mechanisms for ganciclovir resistance in gastrointestinal tumor … · vol. 4, 731-741, marc/i 1998...
TRANSCRIPT
Vol. 4, 731-741, Marc/i 1998 Clinical Cancer Research 731
Mechanisms for Ganciclovir Resistance in Gastrointestinal Tumor
Cells Transduced with a Retroviral Vector Containing the
Herpes Simplex Virus Thymidine Kinase Gene
Lily Yang, Rosa Hwang, Yawen Chiang,
Erlinda M. Gordon, W. French Anderson, and
Dilip Parekh1
Gene Therapy Laboratories, Los Angeles. California [L. Y.. R. H..E. M. G.. W. F. A.!: Department of Surgery. University of Southern
California, Los Angeles. California 90033 [D. P.1: and GeneticTherapy Inc.. Gaithersburg. Maryland 20878 [Y. C.l
ABSTRACT
Transfer of the herpes simplex thymidine kinase (HSV-
TK) gene into tumor cells confers sensitivity to the cells to
the viral drug ganciclovir (GCV). Although the efficacy of
the HSV-TK/GCV approach is well studied, the mechanisms
for the resistance of HSV-TK-transduced tumor cells toGCV are poorly understood. Here, we examined the mech-anisms for GCV resistance in HSV-TK-transduced gastro-
intestinal (GI) cell lines. Our results show that GCV sensi-
tivities vary in vitro and in vivo among the different HSV-TK-transduced GI tumor cell lines. GCV-resistant colonieswere isolated from several different HSV-TK-transduced GI
tumor cell lines after 14 days of GCV treatment.Characterization of GCV-resistant colonies demon-
strated that the HSV-TK gene was either partially or com-
pletely deleted from the resistant HSV-TK-transduced cells.
In the HT.29 RM and MIAPACA-2 RM cells, a 220-bpdeletion of the gene was found, whereas in the HT-29 Ri-
R5-resistant cells, the whole TK gene was found to be absent.
Immunocytochemical studies using a polyclonal antibody to
the TK protein demonstrated that the HSV-TK protein wasabsent in the GCV-resistant, HSV-TK-transduced cells.
Transfection of the resistant cells with an adenoviral vector
containing a HSV-TK gene restored sensitivity to GCV. The
presence of GCV-resistant cells was only demonstrable in GItumor cell lines that also demonstrated a poor bystander
effect. Our results suggest that GCV resistance found in
tumor cells transduced with a retroviral HSV-TK gene is due
to the lack of a functional 1K protein in the tumor cells
rather than any intrinsic resistance of the cells to GCV. Intumor cells with a good bystander effect, the small percent-
age of TK-transduced cells that do not express the TK
protein are probably killed by the bystander effect because
GCV-resistant tumor cells were not found in these cell lines.
GCV-resistant tumor cells were found only in tumor cell
lines with a poor bystander effect, by which, presumably, the
transduced tumor cells lacking a functional TK gene were
not killed by the bystander killing effect.
INTRODUCTION
The transfer of suicide genes such as the HSV-TK2 gene
( 1-5) into tumor cells renders the cells sensitive to nontoxic
prodrugs such as GCV. The HSV-TK/GCV approach for the
treatment of cancer is under clinical investigation in at least 1 1
trials (6). The expressed HSV-TK protein in transduced cells
phosphorylates the prodrug GCV into monophosphorylated
GCV, which is further converted into triphosphorybated GCV by
cellular kinases. Triphosphorylated GCV, when incorporated
into DNA, induces cell death. In mixed cocultures, experiments
HSV-TK-transdueed cells (TK�), as well as surrounding non-
transduced cells (TK�), are killed by GCV treatment (7). This
“bystander effect,” by which adjacent TK cells are killed
during killing of TK4 cells, is thought to be mediated by
transfer of toxic phosphorylated GCV from transduced tumor
cells by intercellular gap junction communication to nontrans-
duced tumor cells (8. 9). The bystander effect is essential for
successful cancer gene therapy because the presence of this
effect makes complete regression of the tumor possible when
only a fraction of the tumor cells in a tumor mass are transduced
(7).
The objective of cancer therapy is to eliminate all tumor
cells in a tumor mass because experience with chemotherapy
has shown that surviving tumor cells grow out to produce
recurrent tumors. In in vitro and in vito experiments, resistance
of HSV-TK-transduced tumor cells to GCV has been described
(10. 1 1). Development of GCV resistance in HSV-TK-trans-
duced cells is of concern because this may limit the efficacy of
the HSV-TK gene transfer approach for the treatment of cancer.
The mechanisms for resistance of HSV-TK-transdueed cells to
GCV is unclear. Here, we investigated the mechanisms of GCV
resistance in HSV-TK-transdueed human 0! tumor cell lines.
Our results show that sensitivity to GCV treatment differed
among the different GI tumor cell lines that are transduced with
a retroviral HSV-TK gene. We found that TK� cell populations
that are resistant to GCV have either a partial or complete
Received 8/8/97: accepted 10/10/97.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertiseme,it in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
I To whom requests for reprints should be addressed, at Department of
Surgery. University of Southern California, 15 10 San Pablo Street, Suite
514, Los Angeles. CA 90033.
2 The abbreviations used are: HSV-TK, herpes simplex virus thymidine
kinase: GCV. ganeiclovir: Gl. gastrointestinal; RT-PCR. reverse
transcriptase-PCR: MIS. (3-4.5-dinuethylthiazol-z-yl)-5-(3-carboxy
methoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium. inner salt; CMV.
cytomegalovirus.
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
732 GCV Resistance and HSV-TK
absence of the HSV-TK gene. In tumor cells with a good
bystander effect. the small percentage of transduced cells that
have lost the HSV-TK gene are killed by the bystander effect.
GCV-resistant tumor cells were found only in tumor cell lines
with poor bystander effects, by which, presumably, the trans-
dueed tumor cells lacking a functional TK protein were not
killed by the bystander killing effect.
MATERIALS AND METHODS
Cell Lines and Vectors. BXPC-3, CAPAN-2,M!APACA-2, PANC-l, PANC-3, and PANC-28 human pan-
creatic cancer cell lines, HT-29 colon cancer cells, and N87
gastric cancer cells were obtained from the American Type
Culture Collection (Roekville, MD). Colon cancer cells KM12C
and KM12 SM were kindly provided by Dr. Fidler from M. D.
Anderson Hospital (Houston, TX). We used retroviral vectors
carrying the HSV-TK gene (GlTklSvNa.7) and a Escherichia
co/i 3-ga/actosidase gene (GlBgSvNa). The GlTklSvNa.7 and
G 1BgSvNa vectors also carry a marker NeoR gene as a positive
selectable marker gene. The producer cell lines for the vectors
were obtained from Genetic Therapy Inc. (Gaithersburg, MD).
The producer cell lines GlTklSvNa.71PA317 and GlBgSvNa/
PA317 were cultured in DMEM with 10% fetal bovine serum,
and the supernatant containing the viral vectors was collected
and frozen in aliquots at -70#{176}C.
The adenoviral vectors Av lTk 1 and Av 1LacZ4 were ob-
tamed from Genetic Therapy Inc. (Gaithersburg, MD). AvlTkl
vector carries the HSV-TK gene, whereas the AvlLaeZ4 carries
the marker LacZ gene.
Transduction of GI Tumor Cell Lines with RetroviralVector GiTkiSvNa.7. GI tumor cells (5 X l0� cells) were
plated in 25-mm2 tissue culture flasks for 24 h. The culture
medium was then replaced with 3 ml ofGlTklSvNa.7 retroviral
vector supernatant with 8 p.g/ml polybrene for 4 h. Forty-eight
h following transduction, a neomycin analogue, G418 (0.3-2
mg/mb). was added to the culture medium to select for trans-
duced BXPC-3, PANC-3, HT-29, and MIAPACA-2 cells. The
GlTklSvNa retroviral vectors used in our studies all carry a
neomycin resistance (NeoR) selectable marker gene; therefore, a
selected population of HSV-TK-transduced cells is established
after selection in G4l8. The cells were selected for 14 days in
G418, and the surviving colonies were subeultured and estab-
lished as permanent HSV-TK-transduced clones (TK� cells).
In Vitro and in Vivo Effects of GCV on HSV-TK�
Tumor Cells. GCV killing curves were determined for both
TK� and TK4 tumor cell lines by culturing the cells in different
concentrations of GCV (Cytovene; Syntex Laboratories, Palo
Alto, CA) for 12-14 days in 6- or 12-well tissue culture plates.
The percentage of cells killed by GCV were estimated by
examining the surviving cell populations under an inverted
Nikon microscope after the resistant colonies were stained with
a 1% methylene blue solution.
To examine for the in x’ivo effect of GCV on TK� GI tumor
cells, BXPC-3 cells (1 X l0�), or HT-29 cells (5 X 106) were
injected s.c. into nude mice (Harlan Sprague-Dawley, Inc. In-
dianapolis, IN). Eleven to 13 days after tumor implantation, the
mice received i.p. injection of GCV (100 mg/kg/day) for 14
days. The size of the tumors was measured every 3 days. and
tumor volumes were calculated.
Establishment of GCV-resistant Mixed and Clonal CellPopulations (TK+/GR) from HT-29 and MIAPACA-2 TK�
Cells. To investigate the mechanisms of GCV resistance in
HSV-TK-transdueed tumor cells, GCV-resistant tumor cell eol-
onies (TK±/G�c) were subcultured out, and GCV-resistant cell
clones were established. MIAPACA-2 and HT-29 TK� cells
were cultured in a medium containing 50 pUg/mb GCV for 3
weeks. The surviving GCV-resistant colonies were harvested
with trypsin-EDTA, and the cells from several different surviv-
ing colonies were collected together and repeatedly passaged to
establish a GCV-resistant cell population (HT-29 RM and MI-
APACA-2 RM cells).
To establish pure ebonal populations of GCV-resistant
TK�eells, HT-29 TK� cells were cultured in a medium con-
taming 90 p.g/ml GCV for 4 weeks, and the surviving HT-29
GCV-resistant colonies were picked up individually and subeul-
tured. Thirty TK� GCV-resistant HT-29 cell clones were estab-
lished, and five resistant clones were randomly selected for
further study (HT-29 Rl-R5 cells).
GCV Killing Curve of Resistant TK� GI Tumor Cells.HT-29 and MIAPACA-2 TK�, TK, and TKf/GR cells (5 X
i0� cells/well) were plated in 12-well tissue cultures plates.
After overnight culture, the culture medium was replaced with
medium containing GCV (0-100 �ig/ml), and the cultures were
treated for 12 days. GCV in the culture medium was changed
daily. At the end of the treatment period, the cells were fixed
with 1% methylene blue solution, and the percentages of sur-
viving cells in each treated well were estimated as a fraction of
untreated control cells (no. of colonies in GCV-treated wells/no.
of colonies in untreated wells X 100).
Detection of HSV-TK Gene and Gene Expression inGCV-resistant Cell Populations by PCR and RT-PCR.Genomic DNA was isolated from TK� , TK�, and TK+/GR cell
populations using an Elu-Quik DNA purification kit (Sehleieher
& Schuell Inc., Keene, NH). The PCR primers used to amplify
TK gene were: TK1, forward primer, 5’-CGTTCTGGCTCCT-
CATGTCG-3’, and reverse primer, 5’-GCCAGCATAGCCAG-
GTCAAG-3’, which amplify a 288-bp region of the TK gene
between 1954 and 2241 bp in the G1Tk1 Sv 1 Na plasmid; and
TK2, forward primer, S’AGCAAGAAGCCACGGAAGTC3’,
and reverse primer, 5’AGGTCGCAGATCGTCGGTAT3’,
which amplify a 990-bp region of HSV-TK gene. The primers
used to amplify a 423-bp Neo’� gene product were: forward
primer, 5’-GGTGGAGAGGCTATFCGGCTATGA-3’, and re-
verse primer, 5’-ATCCTGATCGACAAGACCGGCTFC-3’.
PCR amplification of TK and Neo’� fragments was performed
with a GeneAmp PCR core kit (Perkin-Elmer Corp., Branch-
burg, NJ).
For RT-PCR analysis, total RNA was isolated from the
tumor cells to detect TK and Neo’� gene expression in GCV-
resistant HT-29 cell clones. RNA was isolated using a RNA
STAT-60 kit (Tel-Test, Inc., Friendswood, TX). RT-PCR to
amplify HSV-TK and Neo’� mRNAs was performed using a
Strata script RT-PCR kit according to the manufacturer’s spec-
ifleations (Stratagene, La Jolla, CA). Briefly, 10 �i.g of total
RNA were reverse-transcribed into DNA by using oligo(dT)
primers. Three p.1 of a 50-p.1 eDNA mixture were further
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 733
Table 1 GCV sensitivity of HSV-TK -transduced human GI tumor cellsa
Cell line Type of tumor TK cells, IC50 (p.g/ml) TK� cells, IC50 (�i.g/ml) Fold increase in GCV sensitivity
BXPC-3
PANC-3
tvHAPACA-2
Pancreas
Pancreas
Pancreas
30
150
150
5 X l0�5 X iO�
1
60,00030,000
150HT-29 Colon 100 0.7 143
a This table shows the GCV sensitivity of four representative cell lines among I 1 GI tumor cells examined. IC50 was measured by a MTS cell
proliferation assay after the tumor cells received GCV treatment for 10 days.
amplified with TK or Neo’� primers using Gene Amp PCR core
reagents. All PCR products were visualized by electrophoresis
on a 1% agarose gel with ethidium bromide.
Analysis ofHSV-TK Gene Expression by Northern BlotAnalysis in GCV-resistant Cell Populations. The full-length
transcript of the HSV-TK gene was detected by Northern blot
analysis. Total RNA was isolated using a RNA-STAT 60 kit
(Tel-Test) from TK�,TK, and TK+/GR cells grown to 80%
confluency. Aliquots (20 �i.g) of RNAs were separated by 0.8%
agarose-0.75 M formaldehyde gel electrophoresis and trans-
ferred to nylon transfer membrane (Hybond-N�; Amersham
Life Science Inc., Arlington Heights, IL) by overnight capillary
transfer. A 1189-bp fragment containing the HSV-TK gene was
obtained by cutting the plasmid G1Tk1SvNa (Genetic Therapy
Inc.) with BalII and XhoI. �3-Actin fragments (232 bp) were
obtained from PCR products of DNA of human colon cancer
cell line HT-29 DNA amplified by forward primer 5’-CATT-
GTGATGGACTCCGGAGACGG-3’ and reverse primer 5’-
CATCTCCTGCTCGAAGTCTAGAGC-3’. The fragments
were gel-purified and labeled with [32P]dCTP (Redivue, 3000
Cilmmol; Amersham) with a Megaprime DNA labeling system
(Amersham). Hybridization of the blots was performed at 68#{176}C
with DNA probes in QuickHyb buffer (Stratagene). After sev-
era! washes, a Phosphorlmage system (Molecular Dynamics,
Sunnyvale, CA) was used to detect the positive bands.
Detection of HSV-TK Protein by an Immunocytochem-ical Stain. TK, TK�, and TK±/GR tumor cells (3 X l0�/
well) were plated in Lab-Tek two-well chamber slides (Nunc,
Inc., Naperville, IL) for 72 h. The chambers were then washed
with PBS and fixed with ice-cold acetone for 10 mm. The slides
were air-dried and stained immediately or frozen at -70#{176}Cuntil
use.
Polyclonal rabbit antibody against HSV-TK protein was
kindly obtained from Dr. W. P. Summers at Yale University
(New Haven, CT). To detect the TK protein, the slides were
blocked with 2% normal goat serum for 10 mm and then washed
in PBS. The primary antibody diluted at 1:50 was added onto the
slides for 60 mm. The slides were then washed three times with
PBS, and a secondary antibody goat antirabbit IgG conjugated
with peroxidase (1:100; Vector Laboratories, Burlingame, CA)
was added to the slides for 30 mm. The slides were washed five
times with PBS and developed with a 3,3’-diaminobenzidine
substrate kit (Vector Laboratories). After eounterstaining with
hematoxylin, the slides were examined and photographed under
a Nikon microscope.
Detection of the Bystander Effect in Vitro. Presence of
a bystander effect was examined in vitro using a MIS ealori-
metric cell proliferation assay that measures dehydrogenase
activity in viable cells (Cell Titer-96 aqueous nonradioactive
MTS cell proliferation assay; Promega, Madison, WI). HSV-
TK-transdueed and nontransdueed tumor cells were mixed in
different ratios and cocultured in 96-well tissue culture plates
(3 X i0� cells/well). After a 24-h incubation period, fresh
culture medium with 10 pjg/mi GCV was added to the cells. The
medium was changed to add a fresh lot of GCV each day to the
cells for 14 days. MTS cell proliferation assays were performed
every 3-4 days. Twenty p.1 of assay mix were added to the cell
cultures, and A4� nm was measured on an ELISA reader. The
absorbanee values in this assay reflect viable cell densities.
Treatment of TK+/GR Cells with Adenoviral HSV-TK
Vectors. To confirm our hypothesis that GCV resistance in
TK-transdueed tumor cell lines is due to the loss of production
of the TK protein, we studied the effects of reintroducing the TK
gene into the TK±/GR tumor cells with an adenoviral TK vector
and retreatment of the cells with GCV. An adenoviral vector
was selected for these studies because the TK±/GR resistant
cells were already transduced with a retroviral vector containing
a NeoR gene and, therefore, selection is not necessary after
transducing the cells with an adenoviral vector due to the high
transduction efficiency of the vector.
TK, TK�, and TK±/GR HT-29 and MIAPACA-2 cells
were plated in 96-well tissue culture plates at a cell density of
4 X l0� cells/well. Twenty-four h later, cells were transduced
with either AvlTkl or AvlLaeZ4 vectors at a multiplicity of
infection of 20 for HT-29 and of 200 for MIAPACA-2 for 2 h.
Twenty-four h later, the cells were treated with GCV (10 �i.g/ml)
for 9-10 days. At the end of the treatment period, the surviving
cells in each group were measured by the MTS cell proliferation
assay.
RESULTS
Detection of HSV-TK-transduced Tumor Cells ThatAre Resistant to GCV. GCV treatment effects were studied
in 10 human GI tumor cell lines in vitro by culturing the cells in
35�2 dishes in the presence of logarithmic escalation of
the concentration of GCV. We studied BXPC-3, CAPAN-2,
MIAPACA-2, PANC- 1 , PANC-3, and PANC-28 pancreatic
cancer cells, HT-29, KM12C, and KM12 SM colon cancer cells,
HUTU duodenal cancer cells, and N87 gastric cancer cells. TK�
tumor cells all showed a marked increase in sensitivity to GCV
treatment compared to TK cells (Table 1). However, only
HSV-TK-transduced BXPC-3 and PANC-3 cells were com-
pletely killed by GCV at nontoxie concentrations of GCV (non-
toxic GCV doses were 10-20 �ig/ml). In the other eight cell
lines, TK� cells were not completely killed by GCV, and
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
BXPC-3 TK+-
2.0�
.5.
1.0�
0.5.
-0- TK-
� -a- TK +
�-
EC
0)
Cs
0
2.0’
1.5
E
0) .0Cs‘1�
� 0.5.
0
0.0�
0 � ‘i � � 1� 1� 14
Days after GCV
PANC-3� :
MIAPACA-2 TK+ MIAPACA’2 TK-
4 � )r: �
- .�,-.
1; � ,;�,o � ‘so
0 i 4’ 6’ 8 1#{212} i� i4Days after GCV
2.0 � HT-29
�,ft-_______#{149} -0- TK-
.5 � ,� -0- TK+
iI.0�
0
0.0
2.0
1.5
E� 1.0
Cs
�0.5
0
0.0
. u-I-l-I-I.I- I
0 2 4 6 8 10 12 14
Days after GCV
MIAPACA-2-0- TK-
-0-- TK+
0 2
734 GCV Resistance and HSV-TK
BXPC-3 TK-
,�oI � �
�‘\
Fig. I Emergence of GCV-resistant colonies during treatment of HSV-
TK-transduced GI tumor cells. HSV-TK� and TK BXPC-3, HT-29,
and MIAPACA-2 cells (5 X l0� cells) were cultured in 12-well tissue
culture plates. Culture medium was replaced 24 h later with a medium
containing 0. 5. 10, 30, 50. or 100 �.cg/ml GCV. After 14 days of GCV
treatment, the surviving cells were fixed with a 1% methylene blue
solution. The surviving colonies in each well represent GCV-resistant
cell clones at each GCV concentration.
resistant TK� cell colonies were present after GCV treatment
for 12-15 days (Fig. 1)
Fig. 2 shows the results of GCV treatment of four repre-
sentative TK� cell lines using the MTS cell proliferation assay.
No cytocidal effects were seen on the TK tumor cell lines after
GCV treatment ( 10 p.g/ml) for 12 days. All of the TK� BXPC-3
and PANC-3 cells were killed; however, a small percentage of
Days after GCV
Fig. 2 GCV sensitivity of HSV-TK� and HSV-TK tumor cells. TK�
and TK tumor cells (4 X l0� cells) were cultured in 96-well tissue
culture plates. Culture medium was replaced 24 h after plating with
medium containing 10 p.g/ml GCV, and the cells were treated for 12
days. The surviving cells were assayed by the MTS cell proliferation
assay at days 0, 4, 8. and I 2. The A4� nm values reflect the number of
viable cells in this assay. Similar results were obtained from three
separate experiments (ii = 8 per treatment group).
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
BXPC-3 HT.29
*
,
2.0 3-
.:- 1.5
� 12- __j1.o I� � 1-i�O.5
0.0 0
TK. TK+ TK. TK+
Fig. 3 In vivo treatment of HSV-TK-transdueed BXPC-3 and HT-29
tumors. BXPC-3 (I X l0�) or HT-29 (5 X 106) were injected s.c. into
nude mice (n = 5 per treatment group). TK� tumor cells were injectedon the right flank and TK cells were injected on the left flank in thesame animal. Eleven to 13 days after tumor implantation, all mice
received 100 mg/kg/day GCV i.p. daily for 14 days. Tumor volumes
(em3; length x width X height) in the figure represents the mean tumor
volume on day 5, following the last GCV treatment dose in each group.
HT-29 and MIAPACA-2 cells were still present after 1 2 days of
GCV treatment (Fig. 2). Comparing the absorbanee values of
the cells with TK and TK� cells on day 12 of treatment, about
19.9% of TK� HT-29 and 6.6% of TK� MIAPACA-2 cells
were viable. When compared with the original plating density in
the 96-well culture plates, an estimated 1-6% of the initially
plated HT-29 and MIAPACA-2 TK� tumor cells grew out as
GCV-resistant colonies after 12 days of GCV treatment (Fig. 1).
Similar results were observed in vivo after s.c. implantation
of TK� tumor cells in nude mice (Fig. 3). After being treated
with i.p. injections of GCV (100 mg/kg daily) for 14 days,
tumor-bearing mice had a significant reduction in the size of
their tumors derived from BXPC-3 TK� cells, compared to
control TK cells (P < 0.01, Mann-Whitney, n = 5). On the
other hand, the size of the treated HT-29 TK� tumors was not
different from that of TK controls (P > 0.05, Mann-Whitney,
n 4 per group). These results suggest that resistance to GCV
is observed in vivo in TK� tumor cells, which are also resistant
to GCV in vitro.
GCV Killing Curve of TK�/G’� Cells. The sensitivity
of parent TK� HT-29 and MIAPACA-2 cells is demonstrated
by the killing of95-99% ofthe cells by GCV (Fig. 4A). One and
5% of TK� HT-29 and MIAPACA-2 cells, respectively, were
resistant to GCV at concentrations of up to 100 p.g/ml (Fig. 4A).
The GCV killing curves of the resistant TK+/GR cell lines
HT-29 Ri, HT-29RM, and MIAPACA-2 RM cells were similar
to those of the parent TK HT-29 and MIAPACA-2 cells,
suggesting that the TK±/GR cells have lost their sensitivity to
GCV (Fig. 4).
Integration of the HSV-TK Gene in TK� Parent and
TK+/GR Cell Lines. To understand the mechanisms for GCV
resistance in TK±/GR tumor cells, we investigated the integra-
tion of the HSV-TK gene in the transduced tumor cells.
Presence of the HSV-TK gene was investigated in the TK�
parent cells and in the TK±/GR cell lines HT-29 RM and MI-
APACA-2 RM and the HT-29-resistant clones Rl-R5. The
HSV-TK gene was amplified by PCR using TK1 and TK2 primers
(Fig. SA). We first analyzed the genomie DNA from the tumor cells
by PCR amplification with TK1 primers that amplified a 288-bp
fragment within the TK gene (Fig. 5A). A PCR product was not
detected when the 288-bp fragment inside the HSV-TK gene was
amplified with the TK1 primers in some of the TK±/GR cell lines,
suggesting that the fragment of the TK gene that annealed with the
TK1 primer sequence was deleted. Because the forward TK 1
primer is between 410 and 429 bp and the reverse TK1 primer is
between 678 and 698 bp of the TK gene, the deletion may be
located at either of the primer sequence regions (Fig. SA). The
HSV-TK gene in the GlTklSvNa vector has a splice site located at
the 328 bp. To examine whether the absence of PCR products after
amplification with the TK1 primers was due to the presence of the
spliced TK gene or whether the whole gene was lost from the cells,
we also analyzed genomie DNA by PCR amplification with the
TK2 primer sequences, which amplified 990 bp of the HSV-TK
gene (Fig. 5A).
In the parent HT-29 and MIAPACA-2 TK� cell popuba-
tions, both the 288- and the 990-bp PCR products were ampli-
fled with TK1 and TK2 primer pairs (Fig. SB), demonstrating
integration of the HSV-TK retroviral vector in these cell lines.
In contrast, in the HT-29 RM cells, both the 288- and 990-bp
PCR products were absent; however, a 770-bp PCR product was
found that amplified with the TK2 primer (Fig. SB). In
MIAPACA-2 RM cells, the 288- and 990-bp PCR products were
found; however, a much stronger 770-bp band that amplified
with the TK2 primers was also found (Fig. SB). These results
suggest that a 220-bp fragment of the HSV-TK gene is deleted in
all or majority of the HT-29 RM and MIAPACA-2 RM cells.
In the HT-29-resistant clones Rl-R5, no PCR products
were detected after amplification of genomie DNA with both
TK1 and TK2 primers (Fig. SC). On the other hand, a 423-bp
product that amplified with primers specific for the NeoR gene
was present in all five HT-29-resistant cell clones (Fig. SC).
These results suggest that the HT-29 Rl-R5 cells were trans-
duced with the HSV-TK retroviral vector because the NeaR gene
that conferred the cells’ resistance to G418 was present in the
genomie DNA. Absence of PCR products that amplified with
TK1 and TK2 primers suggest that the TK gene is selectively
lost from the HT-29 Rl-R5 cells.
Expression of the HSV-TK Gene in TK� Parent and
TK+/GR Cell Lines. On Northern blot analysis, TK gene
transcripts were not detected in the TKeells but were detected
in the TK� HT-29 cells and MIAPACA-2 cells. In the HT-29
RM and MIAPACA-2 RM cells, the HSV-TK transcripts were
overexpressed compared to the parent TK� MIAPACA-2 and
HT-29 cells. HT-29 resistant clones Rb and R3 did not express
the 4-kb TK mRNA transcripts (Fig. 6A). Because TK tran-
scripts were not detected in HT-29 Rl-R5 cells, we examined
for expression of Neo�c transcripts in these cells by RT-PCR
(Fig. 6B). Expression of the NeoR gene was demonstrable in
four of the five HT-29-resistant clones (Rl-R3 and RS); how-
ever, the HSV-TK gene was not expressed in any of the five
HT-29-resistant clones on RT-PCR, confirming our findings on
Northern blot analysis. Both the NeoR and the HSV-TK genes
were expressed in the parent HT-29 TK� and MIAPACA TK�
cell lines and were not expressed in the control TK cells (Fig.
6B). Our findings of loss of TK gene expression in the HT-29
Rl-R5 cells but presence of the Neo�c gene expression. taken
together with our findings from analysis of genomie DNA
Clinical Cancer Research 735
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
BXPC-3
(ug/mI)
-0--- TK+
.� TK-
....o....
----a---.
A
� 125 1
:� I -ci--- TK+.� l001Q0
j 75. \\ � TK-
I � � ;� �
GCV concentrationHT-29
� 25
� 00
I�CCV concentration (ug/mI)
MIAPACA-2- 25
� 00
I:J25
B
(.9
B c-I
GCV concentration (ug/ml)
HT.29 RM Jfl’.29 Ri MIAPACA.2 RM
,c-.-�- #{149}�
736 GCV Resistance and HSV-TK
(described above), demonstrate that the HT-29 Rl-R5 cells
were transduced by the HSV-TK retroviral vector; however, the
HSV-TK transgene is absent from these cells.
Detection of the HSV-TK Protein in TK� and TK+/GRGI Tumor Cell Populations. HT-29 TK� and MIAPACA-2
TK� cells were treated with 50 pjg/mi GCV for 12 days in
two-well chamber slides, the surviving GCV-resistant cells were
fixed and stained for the HSV-TK protein using a polyclonal
antibody to the HSV-TK protein. Fig. 7A demonstrates that
almost all of the GCV-resistant MIAPACA-2 and HT-29 cells
do not express the HSV-TK protein because they stained neg-
ative with the HSV-TK antibody. The absence of the TK protein
in the GCV-resistant cell colonies suggests that GCV treatment
of TK� HT-29 and MIAPACA cells selects out a resistant cell
population that does not express the HSV-TK protein. Control
untreated HT-29 TK� and MIAPACA-2 TK� cells strongly
stained for the HSV-TK protein (Fig. 7A).
We also stained the TK±/GR cell lines HT-29 RM,
MIAPACA-2 RM, and HT-29 Rl-RS for the presence of theRM HSV-TK protein and found that the HSV-TK protein was absent
in the TK+/GR cell lines (Fig. 7A), providing further evidence
RI that these cells are resistant to GCV due to loss of the TK
protein in the cells. We stained untreated TK� BXPC-3 cells
with the TK antibody and found that about 10-20% of the TK�
BXPC-3 cell population did not stain or demonstrated a very
weak stain with the TK antibody (Fig. 7B). Presumably, this is
a population of BXPC-3 cells that does not express the TK
protein and may potentially give rise to GCV-resistant cells.
However, GCV-resistant cells were not detected during treat-
-a-- TK+ ment of BXPC-3 cells with GCV, suggesting that TK±/GRBXPC-3 cells are killed during GCV treatment.
�G 1K- Killing of GCV-resistant Cells (TK�/G’� Cells) by
Adenoviral Vectors Carrying a HSV-TK Gene. About 86-
....o,... RM 100% of HT-29, HT-29 RM, and HT-29 Ri cells transduced with
a control AvlLaeZ4 vector were viable after GCV treatment. In
contrast, only 22-32% of the HT-29 parent TK cells and the
HT-29 RM and HT-29 Rl cells transduced with an AvlTk vector
were alive following 9 days of GCV treatment (Fig. 8A). Similar
results was obtained with MIAPACA-2 cells (Fig. 8B). Eighty % of
the control MIAPACA-2 RM cells that were transduced with the
AvlLacZ4 vector survived after treatment with GCV. In contrast,
only 4% of MIAPACA-2 TK and MIAPACA-2 RM cells trans-
duced by the AvlTkl vector were viable after GCV treatment (all
AvlTkl versus AvlLacZ4 groups: P < 0.0001, n = 8 per group).
There were no differences in the sensitivity of TK and �fl(±/GR
cells to GCV in any of the groups (P > 0.05). These results
demonstrate that fl(±/GR cell lines (HT-29 RM and R4 and
MIAPACA-2 RM) are sensitive to GCV if a functional HSV-TK
protein is expressed by these cells.
Bystander Killing Effect in GI Tumor Cells. The de-
gree of the bystander effects varied among the different GI
Fig. 4 Sensitivity of TK±/GR cells to GCV. BXPC-3, HT-29 andMIAPACA-2 TK�, TK, or TK±/GR (�Jy� and HT-29 Ri) cells wereplated in 12-well tissue culture plates (5 X i0� cells/well), and the cell density of control wells as 100%. B, GCV killing curve of TK±/GR
cultures were treated with GCV (0-100 p.g/ml for 12 days). The plates cells. The surviving cell clones in each well represent GCV-resistantwere fixed with 1% methylene blue solution at the end of the treatment. cells after 12 days of GCV treatment. GCV-resistant cell lines
A, percentages of cells in each group surviving 12 days of GCV HT29-RM and MIAPACA-2 RM cells and HT-29 Ri cells demonstrate
treatment. The percentages of surviving cells were estimated using similar killing curves to that of the parent HSV-TK cell lines.
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
TK1
603 bp
310 bp�
1018 bp872 bp -
603 bp -
TK2
- 288 bp
990 bp770 bp
CTK1 Neo
1 kb �500 bp - �. �
II1288bp.� #{149}�.
‘I
#{149}0
.‘�
z
TK2
Clinical Cancer Research 737
A
B
LTR TK SV Neo LTR
‘4’ 1P’��::TK1 primer pair: -.� b��-
TK2 primer pair: _� 990bp
Fig. 5 PCR amplification of genomic DNA for
the HSV-TK gene in TK� and TK_/GR cell lines.
A, two PCR primer pairs (TKI and TK2) wereused to amplify the HSV-TK gene. The TK1
primer pair amplified to a 288-bp fragment within
the TK gene because the forward TK primer is
between 410 and 429 bp and the reverse primer isbetween 678 and 698 bp. The TK2 primer ampli-
fies to the 990 bp of the TK gene because the
forward primer is between 100 and 120 bp and the
reverse primer is between 1052 and 1072 bp. B,tumor DNA was isolated from TK, TK�, andTK�/G’� cells and amplified with TKI and TK2
primers. In the parent HT-29 and MIAPACA-2
TK� cell populations, both 288- and 990-bp PCR
products were amplified with TK1 and TK2
primer pairs. !n the HT-29 RM cells, both the 288-
and 990-bp PCR products were absent; however, a
770-bp PCR product was found that amplifiedwith the TK2 primer. In M!APACA-2 RM cells,
the 288- and 990-bp PCR products were found;
however, a much stronger 770-bp band that am-plified with the TK2 primers was also found. C,the HSV-TK gene was not detected in any of the
five HT-29-resistant cell clones when the genomieDNA was amplified with both TK1 and TK2 prim-
ers (top and bottom, respectively). A 423-bp Neo’�
gene PCR product was detected when genomieDNA was amplified with primers specific for theNeoR gene (top), suggesting that these cell clones
were transduced with the HSV-TK retroviral
vector.
xi � EIa $�
‘C #{149}0‘C ‘e �.� .� Cs 0) Cs � e� e� �
e� � N ,� ,� � N
� � � � � c) LI L) �� � � .� .� .ej� z d
�
-
c�
1018 bp -� : ‘
872 bp -� #{149}� .
I., � � e inI� Ia#{149}0 ‘0 - �#{149}0 ‘C � Cs Cs Os 0) Cs
.� .! Cs Cs Cs e� ei ei e� NN N S S I I I
.,t� � , � � � � �� � I- 1- �-
z z �-� � �
:._990 bp
� �77O bp
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
TK1A
2kb- � - �-actin4kb- - TK
�� � � Cs Cs CS
N� �? �N N
�c_)�_ �t.) Z � � �..� .‘��
B
1 kb500bp - -
1 kb -
5(H) bp
Neo
I. � � I #{149}�
�Cl-�l-
�
.,1� � �
�� a.
Fig. 6 Expression of the HSV-TK gene detected by Northern blot analysis and RT-PCR in TK� and TK±/GR cell lines. A, Northern blot analysis
demonstrating expression of the 4-kb HSV-TK transcripts in HSV-TK�. HT-29 RM. and MIAPACA-2 RM cells. HSV-TK transcripts were not
detected in the HT-29-resistant cell clones (HT-29 Rl and HT-29 R3). �3-Actin was used as an internal loading control. B. RT-PCR products
electrophoresed on a 1% agarose gel stained with ethidium bromide. PCR amplification products for HSV-TK and NeaR transcripts were 288 and 423
bp. respectively. The gel demonstrates that the TK and NeaR transcripts are detected in the parent BXPC-3. HT-29, and PANC-3 TK* cell lines and
the TK producer cell line (positive control) and are negative in the TK cell lines. In the HT-29-resistant clones (Rl-R5). the TK transcripts are absent
but the Neo’� transcripts are detected in Rl-R3 and R-5 cells.
738 GCV Resistance and HSV-TK
+ +
�) CS
- 288 bp
.i�i: -��423bpOs CS CS
�) (‘I (.1 (-1
z� �a.
tumor cell lines (Fig. 9). BXPC-3 cells represents a cell line with
excellent bystander effects. Presence of only 10% of TK�
BXPC-3 cells in a mixed cell population ofTK1 and TK cells
lead to killing of over 90% of the cocultured cells. MIAPACA-2
and HT-29 cells demonstrated poor bystander effects. When
10-75% ofTK� MIAPACA-2 and HT-29 cells were eoeultured
with 25-90% of TK and treated with GCV. little bystander
killing of the TK cells was seen, and the TK cells grew out
as GCV-resistant cells.
DISCUSSION
Near total killing of all of the tumor cells in a tumor mass
is essential for the success of any drug based therapy of cancer
because tumor cells that are resistant to a chemotherapeutie
agent grow out as recurrent tumors. Observations from several
laboratories have documented the development of resistance to
GCV in HSV-TK-transdueed tumor cells (10, 1 1). Moolten and
colleagues ( 10) first demonstrated resistance to GCV in TK-
transduced sarcoma and bymphoma cells in in vitro and in vivo
studies. The resistant cells lacked TK activity, and the authors
postulated that boss of the TK gene was responsible for the
resistance to GCV (10). Similarly, Barba and colleague (I 1)
observed that, following 14 days of GCV treatment, a small
fraction of HSV-TK-transduced rat glioma cells grew out as
resistant colonies. Here, our own experience in 10 human 0!
tumor cell lines demonstrates that the presence of GCV resist-
anee in HSV-TK-transdueed tumor cells is a common phenom-
enon. Of the cell lines studied in vitro (five pancreas, three
colon, one duodenum, and one gastric). complete killing of
HSV-TK-transdueed tumor cells at nontoxie concentrations of
GCV was found in only two pancreas cancer cell lines, BXPC-3
and PANC-3. In the other cell lines, GCV-resistant TK� cells
were identified that demonstrated GCV sensitivity similar to
that of TK tumor cells.
The mechanisms for GCV resistance in HSV-TK-trans-
duced tumor cells are unclear (10, 1 1 ). Elucidation of these
mechanisms may have significant implications for clinical ap-
plication of the HSV-TK/GCV gene therapy approach because
development of therapeutic strategies to augment the efficacy of
this approach may be possible. Two possible mechanisms may
account for the development of GCV resistance in HSV-TK-
transduced tumor cells: absence of a functional protein in the
tumor cell; or development by the tumor cells of specific mech-
anisms to resist GCV analogous to the development of multi-
drug resistance to ehemotherapeutie agents. For example, a
specific mechanism has been identified for resistance to GCV in
human CMV ( 1 2. 13). The UL97 gene in CMV codes for a
phosphotransferase that phosphorylates GCV into its active me-
tabolite (12, 13). Mutations in the UL97 gene have been asso-
ciated with clinical resistance to GCV by CMVs (12, 13). To
distinguish between the two mechanisms, we established HSV-
TK-transdueed cell lines that were resistant to GCV by selecting
the TK� transduced cells in GCV (50-90 p.g/ml) for over 3
weeks. Our studies suggest that partial or complete deletion of
the HSV-TK gene is responsible for GCV resistance in the
TK±/GR cell lines because these changes in the TK transgene
were accompanied by loss of expression of the TK protein in the
transduced tumor cells. Our studies with the HSV-TK adenovi-
rab vector in the TK+/GR cells demonstrate that GCV sensitivity
is restored to the TK±/GR cells following transduction with an
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A
i�” �.
#{149}�‘:4� �
lll-29’lK+,(;(’V l2days
i�s �
4’ � ‘a 5’ �i
�- . If
� _- �II’r.2?L�\1 � ,- . MIA1’ACA-2RM
-‘ . .. � �‘ � 4
,/�t�: � . � �ii .. . ., . ‘1’ �- ‘ ‘ -I� ‘� �� �.. , � I’ :. - � ‘ .,
- � -it. 4 �
� �1 � ,�
� �
j��k � �
Fig. 7 Detection of HSV-TK protein in
TK4. TK, and TK+/GR cell lines by
immunostaining. Cells were cultured in
two-well chamber slides and treated with
GCV. At the end of the treatment period,
the GCV-resistant cells were fixed with
ice-cold acetone and stained with a poly-
clonal antibody to HSV-TK protein. A,
HT-29 TK� and MIAPACA-2 TK� cellsshow the presence of the HSV-TK pro-
tein in greater than 90% of the cells,
whereas TK cells stained negative with
the HSV-TK antibody. The vast majorityof TK� cells (>99%) surviving treat-
ment with 50 p.g/ml GCV for 12 days do
not stain positive with the TK antibody.Arrow, a few HSV-TK protein-positive
cells survived GCV treatment are quies-
cent cells. Established TK+/GK cell lines
(HT-29 RM and MIAPACA-2 RM)
stained negative with the TK antibody.
Original magnification. X50(). B, expres-sion ofHSV-TK protein in BXPC-3 TK�
cells. Only 20-30% of TK� BXPC-3
cells showed a strong TK antibody stain-
ing. Some cells were weakly positive.
whereas 10% of the cells did not havedetectable levels of HSV-TK protein.
Original magnification. X500.
B
Clinical Cancer Research 739
‘‘p. .
BxP�;;�rK- �‘ ��I’;� ‘., #{149},
� ,,�.. :I�� �‘4, &i�II � ‘�
* : � #{149}�‘x�� � ;!�� - �‘I � � �k’’p , . � � - ‘�‘‘
I � �5’ .- �� vV
:‘�� #{149}r � �I , ‘� ..c,,tOrt.4 .. �. -., ., -
�. . .‘I’-� �
����1 � �. � :�.,t - .‘_; S. #{149}‘� � _�!
�, : � � ., 4
(. �. %_‘ ‘
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
HT-29 MIAPACA-2
C
I-
Cri�
xl
asC
.�
I.
HT-29 HT-29RM HT-29R1
I
0
0
a�IcJ
NO AV vector, GCV
AV1LacZ.4, CCV
AV1TK, CCV
NoAVvector,GCv
AVILacZ4,GCV
AV1TK, CCV
1 10 BXPC-3
� +
E..��lt-
a5asasasax
��
�s
Inc-I
+ + 1+ +++ #{149} +� �� �-�a-� �a�a�s�s
�
� � - � � � � ��t-�., �
a5 � a- �s � a� �s
© � In © � v� �� e-l I1� � � � Cs
II
125
I 00
75
50
25
0
MIAPACA.2 M1APACA-2RM
740 GCV Resistance and HSV-TK
The mechanism for the deletion of the TK gene in the 3 S. Forry-Schaudies and Y. Chiang, unpublished observations.
Fig. 8 Transduction of TK+/GR cell lines with an adenoviral TK
vector. HSV-TK, TK�. or TK +/GR HT-29 (A) or MIAPACA-2 cells
(B) were plated in 96-well tissue culture plates and transduced with
AV1TK1 vector supernatant at multiplicity of infection of 20 (HT-29) or
200 (MIAPACA-2). Controls were transduced with an adenoviral vectorcontaining a marker 3-galactosidase gene (AVI. LacZ4). The cells weretreated with 10 p.g/ml GCV for 9-10 days, and the numbers of surviving
cells in each group were measured by the MIS cell proliferation assay
(ii = 8). The absorbance values from the cells without AV vector
treatment but with GCV treatment were defined as 100%.
adenoviral TK vector. These findings further confirm our hy-
pothesis that emergence of GCV resistance in TK�-transdueed
tumor cells is due to the lack of a functional TK protein and not
due to development of specific cellular mechanisms that resist
the actions of GCV in the tumor cells.
The mechanisms by which tumor cells may silence the
expression of transduced foreign genes are not fully understood.
Challita and colleagues (14) have shown that methybation of the
human glucocerebrosidase transgene in hematopoietie stem
cells transduced with retroviral mediated gene transfer lead to
reduced or loss of expression of the transgene. Our study doe-
uments mechanisms for loss of expression of the HSV-TK gene
in transduced 0! tumor cells. We have found partial deletion of
the HSV-TK gene in the vector beads to loss of expression of the
transgene in the HT-29 RM and M!APACA-2 RM cells. In the
HT-29 Rb-RS cells. complete loss of the HSV-TK gene was
found. Because the RM cells were derived from harvesting a
barge number of colonies. whereas the HT-29 Rl-RS ebonab cell
lines were established from randomly picked colonies, it is
possible that loss ofthe TKgene seen in the HT-29 Rl-R5 cells
is more representative than our findings in RM cells.
100
� 90;: 80
� 70� 60�.. 50
� 40.� 3�
� 20
10 -
0-
Fig. 9 In vitro bystander effect in TK-transduced Gb tumor cells.
HSV-TK � and TK cells were mixed in different ratio (total, 4 X l0�cells/per well) in 96-well plate. For example. 25% TK:75% TK�
represents a mixture that contains 1 X l0� TK and 3 X l0� TK� cells.
For each cell mixture, two additional wells with the same amount ofTK of TK� cells were plated as control wells. and the cells were
treated with 10 jig/mI GCV for 12 days. The surviving cells were
measured with the MTS cell proliferation assay. The A4�) �,,s reflect the
number of viable cells. The percentage of TK cells killed by the
bystander effect was calculated as follows: mean absorbance of cocul-
tures - mean absorbanee ofTK� cells/mean absorbanee ofTK cells X
100. Similar results were obtained from three separate experiments.
transduced 0! cell lines is not clear from our studies. The
220-bp deletion in the TK gene may be a consequence of
deletion of the TK gene after integration of the vector into the
host DNA or may be due to the presence of a spliced gene in the
original TK vector that transduced the cell. The HSV-TK gene in
the G1Tk1SvNa vector has a splice site located at bp 328.�
Production and packaging of small amounts of retrovirab vector
particles with a 220-bp deletion of the HSV-TK gene has been
demonstrated.3 Tumor cells transduced with vector particles
containing the spliced gene lack TK activity and were not killed
by GCV.3 Consistent with this observation, the 220-bp deletion
that we have observed in the HT-29 and MIAPACA-2 RM cell
lines is probably due to transduction of the tumor cells with
retroviral particles that contain the spliced TK gene because we
used the GlTkSvNa vector. It is unlikely that the 220-bp debe-
tion in the TK gene that was observed in our study is due to
cellular processes in the tumor cells because a uniform 220-bp
deletion was observed in two different 0! tumor cell lines.
Similarly, the mechanism for the complete absence of the
HSV-TK gene in the HT-29-resistant clones Rl-RS is also
unclear from the present studies; it is unlikely, however, that the
HSV-TK gene was selectively deleted from these tumor cells,
and a likely explanation is that these cells were transduced with
a retroviral vector containing only the NeoR gene. Our studies
suggest that, despite repeated propagation of TK� cell lines in
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 741
culture, only about 1-6% of the cells lose expression of the
transgene. This demonstrates remarkable stability of integration
and expression of the transgene in the vast majority of the cells.
Furthermore, despite prolonged propagation of TK� cells under
continuous exposure to GCV, resistance was not found in cells
expressing the TK protein. These studies demonstrated the re-
markable long-term efficacy of retrovirally mediated gene
transfer.
Loss of all or part of the TK gene is probably not the major
factor responsible for the development of GCV-resistant cells.
Although the same retroviral TK vector supernatant was used to
transduce all of our 01 tumor cell lines, GCV-resistant cells
were not found in TK� BXPC-3 and PANC-3 cell lines. These
two cell lines also demonstrate significant bystander killing
effects; for example, in BXPC-3 cells, admixture of only 10% of
TK� transduced cells with 90% TK cells leads to killing of the
90% of the TK cells during GCV treatment. Our studies
suggest that only about 1-5% of the cells grew out as resistant
colonies in the HT-29 and MIAPACA-2 experiments. If effi-
cient bystander killing is present, then these transduced cells
with an absent HSV-TK gene should have been killed by the
bystander effect. Immunohistochemieal studies confirmed that
about 10% of TK� BXPC-3 cells failed to stain with the TK
antibody, presumably reflecting the population of tumor cells
that were transduced with a vector incorporating a HSV-TK gene
with a splice site. The finding of absence of any GCV-resistant
colonies in TK� BXPC-3 and PANC-3 cultures is probably due
to the efficient bystander killing in these cells. GCV-resistant
colonies were detected in TK� MIAPACA-2 and HT-29 cell
lines; these cells also demonstrate poor bystander effects, and
this may explain why the small percentage of HSV-TK-trans-
duced cells lacking a functional TK protein were not killed and
grew out as resistant colonies. Poor bystander killing may also
explain the development of GCV-resistant cells in previously
reported studies. GCV resistance was found in transduced rat
glioma cells, a cell line with very poor gap junction eommuni-
cation (8, 1 1). We4 and others (8, 9) have found that the
magnitude of the bystander effect correlates with intercellular
gap junction communication between tumor cells. Similarly,
Moobten and colleagues (10) demonstrated recurrence of s.c.
tumors derived from HSV-TK-transduced Ly18 lymphoma cells
after GCV treatment. Lymphocytes lacks gap junction commu-
nication and demonstrates poor bystander killing effects (15).
These studies suggest that the efficiency of bystander killing in
a cell line may determine the development of GCV-resistant
colonies in HSV-TK-transduced cell lines.
Our findings emphasize that the bystander effect plays a
critical role in the success of suicide gene therapy approaches.
At present, there is a strong emphasis on achieving high trans-
duction efficiencies for successful gene therapy of cancer. Our
studies show that, even in a pure selected population of TK�
4 L. Yang, Y. Chiang, H. J. Lenz, K. D. Danenberg, E. M. Gordon, W. F.Anderson, and D. Parekh. Intercellular communication mediates thebystander effect during herpes simplex thymidine kinase/ganciclovir
based gene therapy of human gastrointestinal tumor cells. Hum. GeneTher., in press, 1998.
cells, the few cells that lack an intact TK transgene grow out as
GCV-resistant colonies in the absence of a good bystander
killing effect. Because, at present, 100% transduction of a tumor
mass is not achievable, even with adenoviral vectors, future
strategies aimed at improving both transduction efficiency and
the bystander killing effect will be important for successful
cancer gene therapy.
REFERENCES
I. Martin, L. A., and Lemoine, N. R. Direct cell killing by suicide
genes. Cancer Metastasis Rev., 15: 301-316, 1996.
2. Moolten, F. L. Tumor chemosensitivity conferred by inserted herpes
thymidine kinase genes: paradigm for a prospective cancer controlstrategy. Cancer Res., 46: 5276-5281, 1986.
3. Smythe, R. W., Hwang, H. C., Elshami, A. A., Amin, K. M., Eck,
S. L., Davidson, B. L., Wilson, J. M., Kaiser, L. R., and Albelda, S. M.Treatment of experimental human mesothelioma using adenovirus trans-fer of the herpes simplex thymidine kinase gene. Ann. Surg., 222:
78-86, 1995.
4. Caruso, M., Panis, Y., Gagandeep, S., Houssin, D., Saizmann, J. L.,and Klatzmann, D. Regression of established macroscopic liver metas-taxes after in situ transduction of suicidal gene. Proc. NatI. Acad. Sci.
USA, 80: 7024-7028, 1993.
5. Culver, K. W., Ram, Z., Wallbridge, S., Ishii, H., Oldfield, E. H., and
Blaese, R. M. in vivo gene transfer with retroviral vector producer cellsfor treatment of experimental brain tumors. Science (Washington DC),256: 1550-1552, 1992.
6. Freeman, S. C., Whartenby, K. A., Freeman, J. L., Abboud, C. N.,and Marrogi, A. J. In situ use of suicide tenes for cancer therapy. Semin.Oneol., 23: 31-45, 1996.
7. Freeman, S. M., Abboud, C. N., Whartenby, K. A., Packman, C. H.,Koeplin, D. S., Moolten, F. L., and Abraham, G. N. The “bystandereffect”: tumor regression when a fraction of the tumor mass is geneti-
cabby modified. Cancer Res., 53: 5274-5283, 1993.
8. Fiek, J., Barker, F. 0., II, Dazin, P., Westphale, E. M., Beyer, E. C.,
and Israel, M. A. The extent of heterocellular communication mediatedby gap junctions is predictive of bystander tumor cytotoxicity in vitro.
Proc. NatI. Acad. Sci. USA, 92: 11071-11075, 1995.
9. Pius, J. D. Cancer gene therapy: a bystander effect using the gap
junctional pathway. Mol. Carcinog., 11: 127-130, 1994.
10. Moolten, F. L., Wells, J. M., Heyman, R. A., and Evans, R. M.Lymphoma regression induced by ganciclovir in mice bearing a herpesthymidine kinase transgene. Hum. Gene Ther., 1: 125-134, 1990.
11. Barbar, D., Hardin, J., Ray, J., and Gage, F. H. Thymidine kinase-mediated killing of rat brain tumors. J. Neurosurg, 79: 729-735, 1993.
12. Chou, S., Erice, A., Jordan, M. C., Vercellotti, G. M., Michels,
D. R., Talarico, C. L., Stanat, S. C., and Biron, K. K. Analysis of the
UL97 phosphotransferase coding sequence in clinical cytomegalovirusisolates and identification of mutations conferring ganciclovir resist-ance. J. Infect. Dis., 171: 576-583, 1995.
13. Wolf, D. G.. Smith, I. L., Lee, D. J., Freeman, W. R., Fbores-
Aguilar, M., and Spector, S. A. Mutations in human cytomegalovirusUL97 gene confer clinical resistance to ganeiclovir and can be detecteddirectly in patient plasma. J. Clin. Invest., 95: 257-263, 1995.
14. Challita. P. A., and Kohn, D. B. Lack ofexpression from a retroviralvector after transduction of murine hematopoietic stem cells is associ-
ated with methybation in vivo. Proc. Nail. Acad. Sei. USA, 91: 2567-
2571, 1994.
15. Tiberghien, P., Reynolds, C. W., Keller, J. R., Murphy, B., Lyons,
R., Chiang. Y., Longo, D. L., and Ruscetti, F. W. Growth inhibition ofherpes simplex thymidine kinase (Hs-tk)-transduced primary T lympho-cytes by gancicbovir (GCV): lack of a bystander effect. Proc. Am.Assoc. Cancer Res., 34: 338, 1993.
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1998;4:731-741. Clin Cancer Res L Yang, R Hwang, Y Chiang, et al. simplex virus thymidine kinase gene.cells transduced with a retroviral vector containing the herpes Mechanisms for ganciclovir resistance in gastrointestinal tumor
Updated version
http://clincancerres.aacrjournals.org/content/4/3/731
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
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/4/3/731To request permission to re-use all or part of this article, use this link
Research. on May 26, 2021. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from