Download Bystander Tumoricidal Effect and Gap Junctional

Bystander Tumoricidal Effect and Gap Junctional Communication
in Lung Cancer Cell Lines
Kazuyoshi Imaizumi, Yoshinori Hasegawa, Tsutomu Kawabe, Nobuhiko Emi, Hidehiko Saito,
Keiji Naruse, and Kaoru Shimokata
First Department of Internal Medicine, Second Department of Physiology, and Department of Clinical Preventive Services,
Nagoya University School of Medicine, Nagoya, Japan
Tumor cells expressing the herpes simplex virus-thymidine kinase (HSV-tk) gene become sensitive to ganciclovir (GCV), and the phenomenon by which tumor cells surrounding the HSV-tk expressing cells also
become sensitive to GCV is known as the “bystander effect.” The purpose of this study was to investigate
the bystander effect in human lung-cancer cell lines, and the role of gap-junctional intercellular communication as the mechanism responsible for it. Gap-junctional intercellular communication was measured both
with a dye-transfer assay involving single-cell microinjection of Lucifer Yellow and with a PKH26/calcein-AM double-dye-transfer assay. Significant bystander tumoricidal effect was observed in lung-cancer
cell lines when cultured cells contained only 10% HSV-tk expressing cells. This was also observed to occur with cell lines of different origin or from different species. Although gap-junctional intercellular communication characterized by rapid transfer of Lucifer Yellow was not observed, we did detect gap-junctional communication marked by the slow transfer of calcein-AM in lung-cancer cell lines. However,
neither an inhibitor (1-octanol) nor an enhancer (all trans-retinoic acid [ATRA]) of gap-junctional communication affected the extent of the bystander effect. These findings suggest that low levels of gap-junctional communication may be efficient for producing the bystander effect in lung-cancer cells, or that other
mechanisms may underlie this effect. Although gap-junctional communication may play an important role
in generating the bystander effect in tumor cells expressing the HSV-tk gene, further knowledge of the
mechanism of this effect may help improve the treatment of lung cancer with an HSV-tk system. Imaizumi, K., Y. Hasegawa, T. Kawabe, N. Emi, H. Saito, K. Naruse, and K. Shimokata. 1998. Bystander
tumoricidal effect and gap junctional communication in lung cancer cell lines. Am. J. Respir. Cell
Mol. Biol. 18:205–212.
Lung cancer is one of the leading causes of cancer death,
and failure to achieve sustained therapeutic responses with
conventional medical therapies has led to the testing of
gene therapy as a new approach to treating lung cancer
(1). Recent advances in gene therapy have suggested several prospective approaches to treating human malignancies (2–4). The strategy involving a combination of the
herpes simplex virus-thymidine kinase (HSV-tk) gene as a
chemosensitive suicide gene and the antiviral drug ganciclovir (GCV) is considered one promising method for can(Received in original form October 29, 1996 and in revised form May 12,
1997)
Address correspondence to: Yoshinori Hasegawa, M.D., First Department
of Internal Medicine, Nagoya University School of Medicine, Nagoya 466,
Japan.
Abbreviations: all-trans retinoic acid, ATRA; ganciclovir, GCV; herpes
simplex virus-thymidine kinase, HSV-tk; 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide, MTT.
Am. J. Respir. Cell Mol. Biol. Vol. 18, pp. 205–212, 1998
cer gene therapy (5, 6). The rationale behind this approach
is that GCV is a nucleoside analogue specifically converted by HSV-tk to a toxic form capable of inhibiting
DNA synthesis or disrupting cellular DNA replication, but
is a poor substrate for the tk of mammalian cells. Accordingly, previous studies have demonstrated that HSV-tk genetransduced tumor cells become sensitive to GCV (7). However, non-HSV-tk gene-transduced tumor cells, which are
resistant to GCV, are also killed along with the tumor cells
expressing HSV-tk (4, 5, 8, 9). This phenomenon, known
as the “bystander effect,” may overcome the restriction
that all cells must be transduced by the therapeutic gene in
order to eradicate all tumor cells in cancer gene therapy
(4, 5, 9).
Some possible mechanisms for the bystander effect
have been proposed, and metabolic cooperation, whereby
phosphorylated GCV passes from HSV-tk gene-transduced cells to non-transduced neighboring cells via gap
junctions, is now the foremost hypothetical mechanism for
the bystander effect (4, 10). However, aberrations of gap-
206
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 18 1998
junctional channels have been implicated in cancer (11–
13), and this may be incompatible with clear observation
of the bystander effect in some malignant tumor cells (8, 9,
14, 15). On the basis of such findings, we investigated
whether the bystander effect could be observed in lungcancer cells, and whether gap-junctional communication
plays an important role in the bystander effect in lung-cancer cells.
Materials and Methods
Reagents
GCV (9-[(1,3,-dihydroxy-2-propoxy)methyl]guanine) was
obtained from Tanabe Pharmaceutical Co., Ltd. (Osaka,
Japan). The neomycin analogue G418 was obtained from
GIBCO BRL (Gaithersburg, MD). Lucifer Yellow CH (lithium salt), 1-octanol, and all-trans retinoic acid (ATRA) were
purchased from Sigma Chemical Co. (St. Louis, MO). Cell
labeling dye (PKH26) was purchased from Zynaxis Cell
Science Inc. (Malvern, PA). Cytoplasmic fluorescent dye
(calcein-AM) was purchased from Molecular Probe Inc.
(Eugene, OR).
Tissue Culture and Cell Lines
The human lung cancer cell lines RERF-LC-MS (adenocarcinoma of the lung), RERF-LC-MA (small-cell carcinoma
of the lung), and A549 (alveolar-cell carcinoma of the
lung), and the murine lung-cancer cell line 3LLSA were obtained from the Japanese Cancer Research Resources Bank.
The murine mammary carcinoma cell line MMT060562
was obtained from the American Type Culture Collection.
The human middle-pharyngeal squamous-cell carcinoma
cell line NT-SQ was established in our laboratory. Except
for 3LLSA, these cell lines were maintained in Eagle’s
minimum essential medium (MEM) supplemented with 1%
nonessential amino acids, 1% L-glutamine, 1% sodium
pyruvate, penicillin (50 U/ml), streptomycin (50 mg/ml),
and 10% fetal calf serum (FCS) (MEM/FCS). 3LLSA was
maintained in RPMI 1640 with 1% nonessential amino acids, 1% L-glutamine, 1% sodium pyruvate, penicillin (50
U/ml), streptomycin (50 mg/ml), and 10% FCS (RPMI/
FCS). The cell lines RERF-LC-MS, RERF-LC-MA, 3LLSA,
MMT060562, and NT-SQ were transduced with the recombinant retroviral vector pLTRNL. This was derived
from the Moloney murine leukemia virus (MuLV) backbone, consisting of the long terminal repeat (LTR), HSV-tk
complementary DNA (cDNA), and the neomycin-resistance gene (16). The cell lines were plated in 24-well tissue
culture plates and transduced 24 h later by exposure for
48 h to viral vector derived from the PA317/LTRNL vector-producer line, at a virus titer of 5 3 104 colony forming
units (cfu)/ml in the presence of 4 mg/ml polybrene. The
cells were then cultured with MEM/FCS or RPMI/FCS containing G418 (800 mg/ml and 600 mg/ml, respectively). After 2 wk of selection, G418-resistant clones of RERF-LCMS-LTRNL, RERF-LC-MA-LTRNL, 3LLSA-LTRNL,
MMT-LTRNL, and NT-SQ-LTRNL cells were selected
randomly from the surviving colonies and used in the subsequent experiments. The sensitivity of these cell lines to
GCV was determined with the 3-(4,5-dimethylthiazol-2-
yl)-2,5-diphenyltetrazolium bromide (MTT) assay as previously described (16).
Measurement of the Bystander Effect In Vitro
Experiments designed to investigate the bystander effect
were performed by coculturing the parental wild-type tumor cells with the HSV-tk gene-transduced tumor cells.
The wild-type tumor cells were seeded at 105 cells in 35mm dishes in mixtures with 10 5, 3.3 3 104, 104, and 103
HSV-tk gene-transduced cells (mixture ratios: 1:1, 3:1,
10:1, and 100:1, respectively). To study the effect of physically separated cocultivation on the bystander effect, we
used the Falcon cell culture insert (0.45 mm pore size, 25 mm
diameter) (Becton Dickinson, Inc., Lincoln Park, NJ). The
parental wild-type cells were seeded at 105 cells in the bottom of 35-mm wells. The cell-culture inserts were plated
into the wells, and 10 5 cells of HSV-tk gene-transduced
cells were inoculated into the upper inserts.
After 12 h incubation, the cells were cultured with or
without 5 mg/ml (final concentration) of GCV for 5 days.
Culture medium with or without GCV was replaced every
2 days. The cells were then harvested and the number of
viable cells was determined through the dye-exclusion
method. The parental wild-type cells alone, or HSV-tk
gene-transduced cells alone, were cultured in the medium
with or without GCV as a control.
Lucifer Yellow Dye-Transfer Assay of Gap-Junctional
Intercellular Communication
Gap-junctional intercellular communication was measured
by microinjection of 10% Lucifer Yellow CH in 0.33 M
LiCl. This fluorescent dye was microinjected at constant
pressure into cells of confluent cultures with a microinjector (Microinjector 5242; Eppendorf-Netheler-Hinz GmbH,
Hamburg, Germany) and a manipulator (Micromanipulator, Eppendorf-Netheler-Hinz GmbH). Cultured cells were
washed and the medium was replaced with phosphatebuffered saline (PBS) containing 10% FCS. Injections were
done with an Axiovert 135 Microscope (Karl-Zeiss Co.,
Oberkochen, Germany) with phase and fluorescence optics.
Finely pulled glass capillaries (Femtotips) for the injection
of dye into tissue-culture cells (diameter of opening of tip:
0.5 6 0.2 mm) were purchased from Eppendorf-NethelerHinz GmbH. After microinjection of 10% Lucifer Yellow
CH into a single cell, the spread of fluorescence was monitored with a silicon-intensified target camera (Hamamatsu
C2400; Hamamatsu Photonics, Hamamatsu, Japan). The
total number of fluorescent neighbors of the injected cell
was counted 10 min after the end of injection. To investigate the inhibition or enhancement of gap-junctional communication, we added 1-octanol (1 mM) (17, 18) in 0.087%
dimethyl sulfoxide (DMSO) or ATRA (z 10211 M to 1026
M) (11, 19) in acetone into the culture medium. Such concentrations have been reported to be most appropriate for
inhibition or enhancement of gap-junctional communication (11, 18). Control cells were cultured with medium
containing DMSO or acetone, respectively. After 24 h of
incubation, cells were washed and culture medium was
replaced with PBS containing 1 mM 1-octanol or ATRA
(z 10211 M to 1026 M). Culture medium of control cells
Imaizumi, Hasegawa, Kawabe, et al.: Bystander Effect and Gap Junction in Lung Cancer Cells
was replaced with PBS containing DMSO or acetone. The
microinjection of fluorescent dye was performed under
these conditions.
Flow-cytometric Assay of Gap-Junctional
Intercellular Communication
We measured gap-junctional dye transfer from HSV-tkexpressing cells to wild-type tumor cells with a flow-cytometric assay as described previously (20). Wild-type tumor
cells (RERF-LC-MS, 3LLSA) were harvested in a singlecell suspension by trypnization, washed once with Solution
A (PBS 31 supplemented with 10 mM glucose and 30 mM
4-(2-hydroxyethyl)-1-piperazine-N9-Z-ethane sulfonic acid
[Hepes]). After centrifugation, the cell pellets were suspended at 2 3 107 cells/pml and an equal volume of 4 3
1026 M PKH26 solution was added. Staining with PKH26
was done for 2 min with gentle agitation, and was terminated by incubating with 2 ml FCS. Cells were then washed
three times and cultured in normal medium for 24 h. Staining of HSV-tk-expressing cells (RERF-LC-MS-LTRNL,
3LLSA-LTRNL) with calcein-AM was performed in culture dishes. After gentle washing, a solution of 0.5 mM of
calcein-AM freshly made up in Solution A was added on
top of the cells for 30 min at room temperature. Unincorporated dye was eliminated by three consecutive washes
with medium. Aliquots of 105 PKH26-labeled wild-type tumor cells and 105 calcein-AM-labeled HSV-tk-expressing
cells were cocultured in 35-mm dishes for 4.5 h at 37 8C.
Cocultivations were performed in the pure culture medium,
medium containing 1 mM 1-octanol, or medium containing
ATRA (z 10211 M to 1026 M). The cocultured cells were
then treated with trypsin, centrifuged, and resuspended in
Solution A. Cells were analyzed flow cytometrically with a
dual-laser system on an Epics XL (Coulter Electronics,
Miami, FL) flow cytometer. PKH26 and calcein emission
were excited with a 488-nm argon laser and collected through
575- and 525-nm bandpass filters, respectively. Aliquots of
each stained cell population were mixed just before flow
cytometry to determine the pattern at the start of cocultivation.
207
cells were cocultured with various ratios of HSV-tk gene
transduced cells. As shown in Figure 1, addition of GCV
(5 mg/ml) to the culture medium resulted in the elimination of both HSV-tk gene-transduced cells and nontransduced wild-type cells of human non-small-cell lung cancer
(RERF-LC-MS) and murine lung-cancer cells (3LLSA).
This toxic effect of HSV-tk-positive cells on nearby HSVtk-negative cells was also seen in human small-cell lung
cancer cells (RERF-LC-MA), pharyngeal squamous-cell
carcinoma cells (NT-SQ), and murine mammary carcinoma
cells (MMT060562) (data not shown). These bystander effects were observed even when HSV-tk-positive cells were
as few as 10% of the mixed cell population (P , 0.01).
We next evaluated the bystander effect between the
different cell lines. Human alveolar cell carcinoma cells
(A549) were cocultured with RERF-LC-MS-LTRNL (HSVtk gene-transduced human lung adenocarcinoma cells) or
3LLSA-LTRNL (HSV-tk gene-transduced murine lung-cancer cells) in medium containing GCV (5 mg/ml). As shown
in Figure 2, bystander effects were obvious and similar to
those observed between cell lines of the same origin (Figure 1). We also observed the bystander effect when cells of
the A549 cell line were cocultured with NT-SQ-LTRNL
(HSV-tk gene-transduced pharyngeal squamous cell carci-
Measurement of Bystander Effect with Inhibitor or
Enhancer of Gap-Junctional Communication
Aliquots of 105 MMT060562, RELF-LC-MS, or RELF-LCMA cells were cocultured with 105 MMT-LTRNL, RELFLC-MS-LTRNL, or RELF-LC-MA-LTRNL cells, respectively, in 35-mm dishes (1:1 coculture) with conditioned
medium containing 1-octanol (1 mM). Similarly, 105 wildtype cells were cultured with 104 HSV-tk cells (10:1 coculture) with ATRA (z 10211 M to 1026 M). The cells were
then cultured with GCV (5 mg/ml) for 5 days, after which
living cells were counted as described earlier.
Results
Bystander Tumoricidal Effect on Lung-Cancer
Cell Lines In Vitro
We first investigated the bystander effect of HSV-tk genetransduced cells on nontransduced parental wild-type cells
in lung-cancer cell lines. A fixed number (105) of wild-type
Figure 1. Bystander tumoricidal effect on lung cancer cell lines in
vitro. (A) Results with RERF-LC-MS cells. (B) Results with
3LLSA cells. The wild-type tumor cells were seeded with HSV-tk
gene-transduced cells at the mixture ratios of 1:1, 3:1, 10:1, and
100:1, respectively. After 12 h incubation, they were cultured
with (closed column) or without (open column) 5 mg/ml GCV for
5 days. The number of viable cells was determined with the dyeexclusion method. Results are mean 6 SEM of four independent
experiments. *P , 0.01 compared with wild-type control.
208
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 18 1998
Figure 2. Bystander effect between cancer cells of different tissue
or species origin. Cells of the human alveolar-cell-carcinoma line
A549 (105) were cocultured with 105 cells of HSV-tk gene-transduced RERF-LC-MS-LTRNL human lung adenocarcinoma cells
(A), or with 105 HSV-tk gene-transduced cells of the murine lung
cancer line 3LLSA-LTRNL (B). After 12 h incubation, the cells
were cultured with (closed bar) or without (open bar) 5 mg/ml
GCV for 5 days. The number of viable cells was determined with
the dye-exclusion method. Results are mean 6 SEM of three independent experiments. *P , 0.01 compared with no addition of
GCV.
noma) cells (data not shown). These results indicate that
HSV-tk gene-transduced cells can confer GCV sensitivity
on a wide variety of tumor cells beyond those of identical
tissue origin and identical species.
Analysis of Gap-Junctional Communication for
Lung-Cancer Cell Lines
To investigate the mechanisms of the bystander effect in
lung-cancer cells, we assessed gap-junctional cell-to-cell communication with the single-cell microinjection method, using a fluorescent probe. We first used the murine breast
cancer cell line (MMT060562) as a control cell line, because
the existence of cell-to-cell communication through gap
junctions in MMT060562 cells has been shown previously
(21). Figure 3A shows the extensive spread of the fluorescent dye to neighboring cells among MMT060562 cells. In
contrast to MMT060562 cells, little dye transfer was observed between cells of the lung cancer cell lines RERFLC-MS, RERF-LC-MA, and 3LLSA (Figures 3C and 3D;
Table 1). The number of communicating cells in these cell
lines was significantly smaller than for MMT060562 cells
(Table 1). For further assessment of gap-junctional communication, we adopted the PKH26/calcein-AM doubledye-transfer assay. This method allowed us to use the
longer periods of quantitative observation that result in
the detection of slow and low-level intercellular communication via gap junctions. We assessed the transfer of the
gap-junction-permeable dye calcein-AM from HSV-tkexpressing cells to wild-type cells stained with the membrane-binding dye PKH26 through flow cytometry. In contrast to the single-cell microinjection method with Lucifer
Yellow, the double-dye-transfer assay enabled us to detect
the gap-junctional communication of RERF-LC-MS (Figure 4) and 3LLSA cells (data not shown). Surprisingly, gap-
Figure 3. Fluorescence-optics images of gap-junctional intercellular communication in various cancer cells. Lucifer Yellow dye was
microinjected into cancer cells of
confluent cultures. Ten minutes
after the end of microinjection
of the dye into a single cell, the
spread of fluorescence was monitored with a silicon-intensified target camera. (A) MMT060562 (murine breast carcinoma) cells in
control medium; (B) MMT060562
cells in medium containing 1 mM
1-octanol; (C) RERF-LC-MS (human lung adenocarcinoma) cells
in control medium; (D) 3LLSA
(murine lung carcinoma) cells in
control medium. A through D:
original magnification: 3200.
Imaizumi, Hasegawa, Kawabe, et al.: Bystander Effect and Gap Junction in Lung Cancer Cells
209
junctional communications could be detected with this system even when two different cell lines from different species (human and murine lung cancers) were cocultured
(Figure 5). These findings indicate that gap-junctional communications in lung-cancer cell lines are poor for the rapid
transfer of Lucifer Yellow, although they may permit the
slow transfer of calcein-AM from cell to cell. On the other
hand, as shown in Table 1, the bystander tumoricidal killing effect in these cell lines was obvious (. 80%).
munication. For this purpose, we added 1-octanol (1 mM)
or ATRA (z 10211 M to 1026 M) to the culture medium
24 h before addition of GCV for measurement of the
bystander effect in vitro. Control experiments were done
in medium containing DMSO or acetone. Addition of
1-octanol to the culture medium significantly inhibited
gap-junctional communication between MMT060562 cells
(Figure 3B). It also inhibited gap-junctional communication between RERF-LC-MS cells (Figure 4C) and that between 3LLSA cells (data not shown) as measured with the
double-dye-transfer assay. However, it did not inhibit the
bystander effect in these cell lines (Table 2). On the other
hand, although ATRA is known as a gap-junctional enhancer, it did not enhance the bystander effect on either
RERF-LC-MS cells or 3LLSA cells (Table 2). In our experiments, ATRA at concentrations of 10211 M to 1026 M
did not clearly enhance gap-junctional communication as
measured with the double-dye-transfer assay (Figure 4D),
although it slightly enhanced gap-junctional communication as measured with the Lucifer Yellow microinjection
method (data not shown). On the basis of these findings,
we speculate that the low levels of gap-junctional communication and transfer of phosphorylated GCV may be sufficient to produce the bystander effect, or that multiple
mechanisms including gap-junctional communication may
exist in some malignant cell lines.
Role of Gap-Junctional Communication in the
Bystander Effect
We investigated whether 1-octanol or ATRA influenced
the bystander tumoricidal effect and gap-junctional com-
Role of Soluble Factors in the Bystander Effect
We investigated whether some soluble factors produced
by HSV-tk gene-transduced cells might play an important
role in the bystander effect. Wild-type cells (RERF-LC-
TABLE 1
Gap-junctional communication and bystander effect
Cells
MMT060562
RERF-LC-MS
RERF-LC-MA
3LLSA
Number of
Communicating Cells*
% Bystander
Tumor Killing†
3.2 6 1.1
0.3 6 0.1‡
0.7 6 0.4‡
0.5 6 0.2‡
97.2 6 1.1%
80.0 6 4.1%
95.9 6 1.0%
85.6 6 2.1%
* After microinjection of Lucifer Yellow, the total number of fluorescent
neighbors of the injected cell was counted. A single-cell microinjection was repeated from 30 to 48 times in each experiment. Each value is mean 6 SD.
†
% bystander tumor killing was expressed as [1 2 (number of surviving cells
of coculture experiment treated with GCV/number of cells of wild-type cells
treated with GCV)] 3 100. Each value is mean 6 SE of at least three independent experiments.
‡
Significant difference between MMT060562 cells and other cells (P , 0.05)
by Mann–Whitney U test.
Figure 4. Double-dye-transfer assay of gap-junctional intercellular
communication between RERFLC-MS cells. Wild-type cells were
labeled with the untransferable
fluorescent dye PKH26, and HSVtk expressing cells (RERF-LC-MSLTRNL) were stained with the
transferable dye calcein-AM. After
105 cells of both groups were cocultured for 4.5 h, flow cytometric
analysis was performed. PKH26
fluorescence is shown on the vertical axis and calcein-AM fluorescence on the horizontal axis. (A)
Staining pattern of cells just before cocultivation; (B) cocultivation for 4.5 h in culture medium;
(C) cocultivation for 4.5 h in culture medium containing 1 mM 1octanol; (D) cocultivation for 4.5
h in culture medium containing
ATRA (1026 M).
210
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 18 1998
TABLE 2
Inhibition and enhancement of gap-junctional
communication and bystander effect*
% Bystander Tumor Killing
Agent
1-Octanol, 1 mM
MMT060562
RERF-LC-MS
3LLSA
2
97.2%
(95.2–98.3)
96.8%†
(94.8–98.9)
n.d.‡
1
n.d.
76.8%
(66.2–86.2)
78.1%†
(69.2–86.2)
50.1%
(48.8–66.7)
54.3%†
(41.0–59.0)
88.5%
(81.6–98.0)
85.8%†
(78.7–95.4)
67.1%
(53.1–81.0)
68.0%†
(54.0–82.0)
2
1
ATRA, 1026 M
* Under the gap-junctional inhibitory condition with 1-octanol (1 mM), 105
wild-type cells were cocultured with 105 HSV-tk gene-transduced cells in the
presence of GCV (5 mg/ml). On the other hand, under the gap-junctional enhanced condition with ATRA (1026 M), 105 wild-type cells were cocultured with
104 HSV-tk cells in the presence of GCV (5 mg/ml). Control experiments were
done with addition of DMSO or acetone. % bystander tumor killing was expressed as [1 2 (number of surviving cells in the coculture of experiment
treated with GCV/number of wild-type cells treated with GCV)] 3 100. Values
are means and ranges of at least three independent experiments.
†
No significant difference either with or without agents.
‡
n.d. 5 not done.
Figure 5. Double-dye-transfer assay of gap-junctional intercellular communication between two different cells from different
species. Cells of a human lung-cancer cell line (RERF-LC-MS)
were stained with PKH26 and HSV-tk expressing cells of a murine lung-cancer cell line (3LLSA-LTRNL) were stained with
calcein-AM. (A) Staining pattern of cells just before cocultivation; (B) cocultivation for 4.5 h in culture medium.
transduced human lung-cancer cell lines. The HSV-tk enzyme phosphorylates GCV to a toxic form capable of inhibiting DNA synthesis, and this phosphorylated form of
GCV cannot pass through the cell membrane (4, 15). Consequently, metabolic cooperation, whereby the phosphorylated GCV passes from HSV-tk gene-transduced cells to
nontransduced neighboring cells via gap junctions, has
been proposed as the mechanism of the bystander effect
MS) were cocultured with HSV-tk gene-transduced cells
(RERF-LC-MS-LTRNL) separated from the wild-type
cells by a membrane filter (cell culture insert) (Figure 6A).
Under the condition of physical separation, wild-type cells
were not killed at all, although HSV-tk gene-transduced
cells were killed in the same medium containing GCV
(5 mg/ml). The same results were observed with the
3LLSA (Figure 6B) and RERF-LC-MA cells (data not
shown). If some soluble factors exist and influence the
growth of cells, they could affect the viability of wild-type
cells through membranous pores. These results suggest
that the bystander effect is not mediated by soluble factors, and requires cell–cell contact as previously reported
(22, 23).
Discussion
It is very useful to understand the exact mechanisms responsible for the bystander tumoricidal effect, since this
may permit enhancing this effect to achieve better efficacy
in the treatment of cancer. In this study we addressed the
dependency of the bystander effect on gap-junctional
communication in an in vitro system with HSV-tk gene-
Figure 6. Bystander effect under the condition of physical separation of cells. The parental wild-type cells were seeded in the
bottom of wells, and cell-culture inserts were placed into the
wells. HSV-tk-transduced cells were seeded into the upper inserts. These cells were cultured with or without GCV (5 mg/ml).
(A) Closed column, RERF-LC-MS cells; open column, RERFLC-MS-LTRNL cells; (B) closed column, 3LLSA cells; open column, 3LLSA-LTRNL cells. Results are mean 6 SEM of four independent experiments.
Imaizumi, Hasegawa, Kawabe, et al.: Bystander Effect and Gap Junction in Lung Cancer Cells
(4, 10, 15, 23, 24). The gap-junctional channels allow the
movement of molecules smaller than 1,000 Da, such as
cyclic adenosine monophosphate (cAMP) and inositol trisphosphate, and prevent the movement of protein or nucleic acids (12). Since the molecular weight of the phosphorylated forms of GCV is about 300 Da (4), it is thought
possible that phosphorylated GCV can be transferred to
neighboring cells through gap junctions, and that these
neighboring cells must be killed by GCV administration.
However, previous reports have shown aberrations of
gap-junctional communication in malignant cells (11, 13),
and this may correlate with these cells’ neoplastic transformation (11, 12, 19). In the present study we first measured
gap-junctional communication with a single-cell microinjection method involving Lucifer Yellow fluorescent dye,
which has been frequently used for testing gap-junctional
communication (18, 21). With this method we could not
detect significant gap-junctional communication in several
human or murine lung cancer cell lines, whereas we could
detect the rapid transfer of Lucifer Yellow via gap-junctional communication in the murine mammary carcinomacell line MMT060562. Because a fluorescent-dye-injection
method involving Lucifer Yellow is not appropriate for
longer periods of observation, we used a newly developed
method for assessing gap-junctional communication, the
PKH26/calcein-AM double-dye-transfer assay (20). This
method allowed us to quantitatively evaluate longer periods of observation, which resulted in the detection of slow
and low-level gap-junctional communication. In contrast
to the Lucifer Yellow dye-transfer assay, the PKH26/calcein-AM double-dye-transfer assay enabled us to detect gapjunctional communication in lung-cancer cell lines. Thus,
lung-cancer cell lines lacked the rapid-transfer type of gapjunctional communication, but exhibited the slow-transfer
type of gap-junctional communication assessed with the
PKH26/calcein-AM double-dye-transfer assay. Further,
because the types and the extent of the expression of gapjunctional proteins, such as connexons, varies among cell
lines, the permeability of gap-junction channels between
two different cells, which consist of two different connexons, must be variable (12). We could detect gap-junctional
communication among lung-cancer cell lines from different species with the double-dye-transfer assay but not with
the single-cell microinjection method. This finding indicates that lung-cancer cells of two different species or origins employ the slow-transfer type of gap-junctional communication with one another. We found evidence that the
extent of the bystander effect, even among cell lines from
different species (human and mouse), did not differ from
that between identical cell lines.
Recently, Fick and colleagues reported a correlation
between the extent of gap-junctional communication and
the bystander effect (24). Another report pointed out that
HeLa cells, which showed little gap-junctional communication and little of a bystander effect, began to show an
obvious bystander effect when transfected with the gene
encoding connexin 43 (23). However, neither the gapjunctional inhibitor 1-octanol nor the gap-junctional enhancer ATRA affected the extent of the bystander effect
on lung-cancer cell lines in our experiments. Straight-chain
alcohols containing from seven to nine carbons, such as
211
1-octanol, have been shown to selectively and reversibly
inhibit cell–cell communication via gap junctions in a
number of cell systems (17, 18). In our study, 1-octanol did
not reduce the tumor-cell killing through the bystander effect, although it significantly inhibited gap-junctional communication both in the Lucifer Yellow transfer assay and
the double-dye-transfer assay. Furthermore, ATRA, which
has been shown to enhance gap-junctional communication,
did not enhance the bystander effect at concentrations of
1026 M to 10211 M (11, 19, 25). Thus, we speculate that a
low level of gap-junctional communication may be sufficient to produce the bystander effect in lung-cancer cell
lines, or that mechanisms other than gap-junctional communication may exist. We failed to find any bystander effect with the physical separation of cells with cell-culture
inserts, which indicates that this effect is not mediated by
soluble factors in the medium, and requires direct cell–
cell contact (22). Freeman and colleagues suggested that
phagocytosis by HSV-tk2 cells of apoptotic vesicles generated by dying HSV-tk1 cells plays an important role in the
bystander effect (9). We speculate that multiple pathways
rather than a single pathway may be involved in the mechanism of the bystander effect, even though gap-junctional
communication plays a crucial role in the bystander effect.
In this study, we demonstrated the possibility of successful treatment of lung cancers with an HSV-tk system
because the bystander effect occurs in the lung-cancer cell
lines that were studied. We failed to regulate the bystander effect by modulating gap-junctional communication, and further knowledge of the mechanism responsible
for the bystander effect is needed to achieve effective cancer treatment with an HSV-tk system.
References
1. Lee, C. T., H. L. Chen, and D. P. Carbone. 1995. Gene therapy for lung cancer. Ann. Oncol. 6:S61–S63.
2. Culver, K. W., and R. M. Blaese. 1994. Gene therapy for cancer. Trends
Genet. 10:174–178.
3. Herrmann, F. 1995. Cancer gene therapy: principles, problems, and perspectives. J. Mol. Med. 73:157–163.
4. Culver, K. W., T. M. Vickers, J. L. Lamsam, H. W. Walling, and T. Seregina.
1995. Gene therapy for solid tumors. Br. Med. Bull. 51:192–204.
5. Curiel, D. T., J. M. Pilewski, and S. M. Albelda. 1996. Gene therapy approaches for inherited and acquired lung diseases. Am. J. Respir. Cell Mol.
Biol. 14:1–18.
6. Hasegawa, Y., N. Emi, and K. Shimokata. 1995. Retroviral transfer of
HSV1-TK gene into human lung cancer cell line. J. Mol. Med. 73:107–112.
7. Moolten, F. L. 1986. Tumor chemosensitivity conferred by inserted herpes
thymidine kinase genes: paradigm for a prospective cancer control strategy. Cancer Res. 46:5276–5281.
8. Culver, K. W., Z. Ram, S. Wallbridge, H. Ishii, E. H. Oldfield, and R. M.
Blaese. 1992. In vivo gene transfer with retroviral vector-producer cells for
treatment of experimental brain tumors. Science 256:1550–1552.
9. Freeman, S. M., C. N. Abboud, K. A. Whartenby, C. H. Packman, D. S.
Koeplin, F. L. Moolten, and G. N. Abraham. 1993. The “bystander effect”:
tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res. 53:5274–5283.
10. Pitts, J. D. 1994. Cancer gene therapy: a bystander effect using the gapjunctional pathway. Mol. Carcinogen. 11:127–130.
11. Goldberg, G. S., and J. S. Bertram. 1994. Retinoids, gapjunctional communication and suppression of epithelial tumors. In Vivo 8:745–754.
12. Kumar, N. M., and N. B. Gilula. 1996. The gap junction communication
channel. Cell 84:381–388.
13. Yamasaki, T., K. Enomoto, K. Moritake, and T. Maeno. 1994. Analysis of
intra- and intercellular calcium signaling in a mouse malignant glioma cell
line. J. Neurosurg. 81:420–426.
14. Kuriyama, S., T. Nakatani, K. Masui, T. Sakamoto, K. Tominaga, M.
Yoshikawa, H. Fukui, K. Ikenaka, and T. Tsujii. 1995. Bystander effect
caused by suicide gene expression indicates the feasibility of gene therapy
for hepatocellular carcinoma. Hepatology 22:1838–1846.
212
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 18 1998
15. Bi, W. L., L. M. Parysek, R. Warnick, and P. J. Stambrook. 1993. In vitro evidence that metabolic cooperation is responsible for the bystander effect
observed with HSV tk retroviral gene therapy. Hum. Gene Ther. 4:725–
731.
16. Hasegawa, Y., N. Emi, K. Shimokata, A. Abe, T. Kawabe, T. Hasegawa, T.
Kirioka, and H. Saito. 1993. Gene transfer of herpes simplex virus type I
thymidine kinase gene as a drug sensitivity gene into human lung cancer
cell lines using retroviral vectors. Am. J. Respir. Cell Mol. Biol. 8:655–661.
17. Johnston, M. F., S. A. Simon, and F. Ramon. 1980. Interaction of anaesthetics with electrical synapses. Nature 286:498–499.
18. Reynhout, J. K., P. D. Lampe, and R. G. Johnson. 1992. An activator of protein kinase C inhibits gap junction communication between cultured bovine lens cells. Exp. Cell. Res. 198:337–342.
19. Hossain, M. Z., L. R. Wilkens, P. P. Mehta, W. Loewenstein, and J. S. Bertram. 1989. Enhancement of gapjunctional communication by retinoids
correlates with their ability to inhibit neoplastic transformation. Carcinogenesis 10:1743–1748.
20. Tomasetto, C., M. J. Neveu, J. Daley, P. K. Horan, and R. Sager. 1993.
21.
22.
23.
24.
25.
Specificity of gap junction communication among human mammary cells
and connexin transfectants in culture. J. Cell Biol. 122:157–167.
Enomoto, K., K. Furuya, S. Yamagishi, and T. Maeno. 1992. Mechanically
induced electrical and intracellular calcium responses in normal and cancerous mammary cells. Cell Calcium 13:501–511.
Wu, J. K., W. G. Cano, S. A. G. Meylaerts, P. Qi, F. Vrionis, and V. Cherington. 1994. Bystander tumoricidal effect in the treatment of experimental brain tumors. Neurosurgery 35:1094–1103.
Mesnil, M., C. Piccoli, G. Tiraby, K. Willecke, and H. Yamasaki. 1996. Bystander killing of cancer cells by herpes simplex virus thymidine kinase
gene is mediated by connexins. Proc. Natl. Acad. Sci. USA 93:1831–1835.
Fick, J., F. G. Barker II, P. Dazin, E. M. Westphale, E. C. Beyer, and M. A.
Israel. 1995. The extent of heterocellular communication mediated by gap
junctions is predictive of bystander tumor cytotoxicity in vitro. Proc. Natl.
Acad. Sci. USA 92:11071–11075.
Guo, H., P. Acevedo, F. D. Parsa, and J. S. Bertram. 1992. Gap-junctional
protein connexin 43 is expressed in dermis and epidermis of human skin:
differential modulation by retinoids. J. Invest. Dermatol. 99:460–467.