Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

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). 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 cancer 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 gene-transduced 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-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 lung-cancer cells, and whether gap-junctional communication plays an important role in the bystander effect in lung-cancer cells.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

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 µg/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 µg/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 × 10⁴ colony forming units (cfu)/ml in the presence of 4 µg/ml polybrene. The cells were then cultured with MEM/FCS or RPMI/FCS containing G418 (800 µg/ml and 600 µg/ml, respectively). After 2 wk of selection, G418-resistant clones of RERF-LC-MS-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 35-mm dishes in mixtures with 10⁵, 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 µm 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 10⁵ 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 µg/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 phosphate-buffered 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 ± 0.2 µm) were purchased from Eppendorf-Netheler-Hinz 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 (~ 10¹¹ M to 10⁶ 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 (~ 10¹¹ M to 10⁶ M). Culture medium of control 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-tk-expressing 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 single-cell suspension by trypnization, washed once with Solution A (PBS ×1 supplemented with 10 mM glucose and 30 mM 4-(2-hydroxyethyl)-1-piperazine-N'-Z-ethane sulfonic acid [Hepes]). After centrifugation, the cell pellets were suspended at 2 × 10⁷ cells/pml and an equal volume of 4 × 10⁶ 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 µM 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°C. Cocultivations were performed in the pure culture medium, medium containing 1 mM 1-octanol, or medium containing ATRA (~ 10¹¹ M to 10⁶ 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.

Measurement of Bystander Effect with Inhibitor or Enhancer of Gap-Junctional Communication

Aliquots of 10⁵ MMT060562, RELF-LC-MS, or RELF-LC-MA cells were cocultured with 10⁵ MMT-LTRNL, RELF-LC-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 wild-type cells were cultured with 10⁴ HSV-tk cells (10:1 coculture) with ATRA (~ 10¹¹ M to 10⁶ M). The cells were then cultured with GCV (5 µg/ml) for 5 days, after which living cells were counted as described earlier.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Bystander Tumoricidal Effect on Lung-Cancer Cell Lines In Vitro

We first investigated the bystander effect of HSV-tk gene-transduced cells on nontransduced parental wild-type cells in lung-cancer cell lines. A fixed number (10⁵) of wild-type cells were cocultured with various ratios of HSV-tk gene transduced cells. As shown in Figure 1, addition of GCV (5 µg/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 HSV-tk-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).

View larger version (22K): [in this window] [in a new window]   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 µg/ml GCV for 5 days. The number of viable cells was determined with the dye-exclusion method. Results are mean ± SEM of four independent experiments. *P < 0.01 compared with wild-type control.

We next evaluated the bystander effect between the different cell lines. Human alveolar cell carcinoma cells (A549) were cocultured with RERF-LC-MS-LTRNL (HSV-tk gene-transduced human lung adenocarcinoma cells) or 3LLSA-LTRNL (HSV-tk gene-transduced murine lung-cancer cells) in medium containing GCV (5 µg/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 carcinoma) 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.

View larger version (15K): [in this window] [in a new window]   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 µg/ml GCV for 5 days. The number of viable cells was determined with the dye-exclusion method. Results are mean ± SEM of three independent experiments. *P < 0.01 compared with no addition of GCV.

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 RERF-LC-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 ). For further assessment of gap-junctional communication, we adopted the PKH26/calcein-AM double- dye-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-tk- expressing 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-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 , the bystander tumoricidal killing effect in these cell lines was obvious (> 80%).

View larger version (76K): [in this window] [in a new window]   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: ×200.

View larger version (43K): [in this window] [in a new window]   Figure 4.   Double-dye-transfer assay of gap-junctional intercellular communication between RERF-LC-MS cells. Wild-type cells were labeled with the untransferable fluorescent dye PKH26, and HSV-tk expressing cells (RERF-LC-MS-LTRNL) 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 1-octanol; (D) cocultivation for 4.5 h in culture medium containing ATRA (106 M).

View larger version (25K): [in this window] [in a new window]   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.

Role of Gap-Junctional Communication in the Bystander Effect

We investigated whether 1-octanol or ATRA influenced the bystander tumoricidal effect and gap-junctional communication. For this purpose, we added 1-octanol (1 mM) or ATRA (~ 10¹¹ M to 10⁶ 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 ). In our experiments, ATRA at concentrations of 10¹¹ M to 10⁶ 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 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-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 µg/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).

View larger version (17K): [in this window] [in a new window]   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 µg/ml). (A) Closed column, RERF-LC-MS cells; open column, RERF-LC-MS-LTRNL cells; (B) closed column, 3LLSA cells; open column, 3LLSA-LTRNL cells. Results are mean ± SEM of four independent experiments.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

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-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 (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 carcinoma- cell 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 gap-junctional communication in lung-cancer cell lines. Thus, lung-cancer cell lines lacked the rapid-transfer type of gap-junctional 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 gap-junctional 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 gap-junctional 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 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 10⁶ M to 10¹¹ 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-tk cells of apoptotic vesicles generated by dying HSV-tk⁺ 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.