Specific oncolytic activity of herpesvirus saimiri in pancreatic cancer cells

The potential use of oncolytic viruses in the treatment of cancer has been investigated for some time. A variety of agents have been studied, including some which appear to be selectively replication-competent in cancer cell lines. In this study, we have investigated the ability of herpesvirus saimiri to specifically lyse selected human cancer cell lines. Upon infection with a replication-competent virus carrying the EGFP reporter gene and a neomycin resistance marker, the pancreatic cancer lines MIAPACA and PANC-1 exhibited definite cytopathic effects. In contrast, the colonic carcinoma cell lines SW480 and HCT116 were phenotypically unaltered. In addition, stable cell lines could not be generated from PANC-1 infected cultures, in marked contrast to cultures of cells from other human tissues. Virus recovery assays demonstrated that all of the cell lines produced a small amount of virus post-infection, but that virus replication was minimal after 1 week in culture. In addition, treatment with acyclovir inhibited virus replication but paradoxically increased cytopathic effect. These data suggest that herpesvirus saimiri may have potential as an oncolytic agent for the treatment of pancreatic cancer. © 2000 Cancer Research Campaign

Replication-incompetent viruses have been extensively evaluated as vectors for potential gene delivery to tumour cells, with significant preclinical efficacy having been demonstrated by direct injection of localized tumour masses in animal models (Clayman et al, 1995;Arai et al, 1997;Sanding et al, 1997). However, a major limitation is their apparent inability to effectively treat tumour metastases (Roth and Cristiano, 1997). Selectively replicationcompetent or oncolytic viruses hold promise for the treatment of a variety of cancers. A simple protocol would involve the virus being administered intravenously, potentially exposing many cells to infection. However, the virus would only replicate in (and selectively destroy) malignant cells. Such a procedure would be ideal for the treatment of disseminated disease, the most commonly fatal form of cancer, and would not require the physical targeting of the agent. ONYX-015, for example, is an E1B gene-deleted adenovirus which selectively replicates in and lyses cancer cells expressing mutant p53 (Bischoff, 1996;Heise et al, 1997). This virus has recently shown antitumoural efficacy in one model of metastatic cancer (Heise et al, 1999).
The herpesviruses, primarily herpes simplex virus type 1 (HSV-1), are also under investigation as a potential source of oncolytic viruses (Rampling et al, 1998). We are currently developing gene therapy vectors based on herpesvirus saimiri (HVS). The natural host of this virus is the squirrel monkey (Saimiri sciureus) where it is able to persist without causing any obvious disease. However, if the virus is experimentally transmitted to other New World primates, the animals can develop a number of lymphoproliferative diseases (Fleckenstein and Desrosiers, 1982). HVS strains have been assigned to three subgroups A, B and C, depending on their transforming and pathogenic capabilities and sequence divergence of a terminal region of the viral L-DNA (Medveczky et al, 1984;. In addition, viruses from subgroup C have been shown to be capable of transforming human T-cells in vitro (Beisinger et al, 1992). For reasons of safety, we have therefore evaluated the A-strain of the virus, which is incapable of transforming human T-cells. The virus has been further modified by the removal of the saimiri transforming protein (stpA), rendering it non-pathogenic and non-transforming in all systems tested to date 1985;1986;Koomey et al, 1984).
It has been demonstrated that a selectable HVS genome has the ability to persist episomally in a wide variety of human cell lines for long periods of time, apparently without the production of infectious progeny viruses in many cases (Grassman and Fleckenstein, 1989;Simmer et al, 1991;Stevenson et al, 1999). In this report, we demonstrate that HVS is selectively cytopathic in pancreatic cancer cell lines, suggesting that it may have potential as a specific oncolytic agent for pancreatic cancer.

Cell culture
SW480 and HCT116 cells were routinely grown in RPMI 1640 and McCoy's 5A media, respectively, supplemented with 10% fetal calf serum (FCS) and the antibiotics penicillin and streptomycin. PANC-1 and MIAPACA cells were routinely grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS and the antibiotics penicillin and streptomycin. All tissue culture reagents were obtained from Gibco BRL (Paisley, Scotland). SW480, MIAPACA and PANC-1 cells have a mutation in the p53 gene, in contrast to HCT116 which possess the endogenous wt-p53 gene (Nigro et al, 1989;Redston et al, 1994).

Virus
The marker virus HVS-GFP, containing the enhanced green fluorescent protein (EGFP) and the neomycin resistance gene (NeoR), has been described previously . Summary The potential use of oncolytic viruses in the treatment of cancer has been investigated for some time. A variety of agents have been studied, including some which appear to be selectively replication-competent in cancer cell lines. In this study, we have investigated the ability of herpesvirus saimiri to specifically lyse selected human cancer cell lines. Upon infection with a replication-competent virus carrying the EGFP reporter gene and a neomycin resistance marker, the pancreatic cancer lines MIAPACA and PANC-1 exhibited definite cytopathic effects. In contrast, the colonic carcinoma cell lines SW480 and HCT116 were phenotypically unaltered. In addition, stable cell lines could not be generated from PANC-1 infected cultures, in marked contrast to cultures of cells from other human tissues. Virus recovery assays demonstrated that all of the cell lines produced a small amount of virus post-infection, but that virus replication was minimal after 1 week in culture. In addition, treatment with acyclovir inhibited virus replication but paradoxically increased cytopathic effect. These data suggest that herpesvirus saimiri may have potential as an oncolytic agent for the treatment of pancreatic cancer. © 2000 Cancer Research Campaign Keywords: pancreatic cancer; oncolytic; gene therapy; HVS Briefly, it contains the EGFP gene under the control of a CMVIE promoter and the NeoR gene under the control of an SV40 promoter, cloned between unique repetitive elements of the HVS genome. The addition of these heterologous genes into the HVS genome has no detectable effect on the growth of HVS-GFP in comparison with the wild type virus.

Infection of cells
1 × 10 6 cells were infected at a multiplicity of infection (MOI) of 1 with extracellular stocks of HVS-GFP. Infections were carried out in six-welled dishes by aspirating the culture medium and adding the appropriate amount of virus in 0.5 ml of medium. The cells were then incubated at 37°C for 90 min after which a further 1.5 ml of medium was added. Cultures were examined daily for cytopathic effects (CPE).

Drugs
Acyclovir (ACV) (Sigma, Poole, UK) was dissolved in sterile water and added to the tissue culture medium after infection with the virus. Concentrations of between 0 and 100 µM were investigated.

Virus recovery assays
Supernatant was harvested and replaced, in infected cultures at 2, 5 and 7 days post-infection. 0.5 ml of each sample was used to infect cultures of owl-monkey kidney (OMK) cells, in which HVS is capable of forming plaques. The cultures were examined after 72 h by fluorescence microscopy for the presence of clusters of green cells (indicative of the early stages of plaque formation). This method of enumeration was possible because of the EGFP expressed by the recombinant virus. Numbers of plaques were confirmed by conventional visualization after 1 week.

Cytopathic effects in cell cultures
Examination of cells at 48 h post-infection showed significant cytopathic effects in cultures of pancreatic cancer cells (MIAPACA and PANC-1). This was in marked contrast to the colonic cancer cells (SW480 and HCT116) which exhibited no obvious CPE. Fluorescence microscopy revealed that, in all cases, the vast majority of cells had been successfully infected by the virus (i.e. exhibited a green phenotype) (Figure 1).
Extended analysis revealed that the CPE in the PANC-1 culture was progressive, leading to the total destruction of the culture at 7 days post-infection (Figure 2). Although a similar CPE was noted in the MIAPACA culture a few cells seemed to escape the effect, allowing the monolayer to recover over time. The fact that these cells were green indicated that they had been successfully infected, though not destroyed, by the virus. At no time was any CPE noted in cultures of SW480 and HCT116 cells. This result is consistent with data from other colorectal lines including HT29s, the lung carcinoma cell line A549s (data not shown) and the published results on infection of other human cell lines (Simmer et al, 1991;Stevenson et al, 1999).
The addition of ACV to infected and control cultures of MIAPACA, SW480 and HCT116 had no apparent effect, whereas in the case of infected PANC-1 cultures, increasing concentrations appeared to accelerate the process of monolayer destruction (data not shown).

Virus recovery assay
The results of the virus recovery assay are shown in Figure 3. High numbers of green foci were produced in the OMK monolayers by all of the 2-day infected culture supernatants. However, this was The presence of increasing concentrations of the drug reduced this figure to 18 gfu ml -1 . 20 gfu ml -1 were found in the SW480 culture at the highest concentration (100 µM) of ACV, representing a less significant reduction overall. Lower numbers of gfu ml -1 were detected in the PANC-1 culture and minimal virus was detected in the HCT116 culture at 5 days post-infection. In the case of PANC-1, the presence of 100 µM ACV resulted in no virus being recovered. As expected, virus was not recovered from the 'no-cell' control experiment after 5 days, owing to its instability at 37°C.
At 7 days post-infection, the amount of recovered virus was reduced in every case. No virus was detected in the HCT116 cultures, which remained green confirming persistence of EGFP expression. The presence of ACV prevented any virus being recovered from the PANC-1 cultures. Even in its absence only 2 gfu ml -1 were detected. Lesser amounts of virus could still be recovered from the MIAPACA and SW480 cultures, but the presence of 100 µM ACV reduced this to very low levels in MIAPACA (2 gfu ml -1 ) and to zero in SW480 cultures.

DISCUSSION
The use of a virus as an oncolytic agent for the treatment of cancer is normally dependent on it being selectively replicationcompetent in the tumour cell, to prevent the destruction of healthy tissue. We are currently attempting to develop vectors for the treatment of colorectal and pancreatic cancers based on HVS, a virus which is known to be capable of infecting and persisting within many different human cancer cell lines without apparently causing CPE (Grassman and Fleckenstein, 1989;Simmer et al, 1991;Stevenson et al, 1999). In this study, we have investigated the interactions of an HVS vector expressing EGFP with cell lines representative of the above disease targets. The virus was able to efficiently infect all of the cell lines and, in three out of four instances, stably infected lines were established in which all of the cells expressed the marker protein EGFP. This finding was anticipated. However, surprisingly, the virus caused significant CPE and cell death in the pancreatic cancer cell lines (MIAPACA and PANC-1). This was so severe in the case of PANC-1 that the cell sheet was completely destroyed. Extensive destruction of the MIAPACA cells also occurred. In contrast, the colonic cancer cell lines were apparently unaffected by the infection procedure and exhibited no CPE.
In order to quantitatively assess the amount of virus, if any, produced by these infected cultures, a virus recovery assay was performed. The level of virus production after 5 days was greatest in the pancreatic cell line MIAPACA. However, the destruction of the PANC-1 monolayer at earlier times post-infection suggested that significant virus production may well have occurred prior to the 5-day sample. Unfortunately, as previously discussed, virus production at early stages could not be distinguished from input virus. According to control experiments, input virus was not totally inactivated until 5 days post-infection. Considering the relatively low levels of virus that were produced in any of the cultures it seems possible that the CPE observed in the pancreatic cell lines is not directly linked to the levels of virus production and may instead be caused by the expression of a virus gene product to which the cells are particularly sensitive. The expression of herpes simplex virus type 1 immediate early genes, for example, is known to cause CPE in a variety of human cells (Johnson et al 1994).
One of the advantages of using a herpesvirus-based vector in a clinical setting is that replication of the virus can be controlled by using the drug ACV (Elion, 1980). Our studies have shown that ACV has activity against HVS and could therefore be used as a safeguard in any future clinical investigation. In addition, ACV appeared to accelerate the destruction of the PANC-1 culture. This observation may be due to its activation in cells which were expressing the HVS thymidine kinase gene product, thereby contributing to the death of the cell. A similar phenomenon is widely exploited by 'suicide' gene therapy strategies, involving the use of the HSV-TK gene product (Culver et al, 1992).
In summary, we have shown that HVS appears to be selectively cytopathic in pancreatic tumour cells. This unexpected result suggests its possible application as an oncolytic agent for the treatment of pancreatic cancer. If the virus, in its present replicationcompetent form, was inoculated directly into a tumour it would be unlikely to spread far from the injection site owing to the low levels of virus production from infected human cells. We also note that accidental inoculations into human subjects of the most potently transforming strain of HVS have not resulted in seroconversion of the individuals involved or any recognizable disease (Fleckenstein and Desrosiers, 1982). The strain of virus used in these studies has already been attenuated, being incapable of transforming any known cell type examined to date or causing pathology in any animal model so far tested. HVS therefore merits further evaluation with a view to possible exploitation of these