competent bystander cells protected HSV-TK-transduced cells to a far greater extent than did communication-

incompetent bystander cells (11). We found that bystander cells were

[5FU] faM)10

Fig. 5. Sensitivity of cells to 5-FU. WB-CD. WB-puro, aBl-CD, and aBl-puro cellswere exposed to 5-FU for 48 h. They were then assessed for clonogenicity as describedin "Materials and Methods." Dala points, means of three independent experiments; barn,

SE. •WB-CD; •,aBl-CD; D, WB-puro; O, aBl-puro.

able to protect CD-transduced cells from 5-FC-mediated cytotoxicity.However, in contrast to our results with HSV-TK-transduced cellstreated with GCV, communication-competent WB-puro cells did notconsistently provide better protection than communication-incompetent aBl-puro cells (Fig. 7/4). Furthermore, communication-incompetent aBl-CD cells were protected from I mM 5-FC by coculture witheither aBl-puro or WB-puro cells (Fig. IB). Therefore, although it is

possible that GJIC mediates part of the protection, a substantialcomponent seems unrelated to communication status.

Our findings that the capacity for GJIC had little influence on theresulting clonogenicity of cells bystander cells suggested that thesecretion of 5-FU was chiefly responsible for bystander killing. Therefore, we measured the 5-FU concentrations resulting from treatmentwith 5-FC in medium that we had saved at the time the cocultures

described above were processed for clonogenic survival. This permitted us to construct a plot of the surviving fraction of both theCD-transduced cells and the bystander cells as a function of theextracellular 5-FU concentration generated by that coculture (Fig. 8).

We found that bystander cell cytotoxicity increased only slightly withthe extracellular concentration of 5-FU. However, when the ratio ofCD-transduced cells to bystander cells was 5:95, both WB-CD andaBl-CD cells were killed to a significantly greater extent than thebystander WB-puro and aBl-puro cells for the same extracellular5-FU concentration. Likewise, CD-transduced cells were growth ar

rested to a far greater extent than were bystander cells (data notshown). When the ratio of CD-transduced cells to bystander cells was50:50, extracellular 5-FU concentrations were in the range of 3-6 /AM,

and the loss of clonogenicity of the transduced and bystander cells didnot consistently differ (data not shown).

WB-CD aB1-CD

Fig. 6. Lack of dependence of bystander killing on GJIC. WB-CDcells (A) or aBl-CD cells (B) were cocultured with either WB-puro cells(•)or aB I -puro cells (•}at varying ratios containing either 0. 5, or 50%CD-transduced cells (0:100, 5:95, or 50:50. respectively). Cells wereexposed to 5-FC for 48 h and were then processed to assess theclonogenic fraction of bystander WB-puro or aB 1-puro cells (see Fig. 2).Data points, means of four or five independent experiments; bars. SE.

§

i0.0001

0.01

0.001

0 0.2 0.4 0.6 0.8

[5-FC] (mM)

2591

0.2 0.4 0.6 0.8 1

[5-FC] (mM)

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PREFERENTIAL CYTOTOXICITY OF CD-TRANSDUCED CELLS

Fig. 7. Lack of dependence of CD-transduced cell protection on GJIC.WB-CD cells M) or aBI-CD cells (B) were cultured alone (A), withWB-puro cells (•),or with aBl-puro cells (•).Cocultures contained 5%CD-transduced cells and 95% bystander cells. Cells were exposed lo5-FC for 48 h and were then processed to determine the clonogenicfraction of the CD-transduced cells (Fig. 2). Data points, means of fourindependent experiments; httrs, SE.

aBI-CD

0.0010.2 0.4 0.6 0.8

[5-FC] (mM)

0.00011 0 0.2 0.4 0.6 0.8 1

[5-FC] (mM)

Although the chief purpose of this and our previous investigation(II) was to determine the role of GJIC in the bystander effect of genetherapy, it seemed of interest to begin to assess other potentialmechanisms. Because it has been proposed that the transfer of apop-totic vesicles could mediate bystander killing (21-23), we decided todetermine whether 5-FC produced substantial apoptotic cell killing

under the conditions used for our coculture experiments describedabove. We found that neither WB-CD nor aBI-CD cells showedsignificant cell death from drug-induced apoptosis, assessed as described in "Materials and Methods," after a 48-h treatment with 1 mM

5-FC (Table 1). Parallel experiments were carried out with MCF-7

breast cancer cells as a positive control for our ability to detectapoptosis because these cells have been shown to undergo apoptosisafter a variety of insults (24). In addition, we found that WB-CD andaBI-CD cells were capable of undergoing apoptosis after etoposide

treatment (80 jag/ml for 1 h). These data suggest that, althoughWB-CD and aB 1-CD cells are capable of undergoing apoptosis, theydo not do so to a significant extent after exposure to up to 1 min 5-FC

for =£48h.

DISCUSSION

In this study, we have used communication-competent cells and acommunication-incompetent subclone transduced with CD and anti

biotic resistance genes to assess directly the role of GJIC on the killingof both the transduced and bystander cells. We have found that theCD/5-FC gene therapy strategy expresses little or no dependence on

GJIC, suggesting that bystander killing is determined chiefly bydiffusion of 5-FU into the extracellular space. This result contrasts

with our previous findings derived from analogous experiments usingthe HSV-TK/GCV approach, in which the capacity for GJIC produced

both significant bystander cell killing and protection of the transduced

cell (11). Because most cancer cells lack functional GJIC, our resultssuggest a potential advantage of the CD/5-FC system compared to theHSV-TK approach. However, we have also found that, when trans

duced cells and bystander cells are cocultured at a 5:95 ratio, which isrelevant for in vivo gene therapy, the CD-transduced cells are killed to

a significantly greater extent than the bystander cells for the sameextracellular concentration of 5-FU. This suggests that a potentialdisadvantage of the current CD/5-FC strategy is that the CD-trans

duced cell may be preferentially killed before the bystander tumorcell, thus decreasing the bystander effect and permitting tumor re-

growth.Whereas a number of investigators have evaluated the role of GJIC

in the HSV-TK/GCV strategy (10-11, 25), we are unaware of otherstudies of the influence of GJIC on the efficacy of the CD/5-FC

approach. Our results are consistent with those of Huber and colleagues (6), who found significant regressions of tumors composed ofmixtures of CD-transduced and bystander WiDR cells, which do not

contain gap junctions when evaluated by electron microscopy. It is,therefore, possible that at least a component of the superior resultsthey obtained using the CD/5-FC system compared to the HSV-TK/

GCV strategy was due to the decreased effectiveness of the latterstrategy in tumors that lack the capacity for GJIC.

In this study, we found that increasing concentrations of secreted5-FU produced greater bystander cytotoxicity. In addition, it is likely

that additional factors will influence the bystander effect. For instance, it is possible that the greater diffusion of 5-FU compared to

phosphorylated GCV could actually decrease the selectivity of tumorcell killing. In addition, membrane-impermeant toxic metabolites suchas phosphorylated GCV or FdUMP (in the case of the HSV-TK/GCVand CD/5-FC strategies, respectively) could be transferred from one

cell to another by a mechanism other than GJIC, such as the diffusion

Fig. 8. Preferential killing of CD-transduced cells compared to

bystander cells. This figure shows the data from Figs. 6 and 7 for theratio of 5% CD-transduced cells to 95% bystander cells treated witheither 0.3 (A) or 1 (B) mM 5-FC. The surviving fraction of bystandercells (open symbols'. Fig. 6) and of CD-transduced cells (closedsymbols: Fig. 7) resulting from 5-FC treatment is plotted as a function of 5-FU concentration in the medium of that coculture at thetime of processing for clonogenic survival. Squares. WB; circles,aBl. Data points, means of four or five independent experiments;bars. SE. 0.007

123[5-FU] foM)

0.00014 o 1 2 3

[5-FU] (\iM)

2592

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PREFERENTIAL CYTOTOXICITY OF CD-TRANSDUCED CELLS

Table 1 Induction of apopiosis by 5-FC and etoposide

% apoptotic cells induced

CelllineMCF-7WB-CDaBl-CD5-FC3

±12±10±

1Etoposide48

±127±822

of apoptotic vesicles (21-23). Although neither WB-CD nor aBl-CDcells undergo significant 5-FC-induced apoptosis under the conditions

used in this study, it remains possible that transfer of apoptoticvesicles contributes to the bystander effect in other settings. Furthermore, a bystander effect in vivo could be mediated by nonpharmaco-

logical means, such as stimulation of an immunological response (21,26, 27), which were not examined in this cell culture system. Likewise, the finding that CD-transduced cells are partially protected when

they are cocultured at a 5:95 ratio regardless of communication statuscould reflect 5-FU catabolism by neighboring cells. Therefore, it will

be important to perform further studies of responses in vivo beforedrawing conclusions about which enzyme prodrug strategy is superior.

Perhaps the most important finding of this paper is that CD-

transduced cells are preferentially killed compared to bystander cells.This result is perhaps even more striking, given that WB-CD cells arerelatively resistant to 5-FU compared to the other cell lines used in

this study. It is highly likely that preferential cytotoxicity of transduced cells occurs because they generate high intracellular concentrations of 5-FU, which then becomes diluted before reaching by

stander cells. This result is consistent with the suggestion that CD/5-FC can be used under some conditions as a negative selection

system for selectively removing CD containing 3T3 cells from amixed culture with minimal killing of bystander cells (28). Thus,although successful gene therapy requires that the transduced cellsultimately die, if this occurs prematurely, the great majority of the(nontransduced) tumor cells will not be killed. We have begun toaddress this potential weakness by designing a CD that is secreted.Our preliminary in vitro data suggest that expression of secreted CDproduces more uniform treatment of transduced and bystander cellsthan the expression of intracellular CD used here. Whether the expression of extracellular CD will prove to be superior in vivo to thecurrent approaches using intracellular CD or HSV-TK remains to be

determined.

ACKNOWLEDGMENTS

We thank Dr. Mary Davis and Amy Verhelst for help in preparing thefigures and for reviewing this manuscript and Marlene Langley and BarbaraDavis for secretarial assistance.

REFERENCES

1. Connors. T. A. The choice of prodrugs for gene directed enzyme prodrug therapy ofcancer. Gene Ther., 2: 702-709. 1995.

2. Moolten, F. L. Drug sensitivity ("suicide") genes for selective cancer chemotherapy.

Cancer Gene Ther.. /: 279-287. 1994.3. Deonarain, M. P.. Spooner, R. A., and Epenetos, A. A. Genetic delivery of enzymes

for cancer therapy. Gene Ther., 2: 235-244, 1995.

4. KM, M. S., Kim, J. H., Mullen, C. A., Kim, S. H., and Freytag, S. O. Radiosensi-tization by 5-fluorocytosine of human coiorectal carcinoma cells in culture transducedwith cytosine deaminase gene. Clin. Cancer Res., 2: 53-57, 1996.

5. Hoganson. D. K.. Batra, R. K., Olsen. J. C., and Boucher. R. C. Comparison of theeffects of three different toxin genes and their levels of expression on cell growth andbystander effect in lung adenocarcinoma. Cancer Res., 56: 1315-1323, 1996.

6. Trinh, Q. T., Austin, E. A.. Murray, D. M.. Knick, V. C., and Huber, B. E.Enzyme/prodnig gene therapy: Comparison of cytosine deaminase/5-fluorocytosineversus thymidine kinase/ganciclovir enzyme/prodrug systems in a human coloréela!carcinoma cell line. Cancer Res., 55: 4808-4812, 1995.

7. Hooper, M. L., and Subak-Sharpe, J. H. Metabolic cooperation between cells. Int.Rev. Cytol., 69: 45-104, 1981.

8. Trosko, J. W., Madhukar, B. V., and Chang, C. C. Endogenous and exogenousmodulation of gap junctional intercellular communication: toxicological and pharmacological implications. Life Sci., 53: 1-19, 1993.

9. Pitts, J. D. Cancer gene therapy: a bystander effect using the gap junctional pathway.Mol. Carcinog., //: 127-130, 1994.

10. Pick, J., Barker, F. G., Dazin, P., Westphale, E. M., Beyer, E. C., and Israel, M. A.The extent of heterocellular communication mediated by gap junctions is predictiveof bystander tumor cytotoxicity in vitro. Proc. Nail. Acad. Sci., 92: 11071-11075,

1995.11. Wygoda, M. R., Wilson, M., Davis, M. A., Trosko, J. E., Rehemtulla, A., and

Lawrence, T. S. Protection of HSV-TK-transduced cells from ganciclovir-mediatedcytotoxicity by bystander cells: the Good Samaritan effect. Cancer Res., 57: 1699-

1703. 1997.12. Holder, J. W., Elmore, E., and Barren, J. C. Gap junction function and cancer. Cancer

Res., 53: 3475-3485, 1993.

13. Ram, Z., Culver, K. W., Walbridge, S., Blaese, R. M., and Oldfield, E. H. In situretroviral-mediated gene transfer for the treatment of brain tumors in rats. CancerRes., 53: 83-88, 1993.

14. Loewenslein, W. R., and Rose, B. The cell-cell channel in the control of growth.Semin. Cell Biol., 3: 59-79, 1992.

15. Hotz-Wagenblatt, A., and Shalloway, D. Gap junclional communication and neoplas-tic transformation. Oncogenesis, 4: 541-558, 1993.

16. Oh, S. Y., Dupont, E., Madhukar, B. V., Briand, J. P.. Chang, C. C., Beyer, E.. andTrosko, J. E. Characterization of gap junctional communication-deficient mutants ofa rat liver epithelial cell line. Eur. J. Cell Biol., 60: 250-255, 1993.

17. Austin, E. A., and Huber, B. E. A first step in the development of gene therapy forcoiorectal carcinoma: cloning, sequencing, and expression of Escherichia coli cytosine deaminase. Mol. Pharmacol., 43: 380-387, 1993.

18. Kalimi, G. H., Hampton, L., Trosko, J., Thorgeirsson, S., and Huggett, A. Homologous and heterologous gap-junctional intercellular communication in v-ra/-, v-mvc-,and v-ra/i'v-mvr-transduced rat liver epithelial cell lines. Mol Carcinog., 5: 301-310,

1992.19. Loch-Caruso, R., Pahl, M. S., and Juberg, D. J. Rat myometrial smooth muscle cells

show high levels of gap junctional communication under a variety of culture condi-lions. In Vitro Cell. Dev. Biol., 28A: 97-101, 1992.

20. Lawrence, T. S., Davis, M. A.. Maybaum, J.. Mukhopadhay, S. K., Stetson, P. L.,Normolle, D. P., McKeever, P. E-, and Ensminger, W. D. The potenlial superiorily ofbromodeoxyuridine to iododeoxyuridine as a radiation sensitizer in the treatment ofcoiorectal cancer. Cancer Res., 52: 3698-3704. 1992.

21. Freeman, S. M.. Abboud, C. N., Whartenby, K. A.. Packman, C. H., Koeplin, D. S.,Moollen, F. L., and Abraham. G. N. The "bystander effect": tumor regression when

a fraction of the tumor mass is genetically modified. Cancer Res., 53: 5274-5283,

1993.22. Samejima, Y., and Meruelo, D. "Bystander killing" induces apoptosis and is inhibited

by forskolin. Gene Ther., 2: 50-58, 1995.

23. Colombo, B. M., Benedetti, S., Otlolenghi, S., Mora, M., Pollo, B., Poli, G., andFinocchiaro, G. The "bystander effect": association of U-87 cell death with ganci

clovir-mediated apoptosis of nearby cells and lack of effect in athymic mice. Hum.Gene Ther., 6: 763-772, 1995.

24. Beider, D. R., Tewari, M., Friesen, P. D., Poirier, G., and Dixit, V. M. Thebaculovirus p35 protein inhibits Fas- and tumor necrosis factor-induced apoptosis.J. Biol. Chem., 270: 16526-16528, 1995.

25. Bi, W. L., Parysek, L. M., Wamick, R., and Slambrook, P. J. In vitro evidence lhalmelabolic cooperation is responsible for the bystander effect observed with HSV tkretroviral gene therapy. Hum. Gene Ther., 4: 725-731, 1993.

26. Gagandeep, S., Brew, R., Green, B., Christmas, S. E., Klatzmann. D., Poston, G. J..and Kinsella, A. R. Prodrug-activated gene therapy: involvement of an immunological component in the "bystander effect." Cancer Gene Ther., J: 83-88, 1996.

27. Caruso, M., Pañis,Y., Gagandeep, S., Houssin, D., Salzmann. J. L., and Klatzmann.D. Regression of established macroscopic liver métastasesafter in situ transduction ofa suicide gene. Proc. Nati. Acad. Sci. USA, 90: 7024-7028, 1993.

28. Mullen, C. A.. Kilstrup, M.. and Blaese, R. M. Transfer of the bacterial gene forcytosine deaminase to mammalian cells confers lethal sensitivity to 5-fluorocytosine:a negative selection system. Proc. Nati. Acad. Sci. USA. 89: 33-37, 1992.

2593

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1998;58:2588-2593. Cancer Res   Theodore S. Lawrence, Alnawaz Rehemtulla, Emily Y. Ng, et al.   5-FlucytosineDeaminase Compared to Bystander Cells after Treatment with Preferential Cytotoxicity of Cells Transduced with Cytosine

  Updated version

  http://cancerres.aacrjournals.org/content/58/12/2588

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

on March 3, 2015. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from