Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Malignant mesothelioma (MM) is an uncommon tumor; however, in the past 10 years its incidence has increased (1). This is partly due to previous mining of asbestos, its widespread use in industry, and a relatively long latency period between exposure to asbestos and development of the disease (2, 3). In fact, the rising incidence of MM is predicted to continue for another 20 years (1) or even longer. Current treatment modalities including surgery, chemotherapy, and radiation have had little or no effect on survival once the diagnosis of MM is made (2, 3). Given the failure of conventional treatment and its localized nature (usually in the chest or peritoneal cavity), MM is an ideal target and platform for exploring new treatment strategies such as gene therapy. The herpes simplex virus thymidine kinase (HSV-tk) gene confers drug sensitivity to cells through enzymatic phosphorylation of the antiviral agent acyclovir and its derivative, ganciclovir (GCV), transforming them into highly toxic forms (4, 5). The current approach of delivering the HSV-tk gene to MM with an adenoviral vector has been shown to be effective in cell transduction, cell killing, and increased survival in an animal model, but tumor eradication is not complete (6). Moreover, the adenoviral vector elicits an immune response that may prevent its repeated administration (7).

We have developed an alternative cell-based approach, using a permanently transduced cell line encoding the HSV-tk gene (PA-1-STK cells), to deliver the gene to neighboring tumor cells (11). This cell line has been shown to kill a variety of cell lines in vitro and in vivo by the "bystander" effect, and is currently in a Phase I trial for treating refractory ovarian carcinoma (12). Moolten and colleagues (4) initially demonstrated that tumor cells genetically modified with HSV-tk gene were killed by GCV both in vitro and in vivo. The ability of HSV-tk-gene-transduced cells to kill nearby untransduced tumor cells (i.e., the bystander effect) has allowed the clinical application of HSV-tk-gene therapy. Recently, this effect has been shown to depend on gap-junction formation (13, 14). Moreover, our laboratory has shown that the in vivo efficacy of this approach is dependent on intratumoral release of cytokines such as tumor necrosis factor (TNF), interleukin (IL)-1, and IL-6 (15). This cytokine release within the tumor also leads to upregulation of immune regulatory molecules such as B7.1, B7.2, and intercellular adhesion molecule-1 (ICAM-1) (16, 18).

On the basis of these data, we hypothesized that a cell-based HSV-tk-gene delivery approach would be efficacious in mesothelioma cell killing, and could be applied in the treatment of MM. In this study we examined the efficacy of PA-1-STK cells in killing a human mesothelioma cell line in vitro, and investigated the effect of this killing on survival in an in vivo syngeneic mouse model of MM.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture

The murine mesothelioma cell line AC29 was provided by Dr. Bruce Robinson of the Queen Elizabeth II Medical Center, Perth, Australia (19). The ovarian cancer cell line PA-1 (ATCC CRL-1572; American Type Culture Collection, Rockville, MD) was stably transduced with a retrovirus (STK) encoding the HSV-tk gene to produce in PA-1-STK cells (12). REN cells, a human mesothelioma cell line (7), were provided by Dr. Steven Albelda of the University of Pennsylvania, Philadelphia. All cell lines were grown in Dulbecco's modified Eagle's medium (DMEM)/5% fetal calf serum (FCS) with 1% penicillin-streptomycin.

Cell-mixing Experiments

REN cells or AC29 cells were mixed with different percentages of PA-1 or PA-1-STK cells and seeded at 25,000 cells/well in 96-well plates. Cells were grown overnight and the medium was then changed to DMEM/5% FCS with or without 50 µM GCV. Cells were incubated for 4 d, washed twice with phosphate-buffered saline (PBS), and then seeded in six-well plates and fed with the same media as previously assigned (with or without GCV). Twenty-four hours later, cell colonies were counted to determine control and GCV-resistant colony counts.

Adenovirus Experiments

AdHSV-tk was provided by Dr. David Curiel (20). AdCMVLacZ was used as a control vector. The generation, propagation, and titering of recombinant adenovirus was performed as previously described (21). AC29 cells and REN cells were seeded at 25,000 cells/well in 96-well plates and grown overnight. The medium was removed and cells were infected with AdHSV-tk or AdCMVLacZ at different multiplicities of infection in serum-free medium for 30 min before changing the medium to DMEM/ 5% FCS with or without 50 µM GCV. Cells were incubated for 2 d, washed twice with PBS, and then seeded in six-well plates and fed with the same media as previously assigned. Colony counts were performed as described previously.

In Vivo Experiments

Six- to 8-wk-old male CBA/J mice were purchased from Hilltop Labs (Scottsdale, PA). Mice were acclimatized for 3 d and fed water and food ad libitum. Mice were then challenged with 10⁶ AC29 cells by intraperitoneal injection. Four or 16 d later, mice were randomized to receive intraperitoneal injections of 5 × 10⁶ PA-1 or PA-1-STK cells, or medium alone. The mice were then randomized to receive 150 mg/kg GCV by intraperitoneal injection or an equal volume of PBS twice daily for 7 d as previously described (5). Survival was recorded for up to 12 wk.

Statistical Analysis

Differences between means were determined by analysis of variance (ANOVA) followed by Fisher's protected least significant difference follow-up testing. Differences in survival were determined by contingency-table analysis followed by chi-square testing. A value of P < 0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cell-mixing Experiments

We performed cell-mixing experiments with both AC29 cells and REN cells, as described previously. A mixture of 10% PA-1-STK cells with 90% AC29 cells inhibited GCV-resistant colony formation by more than 40% (Figure 1). As the percentage of PA-1-STK cells was increased, we observed a dose-dependent inhibition in GCV-resistant colony formation. With a dose of 50% PA-1-STK cells, we observed > 90% inhibition of GCV-resistant colony formation. No inhibition of colony formation was seen in the absence of GCV (data not shown). Thus, these data support a bystander effect of PA-1-STK cells and GCV on mesothelioma cells in vitro. To test this concept further, we performed similar experiments with a human mesothelioma cell line (REN cells). We observed an enhanced bystander effect in REN cells. A dose of 50% PA-1-STK cells resulted in > 93% reduction in GCV-resistant colony formation (Figure 2). No effect was seen in experiments done with PA-1 cells lacking the HSV-tk gene (data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Cell-mixing experiments with AC29 cells. Varying percentages of PA-1-STK cells were mixed with AC29 cells, and 25,000 cells/well were seeded into 96-well plates in triplicate, followed by the addition of 50 µM GCV. Cells were harvested 4 d later and plated in six-well plates, again in the presence of 50 µM GCV. GCV-resistant colonies were counted 24 h later. The graph represents data from three experiments; values are means ± SEM.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Cell-mixing experiments with REN cells. Varying percentages of PA-1-STK cells were mixed with REN cells, and 25,000 cells/well were seeded into 96-well plates in triplicate, followed by the addition of 50 µM GCV. Cells were harvested 4 d later and plated in six-well plates, again in the presence of 50 µM GCV. GCV-resistant colonies were counted 24 h later. The graph represents data from three experiments; values are means ± SEM.

We next compared the effect of adenovirus-mediated gene transfer of the HSV-tk gene on inhibition of GCV-resistant colony formation in AC29 cells. There was a clear dose-response relationship between the multiplicity of infection (MOI) of adenovirus and inhibition of GCV-resistant colonies of AC29 cells (Figure 3). There also appeared to be a threshold effect, in that no inhibition of colony formation was observed in doses of 5 MOI or less of AdHSV-tk. Interestingly, it required an MOI of 100 to inhibit colony formation by 65%, and an MOI of 500 to see a 90% inhibition. There was no inhibition of GCV-resistant colony formation in experiments done with AdCMVLacZ (data not shown). However, x-Gal staining of these cells after AdCMVLacZ infection revealed that MOIs of 100 and 500 corresponded to 34 ± 7.2% and 78 ± 8.3% (n = 4 each experiment) cells transduced, respectively. This level of transduction resulted in > 60% and > 88% inhibition of GCV-resistant colonies, respectively. Thus, a bystander effect was observed with AdHSV-tk in vitro, but the effect was not as pronounced as that seen with PA-1-STK cells.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Effect of AdHSV-tk on AC29-cell survival. AC29 cells were seeded at a density of 25,000 cells/well and infected with varying multiplicities of AdHSV-tk or AdCMVLacZ. Twenty-four hours after infection, cells were incubated with 50 µM GCV for 4 d. Cells were harvested and seeded into six-well plates, and GCV-resistant colonies were counted 24 h later. The graph represents data from three experiments; values are means ± SEM. (AdCMVLacZ data not shown.)

In Vivo Studies of PA-1-STK Cells in Mesothelioma

We next tested PA-1-STK cells in an in vivo model of mesothelioma in immunocompetent mice. Mice were implanted with one million AC29 cells. Four days later, mice were randomized to receive intraperitoneal injections of 5 × 10⁶ PA-1 or PA-1-STK cells, or medium alone. Preliminary studies developing this model revealed an estimated doubling time of AC29 cells in vivo of 4.5 d. Thus, at the time of administration of the PA-1-STK cells, the estimated ratio was 28% AC29 cells to 72% PA-1 or PA-1-STK cells. The mice were then randomized to receive 150 mg/kg GCV or an equal volume of PBS twice daily for 7 d as previously described (5). The combination of PA-1-STK cells with in vivo GCV administration resulted in significantly enhanced survival in this model. There was a statistically significant survival benefit that was evident by 7 wk after tumor implantation and that persisted for up to 12 wk (Figure 4). A slight survival advantage was observed in animals treated with PA-1 cells + GCV at 8 wk, although this was not statistically significant. In a follow-up study, we investigated treatment with 5 × 10⁶ PA-1-STK cells in animals with a greater tumor burden. PA-1 or PA-1-STK cells were administered to mice with established tumors (16 d after administration of 10⁶ AC29 cells). Based on pilot experiments, the estimated in vivo doubling time of AC29 cells is 4.5 d. Thus, the estimated tumor burden was 12 × 10⁶ cells. Accordingly, the in vivo percentage of PA-1-STK cells was 30%. Mice were then randomized to receive GCV or PBS as a control. In this subsequent study with a lower in vivo percentage of PA-1-STK cells, we demonstrated no survival benefit, and no mice survived beyond 10 wk (data not shown).

View larger version (16K): [in this window] [in a new window]   Figure 4.   Survival of PA-1-STK/GCV-treated mice. Forty C57Bl/ 6 mice were randomized to receive AC29 cells followed by PA-1 cells or PA-1-STK cells, with or without GCV treatment. Survival was followed to 12 wk. *Significant difference (P < 0.05) in survival as determined by contingency-table analysis followed by chi-square testing.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

For several reasons, MM has been proposed as a good model for investigating gene therapy approaches: (1) it usually presents as localized disease and kills by local invasion. (2) Metastatic disease is uncommon and occurs late, if at all, in the course of MM. (3) The prognosis of the disease has remained essentially unchanged over the past 30 yr because of its resistance to chemotherapy and radiation. With this in mind, Albelda and colleagues have investigated the use of "suicide" gene therapy for MM. In in vivo preclinical studies, these investigators demonstrated efficient gene transfer into mesothelioma tissue, using a lacZ reporter gene (8), as well as dramatic reduction in tumor with an adenovirus encoding the HSV-tk gene followed by GCV treatment (6, 22). This work has led to a U.S. Food and Drug Administration (FDA)-approved Phase I protocol in humans with MM (23). Although gene transfer has been demonstrated in this Phase I human trial, it has required a high dose of vector. Moreover, fever is a common side effect, as is a substantial increase in antiadenoviral antibody titers, which may preclude secondary administration of vector. Given these potential problems with a viral vector approach, we chose to investigate the efficacy of a cell-based approach. Our studies were founded on the rationale that HSV-tk cells can render HSV-tk-gene-transduced cells susceptible to GCV toxicity both in vitro and in vivo (5, 24). Moreover, a cell-based approach with the PA-1-STK cell line has also been FDA/Recombinant DNA Advisory Committee approved for a Phase I trial in patients with advanced ovarian carcinoma (12).

In this report we demonstrate the efficacy of PA-1-STK cells in rendering both murine and human mesothelioma cells susceptible to GCV toxicity in vitro. These data are quite similar to those reported by Freeman and colleagues in both murine fibrosarcoma cell lines and human ovarian cancer in vitro (5). Therefore, PA-1-STK cells are able to confer GCV-mediated toxicity on a wide variety of tumor cells. Moreover, this is one of the first reports demonstrating efficacy of the cell-based HSV-tk approach in an in vivo model of seeded tumor cells. Many prior studies used premixing of HSV-tk⁺ and HSV-tk cells to demonstrate an antitumor effect. The mechanism of this bystander effect has been shown to be mediated partly by gap junctions in vitro (25). Moreover, there are in vivo data suggesting that cytokine release within the tumor is also responsible for this effect (15).

Our studies also show the efficacy of this cell-based approach in an immunocompetent animal model of MM. In studies in which the tumor burden was estimated to be 2 × 10⁶ cells, administration of 5 × 10⁶ PA-1-STK cells followed by GCV treatment produced 80% survival at 12 wk, compared with 10 to 20% survival in control animals. However, in a model of more established disease (2 wk after administration of 10⁶ AC29 cells) in which the estimated tumor burden was 12 × 10⁶ cells, there was no survival benefit with the use of 5 × 10⁶ PA-1-STK cells. This is probably because at the 16-d time point, this dose of PA-1-STK cells represents only 30% of the total tumor burden, and our in vitro data indicate that this dose eradicated only 70% to 75% of the tumor burden (Figure 1). Moreover, we have performed similar studies with irradiated PA-1-STK cells (which are currently being used in a human ovarian cancer trial [12]), and have achieved similar results (data not shown). Because of the lack of immunodeficient mice of the CBA/J background, the role of the immune system in the AC29 model remains unclear. We have recently investigated PA-1-STK cells in a similar model involving AB12 cells in BALB/c mice (19). Although there appears to be survival benefit in this model (data not shown), this tumor is less aggressive than AC29 (19).

A potential risk for the cell-based approach is anaphylactic reaction; however, we did not find this to be a problem despite repeated administration of allogeneic cell lines to patients in different clinical trials. Thus far, 16 patients have been treated with irradiated cells of the PA-1-STK line by intraperitoneal infusion for advanced ovarian cancer without such events. We compared in vitro cytotoxicity of AdHSV-tk and PA-1-STK cells, and found the cell-based approach to be more effective, in terms of the percentages of cells that expressed the tk gene (as estimated with a reporter gene). The maximum cytotoxicity that could be achieved was reached with an MOI of 500, and was 90% after exposure to GCV in AC29 cells, which are reportedly more resistant to adenovirus transduction than are REN cells (7). Smythe and colleagues have described maximum transduction efficiency with an MOI of 100 (7). Thus, part of the diminished effect of AdHSV-tk is probably due to poor transduction efficiency. One potential advantage of the cell-based approach is that it does not require cell transduction, which may be problematic for an adenovirus-based approach. Although adenovirus-mediated HSV-tk-gene delivery has produced impressive reductions of visible tumors, it may also be possible to achieve a similar result with larger numbers of cells through the cell-based approach. This hypothesis is supported by our in vitro and in vivo finding that the proportion of HSV-tk⁺ cells to HSV-tk cells is critical to the efficacy of the cell-based approach. Moreover, both the adenovirus approach and the cell-based approach can be limited by host immunity to the vector or the cells. Although this aspect of this therapy was not addressed in the present study, future investigations are planned to examine secondary administration of PA-1-STK cells.

In summary, we demonstrate improved in vitro efficacy and a survival benefit for animals treated with the gene-modified cancer-cell line PA-1-STK. This novel treatment modality appears to merit further development for the treatment of cancer, and our group was recently given both FDA and RAC approval to begin a Phase I trial of PA-1-STK-cell treatment in patients with MM.